Split (1)

train · ~9.11M rows (showing the first 93.5k)

identifier stringlengths 11 32	pdf_url stringlengths 17 4.62k ⌀	lang stringclasses 120 values	error stringclasses 1 value	title stringlengths 2 500 ⌀	source_name stringlengths 1 435 ⌀	publication_year float64 1.9k 2.02k	license stringclasses 3 values	word_count int64 0 1.64M	text stringlengths 1 9.75M
https://openalex.org/W2007750930	https://discovery.dundee.ac.uk/ws/files/8596405/An_optimized_method.pdf	English	null	An optimized method for the extraction of bacterial mRNA from plant roots infected with Escherichia coli O157:H7	Frontiers in microbiology	2,014	cc-by	7,146	Citation for published version (APA): Holmes, A., Birse, L., Jackson, R. W., & Holden, N. J. (2014). An optimized method for the extraction of bacterial mRNA from plant roots infected with Escherichia coli O157:H7. Frontiers in Microbiology, 5, Article 286. https://doi.org/10.3389/fmicb.2014.00286 University of Dundee An optimized method for the extraction of bacterial mRNA from plant roots infected with Escherichia coli O157:H7 Holmes, Ashleigh; Birse, Louise; Jackson, Robert W.; Holden, Nicola J. University of Dundee Correspondence: p Nicola J. Holden, Cell and Molecular Sciences, The James Hutton Institute, Invergowrie, Dundee, DD2 5DA, UK e-mail: nicola.holden@hutton.ac.uk INTRODUCTION for bacterial adaptation. In addition to challenges with bacterial mRNA stability, extractions from mixed samples bring additional considerations, not least for inhibitory compounds that may affect downstream analysis. Although there are a number of published techniques for the isolation of total RNA from bacteria-infected plant leaves (Schenk et al., 2008; Soto-Suarez et al., 2010; Fink et al., 2012; Goudeau et al., 2013), samples can still be dominated by plant RNA making it challenging to assess bacterial mRNA. In published reports, leaves were typically infected via inﬁltration, introducing a high bacterial inoculum into the sample. In other studies, total RNA was extracted from inoculated plant extracts (Market al.,2005; Hernandez-Moraleset al.,2009; Kyleet al.,2010; Shidore et al., 2012), such as leaf lysates, which helped to reduce the dominance of plant RNA in the sample. The analysis of bacterial gene expression is important in the deter- mination of how adaptation to different environments develops and informs on the roles of different genes during this process. Human pathogens are now recognized to interact with plants and use them as hosts, as a result of recent high-proﬁle outbreaks from contaminated fruit and vegetables (Cooley et al., 2007; Buchholz et al., 2011). Adaptation of food-borne pathogens to secondary hosts has opened up new areas of research and investigation. Cur- rent research suggests that the interaction of human pathogens is more complex than previously perceived in that they can persist for long periods of time (reviewed in Brandl, 2006; Holden et al., 2009; Barak and Schroeder,2012) and invoke an immune response from the plant (Thilmony et al., 2006; Schikora et al., 2008; Roy et al., 2013). Furthermore, the interactions include quite speciﬁc and targeted recognition of the plant host cells by the bacteria (Rossez et al., 2013). Important questions remain as to how the bacteria adapt to secondary hosts, for which analysis of bacterial gene expression is fundamental. There are limited studies where bacterial expression has been assessed directly from the plant root system (roots and rhizo- sphere). However, roots are known to support relatively high densities of bacteria, providing a more favorable habitat than the phyllosphere. METHODS ARTICLE published: 10 June 2014 doi: 10.3389/fmicb.2014.00286 METHODS ARTICLE published: 10 June 2014 doi: 10.3389/fmicb.2014.00286 An optimized method for the extraction of bacterial mRNA from plant roots infected with Escherichia coli O157:H7 Ashleigh Holmes1, Louise Birse1, Robert W. Jackson 2 and Nicola J. Holden1 Ashleigh Holmes1, Louise Birse1, Robert W. Jackson 2 and Nicola J. Holden1* 1 Cell and Molecular Sciences The James Hutton Institute Invergowrie Dundee UK 1 Cell and Molecular Sciences, The James Hutton Institute, Invergowrie, Dundee, UK 2 School of Biological Sciences, The University of Reading, Knight Building, Whiteknights, Reading, UK Analysis of microbial gene expression during host colonization provides valuable informa- tion on the nature of interaction, beneﬁcial or pathogenic, and the adaptive processes involved. Isolation of bacterial mRNA for in planta analysis can be challenging where host nucleic acid may dominate the preparation, or inhibitory compounds affect downstream analysis, e.g., quantitative reverse transcriptase PCR (qPCR), microarray, or RNA-seq. The goal of this work was to optimize the isolation of bacterial mRNA of food-borne pathogens from living plants. Reported methods for recovery of phytopathogen-infected plant material, using hot phenol extraction and high concentration of bacterial inoculation or large amounts of infected tissues, were found to be inappropriate for plant roots inoculated with Escherichia coli O157:H7. The bacterial RNA yields were too low and increased plant material resulted in a dominance of plant RNA in the sample.To improve the yield of bacterial RNA and reduce the number of plants required, an optimized method was developed which combines bead beating with directed bacterial lysis using SDS and lysozyme. Inhibitory plant compounds, such as phenolics and polysaccharides, were counteracted with the addition of high-molecular-weight polyethylene glycol and hexadecyltrimethyl ammonium bromide. The new method increased the total yield of bacterial mRNA substantially and allowed assessment of gene expression by qPCR. This method can be applied to other bacterial species associated with plant roots, and also in the wider context of food safety. Keywords: food-borne pathogens, lettuce, spinach, rhizosphere, mRNA isolation www.frontiersin.org An optimized method for the extraction of bacterial mRNA from plant roots infected with Escherichia coli O157:H7 Holmes, Ashleigh; Birse, Louise; Jackson, Robert W.; Holden, Nicola J. An optimized method for the extraction of bacterial mRNA from plant roots infected with Escherichia coli O157:H7 Holmes, Ashleigh; Birse, Louise; Jackson, Robert W.; Holden, Nicola J. Document Version Publisher's PDF, also known as Version of record Document Version Publisher's PDF, also known as Version of record Link to publication in Discovery Research Portal Citation for published version (APA): Holmes, A., Birse, L., Jackson, R. W., & Holden, N. J. (2014). An optimized method for the extraction of bacterial mRNA from plant roots infected with Escherichia coli O157:H7. Frontiers in Microbiology, 5, Article 286. https://doi.org/10.3389/fmicb.2014.00286 General rights Copyright and moral rights for the publications made accessible in Discovery Research Portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. Take down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Download date: 24. Oct. 2024 Download date: 24. Oct. 2024 METHODS ARTICLE published: 10 June 2014 doi: 10.3389/fmicb.2014.00286 Adam Schikora, IPAZ, JLU Giessen, Germany Germany Reviewed by: Saul Burdman, The Hebrew University of Jerusalem, Israel Thamarai Schneiders, Queen’s University Belfast, UK Correspondence: Nicola J. Holden, Cell and Molecular Sciences, The James Hutton Institute, Invergowrie, Dundee, DD2 5DA, UK e-mail: nicola.holden@hutton.ac.uk Edited by: Edited by: Adam Schikora, IPAZ, JLU Giessen, Germany Adam Schikora, IPAZ, JLU Giessen, Germany Reviewed by: y Saul Burdman, The Hebrew University of Jerusalem, Israel Thamarai Schneiders, Queen’s University Belfast, UK PLANT PROPAGATION AND INFECTION Lettuce (Lactuca sativa) cultivar All Year Round or spinach (Spinacia oleracea) cultivar Amazon seeds (Sutton Seeds, UK) were soaked in sterile distilled water for 2 h before being sur- face sterilized in 2% calcium hypochlorite solution (10 ml) for 10 min. The seeds were then washed vigorously six times with sterile distilled water and germinated on distilled water agar (0.5% w/v) in the dark for 3–5 days, at ∼22◦C. Seedlings were transplanted into 175 ml hydroponic tubes (Greiner, Fricken- hausen, Germany) containing autoclaved perlite and sterile 0.5× Murashige and Skoog (MS) medium (Sigma Aldrich, USA). Seedlings were grown in a cabinet with a light intensity of 150 μmol m2 s−1 (16 h photoperiod) for a further 21 days at 22◦C. To assess bacterial numbers from infected plant material, the roots were washed with PBS to remove excess, non-adherent bacteria. The root sample was homogenized using mortar and pestle and serially diluted for viable counts on selective agar plates. Correspondence: In general, published reports describe that a high number of plant roots is required to retrieve sufﬁcient bacte- rial RNA (Matilla et al., 2007; Hou et al., 2012, 2013; Zy´sko et al., 2012). Further to this, the application of techniques described for RNA puriﬁcation from infected leaves cannot always be suc- cessfully applied to roots. Therefore, optimization of the RNA extraction methods is required to obtain sufﬁcient quantity and qualityof bacterialmRNA,coupledwithareductionintheamount of accompanying plant root RNA. We optimized the method for Bacterial gene expression is typically assessed using quantitative reverse transcriptase PCR (qPCR) or microarray techniques, and more recently RNA-seq (An et al., 2013), which require the isola- tion of high quality and quantity of mRNA. Bacterial mRNA has a short half-life and is less stable than eukaryotic mRNA transcripts, which have capped and polyadenylated RNA tails. Therefore, appropriate measures need to be in place to capture mRNA tran- scriptsthatmaybeinherentlyunstable,butneverthelessimportant June 2014 \| Volume 5 \| Article 286 \| 1 E. coli O157:H7 RNA extraction from plants Holmes et al. loosely attached bacteria. The root sample was placed into a foil packet and immediately stored in liquid nitrogen until all samples had been processed. the food-borne pathogen Escherichia coli O157:H7, frequently associated food-borne outbreaks from consumption of contam- inated spinach and lettuce. Although roots of these plants are not consumed, the pathogen can colonize this habitat successfully (Wright et al., 2013), from where it can contaminate the edible portion, either directly or through cross-contamination during processing. Bead/SDS/phenol method for extraction The optimized method is represented in Figure 1. The samples were removed from the liquid nitrogen, beaten with a spatula to break up the root into smaller pieces, and transferred into an 1.5 ml micro-centrifuge tube (DNase, RNase free: Ambion, Austin, USA) preloaded with ∼250 mg mixture of 1 mm glass and 0.1 mm silica beads (ThermoScientiﬁc Ltd., Waltham, USA), from which the weight was determined. Cooled, sterile equip- ment was used throughout the process. The micro-centrifuge tube was then returned to liquid nitrogen for processing subse- quent samples until all of the samples were collected. For the lysis step, 800 μl of Tris-EDTA (TE) buffer (10 mM Tris-HCl; 1 mM EDTA) supplemented with 500 μg ml−1 lysozyme, 0.1% SDS, and 100 mM β-mercaptoethanol was added to each root sample. The tubes were placed into TissueLyser (Qiagen, Lim- burg, Netherlands) and agitated for 30 s with a 30 s interval on ice for three cycles. After the last cycle, the samples were MATERIALS AND METHODS BACTERIAL STRAINS AND GROWTH CONDITIONS BACTERIAL STRAINS AND GROWTH CONDITIONS E. coli O157:H7 isolate Sakai stx-negative (Hayashi et al.,2001) was routinely cultured overnight in Luria broth at 37◦C, with aeration. For plant-infection assays, the bacteria were sub-cultured at 1:50 dilution into 15 ml rich-deﬁned MOPs media supplemented with 0.2% glucose (RD MOPS glucose; Neidhardt et al., 1974), in a 200 ml Erlenmeyer ﬂask and incubated at 18◦C, with aeration for ∼18 h. Bacterial cultures were diluted to OD600 of 0.2 (equivalent to ∼2 × 108 cfu ml−1) in sterile phosphate buffered saline (PBS) prior to infecting plant roots. RNA EXTRACTION AND PURIFICATION Total RNA was extracted from infected root samples either by the BPEX method (Schenk et al., 2008) or by the Bead/SDS/phenol method (Figure 1): a method which was adapted and optimized from Jahn et al. (2008) and Schenk et al. (2008). BPEX-Schenk method for extraction The BPEX method was originally developed to recover Pseu- domonas syringae phytopathogen mRNA from infected plant material (Schenk et al., 2008), and tested for recovery of E. coli mRNA from infected roots. The main change from the published method was in the tissue type (root vs leaf disks), although the protocol is provided to allow a direct comparison with the opti- mized method. Infected plant tissue was ground to a ﬁne powder in liquid nitrogen with a pre-cooled pestle and mortar. The sam- ple (∼1 g) was transferred to an Eppendorf containing 750 μl of bacteria plant extraction (BPEX) buffer [0.35 M glycine, 0.7 M NaCl, 2% (w/v) polyethylene glycol (PEG) 20000, 40 mM EDTA, 50 mM NaOH, 4% (w/v) SDS] supplemented with 100 mM β- mercaptoethanol just prior to use. The sample was mixed on a vortex mixer in the buffer prior to incubation at 95◦C for 90 s with shaking. An equal volume (750 μl) of phenol/chloroform mix (5:1, pH 4.0) (Sigma, St. Louis, USA) was added and the sample shaken for 5 min to form an emulsion before centrifugation at 16,000 × g for 7 min. The upper phase was collected and added to an equal volume of phenol/chloroform mix (5:1, pH 4.0), shaken, and cen- trifuged again as the previous step. The upper phase was collected and added to an equal volume of phenol/chloroform/isoamyl alcohol (IAA) mix (25:24:1, pH 4.0) (Sigma, St. Louis, USA) shaken and centrifuged again as before. A volume of 495 μl of the upper phase was added to 550 μl chloroform/IAA mix (24:1) and overlayed with 55 μl pre-warmed (55◦C) hexadecyltrimethyl ammonium bromide (CTAB)/NaCl (10% CTAB, 0.7 M NaCl) solution. The suspension was shaken and centrifuged as before. The upper phase was added to 1/4 vol 8 M LiCl solution, mixed by inversion and the RNA precipitated at –20◦C for 30 min. To collect the RNA, the sample was centrifuged at 16,000 × g for 20 min at 4◦C and the RNA pellet resuspended in 100 μl RNase-free water. The RNA was cleaned and DNase treated using RNeasy Plant Mini kit (Qiagen, Limburg, Netherlands) as per the manufacturer’s guidelines. Frontiers in Microbiology \| Plant-Microbe Interaction Sample preparation Sample preparation was common to both methods. Whole plants were gently removed from the hydroponic tubes and washed with sterile PBS to remove as much perlite as possible before bacte- rial infection. Plants were pooled into groups of 5 and the roots were submerged into 20–25 ml bacterial suspension (OD600 = 0.2 equivalent to ∼5 × 107–1 × 108 cfu ml−1), and incubated at 18◦C for 2 h. After incubation, the plants were removed from the bacte- rial suspension and all the tissue above the crown (shoots, leaves, and petioles) were aseptically removed with a sterile scalpel and the roots immediately immersed into 20 ml ice-cold, 95% ethanol: 5% phenol (pH 4.0). The sample was incubated on ice for 5 min and gently agitated on a vortex mixer to remove any excess and June 2014 \| Volume 5 \| Article 286 \| 2 Frontiers in Microbiology \| Plant-Microbe Interaction E. coli O157:H7 RNA extraction from plants Holmes et al. FIGURE 1 \| Bead/SDS/phenol method of RNA extraction from infected plant roots. A schematic of the optimized method (described in detail within Section “Material and Methods”). The ﬁnal concentrations of reagents are indicated in the ﬁgure. Colored text has been used to highlight different stages and treatments (e.g., blue for addition of beads/cold; green for lysis; red for shaking/heat). FIGURE 1 \| Bead/SDS/phenol method of RNA extraction from infected plant roots. A schematic of the optimized method (described in detail within Section “Material and Methods”). The ﬁnal concentrations of reagents are indicated in the ﬁgure. Colored text has been used to highlight different stages and treatments (e.g., blue for addition of beads/cold; green for lysis; red for shaking/heat). indicated in the ﬁgure. Colored text has been used to highlight different stages and treatments (e.g., blue for addition of beads/cold; green for lysis; red for shaking/heat). FIGURE 1 \| Bead/SDS/phenol method of RNA extraction from infected plant roots. A schematic of the optimized method (described in detail within Section “Material and Methods”). The ﬁnal concentrations of reagents are returned to ice before transfer to a heatblock set at 64◦C for 2 min. Sample preparation To extract nucleic acids, the supernatant was collected and pooled after two centrifugation steps: ﬁrst at 100 × g for 1 min to pellet the beads and large fragments of root, followed by a second at 8,600 × g for 2 min to compact the debris fur- ther, yielding more supernatant. One hundred millimolar sodium acetate (pH 5.2) and 2% (w/v) PEG 20000 was added to the supernatant and inverted to mix. An equal volume (∼1 ml) of phenol (pH 4.0) was then added, mixed by inversion, and the sample incubated at 64◦C for 6 min, with the tubes inverted every 40 s. The sample was transferred to an ice bath for 2 min before centrifugation at 16,000 × g for 10 min at 4◦C. The upper aqueous layer was added to an equal volume of chloroform/IAA mix (24:1) and 1/20 volume of pre-warmed (55◦C) CTAB/NaCl solution in a fresh micro-centrifuge tube. The sample was mixed by inversion and then centrifuged at 16,000 × g for 5 min at 4◦C. The upper aqueous layer was added to 1/4 vol 8 M LiCl, mixed by inversion, and then incubated at –20◦C for 20 min to overnight to precipitate the nucleic acid. The nucleic acid was recovered by centrifugation at 16,000 × g for 30 min at 4◦C. The pellet was resuspended in 100 μl RNase-free water and the sample cleaned and DNase treated using the RNeasy Plant Mini kit (Qiagen, Limburg, Netherlands) as per the manufacturer’s instructions. mixed by inversion, and then incubated at –20◦C for 20 min to overnight to precipitate the nucleic acid. The nucleic acid was recovered by centrifugation at 16,000 × g for 30 min at 4◦C. The pellet was resuspended in 100 μl RNase-free water and the sample cleaned and DNase treated using the RNeasy Plant Mini kit (Qiagen, Limburg, Netherlands) as per the manufacturer’s instructions. Total RNA concentration was determined using a Nan- oDrop (Wilmington, USA) spectrophotometer and the relative proportions of ribosomal RNA determined using a BioAnal- yser 2100 (Agilent Technologies, Santa Clara, USA), for both methods. EXTRACTION OF TOTAL RNA FROM PLANT ROOTS INFECTED WITH BACTERIA The samples were run alongside a commercial RNA transcript ladder (0.2, 0.5, 1.0, 2.0, 4.0, and 6.0 kb, Agilent); the 16S and 23S (bacterial), and 18S and 28S (plant) rRNA bands are indicated. Electropherogram of E. coli O157:H7 Sakai infected lettuce roots extracted with (B) the BPEX method (sample derived from lane 1) or with (C) the novel Bead/SDS/phenol method (sample derived from lane 4). The uninfected spinach control was similar to lettuce (lane L) (not shown). FIGURE 2 \|Total RNA from plant root extractions by BPEX and Bead/SDS/phenol methods. Intact plant roots were suspended in 20 ml of bacterial inoculum (OD600 of 0.2) for 2 h at 18◦C. Total RNA was extracted using either the BPEX or the Bead/SDS/phenol method and RNA quality was assessed by spectroscopy on a Bioanalyser 2100 machine (Agilent). (A) A montage of an electropherograph of the total RNA levels following extraction of inoculated plant material. The lanes show RNA from an in vitro culture of E. coli O157:H7 isolate Sakai only (lane S); uninfected lettuce roots (lane L); E. coli O157:H7 Sakai infected lettuce roots (Lane 1, 306 ng μl−1; Lane 4, 119 ng μl−1; Lane 5, 195 ng μl−1); and E. coli O157:H7 Sakai infected spinach roots (Lane 2, 248 ng μl−1; Lane 3, 301 ng μl−1; Lane 6, 141 ng μl−1). Lanes 1, 2, and 3 show extraction using the BPEX method, and Lanes 4, 5, and 6 using the Bead method. The samples were run alongside a commercial RNA transcript ladder (0.2, 0.5, 1.0, 2.0, 4.0, and 6.0 kb, Agilent); the 16S and 23S (bacterial), and 18S and 28S (plant) rRNA bands are indicated. Electropherogram of E. coli O157:H7 Sakai infected lettuce roots extracted with (B) the BPEX method (sample derived from lane 1) or with (C) the novel Bead/SDS/phenol method (sample derived from lane 4). The uninfected spinach control was similar to lettuce (lane L) (not shown). Use of the BPEX method typically yielded 3.6–9.6 μg of total RNA from 250 mg E. coli O157:H7-infected lettuce roots. Anal- ysis of the total RNA showed strong bands corresponding with 18S and 28S rRNA of lettuce (Figure 2A: Lanes 1, 2, 3) with only very faint bacterial RNA corresponding to 16S and 23S rRNA in mixed samples (Figure 2A; Lanes 1, 2, 3). Examination of the elec- tropherogram trace showed that the bacterial rRNA peaks were considerablysmallerthantheplantrRNApeaks,e.g.,BPEXsample 1 (Figure 2B). cDNA SYNTHESIS AND qPCR CONDITIONS cDNA was transcribed from 1 μg total RNA using Superscript II (Invitrogen, Carlsbad, USA) following the random primer proto- col. Amixtureof ten11-meroligonucleotideprimers(Ea1 ±Ea10) June 2014 \| Volume 5 \| Article 286 \| 3 www.frontiersin.org E. coli O157:H7 RNA extraction from plants Holmes et al. FIGURE 2 \|Total RNA from plant root extractions by BPEX and Bead/SDS/phenol methods. Intact plant roots were suspended in 20 ml of bacterial inoculum (OD600 of 0.2) for 2 h at 18◦C. Total RNA was extracted using either the BPEX or the Bead/SDS/phenol method and RNA at 100 nM (Fislage et al., 1997) designed speciﬁcally for Enter- obacteriaceae mRNA was used. Quantitative reverse transcriptase PCR was carried out using speciﬁc primers for E. coli O157:H7 gyrB housekeeping gene (gyrB.RT.F = CATCAGAGAGGTCG- GCTTCC; gyrB.RT.R = CATGGAGCGTCGTTATCCGA) using StepOnePlusTM real-time PCR system (Applied Biosystems, Life Technologies, Carlsbad, USA) and iTaqTM Universal SYBR® Green Supermix (Bio-Rad, Hercules, USA). Each 20 μl reaction vol- ume was composed of iTaq SuperMix, 300 nM of forward and reverse primers and 1 μl of cDNA (diluted 1:4). The PCR pro- gram consisted of an initial denaturation at 95◦C for 10 min, followed by 40 cycles of denaturation at 95◦C for 15 s, and anneal- ing and extension at 60◦C for 1 min. Melt curve analysis was also performed with an initial denaturation at 95◦C for 15 s, fol- lowed by annealing at 60◦C for 1 min with an 0.3◦C increase (step and hold) and ﬁnal step of 95◦C for 15 s. Data were col- lected from three technical replicates and from two biological replicates. Frontiers in Microbiology \| Plant-Microbe Interaction EXTRACTION OF TOTAL RNA FROM PLANT ROOTS INFECTED WITH BACTERIA An optimized method was previously reported for the extraction of P. syringae mRNA from inﬁltrated plant leaves (Schenk et al., 2008). A buffer system was developed that yielded 75–125 μg of total RNA per 150–200 mg sample, from the equivalent of 20 inﬁltrated leaf disks (7 mm). We tested this protocol (termed the “BPEX” method) to determine if it could also be used to recover mRNA of a food-borne pathogen from infected roots. Our work focuses on the bacterial genes induced during the stages of colo- nization of lettuce and spinach roots, and as such the roots in our experiments are not infected via inﬁltration; rather the bacteria colonize the outer surface. This necessitated processing the entire root to recover sufﬁcient RNA. Also, to obtain equivalent bacterial numbers to those reported, where each leaf disk was inﬁltrated with 1 × 108 cfu bacteria, at least 20 root samples were pooled for each extraction. In this method, the samples were ground to a ﬁne powder in liquid nitrogen, prior to incubation in an extraction buffer (BPEX) coupled with cell lysis. RNA was puri- ﬁed using traditional acidic phenol extraction and LiCl-mediated precipitation. FIGURE 2 \|Total RNA from plant root extractions by BPEX and B d/SDS/ h l h d I l d d i FIGURE 2 \|Total RNA from plant root extractions by BPEX and Bead/SDS/phenol methods. Intact plant roots were suspended in 2 FIGURE 2 \|Total RNA from plant root extractions by BPEX and Bead/SDS/phenol methods. Intact plant roots were suspended in 20 ml of bacterial inoculum (OD600 of 0.2) for 2 h at 18◦C. Total RNA was extracted using either the BPEX or the Bead/SDS/phenol method and RNA quality was assessed by spectroscopy on a Bioanalyser 2100 machine (Agilent). (A) A montage of an electropherograph of the total RNA levels following extraction of inoculated plant material. The lanes show RNA from an in vitro culture of E. coli O157:H7 isolate Sakai only (lane S); uninfected lettuce roots (lane L); E. coli O157:H7 Sakai infected lettuce roots (Lane 1, 306 ng μl−1; Lane 4, 119 ng μl−1; Lane 5, 195 ng μl−1); and E. coli O157:H7 Sakai infected spinach roots (Lane 2, 248 ng μl−1; Lane 3, 301 ng μl−1; Lane 6, 141 ng μl−1). Lanes 1, 2, and 3 show extraction using the BPEX method, and Lanes 4, 5, and 6 using the Bead method. DEVELOPMENT OF THE BEAD/SDS/PHENOL PROTOCOL Physical disruption of tissues and cells using inert beads has been used for nucleic acid extraction from ﬁlamentous fungi (Leite et al., 2012), microalgae (Kim et al., 2012), soil and sludge samples (Grifﬁths et al., 2000), and in commercial kits. We postulated that this method of disruption may also aid in bacterial RNA extrac- tion and used a combination of 1 mm glass beads and 0.1 mm silica beads for sample lysis. Since tissue can become over-heated during agitation, which leads to nucleic acid degradation, the sam- ples were “rested” for 30 s on ice between the three treatment cycles. y Bacterial cells were speciﬁcally targeted for lysis in an attempt to reduce plant-derived nucleic acid contamination, with the inclusion of lysozyme in the extraction buffer. However, a com- bination of the hot SDS/phenol extraction protocol described previously (Jahn et al., 2008) with the bead beating step resulted in very low RNA yields; not greater than 0.5 μg. It was possible that contaminants or inhibitors from the plant material, such as polysaccharides, phenolic compounds, and secondary metabolites deleteriously affected RNA recovery. Therefore, the protocol was modiﬁed to include additional steps after the initial lysis and the incorporation of PEG and CTAB/NaCl treatments, followed by LiCl precipitation (Figure 1). The inclusion of high-molecular- weight PEG in the extraction protocol promotes the removal of polysaccharides and phenolic compounds from plant tissues that bind to or co-precipitate with RNA (Gehrig et al., 2000). CTAB is a detergent that acts to separate polysaccharides from nucleic acids (Chang et al., 1993; Jaakola et al., 2001). Lithium chloride is inef- ﬁcient at precipitating DNA, proteins, or carbohydrates, and thus improves RNA yield and purity compared to other nucleic acid precipitation methods, i.e., ethanol and sodium acetate (Barlow et al., 1963; Cathala et al., 1983). EXTRACTION OF TOTAL RNA FROM PLANT ROOTS INFECTED WITH BACTERIA The pre-dominance of plant rRNA in the BPEX- prepared samples indicated that plant mRNA would also dominate over bacterial transcripts. Therefore, to increase the levels of bac- terial mRNA with a concomitant decrease in plant-derived RNA, a modiﬁed method was designed. The aim was threefold: (i) to reduce the number of plants required per extraction; (ii) to reduce the carry-over of plant tissue in the sample; and (iii) incorporate a lysis step which targeted bacterial cells. Frontiers in Microbiology \| Plant-Microbe Interaction June 2014 \| Volume 5 \| Article 286 \| 4 E. coli O157:H7 RNA extraction from plants Holmes et al. Table 1 \| Comparison of Ct values from E. coli O157:H7 isolate Sakai for an in vitro culture and for infected plant roots. Sample Average Ct ± SD Sakai culture 23.435 ± 0.415 Sakai + lettuce roots 23.525 ± 0.295 Sakai + spinach roots 26.997 ± 0.527 In brief, plant roots were suspended in 20 ml bacterial inoculum (OD600 of 0.2) for 2 h at 18◦C, washed vigorously, and RNA prepared. The bacteria-only culture was suspended in 20 ml 1× PBS for 2 h at 18◦C. The gyrB gene was quantiﬁed using a StepOnePlus PCR machine (Applied Biosystems). The values show the average of six samples from two independent experiments. Table 1 \| Comparison of Ct values from E. coli O157:H7 isolate Sakai for an in vitro culture and for infected plant roots. DISCUSSION The isolation of total RNA from plant material is well deﬁned for leaves, but less so for studies involving the roots or rhizosphere. The majority of reports focus on analysis of gene expression for phytopathogens and the number of examples describing the extraction of food-borne pathogen RNA from plants is limited. Here, we describe optimization of a method that resulted in high- quality bacterial mRNA from the infected roots of fresh produce plants: lettuce and spinach. QUANTIFICATION OF BACTERIA Toquantifythebacteriapresentinthesamples,theaveragenumber of E. coli O157:H7 (Sakai) cells present in a 250 mg root sample (typicallyfrom7to8roots)wasdetermined. Theexperimentalset- up was as described for RNA extraction, although the roots were washed with PBS instead of the initial step of mRNA preservation in the ice-cold phenol/ethanol mixture. On average, 5.6 × 107 cfu of E. coli O157:H7 (Sakai) were present in each lettuce sample. The values for spinach roots were similar, which represent between 10 and 15% of the initial inoculum used for infection. These ﬁgures are equivalent to the limit of bacterial gene detection reported by Schenk et al. (2008) of ∼5 107 cfu/sample derived from 20 leaf disks. ANALYSIS OF RNA YIELDS A combination of physical disruption using beads together with the chemical treatments (Figure 1) resulted in total RNA con- centrations that averaged 5.1 μg (range from 1.5 to 8.9 μg) from 250 mg of infected root tissue. Assessment of the RNA showed a substantial increase in bacterial-derived 16S and 23S rRNA in bead-treated samples compared to BPEX samples (Figure 2A – compare lanes 4–6 with lanes 1–3). The optimized method increased the bacterial rRNA to levels that were equivalent or close to plant rRNA (Figure 2C). Reported investigations of gene expression of enteric bacteria, such as Salmonella enterica (Goudeau et al., 2013), E. coli K-12 and E. coli O157:H7 (Kyle et al., 2010; Fink et al., 2012) used large quantities of plant material, e.g., 100 g of leaf tissue coupled with high concentrations of bacteria, approx. 108 cfu per sam- ple. Samples were processed to separate bacteria from plant tissue using physical methods, such as a Stomacher (Kyle et al., 2010; Goudeau et al., 2013) or by shaking (Fink et al., 2012) followed by ﬁltration to remove plant debris. RNA was prepared either using commercial kits (Kyle et al., 2010; Goudeau et al., 2013) or a hot-phenol method (Fink et al., 2012). Some of these studies investigated gene expression on post-harvest material, using pre- packaged commercial leaves rather than propagating plants, where it is possible to obtain the large amount of material required. In contrast, investigation of bacteria gene expression on living plant roots presents challenges in obtaining similar weights of material. A hydroponic system for high-throughput propagation of lettuce plants was described recently (Hou et al., 2012, 2013), where ster- ilized seeds were germinated in 96-well pipette tip boxes. Whole transcriptome analysis of E. coli was examined three days after QUANTITATIVE REVERSE TRANSCRIPTASE PCR (qPCR) ANALYSIS Biol. Rep. 11, 113–116. doi: 10.1007/BF02670468 The optimized method uses a combination of bead-beating, SDS lysis, phenol extraction, and CTAB puriﬁcation to extract high-quality bacterial RNA from as little as 250–300 mg infected roots. Optimization has been carried out for two fresh produce plant species that have previously been associated with large-scale outbreaks of food-borne pathogens. It was interesting to note that the efﬁciency of recovery was lower for spinach root compared to lettuce, although the numbers of bacteria recovered from the plant tissue were similar. Since E. coli O157:H7 Sakai adheres to lettuce and spinach roots in comparable numbers (Wright et al.,2013), we anticipated similar levels of gyrB transcript. The reduction in level may be as a result of plant-associated inhibitors in spinach roots, or because of different plant-derived environmental cues acting on gyrB expression. Therefore, we suggest that the method needs to be validated if it is adopted for other plant species. Veriﬁcation of the technique using qPCR shows that this method could be applied to food safety settings, e.g., for detection of food-borne pathogens in fresh produce. It may be possible to couple this technique to standard methods, which would enable examination of viability and expression of genes of interest, e.g., toxin genes. Furthermore, it facilitates analysis of bacterial gene expression in planta for not only food-borne pathogens but also other plant-associated bacte- ria, providing an insight into the adaptive processes that underpin host–microbe interactions. Cooley, M., Carychao, D., Crawford-Miksza, L., Jay, M. T., Myers, C., Rose, C., et al. (2007). Incidence and tracking of Escherichia coli O157:H7 in a major produce production region in California. PLoS ONE 2:e1159. doi: 10.1371/journal.pone.0001159 Fink, R. C., Black, E. P., Hou, Z., Sugawara, M., Sadowsky, M. J., and Diez- Gonzalez, F. (2012). Transcriptional responses of Escherichia coli K-12 and O157:H7 associated with lettuce leaves. Appl. Environ. Microbiol. 78, 1752–1764. doi: 10.1128/aem.07454-11 Fislage, R., Berceanu, M., Humboldt, Y., Wendt, M., and Oberender, H. (1997). Primer design for a prokaryotic differential display RT-PCR. Nucl. Acids Res. 25, 1830–1835. doi: 10.1093/nar/25.9.1830 Gehrig, H. H., Winter, K., Cushman, J., Borland, A., and Taybi, T. (2000). An improved RNA isolation method for succulent plant species rich in polyphenols and polysaccharides. Plant Mol. Biol. Rep. 18, 369–376. doi: 10.1007/BF02825065 Goudeau, D. M., Parker, C. T., Zhou, Y., Sela, S., Kroupitski, Y., and Brandl, M. T. (2013). QUANTITATIVE REVERSE TRANSCRIPTASE PCR (qPCR) ANALYSIS QUANTITATIVE REVERSE TRANSCRIPTASE PCR (qPCR) ANALYSIS Bacterial gene expression was quantiﬁed from the samples obtained using the optimized method, by carrying out qPCR. The gyrB gene was selected as a housekeeping gene that was expected to be expressed under the conditions tested and expression from the root-derived samples was compared to samples obtained from bacteria grown under in vitro conditions. The amount of gyrB transcript was found to be similar between both lettuce roots and in vitro samples (Table 1), indicating similar copy numbers and stable gyrB expression under in vitro conditions and in plant extracts. It is of note that the amount of detectable gyrB was lower in the infected spinach root extracts compared to lettuce roots (higher Ct value for gyrB). June 2014 \| Volume 5 \| Article 286 \| 5 www.frontiersin.org E. coli O157:H7 RNA extraction from plants Holmes et al. inoculation with the plant roots, although the report does not state the weight of material required for RNA extraction. The main difference in our method was the amount of plant mate- rial required and the time at which the samples were taken. For example, our method allows for gene expression analysis after a short time of interaction, rather than after several days of colo- nization where the bacterial numbers are likely to have increased, dependent on the inoculation time and plant-bacterial system investigated. Brandl, M. T. (2006). Fitness of human enteric pathogens on plants and implications for food safety. Annu. Rev. Phytopathol. 44, 367–392. doi: 10.1146/annurev.phyto.44.070505.143359 Buchholz, U., Bernard, H., Werber, D., Bohmer, M. M., Remschmidt, C., Wilking, H., et al. (2011). German outbreak of Escherichia coli O104:H4 associated with sprouts. N. Engl. J. Med. 365, 1763–1770. doi: 10.1056/NEJMoa1106482 p g Cathala, G., Savouret, J. F., Mendez, B., West, B. L., Karin, M., Martial, J. A., et al. Cathala, G., Savouret, J. F., Mendez, B., West, B. L., Karin, M., Martial, J. A., et al. (1983). A method for isolation of intact, translationally active ribonucleic acid. Cathala, G., Savouret, J. F., Mendez, B., West, B. L., Karin, M., Martial, J. A., et al. (1983). A method for isolation of intact, translationally active ribonucleic acid. DNA 2, 329–335. doi: 10.1089/dna.1983.2.329 (1983). A method for isolation of intact, translationally active ribonucleic acid. DNA 2, 329–335. doi: 10.1089/dna.1983.2.329 Chang, S., Puryear, J., and Cairney, J. (1993). A simple and efﬁcient method for isolating RNA from pine trees. Plant Mol. AUTHOR CONTRIBUTIONS Al authors were involved in the design of the work and interpre- tation of data; Ashleigh Holmes and Louise Birse were involved in data acquisition and analysis; Ashleigh Holmes, Louise Birse, and Nicola J. Holden drafted the work; all authors approved the ﬁnal version. Holden, N., Pritchard, L., and Toth, I. (2009). Colonization outwith the colon: plants as an alternative environmental reservoir for human pathogenic enterobacteria. FEMS Microbiol. Rev. 33, 689–703. doi: 10.1111/j.1574-6976.2008.00153.x Hou, Z., Fink, R. C., Black, E., Sugawara, M., Zhang, Z., Diez-Gonzalez, F., et al. (2012). Gene expression proﬁling of Escherichia coli in response to inter- actions with the lettuce rhizosphere. J. Appl. Microbiol. 113, 1076–1086. doi: 10.1111/j.1365-2672.2012.05412.x ACKNOWLEDGMENTS The work was supported by BBSRC grant BB/I014179/1 and Scot- tish Government RESAS funding. The authors thank Pete Headley and Jenny Morris (James Hutton Institute) for technical assistance; Sonia Humphris and the reviewers for helpful suggestions. Hou, Z., Fink, R. C., Sugawara, M., Diez-Gonzalez, F., and Sadowsky, M. J. (2013). Transcriptional and functional responses of Escherichia coli O157:H7 growing in the lettuce rhizoplane. Food Microbiol. 35, 136–142. doi: 10.1016/j.fm.2013. 03.002 Jaakola, L., Pirttila, A. M., Halonen, M., and Hohtola, A. (2001). Isolation of high quality RNA from bilberry (Vaccinium myrtillus L.) fruit. Mol. Biotechnol. 19, 201–203. doi: 10.1385/mb:19:2:201 QUANTITATIVE REVERSE TRANSCRIPTASE PCR (qPCR) ANALYSIS The Salmonella transcriptome in lettuce and cilantro soft rot reveals a niche overlap with the animal host intestine. Appl. Environ. Microbiol. 79, 250–262. doi: 10.1128/aem.02290-12 Grifﬁths, R. I., Whiteley, A. S., O’Donnell, A. G., and Bailey, M. J. (2000). Rapid method for coextraction of DNA and RNA from natural environments for anal- ysis of ribosomal DNA- and rRNA-based microbial community composition. Appl. Environ. Microbiol. 66, 5488–5491. doi: 10.1128/AEM.66.12.5488-5491. 2000 Hayashi, T., Makino, K., Ohnishi, M., Kurokawa, K., Ishii, K., Yokoyama, K., et al. (2001). Complete genome sequence of enterohemorrhagic Escherichia coli O157:H7 and genomic comparison with a laboratory strain K-12. DNA Res. 8, 11–22. doi: 10.1093/dnares/8.1.11 Hernandez-Morales, A., De La Torre-Zavala, S., Ibarra-Laclette, E., Hernandez- Flores, J., Jofre-Garﬁas, A., Martinez-Antonio, A., et al. (2009). Transcriptional proﬁle of Pseudomonas syringae pv. phaseolicola NPS3121 in response to tissue extracts from a susceptible Phaseolus vulgaris L. cultivar. BMC Microbiol. 9:257. doi: 10.1186/1471-2180-9-257 REFERENCES Genomic analysis reveals the major driving forces of bacterial life in the rhizosphere. Genome Biol. 8, R179. doi: 10.1186/gb-2007-8- 9-r179 Wright, K. M., Chapman, S., Mcgeachy, K., Humphris, S., Campbell, E., Toth, I. K., et al. (2013). The endophytic lifestyle of Escherichia coli O157:H7: quan- tiﬁcation and internal localization in roots. Phytopathology 103, 333–340. doi: 10.1094/PHYTO-08-12-0209-FI Neidhardt, F. C., Bloch, P. L., and Smith, D. F. (1974). Culture medium for enterobacteria. J. Bacteriol. 119, 736–747. Zy´sko, A., Sanguin, H., Hayes, A., Wardleworth, L., Zeef, L. H., Sim, A., et al. (2012). Transcriptional response of Pseudomonas aeruginosa to a phosphate-deﬁcient Zy´sko, A., Sanguin, H., Hayes, A., Wardleworth, L., Zeef, L. H., Sim, A., et al. (2012). Transcriptional response of Pseudomonas aeruginosa to a phosphate-deﬁcient Lolium perenne rhizosphere. Plant Soil 359, 25–44. doi: 10.1007/s11104-011- 1060-z Rossez, Y., Holmes, A., Wolfson, E. B., Gally, D. L., Mahajan, A., Pedersen, H. L., et al. (2013). Flagella interact with ionic plant lipids to mediate adherence of pathogenic Escherichia coli to fresh produce plants. Environ. Microbiol. doi: 10.1111/1462-2920.12315 [Epub ahead of print]. Conflict of Interest Statement: The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. Roy, D., Panchal, S., Rosa, B. A., and Melotto, M. (2013). Escherichia coli O157:H7 induces stronger plant immunity than Salmonella enterica Typhimurium SL1344. Phytopathology 103, 326–332. doi: 10.1094/PHYTO-09-12- 0230-FI Received: 24 March 2014; paper pending published: 27 April 2014; accepted: 23 May 2014; published online: 10 June 2014. Received: 24 March 2014; paper pending published: 27 April 2014; accepted: 23 May 2014; published online: 10 June 2014. Citation: Holmes A, Birse L, Jackson RW and Holden NJ (2014) An optimized method for the extraction of bacterial mRNA from plant roots infected with Escherichia coli O157:H7. Front. Microbiol. 5:286. doi: 10.3389/fmicb.2014.00286 Schenk, A., Weingart, H., and Ullrich, M. S. (2008). Extraction of high-quality bacterial RNA from infected leaf tissue for bacterial in planta gene expression analysis by multiplexed ﬂuorescent Northern hybridization. Mol. Plant Pathol. 9, 227–235. doi: 10.1111/j.1364-3703.2007.00452.x Citation: Holmes A, Birse L, Jackson RW and Holden NJ (2014) An optimized method for the extraction of bacterial mRNA from plant roots infected with Escherichia coli O157:H7. Front. Microbiol. 5:286. doi: 10.3389/fmicb.2014.00286 O157:H7. Front. Microbiol. 5:286. doi: 10.3389/fmicb.2014.00286 Schikora, A., Carreri, A., Charpentier, E., and Hirt, H. (2008). REFERENCES An, S.-Q., Febrer, M., Mccarthy, Y., Tang, D.-J., Clissold, L., Kaithakottil, G., et al. (2013). High-resolution transcriptional analysis of the regulatory inﬂu- ence of cell-to-cell signalling reveals novel genes that contribute to Xanthomonas phytopathogenesis. Mol. Microbiol. 88, 1058–1069. doi: 10.1111/mmi.12229 Jahn, C. E., Charkowski, A. O., and Willis, D. K. (2008). Evaluation of isolation methods and RNA integrity for bacterial RNA quantitation. J. Microbiol. Methods 75, 318–324. doi: 10.1016/j.mimet.2008.07.004 Kim, B.-H., Ramanan, R., Cho, D.-H., Choi, G.-G., La, H.-J., Ahn, C.-Y., et al. (2012). Simple, rapid and cost-effective method for high quality nucleic acids extraction from different strains of Botryococcus braunii. PLoS ONE 7:e37770. doi: 10.1371/journal.pone.0037770 Barak, J. D., and Schroeder, B. K. (2012). Interrelationships of food safety and plant pathology: the life cycle of human pathogens on plants. Annu. Rev. Phytopathol. 50, 241–266. doi: 10.1146/annurev-phyto-081211-172936 Barlow, J. J., Mathias, A. P., Williamson, R., and Gammack, D. B. (1963). A simple method for the quantitative isolation of undegraded high molecular weight ribonucleic acid. Biochem. Biophys. Res. Commun. 13, 61–66. doi: 10.1016/0006-291X(63)90163-3 Kyle, J. L., Parker, C. T., Goudeau, D., and Brandl, M. T. (2010). Transcriptome analysis of Escherichia coli O157:H7 exposed to lysates of lettuce leaves. Appl. Environ. Microbiol. 76, 1375–1387. doi: 10.1128/AEM.02461-09 June 2014 \| Volume 5 \| Article 286 \| 6 Frontiers in Microbiology \| Plant-Microbe Interaction E. coli O157:H7 RNA extraction from plants Holmes et al. Leite, G. M., Magan, N., and Medina, A. (2012). Comparison of different bead-beating RNA extraction strategies: an optimized method for ﬁlamen- tous fungi. J. Microbiol. Methods 88, 413–418. doi: 10.1016/j.mimet.2012. 01.011 Soto-Suarez, M., Bernal, D., Gonzalez, C., Szurek, B., Guyot, R., Tohme, J., et al. (2010). In planta gene expression analysis of Xanthomonas oryzae pathovar oryzae, African strain MAI1. BMC Microbiol. 10:170. doi: 10.1186/1471-2180- 10-170 Mark, G. L., Dow, J. M., Kiely, P. D., Higgins, H., Haynes, J., Baysse, C., et al. (2005). Transcriptome proﬁling of bacterial responses to root exudates identiﬁes genes involved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454–17459. doi: 10.1073/pnas.0506407102 Thilmony, R., Underwood, W., and He, S. Y. (2006). Genome-wide transcriptional analysis of the Arabidopsis thaliana interaction with the plant pathogen Pseu- domonas syringae pv. tomato DC3000 and the human pathogen Escherichia coli y g p p g O157:H7. Plant J. 46, 34–53. doi: 10.1111/j.1365-313X.2006.02725.x Matilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and Ramos- Gonzalez, M. I. (2007). June 2014 \| Volume 5 \| Article 286 \| 7 REFERENCES The dark side of the salad: Salmonella typhimurium overcomes the innate immune response of Arabidopsis thaliana and shows an endopathogenic lifestyle. PLoS ONE 3:e2279. doi: 10.1371/journal.pone.0002279 This article was submitted to Plant–Microbe Interaction, a section of the journal Frontiers in Microbiology. Copyright © 2014 Holmes, Birse, Jackson and Holden. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s)orlicensorarecreditedandthattheoriginalpublicationinthisjournaliscited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Copyright © 2014 Holmes, Birse, Jackson and Holden. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s)orlicensorarecreditedandthattheoriginalpublicationinthisjournaliscited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Shidore, T., Dinse, T., Öhrlein, J., Becker, A., and Reinhold-Hurek, B. (2012). Tran- scriptomic analysis of responses to exudates reveal genes required for rhizosphere competence of the endophyte Azoarcus sp. strain BH72. Environ. Microbiol. 14, 2775–2787. doi: 10.1111/j.1462-2920.2012.02777.x June 2014 \| Volume 5 \| Article 286 \| 7 www.frontiersin.org www.frontiersin.org
https://openalex.org/W2795436427	https://europepmc.org/articles/pmc5893844?pdf=render	English	null	The Role of Platelet Microparticle Associated microRNAs in Cellular Crosstalk	Frontiers in cardiovascular medicine	2,018	cc-by	4,739	Introduction Reviewed by: Paola Patrignani, Università degli Studi G. d'Annunzio Chieti e Pescara, Italy Tarun Tyagi, Yale University United States Reviewed by: Paola Patrignani, Università degli Studi G. d'Annunzio Chieti e Pescara, Italy Tarun Tyagi, Yale University, United States Correspondence: Wai Ho Tang waiho.tang@gwcmc.org Platelet plays a pivotal role in maintaining primary hemostasis, and contributes to the pathogenesis of thrombotic and occlusive vascular disorders such as acute coronary syndrome or stroke (1). Despite lack of nucleus and genomic DNA, there are diverse types of RNAs in platelet, such as protein coding messenger RNAs (mRNAs), microRNAs (miRNAs), YRNAs and circular RNAs, which are inherited from their parent megakaryocytes (2–5). Platelets contain the essential machinery for processing miRNAs, which regulate the expression of proteins through complementary sequence recognition, binding and post-translational repression of mRNA transcripts (6, 7). The function of platelet miRNAs has gained widespread interest in the field of platelet biology since the platelet miRNA expression profile is closely associated with platelet biogenesis and activation. Recent studies have shown that platelets transfer and incorporate specific miRNAs into the recipient cells (8–11), raising the exciting possibility of potential therapeutic targets and disease biomarkers. In this review, we summarize and discuss the mechanistic roles of platelet miRNAs in circulation. Specialty section: This article was submitted to Cardiovascular Metabolism, a section of the journal Frontiers in Cardiovascular Medicine Received: 24 January 2018 Accepted: 15 March 2018 Published: 04 April 2018 Citation: Xia L, Zeng Z and Tang WH (2018) The Role of Platelet Microparticle Associated microRNAs in Cellular Crosstalk. Front. Cardiovasc. Med. 5:29. doi: 10.3389/fcvm.2018.00029 Specialty section: This article was submitted to Cardiovascular Metabolism, a section of the journal Frontiers in Cardiovascular Medicine Specialty section: This article was submitted to Cardiovascular Metabolism, a section of the journal Frontiers in Cardiovascular Medicine The Role of Platelet Microparticle Associated microRNAs in Cellular Crosstalk Luoxing Xia, Zhi Zeng and Wai Ho Tang Institute of Pediatrics, Guangzhou Women and Children's Medical Centre, Guangzhou Medical University, Guangzhou, China Platelet is an anucleate cell containing abundant messenger RNAs and microRNAs (miRNAs), and their functional roles in hemostasis and inflammation remain elusive. Accumulating evidence has suggested that platelets can actively transfer RNAs to hepatocytes, vascular cells, macrophages, and tumor cells. The incorporated mRNAs are translated into proteins, and miRNAs were found to regulate the gene expression, resulting in the functional change of the recipient cells. This novel intercellular communication opens up a new avenue for the pathophysiological role of platelet in platelet-associated vascular diseases. Therefore, understanding the underlying mechanism and identification of the platelet miRNAs involved in this biological process would provide novel diagnostic and therapeutic targets for cardiovascular diseases. Keywords: platelet, microparticle, microRNA, RNA, cardiovascular disease Edited by: Manfredi Tesauro, Università degli Studi di Roma Tor Vergata, Italy Platelet mRNAs and miRNAs Received: 24 January 2018 Accepted: 15 March 2018 Published: 04 April 2018 MicroRNAs are generated either as transcripts that rely on transcription as the promoter for the host protein-encoding gene or by using their own promoter transcript in intergenic region of the genome (12). Anucleate platelets retain megakaryocyte-derived mRNAs (13) and their unique adaptive signals have evolved in maintaining the diversity of genes and proteins (14, 15). Keywords: platelet, microparticle, microRNA, RNA, cardiovascular disease Mini Review published: 04 April 2018 doi: 10.3389/fcvm.2018.00029 Mini Review Platelets and Platelet-Derived Microparticles The horizontal transfer of platelet RNA has also been reported in THP-1 cells (44). Antonina Risitano et al. found that the RNA of PLPs derived from Meg-01 was transferred to THP-1 cells, and those incorporated RNAs were functional in the recipient cells (44). Using microarray analysis of THP-1 cells showed that the genes globins (HGB1/HGB2) and globins (HBA1/HBA2) involved in transfer were up-regulated, but not altered in the control cells (44). In addition, after infused with wildtype platelets, the presence of TLR2-positive monocytes was found in TLR2-deficient mice treated with LPS, indicating that specific platelet RNAs are transferred and incorporated into leukocytes in vivo (44).f Platelets are fragments of cytoplasm, which are derived from the megakaryocytes (29) of the bone marrow and lung (30, 31), then entered into the circulation. Circulating blood platelets have a lifespan of 8–10 days, and play a central role in hemostasis (32, 33). At the site of injury, platelets contact with many adhesion molecules, including collagen, which interacts with platelet surface receptors, leading to the adhesion and activation of platelets. Once activated, platelets release their bioactive components, such as ADP, serotonin and thromboxane A2, leading to further platelet recruitment, aggregation and plug formation (7). Platelet hyperreactivity contributes to the pathogenesis of thrombotic and occlusive vascular disorders, such as atherosclerosis, thrombotic cardiovascular disorders, occlusive thrombus or inflammation and cancer (34–38). y Benoit Laffont et al. demonstrated that PMPs containing miR- 126–3 p was internalized by primary human macrophages (10). The incorporated miR-126–3 p suppressed the expression of four predicted mRNA targets of miR-126–3 p, and two of them were determined at the protein level. The effect of miR-126–3 p was further confirmed by expressing a neutralising miR-126–3 p sponge (10). Furthermore, PMPs induced the upregulation of 34 miRNAs with concomitant downregulation of 367 RNAs, including mRNAs encoding cytokines/chemokines chemokine (C-C motif) ligand 4 (CCL4), colony stimulating factor 1 (CSF1) and tumor necrosis factor (TNF) (10). These changes induced by PMPs were accompanied by a marked increase in the phagocytic capacity of macrophages (10). Microparticles (MPs), also referred to as microvesicles or more rarely ectosomes, are submicron fragments shed from stimulated plasma membrane or apoptotic cells (39). Platelet-derived MPs (PMPs) generated from more than 45% of plasma-borne MPs (39, 40). Based on the size of exosomes (<0.1 µm) and apoptotic bodies (>1 µm), PMPs (0.1–1 µm) can be distinguished from them. Citation: Subsequent analysis of microarray and RNA sequencing have focused on non-healthy individuals, which correlated RNA profiles to specific human diseases (18–23). The stability of platelet miRNAs is unique, especially for the most abundant transcripts, suggesting that the platelet miRNome may be more stable than the miRNomes of other circulating cells or blood pools, such as plasma (24). It is noteworthy that almost half of miRNAs in microparticles, the major proportion of miRNAs in plasma, are produced by platelet (25). Numerous miRNAs have been found to be expressed in platelets and associated with cardiovascular diseases (CVD). Duisters RF et al. (24) reported that miR-30c repressed CTGF expression which was associated with ventricular fibrosis and heart failure. Tang, Y. et al. (26) found that miRNA–150 protected the mouse heart from ischemic injury by regulating EGR2, MYB and P2R × 7. McManus DD et al. (27, 28) demonstrated that miR-328 regulated CACNA1C and CACNB1 expressions, leading to atrial fibrillation and electrical remodeling. Willeit P et al. (29) showed that miR-342 regulated AKT1, glucose metabolism, apoptosis and cell proliferation, resulting in inflammatory stimulation of macrophages. Over the past decade, emerging evidence suggested that platelet miRNAs were potential regulators of platelet protein translation and expression as well as biomarkers for hematologic disease and platelet reactivity (4). In addition, considerable attention has be paid to the horizontal transfer of platelet miRNAs to the recipient cells. (0.1 µm) by stimulating the collagen receptor glycoprotein VI in arthritis pathophysiology (36). Moreover, recent studies have demonstrated that measurement of PMPs can be used as the marker for platelet activation (42, 43). The Horizontal Transfer of Platelet-Derived mRNA/miRNA Into the Vascular Cells Upon activation, platelets can regulate the function of vascular cells by releasing the bioactive molecules. Platelet mRNAs are only associated with low levels of protein translation; however, platelets have a unique membrane structure that allows small molecules to pass through, so their cytoplasmic RNA can be delivered to nucleated cells (44). Antonina Risitano et al. demonstrated that coculture of platelet-like particles (PLPs) derived from megakaryoblastic cell line Meg-01 and human umbilical vein endothelial cells (HUVECs) led to the horizontal transfer of platelet RNA into HUVECs, and those incorporated RNA were functional in the recipient cells (44). Interestingly, platelets were also found to transfer specific miRNA to endothelial cells and regulate their gene expression of the endothelial cells. Benoit Laffont et al. found that platelets stimulated with 0.1 U/ml thrombin preferentially released MPs containing miR-223, and those MPs were internalized by HUVECs (8). Also, PMPs contain Argonaute 2 (Ago2)•miR-223 complexes, suggesting that the incorporated miRNAs were functional in the recipient HUVECs. Indeed, they found that platelet-derived miR- 223 regulated two endogenous endothelial genes FBXW7 and EFNA1 at mRNA and protein levels (8).h Frontiers in Cardiovascular Medicine \| www.frontiersin.org Citation: Xia L, Zeng Z and Tang WH (2018) The Role of Platelet Microparticle Associated microRNAs in Cellular Crosstalk. Front. Cardiovasc. Med. 5:29. doi: 10.3389/fcvm.2018.00029 Xia L, Zeng Z and Tang WH (2018) The Role of Platelet Microparticle Associated microRNAs in Cellular Crosstalk. Front. Cardiovasc. Med. 5:29. doi: 10.3389/fcvm.2018.00029 With the use of high-throughput sequencing, Hélène Plé et al. showed that based on the size of RNA, most of small RNAs found in platelets were miRNAs (16). Platelets have miRNAs, Dicer and Argonaute protein complexes, so pre-RNAs can be processed into miRNAs and control reporter April 2018 \| Volume 5 \| Article 29 1 Frontiers in Cardiovascular Medicine \| www.frontiersin.org Platelet microRNAs in Cellular crosstalk Xia et al. transcripts. Analysed by the deep sequencing techniques, there were about 9,500 transcripts in the platelet of healthy subject (5, 17). Subsequent analysis of microarray and RNA sequencing have focused on non-healthy individuals, which correlated RNA profiles to specific human diseases (18–23). The stability of platelet miRNAs is unique, especially for the most abundant transcripts, suggesting that the platelet miRNome may be more stable than the miRNomes of other circulating cells or blood pools, such as plasma (24). It is noteworthy that almost half of miRNAs in microparticles, the major proportion of miRNAs in plasma, are produced by platelet (25). Numerous miRNAs have been found to be expressed in platelets and associated with cardiovascular diseases (CVD). Duisters RF et al. (24) reported that miR-30c repressed CTGF expression which was associated with ventricular fibrosis and heart failure. Tang, Y. et al. (26) found that miRNA–150 protected the mouse heart from ischemic injury by regulating EGR2, MYB and P2R × 7. McManus DD et al. (27, 28) demonstrated that miR-328 regulated CACNA1C and CACNB1 expressions, leading to atrial fibrillation and electrical remodeling. Willeit P et al. (29) showed that miR-342 regulated AKT1, glucose metabolism, apoptosis and cell proliferation, resulting in inflammatory stimulation of macrophages. Over the past decade, emerging evidence suggested that platelet miRNAs were potential regulators of platelet protein translation and expression as well as biomarkers for hematologic disease and platelet reactivity (4). In addition, considerable attention has be paid to the horizontal transfer of platelet miRNAs to the recipient cells. transcripts. Analysed by the deep sequencing techniques, there were about 9,500 transcripts in the platelet of healthy subject (5, 17). The Horizontal Transfer of Platelet-Derived mRNA Into the Hepatocytes Accumulating evidence shed light on the use of platelet-derived miRNAs as diagnostic markers for various diseases. Platelet- derived miRNAs have been associated with atrial fibrillation, coronary disease, heart failure and vascular disease (50). For example, miR-30c was associated with heart failure (24), miR- 126 was associated with myocardial angiogenesis (28), miR-197 was associated with metabolic syndrome (51), miR-328 was associated with atrial fibrillation (27). Since the blood samples from patients are easily accessible, the identification of platelet- derived miRNAs in diseases will open up a new avenue for the clinical diagnosis and treatment. Marc Kirschbaum et al. showed that either platelets or PLPs derived from the MEG-01 stimulated the proliferation of HepG2 cells (9). They also observed platelet internalization into hepatocytes following a partial hepatectomy in mice (9). PLPs-derived RNA was detected in the cytoplasm of the HepG2 cells. The removal of PLPs-derived RNA by RNA-degrading enzymes partly blocked the stimulating effect of platelets on hepatocyte proliferation (9). Thus, platelets stimulate hepatocyte proliferation via the horizontal transfer of platelet RNA (9). Although no further evidence has been reported that miRNAs were involved in this process, it is plausible that activated platelet could transfer specific miRNAs to hepatocytes and participate into the regulation of liver regeneration. As discussed above, many studies have reported that platelets, MPs and associated miRNAs can be transferred and incorporated into the various cells, and regulate the function of the recipient cells. However, there are still a number of concerns to be addressed. First, it is important to understand how platelets secretes specific miRNAs in response to different stimuli and pathological conditions, in order to precisely deliver therapeutic RNAs to target sites. Second, it is important to know how platelets and PMPs were selectively internalised by the recipient cells, in order to accurately deliver the platelets or PMPs to the specific cells, tissues, or injured sites. If all these unanswered questions could be addressed, using platelets and MPs as drug deliver system would be an efficient and site-specific approach for the treatment of platelet-associated diseases. Platelets and Platelet-Derived Microparticles Despite lack of direct evidence showing the horizontal transfer of platelet miRNAs into VSMCs in vivo, these findings provide some clues for the possible role of platelet-derived exosomes containing miR-21, miR-223 and miR- 339 in regulating the phenotypic switching of VSMCs in diseased condition (46). Recently, James V. Michael et al. reported that platelet MPs suppressed tumor growth and induced tumor cell apoptosis (11). Solid tumor vasculature is highly permeable, allowing the possibility of PMPs-tumor interaction. They showed that PMPs infiltrated solid tumors in human and mice, thereby transferring platelet- derived miRNAs into tumor cells in vitro and in vivo, resulting in tumor cell apoptosis (11). PMPs transfusion suppressed the growth of both lung and colon carcinoma ectopic tumors, whereas blockade of miR-24 in tumor cells accelerated tumor growth. Par4 KO mice had the phenotype of reduced PMPs generation, and the tumor growth in Par4 KO mice was accelerated compared with wild type mice, indicating that PMPs play a pivotal role in the inhibition of tumor growth (11). The direct targets of PMPs-derived miR-24 in tumor cells were found to be the mitochondrial mt-Nd2 and a small noncoding nucleolar RNA Snora75 (11). The inhibition of mt-Nd2 and Snora75 resulted in mitochondrial dysfunction and growth inhibition of tumor cells (11). Thus, further studies are needed to determine the mechanistic role of the incorporated PMPs-derived miRNAs in the tumor progression. The Horizontal Transfer of Platelet-Derived mRNA/miRNA Into the Tumor Cells Tumor growth is a major pathophysiological condition that may be affected by platelet MPs and associated miRNAs (7, 48). PMPs can directly stimulate the tumor growth through the release of potent growth factors in the tumor micro-environment (49). In 2015, Liang et al. first reported that platelet-secreted miRNA exacerbated lung cancer cell invasion (49). They found that miR- 223 significantly increased in MVs derived from platelets of Non- small cell lung cancer patients (49). After cocultured with A549 cells, MVs derived from platelets delivered miR-223 into the recipient cells (49). The incorporated miR-223 exacerbated A549 cells invasion via targeting EPB41L3 (49). Platelets and Platelet-Derived Microparticles PMPs are shed from cytoplasmic membrane of platelet express the surface antigens, such as CD41 and CD62p (P-selectin) which are different from exosomes released from intracellular compartments (41). Boilard et al. showed that platelets generated microparticles Vascular smooth muscle cells (VSMCs) are the major component of the vascular wall and endure both ongoing damage and repair April 2018 \| Volume 5 \| Article 29 2 Platelet microRNAs in Cellular crosstalk Xia et al. Xia et al. during atherosclerosis (45). During repair, newly migrated VSMCs tend to differentiate and frequently undergo apoptosis (46). A recent study has found that platelets may transfer miRNAs to VSMCs via platelet-derived exosomes (47). They also showed that in the murine model of carotid tandem stenosis, the levels of miR- 21, miR-223 and miR-339 were associated with platelet activation (47). These miRNAs were elevated in the pooled mouse plasma exosomes before thrombosis, and were also found to be enriched in thrombin-stimulated platelet-derived exosomes in vitro (47). After cocultured VSMCs with platelet-derived exosomes, the expression of platelet-derived growth factor receptor-beta (PDGFRβ) in VSMCs was decreased (47). Platelet-derived exosomes also inhibited PDGF-stimulated VSMC proliferation. Furthermore, a decrease in PDGFRβ expression was observed in VSMCs around thrombotic areas in vivo (46). Despite lack of direct evidence showing the horizontal transfer of platelet miRNAs into VSMCs in vivo, these findings provide some clues for the possible role of platelet-derived exosomes containing miR-21, miR-223 and miR- 339 in regulating the phenotypic switching of VSMCs in diseased condition (46). during atherosclerosis (45). During repair, newly migrated VSMCs tend to differentiate and frequently undergo apoptosis (46). A recent study has found that platelets may transfer miRNAs to VSMCs via platelet-derived exosomes (47). They also showed that in the murine model of carotid tandem stenosis, the levels of miR- 21, miR-223 and miR-339 were associated with platelet activation (47). These miRNAs were elevated in the pooled mouse plasma exosomes before thrombosis, and were also found to be enriched in thrombin-stimulated platelet-derived exosomes in vitro (47). After cocultured VSMCs with platelet-derived exosomes, the expression of platelet-derived growth factor receptor-beta (PDGFRβ) in VSMCs was decreased (47). Platelet-derived exosomes also inhibited PDGF-stimulated VSMC proliferation. Furthermore, a decrease in PDGFRβ expression was observed in VSMCs around thrombotic areas in vivo (46). References 10. Laffont B, Corduan A, Rousseau M, Duchez AC, Lee CH, Boilard E, et al. Platelet microparticles reprogram macrophage gene expression and function. Thromb Haemost (2016) 115(2):311–23. doi: 10.1160/TH15-05-0389 1. Clancy L, Freedman JE. New paradigms in thrombosis: novel mediators and biomarkers platelet RNA transfer. J Thromb Thrombolysis (2014) 37(1):12–6. doi: 10.1007/s11239-013-1001-1 11. Michael JV, Wurtzel JGT, Mao GF, Rao AK, Kolpakov MA, Sabri A, et al. Platelet microparticles infiltrating solid tumors transfer miRNAs that suppress tumor growth. Blood (2017) 130(5):567–80. doi: 10.1182/blood-2016-11- 751099 2. Sunderland N, Skroblin P, Barwari T, Huntley RP, Lu R, Joshi A, et al. MicroRNA biomarkers and platelet reactivity: the clot thickens. Circ Res (2017) 120(2):418–35. doi: 10.1161/CIRCRESAHA.116.309303 12. Bauersachs J, Thum T. Biogenesis and regulation of cardiovascular microRNAs. Circ Res (2011) 109(3):334–47. doi: 10.1161/CIRCRESAHA.110.228676 3. Nagalla S, Shaw C, Kong X, Kondkar AA, Edelstein LC, Ma L, et al. Platelet microRNA-mRNA coexpression profiles correlate with platelet reactivity. Blood (2011) 117(19):5189–97. doi: 10.1182/blood-2010-09-299719 13. Gnatenko DV, Dunn JJ, Mccorkle SR, Weissmann D, Perrotta PL, Bahou WF. Transcript profiling of human platelets using microarray and serial analysis of gene expression. Blood (2003) 101(6):2285–93. doi: 10.1182/blood-2002-09- 2797 4. Edelstein LC, Bray PF. MicroRNAs in platelet production and activation. Blood (2011) 117(20):5289–96. doi: 10.1182/blood-2011-01-292011 14. Denis MM, Tolley ND, Bunting M, Schwertz H, Jiang H, Lindemann S, et al. Escaping the nuclear confines: signal-dependent pre-mRNA splicing in anucleate platelets. Cell (2005) 122(3):379–91. doi: 10.1016/j. cell.2005.06.015 5. Rowley JW, Oler AJ, Tolley ND, Hunter BN, Low EN, Nix DA, et al. Genome- wide RNA-seq analysis of human and mouse platelet transcriptomes. Blood (2011) 118(14):e101–11. doi: 10.1182/blood-2011-03-339705 15. Weyrich AS, Dixon DA, Pabla R, Elstad MR, Mcintyre TM, Prescott SM, et al. Signal-dependent translation of a regulatory protein, Bcl-3, in activated human platelets. Proc Natl Acad Sci USA (1998) 95(10):5556–61. doi: 10.1073/ pnas.95.10.5556 6. Landry P, Plante I, Ouellet DL, Perron MP, Rousseau G, Provost P. Existence of a microRNA pathway in anucleate platelets. Nat Struct Mol Biol (2009) 16(9):961–6. doi: 10.1038/nsmb.1651 7. Provost P. The clinical significance of platelet microparticle-associated microRNAs. Clin Chem Lab Med (2017) 55(5):657–66. doi: 10.1515/cclm- 2016-0895 16. Plé H, Landry P, Benham A, Coarfa C, Gunaratne PH, Provost P. The repertoire and features of human platelet microRNAs. PLoS ONE (2012) 7(12):e50746. doi: 10.1371/journal.pone.0050746 8. Laffont B, Corduan A, Plé H, Duchez AC, Cloutier N, Boilard E, et al. Conclusion Platelets participate into multiple biological processes, such as inflammation, wound healing, maintenance of blood-lymph barrier and blood clotting where they provide a first and crucial line Frontiers in Cardiovascular Medicine \| www.frontiersin.org Frontiers in Cardiovascular Medicine \| www.frontiersin.org April 2018 \| Volume 5 \| Article 29 3 Xia et al. Platelet microRNAs in Cellular crosstalk Xia et al. Figure 1 \| Summarizing the role of platelet microparticles (MPs) in cellular activation by the transfer of miRNAs. Ecs: endotheilal cells, VSMCs: vascular smooth muscle cells. Since the horizontal transfer of platelet miRNA may represent a novel form of cell-cell communication, which may participate in many pathophysiological processes, growing concern has been received in this field. However, many unanswered questions remain to be addressed about the specificity and selectivity of the horizontal transfer of miRNAs. Therefore, understanding the mechanism by which platelet interacts with other cells, and identifying which miRNAs are involved in the horizontal transfer would provide novel therapeutic targets and diagnostic markers for platelet-associated diseases. Funding of defense against injury, thus maintaining normal homeostasis. As discussed above, recent advance in platelet biology demonstrates that activated platelets release the MPs containing abundant mRNA and miRNA, which are internalised by the recipient cells. The incorporated mRNA or miRNA can modulate the gene expression and regulate the function of the recipient cells, e.g., macrophages, ECs, leukocytes, VSMCs, hepotacytes, and tumor cells (Figure 1). WT is supported by the Guangdong Outstanding Young Scientist Funding (Grant No. 2016A030306049). This work was supported by the National Natural Science Foundation of China (Grant No. 81670117), the Guangdong Province Science and Technology Project (Grant No. 2017A050506008) and by the Guangzhou Science and Technology Project (Grant No. 201707010270). Author Contributions LX wrote this manuscript. ZZ and WT were responsible for editing the manuscript. All authors accept responsibility for the entire content of this submitted manuscript and have approved its submission. Figure 1 \| Summarizing the role of platelet microparticles (MPs) in cellular activation by the transfer of miRNAs. Ecs: endotheilal cells, VSMCs: vascular smooth muscle cells. References Activated platelets can deliver mRNA regulatory Ago2•microRNA complexes to endothelial cells via microparticles. Blood (2013) 122(2):253–61. doi: 10.1182/ blood-2013-03-492801 17. Bray PF, Mckenzie SE, Edelstein LC, Nagalla S, Delgrosso K, Ertel A, et al. The complex transcriptional landscape of the anucleate human platelet. BMC Genomics (2013) 14:1. doi: 10.1186/1471-2164-14-1 18. Freedman JE, Larson MG, Tanriverdi K, O'Donnell CJ, Morin K, Hakanson AS, et al. Relation of platelet and leukocyte inflammatory transcripts to body mass index in the Framingham heart study. Circulation (2010) 122(2):119–29. doi: 10.1161/CIRCULATIONAHA.109.928192 9. Kirschbaum M, Karimian G, Adelmeijer J, Giepmans BN, Porte RJ, Lisman T. Horizontal RNA transfer mediates platelet-induced hepatocyte proliferation. Blood (2015) 126(6):798–806. doi: 10.1182/blood-2014-09-600312 Frontiers in Cardiovascular Medicine \| www.frontiersin.org April 2018 \| Volume 5 \| Article 29 4 Xia et al. Platelet microRNAs in Cellular crosstalk 19. Gnatenko DV, Cupit LD, Huang EC, Dhundale A, Perrotta PL, Bahou WF. Platelets express steroidogenic 17beta-hydroxysteroid dehydrogenases. Distinct profiles predict the essential thrombocythemic phenotype. Thromb Haemost (2005) 94(2):412–21. doi: 10.1160/TH05-01-0037 37. Jain S, Harris J, Ware J. Platelets: linking hemostasis and cancer. Arterioscler Thromb Vasc Biol (2010) 30(12):2362–7. doi: 10.1161/ATVBAHA.110.207514h h 38. Dangwal S, Thum T. MicroRNAs in platelet physiology and pathology. Hamostaseologie (2013) 33(1):17–20. doi: 10.5482/HAMO-13-01-0002 20. Gnatenko DV, Zhu W, Xu X, Samuel ET, Monaghan M, Zarrabi MH, et al. Class prediction models of thrombocytosis using genetic biomarkers. Blood (2010) 115(1):7–14. doi: 10.1182/blood-2009-05-224477 39. Vanwijk MJ, Vanbavel E, Sturk A, Nieuwland R. Microparticles in cardiovascular diseases. Cardiovasc Res (2003) 59(2):277–87. doi: 10.1016/ S0008-6363(03)00367-5 40. Horstman LL, Ahn YS. Platelet microparticles: a wide-angle perspective. Crit Rev Oncol Hematol (1999) 30(2):111–42. doi: 10.1016/S1040-8428(98)00044-4 21. Lood C, Amisten S, Gullstrand B, Jönsen A, Allhorn M, Truedsson L, et al. Platelet transcriptional profile and protein expression in patients with systemic lupus erythematosus: up-regulation of the type I interferon system is strongly associated with vascular disease. Blood (2010) 116(11):1951–7. doi: 10.1182/ blood-2010-03-274605 41. Hristov M, Erl W, Linder S, Weber PC. Apoptotic bodies from endothelial cells enhance the number and initiate the differentiation of human endothelial progenitor cells in vitro. Blood (2004) 104(9):2761–6. doi: 10.1182/blood-2003- 10-3614i 22. Nagalla S, Bray PF. Platelet RNA chips dip into thrombocytosis. Blood (2010) 115(1):2–3. doi: 10.1182/blood-2009-10-246405i 42. Nomura S. Function and clinical significance of platelet-derived microparticles. Int J Hematol (2001) 74(4):397–404. doi: 10.1007/BF02982082 23. Raghavachari N, Xu X, Harris A, Villagra J, Logun C, Barb J, et al. References Amplified expression profiling of platelet transcriptome reveals changes in arginine metabolic pathways in patients with sickle cell disease. Circulation (2007) 115(12):1551–62. doi: 10.1161/CIRCULATIONAHA.106.658641 43. Italiano JE, Mairuhu AT, Flaumenhaft R. Clinical relevance of microparticles from platelets and megakaryocytes. Curr Opin Hematol (2010) 17(6):578–84. doi: 10.1097/MOH.0b013e32833e77ee 24. Duisters RF, Tijsen AJ, Schroen B, Leenders JJ, Lentink V, van der Made I, et al. miR-133 and miR-30 regulate connective tissue growth factor: implications for a role of microRNAs in myocardial matrix remodeling. Circ Res (2009) 104(2):170–8. doi: 10.1161/CIRCRESAHA.108.182535 44. Risitano A, Beaulieu LM, Vitseva O, Freedman JE. Platelets and platelet-like particles mediate intercellular RNA transfer. Blood (2012) 119(26):6288–95. doi: 10.1182/blood-2011-12-396440 45. Uryga AK, Bennett MR. Ageing induced vascular smooth muscle cell senescence in atherosclerosis. J Physiol (2016) 594(8):2115–24. doi: 10.1113/ JP270923 25. Diehl P, Fricke A, Sander L, Stamm J, Bassler N, Htun N, et al. Microparticles: major transport vehicles for distinct microRNAs in circulation. Cardiovasc Res (2012) 93(4):633–44. doi: 10.1093/cvr/cvs007 46. Silverman-Gavrila R, Silverman-Gavrila L, Bendeck MP. Cell division fidelity is altered during the vascular response to injury: its novel role in atherosclerosis progression. Am J Pathol (2013) 182(3):628–39. doi: 10.1016/j. ajpath.2012.11.007h 26. Tang Y, Wang Y, Park KM, Hu Q, Teoh JP, Broskova Z, et al. MicroRNA-150 protects the mouse heart from ischaemic injury by regulating cell death. Cardiovasc Res (2015) 106(3):387–97. doi: 10.1093/cvr/cvv121 27. Mcmanus DD, Lin H, Tanriverdi K, Quercio M, Yin X, Larson MG, et al. Relations between circulating microRNAs and atrial fibrillation: data from the Framingham Offspring Study. Heart Rhythm (2014) 11(4):663–9. doi: 10.1016/j.hrthm.2014.01.018 47. Tan M, Yan HB, Li JN, Li WK, Fu YY, Chen W, et al. Thrombin stimulated platelet-derived exosomes inhibit platelet-derived growth factor receptor- beta expression in vascular smooth muscle cells. Cell Physiol Biochem (2016) 38(6):2348–65. doi: 10.1159/000445588 28. Willeit P, Zampetaki A, Dudek K, Kaudewitz D, King A, Kirkby NS, et al. Circulating microRNAs as novel biomarkers for platelet activation. Circ Res (2013) 112(4):595–600. doi: 10.1161/CIRCRESAHA.111.300539h 48. Goubran H, Sabry W, Kotb R, Seghatchian J, Burnouf T. Platelet microparticles and cancer: an intimate cross-talk. Transfus Apher Sci (2015) 53(2):168–72. doi: 10.1016/j.transci.2015.10.014 29. Machlus KR, Thon JN, Italiano JE. Interpreting the developmental dance of the megakaryocyte: a review of the cellular and molecular processes mediating platelet formation. Br J Haematol (2014) 165(2):227–36. doi: 10.1111/bjh.12758h 49. Liang H, Yan X, Pan Y, Wang Y, Wang N, Li L, et al. Frontiers in Cardiovascular Medicine \| www.frontiersin.org References MicroRNA-223 delivered by platelet-derived microvesicles promotes lung cancer cell invasion via targeting tumor suppressor EPB41L3. Mol Cancer (2015) 14:58. doi: 10.1186/ s12943-015-0327-z 30. Howell WH, Donahue DD. The production of blood platelets in the lungs. J Exp Med (1937) 65(2):177–203. doi: 10.1084/jem.65.2.177 50. Mcmanus DD, Freedman JE. MicroRNAs in platelet function and cardiovascular disease. Nat Rev Cardiol (2015) 12(12):711–7. doi: 10.1038/ nrcardio.2015.101 31. Lefrançais E, Ortiz-Muñoz G, Caudrillier A, Mallavia B, Liu F, Sayah DM, et al. The lung is a site of platelet biogenesis and a reservoir for haematopoietic progenitors. Nature (2017) 544(7648):105–9. doi: 10.1038/nature21706h 51. Karolina DS, Tavintharan S, Armugam A, Sepramaniam S, Pek SL, Wong MT, et al. Circulating miRNA profiles in patients with metabolic syndrome. J Clin Endocrinol Metab (2012) 97(12):E2271–6. doi: 10.1210/jc.2012-1996 32. Harker LA, Finch CA. Thrombokinetics in man. J Clin Invest (1969) 48(6):963– 74. doi: 10.1172/JCI106077 33. Hartwig JH. The platelet: form and function. Semin Hematol (2006) 43(Suppl 1):S94–100. doi: 10.1053/j.seminhematol.2005.11.004 Conflict of Interest Statement: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. 34. Kottke-Marchant K. Importance of platelets and platelet response in acute coronary syndromes. Cleve Clin J Med (2009) 76(Suppl 1):S2–7. doi: 10.3949/ ccjm.76.s1.01l Copyright © 2018 Xia, Zeng and Tang. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. 35. Nurden AT, Platelets NAT. Platelets, inflammation and tissue regeneration. Thromb Haemost (2011) 105(Suppl 1):S13–33. doi: 10.1160/THS10-11-0720 36. Boilard E, Nigrovic PA, Larabee K, Watts GF, Coblyn JS, Weinblatt ME, et al. Platelets amplify inflammation in arthritis via collagen-dependent microparticle production. Science (2010) 327(5965):580–3. doi: 10.1126/ science.1181928 April 2018 \| Volume 5 \| Article 29 Frontiers in Cardiovascular Medicine \| www.frontiersin.org 5
https://openalex.org/W4239173911	https://se.copernicus.org/articles/4/59/2013/se-4-59-2013.pdf	English	null	Segmentation of the Izu-Bonin and Mariana plates based on the analysis of the Benioff seismicity distribution and regional tomography results	null	2,012	cc-by	10,726	Segmentation of the Izu-Bonin and Mariana slabs based on the analysis of the Benioff seismicity distribution and regional tomography results K. Jaxybulatov, I. Koulakov, and N. L. Dobretsov Institute of Petroleum Geology and Geophysics SB RAS, Prospekt Koptyuga, 3, 630090, Novosibirsk, Russia Correspondence to: K. Jaxybulatov (jaxybulatov@gmail.com) Received: 7 June 2012 – Published in Solid Earth Discuss.: 5 July 2012 Revised: 5 December 2012 – Accepted: 18 December 2012 – Published: 31 January 2013 Received: 7 June 2012 – Published in Solid Earth Discuss.: 5 July 2012 Revised: 5 December 2012 – Accepted: 18 December 2012 – Published: 31 January 2013 Received: 7 June 2012 – Published in Solid Earth Discuss.: 5 July 2012 Revised: 5 December 2012 – Accepted: 18 December 2012 – Published: 31 January 2013 Received: 7 June 2012 – Published in Solid Earth Discuss.: 5 July 2012 Revised: 5 December 2012 – Accepted: 18 December 2012 – Published: 31 January 2013 Abstract. We present a new model of P and S velocity anomalies in the mantle down to a depth of 1300 km beneath the Izu-Bonin and Mariana (IBM) arcs. This model is de- rived based on tomographic inversion of global travel time data from the revised ISC catalogue. The results of inversion are thoroughly veriﬁed using a series of different tests. The obtained model is generally consistent with previous studies by different authors. We also present the distribution of relo- cated deep events projected to the vertical surface along the IBM arc system. Unexpectedly, the seismicity forms elon- gated vertical clusters instead of horizontal zones indicating phase transitions in the slab. We propose that these verti- cal seismicity zones mark zones of intense deformation and boundaries between semi-autonomous segments of the sub- ducting plate. The P and S seismic tomography models con- sistently display the slab as prominent high-velocity anoma- lies coinciding with the distribution of deep seismicity. We can distinguish at least four segments which subduct differ- ently. The northernmost segment of the Izu-Bonin arc has the gentlest angle of dipping which is explained by backward displacement of the trench. In the second segment, the trench stayed at the same location, and we observe the accumulation of the slab material in the transition zone and its further de- scending to the lower mantle. In the third segment, the trench is moving forward causing the steepening of the slab. Finally, for the Mariana segment, despite the backward displacement of the arc, the subducting slab is nearly vertical. Segmentation of the Izu-Bonin and Mariana slabs based on the analysis of the Benioff seismicity distribution and regional tomography results Between the Izu-Bonin and Mariana arcs we clearly observe a gap which can be traced down to about 400 km in depth. Based on joint consideration of the tomography results and the seismicity distribution, we propose two different scenarios of the sub- duction evolution in the IBM zone during the recent time, depending on the reference frame of plate displacements. In the ﬁrst case, we consider the movements in respect to the Philippine Plate, and explain the different styles of the sub- duction by the relative backward and forward migrations of the trench. In the second case, all the elements of the sub- duction system move westward in respect to the stable Asia. Different subduction styles are explained by the “anchoring” of selected segments of the slab, different physical properties of the subducting plate and the existence of buoyant rigid blocks related to sea mount and igneous provinces. 1 Introduction Subduction zones are the most spectacular geological objects where intensive geodynamic processes occur. Strong me- chanical deformations and complex geochemical processes in subduction zones are responsible for the diversity of vari- ous geological structures observed on the surface. Investigat- ing the links between surface geology and processes in the subducting slab is an actual problem that involves many spe- cialists from different ﬁelds of geology and geophysics (e.g., Dobretsov et al., 2012). The general concept of subduction is conﬁrmed by a number of different data and accepted by most specialists in the world. However, the details of this process are the subject of intensive debates. In particular, representation of the subduction zones as a two-dimensional conveyer-type process appears to be in most cases over- simpliﬁed and it does not explain many important phenom- ena. For example, the distribution of deep seismicity, arc Solid Earth, 4, 59–73, 2013 www.solid-earth.net/4/59/2013/ doi:10.5194/se-4-59-2013 © Author(s) 2013. CC Attribution 3.0 License. Solid Earth 1.2 Previous slab structure investigations Fig. 1. The main geographic and tectonic units of the studied re- gion. Background is map of topography/bathymetry. The colours indicate the regions of the oceanic crust of different ages according to Peive, (1980): Pre-Cretaceous (green), Pre-Mesozoic (red), Pre- Miocene (yellow), Pre-Quaternary age (white), sea ﬂoor of Paciﬁc Ocean, Pre-Triassic (blue). White dotted lines depict arcs (ancient and present) (Peive, 1980). Names and ages of arcs are indicated. Arrows with numbers present the displacement rates in respect to Philippine plate (DeMets et al., 2010). The red triangles indicate ac- tive volcanoes (including submarine and hydrothermal volcanoes) (Taylor and Nesbitt, 1998; Baker et al., 2008). The structure of the subducting slabs beneath the IBM sys- tem have been investigated in a number of different studies. The ﬁrst information on the shape of the subduction zone was derived from the analysis of depth distribution of seismicity in the Benioff zone (e.g., Katsumata and Sykes 1969; Eng- dahl et al., 1998). More details on the slab structure beneath the IBM area were found in a number of seismic models on the global (e.g., Bijwaard et al., 1998) and regional scales (Widiyantoro, 1999; Widiyantoro and van der Hilst; 1996, Van der Hilst and Seno; 1993, Gorbatov et al., 2003; Miller et al., 2005, 2006a, b, 2011) mostly constructed by a tomo- graphic inversion of body wave travel times. These studies reveal generally consistent information on the shape of the subducting slab, but in details they appear to be considerably different. For example, the penetration of the slab into the lower mantle is still actively debated, and the behaviour of the slab in the contact zone between Izu-Bonin to the Mar- iana segments is not clear yet. The inconsistencies between different models are partly due to the fact that seismic to- mography based on noisy earthquake data is not a perfect tool; often processing of same data by different authors leads to considerably different results. Therefore, any attempt to compute a new model using independent approaches is an important step to corroborate the existing information. volcanism properties and dip angles of slabs may abruptly change along the strike of the subduction zone. There are many open questions related to subduction zones and, to an- swer them, new robust and reliable data on the deep structure are still required. K. Jaxybulatov et al.: Segmentation of the Izu-Bonin and Mariana slabs 60 Fig. 1. The main geographic and tectonic units of the studied re- gion. Background is map of topography/bathymetry. The colours indicate the regions of the oceanic crust of different ages according to Peive, (1980): Pre-Cretaceous (green), Pre-Mesozoic (red), Pre- Miocene (yellow), Pre-Quaternary age (white), sea ﬂoor of Paciﬁc Ocean, Pre-Triassic (blue). White dotted lines depict arcs (ancient and present) (Peive, 1980). Names and ages of arcs are indicated. Arrows with numbers present the displacement rates in respect to Philippine plate (DeMets et al., 2010). The red triangles indicate ac- tive volcanoes (including submarine and hydrothermal volcanoes) (Taylor and Nesbitt, 1998; Baker et al., 2008). as the Kyushu-Palau Ridge with the age of volcanism 37– 25 million years (Peive, 1980; Berckhemer et al., 1982), the West Mariana Ridge with the age of volcanism 23–10 million years and Bonin Ridge. There are also inactive remnants of rift systems, such as the Shikoku Basin, Parece Vela Basin (extension age 25–15 Ma) and the West Philippine Basin (with the age of oceanic plate 100–150 Ma). There is also a presently active back-arc spreading basin of the Mariana Trough, where the extension started about 3 million years ago (Peive, 1980, Fig. 1) and occur at a rate of 15–30 mm yr−1 (Kato et al., 2003). The age of the Paciﬁc Plate along the IBM arc system is generally old and it varies from 146 Ma in the northern segments to 155 Ma at the Mariana trench (e.g., M¨uller et al., 1997). There are more than 20 volcanic islands along the Izu- Bonin arc and more than 70 along the Mariana arc (Tay- lor and Nesbitt, 1998; Baker et al., 2008). Most of the vol- canoes along the IBM arc system are submarine. Figure 1 shows the location of the volcanoes and tectonic units of dif- ferent ages. The volcanoes of the IBM system are generally characterised by low dacite and andesite content (Kelemen et al., 2004) and a frequent occurrence of the boninite series, which were ﬁrst discovered and described in this region (Pe- tersen, 1891; Crawford et al., 1981). These composition fea- tures make the IBM volcanoes particular compared to other subduction zones of the world. Published by Copernicus Publications on behalf of the European Geosciences Union. Published by Copernicus Publications on behalf of the European Geosciences Union. K. Jaxybulatov et al.: Segmentation of the Izu-Bonin and Mariana slabs 2 Data processing and algorithm In this study, we use the travel times of P and S waves from the International Seismological Centre (ISC) for the time pe- riod from 1964 to 2007. The advantage of the ISC dataset is a long registration period and the global data coverage which cannot be achieved by any regional network. At the same time, the data quality in the initial ISC catalogue is low. The reported coordinates of events are strongly biased due to out- dated location algorithm and a big number of outliers in the data. Therefore, before using for tomographic inversion, this catalogue should be reprocessed. Most of scientists use a re- vised EHB catalogue which was created from careful reﬁn- ing of the ISC data by Engdahl et al. (1998). For our stud- ies, we perform our own revision in which the rejection of outliers and relocation of sources were performed using the algorithm described in (Koulakov and Sobolev, 2006). This revised catalogue has been constructed based on the 1D ref- erence model AK 135 (Kennett at al., 1995) and takes into account the ellipticity of the Earth and variable crustal thick- ness according to the global model CRUST2.0 (Bassin et al., 2000). Our approach is less conservative compared to the one used by Engdahl et al. (1998), and it keeps more data, which is favourable for tomographic inversion. y p y The behaviour of the slab beneath the IBM area has been actively discussed in a series of papers by Miller et al. (2005, 2006a, 2006b). They reconstructed the position of the trench in the last 30 Ma and explained the variable shapes of the slab along the IBM arc system by the mutual displacements of the trench and the oceanic plate. The bend around the con- tact zone between Izu-Bonin and Mariana arc is explained by presence of highly buoyant segment in the incoming plate with large igneous Ogasawara Plateau. They explain the ver- tical shape of the Mariana slab by forward displacement of the trench (ﬁgure 10 in Miller et al., 2006b). There are many studies showing the evidence of eastward displacement of the Mariana segment of the trench (Seno and Maruyama, 1984; Seno et al., 1993; Hall et al., 1995a, 1995b; Hall, 2002). 1.3 Kinematics of IBM area of the trench at 25–5 Ma and a forward displacement of the trench at 5–0 Ma. Faccena et al. (2009) show that the trench migration depends on the pull force of the slab, which in- creases when the lithosphere is older. The trench advancing occurs in the case of subduction of older and stronger oceanic lithosphere. Taking into account this effect (also considering other parameters including width and rheology of the litho- sphere) the authors suggest that recent advancing of the IBM trench system occurred after a long-term history of the trench retreat. In turn, the trench displacement (advancing and re- treating) controls the subduction style and the slab shape. It was shown in several papers (Van der Hilst and Seno, 1993; Heuret et al., 2007; Faccena et al., 2009) that in the case of the trench retreating, the subduction becomes shallower, whereas the trench advancing causes steeping of the slab of even backward turning if the bottom of the slab is anchored in the transition zone. The main purpose of this paper is to compute a new 3-D seis- mic model of the mantle beneath the IBM system and to use it to reconstruct the plate interactions and the subduction de- velopment in this area. This reconstruction needs robust in- formation about the plate motions in the region of interest including the surrounding areas in a large extent. We found that the data on the rates of plate motions in and around the Philippine plate given by different authors are often not con- sistent and even contradictory. One of the most important problems is deﬁning an appropriate reference frame which is necessary for accessing the subduction motion with re- spect to the lower mantle. For example, when considering the Philippine plate as a reference point, De Mets et al. (2010) show that the extension along the backarc Mariana Trough causes eastward displacement of the Mariana trench at about 2 cm yr−1 that forms a loop shaped arc (Fig. 1). In this ref- erence frame, the rate of the Paciﬁc plate westward motion near the trench with respect to the Philippine Plate is 4.9 and 4.0 cm yr−1 in the northern and southern segments of the Izu-Bonin arcs; in the Mariana arc it changes from 2.8 to 1.6 cm yr−1 when moving southward (DeMets et al., 1994, 2010). 1.3 Kinematics of IBM area On the other hand, there are several frames based on different reference points which provide considerably differ- ent absolute rates of the plate motions in the IBM region. For example, Schellart et al. (2008) presents an overview of plate motions related to three reference frames which are re- lated to different geological structures. Based on all these frames, they report a westward absolute displacement for the Mariana subduction of both subducting and overriding plates at the rates of 11.8–5 cm yr−1, 8.1 – 1.8 cm yr−1 and 7.1 – 1.5 cm yr−1, respectively. In this study, we present a new seismic model of the man- tle seismic structure down to 1200 km depth beneath the IBM arc system. For this model, we use a more recent version of the ISC catalogue compared to the previous studies which contains considerably larger data amount. We use our own al- gorithm which has some particularities compared to the pre- vious studies. We present several tests which demonstrate the limitations of the resolution and support the robustness of the derived structures. In this paper, we compare our model with the existing ones and discuss possible geodynamical inter- pretation of the results. 1.1 Geological overview Izu-Bonin-Mariana (IBM) arc system is located in the west- ern Paciﬁc (Fig. 1). The arc length is more than 2800 km and it extends from north to south, from Tokyo Bay to Guam. This intraoceanic convergent zone is the result of a multi- stage subduction of the Paciﬁc plate beneath the Philip- pine plate. There are the remnants of ancient arcs, such www.solid-earth.net/4/59/2013/ www.solid-earth.net/4/59/2013/ Solid Earth, 4, 59–73, 2013 61 3 Slab-related seismicity We show the distribution of seismicity from the revised ISC catalogue projected to the vertical proﬁle along the IBM sys- tem in Fig. 2. The map view of the proﬁle location and the band, from which the events were considered, is shown in Fig. 2a. The projection of the events to the proﬁle was per- formed in each zone along the lines with proportional angles in respect to the orientations of opposite sides of the trapez- ium. We present both the individual events with magnitude indications (Fig. 2b) and the integral released energy in cells 50 × 50 km (Fig. 2c). Plots in Fig. 2 display some important features which can be used for geodynamic interpretation. The paradox conclu- sion which can be drawn from this representation of the seis- micity distribution is that the events located below 200 km depth beneath the IBM arc system form rather vertically ori- ented clusters than horizontal ones. This seems to contra- dict a point of view that seismicity in the slab reﬂects phase transitions which occur at certain depth. So we would rather expect horizontal clusters around 100 km, 200 km and other depth levels where the material of the slab is transformed. They can be seen locally, but less prominently than thin ver- tical zones of seismicity. The most clear are vertical clusters of seismicity around the distances of 700, 1380 and 1910 km. The later cluster corresponds to the Mariana segment with the seismicity along a relatively narrow vertical “column” of about 700 km high and only 100–200 km width. In this study, the tomographic inversion is performed sep- arately in three overlapping circular windows with the radius of 10◦, covering the study area. For the IBM area we use the travel times from the revised ISC catalogue which mostly correspond to events located in the study region which were recorded by worldwide stations at all possible epicentral dis- tances. The opposite scheme, with stations in the target area recording the teleseismic events, is almost not used in this case, because of the lack of seismic networks in the study region. The total amount of used rays for the entire region is more than 2 million. The amount of P data is about ten times larger than that of the S-data. K. Jaxybulatov et al.: Segmentation of the Izu-Bonin and Mariana slabs K. Jaxybulatov et al.: Segmentation of the Izu-Bonin and Mariana slabs 62 Regional tomographic inversion in this study is based on the approach proposed by (Koulakov et al., 2002). In later works, this algorithm was modiﬁed and used to study the upper mantle structure beneath some regions (see, for ex- ample, Koulakov et al., 2009a; Koulakov, 2011). This to- mographic inversion algorithm is based on the linearised approach which implies that all calculations are performed on the basis of rays constructed in one-dimensional model AK135 (Kennett at al., 1995). It should be noted that there is no theoretical difﬁculty in using a nonlinear algorithm with a three-dimensional ray tracing (for example, as used in the LOTOS code by Koulakov, 2009b). However, we do not be- lieve that implying the iterative nonlinear inversions would bring principal changes in the results. First, the recovered amplitudes of anomalies in the upper mantle are relatively low (2–3 %) and they do not really lead to a considerable nonlinear effect. Second, the amplitudes derived from the in- version of the ISC data are usually lower than expected in nature because of the large damping which should be applied to reduce the noise-related instability. In this case, even if we iteratively trace the rays in 3-D, they might appear not signiﬁcantly closer to the true paths than the 1D based rays. Third, the linear approach is fast, and it gives us a possibility for performing many trials based on real and synthetic data. This allows us to perform thorough selection of free inver- sion parameters which would be hard in the case of a single nonlinear inversion. ordinates and origin times). The damping coefﬁcients and weighting of different parameters were estimated based on results of synthetic modelling. The inversion of the matrix is performed using the LSQR algorithm (Page and Saunders, 1982; Van der Sluis and Van der Vorst, 1987). The main model was obtained by combining 12 separate inversion re- sults (4 models in each of the three circular areas). 2 Data processing and algorithm Al- ternatively, the westward displacement of the trench is shown by (Carlson and Mortera-Gutierrez, 1990; Schellart et al., 2008; O’Neill et al., 2005; Kreemer et al., 2003; Gripp and Gordon, 2002). It was shown, by several authors, that the trench had several stages of migration at different phases of developing this subduction zone. For example, Sdrolias and Muller (2006) show three main stages of Mariana trench with a stable trench at 50–25 Ma, a backward displacement Solid Earth, 4, 59–73, 2013 www.solid-earth.net/4/59/2013/ K. Jaxybulatov et al.: Segmentation of the Izu-Bonin and Mariana slabs 3 Slab-related seismicity Note that using this ray conﬁguration does not allow obtaining high resolution in tomography reconstruction for the shallowmost sections (above ∼50 km depth) where the ray coverage is poor and there is a strong trade-off between velocity and source pa- rameters. Another striking feature is a large seismicity cluster related to the lower part of the northern segment of the Izu-Bonin plate. Between 0 and 900 km along the proﬁle, it tends to dip from 200 km to 300 km depth for the upper boundary of the cluster, and from 400 km to 500 km for the lower bound- ary. The southern limit of this body roughly ﬁt to the vertical cluster at ∼700 km distance. These observations possibly indicate that along the pro- ﬁle there are at least four separate segments of the subduct- ing plate which may behave differently. The ﬁrst northern segment of the Izu-Bonin plate is represented by gently dip- ping plate with very active seismicity in the depth interval between 300 and 500 km. The next two segments of the Izu- Bonin arc are limited by vertical clusters at 700 km, 1379 km and 1912 km. Unlike the ﬁrst segment, for these two seg- ments, no signiﬁcant deep seismicity is observed. The cluster in the middle at 1379 km is observed down to about 350 km in depth, which may imply that these two segments are only For the inversion, the 3-D velocity distribution is parame- terised with a set of nodes which are distributed in 15 depth levels (50, 100, 150, 220, 290, 360, 430, 500, 570, 650, 710, 800, 900, 1000, 1300 km). In each level, the nodes are in- stalled according to the ray coverage; no nodes are set in areas with insufﬁcient data amount. To minimise any arti- facts related to grid conﬁguration, the velocity models are calculated independently for four differently oriented grids with basic orientations at 0◦, 22◦, 45◦and 67◦and then av- eraged. The inversion was performed simultaneously for P- and S- velocity anomalies and for the source parameters (co- www.solid-earth.net/4/59/2013/ www.solid-earth.net/4/59/2013/ Solid Earth, 4, 59–73, 2013 K. Jaxybulatov et al.: Segmentation of the Izu-Bonin and Mariana slabs K. Jaxybulatov et al.: Segmentation of the Izu-Bonin and Mariana slabs 63 Fig. 2. (A) Distribution of earthquakes in map view. Coloured dots depict earthquake hypocentres. Segments indicate the area in which the seismicity was considered. 3 Slab-related seismicity Line in the centre of the segments is the location of vertical proﬁle for plots B and C. (B) The distribution of earthquakes in the vertical cross-section along the arc; colours and sizes of dots indicate the magnitude of events. (C) Total seismic energy release along the same vertical section in 50 × 50 km cells. Black lines in Plots B and C highlight the cluster of earthquakes discussed in the text. Dotted line depicts the bottom of the seismogenic zone beneath the Mariana Arc. Fig. 2. (A) Distribution of earthquakes in map view. Coloured dots depict earthquake hypocentres. Segments indicate the area in which the seismicity was considered. Line in the centre of the segments is the location of vertical proﬁle for plots B and C. (B) The distribution of earthquakes in the vertical cross-section along the arc; colours and sizes of dots indicate the magnitude of events. (C) Total seismic energy release along the same vertical section in 50 × 50 km cells. Black lines in Plots B and C highlight the cluster of earthquakes discussed in the text. Dotted line depicts the bottom of the seismogenic zone beneath the Mariana Arc. The major feature which is clearly seen in horizontal and vertical sections is a slab-related high-velocity anomaly ob- served beneath the entire arc, except for the shallowmost part in the Mariana region where the resolution is poor. This anomaly is generally consistent in P and S model, although the amount of the S data was about ten times smaller than that of P. The correlation between the P and S models is an infor- mal argument for the reliability of the model; on a regional scale, the major geological structures, such as subduction, usually affect similarly the P and S velocities. Note that this anomaly ﬁts to the deep seismicity distribution in the Benioff zone (see Fig. 4). separated in the upper part and united in the lower part. The southernmost segment corresponds to the Mariana plate. Most of seismicity in the Mariana segment is concentrated in a relatively narrow zone around the mark at 1912 km. Be- tween 2087 km and 3193 km along proﬁle the Mariana Plate demonstrates very minor deep seismicity with decreasing depths (from 500 km to 200 km). 4.1 Computed P- and S-velocity models By observing the vertical and horizontal sections, it helps to determine the shape of the subducting slab. In vertical Sect. 1, we observe the southern part of the Japan arc. The amount and the quality of data in this area are maximal, therefore, the image seems to be well resolved. Section 2 roughly crosses the transition zone of Japan arc to Izu-Bonin. In these sections, the shape of the slab is well repeated in P- and S-anomalies, excluding some features in deepest parts. The recovered slab-related anomaly ﬁts to the distribution of The derived models of seismic P- and S-anomalies are shown in seven horizontal sections (Fig. 3) and in 9 vertical sections for Izu-Bonin-Mariana arc (Fig. 4). The areas, where the dis- tance to the nearest grid node is more than 80 km, are shaded; in these areas the data coverage is not sufﬁcient for the tomo- graphic inversion. The variance reduction after the inversion was more than 50 % which shows relatively high level of co- herent signal. Solid Earth, 4, 59–73, 2013 4.2 Comparison to previous tomography models deep seismicity. In these sections, the slab gently dips with the angle of about 35◦. For Sect. 3, the slab becomes steeper, but in the transition zone between 400 and 600 km depth it turns to be horizontal. In this section, an enigmatic feature is a prominent positive anomaly in the lower mantle which is clearly seen in both P and S models, though with differ- ent shapes. Note that in the neighbouring Sects. 2 and 4, this anomaly is not observed. This anomaly is not distinctly de- tected and discussed in other early tomography models, but in certain papers (e.g., Widiyantoro et al., 1999; Miller et al., 2006) anomalies penetrating into the lower mantle and some- times separate anomalies lower than 660 km are observed. So presence of this feature can be taken into account. The derived model generally agrees with other previously published regional models (Bijwaard et al., 1998; Widiyan- toro, 1999; Van der Hilst and Seno, 1993; Widiyantoro and Van der Hilst, 1996; Gorbatov et al., 2003; Miller et al., 2005, 2006). In Fig. 5, we show an example of comparing of our tomographic images with the model by Miller et al. (2005) which is one of the most recent results for this region. Same as in most of previously published models, our model re- veals almost vertical dipping slab beneath the Mariana arc which penetrates into lower mantle. Generally consistent re- sults are also obtained for the Izu-Bonin arc where shallower subduction is observed. However, in details our model has some particular features. First, our tomographic images pro- vide sharper images of the slab than the previous models. Second, we observe more clearly the penetration of a part of the slab beneath Izu-Bonin arc into the lower mantle (Fig. 4, Sect. 3). Third, tearing between Izu-Bonin and Mariana slab segments down to depths of 500 km is clearly observed in horizontal sections. In contrast to the results by Miller et al. (2005, 2006a, b) who observed the gap in the subduc- tion segments beneath Izu-Bonin and northern Mariana arcs, we can see the continuous transition of the subduction style (see the shapes of the fast anomaly between 400–450 km in In Sects. 4 and 5, the dipping part of the slab becomes steeper and reaches the angle of 60◦–70◦. The horizontal part of the slab above the transition zone is still observed here. www.solid-earth.net/4/59/2013/ K. Jaxybulatov et al.: Segmentation of the Izu-Bonin and Mariana slabs 64 Fig. 3. The results of the real data inversion. The distribution of the P- and S-anomalies are shown at different depths. The purple line indicates tectonic boundaries according to Fig. 1. Fig. 3. The results of the real data inversion. The distribution of the P- and S-anomalies are shown at different depths. The purple line indicates tectonic boundaries according to Fig. 1. www.solid-earth.net/4/59/2013/ 4.2 Comparison to previous tomography models In section 5, the dipping part of the slab disappears and in the map view, we see that the section passes through the gap area between the Izu-Bonin and Mariana plates. In Sects. 8 and 9, we observe the bright images of the Mariana slab which deepens almost vertically. Based on these results, we cannot say deﬁnitively if the slab penetrates into the lower mantle or not. In Sect. 9, the slab seems to stop at 700 km depth, whereas in Sect. 8 for P- and S-anomalies, we observe high- velocity anomaly below 700 km depth. Solid Earth, 4, 59–73, 2013 www.solid-earth.net/4/59/2013/ K. Jaxybulatov et al.: Segmentation of the Izu-Bonin and Mariana slabs 65 Fig. 4. The results of real data inversion. P-and S-anomalies are presented in vertical cross-sections across the Izu-Bonin-Mariana arc; locations of sections are shown in the map on the right. Black dots are the earthquake hypocentres located at a distance of no more than 50 km from the proﬁle. K. Jaxybulatov et al.: Segmentation of the Izu-Bonin and Mariana slabs 65 Fig. 4. The results of real data inversion. P-and S-anomalies are presented in vertical cross-sections across the Izu-Bonin-Mariana arc; locations of sections are shown in the map on the right. Black dots are the earthquake hypocentres located at a distance of no more than 50 km from the proﬁle. Fig. 5. Example of comparing our tomographic images with the model by Miller et al. (2005). Locations of sections are shown in the map on the left. Both P- and S-anomalies are presented. Fig. 5. Example of comparing our tomographic images with the model by Miller et al. (2005). Locations of sections are shown in the map on the left. Both P- and S-anomalies are presented. Solid Earth, 4, 59–73, 2013 www.solid-earth.net/4/59/2013/ www.solid-earth.net/4/59/2013/ 66 K. Jaxybulatov et al.: Segmentation of the Izu-Bonin and Mariana slabs K. Jaxybulatov et al.: Segmentation of the Izu-Bonin and Mariana slabs 66 Fig. 6. Test with even and odd numbers of sources to estimate the level of random noise. The distribution of P- and S-anomalies are shown at different depths. The upper and lower rows correspond to results with odd and even numbers of sources, respectively. Fig. 6. Test with even and odd numbers of sources to estimate the level of random noise. The distribution of P- and S-anomalies are shown at different depths. Fig. 5). Finally, we present more different tests to prove the reliability of our results than in any of previous studies for this region. To evaluate the spatial resolution of the model, we have performed the synthetic checkerboard test. The synthetic model was deﬁned as alternating positive and negative rect- angular anomalies. For the P model the horizontal size of the cell was 1◦× 1◦with an empty space of 0.5◦; for the S model the corresponding parameters were 2◦× 1.5◦and 1◦. With the depth, the sign of anomalies alternated every 200 km. To produce the synthetic dataset, we used the same rays as in the case of real data inversion. The synthetic times were perturbed with random noise with the statistical distri- bution which was estimated after analysis of real residuals in the revised ISC catalogue. The magnitude of the noise was 0.5 s and 0.8 s for the P- and S-data, respectively, which lead to the same variance reduction after inversion (about 50 %) as in the real data case. When reconstructing the synthetic model, the values of the inversion parameters were the same as in the case of real data inversion. Figure 7 displays the re- construction of the checkerboard model at different depths. It shows that in the central part of the area, where the slab is located, the resolution is fair. However, below 700 km in depth the resolution becomes poorer and does not allow the recovering of such small-scale features. 4.2 Comparison to previous tomography models The upper and lower rows correspond to results with odd and even numbers of sources, respectively. Fig. 5). Finally, we present more different tests to prove the reliability of our results than in any of previous studies for this region. 4.3 Veriﬁcation As mentioned previously, the data of ISC catalogue are very noisy. Despite the efforts on pre-processing the data and re- jection of outliers, we should confess that noise is still impor- tant and it can signiﬁcantly affect the inversion. To estimate the inﬂuence of random noise upon the inversion results, we performed the “odd/even” test which consists in independent inversions of two equal subsets of data (e.g., with odd and even numbers of events) using the same inversion parameters as in the case of the entire dataset. If noise effect is signiﬁ- cant, it would generate random anomalies which would be different in two independent results; they should be ignored or smoothed by increasing the damping. The results of this test shown in Fig. 6 demonstrate fairly high consistency of the main patterns which conﬁrm their independence of ran- dom noise in the data. Solid Earth, 4, 59–73, 2013 www.solid-earth.net/4/59/2013/ K. Jaxybulatov et al.: Segmentation of the Izu-Bonin and Mariana slabs 67 K. Jaxybulatov et al.: Segmentation of the Izu-Bonin and Mariana slabs 67 Fig. 7. Synthetic test “Checkerboard”. The distribution of P- and S-anomalies are shown at different depths. The sizes of synthetic anomalies are: 1◦× 1◦for the P-model with the distance between the anomalies of 0.5◦and 2◦× 1.5◦for S-model with the distance between the anomalies of 1◦. Fig. 7. Synthetic test “Checkerboard”. The distribution of P- and S-anomalies are shown at different depths. The sizes of synthetic anomalies are: 1◦× 1◦for the P-model with the distance between the anomalies of 0.5◦and 2◦× 1.5◦for S-model with the distance between the anomalies of 1◦. Fig. 7. Synthetic test “Checkerboard”. The distribution of P- and S-anomalies are shown at different depths. The sizes of synthetic anomalies are: 1◦× 1◦for the P-model with the distance between the anomalies of 0.5◦and 2◦× 1.5◦for S-model with the distance between the anomalies of 1◦. especially for the S-model, however it is retrieved at a correct place. An important test consists in reconstruction of a synthetic model with realistic conﬁguration of anomalies. In Fig. 8 we present a model which reproduces the conﬁguration of the slab in horizontal and vertical sections. The synthetic anoma- lies in this case are free-shaped anomalies along the verti- cal sections. They have ﬁxed shape in the direction across the section within a certain band. This method allows deﬁn- ing the realistic shapes of different segments of the slab. 5 Discussion The main feature of the obtained model is the slab-related high-velocity anomaly which is clearly observed in all hori- zontal sections of P and S model, except for the uppermost part where it is not resolved for the Mariana area. In verti- cal sections we observe complex geometry of the subducting plate which changes the shape along the Izu-Bonin and Mar- iana arcs. Based on these results in Figs. 9 and 10 we propose two scenarios based on different reference frames which ex- plain the present shape of the slab. 4.3 Veriﬁcation In this model, the Izu-Bonin slab is gently dipping and becomes horizontal in the transition zone. On the contrary, The Mari- ana slab consists of three segments and almost vertically dips down to 1200 km. Between the Mariana and Izu-Bonin arcs we deﬁne a small gap. As in the case of the checkerboard test, the synthetic data were perturbed with random noise of 0.5 and 0.8 s magnitude for P and S data, respectively. The re- sults of the reconstruction in horizontal and vertical sections show that the main patterns in the mantle are correctly recon- structed. However, the uppermost structures in the Mariana area are strongly biased because of lack of data and trade- off between velocity structures and source parameters. In the lower mantle, the slab-related anomaly is strongly smeared, 5.1 Scenario #1 The ﬁrst scenario shown in Fig. 9 corresponds to the refer- ence point located in the middle of the Philippine Sea Plate. According to (Seno and Maruyama, 1984; Seno et al., 1993; Hall et al., 1995a, b; Hall, 2002), presently curved shape of www.solid-earth.net/4/59/2013/ Solid Earth, 4, 59–73, 2013 68 K. Jaxybulatov et al.: Segmentation of the Izu-Bonin and Mariana slabs ith slab-shaped anomalies. The results of inversion are shown in horizontal and vertical sec p corresponding to the model deﬁnition at 90 km depth. Fig. 8. Synthetic tests with slab-shaped anomalies. The results of inversion are shown in horizontal and vertical sections. Locations of the proﬁles are shown in map corresponding to the model deﬁnition at 90 km depth. www.solid-earth.net/4/59/2013/ www.solid-earth.net/4/59/2013/ Solid Earth, 4, 59–73, 2013 K. Jaxybulatov et al.: Segmentation of the Izu-Bonin and Mariana slabs atov et al.: Segmentation of the Izu-Bonin and Mariana slabs 69 J y g Fig. 9. Reconstruction of the subduction evolution in different segments of the Izu-Bonin and Mariana arcs. Left: map view of the area. Violet line is the present location of the trench; black line is a possible location of the trench about 12 Ma; dotted line is a possible location of the slab material at about 1000 km depth, background is P-velocity anomalies at 975 km depth; location of proﬁles for schematic representation are shown. Right: different regimes of subduction in four different segments of the arc. Color of patterns corresponds to the age: red is ∼12 Ma, violet is ∼6 Ma and blue is present. Bars with nimbers 1, 2 and 3 correspond to the location of the trench in the three episodes. Black dots depict the present seismicity around the corresponding segment; solid line is the boundary between the upper and lower mantle. PSP is the Philippine Sea Plate, PP is the Paciﬁc Plate, MT is the Mariana trough. Red dot indicates the reference point on the Philippine Sea Plate. Directions of displacements relative to PSP are shown with red arrows and numbers in mm/yr. Fig. 9. Reconstruction of the subduction evolution in different segments of the Izu-Bonin and Mariana arcs. Left: map view of the area. Violet line is the present location of the trench; black line is a possible location of the trench about 12 Ma; dotted line is a possible location of the slab material at about 1000 km depth, background is P-velocity anomalies at 975 km depth; location of proﬁles for schematic representation are shown. Right: different regimes of subduction in four different segments of the arc. Color of patterns corresponds to the age: red is ∼12 Ma, violet is ∼6 Ma and blue is present. www.solid-earth.net/4/59/2013/ It can be seen that since that time in different segments of the trench there were both backward and forward displace- ments. Note that a segment with the forward displacement corresponds to the presence of the Ogasawara large igneous plateau. The change of the subduction shape roughly corre- sponds to transitions between different displacement styles of the trench. Based on the results of tomographic inversion we propose our back-time reconstruction to about 10–12 Ma. In sections to the right of Fig. 9, which roughly correspond to vertical Sects. 2, 3, 4 and 9 in Fig. 4, we present the hypo- thetical evolution of subduction in different segments of IBM For Section A1–A2 we expect a considerable backward displacement of the trench due to eastward movement of Asia. In this section the oceanic lithosphere seems to be united with a large segment beneath Japan where the slab is relatively shallow and turns to be horizontal in the tran- sition zone to a large extent (Zhao et al., 2009). The reason why this old and dense lithosphere beneath Japan behaves in this way is actively discussed in the literature (e.g., Zhao et al., 2009). It might be due to “roof” shape of the slab which gives additional straight and prevents steep bending of the plate. Moving back of the trench in this area causes decreas- ing the dipping angle of the slab, as shown in Section A1–A2 in Fig. 9. Solid Earth, 4, 59–73, 2013 www.solid-earth.net/4/59/2013/ Bars with nimbers 1, 2 and 3 correspond to the location of the trench in the three episodes. Black dots depict the present seismicity around the corresponding segment; solid line is the boundary between the upper and lower mantle. PSP is the Philippine Sea Plate, PP is the Paciﬁc Plate, MT is the Mariana trough. Red dot indicates the reference point on the Philippine Sea Plate. Directions of displacements relative to PSP are shown with red arrows and numbers in mm/yr. the IBM trench (violet line in Fig. 9) was much straighter in the recent past (e.g., 12–25 Ma) as schematically shown with black line in left plot in Fig. 9. An indirect argument for this hypothesis is the nearly linear distribution of high-velocity anomalies at 975 km depth shown in Fig. 9 which probably represent the curvature of the trench in the past (dotted line). Taking the present rate of subduction (∼5 cm yr−1) this depth corresponds to the location of the trench about 20–25 Ma. It can be seen that since that time in different segments of the trench there were both backward and forward displace- ments. Note that a segment with the forward displacement corresponds to the presence of the Ogasawara large igneous plateau. The change of the subduction shape roughly corre- sponds to transitions between different displacement styles of the trench. Based on the results of tomographic inversion we propose our back-time reconstruction to about 10–12 Ma. In sections to the right of Fig. 9, which roughly correspond to vertical Sects. 2, 3, 4 and 9 in Fig. 4, we present the hypo- thetical evolution of subduction in different segments of IBM arc system. We propose that initially the subduction occurred along a weakly curved trench and the shape of the slab was generally similar throughout the IBM system (namely with the dipping angles of 40–50 degrees). the IBM trench (violet line in Fig. 9) was much straighter in the recent past (e.g., 12–25 Ma) as schematically shown with black line in left plot in Fig. 9. An indirect argument for this hypothesis is the nearly linear distribution of high-velocity anomalies at 975 km depth shown in Fig. 9 which probably represent the curvature of the trench in the past (dotted line). Taking the present rate of subduction (∼5 cm yr−1) this depth corresponds to the location of the trench about 20–25 Ma. www.solid-earth.net/4/59/2013/ 70 K. Jaxybulatov et al.: Segmentation of the Izu-Bonin and Mariana sl Fig. 10. Plate kinematics along two vertical proﬁles Red circles indicate the locations where absolute velocities are close to zero O K. Jaxybulatov et al.: Segmentation of the Izu-Bonin and Mariana slabs 70 70 K. Jaxybulatov et al.: Segmentation of the Izu-Bonin and Mariana slab Fig. 10. Plate kinematics along two vertical proﬁles. Red circles indicate the locations where absolute velocities are close to zero. OT Okinawa trough, PPS – Philippine Sea Plate, MT – Mariana trough, PP – Paciﬁc plate. Relative divergence and convergence movemen in spreading and subduction zones, respectively, are indicated with black arrows. Directions of absolute displacements are shown with r arrows and approximate rates are given in mm yr−1 according to different sources (see the text). For Section B1 B2 the situation seems to be signiﬁcantly strength of the lithosphere and made it more buoyant That Fig. 10. Plate kinematics along two vertical proﬁles. Red circles indicate the locations where absolute velocities are close to zero. OT – Okinawa trough, PPS – Philippine Sea Plate, MT – Mariana trough, PP – Paciﬁc plate. Relative divergence and convergence movements in spreading and subduction zones, respectively, are indicated with black arrows. Directions of absolute displacements are shown with red arrows and approximate rates are given in mm yr−1 according to different sources (see the text). For Section B1–B2, the situation seems to be signiﬁcantly different. Here the shallow part of the slab seems to be de- tached from the horizontal body in the transition zone. In this part of the arc the trench remains at the same posi- tion. The continued subduction at a rate of about 4 cm yr−1 (DeMets et al., 2010) causes accumulation of the slab mate- rial in the transition zone (stage 2). When reaching a critical volume, the accumulated denser material starts to descend to the lower mantle. strength of the lithosphere and made it more buoyant. That is why, when this plateau reached the trench, it stopped the sub- duction and leaded to displacement of the trench along the incoming plate movement. In this case, the anchored lower part was ﬁxed, and the displacement of the upper part caused the steepening of the slab. The last Section D1–D2 corresponds to the vertical Sect. 9 (Fig. 4) which passes across the Mariana segment. Accord- ing to Miller et al. K. Jaxybulatov et al.: Segmentation of the Izu-Bonin and Mariana slabs which may be kept in plume traces beneath the plateau. The isolated heavy Mariana plate cannot behave as the northern- most part of the Izu-Bonin plate which is supported by the Japanese shallow slab. Even if in stage 1 the subduction had a normal dipping angle (e.g., 45◦), in stage 2, the shallow part of the slab had to be detached from the deep one, and a new subduction zone originated. The newly formed subduction is mostly driven by high density of the subducted lithosphere which pulls it down. In this case, it behaves not as a normal subduction: the descending of the plate occurs due to both ocean plate movement and trench displacement, as shown in section D1–D2. This is supported by relatively slow rate of the subduction (∼2 cm yr−1) and an approximately same rate of the slab retreat in the Mariana trench (DeMets, 2010). The remnant part of the previous subduction zone could descend to the transition zone and then further could continued sink- ing to the lower mantle. A possible location of sinking of the older subducted slab is in the area, where the resolution of the tomographic model is poor. However, a strongly smeared anomaly of high P-velocity model in the lower mantle can be interpreted as a trace of the previous subduction and it would ﬁt to this hypothesis. Mariana troughs). The rates of shortening and extension in these zones are given as average estimates of several results given in different studies (Suppe J., 1984; Kato et al., 2003; Hsu et al., 2012). The rates of the plates between these zones are presumed to be constant. If we assume that the Asia is ﬁxed in respect to the lower mantle, the entire system in- cluding Philippine Sea and Paciﬁc Plates moves westward, as predicted by Schellart et al. (2008). In this case, the fastest Paciﬁc Plate moves at a rate of 12 cm yr−1. The lower part of the subducting Paciﬁc slab seems to be anchored in the lower mantle. Westward migration of the trench causes steepening of the slab which may lead to vertical and even reverse ori- entation. A similar scenario has been proposed by Schellart (2011) based on numerical modeling. ( ) g The variability of the subduction styles leaded to clear seg- mentation of the Izu-Bonin and Mariana subduction zones with the boundaries marked with vertical seismicity clusters (Fig. 2). 5.2 Scenario #2 A weak point of the scenario shown in Fig. 9 is that ﬁx- ing the reference point in the Philippine Sea Plate needs a very fast eastward movement of Asia. To keep the diver- gence and convergence balances in subduction and spread- ing zones, the rate of eastward displacement of Asia should be about 5–6 cm yr−1, which is similar to the displacement rate of the Paciﬁc Plate. At the same time, much slower rel- ative displacements in continents compared to the fast rates of subduction and spreading processes in oceans suggest that the absolute movements of the entire continents are much slower that those of oceanic plates. Based of this statement, we propose an alternative scheme of plate movement corre- sponding to the ﬁxed location of Asia (Fig. 10). To estimate the total displacement balance between the Paciﬁc Plate and Asia, we need to consider the entire Philippine Plate includ- ing the Ryukyu subduction zone and the Okinawa backarc spreading in the west. The indicated velocities are compiled from different studies including (Suppe J., 1984; Sdrolias et al., 2006; Schellart et al., 2008; Hsu et al., 2012). For the northern part of the IBM system (Section A1–A2 in Fig. 9) we see that in the vertical section the slab penetrates at a high rate (more than 10 cm yr−1) between a ﬁxed Asian litho- sphere and lower mantle having zero velocity in this refer- ence frame. This movement is only possible in the case of a sufﬁciently thick layer in the transition zone between the slab and lower mantle playing the role of lubricants. This explains the shallow angle of dipping the slab and its nearly horizontal shape in the left side of the section. K. Jaxybulatov et al.: Segmentation of the Izu-Bonin and Mariana slabs Similarly as in Miller et al. (2006a), between the Mariana and Izu-Bonin segments we can see the gap corre- sponding to vertical Sect. 6 in Fig. 4 which is observed down to 400 km depth. Initially it was formed by the effect of the Ogasawara plateau. Further opening of this gap occurred due oblique orientation of the Paciﬁc Plate movement in respect to the northern part of the Mariana loop. The strike-slip dis- placement along the trench in this area causes the further de- tachment of the upper part of the slab. The lower part of the slab seems to be united and it probably corresponds to the stage when the trench was not perturbed by relative forward and backward displacements. A problem of this model is the existence of deep seismicity in the gap area where the high- velocity slab is not observed. We propose that the gap area is formed by plastic extension of the subducting plate. In this case the resolution of the tomography model might be not sufﬁcient to resolve the thinned part of the slab in the exten- sion zone. Furthermore, the presence of remnant heat beneath the Ogasawara plateau might lead to degrading some physi- cal properties of the slab and to decreasing seismic velocities. www.solid-earth.net/4/59/2013/ (2006b) and authors referred therein, the Mariana trench is moving backward, and initially the situa- tion was similar to Section A1–A2. The Mariana plate was detached from the Isu-Bonin plate by the Ogasawara plateau which served as a “knife” which cut the lithosphere. We pro- pose that this detachment was facilitated by remnant heat In Section C1–C2 we observe steepening of the slab. In this segment of the arc the trench is though to migrate for- ward (Schellart et al., 2008, Miller et al., 2006) which is probably caused by the presence of the large igneous Oga- sawara Plateau. It is possible that this plateau increased the www.solid-earth.net/4/59/2013/ Solid Earth, 4, 59–73, 2013 www.solid-earth.net/4/59/2013/ 71 K. Jaxybulatov et al.: Segmentation of the Izu-Bonin and Mariana slabs 6 Conclusions The main result of this study is the new model of seismic velocity heterogeneities of P-and S-waves in the mantle be- neath the Izu-Bonin and Mariana arcs which was constructed from tomographic inversion of global travel time data (ISC). This model was thoroughly veriﬁed using a series of differ- ent tests which showed generally high robustness of the re- sults. The main feature of this result is a slab-related anomaly which is similarly expressed in the P and S models. There is a general correlation with previous results derived by other authors; however in details we have detected some new fea- tures. We have found that the slab beneath IBM arc system has strongly variable dipping angle which generally corre- spond to the loops of the trench shape. We also analyze the distribution of the deep seismicity which was projected to the The second section includes two converging (Mariana and Ryukyu subductions) and two diverging zones (Okinawa and www.solid-earth.net/4/59/2013/ K. Jaxybulatov et al.: Segmentation of the Izu-Bonin and Mariana slabs 72 proﬁle along the trench. The striking feature of the seismic- ity distribution is that deep seismic events form clear verti- cal clusters which generally ﬁt to the limits of the IBM arc system segments. We propose that these clusters mark the boundaries between autonomous parts of the slab which may subduct independently with different rates. Carlson, R. L. and Mortera-Gutierrez, C. A.: Subduction hinge mi- gration along the Izu-Bonin-Mariana Arc, Tectonophysics, 181, 331–344, doi:10.1016/00401951(90)90026-5, 1990. Crawford, A. J., Beccaluva, L., and Serri, G.: Tectono-magmatic evolution of the West Philippine – Mariana region and the origin of boninites, Earth Planet. Sci. Lett., 54, 346–356, 1981. DeMets, C., Gordon, R. G., Argus, D. F., and Stein, S.: Effect of re- cent revisions to the geomagnetic reversal time scale on estimates of current plate motions, Geophys. Res. Lett., 21, 2191–2194, Bibcode 1994, GeoRL..21.2191D, doi:10.1029/94GL02118, 1994. We propose the two scenarios based on different reference frames which explain the present shape of subducting plates in the IBM system. The ﬁrst scenario explains the variable dipping angle of the slab beneath the IBM arc system by relative forward and backward displacements of the trench. In the second scenario, the entire system is moving west- ward, and the different angles of the slab can be explained by properties of the subducting lithosphere and the existence of rigid and buoyant blocks related to seamounts and intrusive plateaus. These two scenarios present two extreme cases and the truth might appear to be in the middle. Also, we aware about the uncertainties related to determination of relative rates in divergent and convergent zones which may be signif- icantly different in different studies. In addition, we consider the balances of displacements along 2-D proﬁles, whereas the displacements are clearly 3-D which may also affect our estimates. All these reasons show the importance of imply- ing quantitative estimates based on 3-D numerical modeling. In this case, the tomography results might serve for deﬁning the boundary conditions. DeMets, C., Gordon R. G., and Argus, D. F.: Geologically current plate motions, Geophys. J. Int., 181, 1–80, 2010. Dobretsov, N. L., Koulakov, I. Yu., and Litasov, Yu. D.: Mi- gration paths of magma and ﬂuids and lava compositions in Kamchatka, Russian Geology and Geophysics, 53, 1253–1275, doi:10.1016/j.rgg.2012.10.001, 2012. Engdahl, E. R., Van der Hilst, R. K. Jaxybulatov et al.: Segmentation of the Izu-Bonin and Mariana slabs D., and Buland, R.: Global tele- seismic earthquake relocation with improved travel times and procedures for depth determination, B. Seismol. Soc. Am., 88, 722–743, 1998. Faccenna, C., Di Giuseppe, E., Funiciello, F., Lallemand, S., and van Hunen, J.: Control of seaﬂoor aging on the migration of the Izu–Bonin–Mariana trench, Earth Planet. Sci. Lett., 288, 386– 398, doi:10.1016/j.epsl.2009.09.042, 2009. Gorbatov, A. and Kennett, B. L. N.: Joint bulk-sound and shear tomography for western Paciﬁc subduction zones, Earth Planet. Sci. Lett., 210, 527–543, 2003. Gripp, A. E. and Gordon, R. G.: Young tracks of hotspots and currentplate velocities, Geophys. J. Int., 150, 321–364, doi:10.1046/j.1365-246X.2002.01627.x, 2002. Acknowledgements. We are grateful to Wouter P. Schellart and Francesco P. Lucente for their valuable comments which forced us to explore deep the literature and to learn a lot of important information about the topic of this study. Their criticism was very constructive and important to improve the paper. The comments by the Topical Editor, Susanne Buiter, were very useful to ﬁnalize the paper. This study is supported by the Multidisciplinary project of the SB RAS #20, Project ONZ RAS #7.3, and the Joint research projects of SB RAS and the National Science Council of Taiwan. Hall, R., Ali, J. R., Anderson, C. D., and Baker, S. J.: Origin and motion history of the Philippine Sea Plate, Tectonophysics, 251, 229–250, 1995a. Hall, R., Fuller, M., Ali, J. R., and Anderson, C. D. : The Philippine Sea Plate: Magnetism and reconstructions, in Active Margins and Marginal Basins of the Western Paciﬁc, Geophys. Monogr. Ser., 88, edited by: Taylor, B. and Netland, J., 371–404, AGU, Wash- ington, D. C., 1995b. Hall, R.: Cenozoic geological and plate tectonic evolution of SE Asia and the SW Paciﬁc: Computer-based reconstructions, model and animations, J. Asian Earth Sci., 20, 353–431, 2002. Special Issue: “Subduction zones” Edited by: S. Buiter, F. Funiciello, and J. van Hunen Special Issue: “Subduction zones” Edited by: S. Buiter, F. Funiciello, and J. van Hunen Special Issue: “Subduction zones” Heuret, A., Funiciello, F., Faccenna, C., and Lallemand, S.: Plate kinematics, slab shape and back-arc stress : a comparison be- tween laboratory models and current subduction zones, Earth Planet. Sci. Lett., 256, 473–483, 2007. www.solid-earth.net/4/59/2013/ www.solid-earth.net/4/59/2013/ Solid Earth, 4, 59–73, 2013 K. Jaxybulatov et al.: Segmentation of the Izu-Bonin and Mariana slabs 73 O’Neill, C., Muller D., and Steinberger B.: On the uncertainties in hot spot reconstructions and the signiﬁcance of moving hot spot reference frames, Geochem. Geophys. Geosyst., 6, Q04003, doi:10.1029/2004GC000784, 2005. Kelemen, P. B., Hanghoi, K., and Greene, A. R.: One view of the geochemistry of subduction-related magmatic arcs, with an em- phasis on primitive andesite and lower crust, edited by: Holland, H., and Turekian, K., The Crust: edited by: Rudnick, R. L., Trea- tise on Geochemistry, Vol. 3. Elsevier-Pergamon, Oxford, 593– 659, 2004. Paige, C. C. and Saunders, M. A.: LSQR: An algorithm for sparse linear equations and sparse least squares, ACM trans. Math. Soft., 8, 43–71, 1982. Kennett, B. L. N., Engdahl, E. R., and Buland, B.: Constraints on seismic velocities in the Earth from travel times, Geophys. J. Int., 122, 108–124, 1995. Peive, A. V.: Geology of the Philippine Sea ﬂoor, “Nauka”, Moscow, 261 p., 1980. Koulakov, I., Tychkov, S., Bushenkova, N., and Vasilevskiy, A.: Structure and dynamics of the upper mantle beneath the Alpine- Himalayan orogenic belt from teleseismic tomography, Tectono- physics, 358, 77–96, 2002. Petersen J.: Der Boninit von Peel Island, Jahrb Hamburg, Wiss Anst 8, 341–349, 1891. Schellart, W. P.: A subduction zone reference frame based on slab geometry and subduction partitioning of plate mo- tion and trench migration, Geophys. Res. Lett., 38, L16317, doi:10.1029/2011GL048197, 2011. Koulakov, I. and Sobolev, S. V.: Tomographic Image of Indian Lithosphere Break-off beneath the Pamir Hindukush Region, Geophys. J. Int., 164, 425–440, 2006. Schellart, W. P., Stegman, D. R., and Freeman, J.: Global trench mi- gration velocities and slab migration induced upper mantle vol- ume ﬂuxes: Constraints to ﬁnd an Earth reference frame based on minimizing viscous dissipation, Earth-Sci. Rev., 88, 118–144, doi:10.1016/j.earscirev.2008.01.005., 2008. Koulakov I., Kaban, M. K., Tesauro, M., and Cloetingh, S.: P and S velocity anomalies in the upper mantle beneath Europe from tomographic inversion of ISC data, Geophys. J. Int., 179, 1, 345– 366, doi:10.1111/j.1365-246X.2009.04279.x, 2009a. Sdrolias, S. and Muller, R. D.: Controls on Back-arc Basin Formation, Geochem. Geophy. Geosy., 7, Q04016, doi:10.1029/2005GC001090, 2006. Koulakov I.: LOTOS code for local earthquake tomographic inver- sion. Benchmarks for testing tomographic algorithms, B. Seis- molog. Soc. Am., 99, 194–214, doi:10.1785/0120080013, 2009b. Seno, T. and Maruyama, S.: Paleogeographic reconstruction and origin of the Philippine Sea, Tectonophysics, 102, 53–84, 1984. K. Jaxybulatov et al.: Segmentation of the Izu-Bonin and Mariana slabs Koulakov, I.: High-frequency P and S velocity anomalies in the upper mantle beneath Asia from inversion of world- wide traveltime data, J. Geophys. Res., 116, B04301, doi:10.1029/2010JB007938, 2011. Seno, T., Stein, S., and Gripp, A. E.: A model for the motion of the Philippine Sea Plate consistent with NUVEL-1 and geological data, J. Geophys. Res., 98, 17941–17948, 1993. Kreemer, C., Holt, W. E., and Haines, A. J.: An integrated global model of present-day plate motions and plate boundary deforma- tion, Geophys. J. Int., 154, 8–34, 2003. Suppe, J.: Kinematics of Arc-Continent Collision, Flipping of Sub- duction, and Back-Arc Spreading near Taiwan, Memoir of the Geological Society of China, 6, 21–33, 1984. Miller, M. S., Kennett, B. L., and Lister, G. S.: Imaging changes in morphology, geometry, and physical properties of the subducting Paciﬁc plate along the Izu-Bonin-Mariana arc, Earth Planet. Sci. Lett., 224, 363–370, 2011. Taylor, R. N. and Nesbitt, R. W.: Isotopic characteristics of sub- duction ﬂuids in an intra-oceanic setting, Izu–Bonin Arc, Japan, Earth Planet. Sci. Lett., 164, 79–98, 1998. Van der Hilst, R. D. and Seno, T.: Effects of relative plate motion on the deep structure and penetration depth of slabs below the Izu-Bonin and Mariana island arcs, Earth Planet. Sci. Lett., 120, 375–407, 1993. Miller, M. S., Gorbatov, A., and Kennett, B. L.: Heterogene- ity within the subducting Paciﬁc slab beneath the Izu-Bonin- Mariana arc: Evidence from tomography using 3-D ray trac- ing inversion techniques, Earth Planet. Sci. Lett., 235, 331–342, 2005. Van der Sluis, A. and Van der Vorst, H. A.: Numerical solution of large, sparse linear algebraic systems arising from tomographic problems, in: Seismic tomography, edited by: Nolet, G., 49–83, Reidel, Dortrecht, 1987. Miller, M. S., Gorbatov, A., and Kennet, B. L. N.: Three- dimensional visualization of a near vertical slab tear beneath the southern Mariana arc; Geochem. Geophys. Geosyst., 7 Q06012, doi:10.1029/2005GC001110, 2006a. Widiyantoro, S. and Van der Hilst, R. D. : Structure and evolution of lithospheric slab beneath the Sunda Arc, Indonesia, Science, 271, 1566–1570, 1996. Miller, M. S., Kennett, B. L. N., and Toy, V. G.: Spatial and tem- poral evolution of the subducting Paciﬁc plate structure along the western Paciﬁc margin, J. Geophys. Res., 111, B02401, doi:10.1029/2005JB003705, 2006b. Widiyantoro, S., Kennett, B. L. N., and Van der Hilst, R. D.: Seismic tomography with P and S data reveals lateral variations in the rigidity of slabs, Earth Planet. Sci. References Hsu, Y.-J., Ando, M., Yu, S.-B., and Simons, M.: The potential for a great earthquake along the southernmost Ryukyu subduction zone, Geophys. Res. Lett., 39, L14302, doi:10.1029/2012GL052764, 2012. Baker, E. T., Embley, R. W., Walker, S. L., Resing, J. A., Lupton, J. E., Nakamura, K.-I., de Rode, C. E. J., and Massoth, G. J.: Hy- drothermal activity and volcano distribution along the Mariana arc, J. Geophys. Res., 113, B08S09, doi:10.1029/2007JB005423, 2008. Kato, T., Beavan J., Matsushima T., Kotake Y., Camacho J. T., and Nakao S.: Geodetic evidence of back-arc spread- ing in the Mariana Trough, Geophys. Res. Lett., 30, 1625, doi:10.1029/2002GL016757, 2003. Bassin, C., Laske, G., and Masters, G.: The current limits of resolu- tion for surface wave tomography in North America, EOS, Trans. Am. Geophys. Un., 81, F897, 2000. Katsumata, M. and Sykes, L. R.: Seismicity and tecton- ics of the western Paciﬁc: Izu-Mariana-Caroline and Ryukyu-Taiwan regions, J. Geophys. Res., 74, 5923–5948, doi:10.1029/JB074i025p05923, 1969. Berckhemer H. and Hs¨u K.: Alpine-Mediterranean geodynamics, American Geophys. Union, Washington, 1982. Bijwaard, H., Spakman, W., and Engdahl, E. R.: Closing the gap between regional and global travel time tomography, J. Geophys. Res., 103, 30055–30078, doi:10.1029/98JB02467, 1998. Solid Earth, 4, 59–73, 2013 www.solid-earth.net/4/59/2013/ www.solid-earth.net/4/59/2013/ K. Jaxybulatov et al.: Segmentation of the Izu-Bonin and Mariana slabs Lett., 173, 91–100, 1999. M¨uller, R. D., Roest, W. R., Royer, J. Y., Gahagan, L. M., and Sclater, J. G.: Digital isochrons of the World’s ocean ﬂoor, J. Geophys. Res., 102, 3211–3214, 1997. Zhao, D., Tian, Y., Lei, J., Liu, L., and Zheng, S.: Seismic image and origin of the Changbai intraplate volcano in East Asia: Role of big mantle wedge above the stagnant Paciﬁc slab, Phys. Earth Planet. Int., 173, 197–206, 2009. Solid Earth, 4, 59–73, 2013 www.solid-earth.net/4/59/2013/
https://openalex.org/W4282930972	https://zenodo.org/record/6644564/files/ZDPP17077.pdf	Quechua	null	IMKONIYATI CHEKLANGAN O'QUVCHILARGA "KO'RSATKICHLI FUNKSIYA" MAVZUSINI INTERFAOL METODLAR ASOSIDA O'QITISH METODIKASI	Zenodo (CERN European Organization for Nuclear Research)	2,022	cc-by	1,232	NKSIYA” NKSIYA” IMKONIYATI CHEKLANGAN O’QUVCHILARGA “KO’RSATKICHLI FUNKSIYA” MAVZUSINI INTERFAOL METODLAR ASOSIDA O’QITISH METODIKASI Umarova Guljahon Erdanovna1 matematika fan o’qituvchisi Xalilxodjayeva Shaxnoza Gafurovna2 biznes asoslari fani o’qituvchisi 1-2Toshkent imkoniyati cheklangan shaxslar uchun ixtisoslashtirilgan1-son kasb-hunar maktabi https://doi.org/10.5281/zenodo.6644564 IMKONIYATI CHEKLANGAN O’QUVCHILARGA “KO’RSATKICHLI FUNKSIYA” MAVZUSINI INTERFAOL METODLAR ASOSIDA O’QITISH METODIKASI Umarova Guljahon Erdanovna1 matematika fan o’qituvchisi Xalilxodjayeva Shaxnoza Gafurovna2 biznes asoslari fani o’qituvchisi 1-2Toshkent imkoniyati cheklangan shaxslar uchun ixtisoslashtirilgan1-son kasb-hunar maktabi https://doi.org/10.5281/zenodo.6644564 «Zamonaviy dunyoda pedagogika va psixologiya» nomli ilmiy, masofaviy, onlayn konferensiya 5. Eng muhimi o‘quv jarayonida didaktik motivlar vujudga keladi. Ya’ni, o‘quvchining ehtiyoji, xohish-istagi qondiriladi. IMKONIYATI CHEKLANGAN O’QUVCHILARGA “KO’RSATKICHLI FUNKSIYA” MAVZUSINI INTERFAOL METODLAR ASOSIDA O’QITISH METODIKASI Umarova Guljahon Erdanovna1 matematika fan o’qituvchisi Xalilxodjayeva Shaxnoza Gafurovna2 biznes asoslari fani o’qituvchisi 1-2Toshkent imkoniyati cheklangan shaxslar uchun ixtisoslashtirilgan1-son kasb-hunar maktabi https://doi.org/10.5281/zenodo.6644564 Annotatsiya Ushbu maqolada imkoniyati cheklangan o’quvchilarga “ko’rsatkichli funksiya” mavzusini interfaol metodlar asosida o’qitish metodikasi haqida fikr yuritiladi. Ushbu maqola imkoniyati cheklangan o’quvchilarga "ko’rsatkichli funksiya mavzusini interfaol metodlar asosida o’qitish metodikasi" moduliga bag'ishlangan bo'lib, unda o'qituvchilarning pedagogik mahoratini rivojlantirishga qaratilgan innovatsion ta'lim metodlari bayon qilingan. Interfaol ta’lim metodlarini dars jarayonlarida tatbiq etish asosida ta’lim samaradorligini oshirish va ta’limni individuallashtirish, shaxsga yo‘naltirish, ta’limda texnologik yondashuvni shakllantirish, hamda faol usullarni ta’lim jarayonlarida qo‘llash orqali o‘quvchilarni mustaqil ravishda o‘zlashtirish, tahlil qilish, ijodiy mulohaza yuritishga, shaxsiy xulosalar chiqarish, erkin fikr yuritishga o‘rgatish – pedagogning dasturilamal vazifasidir. Imkoniyati cheklangan o’quvchilarga "Ko’rsatkichli funksiya" mavzusini interfaol ta’lim metodlari asosida tashkil qilish orqali ta’lim samaradorligini oshirish mumkin. Imkoniyati cheklangan o’quvchilar o’qiyotgan kasb-hunar maktablarida ko’rsatkichli funksiya mavzularini interfaol usullardan foydalanib dars jarayonida qo’llab olib borish ta’lim samaradorligini oshiradi. Bu borada zamonaviy ta’limni tashkil etishga qo‘yiladigan muhim talablardan biri ortiqcha ruhiy va jismoniy kuch sarf etmay qisqa vaqt ichida yuksak natijalarga erishishdir. Interfaol usullari asosida dars jarayoni tashkil etilganda: 1. O‘quvchining o‘zaro faolligi oshadi, hamkorlik, ijodkorlikda ishlash ko‘nikmalari shakllanadi. 1. O‘quvchining o‘zaro faolligi oshadi, hamkorlik, ijodkorlikda ishlash ko‘nikmalari shakllanadi. 2. O‘quv, reja, dastur, darslik, standart, me’yor, qo‘llanmalar, mavzu mazmuni bilan ishlash malakalari shakllanadi. 3. Ta’lim mazmunini, ma’ruzasini mustaqil mutolaa qilish, ishlash, o‘zlashtirish kundalik shaxsiy ishlariga aylanadi. 4. O‘quvchi erkin fikr bildirish, o‘z fikrini himoya qilish, isbotlash, tasdiqlay olishga odatlanadi. 315 Zamonaviy dunyoda pedagogika va psixologiya» omli ilmiy, masofaviy, onlayn konferensiya 5. Eng muhimi o‘quv jarayonida didaktik motivlar vujudga keladi. Ya’ni, o‘quvchining ehtiyoji, xohish-istagi qondiriladi. Kichik g r hda ishlash Kichik guruhda ishlash Bu bosqichda imkoniyati cheklangan o’quvchilar kichik guruhlarga bo’linadi va har bir guruhga nom beriladi. Kichik guruhlarga guruh sardori tayinlanadi. Har bir guruh sardori o’z guruhi nomini ta’riflab beradi. Dars jarayonida ta’lim metodlarini qo’llash Dars jarayonida ta’lim metodlarini qo’llash Metodlar Klaster Aqliy hujum B.B.B jadval Amalga oshirish bosqichi • O’tilgan mavzuni so’rashda • Mavzu bayonida • Yangi mavzuni mustahkamlashda 31 Metodlar Klaster Aqliy hujum B.B.B jadval Amalga oshirish bosqichi • O’tilgan mavzuni so’rashda • Mavzu bayonida • Yangi mavzuni mustahkamlashda O’tgan mavzuni mustahkamlash 1-guruh 2-guruh 1.Cos 750 ni hisoblang 1. Sin 1050 ni hisoblang 2. Tenglamani yeching 2.Tenglamani yeching B.B.B. texnikasi № Mavzu savoli Bilama n Bilishni xohlaym an Bildim 1 2 3 4 5 1 Ko’rsatkichli funksiya? 2 Ko’rsatkichli funksiyaning aniqlanish 2 2 sin  x ;1 2 sin   x ;1 3 sin  x 2 3  x cos • Mavzu bayonida 2-guruh 1. Sin 1050 ni hisoblang 2.Tenglamani yeching 2 2 sin  x ;1 3 sin  x ;1 2 sin   x 2 3  x cos B.B.B. texnikasi № Mavzu savoli Bilama n Bilishni xohlaym an Bildim 1 2 3 4 5 1 Ko’rsatkichli funksiya? 2 Ko’rsatkichli funksiyaning aniqlanish B.B.B. texnikasi Mavzu savoli 316 monaviy dunyoda pedagogika va psixologiya» i ilmiy, masofaviy, onlayn konferensiya sohasi? 3 Ko’rsatkichli funksiyaning qiymatlar sohasi? 4 Ko’rsatkichli funksiyaning o`sish oralig`i? 5 Ko’rsatkichli funksiyaning kamayish oralig`i? 6 Xossalari? 1-guruh. Funksiya grafigini chizing 1) y=3 x 2-guruh. Funksiya grafigini chizing 2) y= 2 x 2-guruh. Ko’rsatkichli tenglamani yeching . 6 6 2 1 1 3 x x   1-guruh. Ko’rsatkichli tenglamani yeching ; 6,0 6,0 5 2 3   x x 1-guruh. Ko’rsatkichli tenglamani yeching 3 108 32 1 2    x x 2-guruh. Ko’rsatkichli tenglamani yeching 30 2 2 2 3 2 3     x x 1-guruh. Ko’rsatkichli tengsizlikni yeching 2-guruh. Ko’rsatkichli tengsizlikni hi 1-guruh. Funksiya grafigini chizing 2-guruh. Funksiya grafigini chizing 1-guruh. Ko’rsatkichli tenglamani yeching ; 6,0 6,0 5 2 3   x x 1-guruh. Ko’rsatkichli tenglamani yeching 3 108 32 1 2    x x 2-guruh. Ko’rsatkichli tenglamani yeching 30 2 2 2 3 2 3     x x 1-guruh. Ko’rsatkichli tengsizlikni yeching 8x-1≤ 2 2-guruh. Ko’rsatkichli tengsizlikni yeching 32x ≥ 1 317 «Zamonaviy dunyoda pedagogika va psixologiya» nomli ilmiy, masofaviy, onlayn konferensiya Uyga vazifa uchun misollar berish. Uyga vazifa uchun misollar berish. 1. Tenglamalarni yeching: 1. Tengsizlikni yeching. . 27 9 1 )6 ,8 4 1 )5 , 3 1 3 )4 , 2 1 9 )3 , 27 3 1 )2 , 4 1 2 )1 1 2 1 2 3                           x x x x x x Xulosa qilib shuni aytish joizki, ushbu maqolada "Ko’rsatkichli funksiya" modulini ta’lim jarayonlarida interfaol metodlardan foydalanish asosida olib borish ta’lim samaradorligini oshirish muammosini yechishga o‘quvchilarning bilim malaka, ko‘nikma, shaxsiy sifatlarini shakllantirish, ta’lim jarayoni qatnashchilarining mavzu yuzasidan bilim darajasini oshirish, hamkorlik va ijodkorligini oshirishga yordam beradi. Dars jarayonida ta’lim metodlarini qo’llash Tenglamalarni yeching: 1) ; 2) ; 3) ; 4) ; 5) ; 6) ; 7) ; 8) ; 9) ; 2 8 1   x 1 32  x 1 5,0 2  x x 4 1 2 1 2   x 9 3 1 1 2         x 5 25  x 16 2 4 2   x 9 27 3  x 3 2 1 7,0 7,0   x x 7) ; 8) ; 9) ;   16 2 4 2   x 9 27 3  x 3 2 1 7,0 7,0   x x 10) ; 11) ; 12) 256 81 8 9 9 2              x x 216 3 2   x x . 40 5 3 5 1 2 2     x x 13) ; 14) ; 15) ; 16) ; 17) ; 18) ; 19) ; 56 7 7 1    x x 1 1 4 5   x x 1 2 1 3 2    x x 1 1 1 1 2 7 2 3 3         x x x x x x x x 7 5 7 3 3 1 1 1        0 6 2 4    x x 0 10 10 11 100     x x Tengsizlikni yeching. Foydalanilgan adabiyotlar ro'yxati: 1. Vafoev R., Xusanov J.. boshqalar. Algebra va analiz asoslari. Akaddemik litseylar uchun qo’llanma . – T . O’qituvchi. 2001. 1. Vafoev R., Xusanov J.. boshqalar. Algebra va analiz asoslari. Akaddemik litseylar uchun qo’llanma . – T . O’qituvchi. 2001. 2. Abduhamidov A., Nasimov A. va boshqalar . Algebra va analiz asoslari. I k. Akademik litseylar uchun qo’llanma T.O’qituvchi 2001. Mirzaaxmedov A., Ismailov Sh.N,. Amanov A.Q Matematika 10 II-qism 318
https://openalex.org/W3208924269	https://www.degruyter.com/document/doi/10.1515/nanoph-2021-0319/pdf	English	null	True- and quasi-bound states in the continuum in one-dimensional gratings with broken up-down mirror symmetry	Nanophotonics	2,021	cc-by	11,833	technology and relying on the high quality factor induced by BIC. Keywords: bound states in the continuum; Fano reso- nance; microcavities. Abstract: There are many reports in the literature of bound states in the continuum (BICs) in systems with up–down mirror symmetry. Semiconductor-based technology re- quires bulk semiconductor substrates, which impose symmetry breaking in the vertical direction. In this paper, we explore the possibility of realizing BICs in a high refractive index subwavelength one-dimensional grating placed on a substrate with a refractive index that varies from 1 to almost the refractive index of the grating, while the refractive index above the grating is 1. We demonstrate that in gratings with broken up–down mirror symmetry not only symmetry-protected BICs can arise, but also Frie- drich–Wintgen (FW) and interference-based (IB) BICs with diverging quality factors. The limit of the refractive index difference between the grating and the substrate support- ing the BIC was found to be as little as 0.03. We also present a study of configurations composed of a finite number of grating stripes, with refractive indices corresponding to GaAs in the grating and Al-rich AlGaAs in the substrate. We demonstrate that such an all-semiconductor configuration enables Q-factors above 104 when composed of fewer than 20 periods and nearly exponential Q-factor growth with increasing numbers of grating periods. The results of this study pave the way for a new class of micro- and nano- optical cavities realised in standard all-semiconductor Nanophotonics 2021; 10(16): 3979–3993 Open Access. © 2021 Weronika Głowadzka et al., published by De Gruyter. This work is licensed under the Creative Commons Attribution 4.0 International License. Bound states in gratings with broken symmetry https://doi.org/10.1515/nanoph-2021-0319 Received June 24, 2021; accepted October 8, 2021; published online October 25, 2021 https://doi.org/10.1515/nanoph-2021-0319 Received June 24, 2021; accepted October 8, 2021; published online October 25, 2021 Research Article Weronika Głowadzka, Michał Wasiak and Tomasz Czyszanowski* ronika Głowadzka et al., published by De Gruyter. This work is licensed under the Creative Commons Attribution 4.0 1 Introduction When there is a lack of vertical symmetry [13–15] or in periodic structures with ﬁnite dimensions, which is the case with real-world realizations, the BIC transforms into a so-called quasi-BIC (qBIC) [16, 17] of ﬁnite but still very high quality factor. In VS gratings, three types of BICs can be formed at ky = 0. The ﬁrst appears when the mode is antisymmetric and the grating is subwavelength. The grating admits only a single leaky channel, related to zero diffraction order emission. Zero diffraction order emission is prohibited in the case of antisymmetric modes; hence, antisymmetric modes cannot be radiated from the grating. This type of BIC is called symmetry-protected (SP) [16] and is robust against changes in the grating parameters as long as the grating is subwavelength and ky = 0. In the case of antisymmetric modes in the grating with a period larger than the wave- length (diffraction regime), or in the case of symmetric modes regardless of the period size, at least one emission channel is open and symmetry protection is no longer possible [12, 18]. When there is only a single radiation channel, judicious tuning of the grating parameters can lead to destructive interference at both interfaces of the grating, producing an IB BIC [19, 20]. When tuning the grating parameters leads the dispersion curves of strongly coupled nonorthogonal resonances to approach each other, inducing avoided crossing with noticeable separa- tion of the frequencies [21], the FW BIC appears [22]. The IB and FW types of BIC (that are also named accidental BICs) may occur for modes with ky = 0 as well as ky ≠0 [23]. quality factor. In this paper, we investigate qBICs of finite Q-factor and BICs of inﬁnite Q-factor (also known as true BICs, to emphasize their difference with respect to qBICs) in binary subwavelength gratings implemented on the surface of a semi-inﬁnite substrate (Figure 1). In this conﬁguration, the stripes are positioned periodically along the y direction and are inﬁnite along the x direction (see the coordinate system in Figure 1). The wavevectors (k0) of the resonant states (modes) are composed in general of kx, ky and kz compo- nents. However, due to the inﬁnite x dimension in our analysis, the wavevector is composed of two nonzero components: k0 = (0, ky, kz). The norm of k0 is a wave number (k0) related to the resonant wavelength λ = 2π/k0. 1 Introduction We limit our study to the modes at Γ-point (kx = ky = 0). We consider transverse-electric (TE) polarization of the elec- tromagnetic wave, in which the electric component of the electromagnetic ﬁeld is parallel to the grating stripes. We name the modes TEnm, where n indicates the number of zeros in the distribution of the electric ﬁeld in the vertical direction of the mode in a single grating segment (the segment is deﬁned in Section 2) and m is the number of zeros in the lateral direction perpendicular to the stripes. Zeroes at the segment boundaries (if present) are regarded as one. We do not consider transverse magnetic (TM) po- larization that is orthogonal to TE, as it would require a separate analysis and conclusions from the analysis of TE polarization cannot be directly applied to TM polarization. Modes propagating in the grating of which the lateral dis- tribution (along y direction) of the electric ﬁeld is sym- metric with respect to all symmetry x–z planes of the grating are referred to as symmetric modes (s-TE) (see Figure 4 for examples of symmetric modes). Analogously, modes with antisymmetric lateral distribution of the elec- tric ﬁeld with respect to all symmetry x–z planes of the grating are referred to as antisymmetric (a-TE) (see Figure 7 for examples of antisymmetric modes). When the refractive index of the substrate (ns) is equal to the refractive index of superstrate (na), the grating is vertically symmetric (VS). A VS grating with a superstrate and substrate both with a refractive index of 1 is a special case referred to as a membrane. The gratings with broken up-down mirror symmetry we call vertically nonsymmetric (VN) gratings. In the case of the VN grating, the subwavelength regime VS gratings in which BICs are observed are very often impractical, as they are thin membranes suspended in air [24] or implemented on a low refractive index substrate and then covered with a material that has an identical or almost identical refractive index to the substrate [16]. Such designs Figure 1: Schematic 3D picture of a grating with broken up–down mirror symmetry, consisting of a high refractive index periodic membrane (ng) placed on a substrate with a refractive index ns, where ng > ns. 1 Introduction The interaction of light and matter within subwavelength volumes is an important phenomenon in numerous branches of physics and technology, including light detectivity [1], stimulated emission in lasers [2], nonlinear optics [3], single-photon generation [4], observation of polariton–excitons [5], quantum electrodynamics, and others. A prominent example is Bose–Einstein condensa- tion in photonic systems [6], which is observed along with coherent lasing below population inversion [7]. Strong light–matter interaction occurs in metallic periodic struc- tures that utilize subwavelength localization of the light by means of surface plasmon polaritons [8]. However, light localization in these structures is inevitably associated with light absorption by metal in the visible and infrared region, which deteriorates the efﬁciency of plasmonic de- vices. In dielectric high refractive index periodic structures, strong light conﬁnement in a small volume is enabled in the absence of absorption by means of Fano resonance [9, 10]. Fano resonance is manifested by an abrupt transition from total reﬂection to total transmission in the spectrum of the light interacting with a structure. When the spectral width of the dissipative Fano resonance becomes inﬁnitely narrow, its quality factor tends to inﬁnity and the reso- nance becomes a bound state in the continuum (BIC) [11, 12]. A BIC is a nonradiating resonant state of inﬁnite quality (Q) factor localised in an open photonic system. Other resonant states of the system correspond to the continuum of optical states that carry electromagnetic energy away *Corresponding author: Tomasz Czyszanowski, Institute of Physics, Lodz University of Technology, 219 Wólczańska Street, 90-924, Łódź, Poland, E-mail: tomasz.czyszanowski@p.lodz.pl. https://orcid.org/ 0000-0002-0283-5074 Weronika Głowadzka and Michał Wasiak, Institute of Physics, Lodz University of Technology, 219 Wólczańska Street, 90-924, Łódź, Poland. https://orcid.org/0000-0002-0353-6690 (W. Głowadzka). https://orcid.org/0000-0002-9569-4265 (M. Wasiak) Weronika Głowadzka and Michał Wasiak, Institute of Physics, Lodz University of Technology, 219 Wólczańska Street, 90-924, Łódź, Poland. https://orcid.org/0000-0002-0353-6690 (W. Głowadzka). https://orcid.org/0000-0002-9569-4265 (M. Wasiak) 3980 W. Głowadzka et al.: Bound states in gratings with broken symmetry 3980 relative to air is referred to as the air-subwavelength (1 < L/λ) and subwavelength regime relative to substrateisreferred to as substrate-subwavelength (ns −1 < L/λ), where L is period of the grating. from the photonic structure. BICs have been found theo- retically in inﬁnite structures that are periodic in at least one dimension and symmetric in two other dimensions. make it difﬁcult to fabricate electrically driven devices, since the air or low refractive index materials (typically dielectrics) are nonconducting. In comparison, all- semiconductor structures with broken vertical symmetry, such as for example a grating implemented on a bulk substrate, offer several advantages. Such structures can be robust and immune to mechanical failure. They can also be used as electrically driven devices. In contrast to sus- pended structures, their fabrication does not require so- phisticated critical point drying after selective wet etching to release the membrane during processing. A given peri- odic structure on a solid semiconductor substrate can cover an arbitrarily large area, whereas large area suspended structures are always at risk of collapsing [25]. The spatial parameters of the grating can be precisely controlled via electron beam lithography for research applications or via nanoimprint lithography [26] on a mass production scale. However, such all-semiconductor structures have the obvious disadvantages of broken vertical symmetry and lower contrast between the refractive index of the periodic structure and the substrate, which contribute to light leakage into the substrate. case of modes with symmetric lateral distribution of the electric ﬁeld in a single segment (s-TE). In Section 4, we consider modes with antisymmetric lateral distribution of the electric ﬁeld (a-TE) in a single segment of a VN grating. In Section 5, we provide a broad overview of high Q-factor resonances for all mode symmetries in VN gratings with substrate refractive indices varying from 1 to the refractive index of the grating. The structures analysed in Sections 3– 5 are inﬁnitely periodic gratings. In Section 6, we give an example of a conﬁguration with a ﬁnite number of grating stripes and a low refractive index contrast between the grating and the substrate, corresponding to the contrast achieved in arsenide-based materials in the near infrared region. This example compares the Q-factors of two ﬁnite conﬁgurations with parameters corresponding to true BIC and qBIC in inﬁnite arrangement. It demonstrates the possibility of achieving a very high Q-factor in a ﬁnite all- semiconductor VN grating and shows a possible way of constructing current injection devices. 2 Methods and structures To the best knowledge of the authors, this is the first study to investigate the influence of up–down symmetry breaking and refractive index contrast on the creation of BICs and high Q-factor Fano resonances in a one- dimensional grating. Analyses of high Q-factor reso- nances in VN conﬁgurations are sparsely documented in the literature. Two studies have provided valuable insights into the mechanisms that cause BICs to collapse into ﬁnite Q-factor Fano resonances, due to lack of vertical symmetry [13, 14]. Neither of these studies, however, showed that true IB and FW BICs can be formed in VN gratings. The main purpose of the numerical analysis presented in the present paper is to investigate the inﬂuence of up–down mirror symmetry breaking and refractive index contrast on the creation of accidental (IB and FW) true and quasi BICs in a one-dimensional grating. The gratings with broken up– down mirror symmetry and reduced contrast between the refractive index of the grating and surroundings represent real-life realizations of gratings implemented directly on a substrate in all-semiconductor conﬁgurations. Although our work concerns a one-dimensional grating as an example structure, the conclusions are expected to be of more general signiﬁcance and may be applicable to various one- and two-dimensional periodic conﬁgurations imple- mented on substrates with broken up–down mirror symmetry. The method we used to numerically solve Maxwell equa- tions is based on the Admittance Method [27], which is applied for the z direction, combined with Fourier series expansion of the electromagnetic ﬁeld and refractive index in the lateral y direction of the structure. We assume that the structure and the electromagnetic ﬁeld are homoge- neous along the x axis. Consequently, the structure is pe- riodic along the y axis and consists of inﬁnitely repeated copies of the rectangular (in the (y, z) plane) computational window of width W in the y direction in which the structure is deﬁned. The computed electromagnetic wave along the y direction is in the form given by Bloch’s theorem: Ψ(y) = eikyyf(y) (1) (1) where f(y) is a periodic function with the period W as the computational window and ky is an arbitrarily chosen number ranging from –π/W to π/W. We simulate two types of conﬁguration. The ﬁrst, which we call inﬁnite, the computational window W is equal to the period of the grating structure (L). 1 Introduction L is the width of the grating segment, H is the height of the grating stripes, a is the width of the stripes and F is the duty cycle determined as the ratio a/L. Figure 1: Schematic 3D picture of a grating with broken up–down mirror symmetry, consisting of a high refractive index periodic membrane (ng) placed on a substrate with a refractive index ns, where ng > ns. L is the width of the grating segment, H is the height of the grating stripes, a is the width of the stripes and F is the duty cycle determined as the ratio a/L. W. Głowadzka et al.: Bound states in gratings with broken symmetry 3981 2 Methods and structures In the second case, referred to as ﬁnite, the computational window comprises numerous segments of the same width (L) of the grating structure. Therefore, L is not a period of the whole structure. Unless stated otherwise, the calculations are performed for ky = 0. We use two models that represent the electromagnetic ﬁeld in the form of Bloch waves, as shown in (1). The Plane Wave Reﬂection Transformation Method (PWRTM) [28] is used to determine the power reﬂectance and transmittance of each where f(y) is a periodic function with the period W as the computational window and ky is an arbitrarily chosen number ranging from –π/W to π/W. We simulate two types of conﬁguration. The ﬁrst, which we call inﬁnite, the computational window W is equal to the period of the grating structure (L). In the second case, referred to as ﬁnite, the computational window comprises numerous segments of the same width (L) of the grating structure. Therefore, L is not a period of the whole structure. Unless stated otherwise, the calculations are performed for ky = 0. We use two models that represent the electromagnetic ﬁeld in the form of Bloch waves, as shown in (1). The Plane Wave Reﬂection Transformation Method (PWRTM) [28] is used to determine the power reﬂectance and transmittance of each In Section 2, we present the grating conﬁgurations considered in this work and the numerical method used for their analysis. In Section 3, we investigate VN grating in the 3982 W. Głowadzka et al.: Bound states in gratings with broken symmetry 3982 diffraction order. The Plane Wave Admittance Method (PWAM) [27] is used to ﬁnd electromagnetic resonances occurring in the structure. The complex wavenumbers of the modes (k0 = k0re + ik0im) can be determined by PWAM, with real parts (k0re) corresponding to the values of the free space wavevectors of the optical modes and imaginary parts (k0im) that are proportional to the optical losses of the modes. The Q-factor of the modes is deﬁned as Q = 0.5k0re/k0im. In both methods, the only approximation used is truncation of the basis size. Our model has been shown to have high reli- ability by comparison with experimental results [29, 30]. In Sections 3 and 4, the computational window comprises a single grating segment. 2 Methods and structures To simulate structures with ﬁnite gratings, the computational window in the lateral direction contains a ﬁnite number of grating stripes and a region with imposed absorbing boundary conditions is positioned at a distance of three segment widths from the grating. The number of plane waves used in the method (the size of the basis) was approximated according to the formula W/L × 30, where L is the width of the grating segment and W is the computational window. refractive index is the weighted refractive index over the normalized optical ﬁeld intensity of the mode [32]. There- fore, its value is smaller than ng. When F is 1, then neff g = ng for H →∞and all the modes in the range expressed by formula (2) are bound states with inﬁnite Q. When F < 1, there is a reduction in neff g compared to F = 1. In addition, the s-TE modes can couple to the free space, turning them into dissipative modes with a ﬁnite Q-factor [11]. The light intensity maximum of these dissipative modes is located in the grating and the light does not vanish outside the grating. We restrict our considerations to air- subwavelength gratings (L < λ), which limits m to 6 in our example (since m 2neff g > 1 for m ≥8 and ng eff ≤ng). In this spectral region, in the case of VS grating although the modes couple with the plane wave in air, light dissipation can be eliminated by tuning the grating parameters and BICs of two types emerge: FW and IB (see Figure S1b in the Supplementary Materials). The FW BICs appear in the vi- cinity of avoided crossing, between pairs of nonorthogonal modes [21]: s-TEn2 and s-TEn4, as well as s-TEn4 and s-TEn6. The other BICs are IB BICs, which are present in the dispersion curves of s-TEnm modes for m ≥2. In the calculations, the refractive index of the grating (ng) is 3.52 (Figure 1). This level of refractive index corre- sponds to GaAs at 1 µm. It is also close to silicon and numerous alloys of phosphide- and antimonide-based materials in the infrared region. The refractive index of the substrate (ns) varies. In Sections 3 and 4 we consider a VN grating where ns = 2.5. In Section 5, ns is varied in the range from 1 to values approaching 3.52. 2 Methods and structures Finally, in Section 6, the refractive index of the substrate is ns = 2.95, which corre- sponds to the refractive index of Al-rich AlGaAs at 1 µm. We consider possible variations in all the grating parameters shown in Figure 1: the grating segment (L), the etching depth (H) and the grating duty cycle (F) deﬁned as the ratio of the grating stripe width (a) to the segment width (L). Figures 2 and S1b in the Supplementary Materials illustrate the dispersion curves of s-TE modes for the VN grating and membrane, respectively. The Q-factors vary along the curves and reach maxima at the points indicated by circles. Colours correspond to the Q-factor value. The positions and values of the Q-factor maxima were deter- mined using an optimization algorithm, searching for the minimal k0im that takes the real part of the wavenumber 1.0 0.8 0.6 0.4 0 1 2 3 C4 10 0 10 2 10 4 Q H/L C2 10 6 Figure 2: Q-factor (Q) of resonances expressed by colours in the domain of the grating thickness (H) and wavelength (λ) for a grating implemented on a substrate with a refractive index of 2.5. L is the grating segment width and the duty cycle F = 0.9. Circles represent local maxima of the Q-factor. Vertical dashed lines indicate s-TEnm mode cutoffs for m = 2 and m = 4. 1.0 0.8 0.6 0.4 0 1 2 3 C4 10 0 10 2 10 4 Q H/L C2 10 6 3 Laterally symmetric modes H/L The modes propagating in the grating can be described according to waveguide theory [31], by determining their wavelength range within the slab periodic waveguide: m 2neff g < L λ < m 2ns = Cm (2) (2) In formula (2), the short wavelength limit of L/λ is the cutoff (Cm) of the TEnm modes. The cutoff is determined by the refractive index of the substrate (ns) and is equal for all s-TEnm and a-TEnm modes with equal m. The long wave- length limit corresponds to H →∞and depends on the effective refractive index of the grating (neff g ). The effective Figure 2: Q-factor (Q) of resonances expressed by colours in the domain of the grating thickness (H) and wavelength (λ) for a grating implemented on a substrate with a refractive index of 2.5. L is the grating segment width and the duty cycle F = 0.9. Circles represent local maxima of the Q-factor. Vertical dashed lines indicate s-TEnm mode cutoffs for m = 2 and m = 4. W. Głowadzka et al.: Bound states in gratings with broken symmetry 3983 represent dispersion curves and Q-factor in the case of the VN grating. Although the curves relate to a speciﬁc mode, the behaviour shown in Figure 3(a) is representative of the dispersion curves for all s-TEnm modes. As can be clearly seen, breaking the vertical symmetry eliminates anti- crossings from the dispersion curve, due to reduced coupling between the modes in the VN grating (this issue is discussed in greater detail in the Supplementary Materials S1). However, a slight ﬂuctuation in the intersections with s-TEn2 modes remains. Although avoided crossing van- ishes, the corresponding maximum Q-factor remains in the vicinity of the crossing with s-TEn2 modes. In the case of the membrane, the corresponding Q-factor maximum is inﬁn- ite and indicates an FW BIC. In the case of the VN grating, the position of the maximum is very close to that of the FW BIC in the membrane. However, its Q-factor is reduced to a ﬁnite value, indicating a quasi-FW (qFW) BIC. The BICs in the membrane that are not positioned in the vicinity of the avoided crossing are IB BICs. Breaking the vertical sym- metry of the grating also reduces their Q-factor, turning them into quasi-IB (qIB) BICs. 3 Laterally symmetric modes Numbers (1)–(5) indicate the resonant wavelengths and H/L of the VN grating for which the distribution of the modes is illustrated in Figure 4. Distributions numbered (2) and (4) indicate the qFW BIC and qIB BICs, respectively. As both conﬁgurations are positioned outside the substrate-subwavelength regime, the light distribution reveals a typical checkerboard pattern in the substrate. In both cases, this indicates the presence of at least two diffraction orders emitted towards substrate. On the air side, the uniform distribution of light intensity indicates zeroth diffraction order emission. and H as variables. A comparison of the two maps indicates that in a VN grating with Q < 106 the breaking of up-down mirror symmetry by the presence of the substrate turns the BICs present in the membrane into qBICs (for examples of modes distributions in VN grating see Figure S2 in the Supplementary Materials). In the example presented in Figure 2, the presence of the substrate with a refractive index of ns = 2.5 sets the diffraction regime in the range ns −1 = 0.4 < L/λ. For ns −1 > L/λ only the 0th diffraction order is emitted towards the superstrate and substrate. However, vertical symmetry breaking introduces different conditions for destructive interference in the far ﬁeld in both the superstrate and the substrate. These conditions cannot be fulﬁlled simultaneously, contributing to light dissipation. In the range ns −1 < L/λ all the modes emit higher than 0th diffraction orders towards the substrate, contributing to open additional emission channels besides the 0th diffraction order. Both mechanisms, breaking the up– down mirror symmetry and opening additional leakage channels, contribute to the absence of true BICs in the VN grating, as discussed in [13, 14]. The qBICs with the largest Q-factors are observed in the case of low-order vertical modes (small n indices) and in the case of large H, which clearly increases the Q-factor of cavity by increasing cavity length. Although this tendency is shown for a speciﬁc VN conﬁguration, it is generalized for all VN conﬁgurations considered in Section 5. Figure 3(a) provides a closer inspection of the evolu- tion of the dispersion curve of an arbitrarily selected s-TE04 mode. Grey lines and dark red points indicate membrane dispersion curves and BICs, respectively. Coloured lines Figure 3: Dispersion curves of symmetric modes in the VN grating (coloured curves) and the membrane (grey curves) (a). 3 Laterally symmetric modes Głowadzka et al.: Bound states in gratings with broken symmetry Figure 4: Lateral profiles of the electric field (black line in top panels) and light intensity distributions (bottom panels) of the s-TE04 mode within a single segment of VN grating. Numbers (1) to (5) correspond to the grating conﬁgurations indicated by the same numbers in Figure 3(a) and (b). The proﬁles of the electric ﬁeld are cross-sections through the light intensity maximum of the modes. In the top panels, blue lines represent the proﬁle of the refractive index in the grating and grey dashed lines indicate 0 of the electric ﬁeld. In the bottom panels, light intensity is presented on a logarithmic scale of colours. Figure 4: Lateral profiles of the electric field (black line in top panels) and light intensity distributions (bottom panels) of the s-TE04 mode within a single segment of VN grating. Numbers (1) to (5) correspond to the grating conﬁgurations indicated by the same numbers in Figure 3(a) and (b). The proﬁles of the electric ﬁeld are cross-sections through the light intensity maximum of the modes. In the top panels, blue lines represent the proﬁle of the refractive index in the grating and grey dashed lines indicate 0 of the electric ﬁeld. In the bottom panels, light intensity is presented on a logarithmic scale of colours. Figure 3(b) compares Q-factors along the dispersion curve of the s-TE04 mode in the membrane and in the VN grating for a broad range of H/L. As can be seen, the qFW BICs in the VN grating form at nearly the same H/L as the FW BICs in the membrane. The difference in the refractive index of the substrate modiﬁes the VN and membrane dispersion curves marginally. Therefore, the dispersion curve intersections do not change their positions signiﬁ- cantly in the H/L–L/λ plane when the two conﬁgurations are compared. A greater difference in positions of IB and qIB BICs is observed. Figure 3(c) enables a more in-depth examination of the mechanism of qBIC formation in the VN grating illustrating the ratio of the electromagnetic ﬁeld energy in the cavity relative to the light energy in the substrate (ηs) and air (ηa) calculated far from the cavity. Both ηs and ηa are superimposed on the Q-factor function for the VN grating. The ﬁgure indicates a strong correlation be- tween the positions of the qBIC and ηs maxima. 3 Laterally symmetric modes Moreover, ηs is signiﬁcantly smaller than ηa, revealing that the main leakage channel of the electromagnetic ﬁeld from the cavity leads through the substrate, where the energy ﬂow cannot be completely inhibited due to the existence of two diffrac- tion orders. A comparison of ηa and the membrane Q-factor in Figure 3(b) (black line) shows that the positions of the ηa maxima and membrane BICs coincide very closely. How- ever, not all ηa maxima in the VN grating diverge to inﬁnity (as occurs in the membrane), indicating that completely destructive air-side interference is hindered in conﬁgura- tions with broken up-down mirror symmetry. hand, only the zeroth diffraction order can be emitted inthe subwavelength regime of the membrane (L < λ). Therefore, any emission of antisymmetric modes in the VS grating is prohibited; making those modes bound states [16]. Imple- menting a substrate with refractive index ns modiﬁes the substrate-subwavelength condition (L/λ < ns −1) in which SP BICs exist [14]. Modes that fulﬁl the condition ns −1 < L/λ < 1 emit the nonzero diffraction order into the substrate, becoming dissipative resonances. Figure 5 illustrates the dispersion curves of a-TEnm modes with even m in the grating implemented on a substrate with ns = 2.5, in two conﬁgurations differing in F. The distributions of the electric ﬁeld and light intensity for the selected modes in the grating conﬁguration with F = 0.9 are illustrated in Figure S4 in the Supplementary Materials. The zeros of the electric ﬁeld of the modes on the axes of the grating period symmetry indicate their antisymmetric distributions. The substrate-subwavelength condition (L/λ < ns −1) is below the long wavelength limit of the a-TEn4 mode ( 1 ns < 2 neff g ). There- fore, only a-TEn2 can be an SP BIC in this conﬁguration. The substrate-subwavelength condition is equivalent to the cutoff condition of an a-TEn2 mode. Hence, the a-TEn2 modes with wavelengths shorter than the cutoff limit become leaky. On the other hand, as long as neff g > ns, an a-TEn2 mode can always be in a bound state within a certain wavelength range. Different values for F induce a change in the effective index of the grating (neff g ), affecting the num- ber of modes observed in the range L < λ. 3 Laterally symmetric modes The two sloping curves in (a) represent s-TE04 modes in both conﬁgurations. The nearly horizontal curves represent s-TEn2 modes that are leaky in the case of the VN grating. Dark red circles in grey curves represent BICs in the membrane. Colours in the dispersion curves of VN grating correspond to the Q-factor value. This map is a magniﬁcation of an arbitrarily chosen section of the maps presented in Figure 2; (b) Q-factor versus etching depth (H/L) for the s-TE04 mode in the membrane (black line) and VN grating (red line) with Friedrich–Wintgen (FW) and interference-based (IB) BICs and qBICs. Numbers (1)–(5) indicate the parameters of the VN grating and corresponding Q-factors of the mode illustrated in Figure 4. (c) The ratio (η) of the electromagnetic ﬁeld energy in the cavity relative to the light energy in the substrate (blue line and ηs in the text) and air (red line and ηa in the text). The grey line indicates the Q-factor function for the VN grating, which is the same as in b) (red line). FW and IB quasi-BICs are indicated as qF and qI. Figure 3: Dispersion curves of symmetric modes in the VN grating (coloured curves) and the membrane (grey curves) (a). The two sloping curves in (a) represent s-TE04 modes in both conﬁgurations. The nearly horizontal curves represent s-TEn2 modes that are leaky in the case of the VN grating. Dark red circles in grey curves represent BICs in the membrane. Colours in the dispersion curves of VN grating correspond to the Q-factor value. This map is a magniﬁcation of an arbitrarily chosen section of the maps presented in Figure 2; (b) Q-factor versus etching depth (H/L) for the s-TE04 mode in the membrane (black line) and VN grating (red line) with Friedrich–Wintgen (FW) and interference-based (IB) BICs and qBICs. Numbers (1)–(5) indicate the parameters of the VN grating and corresponding Q-factors of the mode illustrated in Figure 4. (c) The ratio (η) of the electromagnetic ﬁeld energy in the cavity relative to the light energy in the substrate (blue line and ηs in the text) and air (red line and ηa in the text). The grey line indicates the Q-factor function for the VN grating, which is the same as in b) (red line). FW and IB quasi-BICs are indicated as qF and qI. 3984 W. 3 Laterally symmetric modes As a consequence, a-TEn2 and a-TEn4 are present for F = 0.7 (Figure 5(a)), whereas a-TEn6 is additionally observed for F = 0.9 (Figure 5(b)). If we compare the Q-factors for the two grating conﬁgurations of a-TEn4, it will be noticed that the average values are lower in the case of F = 0.7 (predominantly colour blue in the a-TEn4 dispersion curves) in comparison to F = 0.9 (predominantly green). This relates to stronger 4 Laterally antisymmetric modes The zeroth diffraction order is not emitted by the anti- symmetric modes a-TEn2, a-TEn4, and a-TEn6. On the other W. Głowadzka et al.: Bound states in gratings with broken symmetry 3985 3985 Figure 5: Q-factors of a-TE modes indicated by colours for the VN grating implemented on a substrate with a refractive index of 2.5 and F = 0.7 (a) and F = 0.9 (b). Q-factor functions depicted in the domain of the wavelength (λ) and the height of the grating (H). Both are normalised with respect to the length of the grating segment (L). Large dark red points represent inﬁnite values corresponding to BICs. qBICs positioned in the dissipative regime are indicated by large points in other colours. Figure 5: Q-factors of a-TE modes indicated by colours for the VN grating implemented on a substrate with a refractive index of 2.5 and F = 0.7 (a) and F = 0.9 (b). Q-factor functions depicted in the domain of the wavelength (λ) and the height of the grating (H). Both are normalised with respect to the length of the grating segment (L). Large dark red points represent inﬁnite values corresponding to BICs. qBICs positioned in the dissipative regime are indicated by large points in other colours. a-TEn2 modes. Every second BIC relates to avoided crossings inthemembrane withF =0.7, asillustrated by thegrey curves in Figure 6(a). This suggests that every second BIC is of the FW-type and the others are IB-type. coupling of the grating modes with the continuum in the case of smaller F, which is enabled by larger air slits in the grating. Surprisingly, the local maxima of the Q-factors present in the dispersion curves of the a-TEn4 modes reveal the opposite tendency. In the case of F = 0.9, the Q-factors of the qBICs scarcely reach the level of 1012 and are usually at the level of 1010. However, in the case of F = 0.7 the local maxima of the Q-factors are more than 1012 and a signiﬁcant number of the maxima reach values exceeding the numer- ical capabilities of computers, indicating inﬁnite values and the presence of FW or IB BICs in the VN grating. This example shows the possibility of true BICs appearing in structures with broken up-down mirror symmetry. Figure 6(b) compares the Q-factors of the a-TE04 mode for two conﬁgurations with F = 0.7 (black) and F = 0.9 (red). 4 Laterally antisymmetric modes It conﬁrms that weaker coupling with the substrate in the case of F = 0.9 results in Q-factors in the range of 105–106 in the minima of the Q-factor function. The Q-factors of the a-TEn2 modes are at the level of 102–103 (Figures 6(a) and 5(b)), which creates a larger difference between the imag- inary wavenumbers of a-TE14 and a-TEn2 modes in com- parison to F = 0.7 (see Figure 5(a)). As a result, no true BICs are indicated in the conﬁguration with F = 0.9. In the VN configuration with ns = 2.5 and F = 0.7, when ns −1 < L/λ < 2ns −1 the presence of a true BIC can be explained by the existence of a single leakage channel related to the emission of the ﬁrst diffraction order towards the substrate. The light emitted towards the air is prohibited for L < λ. Figure 6(a) shows the intersections of the a-TE14 and aTEn2 modes. Panels (1), (3), and (5) in Figure 7 conﬁrm the ex- istence of a single emission channel towards the substrate related to the ﬁrst diffraction order. As expected [33], when the real parts of the wavelengths of two modes cross and their Q-factors are close, strong mode repulsion occurs in the imaginary part of the wavenumber, in the vicinity of the crossing. This mechanism is manifested as the cancelation of light emission towards the substrate and the appearance of BICs, as shown in panels (2) and (4) of Figure 7 (additionally conﬁrmed by Figure S5 in Supplementary Materials). The BICs are positioned in the proximity of the crossings with the 5 Influence of the substrate refractive index on Q-factor In the top panels, blue lines represent the proﬁle of the refractive index in the grating and grey dashed lines indicate 0 of the electric ﬁeld. In the bottom panels, light intensity is presented on a logarithmic scale of colours. Figure 7: y-Directional proﬁles of the electric ﬁeld (black line in top panels) and light intensity distributions (bottom panels) for an a-TE04 mode within a single segment of the VN grating in the y–z plane. Numbers (1)–(5) correspond to the grating conﬁgurations indicated by the same numbers in Figure 6(b). The proﬁles of the electric ﬁeld are cross-sections through the light intensity maximum of the modes. In the top panels, blue lines represent the proﬁle of the refractive index in the grating and grey dashed lines indicate 0 of the electric ﬁeld. In the bottom panels, light intensity is presented on a logarithmic scale of colours. [11, 12] that are not analysed in this paper. Increasing the refractive index of the substrate in the VN grating qBICs reduces their Q-factors abruptly. Increasing F increases the Q-factor of qBICs by reducing the ability of the grating modes to couple with the continuum of states. Increase of F increases also mode conﬁnement due to increase of the effective refractive index. In Figure 3(b), we showed that qFW BICs reach higher Q-factors in comparison to qIB BICs. Two mechanisms supporting the high Q-factors of sym- metric modes can be observed, which were partly noticed in our analysis of Figure 2. The ﬁrst relates to the occur- rence of high Q-factor qBICs in the case of s-TEnm modes with small n indices and large H. These qBICs are therefore positioned close to the long-wavelength limits. They reach Q ≈106 (bright blue colour in Figure 8(d)–(f)) even in the an example, the L/λ and Q-factors of all the qBICs indicated by coloured circlesinFigure 2 areimplementedinFigure8(f) along the red dashed line, corresponding to the refractive index ns = 2.5. The same procedure is applied to all considered ns and mode symmetries. In Figures 8 and 9 the points representing larger Q-factors overlap lower values. The dashed black lines in Figures 8(f) and 9(f) represent the borders of regions in which particular TEnm modes are conﬁned inthegrating. 5 Influence of the substrate refractive index on Q-factor The refractive index contrast between the grating and the substrate is an important design parameter, as it de- termines the materials from which the grating and the substrate can be fabricated. Figures 8 and 9 show maps of Q-factors in the domain of L/λ and the refractive index of the substrate (ns) for different values of F. Each map is composed of points produced by searching for the Q-factor maxima for given ns in the H/L range from 0 to 3 and in the L/λ range from 0.2 to 1. No information on the H/L corre- sponding to Q-factor maxima is displayed in the maps. As 3986 W. Głowadzka et al.: Bound states in gratings with broken symmetry 3986 W. Głowadzka et al.: Bound states in gratings with broken symmetry Figure 6: Dispersion curves for antisymmetric modes with F = 0.7 in the VN grating (coloured curves) and in the membrane (grey curves) (a). Colours in the dispersion curves correspond to the Q-factor values. Sloping curves represent an a-TE04 mode and nearly horizontal curves represent a-TEn2 modes. (b) Q-factor versus the stripe height (H/L) for the a-TE04 mode in VN gratings where F = 0.7 (black curve) and F = 0.9 (red curve), with Friedrich–Wintgen (FW) and interference-based (IB) BICs and qBICs. Numbers (1)–(5) indicate structures and corresponding modes illustrated in Figure 7. Figure 6: Dispersion curves for antisymmetric modes with F = 0.7 in the VN grating (coloured curves) and in the membrane (grey curves) (a). Colours in the dispersion curves correspond to the Q-factor values. Sloping curves represent an a-TE04 mode and nearly horizontal curves represent a-TEn2 modes. (b) Q-factor versus the stripe height (H/L) for the a-TE04 mode in VN gratings where F = 0.7 (black curve) and F = 0.9 (red curve), with Friedrich–Wintgen (FW) and interference-based (IB) BICs and qBICs. Numbers (1)–(5) indicate structures and corresponding modes illustrated in Figure 7. Figure 7: y-Directional proﬁles of the electric ﬁeld (black line in top panels) and light intensity distributions (bottom panels) for an a-TE04 mode within a single segment of the VN grating in the y–z plane. Numbers (1)–(5) correspond to the grating conﬁgurations indicated by the same numbers in Figure 6(b). The proﬁles of the electric ﬁeld are cross-sections through the light intensity maximum of the modes. 5 Influence of the substrate refractive index on Q-factor Thenumbers inthe regionsenclosed by the dashed black lines indicate the m indices of the TEnm modes.The hyperbolic blackdashedcurves showthe cutoffs of these modes and the vertical black dashed lines show their long-wavelength limits (see formula (2)). s-TE modes (Figure 8) can form BICs only in the case of the membrane and other conﬁgurations of VS gratings W. Głowadzka et al.: Bound states in gratings with broken symmetry 3987 Figure 8: BICs (dark red) and Q-factor values of qBICs given by colours in the domain of the normalized wavelength (L/λ) and the substrate refractive index (ns), for different duty cycle (F) values of antisymmetric modes. The dashed black lines in (f) show the regions where the groups of a-TEnm modes indicated by the m index reach their Q-factor maxima. Figure 8: BICs (dark red) and Q-factor values of qBICs given by colours in the domain of the normalized wavelength (L/λ) and the substrate refractive index (ns), for different duty cycle (F) values of antisymmetric modes. The dashed black lines in (f) show the regions where the groups of a-TEnm modes indicated by the m index reach their Q-factor maxima. thanthatpresentedinFigure3,whichinducesrecordQ-factor values for qBICs exceeding 1010. These regions are limited in ns as they require the cutoff for m index modes to be larger than the long wavelength limit of m + 2 index modes. case of substrate refractive indices as large as 3. The disadvantage of these conﬁgurations is the large H, which could hamper their real-world realization. Another mech- anism is crossing or anticrossing in the modes inducing qFW BICs. In Figure 3(a), the appearance of a qFW BIC was associated with crossing of the dispersion curves of the s-TE04 mode, which was conﬁned in the grating, and a leaky s-TE12 mode. Despite the weak overlap of the two modes, the repulsion of the imaginary parts of their wavenumbers results in a qFW BIC with a Q-factor larger than the Q-factors of neighbouring qIB BICs. In this example, we did not show the enhancement of Q-factor that occurs in the case of crossing of the dispersion curves for two modes conﬁned in the grating. Figure 8(d)–(f) shows broad regions in the ns–L/λ plane, where crossing between modes conﬁned in the grating occurs (e.g. the areas marked “2, 4” and “4, 6” in Figure 9(f)). 5 Influence of the substrate refractive index on Q-factor When F < 0.5, the a-TEn4 mode remains outside the air- subwavelength region. No BICs can be observed in these conﬁgurations, except for SP BICs. When F = 0.9, only a very few non-SP BICs can be observed in the region of crossing a-TEn4 and a-TEn6 modes. The coexistence of a-TEn4 and a-TEn6 modes for F = 0.8 and 0.9 is also worth noting, due to the presence of extremely high Q-factor resonances that are mostly qFW BICs. However, these resonances are limited to the refractive indices of the substrate below 2.5. We also found a resonance of Q ≈6000 related to a qFW BIC that exists in the lowest refractive index contrast conﬁguration, where ns = 3.5 and F = 0.9. The geometrical parameters of the grating correspond to our limit value of H/L = 3 and the long- wavelength limit of the a-TE12 mode L/λ = 1.8. Finally, qBICs with Q-factors of 103 in the case of F = 0.5 can be noticed even for ns = ng. The origin of these low Q-factor resonances, which exist even though the waveguiding condition is not fulﬁlled, will be the subject of future work. F = 0.9, and H/L = 2.3 at L/λ = 1.8. Remarkably, resonances of extremely high Q-factors exceeding the precision of numeri- cal calculations can be found in the region ns −1 < L/λ < 2ns −1 where symmetry protection does not hold and only one emission channel related to the ﬁrst diffraction order exists. These BICs can be found for F < 0.9 where crossings between a-TEn4 and leaky a-TEn2 induce strong repulsion of the imaginary parts of their wavenumbers, resulting in inﬁnite Q- factor resonances from the a-TEn4 modes. This tendency conﬁrms the results presented in Figures 5(a) and 6(a). When F < 0.5, the a-TEn4 mode remains outside the air- subwavelength region. No BICs can be observed in these conﬁgurations, except for SP BICs. When F = 0.9, only a very few non-SP BICs can be observed in the region of crossing a-TEn4 and a-TEn6 modes. The coexistence of a-TEn4 and a-TEn6 modes for F = 0.8 and 0.9 is also worth noting, due to the presence of extremely high Q-factor resonances that are mostly qFW BICs. However, these resonances are limited to the refractive indices of the substrate below 2.5. We also 5 Influence of the substrate refractive index on Q-factor The in- tersections of such modes lead to much stronger repulsion Figures 8 and 9 support our earlier observation that the spectral positions of the mode cutoffs are independent of the effective refractive index of the grating, since the hy- perbolic function representing the cutoff is the same in each map representing different F. However, the long- wavelength limit of the modes shifts toward longer wave- lengths as F increases. The a-TE modes reach longer wavelengths than the s-TE modes, as a consequence of zeros in the light distribution in the air slits of the a-TE modes. The a-TE modes (Figure 9) are SP BICs when L/λ < ns −1. This condition can be fulﬁlled by conﬁgurations with high values of ns only for high values of F. We observed an SP BIC for a maximal refractive index of thesubstratewhenns = 3.48, 3988 W. Głowadzka et al.: Bound states in gratings with broken symmetry 3988 W. Głowadzka et al.: Bound states in gratings with broken symmetry Figure 9: BICs (dark red) and Q-factor values of qBICs given by colours in the domain of the normalized wavelength (L/λ) and the substrate refractive index (ns), for different duty cycle (F) values of antisymmetric modes. The dashed black lines in f) show the regions where the groups of a-TEnm modes indicated by the m index reach their Q-factor maxima. Figure 9: BICs (dark red) and Q-factor values of qBICs given by colours in the domain of the normalized wavelength (L/λ) and the substrate refractive index (ns), for different duty cycle (F) values of antisymmetric modes. The dashed black lines in f) show the regions where the groups of a-TEnm modes indicated by the m index reach their Q-factor maxima. F = 0.9, and H/L = 2.3 at L/λ = 1.8. Remarkably, resonances of extremely high Q-factors exceeding the precision of numeri- cal calculations can be found in the region ns −1 < L/λ < 2ns −1 where symmetry protection does not hold and only one emission channel related to the ﬁrst diffraction order exists. These BICs can be found for F < 0.9 where crossings between a-TEn4 and leaky a-TEn2 induce strong repulsion of the imaginary parts of their wavenumbers, resulting in inﬁnite Q- factor resonances from the a-TEn4 modes. This tendency conﬁrms the results presented in Figures 5(a) and 6(a). 6 Example: modal characteristics of a 2D finite structure To support the simulations of an infinite structure pre- sented in Sections 3–5, we perform a numerical analysis of a two-dimensional subwavelength grating composed of ﬁnite number of stripes. The grating and the substrate are W. Głowadzka et al.: Bound states in gratings with broken symmetry 3989 made from uniform materials with refractive indices of 3.52 and 2.95, respectively. These values are close to the refractive indices of GaAs and Al-rich AlGaAs in the near infra-red range. Such a combination of materials enables real-world realization of all-semiconductor structures. The refractive indices are real numbers (with a zero imaginary part), which relates to the absence of gain and internal absorption. To demonstrate the difference in the impact of BIC and qBIC on the Q-factor of a ﬁnite structure, we chose two conﬁgurations. In the ﬁrst, F = 0.8, which enables a true BIC in an inﬁnite conﬁguration (see Figure 9(e)). In the second, F = 0.9, which provides a qBIC in an inﬁnite structure (see Figure 9(f)). Figure 10 shows the Q-factors of modes for both conﬁgurations in the domain of the height of the grating stripes, for an arbitrarily chosen range of the resonant wavelengths. The dispersion maps of the inﬁnite conﬁgurations are presented in Figure S6 in the Supple- mentary Materials. concerning electronic states in low-dimensional structures in solid state physics, we name the slowly varying function the “envelope”. The distribution of a mode in a ﬁnite structure can be seen approximately as a product of the distribution of a TEnm mode and an envelope. If we denote by p the number of zeros in the envelope function, we can denote the ﬁnite-structure mode as TEnmp. The dispersion lines of fundamental lateral modes (p = 0) presented in Figure 10 overlap closely with the dispersion lines of their inﬁnite counterparts, which are depicted in Figure S6 in Supplementary Materials. Figure 10 presents the spectral range in which symmetric and antisymmetric modes can be observed (s-TE modes are absent in the spectral range presented in Figure 10(a)). Direct comparison indicates that a-TEn4p modes reveal Q- factors 100 times higher than those of the s-TEn4p modes in the ﬁnite structure. This is a result of the several orders of magnitude higher Q-factors of a-TEn4 with respect to s-TEn4, as illustrated in Figure S6 in the Supplementary Materials. The ratio of the Q-factors of conﬁguration 1 and 2 is approximately equal throughout the whole range of the investigated number of grating stripes. However, the Q- factor of conﬁguration 2 is expected to show asymptotic behaviour when the Q-factor approaches the limit of 1012, related to the inﬁnite structure. The memory of the BIC is sustained in the ﬁnite structure due to reduced light emission in the vertical direction in conﬁguration 1 with respect to conﬁguration 2. Figure S10 in the Supplementary Materials enables qualitative comparison of light emitted vertically by conﬁgurations 1 and 2, indicating noticeably larger light emission by conﬁguration 2 in three variants composed of 20, 60, and 100 stripes. Lateral emission, which is the second source of the cavity losses, is compa- rable in both conﬁgurations as their effective masses are of comparable value, as discussed in the Supplementary Materials S2. bl l ld l f b d d Figure 11: Light intensity distribution within a VN grating composed of 20 periods for selected modes (a-TE040 (1), a-TE041 (2), a-TE140 (3) and a-TE141 (4)) in the whole grating (left column) and corresponding light distributions in the four central stripes (right column) as indicated in (1). The grating region is indicated in the left column by a white bordered rectangle, and the four central grating stripes are shown in right column. The parameters that are not shown in Figure 10 are F = 0.8, ns = 2.95. increases in the index p. In the case of antisymmetric modes, this tendency is reversed, which results from the opposite signs of the effective mass of the cavity modes (see Supplementary Materials S2). Comparison of the dispersion maps in the case of the finite structure shown in Figure 10, as well as in periodic structures (see Figure S6 Supplementary Materials), re- veals a correlation between qBICs in the inﬁnite conﬁgu- ration and the maxima of the Q-factors in ﬁnite conﬁgurations. In ﬁnite conﬁgurations, apart from the obviously smaller Q-factor, there is a relatively extended range of large Q-factors with respect to H. We relate this behaviour to the fact that for a qBIC occurring at a Γ−point in inﬁnite grating, a change in the geometric parameter, e.g. H, induces the qBIC to shift beyond the Γ−point [29]. 3990 Figure 11: Light intensity distribution within a VN grating composed of 20 periods for selected modes (a-TE040 (1), a-TE041 (2), a-TE140 (3) and a-TE141 (4)) in the whole grating (left column) and corresponding light distributions in the four central stripes (right column) as indicated in (1). The grating region is indicated in the left column by a white bordered rectangle, and the four central grating stripes are shown in right column. The parameters that are not shown in Figure 10 are F = 0.8, ns = 2.95. geometric parameters: H/L = 1.39, L/λ = 0.62, and F = 0.8 (conﬁguration 1) which corresponds to a true BIC in an inﬁnite structure (see Figure S6a in the Supplementary Materials). The considered structures differ with respect to the number of grating stripes. All conﬁgurations reveal emission in the vertical as well as lateral directions. How- ever, normal emission reduces with higher numbers of stripes. Figure 13 illustrates the Q-factor of the a-TE040 mode (black line) as a function of the number of grating stripes, in a conﬁguration with the same geometrical parameters as in Figure 12. For the sake of comparison, Figure 13 also shows the Q-factor of the same mode in a structure with the following geometric parameters: H/L = 1.36, L/λ = 0.58, and F = 0.9 (conﬁguration 2). In the case of the inﬁnite conﬁg- uration 2, there is a very large but ﬁnite Q-factor of 1012 (see Figure S6c in Supplementary Materials). The calculations for ﬁnite conﬁgurations 1 and 2 were limited to the maximal size of a structure composed of ∼100 stripes. This limitation is due to computational capability. In conﬁguration 1, Q-factors above 1000 can be achieved for structures composed of as few as 6 grating stripes, enabling very small mode volumes required for example for applications related to single photon devices. A Q-factor level above 104, which is required for achieving stimulated emission [29], is achieved by conﬁguration 1 composed of 20 stripes. Increasing the number of grating stripes increases the Q- factor nearly exponentially. When the Q-factors of the two structures are compared, conﬁguration 1 reveals an approximately 3–5 times larger Q-factor in comparison to conﬁguration 2, which indicates the inﬂuence of the true BIC in the inﬁnite conﬁguration on its ﬁnite counterpart. 6 Example: modal characteristics of a 2D finite structure Although zeroth diffraction order emission is prohibited in the case of a-TEn4 modes in the inﬁnite grating, it is permitted in the ﬁnite conﬁguration since ﬁnite structures do not totally cancel the emission. More- over, symmetric TEnmp modes of the same n and m indices reveal decreases in the wavelength (increased energy) with Figure 11 illustrates the distributions of the arbitrarily chosen modes indicated by numbers in Figure 10(a). On the scale of a single segment, the intensity distribution re- sembles that of a TEnm mode in the corresponding periodic structure. However, on the scale of the whole structure, another dependence is superimposed on the TEnm distri- bution, which varies slowly (compared to the single- segment scale). Analogously to similar considerations Figure 10: Q-factor of resonances in the domain of the VN grating thickness (H/L) and the normalized wavelength (L/λ). The VN grating is composed of 20 periods implemented on a substrate with a refractive index of 2.95 and a duty cycle F = 0.8 (a) and F = 0.9 (b). Numbers in brackets indicate the parameters of modes presented in Figure 11. Ellipses indicate dispersion curves of modes with the same light distribution within a single segment, corresponding to a-TEn4 or s-TEn4 modes. The dispersion maps of the inﬁnite conﬁguration with F = 0.9 are presented in Figure S6b, c and with F = 0.8 in Figure S6a in the Supplementary Materials. Figure 10: Q-factor of resonances in the domain of the VN grating thickness (H/L) and the normalized wavelength (L/λ). The VN grating is composed of 20 periods implemented on a substrate with a refractive index of 2.95 and a duty cycle F = 0.8 (a) and F = 0.9 (b). Numbers in brackets indicate the parameters of modes presented in Figure 11. Ellipses indicate dispersion curves of modes with the same light distribution within a single segment, corresponding to a-TEn4 or s-TEn4 modes. The dispersion maps of the inﬁnite conﬁguration with F = 0.9 are presented in Figure S6b, c and with F = 0.8 in Figure S6a in the Supplementary Materials. 3990 W. Głowadzka et al.: Bound states in gratings with broken symmetry 7 Conclusions In this paper, we have presented the results of an extensive numerical analysis of Fano resonance, which is respon- sible for the creation of BICs in one-dimensional gratings with broken up-down mirror symmetry. Such configura- tions are composed of a grating implemented on a sub- strate with a different refractive index to that of the air on the opposite side of the grating. The obvious disadvantage of such a configuration is the collapse of nonsymmetry- protected BICs (FW BICs and IB BICs), as has been explored in [13, 14]. On the other hand, a very important advantage, which may justify a compromise with regard to the collapse of BICs, is the possibility of realizing all-semiconductor conﬁgurations that in the future could enable real-world realizations of semiconductor-based electrically driven devices exploiting quasi-BICs. Figure 13: Q-factor and relative wavelength (ΔL/λ) versus the number of grating stripes for conﬁguration 1 (black line, H/L = 1.39, L/λ ≈0.62 and F = 0.8) and conﬁguration 2 (blue line, H/L ≈1.36, L/ λ ≈0.58 and F = 0.9). ΔL/λ is a differencebetween the L/λ in the ﬁnite stripe grating and that of the inﬁnite grating. to effects not considered in this analysis related to material absorption, surface roughness, disorder effects, and radi- ation in the x direction, which limit the Q-factor to 105 [34, 35]. If taken into consideration in the analysis, these effects would reduce the Q-factor of both conﬁgurations to a comparable extent. However, the ratio of Q-factors of the two analysed cavities is expected to be at the same of 3–5 level if both conﬁgurations are realized. to effects not considered in this analysis related to material absorption, surface roughness, disorder effects, and radi- ation in the x direction, which limit the Q-factor to 105 [34, 35]. If taken into consideration in the analysis, these effects would reduce the Q-factor of both conﬁgurations to a comparable extent. However, the ratio of Q-factors of the two analysed cavities is expected to be at the same of 3–5 level if both conﬁgurations are realized. We studied the formation of BICs and quasi-BICs in an infinite one-dimensional grating with a refractive index of 3.52 implemented on a substrate. The refractive index of the substrate was modified from 1 to the refractive index of the grating. The refractive index on the superstrate was kept at 1. Modes in ﬁnite structures, particularly those composed of a low number of stripes, contain wavevector distribu- tions with signiﬁcantly divergent transverse wavevector component. Thus, large Q-factors can be maintained for a wide range of H, due to the presence of wavevectors whose transverse components are close to those corresponding to qBIC in an inﬁnite structure. Figure 12 illustrates the light intensity distribution of a-TE040 mode in conﬁgurations with the following Possible real-world realization of BIC-based periodic cavities would entail further reduction of the Q-factor, due W. Głowadzka et al.: Bound states in gratings with broken symmetry 3991 Figure 12: Light intensity distribution on a logarithmic scale of colours for the a-TE040 mode of a VN grating with ns = 2.95 composed of 20 (a), 60 (b), and 100 (c) grating stripes with L/λ ≈0.62. The other grating parameters are F = 0.8 and H/L = 1.39. The white borders of the rectangles indicate the grating region. Figure 12: Light intensity distribution on a logarithmic scale of colours for the a-TE040 mode of a VN grating with ns = 2.95 composed of 20 (a), 60 (b), and 100 (c) grating stripes with L/λ ≈0.62. The other grating parameters are F = 0.8 and H/L = 1.39. The white borders of the rectangles indicate the grating region. Figure 13: Q-factor and relative wavelength (ΔL/λ) versus the number of grating stripes for conﬁguration 1 (black line, H/L = 1.39, L/λ ≈0.62 and F = 0.8) and conﬁguration 2 (blue line, H/L ≈1.36, L/ λ ≈0.58 and F = 0.9). ΔL/λ is a differencebetween the L/λ in the ﬁnite stripe grating and that of the inﬁnite grating. and air, which enhanced the conﬁnement of light in the grating. Figure S11 in the Supplementary Materials in- dicates that a working device would remain stable under the changing thermal conditions that arise during contin- uous wave operation. Thermal shift of the resonant wave- length and Q-factor are insigniﬁcant in terms of the efﬁciency of the operating device. References [1] K. Kishino, M. S. Unlu, J. I. Chyi, J. Reed, L. Arsenault, and H. Morkoc, “Resonant cavity-enhanced (RCE) photodetectors,” IEEE J. Quant. Electron., vol. 27, pp. 2025–2034, 1991. [2] A. J. Campillo, J. D. Eversole, and H.-B. Lin, “Cavity quantum electrodynamic enhancement of stimulated emission in microdroplets,” Phys. Rev. Lett., vol. 67, pp. 437–440, 1991. We also performed an exemplary analysis of a finite structure composed of a finite number of grating stripes with a refractive index of 3.52 and a substrate with a refractive index of 2.95, both corresponding to the refractive indices of arsenide-based materials. In such a relatively low contrast structure, we calculated that increasing the number of grating stripes enables a nearly exponential increase in the Q-factor. We also showed the impact of the true BIC and the quasi BIC in certain inﬁnite conﬁgurations on the Q-factor of their ﬁnite versions demonstrating 3–5 larger Q-factor in the case of true-BIC-based ﬁnite conﬁguration. We also showed that the Q-factor of the resonance is robust against tem- perature changes that occur in electrically-driven semi- conductor devices under continuous wave operation. [3] J. Bravo-Abad, A. Rodriguez, P. Bermel, S. G. Johnson, J. D. Joannopoulos, and M. Soljačić, “Enhanced nonlinear optics in photonic-crystal microcavities,” Opt. Express, vol. 15, pp. 16161–16176, 2007. [4] C. Schneider, T. Heindel, A. Huggenberger, et al., “Microcavity enhanced single photon emission from an electrically driven site-controlled quantum dot,” Appl. Phys. Lett., vol. 100, 2012, Art no. 091108. [5] H. Deng, G. Weihs, C. Santori, J. Bloch, and Y. Yamamoto, “Condensation of semiconductor microcavity exciton polaritons,” Science, vol. 298, pp. 199–202, 2002. [6] J. Klaers, J. Schmitt, F. Vewinger, and M. Weitz, “Bose–Einstein condensation of photons in an optical microcavity,” Nature, vol. 468, pp. 545–548, 2010. [7] J. Kasprzak, M. Richard, S. Kundermann, et al., “Bose–Einstein condensation of exciton polaritons,” Nature, vol. 443, pp. 409–414, 2006. More generally, we demonstrated that in structures with broken up-down mirror symmetry not only symmetry- protected BICs can appear, as expected, but also FW and IB BICs, which had not been reported previously. The two conditions necessary to observe BICs are that the effective refractive index of the grating should be larger than the refractive index of the substrate and that the grating should be subwavelength with respect to air (or more generally with respect to the lower refractive index that surrounds the grating). 7 Conclusions We showed that reducing the up-down mirror symmetry leads to an abrupt collapse of BICs in the case of symmetric modes. FW and IB BICs transform to quasi-FW BICs and quasi-IB BICs, with a noticeable preponderance Finally, we examined the influence of changes in the refractive index induced by temperature variation, which reproduced the effect of temperature rising during opera- tion of the device. We assumed a temperature dependence of 2 × 10−4 K−1 for the refractive index of the grating and the substrate, which is close to the thermal dependence of the refractive index of GaAs. Increasing the temperature increased the refractive index contrast between the grating 3992 W. Głowadzka et al.: Bound states in gratings with broken symmetry 3992 of quasi-FW BICs compared to quasi-IB BICs in terms of Q-factor. Doctoral Candidate in the Interdisciplinary Doctora School at Lodz University of Technology, Poland. Conﬂict of interest statement: The authors declare no conﬂicts of interest regarding this article. Doctoral Candidate in the Interdisciplinary Doctoral School at Lodz University of Technology, Poland. Conﬂict of interest statement: The authors declare no conﬂicts of interest regarding this article. Analogous analysis in the case of nonsymmetric modes indicated existence of FW and IB BICs in the structures with broken up–down mirror symmetry. Their existence can be explained by destructive interference occurring in the single emission channel, which is related to the first diffraction order being emitted towards the substrate. These BICs are present in the structures with grating duty cycles ranging from 0.5 to 0.8. They result from the repulsion of low Q-factor modes, although avoi- ded crossing is absent. W. Głowadzka et al.: Bound states in gratings with broken symmetry [27] M. Dems, R. Kotyński, and K. Panajotov, “Planewave admittance method – a novel approach for determining the electromagnetic modes in photonic structures,” Opt. Express, vol. 13, pp. 3196–3207, 2005. [17] A. Taghizadeh and I.-S. Chung, “Dynamical dispersion engineering in coupled vertical cavities employing a high- contrast grating,” Sci. Rep., vol. 7, 2017, Art no. 2123. [18] J. Lee, B. Zhen, S.-L. Chua, et al., “Observation and differentiation of unique high Q optical resonances near zero wave vector in macroscopic photonic crystal slabs,” Phys. Rev. Lett., vol. 109, 2012, Art no. 067401. [28] M. Dems, “Modelling of high-contrast grating mirrors. the impact of imperfections on their performance in VCSELs,” Opto- Electron. Rev., vol. 19, pp. 340–345, 2011. [19] S. Dai, L. Liu, D. Han, and J. Zi, “From topologically protected coherent perfect reﬂection to bound states in the continuum,” Phys. Rev. B, vol. 98, 2018, Art no. 081405. [29] T. Czyszanowski, M. Gębski, E. Pruszyńska-Karbownik, M. Wasiak, and J. A. Lott, “Monolithic high-contrast grating planar microcavities,” Nanophotonics, vol. 9, pp. 913–925, 2020. [20] C. J. Chang-Hasnain and W. Yang, “High-contrast gratings for integrated optoelectronics,” Adv. Opt Photon, vol. 4, pp. 379–440, 2012. [30] M. Marciniak, A. Broda, M. Gębski, et al., “Tuning of reﬂection spectrum of a monolithic high-contrast grating by variation of its spatial dimensions,” Opt. Express, vol. 28, p. 20967–20977, 2020. [21] S. I. Azzam, V. M. Shalaev, A. Boltasseva, and A. V. Kildishev, “Formation of bound states in the continuum in hybrid plasmonic- photonic systems,” Phys. Rev. Lett., vol. 121, 2018, Art no. 253901. [31] A. Snyder and J. Love, Optical Waveguide Theory, US, Springer, 1984. [32] J. Winn, J. Joannopoulos, S. Johnson, and R. Meade, Photonic Crystals: Molding the Flow of Light – Second Edition, Princeton, Princeton University Press, 2008. [22] J. M. Dobrindt, I. Wilson-Rae, and T. J. Kippenberg, “Parametric normal-mode splitting in cavity optomechanics,” Phys. Rev. Lett., vol. 101, 2008, Art no. 263602. [33] M. Müller, F.-M. Dittes, W. Iskra, and I. Rotter, “Level repulsion in the complex plane,” Phys. Rev. E, vol. 52, pp. 5961–5973, 1995. [23] C. W.Hsu, B. Zhen,J.Lee,etal.,“Observationoftrappedlightwithin the radiation continuum,” Nature, vol. 499, pp. 188–191, 2013. [34] Z. Liu, Y. Xu, Y. Lin et al., “High-Q quasibound states in the continuum for nonlinear metasurfaces,” Phys. Rev. Lett., vol. 123, 2019, Art no. 253901. [24] A. Kodigala, T. References This conclusion may also apply to two- dimensional gratings, as they obey analogous wave- guiding phenomena. The calculations presented in this paper relate to a refractive index of the grating corre- sponding to narrow band gap semiconductors operating in the infrared spectrum, but the same mechanisms are ex- pected to occur in many other materials where different refractive indices induce waveguiding in the grating. [8] O. Hess, J. Pendry, S. Maier, R. Oulton, J. Hamm, and K. Tsakmakidis, “Active nanoplasmonic metamaterials,” Nat. Mater., vol. 11, pp. 573–584, 2012. [9] M. Limonov, M. Rybin, A. Poddubny, and Y. Kivshar, “Fano resonances in photonics,” Nat. Photonics, vol. 11, pp. 543–554, 2017. [10] V. Karagodsky and C. J. Chang-Hasnain, “Physics of near- wavelength high contrast gratings,” Opt. Express, vol. 20, pp. 10888–10895, 2012. [11] D. C. Marinica, A. G. Borisov, and S. V. Shabanov, “Bound states in the continuum in photonics,” Phys. Rev. Lett., vol. 100, 2008, Art no. 183902. [12] Y. Plotnik, O. Peleg, F. Dreisow, et al., “Experimental observation of optical bound states in the continuum,” Phys. Rev. Lett., vol. 107, 2011, Art no. 183901. [13] S.-G. Lee, S.-H. Kim, andC.-S. Kee,“Bound statesin the continuum (BIC) accompanied by avoided crossings in leaky-mode photonic lattices,” Nanophotonics, vol. 9, pp. 4373–4380, 2020. [14] Z. F. Sadrieva, I. S. Sinev, K. L. Koshelev, et al., “Transition from optical bound states in the continuum to leaky resonances: role of substrate and roughness,” ACS Photonics, vol. 4, pp. 723–727, 2017. Author contribution: All the authors have accepted responsibility for the entire content of this submitted manuscript and approved submission. [15] F. Wu, J. Wu, Z. Guo, et al., “Giant enhancement of the goos- hänchen shift assisted by quasibound states in the continuum,” Phys. Rev. Appl., vol. 12, 2019, Art no. 014028. Research funding: This work is supported by the Polish National Science Centre within the projects Opus (2018/29/ B/ST7/01927) and Preludium (2018/29/N/ST7/02460). This work has been completed while the ﬁrst author was the [16] C. W. Hsu, B. Zhen, A. D. Stone, J. Joannopoulos, and M. Soljačić, “Bound states in the continuum,” Nat. Rev. Mater., vol. 1, 2016, Art no. 16048. 3993 Lepetit, Q. Gu, B. Bahari, Y. Fainman, and B. Kanté, “Lasing action from photonic bound states in continuum,” Nature, vol. 541, pp. 196–199, 2017. [35] J. Jin, X. Yin, L. Ni, M. Soljačić, B. Zhen, and C. Peng, “Topologically enabled ultrahigh-Q guided resonances robust to out-of-plane scattering,” Nature, vol. 574, pp. 501–504, 2019. [25] S. Kim, “Coherent nonlinear phenomena in subwavelength- grating based microcavities,” Ph.D. Thesis, The University of Michigan, 2019. [26] S. V. Sreenivasan, “Nanoimprint lithography steppers for volume fabrication of leading-edge semiconductor integrated circuits,” Microsys. Nanoeng., vol. 3, pp. 1–19, 2017. [26] S. V. Sreenivasan, “Nanoimprint lithography steppers for volume fabrication of leading-edge semiconductor integrated circuits,” Microsys. Nanoeng., vol. 3, pp. 1–19, 2017. Supplementary Material: The online version of this article offers supplementary material (https://doi.org/10.1515/nanoph-2021-0319).
https://openalex.org/W2614916965	https://dash.harvard.edu/bitstream/1/41461301/1/87139%20pnas.201701805.pdf	English	null	The general form of Hamilton’s rule makes no predictions and cannot be tested empirically	Proceedings of the National Academy of Sciences of the United States of America	2,017	cc-by	8,280	Terms of Use This article was downloaded from Harvard University’s DASH repository, and is made available under the terms and conditions applicable to Other Posted Material, as set forth at http:// nrs.harvard.edu/urn-3:HUL.InstRepos:dash.current.terms-of-use#LAA Permanent link http://nrs.harvard.edu/urn-3:HUL.InstRepos:41461301 Citation Nowak, Martin A., Alex McAvoy, Benjamin Allen, and Edward O. Wilson. 2017. “The General Form of Hamilton’s Rule Makes No Predictions and Cannot Be Tested Empirically.” Proceedings of the National Academy of Sciences 114 (22): 5665–70. https://doi.org/10.1073/pnas.1701805114. The general form of Hamilton’s rule makes no predictions and cannot be tested empirically aProgram for Evolutionary Dynamics, Harvard University, Cambridge, MA 02138; bDepartment of Mathematics, Harvard University, Cambridge, MA 02138; cDepartment of Organismic and Evolutionary Biology, Harvard University, Cambridge, MA 02138; dDepartment of Mathematics, Emmanuel College, Boston, MA 02115; and eMuseum of Comparative Zoology, Harvard University, Cambridge, MA 02138 uted by Edward O. Wilson, April 13, 2017 (sent for review February 2, 2017; reviewed by Michael Doebeli and Jan Rychtar whether or not the trait has increased. In particular, the parame- ters B and C depend on the change in average trait value. Hamilton’s rule asserts that a trait is favored by natural selection if the beneﬁt to others, B, multiplied by relatedness, R, exceeds the cost to self, C. Speciﬁcally, Hamilton’s rule states that the change in average trait value in a population is proportional to BR −C. This rule is commonly believed to be a natural law mak- ing important predictions in biology, and its inﬂuence has spread from evolutionary biology to other ﬁelds including the social sci- ences. Whereas many feel that Hamilton’s rule provides valuable intuition, there is disagreement even among experts as to how the quantities B, R, and C should be deﬁned for a given system. Here, we investigate a widely endorsed formulation of Hamilton’s rule, which is said to be as general as natural selection itself. We show that, in this formulation, Hamilton’s rule does not make predictions and cannot be tested empirically. It turns out that the parameters B and C depend on the change in average trait value and therefore cannot predict that change. In this formula- tion, which has been called “exact and general” by its proponents, Hamilton’s rule can “predict” only the data that have already been given. The second astonishing fact of HRG is that the prediction, which exists only in retrospect, is not based on relatedness or any other aspect of population structure. A common interpre- tation of the terms in Hamilton’s rule is that R quantiﬁes the population structure, whereas B and C characterize the nature of the trait. But the derivation shows that this interpretation is wrong. All three terms, B, R, and C, are functions of popula- tion structure, whereas the overall value of BR −C is function- ally independent of population structure. Any information about who interacts with whom cancels out when calculating the value of BR −C. Author contributions: M.A.N., A.M., B.A., and E.O.W. designed research, performed research, analyzed data, and wrote the paper. Reviewers: M.D., University of British Columbia; and J.R., The University of North Carolina at Greensboro. The authors declare no conﬂict of interest. Freely available online through the PNAS open access option. 1To whom correspondence should be addressed. Email: ewilson@oeb.harvard.edu. The general form of Hamilton’s rule makes no predictions and cannot be tested empirically The third fact of HRG is that no conceivable experiment exists that could test (or invalidate) this rule. All input data, whether they come from biology or not, are formally in agreement with HRG. This agreement is not a consequence of natural selection, but a statement about a relationship between slopes in multivari- ate linear regression. This relationship between slopes has been known in statistics at least since 1897 (15). Derivation of HRG We recapitulate the derivation of HRG given in refs. 4, 5, 7, 9, and 10. We also provide explicit algebraic formulas for B and C that result from this derivation. H H amilton’s rule is a widely known concept in evolutionary biology. It has become standard textbook knowledge and is encountered in undergraduate education. For many, Hamilton’s rule expresses the intuition that cooperation evolves more easily when there are frequent interactions among relatives, because relatives are likely to share the cooperative trait. However, Hamilton’s rule goes beyond this intuition by positing a quan- titative condition, BR −C > 0, which is said to predict whether or not a trait will be selected. Speciﬁcally, it is claimed that the change in average trait value from one time point to the next is proportional to BR −C. We imagine a population of n individuals at a given point in time. Each individual i has a ﬁtness value, wi. The list w = (w1, w2, . . . , wn) is the collection of ﬁtness values in the population, which can be interpreted as expected or realized number of offspring in the next generation. If the total popu- lation size is constant, which is assumed for simplicity, then the average ﬁtness is ¯w = 1. www.pnas.org/cgi/doi/10.1073/pnas.1701805114 Share Your Story The Harvard community has made this article openly available. Please share how this access benefits you. Submit a story . The Harvard community has made this article openly available. Please share how this access benefits you. Submit a story . Accessibility Signiﬁcance We immediately encounter the question of how the “bene- ﬁt,” B, the “relatedness,” R, and the “cost,” C, are calculated for a given system. Surprisingly, there is no consensus about the correct method. A variety of derivations have been proposed over the years (1–10), which deﬁne B, R, and C in distinct (nonequivalent) ways. In the empirical literature, peer reviewers often disagree over which method should be used in a particular manuscript (11). Hamilton’s rule is a well-known concept in evolutionary biol- ogy. It is usually perceived as a statement that makes pre- dictions about natural selection in situations where inter- actions occur between genetic relatives. Here, we examine what has been called the “exact and general” formulation of Hamilton’s rule. We show that in this formulation, which is widely endorsed by proponents of inclusive ﬁtness theory, Hamilton’s rule does not make any prediction and cannot be tested empirically. This formulation of Hamilton’s rule is not a consequence of natural selection and not even a statement speciﬁcally about biology. We give simple and transparent expressions for the quantities of beneﬁt, cost, and related- ness that appear in Hamilton’s rule, which reveal that these quantities depend on the data that are to be predicted. p ( ) A number of recent papers (7, 9, 10) have endorsed a particu- lar formulation (4, 5) as the exact, general, and even “canonical” version of Hamilton’s rule. This formulation, called “Hamilton’s rule—general” (HRG) (12, 13), is claimed to be as general as natural selection itself (7, 14). The derivation, which we recapit- ulate below, is simple and contains only a few steps. p y p The mathematical investigation of HRG reveals three aston- ishing facts. First, HRG is logically incapable of making any pre- diction about any situation because the beneﬁt, B, and the cost, C, cannot be known in advance. They depend on the data that are to be predicted. At the outset of an experiment, B and C are unknown, and so there is no way to say what Hamilton’s rule would predict. Once the experiment is done, HRG will produce B and C values in retrospect such that BR −C is positive if the trait in question has increased and negative if it has decreased. But these “predictions” are merely rearrangements of the data that have been collected and already contain information about EVOLUTION 1To whom correspondence should be addressed. Author contributions: M.A.N., A.M., B.A., and E.O.W. designed research, performed research, analyzed data, and wrote the paper. Signiﬁcance Each individual is labeled by its ﬁtness value, wi. (B) The equivalent table representation shows the input data for ﬁtness, w; trait value, g; and trait value of inter- action partners, h. (C) The change in trait value, ∆¯g, is proportional to the slope of the regression w vs. g. (D) Slope of the regression w vs. h. (E) There are two different relatedness values for the slopes of h vs. g and g vs. h. (F) The multivariate linear regression of w vs. g and h gives the slopes, Mh = B and Mg = −C, that are called beneﬁt and cost. The slopes are related as follows: mwg = Mg + Mhmhg and mwh = Mh + Mgmgh. The ﬁrst of these two equations is Hamilton’s rule. This relationship between slopes is a conse- quence of multivariate linear regression and has been known in statistics at least since 1897 (15). C = −vh∆¯g −Rchwvg vgvh −R2v 2g . [4b] [4b] The parameters B and C depend on R and therefore on popula- tion structure. Therefore, “beneﬁt” and “cost” are not just prop- erties of a trait but depend on relatedness. It follows that chang- ing relatedness would typically also change beneﬁt and cost. This dependency is at odds with many empiricists’ intuition about Hamilton’s rule. Both B and C also depend on ∆¯g, which is the change in aver- age trait value. The change in average trait value is supposed to be predicted by Hamilton’s rule, but the parameters B and C depend on this quantity. Therefore, it makes no sense to claim that HRG makes a prediction about ∆¯g. p g The derivation is completed by calculating the term, BR −C, which leads to Each individual is assigned a trait value, gi, which can indi- cate the presence or absence of a genetic mutation, or it can quantify the genetic predisposition to some phenotype. For brevity, we use the term “trait” to indicate the genetic propen- sity toward a particular trait. The list g = (g1, g2, . . . , gn) is the collection of trait values in the population. The average trait value is ¯g = Pn i=1 gi/n. The ﬁrst and second lists together specify the average trait value in the next generation, which is ¯g′ = Pn i=1 giwi/n. Signiﬁcance For each individual i, the quantity hi represents the average trait value of i’s interac- tion partners. The list h = (h1, h2, . . . , hn) summarizes the inter- actions in the population at the given time. p p g Some further notation is needed. The variance of a list of num- bers is vg = Pn i=1 g2 i /n −¯g2. The covariance of two lists, g and w, is cgw = Pn i=1 giwi/n −¯g ¯w. Note that cgg = vg. D D C E C D F , gw P i=1 gi i/ g gg g Because ¯w = 1, the change in average trait value is , gw P i=1 gi i/ g gg g Because ¯w = 1, the change in average trait value is ∆¯g = cgw. [1] [1] HRG deﬁnes the parameter R as the slope of a best-ﬁt line for the data in the plane h vs. g (Fig. 1E). The formula for the slope is HRG deﬁnes the parameter R as the slope of a best-ﬁt line for the data in the plane h vs. g (Fig. 1E). The formula for the slope is R = cgh vg . [2] [2] The covariance, cgh, contains terms of the form gihi, which is the product of the trait value of individual i and the average trait value of the interaction partners of individual i. Therefore, cgh and consequently R depend on population structure. F F E E The derivation continues by calculating a best-ﬁt plane to the data given in the 3D space w vs. g and h (Fig. 1F). The param- eters B and C are deﬁned by the slopes of this plane. The alge- braic formulas are B = vgchw −cghcgw vgvh −c2 gh ; [3a] B = vgchw −cghcgw vgvh −c2 gh ; [3a] C = −vhcgw −cghchw vgvh −c2 gh . [3b] [3a] C = −vhcgw −cghchw vgvh −c2 gh . [3b] [3b] These expressions can be written as These expressions can be written as B = chw −R∆¯g vh −R2vg ; [4a] [4a] Fig. 1. Hamilton’s rule—general (HRG) is a relationship among slopes in mul- tivariate linear regression. (A) A population of individuals with pairwise interactions. Blue indicates the presence of a trait, gi = 1, and red indicates the absence of the trait, gi = 0. Links imply interactions. 5666 \| www.pnas.org/cgi/doi/10.1073/pnas.1701805114 Signiﬁcance Email: ewilson@oeb.harvard.edu. PNAS \| May 30, 2017 \| vol. 114 \| no. 22 \| 5665–5670 A B C D E F Fig. 1. Hamilton’s rule—general (HRG) is a relationship among slopes in mul- tivariate linear regression. (A) A population of individuals with pairwise interactions. Blue indicates the presence of a trait, gi = 1, and red indicates the absence of the trait, gi = 0. Links imply interactions. Each individual is labeled by its ﬁtness value, wi. (B) The equivalent table representation shows the input data for ﬁtness, w; trait value, g; and trait value of inter- action partners, h. (C) The change in trait value, ∆¯g, is proportional to the slope of the regression w vs. g. (D) Slope of the regression w vs. h. (E) There are two different relatedness values for the slopes of h vs. g and g vs. h. (F) The multivariate linear regression of w vs. g and h gives the slopes, Mh = B and Mg = −C, that are called beneﬁt and cost. The slopes are related as follows: mwg = Mg + Mhmhg and mwh = Mh + Mgmgh. The ﬁrst of these two equations is Hamilton’s rule. This relationship between slopes is a conse- quence of multivariate linear regression and has been known in statistics at least since 1897 (15). A B D Although the ﬁrst two lists determine the value of BR −C, as well as the overall genetic change in the population, a third list is required to determine the individual values of B, R, and C. B H il ’ l i d ib h h Although the ﬁrst two lists determine the value of BR −C, as well as the overall genetic change in the population, a third list is required to determine the individual values of B, R, and C. Because Hamilton’s rule is meant to describe phenomena such as kin selection or social evolution, the third list contains informa- tion about interactions between individuals. For each individual i, the quantity hi represents the average trait value of i’s interac- tion partners. The list h = (h1, h2, . . . , hn) summarizes the inter- actions in the population at the given time. B A Because Hamilton s rule is meant to describe phenomena such as kin selection or social evolution, the third list contains informa- tion about interactions between individuals. Signiﬁcance The difference between these two quantities is the change in average trait value from one generation to the next: ∆¯g = ¯g′ −¯g. W ill h h li f ll if h i l l BR −C = cgw vg . [5 [5] Note that all hi values have canceled out in this calculation. Therefore, the prediction BR −C, which is equal to cgw/vg, has discarded all information about the h list, which denotes the trait list of interaction partners. Whereas the individual values of B, of R, and of C contain the h list, the value of BR −C does not. Any collection of numbers that is used for the h list, as long as the denominators in Eqs. 3a and 3b are nonzero, gives the same value of BR −C. For example, one can use the digits of π and obtain the same prediction for the change in average trait value (Fig. 2). The numbers of the h list affect the individual values of B and C, but they do not affect the value of BR −C. We will see that these two lists fully specify the numerical value of BR −C, which is equal to a positive quantity times ∆¯g. Thus, BR −C and ∆¯g have the same sign. Whenever a trait increases (or decreases), the sign of BR −C is positive (or negative). However, this result does not produce a prediction because both BR −C and ∆¯g are calculated from the same lists of numbers. Nowak et al. A B C Fig. 2. The retrospective prediction of Hamilton’s rule is based on the numerical value of the term BR −C. (A and B) The numerical value of BR −C does not depend (functionally) on the h list, which speciﬁes the trait values of the interaction partners. Any (generic) choice of numbers can be used for the h list, for example the digits of π, and the value of BR −C remains the same. The prediction of this form of Hamilton’s rule therefore does not use information about who interacts with whom or on whether interactions are between relatives or not. (C) The numerical value of R depends on the g and h lists. The numerical values of B and C depend on all three lists. In particular, B and C also depend on the change in trait value. Slopes and Statistics The relationship between the slopes of the various linear regres- sions, expressed by HRG (Eq. 5), is not a consequence of biology and not a discovery of inclusive ﬁtness theory. It is a fact of multi- variatestatisticalanalysis,theproofofwhichwerecallinAppendix. y p pp In a linear regression of w vs. g, the slope of the line is mwg = cwg/vg. Likewise, a linear regression of w vs. h gives slope mwh = cwh/vh. Furthermore, the linear regressions of h vs. g and of g vs. h lead to the slopes mhg = chg/vg and mgh = cgh/vh, respectively. Signiﬁcance Therefore, they cannot be used to predict that change in any meaningful way. C B Fig. 2. The retrospective prediction of Hamilton’s rule is based on the numerical value of the term BR −C. (A and B) The numerical value of BR −C does not depend (functionally) on the h list, which speciﬁes the trait values of the interaction partners. Any (generic) choice of numbers can be used for the h list, for example the digits of π, and the value of BR −C remains the same. The prediction of this form of Hamilton’s rule therefore does not use information about who interacts with whom or on whether interactions are between relatives or not. (C) The numerical value of R depends on the g and h lists. The numerical values of B and C depend on all three lists. In particular, B and C also depend on the change in trait value. Therefore, they cannot be used to predict that change in any meaningful way. However, it is well understood in statistics that relationships such as Eq. 6 do not themselves imply causality (18, 19). Whereas there can be causal relationships between dependent and inde- pendent variables in a linear model, these relationships cannot be deduced from linear regression alone (20, 21). Therefore, with- out further assumptions or information, the meanings attached to the terms in HRG have no basis in mathematics or statistics. Moreover, the derivation of HRG does not take into account any aspect of the mechanism that leads to a change in trait value and therefore cannot return a description of that mechanism (Fig. 3) (22). It merely deﬁnes quantities B, R, and C as functions of w, g, and h such that BR −C is proportional to the change in trait frequency, ∆g. The formal correctness of HRG is established by Eqs. 1 and 5. Because the variance vg must be positive for the formalism to make sense, the sign of BR−C is the same as the sign of ∆¯g. But we have seen that B and C depend on ∆¯g. Therefore, BR −C cannot predict or explain ∆¯g. The expression BR −C is simply an extended form of writing the ratio ∆¯g/vg, in which we add an arbitrary h list that subsequently cancels out again. [6b] mwh = Mh + Mgmgh. [6b] See Fig. 1 for a graphical depiction of each term. Inclusive ﬁtness theorists call Eq. 6a “Hamilton’s rule” by set- ting B = Mh, C = −Mg, and R = mhg. These quantities are inter- preted as beneﬁt, cost, and relatedness (7, 9) and are used to clas- sify behaviors as “altruistic,” “mutually beneﬁcial,” “selﬁsh,” or “spiteful” (17). If B and C are positive, the situation is classiﬁed as altruism; if furthermore BR −C > 0, then it is concluded that the altruistic trait increases because interactions occur between close relatives. EVOLUTION Nowak et al. Beneﬁt and Cost Need Not Make Sense Although HRG does not explain or predict the change in aver- age trait value, it could be the case that the parameters B and C provide some biological insights. In a previous paper, we showed that it is easy to envisage biological processes that are mischar- acterized by the resulting values of B and C (22). In Fig. 3, we provide further examples of this kind. p y In three dimensions, the multivariate regression of w vs. both g and h leads to a plane given by two slopes Mh and Mg. It is standard textbook knowledge (16) that the relationship between the slopes is given by the linear system of equations p p The parameters B and C also behave unbiologically in the fol- lowing way: A small change in a single gi or hi value can make B and C jump from a very large negative value to a very large positive value (Fig. 4). For example, a small deviation in the mea- surement of an empirical system could change the assessment of a behavior from tremendously helpful (say, B = 1010) to tremen- dously harmful (B = −1010). Such quantities are not biologically meaningful. mwg = Mg + Mhmhg; [6a] mwh = Mh + Mgmgh. [6b] mwg = Mg + Mhmhg; [6a] [6a] Discussion The existence of these conﬂicting deﬁnitions makes it impossi- ble to meaningfully test or falsify Hamilton’s rule. Any theoretical or empirical result that appears to violate Hamilton’s rule can be reanalyzed using HRG to show that the outcome is “as predicted by Hamilton’s rule.” Indeed, this pattern has been repeated many times in the literature (7, 14, 28, 32–34). It appears that there are no real or hypothetical data that the inclusive ﬁtness community would accept as a violation of Hamilton’s rule. The predictive power of HRG is equivalent to the following example: If you give me the shoe sizes and heights of a group of people, then I can predict the heights. My algorithm also works if you gave me the wrong shoe sizes. y g g That HRG has no predictive power has been previously noted by ourselves (22) and others (12, 23, 24), yet HRG is credited with making a variety of empirical predictions (7, 9, 14). g y p p ( ) Much like the Price equation (25–27), HRG provides a func- tional relationship between quantities that are obtained from a population at two successive points in time. Whereas the change in trait frequency, ∆g, need not be independent of h in a statis- tical sense, the derivation of HRG takes neither statistical rela- tionships nor any information about suitability of a linear model into account (7, 22). Starting from three lists, w, g, and h, it ﬁts a linear model of w vs. g and h and ﬁnds B, C, and R such that BR −C is proportional to ∆g. Notably, B and C are themselves functions of ∆g. The result is an algebraic expression for BR−C that is (functionally) independent of h (Fig. 2). In analogy to our previous example, although shoe size and height could be cor- related, if we already know the heights, the shoe sizes are not needed to determine heights. p Some papers attempt to empirically test Hamilton’s rule (35– 40). Tests of Hamilton’s rule are typically done by experimentally determining the beneﬁts and costs of a phenotype and quanti- fying relatedness using genetic markers or pedigree. But such a procedure—while scientiﬁcally reasonable—tests only HRS, which is not the exact and general version of Hamilton’s rule. We are aware of only one paper (23) that attempts to apply HRG to an empirical system. Discussion In all four cases the regression method yields R = 7/15. In A, C, and D, we have BR −C < 0 (blue decreases in frequency); in B, BR −C > 0 (blue increases in frequency). to make important, testable predictions for the evolution of social behavior: A trait is selected if beneﬁt times relatedness exceeds cost. Clearly, such a simple and seemingly plausible statement has great intuitive appeal. In short, there is a startling discrepancy between the common intuitive understanding of Hamilton’s rule and the derivation of this rule that has been described as exact and general. In some cases, this discrepancy can be seen within a single paper. For example, ref. 7 uses 18 different variations of “Hamilton’s rule correctly predicts...” in reference to HRG, which makes no pre- diction at all. However, any intuition can only be as good as its mathemat- ical or biological underpinning. The purpose of this article has been to clarify the mathematical derivation of Hamilton’s rule that has been endorsed as exact, general, and canonical by the inclusive ﬁtness community (HRG) (4, 5, 7, 9, 10). The derivation of HRG is encapsulated in Eqs. 1–5. Any collection of triples can be used as input data and will turn out to be in “agreement with Hamilton’s rule” as long as the relevant denominators are nonzero. If the denominators are zero, the quantities B and C are undeﬁned. The data can come from any experiment, from any theory, from a deliberate or erroneous variation of either, or be completely imaginary. All such data, biological or not, will behave as “predicted by Hamilton’s rule.” Clearly, HRG is not a state- ment about biology and not a consequence of natural selection. Although HRG is the only formulation of Hamilton’s rule that is claimed to be exact and general, there are other approaches that deﬁne beneﬁt, cost, and relatedness in different ways. For example, beneﬁt and cost can be properties of individual phe- notypes, and relatedness can be deﬁned using common ancestry (1, 3, 12, 13, 24, 28). This approach, “Hamilton’s rule—special” (HRS) (12, 13), has the advantage of making testable predic- tions, because the beneﬁt and cost of a phenotype can be deter- mined in advance. However, it is easy to show that HRS holds only for special cases and not in general (12, 28–31). Discussion Hamilton’s rule is commonly thought to capture the idea that cooperative behaviors can be selected if the beneﬁts go to close relatives, because these relatives are likely to share genes for cooperation. In this understanding, Hamilton’s rule is believed PNAS \| May 30, 2017 \| vol. 114 \| no. 22 \| 5667 A B C D Fig. 3. The parameters B and C mischaracterize the underlying biology in simple examples. In all four populations of size N = 8, there are two types of individuals: blue, which indicates presence of a trait (g = 1), and red, which indicates absence of the trait (g = 0). Each individual is labeled by its ﬁt- ness, which is later normalized to ensure constant population size. Arrows indicate interaction partners. (A) Both blue and red have baseline ﬁtness 3. Blue harms blue by reducing its ﬁtness by 2 at no per- sonal cost. Blue helps red by increasing its ﬁtness by 2 at no cost. Red does nothing to its interac- tion partners. The regression method yields B, C > 0, misclassifying blue as altruism. (B) Blue has base- line ﬁtness 1, and red has baseline ﬁtness 2. Blue increases the ﬁtness of blue by 3 and decreases the ﬁtness of red by 1, both at no cost. Red does noth- ing to its interaction partners. The regression method yields B > 0 and C < 0, misclassifying blue as mutu- ally beneﬁcial. This classiﬁcation is incorrect because blue harms red. (C) Blue has baseline ﬁtness 2, and red has baseline ﬁtness 3. Blue increases the ﬁt- ness of blue by 1 and decreases the ﬁtness of red by 2, both at no cost. Red decreases the ﬁtness of blue by 1 at no cost and does nothing to red. The regression method yields B < 0 and C > 0, misclassi- fying blue as spiteful. (D) Blue has baseline ﬁtness 1 and red has baseline ﬁtness 2. Blue does nothing to blue, and red increases the ﬁtness of blue by 2 at no cost. Red does nothing to red. The regression method yields B, C < 0, misclassifying blue as selﬁsh. In all four cases the regression method yields R = 7/15. In A, C, and D, we have BR −C < 0 (blue decreases in frequency); in B, BR −C > 0 (blue increases in frequency). A B C D Fig. Discussion 3. The parameters B and C mischaracterize the underlying biology in simple examples. In all four populations of size N = 8, there are two types of individuals: blue, which indicates presence of a trait (g = 1), and red, which indicates absence of the trait (g = 0). Each individual is labeled by its ﬁt- ness, which is later normalized to ensure constant population size. Arrows indicate interaction partners. (A) Both blue and red have baseline ﬁtness 3. Blue harms blue by reducing its ﬁtness by 2 at no per- sonal cost. Blue helps red by increasing its ﬁtness by 2 at no cost. Red does nothing to its interac- tion partners. The regression method yields B, C > 0, misclassifying blue as altruism. (B) Blue has base- line ﬁtness 1, and red has baseline ﬁtness 2. Blue increases the ﬁtness of blue by 3 and decreases the ﬁtness of red by 1, both at no cost. Red does noth- ing to its interaction partners. The regression method yields B > 0 and C < 0, misclassifying blue as mutu- ally beneﬁcial. This classiﬁcation is incorrect because blue harms red. (C) Blue has baseline ﬁtness 2, and red has baseline ﬁtness 3. Blue increases the ﬁt- ness of blue by 1 and decreases the ﬁtness of red by 2, both at no cost. Red decreases the ﬁtness of blue by 1 at no cost and does nothing to red. The regression method yields B < 0 and C > 0, misclassi- fying blue as spiteful. (D) Blue has baseline ﬁtness 1 and red has baseline ﬁtness 2. Blue does nothing to blue, and red increases the ﬁtness of blue by 2 at no cost. Red does nothing to red. The regression method yields B, C < 0, misclassifying blue as selﬁsh. In all four cases the regression method yields R = 7/15. In A, C, and D, we have BR −C < 0 (blue decreases in frequency); in B, BR −C > 0 (blue increases in frequency). B A D fying blue as spiteful. (D) Blue has baseline ﬁtness 1 and red has baseline ﬁtness 2. Blue does nothing to blue, and red increases the ﬁtness of blue by 2 at no cost. Red does nothing to red. The regression method yields B, C < 0, misclassifying blue as selﬁsh. 5668 \| www.pnas.org/cgi/doi/10.1073/pnas.1701805114 Discussion They ﬁnd, as we have shown here, that HRG does not predict any aspect of their system, but yields only a value of BR−C that coincides with the result they have already obtained. The biological question at hand is how population structure affects the evolution of social behavior, which is a deep and important question that has been studied extensively (41–49). 5668 \| www.pnas.org/cgi/doi/10.1073/pnas.1701805114 Nowak et al. Nowak et al. A B C D E F Fig. 4. The numerical values of the parameters B (“beneﬁt to interaction partners”) and C (“cost to self”) are not robust. An inﬁnitesimally small change in population structure or trait value can modify B and C from arbitrarily large positive values to arbitrarily large negative values. Blue corresponds to g = 1, indicating the presence of a focal trait, and red corresponds to g = 0, indicating its absence. In A, the trait values (blue/red) are held constant and the weight of a single interaction is perturbed by ε. B shows the resulting values of w, g, and h. C illustrates the effects of this perturbation on B and C as ε →0 (when x = 0.5, w1 = 1.6, and w2 = w3 = w4 = 0.8). In D, all players have a trait value of x (where 0 < x < 1), and a single player’s trait value is slightly perturbed by ε. E gives the table representation of this population, and in F, we see the erratic behavior of B (beneﬁt) and C (cost) as ε →0 (depending on w). In both populations, the limits of B and C are different from the left (ε small and negative) and from the right (ε small and positive). For ε = 0, the parameters B and C are undeﬁned in both populations. A B B B C B A D E F E F D E Fig. 4. The numerical values of the parameters B (“beneﬁt to interaction partners”) and C (“cost to self”) are not robust. An inﬁnitesimally small change in population structure or trait value can modify B and C from arbitrarily large positive values to arbitrarily large negative values. Blue corresponds to g = 1, indicating the presence of a focal trait, and red corresponds to g = 0, indicating its absence. Proposition. These slopes satisfy the following equations: The intuition that a cooperative gene can spread by preferen- tially conferring beneﬁts on cobearers of this gene is correct. However, Hamilton’s rule, in its exact and general formulation, is unrelated to this biological intuition and (in general) neither predicts nor explains the evolution of social behavior. mwg = Mg + Mhmhg; [7a] mwh = Mh + Mgmgh. [7b] [7a] [7b] mwh = Mh + Mgmgh. [7b] Proof. Suppose that {(xi1, . . . , xiℓ) , yi}n i=1 is a collection of n data points, which can be expressed in matrix form as Proof. Suppose that {(xi1, . . . , xiℓ) , yi}n i=1 is a collection of n data points, which can be expressed in matrix form as Indeed, we should not expect that interplay of population structure and social behavior can be reduced to a simple rule with three parameters. Social interactions, which are typically multi- lateral (50) and nonlinear (51, 52), cannot be expressed by a sin- gle beneﬁt and cost. Complex population structures (43, 46, 53, 54) cannot be captured by a single relatedness quantity. Assort- ment among relatives often has a positive effect on cooperation (41, 44–47), but in other cases it has a negative effect (48, 55) or no effect at all (42, 45). A good understanding of these questions, like all great problems in science, will require careful empirical observation in concert with meaningful mathematics. X :=     1 x11 · · · x1ℓ 1 x21 · · · x2ℓ ... ... ... ... 1 xn1 · · · xnℓ    ; y :=    y1 ... yn   . [8] [8] If we are to ﬁt a linear model of the form y = x Tβ for some coef- ﬁcient vector, β, then it is a well-known result in statistics (20) that the least-squares solution satisﬁes the equation X TX β = X Ty. [9] Discussion In A, the trait values (blue/red) are held constant and the weight of a single interaction is perturbed by ε. B shows the resulting values of w, g, and h. C illustrates the effects of this perturbation on B and C as ε →0 (when x = 0.5, w1 = 1.6, and w2 = w3 = w4 = 0.8). In D, all players have a trait value of x (where 0 < x < 1), and a single player’s trait value is slightly perturbed by ε. E gives the table representation of this population, and in F, we see the erratic behavior of B (beneﬁt) and C (cost) as ε →0 (depending on w). In both populations, the limits of B and C are different from the left (ε small and negative) and from the right (ε small and positive). For ε = 0, the parameters B and C are undeﬁned in both populations. Proposition. These slopes satisfy the following equations: Proposition. These slopes satisfy the following equations: Nowak et al. ACKNOWLEDGMENTS. The authors thank Donald Rubin, Karl Sigmund, Corina Tarnita, and John Wakeley for helpful discussions. Nature 434:69–72. 9. Marshall JAR (2015) Social Evolution and Inclusive Fitness Theory: An Introduction (Princeton Univ Press, Princeton, NJ). 39. Richards MH, French D, Paxton RJ (2005) It’s good to be queen: Classically eusocial colony structure and low worker ﬁtness in an obligately social sweat bee. Mol Ecol 14:4123–4133. 10. Rousset F (2015) Regression, least squares, and the general version of inclusive ﬁtness. Evolution 69:2963–2970. 40. Bourke AFG (2014) Hamilton’s rule and the causes of social evolution. Philos Trans R Soc B 369:20130362. 11. Nonacs P, Richards MH (2015) How (not) to review papers on inclusive ﬁtness. Trends Ecol Evol 20:1–2. 12. Birch J (2014) Hamilton’s rule and its discontents. Br J Philos Sci 65:3 41. Nowak MA, May RM (1992) Evolutionary games and spatial chaos. Nature 359: 826–829. 13. Birch J, Okasha S (2015) Kin selection and its critics. BioScience 65:22–32. 42. Taylor PD (1992) Altruism in viscous populations—an inclusive ﬁtness model. Evol Ecol 6:352–356. 14. Abbot P, et al. (2011) Inclusive ﬁtness theory and eusociality. Nature 471:E1–E4. 15. Yule GU (1897) On the theory of correlation. J R Stat Soc 60:812–85 43. Santos FC, Pacheco JM (2005) Scale-free networks provide a unifying framework for the emergence of cooperation. Phys Rev Lett 95:98104. 16. Cox DR, Wermuth N (1996) Multivariate Dependencies: Models, Analysis and Interpretation (Taylor & Francis, Boca Raton, FL). 17. West SA, Grifﬁn AS, Gardner A (2007) Social semantics: Altruism, cooperation, mutu- alism, strong reciprocity and group selection. J Evol Biol 20:415–432. 44. Kerr B, Neuhauser C, Bohannan BJM, Dean AM (2006) Local migration promotes competitive restraint in a host–pathogen‘tragedy of the commons’. Nature 442:75– 78. 18. Seber GAF, Lee AJ (2003) Linear Regression Analysis (Wiley-Blackwell, Hoboken, NJ). 45. Ohtsuki H, Hauert C, Lieberman E, Nowak MA (2006) A simple rule for the evolution of cooperation on graphs and social networks. Nature 441:502–505. 19. Cox DR, Wermuth N (2004) Causality: A statistical view. Int Stat Rev 72:285–305. 20. Hastie T, Friedman J, Tibshirani R (2001) The Elements of Statistical Learning (Springer, New York). 46. Nowak MA, Tarnita CE, Antal T (2010) Evolutionary dynamics in structured popula- tions. Philos Trans R Soc Lond B Biol Sci 365:19–30. 21. Berk R (2004) Regression Analysis: A Constructive Critique (SAGE Publications, Thou- sand Oaks, CA). 47. D´ebarre F, Hauert C, Doebeli M (2014) Social evolution in structured populations. Nat Commun 5:3409. 22. ACKNOWLEDGMENTS. The authors thank Donald Rubin, Karl Sigmund, Corina Tarnita, and John Wakeley for helpful discussions. [13b] ACKNOWLEDGMENTS. The authors thank Donald Rubin, Karl Sigmund, Corina Tarnita, and John Wakeley for helpful discussions. 1. Hamilton WD (1964) The genetical evolution of social behaviour. I. J Theor Biol 7: 1–16. 1. Hamilton WD (1964) The genetical evolution of social behaviour. I. J Theor Biol 7: 1–16. 1. Hamilton WD (1964) The genetical evolution of social behaviour. I. J Theor Biol 7: 1–16. 32. van Veelen M (2009) Group selection, kin selection, altruism and cooperation: When inclusive ﬁtness is right and when it can be wrong. J Theor Biol 259:589–600. 2. Hamilton WD (1970) Selﬁsh and spiteful behaviour in an evolutionary model. Nature 228:1218–1220. 2. Hamilton WD (1970) Selﬁsh and spiteful behaviour in an evolutionary model. Nature 228:1218–1220. 33. Fletcher JA, Doebeli M (2009) A simple and general explanation for the evolution of altruism. Proc R Soc Lond B Biol Sci 276:13–19. 33. Fletcher JA, Doebeli M (2009) A simple and general explanation for the evolution of altruism. Proc R Soc Lond B Biol Sci 276:13–19. 34. Marshall JAR (2014) Generalizations of Hamilton’s rule applied to non-additive public goods games with random group size. Front Ecol Evol 2:40. 34. Marshall JAR (2014) Generalizations of Hamilton’s rule applied to non-additive public goods games with random group size. Front Ecol Evol 2:40. 3. Michod RE, Hamilton WD (1980) Coefﬁcients of relatedness in sociobiology. Nature 288:694–697. 4. Queller DC (1992) A general model for kin selection. Evolution 46:376–380. 35. Stark RE (1992) Cooperative nesting in the multivoltine large carpenter bee xylocopa sulcatipes maa (apoidea: Anthophoridae): Do helpers gain or lose to solitary females? Ethology 91:301–310. . Frank SA (1998) Foundations of Social Evolution (Princeton Univ Press 6. Grafen A (1985) A geometric view of relatedness. Oxford Surveys in Evolutionary Biology, eds Dawkins R, Ridley M (Oxford Univ Press, Oxford, UK), Vol 2, pp 28–89. 36. Nonacs P, Reeve HK (1995) The ecology of cooperation in wasps: Causes and conse- quences of alternative reproductive decisions. Ecology 76:953–967. 7. Gardner A, West SA, Wild G (2011) The genetical theory of kin selection. J Evol Biol 24:1020–1043. quences of alternative reproductive decisions. Ecology 76:953–967. 37. Loeb MLG (2003) Evolution of egg dumping in a subsocial insect. Am Nat 161: 129–142. 8. Lehmann L, Rousset F (2014) The genetical theory of social behaviour. Philos Trans R Soc Lond B Biol Sci 369:20130357. 38. Krakauer AH (2005) Kin selection and cooperative courtship in wild turkeys. Appendix 7 is if mhgmgh = 1, which happens if and only if there exist constants k1 and k2 (not both zero) for which k1 (gi −g) = k2 hi −h for each i. In this case, the values of Mg and Mh, and therefore also the values of B and C in HRG, are undeﬁned. Moreover, we also see from Eq. 9 that Moreover, we also see from Eq. 9 that n 0 0 0 g · g g · h 0 g · h h · h ! 0 Mg Mh ! = 0 w · g w · h ! , n 0 0 0 g · g g · h 0 g · h h · h ! 0 Mg Mh ! = 0 w · g w · h ! , [11] e u de ed. Introducing R′ = mgh, we can write c R∆g Introducing R′ = mgh, we can write Introducing R′ = mgh, we can write [11] B = chw −R∆g vh (1 −RR′); [14a] [14a] □ which, together with Eq. 10, gives Eq. 7. which, together with Eq. 10, gives Eq. 7. , g q , g q Note that Eq. 7 can be written in matrix form as , g q , g q Note that Eq. 7 can be written in matrix form as C = −∆g −R′chw vg (1 −RR′), [14b] mwg mwh = 1 mhg mgh 1 Mg Mh . [12] C = −∆g −R′chw vg (1 −RR′), hi h i i l i E 4 [14b] mwg mwh = 1 mhg mgh 1 Mg Mh . [12] whic [12] mwg mwh = 1 mhg mgh 1 Mg Mh . [12] C = − g hw vg (1 −RR′), which gives a symmetric alternative to Eq. 4. which gives a symmetric alternative to Eq. 4. mwh mgh h ̸ 1 i hi i h which gives a symmetric alternative to Eq. 4. Appendix Provided mhgmgh ̸= 1, we can invert this matrix to see that mhgmgh ̸= 1, we can invert this matrix to see that Mg = mwg −mwhmhg 1 −mhgmgh ; [13a] g y q It is interesting to note that the relationships between linear regressions with one and two explanatory variables, which are captured in Eq. 13 and give rise to HRG, appeared in the statis- tics literature as far back as 1897 (15). g y q It is interesting to note that the relationships between linear regressions with one and two explanatory variables, which are captured in Eq. 13 and give rise to HRG, appeared in the statis- tics literature as far back as 1897 (15). Mg = mwg −mwhmhg 1 −mhgmgh ; [13a] M mwh −mwgmgh [13b] It is interesting to note that the relationships between linear regressions with one and two explanatory variables, which are captured in Eq. 13 and give rise to HRG, appeared in the statis- tics literature as far back as 1897 (15). ACKNOWLEDGMENTS The authors thank Donald Rubin Karl Sigmund Mg = mwg −mwhmhg 1 −mhgmgh ; [13a] Mh = mwh −mwgmgh 1 −mhgmgh . [13b] Mg = mwg −mwhmhg 1 −mhgmgh ; [13a] Mh = mwh −mwgmgh 1 −mhgmgh . [13b] wh −mwgmgh 1 −mhgmgh . [13b] ACKNOWLEDGMENTS. The authors thank Donald Rubin, Karl Sigmund, Corina Tarnita, and John Wakeley for helpful discussions. Appendix [9] Here, we recapitulate the derivation of HRG as a simple conse- quence of a well-known result in statistics. Because all of the terms appearing in Eq. 7 are slopes, we may assume without a loss of generality that g = h = w = 0. From Eq. 9, we obtain Because all of the terms appearing in Eq. 7 are slopes, we may assume without a loss of generality that g = h = w = 0. From Eq. 9, we obtain For any collection of n data points in the form of triples, {(wi, gi, hi)}n i=1, suppose that (i) mwg is the slope the least- squares regression line w vs. g, (ii) mwh is the slope of w vs. h, (iii) mhg is the slope of h vs. g, and (iv) mgh is the slope of g vs. h. Furthermore, for the least-squares plane expressing w vs. both g and h, let (v) Mg be the slope of the line obtained by holding h constant, and let (vi) Mh be the slope of the line obtained by holding g constant (see Fig. 1 for details). mwg = w · g g · g ; [10a] mwh = w · h h · h ; [10b] [10a] [10a] [10b] PNAS \| May 30, 2017 \| vol. 114 \| no. 22 \| 5669 PNAS \| May 30, 2017 \| vol. 114 \| no. 22 \| 5669 Nowak et al. mhg = g · h g · g ; [10c] mgh = g · h h · h . [10d] e from Eq. 9 that 0 ! 0 ! 0 ! Because B = Mh and C = −Mg in HRG, we have Eq. 3. The only situation in which we cannot solve explicitly for Mg and Mh in Eq. 7 is if mhgmgh = 1, which happens if and only if there exist constants k1 and k2 (not both zero) for which k1 (gi −g) = k2 hi −h for each i. In this case, the values of Mg and Mh, and therefore also the values of B and C in HRG, are undeﬁned Because B = Mh and C = −Mg in HRG, we have Eq. 3. and Mh in Eq. ACKNOWLEDGMENTS. The authors thank Donald Rubin, Karl Sigmund, Corina Tarnita, and John Wakeley for helpful discussions. Allen B, Nowak MA, Wilson EO (2013) Limitations of inclusive ﬁtness. Proc Natl Acad Sci USA 110:20135–20139. 48. Allen B, Nowak MA (2015) Games among relatives revisited. J Theor Biol 378:103–116. 23. Chuang JS, Rivoire O, Leibler S (2010) Cooperation and Hamilton’s rule in a simple synthetic microbial system. Mol Syst Biol 6:398. 49. Van Cleve J (2015) Social evolution and genetic interactions in the short and long term. Theor Popul Biol 103:2–26. 50. Tarnita CE (2017) The ecology and evolution of social behavior in microbes. J Exp Biol 220:18–24. 24. van Veelen M, Allen B, Hoffman M, Simon B, Veller C (2016) Hamilton’s rule. J Theor Biol 414:176–230. 51. Gore J, Youk H, van Oudenaarden A (2009) Snowdrift game dynamics and facultative cheating in yeast. Nature 459:253–256. 25. Price GR (1970) Selection and covariance. Nature 227:520–521. 26. van Veelen M (2005) On the use of the price equation. J Theor Biol 237:412–426. 26. van Veelen M (2005) On the use of the price equation. J Theor B 52. Archetti M, Ferraro DA, Christofori G (2015) Heterogeneity for IGF-II production main- tained by public goods dynamics in neuroendocrine pancreatic cancer. Proc Natl Acad Sci USA 112:1833–1838. 27. Frank SA (2012) Natural selection. IV. The price equation. J Evol Biol 25:1002–1019. 28. Nowak MA, Tarnita CE, Wilson EO (2010) The evolution of eusociality. Nature 466:1057–1062. 53. Maciejewski W, Fu F, Hauert C (2014) Evolutionary game dynamics in populations with heterogenous structures. PLoS Comput Biol 10:e1003567. 29. Charlesworth B (1978) Some models of the evolution of altruistic behaviour between siblings. J Theor Biol 72:297–319. 30. Cavalli-Sforza LL, Feldman MW (1978) Darwinian selection and “altruism”. Theor Popul Biol 14:268–280. 54. Allen B, et al. (2017) Evolutionary dynamics on any population structure. Nature 544:227–230. 31. Karlin S, Matessi C (1983) The eleventh RA Fisher memorial lecture: Kin selection and altruism. Proc R Soc Lond B Biol Sci 219:327–353. 55. Hauert C, Doebeli M (2004) Spatial structure often inhibits the evolution of coopera- tion in the snowdrift game. Nature 428:643–646. 5670 \| www.pnas.org/cgi/doi/10.1073/pnas.1701805114 Nowak et al.
https://openalex.org/W4284878888	https://tuprints.ulb.tu-darmstadt.de/21867/1/materials-15-04737-v2.pdf	English	null	Robust Determination of Fatigue Crack Propagation Thresholds from Crack Growth Data	Materials	2,022	cc-by	13,975	Citation: Schönherr, J.A.; Duarte, L.; Madia, M.; Zerbst, U.; Geilen, M.B.; Klein, M.; Oechsner, M. Robust Determination of Fatigue Crack Propagation Thresholds from Crack Growth Data. Materials 2022, 15, 4737. https://doi.org/10.3390/ ma15144737 Academic Editors: Lucjan ´Snie˙zek, Jaroslaw Galkiewicz and Sebastian Lipiec Received: 1 June 2022 Accepted: 1 July 2022 Published: 6 July 2022 Keywords: fatigue crack propagation threshold; ISO 12108; ASTM E647; data evaluation methods; experimental determination Citation: Schönherr, J.A.; Duarte, L.; Madia, M.; Zerbst, U.; Geilen, M.B.; Klein, M.; Oechsner, M. Robust Determination of Fatigue Crack Propagation Thresholds from Crack Growth Data. Materials 2022, 15, 4737. https://doi.org/10.3390/ ma15144737 Article Robust Determination of Fatigue Crack Propagation Thresholds from Crack Growth Data Josef Arthur Schönherr 1,* , Larissa Duarte 2, Mauro Madia 2, Uwe Zerbst 2, Max Benedikt Geilen 1 , Marcus Klein 1 and Matthias Oechsner 1 herr 1,* , Larissa Duarte 2, Mauro Madia 2, Uwe Zerbst 2, Max Benedikt Geilen 1 , nd Matthias Oechsner 1 hur Schönherr 1,* , Larissa Duarte 2, Mauro Madia 2, Uwe Zerbst 2, Max Benedikt Geilen 1 , lein 1 and Matthias Oechsner 1 1 Center for Structural Materials (MPA-IfW), Technical University of Darmstadt, 64283 Darmstadt, Germany; max.geilen@tu-darmstadt.de (M.B.G.); marcus.klein@tu-darmstadt.de (M.K.); matthias.oechsner@tu-darmstadt.de (M.O.) 1 Center for Structural Materials (MPA-IfW), Technical University of Darmstadt, 64283 Darmstadt, Germany; max.geilen@tu-darmstadt.de (M.B.G.); marcus.klein@tu-darmstadt.de (M.K.); matthias.oechsner@tu-darmstadt.de (M.O.) ( ) 2 Bundesanstalt für Materialforschung und -Prüfung (BAM), Division 9.4, 12205 Berlin, Germany; larissa.duarte@bam.de (L.D.); mauro.madia@bam.de (M.M.); uwe.zerbst@bam.de (U.Z.) * Correspondence: josef.schoenherr@tu-darmstadt.de; Tel.: +49-6151-16-20348 * Correspondence: josef.schoenherr@tu-darmstadt.de; Tel.: +49-6151-16-20348 Abstract: The robust determination of the threshold against fatigue crack propagation ∆Kth is of paramount importance in fracture mechanics based fatigue assessment procedures. The standards ASTM E647 and ISO 12108 introduce operational deﬁnitions of ∆Kth based on the crack propagation rate da/dN and suggest linear ﬁts of logarithmic ∆K– da/dN test data to calculate ∆Kth. Since these ﬁts typically suffer from a poor representation of the actual curvature of the crack propagation curve, a method for evaluating ∆Kth using a nonlinear function is proposed. It is shown that the proposed method reduces the artiﬁcial conservativeness induced by the evaluation method as well as the susceptibility to scatter in test data and the inﬂuence of test data density. materials materials 1.1. Procedure Suggested by Both ASTM and ISO Standards According to ASTM and ISO standards, the threshold stress intensity factor ranges are evaluated by determining ‘the best ﬁt straight line’ [4,6] to log ∆K–log da/dN data, log10 ∆K = P1 · log10 da/dN + P0, (1) (1) log10 ∆K = P1 · log10 da/dN + P0, where P0, P1 are ﬁtting parameters, and then calculating the stress intensity ranges corre- sponding to da/dNth,ASTM and da/dNth,ISO, respectively. Both standards deﬁne a mini- mum number of ﬁve data points, approximately equally spaced in da/dN. The ﬁtting interval includes data pairs between 10−7 mm/cycle and 10−6 mm/cycle for ASTM and between 10−8 mm/cycle and 10−7 mm/cycle for ISO, even though both standards allow for using additional data with lower fatigue crack propagation rates, but require documenting the modiﬁed range within the test protocol. Since nowadays FCG tests typically yield far more than ﬁve data points within one decade of da/dN data, there is plenty of room for interpretation of the suggested methods. Probably, the most straightforward interpretation (named “interpretation one” in the following) is to just take all data points within the speciﬁed ranges (as long as they are approximately equally distributed in da/dN direc- tion) and then identify the best ﬁt straight line for example by utilizing the least-squares method; see Figure 1a. Thereby, Equation (1) is ﬁtted to the (logarithmic) test data, using ∆K as the dependent variable (i.e., the direction of the estimated error). The optimal set of parameters obtained by the least-squares parameter optimization returns the “best ﬁt”. Another interpretation (named “interpretation two” in the following) might be that one may freely select n ≥5 approximately equally spaced points that lie within the deﬁned boundaries, for example starting with the point next to the desired threshold fatigue crack propagation rate. Then, the ﬁt showing the maximum Pearson correlation coefﬁcient is selected; see Figure 1b. The results differ and, by adding data generated at lower decades of da/dN, even a non-conservative (= higher) FCG threshold stress intensity range might be calculated, see Section 3.1.3. Furthermore, neither of both interpretations (nor any other straight line) is able to reﬂect the curvature of the depicted test data. 3 4 ∆K in MPa·m1/2 10−8 10−7 da/dN in mm/cycle Experimental dataset Fitting dataset Fitting curve (Eq. (1), all data points) ∆Kth,ASTM = 2.72MPa·m1/2 3 4 ∆K in MPa·m1/2 10−8 10−7 da/dN in mm/cycle Experimental dataset Fitting dataset Fitting curve (Eq. 1. Procedures for the Determination of the Fatigue Crack Propagation Threshold from Crack Propagation Data Typically, the outcome of fatigue crack growth (FCG) tests for the determination of the fatigue crack propagation threshold ∆Kth are the crack length a and load history data (e.g., minimum and maximum force) on dependence of the number of cycles N. Consider- ing linear elastic fracture mechanics, the stress intensity factor range ∆K can be calculated from the load and crack length history [1–3]. By using, e.g., the numerical differentiation technique like the secant method or the incremental polynomial method [4], fatigue crack propagation rate da/dN can be computed from the crack length readings and the cycles count. Due to measurement inaccuracy, inﬂuence of the testing environment, material inho- mogeneities and other effects, test data are always affected by scatter. Furthermore, in most cases, there is no distinct reading at da/dN = da/dNth,ASTM and da/dN = da/dNth,ISO, respectively. Hence, the direct determination of the corresponding stress intensity ranges ∆Kth,ASTM and ∆Kth,ISO is not possible. Therefore, data ﬁtting including inter- or extrapola- tion techniques to determine ∆Kth,ASTM and ∆Kth,ISO are needed, which will be discussed in the following. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional afﬁl- iations. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional afﬁl- iations. Especially in the presence of extrinsic effects, i.e., e.g., crack closure effects [5], the ac- curate determination of the fatigue crack propagation threshold ∆Kth is not trivial. The goal of this contribution is to investigate different methods for the evaluation of the fatigue crack propagation threshold and discuss their robustness and conservativeness. Copyright: © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). Following the two most well-known standards regarding FCG tests, namely ASTM E647 [4] and ISO 12108 [6], ∆Kth is deﬁned as the asymptotic value of stress intensity factor range ∆K at which the fatigue crack propagation rate da/dN approaches zero. The technical (or also operational) deﬁnition of ∆Kth for most materials is given at ﬁnite crack growth rates https://www.mdpi.com/journal/materials Materials 2022, 15, 4737. https://doi.org/10.3390/ma15144737 Materials 2022, 15, 4737 2 of 21 da/dNth,ASTM = 10−7 mm/cycle according to ASTM and da/dNth,ISO = 10−8 mm/cycle according to ISO. 1. Procedures for the Determination of the Fatigue Crack Propagation Threshold from Crack Propagation Data Since neither the fatigue crack growth rate nor the fatigue crack growth threshold stress intensity factor range can be measured directly, the data evaluation is of paramount importance. . Procedure Suggested by Both ASTM and ISO Standards 1.2. Procedures Suggested in the Literature Bucci [7] considered a four-parameter Weibull function (see also [8,9]), ﬁtted to the entire data set, see Figure 2a. He stated that the Weibull approach addresses nonlinearity in a better way, but if there are enough data points in the near-threshold region, straight line ﬁts are performing quite well and are easier to use. In Figure 2a, the test data have been ﬁtted using the four-parameter Weibull function 1 −∆K Kb = exp − da/dN −e v −e k! (2) (2) and ∆Kth,ASTM has been determined at da/dNth,ASTM. Smith and Hoeppner [9] observed that the least-squares method is not suitable to ﬁt the instability parameter Kb, the thresh- old parameter e, the characteristic value v and the shape parameter k to da/dN–∆K data accurately within the threshold region due to the differences in orders of magnitude of the non-logarithmic test data (da/dN ranges between 10−7 mm/cycle and 10−6 mm/cycle vs. ∆K ranges between 100 MPa·m1/2 and 101 MPa·m1/2, approximately). Therefore, they proposed to use a least-squares optimization to calculate a preliminary optimized pa- rameter set and afterwards improve the ﬁtting results by optimizing only e and v in an orthogonal distance regression (ODR) [10]. Since an ODR optimization is connected to a quite high computational cost and compared to the 1990s the computation power in- creased dramatically until nowadays, this might have been the reason why in [9] only the two parameters showing the major contribution with regard to the mentioned errors were optimized using an ODR. In fact, it has been observed in this work that the curve ﬁt may be improved by optimizing all four parameters instead of just two in an ODR after least-squares minimization. As one can clearly see, the overall agreement to the test data is quite good but locally diverges slightly. This can be observed in Figure 2a as well as in Figure 2b, where the four-parameter Weibull ﬁt predicts a too steep curvature towards the ISO operational threshold deﬁnition (da/dNth = 10−8 mm/cycle) and therefore leads to non-conservative results for ∆Kth,ISO. Since non-conservative values are not acceptable, the Weibull ﬁt has not been considered any further. y Another well established method to evaluate the threshold stress intensity range was proposed by Döker [11]. The method uses a straight line ﬁt applied to the (non- logarithmic) da/dN–∆K data in the range 5 × 10−8mm/cycle ≤da/dN ≤10−6 mm/cycle, see Figure 3a. 1.1. Procedure Suggested by Both ASTM and ISO Standards (1), ﬁrst n data points) ∆Kth,ASTM = 2.77MPa·m1/2 (a) (b) Figure 1. FCG test conducted with conventional K-decreasing at R = 0.8 in lab air, see corresponding paragraph in Section 3.1.1: (a) linear ﬁt over all data points with 10−7 mm/cycle ≤da/dN ≤ 10−6mm/cycle; (b) best linear ﬁt over the ﬁrst n data points with da/dN ≥10−7 mm/cycle, where n = 112 (out of 177 within the interval) is selected yielding the highest Pearson correlation coefﬁcient. The results obtained following interpretation are strongly affected by the test data curvature because of its averaging character. The higher the curvature, the more conserva- tive is the calculated ∆Kth. In contrast, interpretation two is less affected by data curvature 3 4 ∆K in MPa·m1/2 10−8 10−7 da/dN in mm/cycle Experimental dataset Fitting dataset Fitting curve (Eq. (1), all data points) ∆Kth,ASTM = 2.72MPa·m1/2 (a) 3 4 ∆K in MPa·m1/2 10−8 10−7 da/dN in mm/cycle Experimental dataset Fitting dataset Fitting curve (Eq. (1), ﬁrst n data points) ∆Kth,ASTM = 2.77MPa·m1/2 (b) (a) (b) Figure 1. FCG test conducted with conventional K-decreasing at R = 0.8 in lab air, see corresponding paragraph in Section 3.1.1: (a) linear ﬁt over all data points with 10−7 mm/cycle ≤da/dN ≤ 10−6mm/cycle; (b) best linear ﬁt over the ﬁrst n data points with da/dN ≥10−7 mm/cycle, where n = 112 (out of 177 within the interval) is selected yielding the highest Pearson correlation coefﬁcient. The results obtained following interpretation are strongly affected by the test data curvature because of its averaging character. The higher the curvature, the more conserva- tive is the calculated ∆Kth. In contrast, interpretation two is less affected by data curvature The results obtained following interpretation are strongly affected by the test data curvature because of its averaging character. The higher the curvature, the more conserva- tive is the calculated ∆Kth. In contrast, interpretation two is less affected by data curvature Materials 2022, 15, 4737 3 of 21 because of its deﬁnition using the correlation coefﬁcient of the straight line. In case of a more pronounced scatter in the vicinity of da/dNth, interpretation two would have accounted for a higher number of data points and therefore be more conservative. 1.1. Procedure Suggested by Both ASTM and ISO Standards It is trivial to state that a straight line is unable to characterize the curvature of the crack growth curve and therefore leads to conservative ∆Kth results for strictly monotonic increasing FCG data ﬁtted within the recommended ranges. In turn, the evaluation error depends on the degree of the curvature, and it is therefore sensitive to the number (and location) of data pairs taken into account as well as to the curve’s gradient, see also [7]. Therefore—as also mentioned in the ASTM standard [4]—a nonlinear relationship between ∆K and da/dN might be beneﬁcial to obtain a good ﬁt. 1.2. Procedures Suggested in the Literature The threshold deﬁnition is independent from the ASTM or ISO standards, and ∆Kth is evaluated at da/dNth = 0. The resulting threshold stress intensity factor range is located somewhere between the values one would expect using the ISO or ASTM operational deﬁnitions. Considering that only the ASTM E647 standard [12] is cited within the original publication, this method shows quite conservative results using the ASTM operational deﬁnition of ∆Kth. The straightforward extension to the ISO threshold by shifting the ﬁt range by one decade of da/dN would require test results as low as da/dN = 5 × 10−9 mm/cycle, which would be very time consuming to obtain and there- fore not practicable. The curve behavior is poorly described, both for linear, Figure 3a, Materials 2022, 15, 4737 4 of 21 and double-logarithmic, Figure 3b, scaled da/dN–∆K data. Therefore, further analyses with this approach have also been discarded. and double-logarithmic, Figure 3b, scaled da/dN–∆K data. Therefore, further analyses with this approach have also been discarded. Furthermore, there is a multitude of different crack propagation laws aiming at describ- ing the whole FCG data starting from the threshold regime over the range where the FCG curve grows linear in a double-logarithmic scaled plot (also known as Paris regime) and some even include the region of instable crack growth; see [13–19]. Several of these models contain ∆Kth as a model parameter, which should not be confused with the operational deﬁnitions of the threshold stress intensity factor range included in the standards. 3 4 ∆K in MPa·m1/2 10−8 10−7 da/dN in mm/cycle Experimental dataset Fitting dataset Fitting curve (Eq. (2), Weibull) ∆Kth,ASTM = 2.79MPa·m1/2 3 4 ∆K in MPa·m1/2 10−8 10−7 10−6 da/dN in mm/cycle Experimental dataset Fitting dataset Fitting curve (Eq. (2), Weibull) ∆Kth,ISO = 2.28MPa·m1/2 (a) (b) Figure 2. Application of a four-parameter Weibull function on all test data, referring to: (a) a threshold deﬁnition at da/dNth,ASTM = 10−7 mm/cycle. K-decreasing test at R = 0.8 in lab air, see correspond- ing paragraph in Section 3.1.1; (b) a slightly non-conservative FCG threshold determination at da/dNth,ISO = 10−8 mm/cycle. Kmax = const., Rmax ≈0.8, lab air, see corresponding paragraph in Section 3.1.1. 3 4 ∆K in MPa·m1/2 10−8 10−7 10−6 da/dN in mm/cycle Experimental dataset Fitting dataset Fitting curve (Eq. (2), Weibull) ∆Kth,ISO = 2.28MPa·m1/2 (b) 3 4 ∆K in MPa·m1/2 10−8 10−7 da/dN in mm/cycle Experimental dataset Fitting dataset Fitting curve (Eq. 1.2. Procedures Suggested in the Literature (2), Weibull) ∆Kth,ASTM = 2.79MPa·m1/2 (a) (b) (a) Figure 2. Application of a four-parameter Weibull function on all test data, referring to: (a) a threshold deﬁnition at da/dNth,ASTM = 10−7 mm/cycle. K-decreasing test at R = 0.8 in lab air, see correspond- ing paragraph in Section 3.1.1; (b) a slightly non-conservative FCG threshold determination at da/dNth,ISO = 10−8 mm/cycle. Kmax = const., Rmax ≈0.8, lab air, see corresponding paragraph in Section 3.1.1. 3 4 5 ∆K in MPa·m1/2 0 2×10−7 3×10−7 4×10−7 5×10−7 6×10−7 7×10−7 8×10−7 da/dN in mm/cycle Experimental dataset (linear) Fitting dataset Fitting curve (linear, D¨oker) ∆Kth = 2.64MPa·m1/2 3 4 ∆K in MPa·m1/2 10−8 10−7 da/dN in mm/cycle Experimental dataset (log–log) Fitting dataset Fitting curve (linear, D¨oker) ∆Kth = 2.64MPa·m1/2 (a) (b) Figure 3. FCG threshold determination using a linear ﬁt, applied to regular (non-logarithmic) test data, following [11]. K-decreasing test at R = 0.8 in lab air, see corresponding paragraph in Section 3.1.1. (a) linear-scaled data; (b) double-logarithmic scaled data. The method clearly provides a poor data ﬁt and determination of ∆Kth at da/dNth = 0. 3 4 ∆K in MPa·m1/2 10−8 10−7 da/dN in mm/cycle Experimental dataset (log–log) Fitting dataset Fitting curve (linear, D¨oker) ∆Kth = 2.64MPa·m1/2 (b) 3 4 5 ∆K in MPa·m1/2 0 2×10−7 3×10−7 4×10−7 5×10−7 6×10−7 7×10−7 8×10−7 da/dN in mm/cycle Experimental dataset (linear) Fitting dataset Fitting curve (linear, D¨oker) ∆Kth = 2.64MPa·m1/2 (a) (b) (a) Figure 3. FCG threshold determination using a linear ﬁt, applied to regular (non-logarithmic) test data, following [11]. K-decreasing test at R = 0.8 in lab air, see corresponding paragraph in Section 3.1.1. (a) linear-scaled data; (b) double-logarithmic scaled data. The method clearly provides a poor data ﬁt and determination of ∆Kth at da/dNth = 0. 2.1. Investigated Fitting Functions The most straightforward approach was to extend Equation (1) to a more general polynomial form as The most straightforward approach was to extend Equation (1) to a more general polynomial form as log10 ∆K = I ∑ i=0 Pi · log10 da/dN i, (3) (3) where I denotes the polynomial degree and Pi are ﬁt parameters. For I = 1, Equation (3) equals Equation (1). Regarding I ≥2, in case there are only very few (or even no) data points available in the vicinity of da/dNth, the ﬁt may yield a curve having an inﬂection point. Hence, the ﬁt does not represent the data in a satisfactory manner, and (for I = 2) it is even possible that the ﬁt does not intersect the da/dNth line. Therefore, it is not recommended to use polynomial ﬁts of the type given in Equation (3) with I ̸= 1. The use of non-logarithmic data, like in the approach of Döker [11], leads to the same problems. Hence, further approaches have been investigated. The log −log data depicted in Figure 1 considered swapping the axes, suggesting a hyperbolic trend, leading to the ﬁt function log10 ∆K = P1 · −log10 da/dN −1 + P2, (4) (4) with the ﬁtting parameters P1 and P2. Fitting Equation (4) to the dataset leads to results comparable to the linear ﬁt using all data points in accordance with the standards, since the exponent −1 does not represent the curvature of the crack growth curve, see Figure 4a. By extending Equation (4) to a variable exponent as log10 ∆K = P1 · −log10 da/dN −P3 + P2, (5) (5) with P3 ≥1, a much better ﬁtting to the test data is possible, see Figure 4b. Rounding the ﬁt result of P3 ≈4.60 to the previous and next integer number, namely P3 = 4, log10 ∆K = P1 · −log10 da/dN −4 + P2, (6) (6) see Figure 4c and P3 = 5, log10 ∆K = P1 · −log10 da/dN −5 + P2, (7) (7) see Figure 4d provides two straightforward variants of Equation (5) with only two free parameters. The curvature of the ﬁtting curve increases with increasing P3. It is worth noting that, although P3 = 5 leads to non-conservative estimation of the FCG threshold according to ISO, it is still conservative following the ASTM operational deﬁnition. 2. Experimental Procedure Since none of the methods shown in Sections 1.1 and 1.2 allowed universal application for both standards and simultaneously led to robust and not overly conservative results, further ﬁtting functions have been investigated. A comprehensive set of fatigue crack growth data recorded at BAM Berlin and MPA-IfW Darmstadt has been used to calibrate and validate the evaluation procedure. The ﬁtting has been performed using least squares minimization, where the threshold stress intensity factor range has been used as dependent variable, if not otherwise mentioned. Materials 2022, 15, 4737 5 of 21 2.1. Investigated Fitting Functions 2.1. Investigated Fitting Functions 3 4 ∆K in MPa·m1/2 10−8 10−7 da/dN in mm/cycle Experimental dataset Fitting dataset Fitting curve (Eq. (5), P3 = 1) ∆Kth,ASTM = 2.75MPa·m1/2 3 4 ∆K in MPa·m1/2 10−8 10−7 da/dN in mm/cycle Experimental dataset Fitting dataset Fitting curve (Eq. (5), P3 = free) ∆Kth,ASTM = 2.80MPa·m1/2 (a) (b) Figure 4. Cont. 3 4 ∆K in MPa·m1/2 10−8 10−7 da/dN in mm/cycle Experimental dataset Fitting dataset Fitting curve (Eq. (5), P3 = free) ∆Kth,ASTM = 2.80MPa·m1/2 (b) 3 4 ∆K in MPa·m1/2 10−8 10−7 da/dN in mm/cycle Experimental dataset Fitting dataset Fitting curve (Eq. (5), P3 = 1) ∆Kth,ASTM = 2.75MPa·m1/2 (b) (a) Figure 4. Cont. Figure 4. Cont. Materials 2022, 15, 4737 6 of 21 3 4 ∆K in MPa·m1/2 10−8 10−7 da/dN in mm/cycle Experimental dataset Fitting dataset Fitting curve (Eq. (6), P3 = 4) ∆Kth,ASTM = 2.79MPa·m1/2 3 4 ∆K in MPa·m1/2 10−8 10−7 da/dN in mm/cycle Experimental dataset Fitting dataset Fitting curve (Eq. (7), P3 = 5) ∆Kth,ASTM = 2.80MPa·m1/2 (c) (d) Figure 4. FCG threshold evaluation according to ASTM (da/dNth,ASTM = 10−7 mm/cycle) for the data set presented in Figure 1 using Equation (5) with ﬁxed or free parameter P3: (a) ﬁxed, P3 = 1; (b) free, optimized value P3 ≈4.60; (c) ﬁxed, P3 = 4; (d) ﬁxed, P3 = 5. 3 4 ∆K in MPa·m1/2 10−8 10−7 da/dN in mm/cycle Experimental dataset Fitting dataset Fitting curve (Eq. (7), P3 = 5) ∆Kth,ASTM = 2.80MPa·m1/2 (d) 3 4 ∆K in MPa·m1/2 10−8 10−7 da/dN in mm/cycle Experimental dataset Fitting dataset Fitting curve (Eq. (6), P3 = 4) ∆Kth,ASTM = 2.79MPa·m1/2 (c) (d) (c) Figure 4. FCG threshold evaluation according to ASTM (da/dNth,ASTM = 10−7 mm/cycle) for the data set presented in Figure 1 using Equation (5) with ﬁxed or free parameter P3: (a) ﬁxed, P3 = 1; (b) free, optimized value P3 ≈4.60; (c) ﬁxed, P3 = 4; (d) ﬁxed, P3 = 5. 2.2. Quantitative Data Analysis The main datasets investigated stemmed from a total of 47 specimens manufactured from 12 mm thick S690QL hot-rolled plates, tested at BAM Berlin and MPA-IfW Darmstadt. The materials chemical composition and mechanical properties are given in Tables 1 and 2. The microscopic analysis on etched samples showed a ﬁne grained quenched and tempered martensitic-bainitic microstructure, see Figure 5. Table 1. Chemical composition (in weight percent) obtained by means of spark optical emission spectrometry. C Si Mn P S Cr Mo Ni Al Cu Nb Fe 0.16 0.23 1.15 0.01 < 0.01 0.41 0.18 0.04 0.08 0.02 0.04 97.65 Table 2. Mechanical properties. σy in MPa σu in MPa E in GPa A in % KV2 in J (Orientation: T–L [20]) 810 825 207 16 126 Figure 5. S690QL microstructure (T–L plane). All tests were performed using SENB specimens [6] with a cross section of 19 mm × 6 mm. The specimens were oriented, such that the direction of crack propagation is parallel to the rolling direction, i.e., the orientation T–L according to [20]. The test data have been obtained on three resonance testing machines equipped with an eight-point bending ﬁxture, see [6]. These were a RUMUL MIKROTRON 654 with a maximum load capacity of 20 kN and an average testing frequency of about 108 Hz, a RUMUL TESTRONIC with a able 1. Chemical composition (in weight percent) obtained by means of spark optical emission spectrometry. Table 2. Mechanical properties. Figure 5. S690QL microstructure (T–L plane). Figure 5. S690QL microstructure (T–L plane). Figure 5. S690QL microstructure (T–L plane). All tests were performed using SENB specimens [6] with a cross section of 19 mm × 6 mm. The specimens were oriented, such that the direction of crack propagation is parallel to the rolling direction, i.e., the orientation T–L according to [20]. The test data have been obtained on three resonance testing machines equipped with an eight-point bending ﬁxture, see [6]. These were a RUMUL MIKROTRON 654 with a maximum load capacity of 20 kN and an average testing frequency of about 108 Hz, a RUMUL TESTRONIC with a Materials 2022, 15, 4737 7 of 21 7 of 21 maximum load capacity of 100 kN and an average testing frequency of about 60 Hz, both at BAM Berlin, and a RUMUL TESTRONIC with a maximum load capacity of 250 kN and an average testing frequency of about 90 Hz at MPA-IfW Darmstadt. 3.1.1. Evaluation for the Intervals Suggested by the Standards In order to assess the performance of the polynomial functions with negative exponent Equation (5) in comparison to the ﬁt suggested by the standards, see Equation (1), test data obtained at either a ﬁxed load ratio R = 0.8 (K-decreasing procedures) or at Rmax ≈0.8 (Kmax tests) have been considered, see ([4], Section 8.6). To minimize inﬂuences of the extrapolation method, the smallest recorded crack propagation rate has been required to be smaller than 1.1 · da/dNth, i.e., min(da/dN) ≤1.1 × 10−8 mm/cycle for the ISO and min(da/dN) ≤1.1 × 10−7 mm/cycle for the ASTM operational deﬁnition of ∆Kth. It shall be noted that these boundaries have been used only for comparability between tests within this work and neither deﬁne the actual application boundaries of the method regarding data extrapolation nor represent a general recommendation. Since the ﬁxed parameters P3 = 4 in Equation (6) and P3 = 5 in Equation (7) provided a better description of the data curvature and a less conservative determination of ∆Kth in addition to the more general function with variable P3, these have been compared with the two interpretations of the linear ﬁt method suggested by the standards, see Figure 1a,b. The range used to ﬁt the data was in all cases ﬁxed to one decade of da/dN data, starting from da/dNth and therefore equal to the ranges suggested by the standards in order to ensure comparability between the methods. 2.2. Quantitative Data Analysis The crack length was monitored using direct current potential drop techniques with current reversal and active temperature compensation (current source: HP 6033A, nanovolt meter: Keithley 2182A) at BAM and specimen compliance techniques (clip-gage: Sandner EXR10-0.5o) at MPA-IfW. The crack length was corrected a posteriori by means of optical measurements on the broken open fracture surfaces. Then, the crack propagation rates have been calculated using the slope of piecewise straight line ﬁts performed on ﬁltered test data. Each segment of the piecewise function referred to a crack extension of 0.02 mm. The corresponding stress intensity factors have been calculated using the formulations reported in [6]. Since in these tests the same specimen types, manufactured from the same material batch in the same specimen orientation, have been used, a high repeatability was expected. Consequently, the standard deviation in ∆Kth calculated for each method was the result of the data scatter within the test (stemming from small variations in environment conditions, material inhomogeneities, specimen misalignment, errors in the calculation of da/dN, etc.) and an error induced by the ﬁt used to evaluate ∆Kth. Since the ﬁrst part is independent of the ﬁtting procedure, the differences in standard deviations are a measure for the ﬁt robustness, whereas the corresponding mean value gives information on the ﬁt quality and therefore the inter- and extrapolation error, respectively. 3. Results and Discussion 3.1. Application to Data Obtained at R ≈0.8 3.1. Application to Data Obtained at R ≈0.8 First, data obtained at a load ratio of approximately R ≈0.8, which produce only a negligible inﬂuence of crack closure effects have been investigated. 3.1.1. Evaluation for the Intervals Suggested by the Standards Conventional K-Decreasing at R = 0.8 (∆KLR) The ﬁrst datasets investigated stemmed from a total of nine SENB specimens made of S690QL. These tests have been conducted using the standard K-decreasing (or load shedding) procedure suggested by the standards at constant R = 0.8, tested in lab air. Further information on the experimental procedure may be found in [21]. All nine speci- mens contained data points below da/dN = 1.1 × 10−7 mm/cycle and are valid for ASTM operational threshold evaluation, and four of them were also valid according to the ISO deﬁnition. The threshold stress intensity factor range results with corresponding standard deviations are presented in Table 3. Materials 2022, 15, 4737 8 of 21 Table 3. Comparison of effective threshold values for S690QL obtained with various ﬁt methods. Tests conducted using the K-decreasing procedure at R = 0.8 in lab air. Method ∆Kth,ASTM in MPa·m1/2 ∆Kth,ISO in MPa·m1/2 Linear, all data points (Equation (1)) 2.71 ± 0.06 2.27 ± 0.04 Linear, ﬁrst n data points (Equation (1)) 2.76 ± 0.05 2.29 ± 0.04 Polynomial, neg. exp. (Equation (5)) 2.78 ± 0.04 2.34 ± 0.04 Polynomial, neg. exp. P3 = 4 (Equation (6)) 2.77 ± 0.05 2.32 ± 0.04 Polynomial, neg. exp. P3 = 5 (Equation (7)) 2.78 ± 0.05 2.32 ± 0.03 Table 3. Comparison of effective threshold values for S690QL obtained with various ﬁt methods Tests conducted using the K-decreasing procedure at R = 0.8 in lab air. Table 3. Comparison of effective threshold values for S690QL obtained with various ﬁt methods. Tests conducted using the K-decreasing procedure at R = 0.8 in lab air. The comparison between the determined threshold stress intensity factor ranges (on single specimen basis) and the value read out from test data proved conservative for every single method and dataset. Considering the mean value of the threshold stress intensity ranges, both linear ﬁts showed a more pronounced underestimation of ∆Kth, inducing an artiﬁcial conservativeness, as already observed in Section 2.1. This is proven true especially for the linear ﬁt Equation (1) performed on all data points. For the linear ﬁt over the ﬁrst n points, the artiﬁcial conservativeness of ∆Kth,ASTM was found to be comparatively higher than for ∆Kth,ISO. The results obtained using Equation (5) agreed fairly well, whereas the ﬁxed exponents P3 = 4 and P3 = 5 showed a slightly higher conservativeness for ∆Kth,ISO compared to the three-parameter version of Equation (5). Conventional K-Decreasing at R = 0.8 (∆KLR) The standard deviation is very low and comparable for all tests. Load Shedding at Constant Kmax (Rmax ≈0.8) Load Shedding at Constant Kmax (Rmax ≈0.8) A set of nine SENB prepared from the same material batch has been investigated using a load shedding scheme at constant Kmax with a final load ratio Rmax ≈0.8 at about ∆Kth,ISO and R ≈0.72 . . . 0.76 at ∆Kth,ASTM. All nine specimens provided data for evaluating ∆Kth,ASTM and among them four were also valid for ∆Kth,ISO evaluation, see Table 4. Table 4. Comparison of effective threshold values for S690QL obtained with various ﬁt methods. Tests conducted using the constant Kmax procedure (Rmax ≈0.8) in lab air. Method ∆Kth,ASTM in MPa·m1/2 ∆Kth,ISO in MPa·m1/2 Linear, all data points (Equation (1)) 2.67 ± 0.03 2.22 ± 0.01 Linear, ﬁrst n data points (Equation (1)) 2.77 ± 0.03 2.24 ± 0.01 Polynomial, neg. exp. (Equation (5)) 2.79 ± 0.03 2.29 ± 0.03 Polynomial, neg. exp. P3 = 4 (Equation (6)) 2.77 ± 0.03 2.27 ± 0.02 Polynomial, neg. exp. P3 = 5 (Equation (7)) 2.79 ± 0.03 2.27 ± 0.02 Table 4. Comparison of effective threshold values for S690QL obtained with various ﬁt methods. Tests conducted using the constant Kmax procedure (Rmax ≈0.8) in lab air. Constant Force Range (∆F-Constant) at R = 0.8 The fourth and last test to determine the threshold at R ≈0.8 in ambient air has been based on another set of four SENB specimens produced from the same material batch, using conventional precracking followed by a test at constant force amplitude (∆F-constant) at a load ratio R = 0.8. All four specimens have been considered valid for evaluating ∆Kth,ASTM and three for ∆Kth,ISO, see Table 6. Here, the same observations as above apply. Table 6. Comparison of threshold values for S690QL obtained with various ﬁt methods. Tests conducted using ∆F-constant at R = 0.8 in lab air. Method ∆Kth,ASTM in MPa·m1/2 ∆Kth,ISO in MPa·m1/2 Linear, all data points (Equation (1)) 2.58 ± 0.03 2.18 ± 0.06 Linear, ﬁrst n data points (Equation (1)) 2.64 ± 0.03 2.21 ± 0.06 Polynomial, neg. exp. (Equation (5)) 2.69 ± 0.02 2.31 ± 0.02 Polynomial, neg. exp. P3 = 4 (Equation (6)) 2.67 ± 0.02 2.25 ± 0.03 Polynomial, neg. exp. P3 = 5 (Equation (7)) 2.69 ± 0.02 2.25 ± 0.02 Table 6. Comparison of threshold values for S690QL obtained with various ﬁt methods. Tests conducted using ∆F-constant at R = 0.8 in lab air. Method Since the crack propagation data obtained from different procedures at R ≈0.8 nearly coincide and exhibit a very low scatter, the same conclusion as in the case of conventional K-decreasing tests can be drawn. Compression Precracking Load Reduction (CPLR) at R = 0.8 Compression Precracking Load Reduction (CPLR) at R = 0.8 A further set of four SENB extracted from the same batch has been investigated using compression precracking followed by a K-decreasing test with a constant load ratio R = 0.8. Applying the same criteria for selecting valid data sets for comparison returned four specimens for evaluating ∆Kth,ASTM and two for ∆Kth,ISO, see Table 5. Here, the standard deviation for ISO is omitted due to the insufﬁcient number of available data sets that include points below da/dN = 1.1 × 10−8 mm/cycle. The results conﬁrm those previously shown in Table 4. Materials 2022, 15, 4737 9 of 21 Table 5. Comparison of threshold values for S690QL obtained with various ﬁt methods. Tests conducted using Compression Precracking Load Reduction at R = 0.8 in lab air. Table 5. Comparison of threshold values for S690QL obtained with various ﬁt methods. Tests conducted using Compression Precracking Load Reduction at R = 0.8 in lab air. Table 5. Comparison of threshold values for S690QL obtained with various ﬁt methods. Tests conducted using Compression Precracking Load Reduction at R = 0.8 in lab air. Method ∆Kth,ASTM in MPa·m1/2 ∆Kth,ISO in MPa·m1/2 Linear, all data points (Equation (1)) 2.67 ± 0.02 2.22 Linear, ﬁrst n data points (Equation (1)) 2.74 ± 0.03 2.23 Polynomial, neg. exp. (Equation (5)) 2.77 ± 0.04 2.29 Polynomial, neg. exp. P3 = 4 (Equation (6)) 2.73 ± 0.02 2.26 Polynomial, neg. exp. P3 = 5 (Equation (7)) 2.74 ± 0.03 2.27 Summary Robustness of the Fitting Methods in Handling Data Subjected to Augmented Artiﬁcial Scatter In order to assess the ability and robustness of the ﬁtting methods to handle scattered data, artiﬁcial scatter has been added to the test data presented in Figure 1. The additional scatter has been generated by sampling random values from a normal distribution with a mean µ = 1 and a standard deviation SD = 0.02 and multiplying them with ∆K data, whereas da/dN remained unchanged. The comparison of the linear fit using all scattered data points within the defined interval (∆Kth,ASTM = 2.69 MPa·m1/2, Figure7a)withtheoriginaldataset(∆Kth,ASTM = 2.72 MPa·m1/2, Figure 1a) did not reveal a notable difference. The polynomials with negative expo- nents were also almost insensitive to scatter. The three-parameter polynomial provided ∆Kth,ASTM = 2.78 MPa·m1/2 for the scattered data (Figure 7d) compared to ∆Kth,ASTM = 2.80 MPa·m1/2 for the original dataset (Figure 4b). The two-parameter polynomial with P3 = 4 showed a similar trend with ∆Kth,ASTM = 2.79 MPa·m1/2 for scattered data (Figure 7c) in comparison to ∆Kth,ASTM = 2.79 MPa·m1/2 for the original dataset (Figure 4c). 3 4 ∆K in MPa·m1/2 10−8 10−7 da/dN in mm/cycle Artiﬁcially scattered experimental dataset Fitting dataset Fitting curve (Eq. (1), all data points) ∆Kth,ASTM = 2.69MPa·m1/2 3 4 ∆K in MPa·m1/2 10−8 10−7 da/dN in mm/cycle Artiﬁcially scattered experimental dataset Fitting dataset Fitting curve (Eq. (1), ﬁrst n data points) ∆Kth,ASTM = 2.72MPa·m1/2 (a) (b) 3 4 ∆K in MPa·m1/2 10−8 10−7 da/dN in mm/cycle Artiﬁcially scattered experimental dataset Fitting dataset Fitting curve (Eq. (6)) ∆Kth,ASTM = 2.79MPa·m1/2 3 4 ∆K in MPa·m1/2 10−8 10−7 da/dN in mm/cycle Artiﬁcially scattered experimental dataset Fitting dataset Fitting curve (Eq. (5)) ∆Kth,ASTM = 2.78MPa·m1/2 (c) (d) Figure 7. FCG data reported in Figure 1 with additional artiﬁcial scatter. The da/dNth,ASTM has been evaluated using four different methods: (a) linear, all data points Equation (1); (b) linear, ﬁrst n = 176 data points (out of 177 within the interval) Equation (1); (c) polynomial, negative exponent, P3 = 4 Equation (6); (d) polynomial, negative exponent Equation (5). 3 4 ∆K in MPa·m1/2 10−8 10−7 da/dN in mm/cycle Artiﬁcially scattered experimental dataset Fitting dataset Fitting curve (Eq. (1), all data points) ∆Kth,ASTM = 2.69MPa·m1/2 (a) 3 4 ∆K in MPa·m1/2 10−8 10−7 da/dN in mm/cycle Artiﬁcially scattered experimental dataset Fitting dataset Fitting curve (Eq. Summary The linear ﬁts induce artiﬁcial conservativeness in the evaluation of fatigue crack propagation thresholds obtained at R ≈0.8. Nevertheless, this behavior is observed less pronounced for the linear ﬁt incorporating only the ﬁrst n points. In all cases, the three- parameter polynomial Equation (5) provides less conservative results. Fixing its parameter P3 to P3 = 4 or P3 = 5 sometimes induces conservativeness, but in most of the cases less pronounced than the linear ﬁts. p Figure 6 summarizes the evaluation of the four datasets presented earlier in Section 3.1.1. The threshold stress intensity factor ranges determined according to the ASTM operational deﬁnition (see Figure 6a) as well as those following the ISO operational deﬁnition (see Figure 6b) for the four test methods (∆KLR, Kmax, CPLR and ∆F-constant) agree fairly well within each data evaluation method. ∆KLR Kmax CPLR ∆F−const. 2.0 2.2 2.4 2.6 2.8 3.0 ∆Kth,ASTM in MPa·m1/2 ∆KLR Kmax CPLR ∆F−const. 2.0 2.2 2.4 2.6 2.8 3.0 ∆Kth,ISO in MPa·m1/2 Linear, all data points (Eq. (1)) Linear, ﬁrst n data points (Eq. (1)) Polynomial, neg. exponent (Eq. (5)) Polynomial, neg. exponent, P3 = 4 (Eq. (6)) Polynomial, neg. exponent, P3 = 5 (Eq. (7)) (a) (b) Figure 6. Fatigue crack propagation thresholds obtained at R ≈0.8 applying the ﬁt methods to the four datasets presented in Section 3.1.1: (a) according to the ASTM operational deﬁnition; (b) according to the ISO operational deﬁnition. ∆KLR Kmax CPLR ∆F−const. 2.0 2.2 2.4 2.6 2.8 3.0 ∆Kth,ISO in MPa·m1/2 Linear, all data points (Eq. (1)) Linear, ﬁrst n data points (Eq. (1)) Polynomial, neg. exponent (Eq. (5)) Polynomial, neg. exponent, P3 = 4 (Eq. (6)) Polynomial, neg. exponent, P3 = 5 (Eq. (7)) (b) ∆KLR Kmax CPLR ∆F−const. 2.0 2.2 2.4 2.6 2.8 3.0 ∆Kth,ISO in MPa·m1/2 Linear, all data points (Eq. (1)) Linear, ﬁrst n data points (Eq. (1)) Polynomial, neg. exponent (Eq. (5)) Polynomial, neg. exponent, P3 = 4 (Eq. (6)) Polynomial, neg. exponent, P3 = 5 (Eq. (7)) (b) ∆KLR Kmax CPLR ∆F−const. 2.0 2.2 2.4 2.6 2.8 3.0 ∆Kth,ASTM in MPa·m1/2 (a) (a) (b) Figure 6. Fatigue crack propagation thresholds obtained at R ≈0.8 applying the ﬁt methods to the four datasets presented in Section 3.1.1: (a) according to the ASTM operational deﬁnition; (b) according to the ISO operational deﬁnition. Materials 2022, 15, 4737 10 of 21 10 of 21 3.1.2. 3.1.3. Inﬂuence of an Augmented Fit Interval 3.1.3. Inﬂuence of an Augmented Fit Interval Both ASTM E647 and ISO 12108 suggest a ﬁt interval of one decade of da/dN data, starting from da/dNth,ASTM and da/dNth,ISO, respectively. Nevertheless, both allow for use data obtained at lower fatigue crack propagation rates for determining the threshold stress intensity factor range. Therefore, the impact of an augmented ﬁt interval on the deter- mination of fatigue crack propagation thresholds has been investigated. Since no datasets with crack propagation rates momentously below da/dNth,ISO = 1 × 10−8 mm/cycle were available, only the threshold following the ASTM operational deﬁnition has been consid- ered. Therefore, the data shown in Table 3, which have been obtained using a ﬁt interval of 10−7 mm/cycle ≤da/dN ≤10−6 mm/cycle, have been compared with threshold stress intensity factor ranges obtained with augmented intervals. The upper bound has been held constant, whereas the lower bound has been varied from 2.5 × 10−8 mm/cycle up to 10−7 mm/cycle. The respective threshold stress intensity factor ranges have been named as ∆Kth,ASTM,2.5, ∆Kth,ASTM,5, ∆Kth,ASTM,7.5 and ∆Kth,ASTM,10, according to the lower FCG propagation rate bounds (2.5, 5, 7.5 and 10 × 10−8 mm/cycle, see Figure 8). ∆Kth,ASTM,10∆Kth,ASTM,7.5 ∆Kth,ASTM,5 ∆Kth,ASTM,2.5 2.5 2.6 2.7 2.8 2.9 3.0 3.1 ∆Kth,ASTM in MPa·m1/2 Linear, all data points (Eq. (1)) Linear, ﬁrst n data points (Eq. (1)) Polynomial, neg. exponent (Eq. (5)) Polynomial, neg. exponent, P3 = 4 (Eq. (6)) Polynomial, neg. exponent, P3 = 5 (Eq. (7)) Figure 8. Comparison of threshold values obtained with various ﬁt methods using vary- ing ﬁt intervals from 2.5 × 10−8 mm/cycle ≤da/dN ≤10−6 mm/cycle (∆Kth,ASTM,2.5) up to 10 × 10−8 mm/cycle ≤da/dN ≤10−6 mm/cycle (∆Kth,ASTM,10). The results refer to eight speci- mens made of S690QL, tested using a K-decreasing procedure at R = 0.8 in lab air. Figure 8. Comparison of threshold values obtained with various ﬁt methods using vary- ing ﬁt intervals from 2.5 × 10−8 mm/cycle ≤da/dN ≤10−6 mm/cycle (∆Kth,ASTM,2.5) up to 10 × 10−8 mm/cycle ≤da/dN ≤10−6 mm/cycle (∆Kth,ASTM,10). The results refer to eight speci- mens made of S690QL, tested using a K-decreasing procedure at R = 0.8 in lab air. Regarding both linear ﬁts, there is a clear tendency that, with augmenting the ﬁt interval towards lower minimal crack propagation rates (displayed in Figure 8 from left to right), there is an increase in ∆Kth,ASTM and therefore a decrease in conservativeness. Summary (1), ﬁrst n data points) ∆Kth,ASTM = 2.72MPa·m1/2 (b) (a) 3 4 ∆K in MPa·m1/2 10−8 10−7 da/dN in mm/cycle Artiﬁcially scattered experimental dataset Fitting dataset Fitting curve (Eq. (6)) ∆Kth,ASTM = 2.79MPa·m1/2 (c) (b) 3 4 ∆K in MPa·m1/2 10−8 10−7 da/dN in mm/cycle Artiﬁcially scattered experimental dataset Fitting dataset Fitting curve (Eq. (5)) ∆Kth,ASTM = 2.78MPa·m1/2 (d) (a) (b) (d) (c) Figure 7. FCG data reported in Figure 1 with additional artiﬁcial scatter. The da/dNth,ASTM has been evaluated using four different methods: (a) linear, all data points Equation (1); (b) linear, ﬁrst n = 176 data points (out of 177 within the interval) Equation (1); (c) polynomial, negative exponent, P3 = 4 Equation (6); (d) polynomial, negative exponent Equation (5). In contrast, the best linear ﬁt over the ﬁrst n points showed a pronounced sensi- tivity to scattered data. The best ﬁt interval coincided almost with all data points (see Figures 7b and 1b). Consequently, ∆Kth,ASTM = 2.72 MPa·m1/2 calculated for scattered data was more conservative than ∆Kth,ASTM = 2.77 MPa·m1/2 calculated in case of the original dataset. Materials 2022, 15, 4737 11 of 21 11 of 21 3.1.4. Data Extrapolation The investigations presented in Section 3.1.1 included only data at crack propagation rates as low as 1.1 · da/dNth. Nevertheless, it shall be noted that no data might be available at low crack propagation rates, especially for the ISO operational deﬁnition with a threshold crack propagation rate as low as da/dNth,ISO = 10−8 mm/cycle. Therefore, to ensure a reliable and conservative evaluation of the fatigue crack propagation thresholds, a robust extrapolation technique is needed. To assess the goodness of the extrapolation, the test results given in Section 3.1.1 have been compared to artiﬁcially censored datasets, using only data with da/dN ≥3 · da/dNth. Hence, the resulting ﬁt intervals after censoring were 3 × 10−8 mm/cycle ≤da/dN ≤10−7 mm/cycle for ISO operational deﬁnition and 3 × 10−7mm/cycle ≤da/dN ≤10−6 mm/cycle for ASTM operational deﬁnition. Since extrapolation is very sensitive to the data range available, in order to have a reliable comparison, the investigations have been restricted to datasets that had data within the whole censored interval (including the upper bound). Hence, the number of tests with valid data for ASTM threshold determination is reduced in comparison to Section 3.1.1. The comparison has been based on the change in ∆Kth induced by censoring the FCG data. ∆Kth,cens denotes the fatigue crack propagation threshold obtained for the censored version of the data set used to evaluate ∆Kth. It follows that ∆Kth −∆Kth,cens values greater than or equal to zero are considered conservative, whereas values lower than zero are non- conservative. The minimum difference throughout all specimens shows whether all tests are extrapolated conservatively, whereas the mean value can be regarded as an index of the goodness of the extrapolation. The results are given in Table 7. Conservative extrapolation has been obtained for both linear ﬁts and for the polynomial with negative exponent ﬁxed to P3 = 4, whilst the versions with P3 = 5 or free P3 returned a very limited number of negative results, meaning non-conservative extrapolation results. Even though the latter are just slightly non-conservative and rare, the occurrence of a non-conservative extrapolation should be avoided whenever possible. 3.1.3. Inﬂuence of an Augmented Fit Interval In contrast, the results for the polynomial with negative exponent, Equations (5)–(7), are almost insensitive to the interval augmentation, but for P3 = 4 and P3 = 5, a reduction in standard deviation can be observed with increasing the interval. For Equation (6), the optimal lower bound was found at 2.5 × 10−8 mm/cycle and 5 × 10−8 mm/cycle with equal magnitudes in mean and standard deviation. Using the linear functions with augmented intervals increases the risk of non-conser- vative extrapolation, as one can see comparing the values for ∆Kth,ASTM,2.5, where the linear functions provided the highest threshold stress intensity factor ranges among all five methods under comparison. This issue can be clearly understood looking at the evaluation depicted in Figure 9. In particular, Figure 9a shows that the ∆Kth,ASTM calculated using the linear fit over all data points is on the right-hand side of the dataset, i.e., in the non- conservative region. In contrast, the polynomial with P3 = 4 does not show this issue (Figure 9b). Materials 2022, 15, 4737 12 of 21 3 4 ∆K in MPa·m1/2 10−8 10−7 da/dN in mm/cycle Experimental dataset Fitting dataset Fitting curve (Eq. (1), augmented interval) ∆Kth,ASTM = 2.84MPa·m1/2 3 4 ∆K in MPa·m1/2 10−8 10−7 da/dN in mm/cycle Experimental dataset Fitting dataset Fitting curve (Eq. (6), augmented interval) ∆Kth,ASTM = 2.80MPa·m1/2 (a) (b) Figure 9. Inﬂuence of augmented ﬁt interval on various ﬁts, using an interval 2.5 × 10−8 mm/cycle ≤ da/dN ≤10−6 mm/cycle: (a) the linear ﬁt using all data within the speciﬁed range leads to a non- conservative result; (b) the polynomial ﬁt with negative exponent P3 = 4 gives a conservative result. The same dataset as reported in Figure 1 has been used. 3 4 ∆K in MPa·m1/2 10−8 10−7 da/dN in mm/cycle Experimental dataset Fitting dataset Fitting curve (Eq. (6), augmented interval) ∆Kth,ASTM = 2.80MPa·m1/2 (b) 3 4 ∆K in MPa·m1/2 10−8 10−7 da/dN in mm/cycle Experimental dataset Fitting dataset Fitting curve (Eq. (1), augmented interval) ∆Kth,ASTM = 2.84MPa·m1/2 (a) (b) (a) Figure 9. Inﬂuence of augmented ﬁt interval on various ﬁts, using an interval 2.5 × 10−8 mm/cycle ≤ da/dN ≤10−6 mm/cycle: (a) the linear ﬁt using all data within the speciﬁed range leads to a non- conservative result; (b) the polynomial ﬁt with negative exponent P3 = 4 gives a conservative result. The same dataset as reported in Figure 1 has been used. 3.1.5. Application to the Full Dataset Comparison of threshold values determined from artiﬁcial scattered test data, obtained at R ≈0.8 with various ﬁt methods using all 29 specimen data sets for ∆Kth,ASTM and ∆Kth,ISO; results for S690QL in lab air. Table 9. Comparison of threshold values determined from artiﬁcial scattered test data, obtained at R ≈0.8 with various ﬁt methods using all 29 specimen data sets for ∆Kth,ASTM and ∆Kth,ISO; results for S690QL in lab air. Method ∆Kth,ASTM in MPa·m1/2 ∆Kth,ISO in MPa·m1/2 Linear, all data points (Equation (1)) 2.65 ± 0.06 2.16 ± 0.08 Linear, ﬁrst n data points (Equation (1)) 2.68 ± 0.07 2.23 ± 0.05 Polynomial, neg. exp. P3 = 4 (Equation (6)) 2.76 ± 0.07 2.27 ± 0.04 3.1.5. Application to the Full Dataset Based on the conclusions drawn in Section 3.1.4, where a robust conservative extrap- olation could be obtained for the linear ﬁts as well as for the polynomial with negative exponent ﬁxed to P3 = 4, the full dataset, including all datasets that may be extrapolated, has been reevaluated. Because an augmented ﬁt interval may lead to non-conservative results in case of linear ﬁts, the intervals suggested in the standards have been used (see Section 1.1). In contrast, a beneﬁcial effect of an augmented interval has been observed when using Equation (6), see Section 3.1.3. Hence, an augmented interval for calculat- ing ∆Kth,ASTM has been used in this case. Consequently, the ﬁtting intervals have been deﬁned as da/dN = 5 × 10−8 mm/cycle ≤da/dN ≤10−6 mm/cycle for ASTM and da/dN = 10−8 mm/cycle ≤da/dN ≤10−7 mm/cycle for ISO. The results are presented in Table 8. The linear ﬁt over the full interval resulted in the highest conservativeness in combination with a small standard deviation. Using only the ﬁrst n points of the interval to generate the linear ﬁt reduced the conservativeness with the drawback of increasing the standard deviation, especially for ASTM. By using an appropriate nonlinear function like Equation (6), the conservativeness as well as the standard deviation can be reduced. Table 8. Comparison of threshold values obtained at R ≈0.8 with various ﬁt methods using all 29 specimen data sets for ∆Kth,ASTM and ∆Kth,ISO; results for S690QL in lab air. Table 8. Comparison of threshold values obtained at R ≈0.8 with various ﬁt methods using all 29 specimen data sets for ∆Kth,ASTM and ∆Kth,ISO; results for S690QL in lab air. Method ∆Kth,ASTM in MPa·m1/2 ∆Kth,ISO in MPa·m1/2 Linear, all data points (Equation (1)) 2.67 ± 0.06 2.22 ± 0.04 Linear, ﬁrst n data points (Equation (1)) 2.73 ± 0.08 2.24 ± 0.05 Polynomial, neg. exp. P3 = 4 (Equation (6)) 2.76 ± 0.05 2.28 ± 0.04 In order to assess the methods capabilities when dealing with an augmented scatter in ∆K, artiﬁcial scatter, as described in Section 3.1.2, is added to each specimen data set. The results are given in Table 9. Like already observed in Figure 7, the two-parameter polynomial Equation (6) is almost insensitive to scatter in test data, whereas both linear ﬁts partially suffer from pronounced susceptibility to scattered data. Table 9. 3.1.4. Data Extrapolation Regarding the “mean( · )” columns, denoting the ex- trapolation error, the linear functions performed far worse than the polynomial Equation (5) with P3 = 4, especially with regard to the ∆Kth,ASTM −∆Kth,cens,ASTM values, where the mean and minimum extrapolation error were 0.20 MPa·m1/2 and 0.17 MPa·m1/2, regarding the ﬁt over all data points and 0.23 MPa·m1/2 and 0.18 MPa·m1/2 for the ﬁt over the ﬁrst n points, respectively, compared to 0.03 MPa·m1/2 and 0.01 MPa·m1/2 for the polynomial with P3 = 4. Hence, in case extrapolation is inevitable, the polynomial with P3 = 4 brings a notable improvement over the linear ﬁts, suggested by the standards. With regard to both standards, neither a statement on the minimum required crack propagation rate for a valid evaluation of the FCG threshold nor any comments on the legitimacy of a potentially necessary data extrapolation is given. Materials 2022, 15, 4737 13 of 21 13 of 21 Table 7. Comparison of extrapolation errors obtained with various ﬁt methods. The data refer to K-decreasing tests at R = 0.8 and Kmax tests conducted on S690QL in lab air. The evaluation comprised ten tests for ASTM threshold and thirteen specimens for ISO threshold. All values are given in MPa·m1/2. ∆Kth,ASTM −∆Kth,cens,ASTM ∆Kth,ISO −∆Kth,cens,ISO Method mean( · ) min( · ) mean( · ) min( · ) Linear, all data points (Equation (1)) 0.20 0.17 0.07 0.02 Linear, ﬁrst n data points (Equation (1)) 0.23 0.18 0.07 0.02 Polynomial, neg. exp. (Equation (5)) 0.04 −0.02 0.03 −0.05 Polynomial, neg. exp. P3 = 4 (Equation (6)) 0.03 0.01 0.03 0.00 Polynomial, neg. exp. P3 = 5 (Equation (7)) 0.01 −0.02 0.02 −0.02 Method Materials 2022, 15, 4737 14 of 21 14 of 21 3.2. Deﬁnition of the Fitting Function and Interval for the Determination of Thresholds Obtained at R ≈0 8 Based on the results presented in Section 3.1, we suggested to use Equation (6) for ﬁt- ting the test data obtained at R ≈0.8. This polynomial exhibits only a minimal dependency on scatter in test data and returns robust results that simultaneously show only a small, but persistent conservativeness. Furthermore, data extrapolation has been shown to be valid if the lowest available crack propagation rate fulﬁlls da/dNmin ≤3 · da/dNth. Since an augmented ﬁt interval shows an additional reduction in standard devia- tion, we suggested to use an augmented interval of 5 × 10−8 mm/cycle ≤da/dN ≤ 10−6 mm/cycle compared to the one proposed in the ASTM standard (10−7 mm/cycle ≤ da/dN ≤10−6 mm/cycle). Regarding the ISO operational deﬁnition, we recommended to use the suggested interval of 10−8 mm/cycle ≤da/dN ≤10−7 mm/cycle, since crack propagation rates much lower than 10−8 mm/cycle are time consuming using conventional FCG testing. Using the herein proposed method for evaluating ∆Kth, the artiﬁcial conservativeness can be reduced and the ﬁts’ robustness improved compared to the linear functions, as shown in the previous paragraphs. Since the evaluation following the proposed method involves a very low effort compared to the conducted experiments, the application thereof is generally preferable over the linear ﬁts. 3.3. Evaluation of the Fatigue Crack Propagation Threshold at R = −1 In contrast to tests carried out at R = 0.8, specimens tested at lower load ratios like R = −1 may exhibit a distinct inﬂuence of extrinsic effects such as crack closure. This can be observed in the example reported in Figure 10: due to the progressive development of crack closure during the load shedding test, the crack propagation rate decreases rapidly in the near-threshold regime, leading to a steep crack propagation curve towards the threshold. This has major consequences on the evaluation of the fatigue crack propagation threshold due to the fact that much fewer experimental points are available in the selected ﬁtting intervals. g The linear ﬁt over all points within the interval (Figure 10a) is not capable of handling the pronounced curvature. Hence, ∆Kth,ASTM is calculated overly conservatively. This holds true also for the polynomial with P3 = 4, Equation (6) (Figure 10c). In contrast, the linear ﬁt using the ﬁrst n points (Figure 10b) shows a fairly good threshold approximation. The best results are obtained using the three-parameter polynomial Equation (5) (Figure 10d), which adapts to the curvature displayed by the data fairly well and results in the least (but still) conservative ∆Kth,ASTM value. 12 13 14 ∆K in MPa·m1/2 10−7 10−6 da/dN in mm/cycle Experimental dataset Fitting dataset Fitting curve (Eq. (1), all data points) ∆Kth,ASTM = 11.45MPa·m1/2 12 13 14 ∆K in MPa·m1/2 10−7 10−6 da/dN in mm/cycle Experimental dataset Fitting dataset Fitting curve (Eq. (1), ﬁrst n data points) ∆Kth,ASTM = 11.87MPa·m1/2 (a) (b) Figure 10. Cont. 12 13 14 ∆K in MPa·m1/2 10−7 10−6 da/dN in mm/cycle Experimental dataset Fitting dataset Fitting curve (Eq. (1), ﬁrst n data points) ∆Kth,ASTM = 11.87MPa·m1/2 12 13 14 ∆K in MPa·m1/2 10−7 10−6 da/dN in mm/cycle Experimental dataset Fitting dataset Fitting curve (Eq. (1), all data points) ∆Kth,ASTM = 11.45MPa·m1/2 (a) (b) Figure 10. Cont. Materials 2022, 15, 4737 15 of 21 12 13 14 ∆K in MPa·m1/2 10−7 10−6 da/dN in mm/cycle Experimental dataset Fitting dataset Fitting curve (Eq. (6)) ∆Kth,ASTM = 11.77MPa·m1/2 12 13 14 ∆K in MPa·m1/2 10−7 10−6 da/dN in mm/cycle Experimental dataset Fitting dataset Fitting curve (Eq. (5)) ∆Kth,ASTM = 11.89MPa·m1/2 (c) (d) Figure 10. FCG data showing a distinct effect of fatigue crack closure leading to crack arrest. The tests have been conducted following a compression precracking load reduction procedure at R = −1 in lab air. 3.3. Evaluation of the Fatigue Crack Propagation Threshold at R = −1 Different ﬁtting strategies have been used: (a) the linear ﬁt using all data in the interval 10−7 mm/cycle ≤da/dN ≤10−6 mm/cycle provided an overly conservative ∆Kth,ASTM value; (b) the linear ﬁt using the ﬁrst n = 5 data points (out of 29 within the interval) showed a fairly good ∆Kth,ASTM approximation; (c) Equation (6), using a ﬁxed exponent P3 = 4, was not capable of reproducing the curvature and therefore gave a very conservative result; (d) a very good result could be achieved using the three-parameter variant of the polynomial (P3 ≈24.1). 12 13 14 ∆K in MPa·m1/2 10−7 10−6 da/dN in mm/cycle Experimental dataset Fitting dataset Fitting curve (Eq. (5)) ∆Kth,ASTM = 11.89MPa·m1/2 (d) 12 13 14 ∆K in MPa·m1/2 10−7 10−6 da/dN in mm/cycle Experimental dataset Fitting dataset Fitting curve (Eq. (6)) ∆Kth,ASTM = 11.77MPa·m1/2 (c) (d) (c) Figure 10. FCG data showing a distinct effect of fatigue crack closure leading to crack arrest. The tests have been conducted following a compression precracking load reduction procedure at R = −1 in lab air. Different ﬁtting strategies have been used: (a) the linear ﬁt using all data in the interval 10−7 mm/cycle ≤da/dN ≤10−6 mm/cycle provided an overly conservative ∆Kth,ASTM value; (b) the linear ﬁt using the ﬁrst n = 5 data points (out of 29 within the interval) showed a fairly good ∆Kth,ASTM approximation; (c) Equation (6), using a ﬁxed exponent P3 = 4, was not capable of reproducing the curvature and therefore gave a very conservative result; (d) a very good result could be achieved using the three-parameter variant of the polynomial (P3 ≈24.1). 3.3.1. Handling of Data Affected by Extrinsic Mechanisms 3.3.1. Handling of Data Affected by Extrinsic Mechanisms Extrinsic mechanisms affect the cyclic deformation in the crack wake and therefore inﬂuence the crack growth rate [22]. The results about tests carried out at R = 0.1 and R = −1 showed a kink of the crack propagation curve in the near-threshold regime. For an in detail discussion on these ﬁndings, see [21]. The results depicted in Figure 11 about two specimens tested in lab air, using a conventional K-decreasing procedure at R = −1, pose the question of how to analyze the data to evaluate the fatigue crack propagation threshold. In fact, both specimens show a distinct kink in the FCG data at about da/dN ≈ 10−7 mm/cycle. In such cases, the calculation of a threshold stress intensity factor range is—regardless of the standard used—questionable, since no asymptotic behavior of the da/dN-∆K curve towards ∆Kth is observable (see Section 1). 10 11 12 13 14 15 16 17 18 ∆K in MPa·m1/2 10−7 10−6 da/dN in mm/cycle 11 12 13 14 15 16 17 18 ∆K in MPa·m1/2 10−7 10−6 da/dN in mm/cycle (a) (b) Figure 11. FCG data showing a kink at about the ASTM threshold (da/dNth,ASTM = 10−7 mm/cycle): (a) kinking starts below da/dN = 10−7 mm/cycle; (b) kinking starts slightly above 10−7 mm/cycle. The data refer to conventional K-decreasing at R = −1 in lab air. Both tests have been interrupted, since neither crack arrest has been observed nor da/dNth,ISO has been reached. 10 11 12 13 14 15 16 17 18 ∆K in MPa·m1/2 10−7 10−6 da/dN in mm/cycle (a) 11 12 13 14 15 16 17 18 ∆K in MPa·m1/2 10−7 10−6 da/dN in mm/cycle (b) (b) (a) Figure 11. FCG data showing a kink at about the ASTM threshold (da/dNth,ASTM = 10−7 mm/cycle): (a) kinking starts below da/dN = 10−7 mm/cycle; (b) kinking starts slightly above 10−7 mm/cycle. The data refer to conventional K-decreasing at R = −1 in lab air. Both tests have been interrupted, since neither crack arrest has been observed nor da/dNth,ISO has been reached. No general rule is applicable to the evaluation of the fatigue crack propagation thresholds, but each dataset shall be analysed separately. For instance, Figure 11a shows that the determi- nation of ∆Kth,ASTM would have been possible, since the kink started below da/dNth,ASTM = Materials 2022, 15, 4737 16 of 21 16 of 21 10−7 mm/cycle. 3.3.3. Validation of the Proposed Method After examining all the possible issues which might inﬂuence the robust determination of the fatigue crack propagation threshold, the methods have been validated against various datasets. Conventional K-Decreasing at R = −1 in Lab Air 3.3.1. Handling of Data Affected by Extrinsic Mechanisms If the aim of this K-decreasing test would have been to determine ∆Kth,ASTM, one might have stopped the test after reaching da/dN < da/dNth,ASTM = 10−7 mm/cycle for the first time. Hence, one would have never observed the effect of corrosion on FCG data starting just below ∆K ≈12MPa·m1/2. Nevertheless, as the crack propagation data did not display a pronounced threshold behavior, a value for ∆Kth,ASTM should not be provided. This holds true also for the dataset depicted in Figure 11b, in which corrosion effects started above da/dNth,ASTM = 10−7 mm/cycle. In these cases, we recommend providing the last da/dN–∆K reading recorded within the test as a pure indication of the lowest stress intensity factor range obtained in the load reduction test, which, nevertheless, must not be taken as a fatigue crack propagation threshold. Hence, an automated evaluation of FCG data should only be performed after checking the crack propagation data in the near-threshold regime. 3.3.2. Inﬂuence of a Lower Data Density The majority of test results presented herein show very low noise in conjunction with a fairly high data density. Both properties depend on the quality of raw data and the methodology used to calculate da/dN and ∆K. In Figure 12, the test data presented in Figure 10 are reduced by a factor of two by skipping every second data point. Since still 14 data points distributed over the whole interval of 10−6 mm/cycle ≤da/dN ≤ 10−7 mm/cycle are left, this data set clearly fulﬁlls the requirement of providing at least ﬁve points, deﬁned in [4]. 12 13 14 ∆K in MPa·m1/2 10−7 10−6 da/dN in mm/cycle Artiﬁcially censored experimental dataset Fitting dataset Fitting curve (Eq. (1), ﬁrst n data points) ∆Kth,ASTM = 11.77MPa·m1/2 12 13 14 ∆K in MPa·m1/2 10−7 10−6 da/dN in mm/cycle Artiﬁcially censored experimental dataset Fitting dataset Fitting curve (Eq. (5)) ∆Kth,ASTM = 11.87MPa·m1/2 (a) (b) Figure 12. Test data presented in Figure 10 reduced by 50% to assess the inﬂuence of a lower data density next to da/dNth,ASTM. (a) the linear ﬁt using the best ﬁrst n data points gives an overly conservative ∆Kth,ASTM value; (b) using the three-parameter polynomial Equation (5) shows a very good agreement with test data. 12 13 14 ∆K in MPa·m1/2 10−7 10−6 da/dN in mm/cycle Artiﬁcially censored experimental dataset Fitting dataset Fitting curve (Eq. (1), ﬁrst n data points) ∆Kth,ASTM = 11.77MPa·m1/2 (a) 12 13 14 ∆K in MPa·m1/2 10−7 10−6 da/dN in mm/cycle Artiﬁcially censored experimental dataset Fitting dataset Fitting curve (Eq. (5)) ∆Kth,ASTM = 11.87MPa·m1/2 (b) (b) (a) Figure 12. Test data presented in Figure 10 reduced by 50% to assess the inﬂuence of a lower data density next to da/dNth,ASTM. (a) the linear ﬁt using the best ﬁrst n data points gives an overly conservative ∆Kth,ASTM value; (b) using the three-parameter polynomial Equation (5) shows a very good agreement with test data. Here, the advantages of an appropriate nonlinear ﬁt functions apply. The linear ﬁt using the ﬁrst n points shows a pronounced underestimation of the threshold stress intensity range, whereas the three-parameter polynomial’s sensitivity to the number of data points is very limited. Conventional K-Decreasing at R = −1 in Lab Air The ﬁrst datasets investigated stemmed from a total of eight SENB specimens made of S690QL, whereof three showed a kink (see Figure 11) and therefore have not been consid- ered in the analysis. These tests have been conducted using the K-decreasing procedure included in the standards [4,6] at constant load ratio R = −1 in lab air. All ﬁve remaining datasets contained data points below da/dN = 1.1 × 10−7 mm/cycle and therefore have Materials 2022, 15, 4737 17 of 21 17 of 21 been considered valid for ASTM operational threshold evaluation, while two of them were also valid regarding the ISO deﬁnition. The evaluated fatigue crack propagation thresholds, including the corresponding standard deviations, are presented in Table 10. Note that, in case of ISO, no standard deviation has been calculated due to insufﬁcient data available. been considered valid for ASTM operational threshold evaluation, while two of them were also valid regarding the ISO deﬁnition. The evaluated fatigue crack propagation thresholds, including the corresponding standard deviations, are presented in Table 10. Note that, in case of ISO, no standard deviation has been calculated due to insufﬁcient data available. Table 10. Comparison of threshold values obtained with various ﬁt methods in case of K-decreasing tests conducted on S690QL at R = −1 in lab air. tests conducted on S690QL at R = −1 in lab air. Method ∆Kth,ASTM in MPa·m1/2 ∆Kth,ISO in MPa·m1/2 Linear, all data points (Equation (1)) 13.43 ± 0.91 12.78 Linear, ﬁrst n data points (Equation (1)) 13.67 ± 0.80 12.89 Polynomial, neg. exp. (Equation (5)) 13.69 ± 0.82 12.92 Polynomial, neg. exp. P3 = 4 (Equation (6)) 13.56 ± 0.85 12.90 Polynomial, neg. exp. P3 = 5 (Equation (7)) 13.57 ± 0.85 12.90 The conservativeness of each method has been analyzed: the two-parameter polyno- mials Equation (6) and Equation (7), and especially the linear ﬁt over all points, showed a pronounced underestimation of ∆Kth,ASTM in conjunction with a higher standard devia- tion compared to the linear ﬁt over the ﬁrst n points or the three-parameter polynomial Equation (5). The observations with regard to artiﬁcial conservativeness also apply to the results for ∆Kth,ISO. Comparing the latter two methods, both show comparable results, both regarding the mean ∆Kth value and the standard deviation. Compression Precracking Load Reduction at R = −1 in Lab Air Compression Precracking Load Reduction at R = −1 in Lab Air The datasets investigated stemmed from a total of ten SENB specimens made of S690QL. All specimens showed a distinct threshold behavior and therefore all have been considered in the analysis. These tests have been conducted using compression precracking followed by a load reduction procedure at constant R = −1 in lab air. All ten datasets contained data points below da/dN = 1.1 × 10−7 mm/cycle and therefore have been con- sidered valid for ASTM operational threshold evaluation, whereas none of them contained points in order to calculate a threshold value according to ISO. The threshold stress intensity factor ranges with corresponding standard deviations are presented in Table 11. Table 11. Comparison of threshold values obtained with various ﬁt methods in case of compression precracking load reduction tests carried out at R = −1 in lab air. Method ∆Kth,ASTM in MPa·m1/2 Linear, all data points (Equation (1)) 9.88 ± 0.99 Linear, ﬁrst n data points (Equation (1)) 10.52 ± 0.89 Polynomial, neg. exp. (Equation (5)) 10.54 ± 0.91 Polynomial, neg. exp. P3 = 4 (Equation (6)) 10.20 ± 0.94 Polynomial, neg. exp. P3 = 5 (Equation (7)) 10.22 ± 0.93 Table 11. Comparison of threshold values obtained with various ﬁt methods in case of compression precracking load reduction tests carried out at R = −1 in lab air. The same conclusions as in the previous paragraph can be drawn: the linear ﬁt over the ﬁrst n points and the three-parameter polynomial Equation (5) provided the best results. 3.4. Deﬁnition of the Fitting Function and Interval for Tests Conducted at R ≪0.8 Data obtained at load ratios momentously lower than R = 0.8 might be affected by phenomena like crack-closure or corrosion, which make the determination of the fatigue crack propagation threshold difﬁcult. When a steep gradient or a kink in the near-threshold data are observed, no general or automatic extrapolation of these test datasets without further investigation is advisable. Furthermore, the augmentation of ﬁt intervals to crack propagation rates smaller than da/dNth may lead to non-conservative results and therefore it is not recommended. If a valid threshold behavior is observed, the evaluation using the three-parameter polyno- mial Equation (5) provided constantly conservative, but not overly conservative, results that Materials 2022, 15, 4737 18 of 21 18 of 21 are neither sensitive to the data density nor to scatter. 3.5. Application of the Fitting Methods to the IBESS Dataset The proposed method has been validated further against the data from the IBESS project [23,24].Within the IBESS project, fatigue crack propagation tests have been per- formed on two different structural steels, S355NL and S960QL, using both standard K- decreasing procedures and CPLR tests at constant load ratios varying between R = −1 and R = 0.7. The number of specimens tested at each stress ratio was limited; therefore, the present validation considered just those tests for which a meaningful data analysis could be performed. In particular, three tests at R = 0 for the S355NL and three tests at R = 0.5 for the S960QL have been considered. Furthermore, according to the recommendations on the ﬁtting intervals given in this work, the data have been further narrowed. In case of the S355NL, all three datasets have been considered valid with respect to the ASTM ﬁtting interval, whereas just two among them could be used for the determination of the fatigue crack propagation threshold according to ISO. For the S960QL, only two tests for ASTM and one for ISO have been included in the analysis. The results displayed in Tables 12 and 13 conﬁrm the conclusions drawn for the S690QL: the linear ﬁt using the ﬁrst n points and the polynomial with three parameters reduce the conservativeness in the evaluation of the fatigue crack propagation thresholds. The method proposed in the standards always provides the most conservative results. Furthermore, for the datasets with enough valid data in the ﬁtting interval (∆Kth,ASTM), the polynomial with three parameters provided the smallest standard deviation. Table 12. Comparison of the threshold values for the S355NL tested at R = 0 obtained with various ﬁtting methods—a total of three sets are eligible for ∆Kth,ASTM evaluation and two for ∆Kth,ISO. Method ∆Kth,ASTM in MPa·m1/2 ∆Kth,ISO in MPa·m1/2 Linear, all data points (Equation (1)) 5.96 ± 0.40 5.75 Linear, ﬁrst n data points (Equation (1)) 6.02 ± 0.38 5.86 Polynomial, neg. exp. (Equation (5)) 6.08 ± 0.33 5.88 Table 13 Comparison of the threshold values for the S960QL tested at R = 0 5 obtained with various Table 12. Comparison of the threshold values for the S355NL tested at R = 0 obtained with various ﬁtting methods—a total of three sets are eligible for ∆Kth,ASTM evaluation and two for ∆Kth,ISO. Compression Precracking Load Reduction at R = −1 in Lab Air Therefore, we recommend this function over the linear fit over the first n points, which indeed performed well on regular shaped datasets. The fitting intervals shall follow the recommendations provided in the standards (10−7 mm/cycle ≤da/dN ≤10−6 mm/cycle for ASTM and 10−8 mm/cycle ≤da/dN ≤ 10−7 mm/cycle for ISO). In case no definite threshold is reached, i.e., no crack arrest is observed (due for instance to anti-shielding effects, see Figure 11a), ∆Kth cannot be determined. Therefore, we recommended providing the last da/dN- ∆K data pair for pure orientation, which shall not be intended as substitute for ∆Kth. 3.5. Application of the Fitting Methods to the IBESS Dataset Method ∆Kth,ASTM in MPa·m1/2 ∆Kth,ISO in MPa·m1/2 Linear, all data points (Equation (1)) 5.96 ± 0.40 5.75 Linear, ﬁrst n data points (Equation (1)) 6.02 ± 0.38 5.86 Polynomial, neg. exp. (Equation (5)) 6.08 ± 0.33 5.88 Table 13. Comparison of the threshold values for the S960QL tested at R = 0.5 obtained with various ﬁtting methods—a total of two sets eligible for ∆Kth,ASTM evaluation and one for ∆Kth,ISO. Method ∆Kth,ASTM in MPa·m1/2 ∆Kth,ISO in MPa·m1/2 Linear, all data points (Equation (1)) 3.30 3.04 Linear, ﬁrst n data points (Equation (1)) 3.31 3.10 Polynomial, neg. exp. (Equation (5)) 3.31 3.08 Table 12. Comparison of the threshold values for the S355NL tested at R = 0 obtained with various ﬁtting methods—a total of three sets are eligible for ∆Kth,ASTM evaluation and two for ∆Kth,ISO. Table 13. Comparison of the threshold values for the S960QL tested at R = 0.5 obtained with various ﬁtting methods—a total of two sets eligible for ∆Kth,ASTM evaluation and one for ∆Kth,ISO. 4. Conclusions The present paper compared several methods for the evaluation of the fatigue crack propagation thresholds. New ﬁtting strategies have been introduced and calibrated on a large dataset of crack growth data for the S690QL. The goodness of the ﬁtting methods has been validated further against a dataset for S355NL and S960QL. g The following conclusions can be drawn: The following conclusions can be drawn: • The ASTM E647 and ISO 12108 standards suggest to ﬁt log ∆K over log da/dN data using a linear ﬁt, but leave plenty of room for interpretation with respect to the choice of the points in the ﬁtting interval; Materials 2022, 15, 4737 19 of 21 19 of 21 • When using all data points within the suggested ﬁtting intervals, the most conservative values of ∆Kth are obtained. However, the ﬁt is not very subjected to scattered data; • To use only the ﬁrst n data points starting from the threshold crack propagation rate in order to ensure the best linear ﬁt reduces the conservativeness at the cost of a more pronounced susceptibility to scatter and lower density of the data; • The proposed ﬁtting polynomials provided an improvement with respect to the goodness of the ﬁt and susceptibility to scatter; • An extrapolation of data was possible within given bounds for the structural steel S690QL, tested in lab air at room temperature at R ≈0.8. Further tests comprising changes in materials, temperatures and the test environment should be conducted to assess the validity ranges; • Tests subjected to crack closure phenomenon cannot be assessed in a fully automatic manner and require a manual dataset evaluation. • Tests subjected to crack closure phenomenon cannot be assessed in a fully automatic manner and require a manual dataset evaluation. Author Contributions: Conceptualization, J.A.S., L.D., M.B.G., M.M., U.Z. and M.K.; methodology, J.A.S., L.D., M.B.G., M.M., U.Z. and M.K.; software, J.A.S. and L.D.; validation, J.A.S. and L.D.; formal analysis, J.A.S., L.D., M.B.G. and M.M.; investigation, J.A.S., L.D., M.B.G., M.M., U.Z. and M.K.; resources, M.O.; data curation, J.A.S. and L.D.; writing—original draft preparation, J.A.S.; writing— review and editing, L.D., M.B.G., M.M., U.Z., M.K. and M.O.; visualization, J.A.S.; supervision, M.M., U.Z., M.K. and M.O.; project administration, M.M., U.Z., M.K. and M.O.; funding acquisition, U.Z. and M.O. All authors have read and agreed to the published version of the manuscript. Author Contributions: Conceptualization, J.A.S., L.D., M.B.G., M.M., U.Z. 4. Conclusions and M.K.; methodology, J.A.S., L.D., M.B.G., M.M., U.Z. and M.K.; software, J.A.S. and L.D.; validation, J.A.S. and L.D.; formal analysis, J.A.S., L.D., M.B.G. and M.M.; investigation, J.A.S., L.D., M.B.G., M.M., U.Z. and M.K.; resources, M.O.; data curation, J.A.S. and L.D.; writing—original draft preparation, J.A.S.; writing— review and editing, L.D., M.B.G., M.M., U.Z., M.K. and M.O.; visualization, J.A.S.; supervision, M.M., U.Z., M.K. and M.O.; project administration, M.M., U.Z., M.K. and M.O.; funding acquisition, U.Z. and M.O. All authors have read and agreed to the published version of the manuscript. Author Contributions: Conceptualization, J.A.S., L.D., M.B.G., M.M., U.Z. and M.K.; methodology, J.A.S., L.D., M.B.G., M.M., U.Z. and M.K.; software, J.A.S. and L.D.; validation, J.A.S. and L.D.; formal analysis, J.A.S., L.D., M.B.G. and M.M.; investigation, J.A.S., L.D., M.B.G., M.M., U.Z. and M.K.; resources, M.O.; data curation, J.A.S. and L.D.; writing—original draft preparation, J.A.S.; writing— review and editing, L.D., M.B.G., M.M., U.Z., M.K. and M.O.; visualization, J.A.S.; supervision, M.M., U.Z., M.K. and M.O.; project administration, M.M., U.Z., M.K. and M.O.; funding acquisition, U.Z. and M.O. All authors have read and agreed to the published version of the manuscript. Funding: This work is part of the research project IGF 20530 N / 1263 “Ermittlung des intrinsischen Schwellenwerts und dessen Validierung als Werkstoffparameter” from the Research Association for Steel Application (FOSTA), Düsseldorf, which is supported by the Federal Ministry of Economic Affairs and Climate Action through the German Federation of Industrial Research Associations (AiF) as part of the program for promoting industrial cooperative research (IGF) on the basis of a decision by the German Bundestag. The project is carried out at BAM Berlin and MPA-IfW Darmstadt. Institutional Review Board Statement: Not applicable. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: Not applicable. Acknowledgments: We acknowledge support by the Deutsche Forschungsgemeinschaft (DFG—German Research Foundation) and the Open Access Publishing Fund of the Technical University of Darmstadt. References 1. Paris, P.C.; Sih, G.C. Stress Analysis of Cracks; Technical Report; NASA: Washington, DC, USA, 1964. 1. Paris, P.C.; Sih, G.C. Stress Analysis of Cracks; Technical Report; NASA: Washington, DC, USA, 1 y f p g 2. Knott, J.F. Fundamentals of Fracture Mechanics; Butterworth: London, UK, 1973. 3. Tada, H.; Paris, P.C.; Irwin, G.R. The Stress Analysis of Cracks Handbook, 3rd ed.; ASME Press: New York, NY, USA, 2000. 3. Tada, H.; Paris, P.C.; Irwin, G.R. The Stress Analysis of Cracks Handbook, 3rd ed.; ASME Press: New York, NY, USA, 2000. 4 ASTM E647-15e1; Standard Test Method for Measurement of Fatigue Crack Growth Rates ASTM International: West 3. Tada, H.; Paris, P.C.; Irwin, G.R. The Stress Analysis of Cracks Handbook, 3rd ed.; ASME Press: New York, NY, USA, 2000. 4. ASTM E647-15e1; Standard Test Method for Measurement of Fatigue Crack Growth Rates. ASTM International: West Con shohocken, PA, USA, 2015. [CrossRef] Mechanisms of Fatigue-Crack Propagation in Ductile and Brittle Solids. Int. J. Fract. 1999, 100, 55–83. [CrossR 5. Ritchie, R. Mechanisms of Fatigue-Crack Propagation in Ductile and Brittle Solids. Int. J. Fract. 199 g p g 18; Metallic Materials—Fatigue Testing—Fatigue Crack Growth Method. ISO: Geneva, Switzerland, 2018. ISO 12108:2018; Metallic Materials—Fatigue Testing— 6. ISO 12108:2018; Metallic Materials—Fatigue Testing—Fatigue Crack Growth Method. ISO: Geneva 7. Bucci, R.J. Development of a proposed ASTM standard test method for near-threshold fatigue crack growth rate measurement. In Fatigue Crack Growth Measurement and Data Analysis. ASTM STP 738; Hudak, S.J., Bucci, R.J., Eds.; ASTM International: West Conshohocken, PA, USA, 1981; pp. 5–28. [CrossRef] 7. Bucci, R.J. Development of a proposed ASTM standard test method for near-threshold fatigue crack growth rate measurement. In Fatigue Crack Growth Measurement and Data Analysis. ASTM STP 738; Hudak, S.J., Bucci, R.J., Eds.; ASTM International: West Conshohocken, PA, USA, 1981; pp. 5–28. [CrossRef] Conshohocken, PA, USA, 1981; pp. 5–28. [CrossRef pp er, D. A Weibull Analysis of Fatigue-Crack Propagation Data from a Nuclear Pressure Vessel Steel. Eng. Fract 184. [CrossRef] , , [ ] 9. Smith, F.; Hoeppner, D.W. Use of the Four Parameter Weibull Function for Fitting Fatigue and Compliance Calibration Data. Eng. Fract. Mech. 1990, 36, 173–178. [CrossRef] 9. Smith, F.; Hoeppner, D.W. Use of the Four Parameter Weibull Function for Fitting Fatigue and Comp Fract. Mech. 1990, 36, 173–178. [CrossRef] D.W. Use of the Four Parameter Weibull Function for Fitting Fatigue and Compliance Calibration Data. Eng 6, 173–178. 4. Conclusions The following symbols and abbreviations are used in this manuscript: ∆F applied force range ∆K applied stress intensity factor range ∆KLR K-decreasing FCG test procedure at constant load ratio ∆Kth fatigue crack propagation threshold ∆Kth,ASTM ∆Kth referring to the ASTM operational deﬁnition ∆Kth,ISO ∆Kth referring to the ISO operational deﬁnition ∆Kth,cens ∆Kth obtained for the censoring of the data set µ mean value of the distribution σu ultimate tensile strength σy upper yield strength a crack size ∆F applied force range ∆K applied stress intensity factor range ∆KLR K-decreasing FCG test procedure at constant load ratio ∆Kth fatigue crack propagation threshold ∆Kth,ASTM ∆Kth referring to the ASTM operational deﬁnition ∆Kth,ISO ∆Kth referring to the ISO operational deﬁnition ∆Kth,cens ∆Kth obtained for the censoring of the data set µ mean value of the distribution σu ultimate tensile strength σy upper yield strength a crack size ∆F applied force range ∆K applied stress intensity factor range ∆KLR K-decreasing FCG test procedure at constant load ratio ∆Kth fatigue crack propagation threshold ∆Kth,ASTM ∆Kth referring to the ASTM operational deﬁnition ∆Kth,ISO ∆Kth referring to the ISO operational deﬁnition ∆Kth,cens ∆Kth obtained for the censoring of the data set µ mean value of the distribution σu ultimate tensile strength σy upper yield strength a crack size Materials 2022, 15, 4737 20 of 21 A elongation at break da/dN fatigue crack propagation rate da/dNth da/dN referring to an operational threshold deﬁnition da/dNth,ASTM ASTM operational threshold deﬁnition of da/dNth da/dNth,ISO ISO operational threshold deﬁnition of da/dNth e Weibull threshold parameter E Young’s modulus I degree of the polynomial k Weibull shape parameter Kb Weibull instability parameter Kmax maximum stress intensity factor in a loading cycle KV2 impact energy n number of data points N number of loading cycles Pi ﬁtting parameters, i ∈N R stress ratio Rmax maximum stress ratio within Kmax-FCG test v Weibull characteristic value ASTM American Society for Testing and Materials BAM Bundesanstalt für Materialforschung und -prüfung CPLR compression precracking load reduction FCG fatigue crack growth ISO International Organization for Standardization MPA-IfW Materialprüfungsanstalt Darmstadt, Institut für Werkstoffkunde ODR orthogonal distance regression SD standard deviation SENB single edge notch bending 24. Kucharczyk, P.; Madia, M.; Zerbst, U.; Schork, B.; Gerwin, P.; Münstermann, S. Fracture-mechanics based prediction of the fatigue strength of weldments. Material aspects. Eng. Fract. Mech. 2018, 198, 79–102. [CrossRef] 22. Pippan, R.; Hohenwarter, A. Fatigue crack closure: A review of the physical phenomena. Fatigue Fract. Eng. Mater. Struct. 2017, 40, 471–495. [CrossRef] [PubMed] 23 Z b U A l i h b h h i h E i l d S h i f i k i S h iß bi d (IBESS A3) T h i l References [CrossRef] 10. Boggs, P.T.; Rogers, J.E. Orthogonal distance regression. In Statistical Analysis of Measurement Error Models and Applications; American Mathematical Society: Providence, RI, USA, 1990; Volume 112, pp. 183–194. [CrossRef] y pp 11. Döker, H. Fatigue Crack Growth Threshold: Implications, Determination and Data Evaluation. Int. J. Fatigue 1997, 19, 145–149. [CrossRef] 11. Döker, H. Fatigue Crack Growth Threshold: Implications, Determination and Data Evaluation [CrossRef] 12. ASTM E 647. Annual Book of ASTM Standards; Section 4; American Society of Testing and Materials: Philadelphia, PA, USA, 1995; Volume 03.01, pp. 578–614. 13. Klesnil, M.; Lukáš, P. Inﬂuence of Strength and Stress History on Growth and Stabilisation of Fatigue Cracks. Eng. Fract. Mech. 1972, 4, 77–92. [CrossRef] 14. Romvari, P.; Tot, L.; Nad’, D. Analysis of Irregularities in the Distribution of Fatigue Cracks in M 12, 1481–1492. [CrossRef] 15. Zheng, X. A Simple Formula for Fatigue Crack Propagation and a New Method for the Determination of ∆Kth. Eng. Fract. Mech. 1987, 27, 465–475. [CrossRef] , , [ ] 16. Ramsamooj, D.; Shugar, T. Model Prediction of Fatigue Crack Propagation in Metal Alloys in Laboratory Air. Int. J. Fatigue 2001, 23 287 300 [C R f] [ ] 16. Ramsamooj, D.; Shugar, T. Model Prediction of Fatigue Crack Propagation in Metal Alloys in Laboratory Air. Int. J. Fatigue 2001, 23, 287–300. [CrossRef] 17. Day, B.; Goswami, T. Weibull Model Development for Fatigue Crack Growth. J. Mech. Behav. Mater. 200 17. Day, B.; Goswami, T. Weibull Model Development for Fatigue Crack Growth. J. Mech. Behav. Mater. 2002, 13, 283–296. [CrossRef] 21 of 21 21 of 21 Materials 2022, 15, 4737 18. Paolino, D.S.; Cavatorta, M.P. Sigmoidal Crack Growth Rate Curve: Statistical Modelling and Applications: SIGMOIDAL CRACK-GROWTH-RATE CURVE. Fatigue Fract. Eng. Mater. Struct. 2013, 36, 316–326. [CrossRef] g g [ ] 19. Maierhofer, J.; Pippan, R.; Gänser, H.P. Modiﬁed NASGRO Equation for Physically Short Cracks. Int. J. Fatigue 2014, 59, 200–207. [CrossRef] 20. E1823-20; Standard Terminology Relating to Fatigue and Fracture Testing. ASTM International: West Conshohocken, PA, USA, 2020. [CrossRef] 21. Duarte, L.; Schönherr, J.A.; Madia, M.; Zerbst, U.; Geilen, M.B.; Klein, M.; Oechsner, M. Recent Developments in Fatigue Crack Propagation Threshold Determination. Int. J. Fatigue 2022. submitted. 22. Pippan, R.; Hohenwarter, A. Fatigue crack closure: A review of the physical phenomena. Fatigue Fract. Eng. Mater. Struct. 2017, 40, 471–495. [CrossRef] [PubMed] 23. Zerbst, U. 23. Zerbst, U. Analytische bruchmechanische Ermittlung der Schwingfestigkeit von Schweißverbindung Report; Bundesanstalt für Materialforschung und -prüfung (BAM): Berlin, Germany, 2016. References Analytische bruchmechanische Ermittlung der Schwingfestigkeit von Schweißverbindu Report; Bundesanstalt für Materialforschung und -prüfung (BAM): Berlin, Germany, 2016. 24. Kucharczyk, P.; Madia, M.; Zerbst, U.; Schork, B.; Gerwin, P.; Münstermann, S. Fracture-mechanics based prediction of the fatigue strength of weldments. Material aspects. Eng. Fract. Mech. 2018, 198, 79–102. [CrossRef]
https://openalex.org/W1986376280	https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0095458&type=printable	English	null	Rapid Development of Microsatellite Markers for Callosobruchus chinensis Using Illumina Paired-End Sequencing	PloS one	2,014	cc-by	7,649	Rapid Development of Microsatellite Markers for Callosobruchus chinensis Using Illumina Paired-End Sequencing Can-xing Duan, Dan-dan Li, Su-li Sun, Xiao-ming Wang, Zhen-dong Zhu* Can-xing Duan, Dan-dan Li, Su-li Sun, Xiao-ming Wang, Zhen-dong Zhu* The National Key Facility for Crop Gene Resources and Genetic Improvement, Institute of Crop Science, Chinese Acad Can-xing Duan, Dan-dan Li, Su-li Sun, Xiao-ming Wang, Zhen-dong Zhu* The National Key Facility for Crop Gene Resources and Genetic Improvement, Institute of Crop Science, Chinese Academy of Agricultural Sciences, Beijing, People’s Republic of China Can-xing Duan, Dan-dan Li, Su-li Sun, Xiao-ming Wang, Zhen-dong Zhu The National Key Facility for Crop Gene Resources and Genetic Improvement, Institute of Crop Science, Chinese Academy of Agricultural Sciences, Beijing, People’s Republic of China Abstract doi:10.1371/journal.pone.0095458 Editor: Kun Yan Zhu Kansas State University United States of America Editor: Kun Yan Zhu, Kansas State University, United States of America Editor: Kun Yan Zhu, Kansas State University, United States of America Received October 30, 2013; Accepted March 27, 2014; Published May 16, 2014 Received October 30, 2013; Accepted March 27, 2014; Published May 16, 2014 Copyright: 2014 Duan et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: This work was supported by the Public Welfare Special Fund (2060302-2-13) from Institute of Crop Sciences, CAAS and Modern Agro-industry Technology Research System (CARS-09) from the Ministry of Agriculture of China. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: zhuzhendong@caas.cn Abstract Background: The adzuki bean weevil, Callosobruchus chinensis L., is one of the most destructive pests of stored legume seeds such as mungbean, cowpea, and adzuki bean, which usually cause considerable loss in the quantity and quality of stored seeds during transportation and storage. However, a lack of genetic information of this pest results in a series of genetic questions remain largely unknown, including population genetic structure, kinship, biotype abundance, and so on. Co-dominant microsatellite markers offer a great resolving power to determine these events. Here, we report rapid microsatellite isolation from C. chinensis via high-throughput sequencing. Principal Findings: In this study, 94,560,852 quality-filtered and trimmed reads were obtained for the assembly of genome using Illumina paired-end sequencing technology. In total, the genome with total length of 497,124,785 bp, comprising 403,113 high quality contigs was generated with de novo assembly. More than 6800 SSR loci were detected and a suit of 6303 primer pair sequences were designed and 500 of them were randomly selected for validation. Of these, 196 pair of primers, i.e. 39.2%, produced reproducible amplicons that were polymorphic among 8 C. chinensis genotypes collected from different geographical regions. Twenty out of 196 polymorphic SSR markers were used to analyze the genetic diversity of 18 C. chinensis populations. The results showed the twenty SSR loci were highly polymorphic among these populations. Conclusions: This study presents a first report of genome sequencing and de novo assembly for C. chinensis and demonstrates the feasibility of generating a large scale of sequence information and SSR loci isolation by Illumina paired- end sequencing. Our results provide a valuable resource for C. chinensis research. These novel markers are valuable for future genetic mapping, trait association, genetic structure and kinship among C. chinensis. Citation: Duan C-x, Li D-d, Sun S-l, Wang X-m, Zhu Z-d (2014) Rapid Development of Microsatellite Markers for Callosobruchus chinensis Using Illumina Paired-End Sequencing. PLoS ONE 9(5): e95458. doi:10.1371/journal.pone.0095458 Citation: Duan C-x, Li D-d, Sun S-l, Wang X-m, Zhu Z-d (2014) Rapid Development of Microsatellite Markers for Callosobruchus chinensis Using Illumina Paired-End Sequencing. PLoS ONE 9(5): e95458. Introduction Population code Number of individuals Occurrence location Collection time Host DZ 30 Dengzhou, Henan 2012 mungbean FC 30 Fuchuan, Guangxi 2012 mungbean HH 30 Hohhot, Inner Mongolia 2012 mungbean MT 30 Mengtougou, Beijing 2013 adzuki bean QD 30 Qingdao, Shandong 2013 mungbean TS 30 Tangshan, Hebei 2013 adzuki bean UM 30 Urumuqi, Xinjiang 2012 mungbean WL 30 Wulong, Chongqing 2012 mungbean doi:10.1371/journal.pone.0095458.t001 doi:10.1371/journal.pone.0095458.t001 Illumina platform was at first confined to re-sequencing applica- tions which rely on a reference sequence. Recently, with increased read length by Illumina technology improvement and develop- ment of new computational tools, short reads can be assembled and used for transcriptome or genome analysis. Illumina sequences now can produce moderately long reads (up to 150 bp with the GAIIx, and 100 bp with the HiSeq 2000) and accommodate paired-end sequencing from both ends of ,200– 600 bp fragments. There have also been massive increases in the number of reads obtained per Illumina sequencing run. Further- more, compared with relatively constant cost of 454 sequencing, the cost of obtaining Illumina sequence data has dropped substantially [22]. De novo assemblies using sequence reads from Illumina technology have been reported in many non-model organisms without a reference sequence [23–27]. To take advantage of these advances, we utilized Illumina paired-end sequencing to isolate SSR loci rapidly and massively. methods use a few specific repeated motifs, generally selected without prior knowledge of their abundance in the genome [12], hence introducing potential bias in genome representativeness. Recently developed next-generation sequencing (NGS) plat- forms such as SOLiD (ABI, Norwalk, CT), 454 GS FLX (Roche, Penzberg, Germany), and Illumina Genome Analyzer (Illumina, San Diego, CA) can facilitate high-throughput genome sequencing of both noncoding and coding regions, including large scale resequencing in well-characterized species or de novo transcriptome sequencing for species without reference sequences. These technologies are capable of generating a large number of reads at a relatively low cost. The huge amounts of sequence data produced on these high-throughput platforms can be used to identify a great quantity of genetic markers [13–17]. In the past few years, the 454 pyrosequencing has been preferred for non-model organisms [18–21] due to its longer read length than Illumina sequencing. With shorter read length, Table 2. The information of 18 Callosobruchus chinensis populations used in genetic diversity analysis. Introduction all eukaryotes tested and are often abundant and evenly dispersed [6–8]. Microsatellite sequences are usually characterized by a high degree of length polymorphism, and are ideal single-locus co- dominant markers for genetical studies. Microsatellites are the most frequently used DNA markers in many areas of research because of their high abundance, multi-allelic nature, high polymorphism rates, and rapid detection of the alleles by wide variety of methods [9–11]. The adzuki bean weevil, Callosobruchus chinensis (L.), is one of the most destructive pests of leguminous stored seeds in Asia. C. chinensis larvae utilize a variety of dried legume seeds as their hosts, primarily Vigna species and genera such as Cajanus and Lens [1], [2]. This insect often causes considerable damage to stored mungbean (Vigna radiata L.), cowpea (Vigna unguiculata L.), adzuki bean (Vigna angularis Ohwi & Ohashi), faba bean (Vicia faba L.), pea (Pisum sativum L.), chickpea (Cicer arietinum L.), soybean (Glycine max L. Merr.) and lotus seed (Nelumbo nucifere Gaertn.). Usually, this beetle causes 32,64% Vigna seeds loss during transportation and storage and infestation rate may reach 100% with 3–5 months of storage [3], [4], which results in loss of mass and low germination rate and makes them unfit for human consumption or for agricultural and commercial use [5]. However, their availability and quality are limited by the difficulties of de novo development in organisms without available genomic information. Up to date, none of SSR loci as well as other makers such as AFLP, RAPD and AFLP in C. chinensis has been isolated and identified, which results in a large number of genetic information related to C. chinensis has not been revealed. The most commonly used procedure for microsatellite isolation is enrich- ment of genomic DNA in microsatellite motifs, followed by cloning and sequencing of the enriched DNA library by the Sanger method, which is difficult, time-consuming and costly. Enrichment Microsatellites or simple sequence repeats (SSRs) are tandemly repeated motifs of one to six bases found in the nuclear genomes of May 2014 \| Volume 9 \| Issue 5 \| e95458 PLOS ONE \| www.plosone.org 1 May 2014 \| Volume 9 \| Issue 5 \| e95458 Rapid Development of Microsatellite Markers Rapid Development of Microsatellite Markers Table 1. The information of 8 Callosobruchus chinensis genotypes used in polymorphism test of SSR markers. Table 1. The information of 8 Callosobruchus chinensis genotypes used in polymorphism test of SSR markers. doi:10.1371/journal.pone.0095458.t002 Introduction Population code Number of individuals Occurrence location Collection time Host BB 20 Bobai, Guangxi 2012 mungbean BD 20 Baoding, Hebei 2013 adzuki bean DX 20 Daxin, Guangxi 2012 mungbean DZ 20 Dengzhou, Henan 2012 mungbean HF 20 Hefei, Anhui 2012 mungbean HH 20 Hohhot, Inner Mongolia 2012 mungbean JC 20 Jiangchuan, Yunnan 2012 mungbean LC 20 Lingchuan, Shanxi 2013 adzuki bean LL 20 Luliang, Yunnan 2012 mungbean LS 20 Lishui, Jiangsu 2013 adzuki bean MC 20 Mengcheng, Anhui 2012 mungbean QD 20 Qingdao, Shandong 2013 mungbean QJ 20 Qianjiang, Chongqing 2012 mungbean RD 20 Rudong, Jiangsu 2012 adzuki bean TS 20 Tangshan, Hebei 2013 adzuki bean UM 20 Urumuqi, Xinjiang 2012 mungbean XY 20 Xinye, Henan 2012 adzuki bean YY 20 Yangyuan, Hebei 2013 mungbean doi:10.1371/journal.pone.0095458.t002 Table 2. The information of 18 Callosobruchus chinensis populations used in genetic diversity analysis. May 2014 \| Volume 9 \| Issue 5 \| e95458 PLOS ONE \| www.plosone.org 2 Rapid Development of Microsatellite Markers Table 3. Summary of adzuki bean weevil sequencing data output. Samples ID Raw reads Effective rate (%) Clear reads Q20 (%) Q30 (%) GC Content (%) Bpis1 52,566,784 88.64 46,595,197 98.81; 96.42 96.38; 91.83 36.76; 36.75 Bpis2 54,321,240 88.30 47,965,654 98.73; 96.27 96.16; 91.47 36.75; 36.74 Notes: Effective rate (%) = clear reads/raw reads6100, Q20 and Q30 refer to the values of quality of sequencing data. doi:10.1371/journal.pone.0095458.t003 enzymes were removed. After adenylation of 39 ends of DNA fragments, Illumina PE adapter oligonucleotides were ligated to prepare for hybridization. In order to select DNA fragments of preferentially 500 bp in length, agarose electrophoresis was performed (120V, 40 min, 1.5% agarose) and adapter-ligated constructs derived from 620 bp to 670 bp was separated. After purification using spin column (QIAGEN, Dusseldorf, Germany), DNA fragments with ligated adapter molecules on both ends were selectively enriched using Illumina PCR Primer (F: 59-AATGA- TACGGCGACCACCGAGA-39 and R: 59-CAAGCAGAA- GACGGCATACGAGT-39) Cocktail in a PCR reaction with 10 cycles. Products were purified using AMPure XP system (Beckman Coulter, Beverly, MA) and quantified using the Agilent high sensitivity DNA assay on the Agilent Bioanalyzer 2100 system. To the best of our knowledge, this is the first systematic report on development of microsatellite markers of adzuki bean weevil with de novo assembly using Illumina paired-end sequencing. The present study identified a large scale of putative SSR loci efficiently and inexpensively using Illumina HiSeq 2500 platform. Insect Material The adzuki bean weevil male adults for genome sequencing were collected from infected mungbean stored at the Institute of Crop Science (ICS), Chinese Academy of Agricultural Sciences (CAAS), Beijing. The C. chinensis genotypes for polymorphism detection and genetic diversity were collected from eight and eighteen different regions, respectively (Table 1 and Table 2). DNA Isolation, Quantification and Qualification Total genomic DNA of single male adult with belly removal was extracted with TIANamp Genomic DNA Kit according to the manufacturer’s instructions (TIANGEN, Beijing, China). The quality of extracted DNA was monitored on 1.5% agarose gels. DNA purity was checked using the NanoPhotometer spectropho- tometer (IMPLEN, San Diego, CA). DNA concentration was measured using Qubit DNA Assay Kit in Qubit 2.0 Flurometer (Life Technologies, San Diego, CA). Fragment distribution of DNA library was measured using the DNA Nano 6000 Assay Kit of Agilent Bioanalyzer 2100 system (Agilent Technologies, San Diego, CA). Data Filtering, De Novo Assembly and Decontamination Before the genome assembly, we carried out a stringent filtering process of raw sequencing reads. The reads with more than 10% of bases with a quality score of Q,20, or the paired reads if the N content of a single read exceeds 10% of the read length, ambiguous sequences represented as ‘‘N’’ and adaptor contami- nation were removed. De novo genome assembly was performed by de Bruijn graph and SOAPdenovo [28] software package with the default settings except K-mer value. The high-quality reads were loaded into the computer, and then de Bruijn graph data structure was used to represent the overlap among the reads. After adjusting for various parameters, an 87-mer assembly was finally selected. In order to ensure the accuracy and validity of the SSR search, the contigs less than 500 bp length were filtered out. Introduction The resulting SSR sequences were characterized and validated through successful amplification of randomly selected target loci across a selection of adzuki bean weevil genotypes from diverse geographic regions. Clustering and Sequencing The clustering of the index-coded samples was performed on a cBot Cluster Generation System using TruSeq PE Cluster Kit v3- cBot-HS (Illumia, San Diego, CA) according to the manufacturer’s instructions. After cluster generation, the library preparations were sequenced on an Illumina Hiseq 2500 PE150 platform (Illumina, San Diego, CA) and 150 bp paired-end reads were generated. The sequencing data have been submitted to the National Center for Biotechnology Information (NCBI) short read archive (SRA) and given the accession numbers SRR949786 and SRR952345. DNA Library Preparation for Sequencing A total amount of 3 mg genomic DNA per sample was used as input material for the DNA sample preparation. Sequencing libraries were generated using Illumina Truseq DNA Sample Preparation Kit (Illumina, San Diego, CA) following manufactur- er’s recommendations and x index codes were added to attribute sequences to each sample. Briefly, fragmentation was carried out by hydrodynamic shearing system (Covaris, Woburn, MA) to generate ,800 bp fragments. Remaining overhangs were con- verted into blunt ends via exonuclease/polymerase activities and In view of the fact that some Wolbachia strains are the endosymbionts living in the cytoplasm of bean beetles, such as C. chinensis and Callosobruchus analis [29], [30], the sequencing data Table 4. Re-assembly of adzuki bean weevil genome after decontamination. Sample ID Total length(bp) Contigs number N50 (bp) N90 (bp) GC(%) Bpis 499,585,451 403,617 1,439 608 36.48 doi:10.1371/journal.pone.0095458.t004 PLOS ONE \| www.plosone.org 3 May 2014 \| Volume 9 \| Issue 5 \| e95458 Table 4. Re-assembly of adzuki bean weevil genome after decontamination. Table 4. Re-assembly of adzuki bean weevil genome after decontamination. Table 4. Re-assembly of adzuki bean weevil genome after decontamination. May 2014 \| Volume 9 \| Issue 5 \| e95458 May 2014 \| Volume 9 \| Issue 5 \| e95458 PLOS ONE \| www.plosone.org PLOS ONE \| www.plosone.org 3 Rapid Development of Microsatellite Markers Table 5. Final assembly of adzuki bean weevil genome. Sample ID Total length(bp) Contigs number N50 (bp) N90 (bp) GC(%) Bpis 497,124,785 403,113 1,432 606 36.48 doi:10.1371/journal.pone.0095458.t005 Table 5. Final assembly of adzuki bean weevil genome. Sample ID Total length(bp) Contigs number N50 (bp) N90 (bp) GC(%) Bpis 497,124,785 403,113 1,432 606 36.48 doi:10.1371/journal.pone.0095458.t005 non-denaturing polyacrylamide gels and silver stained [33]. The band sizes were determined roughly based on 100 bp DNA ladder (TIANGEN, Beijing, China). of C. chinensis were likely to be contaminated with Wolbachia’s sequences, hence decontamination had to be performed. First, re- assembly should be performed after removal of the Wolbachia sequences from the valid sequencing data. Specifically, based on the distribution plot of contig depth and GC content, contigs with a depth greater than 506 (contamination contigs) were extracted and cut into the 87-mer fragments. The fragments were aligned with the valid sequence data and the matched reads were removed. Second, the reassembled C. chinensis genome was aligned with records in bacterial and fungal genome databases and the matched contigs were also removed. DNA Library Preparation for Sequencing The genotyping data was subsequently used to determine genetic relationships among 18 diverse C. chinensis populations collected from different regions (Table 2). The number of alleles (Na), expected heterozygosity (He), observed heterozygosity (Ho) and Shannon’s information index (I) were calculated using POPGEN1.32 software [34], [35]. Using the MEGA 4 software, based on Nei’s genetic distance, a phylogenetic tree was constructed using the neighbor-joining clustering method [36]. Illumina Sequencing and Quality Control The paired-end sequencing yielded 26150-bp reads from either end of the DNA fragment. In this study, a total of 32 Gb of raw sequencing data, comprising 106,888,024 raw sequencing reads were generated from a 300 bp insert library. After stringent quality assessment and data filtering, 94,560,852 (88.47%) clean reads with Q20 bases (those with a base quality greater than 20) were deemed as high quality reads and used for further analysis. The sequences have been deposited in NCBI. Output data are shown in Table 3, in which Bpis2 was an additional sequencing data, indicating it was a high-quality (Q20$96%, Q30$91%) sequencing and had a normal sequencing error rate (0.04–0.06%) (Figure S1 and Figure S2). SSR Marker Polymorphism Detection and Assessment Five hundred randomly selected primer pairs were synthesized (Sangon Biotech, Beijing, China) for polymorphism detection among 8 C. chinensis genotypes (Table 1). Polymerase chain reactions (PCR) were carried out in 10 mL reaction volumes containing 10 mmol/L Tris-HCl pH 8.3, 50 mmol/L KCl, 1.5 mM MgCl, 250 mM each of dNTPs, 0.2 M of each primers, 0.5 U Taq polymerase (Dingguo, Beijing, China) and 30 ng of DNA template. Reactions were performed using a GeneAmp PCR System 9700 thermal cycler (ABI, Norwalk, CT) programmed as 94uC for 4 min, followed by 35 cycles of 94uC for 30 s, 50–60uC (it depends optimized annealing temperature) for 30 s and 72uC for 60 s. The final extension was performed at 72uC for 10 min. The PCR products were analyzed by electrophoresis on 8.0% In order to detect whether the sequences were contaminated by endosymbiont sequences, part of the valid sequencing data were analyzed by BLAST against NCBI nucleotide database (NT) and found that 1.54% and 1.57% of the sequencing data were associated with the genus Wolbachia, respectively, indicating that bacterial contamination was present in the samples. Therefore, decontamination has to be performed in order for veritable assembly of the C. chinensis genome. Results Potential SSR markers were detected among the assembled genome using the MISA tool [31] (http://pjrc.ipk-gatersleben.de/ misa/). We searched for SSRs with motifs ranging from monoto hexa-nucleotides in size. The minimum of repeat units were set as follows: ten repeat units for mono-nucleotide, six for dinucleotides and five for tri-, tetra-, penta- and hexa-nucelotides. Mono- nucleotide repeats were discarded in that it was difficult to distinguish genuine mono-nucleotide repeats from polyadenylation products and some mono-nucleotide repeats were generated by base mismatch or sequencing errors. Primer pairs were designed using Primer3 [32] (http://fokker.wi.mit.edu/primer3/) with default parameters. Genome De Novo Assembly and Decontamination The 94,560,852 high-quality clean reads were used to assemble the genome of C. chinensis with SOAPdenovo software. According to the overlapping information of high-quality reads, an 87-mer Table 6. The base content in adzuki bean weevil genome. Base Type Number (bp) % of Genome Adenine (A) 157,892,519 31.76 Thymine (T) 157,873,741 31.76 Cytosine (C) 90,693,363 18.24 Guanine (G) 90,665,162 18.24 GC 36.48% Total 497,124,785 doi:10.1371/journal.pone.0095458.t006 PLOS ONE \| www.plosone.org 4 May 2014 \| Volume 9 \| Issue 5 \| e95458 Table 6. The base content in adzuki bean weevil genome. May 2014 \| Volume 9 \| Issue 5 \| e95458 Rapid Development of Microsatellite Markers Figure 1. Distribution of sequencing depth. doi:10.1371/journal.pone.0095458.g001 Figure 1. Distribution of sequencing depth. doi:10.1371/journal.pone.0095458.g001 assembly was finally selected and the assembled C. chinensis genome was generated with the total length of 499,585,451 bp, comprising of 405,744 contigs with an average length of 1237 bp, a N50 of 1439 bp, N90 of 608 bp and 36.47% of GC content. contigs sequence of each part to the nucleotide database showed that contigs with a depth greater than 50 were contaminated with bacteria. To facilitate better and authentic assembly of the C. chinensis genome, decontamination analysis was carried out (Table S1 and Table 4). Finally, the assembled C. chinensis genome was obtained with the total length of 497,124,785 bp, comprising of 403,113 contigs (Table 5). The Adenine, thymine, cytosine and guanine content and the abundance in C. chinensis genome are shown in Table 6. The coverage of contigs assembled with decontaminated clean reads reaches 99.12%, with an average sequencing depth of about 206(Figure 1). A scatterplot of the sequencing depth against the GC content was generated and the vast majority of the points appeared in a relatively narrow range, indicating the final assembled genome was free of contamination (Figure 2). The GC content scatterplot showing only one species will show a concentrated distribution. Taking account of sequencing data probably contained bases from the genus Wolbachia, the GC distribution was analyzed in order to determine whether the assembly contained contamination from non-target species or not. In the scatterplot of contigs depth against GC content, the points were divided into three parts (Figure S3). A BLAST analysis of the Figure 2. Distribution between GC content and sequencing depth. doi:10.1371/journal.pone.0095458.g002 May 2014 \| Volume 9 \| Issue 5 \| e95458 Identification of SSR Loci After MISA analysis, a total of 6,593 potential simple sequence repeats (SSRs) were identified with the length of 12–64 bp. The majority of the SSR sequences were from 12 to 25 bp in length, accounting for 98.62% of the total identified SSR loci (Figure 3). Of these, the most frequently detected SSR sequences in C. chinensis consisted of the motif of trinucleotide repeats, with 70.02% of abundance, followed by tetranucleotide (16.35%) and dinucleotide repeats (11.80%) (Figure 4A). The length of SSR motifs ranged from 2 to 6 bp and the most common type of SSR motifs comprised tetranucleotide repeats, followed by pentanu- cleotide and trinucleotide (Figure 4B). More than 6800 SSR loci were detected and a total of 6,303 primer pairs were designed by Primer3 (Table S2) for future assessment of validity by locus amplification. Figure 2. Distribution between GC content and sequencing depth. doi:10.1371/journal.pone.0095458.g002 May 2014 \| Volume 9 \| Issue 5 \| e95458 PLOS ONE \| www.plosone.org PLOS ONE \| www.plosone.org 5 Rapid Development of Microsatellite Markers Figure 3. Distribution between SSR length and number of SSRs. doi:10.1371/journal.pone.0095458.g003 distance of 0.14 or so. The majority of populations were grouped into cluster I (RD, HF, OJ, BD, MC, DZ, LS, LC, UM, CQ, QD, YY and DX), while cluster II and cluster III included merely 2 (XY and HH) and 3 (TS, JC and BB) populations, respectively. distance of 0.14 or so. The majority of populations were grouped into cluster I (RD, HF, OJ, BD, MC, DZ, LS, LC, UM, CQ, QD, YY and DX), while cluster II and cluster III included merely 2 (XY and HH) and 3 (TS, JC and BB) populations, respectively. Discussion This study demonstrated that Illumina paired-end sequencing is capable of identifying massive numbers of potentially PCR- amplifiable SSR loci with relative low-cost. We find that on a read- by-read basis, Illumina paired-end sequences are effective for developing SSR markers in non-model organism such as C. chinensis. Compared with Roche GS FLX, the Illumina platform was originally utilized in the organisms with reference genomes [37–40]. With the rapid development of various NGS technologies along with a series of novel assembly methods in recent years, de novo sequencing and assembly of genome or transcriptome have been successfully used for model [41], [42] and non-model organisms [43–46]. Consistent with previous reports, the results from this research also suggested that short reads from Illumina sequencing can be effectively assembled and used for SSR marker development or gene identification in non-model organisms. In this study, more than 94 million high-quality clean reads were used to assemble the genome of C. chinensis. This large dataset resulted in a sequencing depth with an average of 20 folds. Finally, the adzuki bean weevil genome with length of 497,124,785 bp was obtained, comprising 403,113 contigs by means of quality-filtering, assembly, decontamination and reassembly. Figure 3. Distribution between SSR length and number of SSRs. doi:10.1371/journal.pone.0095458.g003 doi:10.1371/journal.pone.0095458.t007 Validation of SSR Assay Five hundred primer pairs were randomly selected to evaluate the rate of amplication and polymorphism in 8 diverse C. chinensis genotypes, of which 403 pairs (80.6%) successfully amplified clear fragments and 365 pairs (73.0%) produced PCR amplicons with the expected size. Of these, 196 pairs (39.2%) were polymorphic among 8 genotypes of C. chinensis (Figure 5 and Table S3). The genetic relationship among 18 diverse C. chinensis populations from different regions were performed using 20 polymorphic SSR pairs randomly selected from 196 pairs (Figure 6 and Table 7). The number of alleles (Na) per locus ranged from 4 to 13 with an average of 7.35, the expected heterozygosity (He) ranged from 0.0591 to 0.7868, and the observed heterozygosity (Ho) varied from 0.0075 to 0.6815. Shannon’s information index (I) values ranged from 0.1659 to 1.6831 with an average of 0.9545 (Table 8). Based on to Nei’s genetic distance, the dendrogram generated from the cluster analysis showed the genetic relationship among 18 C. chinensis populations. As shown in Figure 7, the 18 populations were classified into three distinct genetic groups at the genetic No molecular maker is available in C. chinensis to date. In the previous research, based on the sequence variations of mitochon- drial DNA (mtDNA), cytochrome coxidase I gene (COI), and the second internal transcribed spacer of nuclear ribosomal DNA (rDNA ITS2) rather than molecular marker, genetic relationship among C. chinensis populations was analyzed [47], [48], which was relatively time-consuming, costly and inconvenient. Nowadays, polymorphic SSR markers play an important role in genetic diversity research, linkage map construction, identification of varieties, comparative genomics and related analysis. The genome sequencing and assembly provided plenty of sequences for developing numerous markers in C. chinensis. We performed a general screen on C. chinensis genome for the presence of microsatellites. Based on the de novo assembled genome, a total Figure 4. Distribution of SSR motif. (A) Distribution between SSR motif and number of SSRs. (B) Distribution between SSR motif and number of SSR motif type. doi:10.1371/journal.pone.0095458.g004 Figure 4. Distribution of SSR motif. (A) Distribution between SSR motif and number of SSRs. (B) Distribution between SSR motif and number of SSR motif type. doi:10.1371/journal.pone.0095458.g004 May 2014 \| Volume 9 \| Issue 5 \| e95458 May 2014 \| Volume 9 \| Issue 5 \| e95458 PLOS ONE \| www.plosone.org 6 Rapid Development of Microsatellite Markers Figure 5. Polymorphism of primer CCM46 among 8 Callosobruchus chinensis genotypes. Validation of SSR Assay doi:10.1371/journal.pone.0095458.g005 Figure 6. SSR PCR amplification patterns of 20 Callosobruchus chinensis individuals from RD population using primer CCM46. (M is the 100 bp DNA ladder marker). doi:10.1371/journal.pone.0095458.g006 Figure 5. Polymorphism of primer CCM46 among 8 Callosobruchus chinensis genotypes. doi:10.1371/journal.pone.0095458.g005 Figure 5. Polymorphism of primer CCM46 among 8 Callosobruchus chinensis genotypes. doi:10.1371/journal.pone.0095458.g005 Figure 6. SSR PCR amplification patterns of 20 Callosobruchus chinensis individuals from RD population using primer CCM46. (M is the 100 bp DNA ladder marker). doi:10.1371/journal.pone.0095458.g006 Figure 6. SSR PCR amplification patterns of 20 Callosobruchus chinensis individuals from RD population using primer CCM46. (M is the 100 bp DNA ladder marker). doi:10.1371/journal.pone.0095458.g006 Table 7. Characteristics of 20 polymorphic SSR markers used in genetic diversity analysis (F = forward primer, R = reverse primer, Size = size of cloned allele, Ta = annealing temperature). Table 7. Characteristics of 20 polymorphic SSR markers used in genetic diversity analysis (F = forward primer, R = reverse primer, Size = size of cloned allele, Ta = annealing temperature). Table 7. Characteristics of 20 polymorphic SSR markers used in genetic diversity analysis (F = forward primer, R = reverse primer, Size = size of cloned allele, Ta = annealing temperature). Primer F (59–39) R (59–39) Size (bp) Ta (6C) CCM1 TTGGTGCTAACGTACATGAAATGAA CTCGCATAACGACTTCATGGATTT 155 58 CCM2 CAGGACTGGACTGGACATGTGATA TGCTCAGCCTAATTCAACCTGTTT 156 55 CCM19 AGCTATCCTCATAATTGCTCACGG CCTCCTTATAAGTGCATTTGTTTGC 138 56 CCM25 AAACTCTTGCGTTAATTGTGCCAT TTTACTTGATGACTTGGCAGAGCC 99 55 CCM27 TGTGTTCTATCTAGCGTGACTGGG ATGCTTGTACATTTTGACTTGCCA 158 55 CCM29 CAATGGCATGTCCACCACTACT GATAAGCAATTGCCTCTTATCCCA 146 55 CCM40 AGGGGATATCGTGTTCTCTGTTCA GATCGAACCTCGAAGCAGAAGTT 115 55 CCM46 AACGGACAGACAGAGAGATAGTCAA TCTGTTTGTTTGCCTGTTTGC 134 55 CCM49 GTTGACCGTTGGACAATAAAGCTA TGTCACTACAATCGATTCGTCTCAT 145 55 CCM56 CCTAAGAGTTTCGTTTTTGCATGG GGCAAATTGAGCTCAGCATTAGAT 110 55 CCM65 GTACCGCGGGTACCTACGTACTTT GTGTTTACCGTTTTCAAACGCAG 81 55 CCM66 GGGCTAATCGCACAACTGTAAATC AAATAGCCAGGAGCTAGACTTCGC 141 58 CCM83 AGGGTTTTTCCCGTTTAACGATATT CGGATCTGGAAGCAGTTTTGTTTA 140 57 CCM108 ACTTAAATTACCCGGTGACGTGAA AACTTTTGTAAATCGCGCTCAAAG 131 55 CCM118 TTAATTCTGGTCCTGTTCCGTTTT TCTTCTAATCCCATGACTAGCTTCAC 130 55 CCM125 AGCGTGTTAAAAAGAGAGTGGACG TATTATTCTGCAGATGTGCCGCT 144 57 CCM149 CTTTGCAGTTACGAGAGACAGGGT TGCCAACCTTCTATAAGAACACGG 139 57 CCM154 TTGATCCTTTTCCTCCCAATAAAA TTCGTCCTCATATAGCACAAATGTT 145 53 CCM180 TACAACTGTAAATGGCGTTTGGCT GGCATTTCAACAAAAGTACGCTTC 116 55 CCM189 GGAAGTTGCATATGTGAAGTCCAA AGGATGCATTCTGTGTCATCTGTT 107 56 doi:10.1371/journal.pone.0095458.t007 Table 7. Characteristics of 20 polymorphic SSR markers used in genetic diversity analysis (F = forward primer, R = reverse primer, Size = size of cloned allele, Ta = annealing temperature). May 2014 \| Volume 9 \| Issue 5 \| e95458 7 PLOS ONE \| www.plosone.org PLOS ONE \| www.plosone.org 7 Rapid Development of Microsatellite Markers Table 8. Informativeness of SSR loci following amplification from 18 geographically diverse Callosobruchus chinensis populations in China. Acknowledgments We thank Beijing Novogene Bioinformatics Co., Ltd for the assistance in raw data processing. We also thank Dr. Tao Yang, Dr. Gang-gang Guo and Dr. Jing Wu for useful suggestions in data analyses and submission. We thank Beijing Novogene Bioinformatics Co., Ltd for the assistance in raw data processing. We also thank Dr. Tao Yang, Dr. Gang-gang Guo and Dr. Jing Wu for useful suggestions in data analyses and submission. All in all, development of SSR markers is useful to elucidate the population genetic structure, and to monitor migration and biotype abundance among adzuki bean weevil or sibling species such as cowpea weevil (Callosobruchus maculatus Fabricius) popula- tions. In the meantime, these markers could be used for linkage map construction, gene mapping. Furthermore, the assembled genome will be beneficial to further research such as gene identification, annotation, and so on, which will certainly Conclusion In this work, we reported a major advance in the identification of large numbers of informative SSR loci in C. chinensis. This is a great attempt to study genomic information and develop SSR markers of C. chinensis without reference genome using Illumina paired-end sequencing technology. Based on the de novo assembly genome, we developed 6,303 SSR markers, and some of which were verified their validity among diverse C. chinensis genotypes and populations. These results fully demonstrate that Illumina paired-end sequencing is a rapid and cost-effective approach for SSR markers development in non-model organism. In this study, 6,303 SSR markers were developed and a total of 500 primer pairs were used for assessment of the assembly quality and validity of markers. Of these, 365 pairs produced PCR amplicons with the expected size and 196 pairs were polymorphic among 8 C. chinensis genotypes. Most of the SSR primers generated high quality amplicons, suggesting that the genome sequencing and assembly were accurate, and SSR makers are valid and applicable. 2. Shinoda K, Yoshida T, Okamoto T (1991) Two wild leguminous host plants of the adzuki bean weevil, Callosobruchus chinensis (L.) (Coleoptera: Bruchidae). Appl Entomol Zool 26: 91–98. 1. Nahdy MS, Ellis RH, Silim SN, Smith J (1998) Field infestation of pigeonpea (Cajanus cajan (L.) Millsp.) by Callosobruchus chinensis (L.) in Uganda. J Stored Prod Res 34: 207–216. Validation of SSR Assay Locus Na He Ho I CCM1 9.0000 0.7455 0.2171 1.5977 CCM2 6.0000 0.0801 0.0597 0.2283 CCM19 4.0000 0.2564 0.2340 0.5288 CCM25 8.0000 0.5384 0.1929 1.0222 CCM27 13.0000 0.4222 0.2587 0.8954 CCM29 4.0000 0.3424 0.3885 0.6624 CCM40 6.0000 0.4502 0.3372 0.9405 CCM46 7.0000 0.6956 0.5059 1.4290 CCM49 7.0000 0.5437 0.5075 0.9823 CCM56 13.0000 0.6890 0.6815 1.5153 CCM65 4.0000 0.5212 0.3583 0.8078 CCM66 6.0000 0.2339 0.0800 0.5297 CCM83 9.0000 0.3247 0.2939 0.6864 CCM108 8.0000 0.5260 0.3360 0.9079 CCM118 9.0000 0.7868 0.3386 1.6831 CCM125 7.0000 0.4158 0.0075 0.7910 CCM149 5.0000 0.0591 0.0361 0.1659 CCM154 6.0000 0.6179 0.3431 1.1863 CCM180 9.0000 0.7292 0.4826 1.4833 CCM189 7.0000 0.5422 0.3034 1.0475 Notes: Number of alleles (Na), expected heterozygosity (He), observed heterozygosity (Ho) and Shannon’s information index (I). doi:10.1371/journal.pone.0095458.t008 Table 8. Informativeness of SSR loci following amplification from 18 geographically diverse Callosobruchus chinensis populations in China. Notes: Number of alleles (Na), expected heterozygosity (He), observed heterozygosity (Ho) and Shannon’s information index (I). doi:10.1371/journal.pone.0095458.t008 Figure 7. Dendrogram for 18 populations of Callosobruchus chinensis in China based on 20 SSR loci (Population code see table 2). doi:10.1371/journal.pone.0095458.g007 Figure 7. Dendrogram for 18 populations of Callosobruchus chinensis in China based on 20 SSR loci (Population code see table 2). doi:10.1371/journal.pone.0095458.g007 Figure 7. Dendrogram for 18 populations of Callosobruchus chinensis in China based on 20 SSR loci (Population code see table 2). doi:10.1371/journal.pone.0095458.g007 May 2014 \| Volume 9 \| Issue 5 \| e95458 PLOS ONE \| www.plosone.org 8 Rapid Development of Microsatellite Markers accelerate the research progress in molecular biology and genetics of adzuki bean weevil and find new measures to control it. of 6,593 potential SSRs were isolated and identified with the most common SSR motifs of trinucleotide, tetranucleotide and dinu- cleotide repeats. The majority of the SSR sequences ranged from 12 to 25 bp in length, which was reasonable and similar to other reports [49], [50]. Given the types of SSR motifs and the large number of SSR loci identified in our study, future SSR marker isolation may be preferentially focused on those composing of trinucleotide repeats which have been demonstrated to be highly polymorphic and stably inherited in the human genome [51–53]. The tri-, tetra-, and dinucleotide repeats mostly contributed to the major proportion of SSRs in this study, while a very small share was contributed by penta- and hexa-nucleotide repeats. Author Contributions Conceived and designed the experiments: C-XD Z-DZ. Performed the experiments: D-DL C-XD. Analyzed the data: C-XD S-LS. Contributed reagents/materials/analysis tools: C-XD Z-DZ X-MW. Wrote the paper: C-XD S-LS. Conceived and designed the experiments: C-XD Z-DZ. Performed the experiments: D-DL C-XD. Analyzed the data: C-XD S-LS. Contributed reagents/materials/analysis tools: C-XD Z-DZ X-MW. Wrote the paper: C-XD S-LS. Conceived and designed the experiments: C-XD Z-DZ. Performed the experiments: D-DL C-XD. Analyzed the data: C-XD S-LS. Contributed reagents/materials/analysis tools: C-XD Z-DZ X-MW. Wrote the paper: C-XD S-LS. 12. Castoe TA, Poole AW, Gu W, Jason de Koning AP, Daza JM, et al. (2010) Rapid identification of thousands of copperhead snake (Agkistrodon contortrix) microsatellite loci from modest amounts of 454 shotgun genome sequence. Mol Ecol Resour 10: 341–347. 11. Dutta S, Kumawat G, Singh BP, Gupta DK, Singh S, et al. (2011) Development of genic-SSR markers by deep transcriptome sequencing in pigeonpea [Cajanus cajan (L.) Millspaugh]. BMC Plant Biology 11: 17. Figure S1 Distribution of sequencing quality (Q20 and Q30 refer to the values of quality of sequencing data). (PDF) Genetic diversity analysis among 18 different geologic popula- tions based on 20 polymorphic SSR loci confirmed validity and usability of polymorphic SSR markers (Table 7). The average number of alleles and Shannon’s information index were 7.35 and 0.9545, respectively among the 20 microsatellite loci indicating that these primers provided abundant polymorphic information in 18 populations and uniform distribution of allele frequencies. Hence the genetic diversity analysis based on these SSR markers is reliable and informative. Gene heterozygosity is considered to be an optimum parameter to measure genetic variation within populations [54]. The average expected heterozygosity was 0.4760, suggesting moderate genetic diversity among the samples, which was likely to be caused by their high adaptability and frequent communication. Based on 20 polymorphic SSR loci, genetic relationship among 18 populations was clearly shown in dendrogram graph in spite that the level of genetic structure variation among populations was modest (Figure 7), indicating relatively high resolution power of SSR makers and applicability in phylogenetics. Figure S2 Distribution of sequencing error rate. (PDF) Figure S2 Distribution of sequencing error rate. (PDF) Figure S3 Distribution of GC content and contig depth. (PDF) Table S1 Decontamination of valid sequencing data. (DOC) 4. Chaubey MK (2008) Fumigant toxicity of essential oils from some common spices against pulse beetle, Callosobruchus chinensis (Coleoptera: Bruchidae). J Oleo Sci 57: 171–179. 3. Mishra D, Shukla AK, Tripathi KK, Singh A, Dixit AK, et al. (2006) Efficacy of application of vegetable seed oils as grain protectant against infestation by Callosobruchus chinensis and its effect on milling fractions and apparent degree of dehusking of legume-pulses. J Oleo Sci 56: 1–7. 5. Talekar NS (1988) Biology, damage and control of bruchid pests of mungbean. In Mungbean: Proceedings of the 2nd International Symposium. Edited by Shanmugasundaram S and McLean BT. AVRDC, Tainan, Taiwan. 329–342. 10. Choudhary S, Sethy NK, Shokeen B, Bhatia S (2009) Development of chickpea EST-SSR markers and analysis of allelic variation across related species. Theor Appl Genet 118: 591–608. 6. Tautz D, Renz M (1984) Simple sequence repeats are ubiquitous repetitive components of eukaryotic genomes. Nucleic Acids Res 12: 4127–4138. Table S1 Decontamination of valid sequencing data. (DOC) Table S2 The primer pairs of Callosobruchus chinensis designed successfully by Primer3. (XLS) 9. Sunnucks P (2000) Efficient genetic markers for population biology. Trends Ecol Evol 15: 199–203. Supporting Information Figure S1 Distribution of sequencing quality (Q20 and Q30 refer to the values of quality of sequencing data). (PDF) Figure S1 Distribution of sequencing quality (Q20 and Q30 refer to the values of quality of sequencing data). (PDF) 8. Ellegren H (2004) Microsatellites: simple sequences with complex evolution. Nat Rev Genet 5: 435–445. 5. Talekar NS (1988) Biology, damage and control of bruchid pests of mungbean. In Mungbean: Proceedings of the 2nd International Symposium. Edited by Shanmugasundaram S and McLean BT. AVRDC, Tainan, Taiwan. 329–342. 6. Tautz D, Renz M (1984) Simple sequence repeats are ubiquitous repetitive components of eukaryotic genomes. Nucleic Acids Res 12: 4127–4138. Rapid Development of Microsatellite Markers Berger MF, Levin JZ, Vijayendran K, Sivachenko A, Adiconis X, et al. (2010) Integrative analysis of the melanoma transcriptome. Genome Res 20: 413–427. y 20. Wang H, Zhang H, Wong YH, Voolstra C, Ravasi T, et al. (2010) Rapid transcriptome and proteome profiling of a non-model marine invertebrate, Bugula neritina. Proteomics 10: 2972–2981. 42. Wang B, Guo G, Wang C, Lin Y, Wang X, et al. (2010) Survey of the transcriptome of Aspergillus oryzae via massively parallel mRNA sequencing. Nucleic Acids Res 38: 5075–5087. g 21. Shahin A, van Gurp T, Peters SA, Visser RG, van Tuyl JM, et al. (2012) SNP markers retrieval for a non-model species: a practical approach. BMC Research Notes 5: 79. 43. Collins LJ, Biggs PJ, Voelckel C, Joly S (2008) An approach to transcriptome analysis of non-model organisms using short-read sequences. Genome informatics International Conference on Genome Informatics 21: 3–14. 22. Castoe TA, Poole AW, de Koning AP, Jones KL, Tomback DF, et al. (2012) Rapid microsatellite identification from Illumina paired-end genomic sequenc- ing in two birds and a snake. PloS One 7: e30953. 44. Wang QQ, Liu F, Chen XS, Ma XJ, Zeng HQ, et al. (2010) Transcriptome profiling of early developing cotton fiber by deep-sequencing reveals significantly differential expression of genes in a fuzzless/lintless mutant. Genomics 96: 369– 376. 23. Feldmeyer B, Wheat CW, Krezdorn N, Rotter B, Pfenninger M (2011) Short read Illumina data for the de novo assembly of a non-model snail species transcriptome (Radix balthica, Basommatophora, Pulmonata), and a comparison of assembler performance. BMC Genomics 12: 317. 45. Zenoni S, Ferrarini A, Giacomelli E, Xumerle L, Fasoli M, et al. (2010) Characterization of transcriptional complexity during berry development in Vitis vinifera using RNA-Seq. Plant Physiol 152: 1787–1795. p 24. Mizrachi E, Hefer CA, Ranik M, Joubert F, Myburg AA (2010) De novo assembled expressed gene catalog of a fast-growing Eucalyptus tree produced by Illumina mRNA-Seq. BMC Genomics 11: 681. 46. Iorizzo M, Senalik DA, Grzebelus D, Bowman M, Cavagnaro PF, et al. (2011) De novo assembly and characterization of the carrot transcriptome reveals novel genes, new markers, and genetic diversity. BMC Genomics 12: 389. q 25. Li D, Deng Z, Qin B, Liu X, Men Z (2012) De novo assembly and characterization of bark transcriptome using Illumina sequencing and development of EST-SSR markers in rubber tree (Hevea brasiliensis Muell. Arg.). BMC Genomics 13: 192. 47. Rapid Development of Microsatellite Markers XuJ J (2007) Molecular detection of geographic populations and interspecies genetic relationships of several important bean weevils (Coleoptera: Bruchidae). Dissertation, Yangzhou University, Yangzhou. 26. Fu N, Wang Q, Shen HL (2013) De novo assembly, gene annotation and marker development using Illumina paired-end transcriptome sequences in celery (Apium graveolens L.). PloS one 8: e57686. 48. Tuda M, Wasano N, Kondo N, Horng SB, Chou LY, et al. (2004) Habitat- related mtDNA polymorphism in the stored-bean pest Callosobruchus chinensis (Coleoptera: Bruchidae). B Entomol Res 94: 75–80. 27. Wang Z, Fang B, Chen J, Zhang X, Luo Z, et al. (2010) De novo assembly and characterization of root transcriptome using Illumina paired-end sequencing and development of cSSR markers in sweet potato (Ipomoea batatas). BMC Genomics 11: 726. 49. An ZW, Zhao YH, Cheng H, Li WG, Huang HS (2009) Development and application of EST-SSR markers in Hevea brasiliensis Muell. Arg Hereditas 31: 311–319. 50. Yang T, Bao SY, Ford R, Jia TJ, Guan JP, et al. (2012) High-throughput novel microsatellite marker of faba bean via next generation sequencing. BMC Genomics 13: 602. 28. Li R, Zhu H, Ruan J, Qian W, Fang X, et al. (2010) De novo assembly of human genomes with massively parallel short read sequencing. Genome Res 20: 265– 272. 51. Edwards A, Civitello A, Hammond HA, Caskey CT (1991) DNA typing and genetic mapping with trimeric and tetrameric tandem repeats. Am J Hum Genet 49: 746–756. 29. Kondo N, Shimada M, Fukatsu T (1999) High prevalence of Wolbachia in the adzuki bean beetle Callosobruchus chinensis (Coleoptera: Bruchidae). Zool Sci 16: 955–962. 52. Gastier JM, Pulido JC, Sunden S, Brody T, Buetow KH, et al. (1995) Survey of trinucleotide repeats in the human genome: assessment of their utility as genetic markers. Hum Mol Genet 4: 1829–1836. 30. Kageyama D, Narita S, Imamura T, Miyanoshita A (2010) Detection and identification of Wolbachia endosymbionts from laboratory stocks of stored- product insect pests and their parasitoids. J Stored Prod Res 46: 13–19. 53. Sheffield VC, Weber JL, Buetow KH, Murray JC, Even DA, et al. (1995) A collection of tri- and tetranucleotide repeat markers used to generate high quality, high resolution human genome-wide linkage maps. Hum Mol Genet 4: 1837–1844. 31. Thiel T, Michalek W, Varshney RK, Graner A (2003) Exploiting EST database for the development and characterization of gen-derived SSR-markers in barley (Hordeum vulgare L.). Theor Appl Genet 106: 411–422. Rapid Development of Microsatellite Markers 13. Ossowski S, Schneeberger K, Clark RM, Lanz C, Warthmann N, et al. (2008) Sequencing of natural strains of Arabidopsis thaliana with short reads. Genome Res 18: 2024–2033. 33. Creste S, Neto AT, Figueira A (2001) Detection of single sequence repeat polymorphism in denaturating polyacrylamide sequencing gels by silver staining. Plant Mol Biol Rep 19: 299–306. 14. Varshney RK, Nayak SN, May GD, Jackson SA (2009) Next-generation sequencing technologies and their implications for crop genetics and breeding. Trends Biotechnol 27: 522–530. p 34. Kimura M, Crow JF (1964) The number of alleles that can be maintained in a finite population. Genetics 49: 725–738. 35. Lewontin RC (1972) The apportionment of human diversity. Evolutionary Biology 6: 381–398. 15. Bai X, Zhang W, Orantes L, Jun TH, Mittapalli O, et al. (2010) Combining next-generation sequencing strategies for rapid molecular resource development from an invasive aphid species, Aphis glycines. PLoS One 5: e11370. 36. Saitou N, Nei M (1987) The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol Biol Evol 4: 406–425. 16. Garg R, Patel RK, Tyagi AK, Jain M (2011) De novo assembly of chickpea transcriptome using short reads for gene discovery and marker identification. DNA Res 18: 53–63. 37. Wang J, Wang W, Li R, Li Y, Tian G, et al. (2008) The diploid genome sequence of an Asian individual. Nature 456: 60–65. 38. Rosenkranz R, Borodina T, Lehrach H, Himmelbauer H (2008) Characterizing the mouse ES cell transcriptome with Illumina sequencing. Genomics 92: 187– 194. 17. Malausa T, Gilles A, Meglecz E, Blanquart H, Duthoy S, et al. (2011) High- throughput microsatellite isolation through 454 GS-FLX Titanium pyrose- quencing of enriched DNA libraries. Mol Ecol Resour 11: 638–644. 39. Mortazavi A, Williams BA, McCue K, Schaeffer L, Wold B (2008) Mapping and quantifying mammalian transcriptomes by RNA-Seq. Nature Methods 5: 621– 628. 18. Natarajan P, Parani M (2011) De novo assembly and transcriptome analysis of five major tissues of Jatropha curcas L. using GS FLX titanium platform of 454 pyrosequencing. BMC Genomics 12: 191. 40. Nagalakshmi U, Wang Z, Waern K, Shou C, Raha D, et al. (2008) The transcriptional landscape of the yeast genome defined by RNA sequencing. Science 320: 1344–1349. py q g 19. Parchman TL, Geist KS, Grahnen JA, Benkman CW, Buerkle CA (2010) Transcriptome sequencing in an ecologically important tree species: assembly, annotation, and marker discovery. BMC Genomics 11: 180. 41. References 7. Lagercrantz U, Ellegren H, Andersson L (1993) The abundance of various polymorphic microsatellite motifs differs between plants and vertebrates. Nucleic Acids Res, 21: 1111–1115. 7. Lagercrantz U, Ellegren H, Andersson L (1993) The abundance of various polymorphic microsatellite motifs differs between plants and vertebrates. Nucleic Acids Res, 21: 1111–1115. 8. Ellegren H (2004) Microsatellites: simple sequences with complex evolution. Nat Rev Genet 5: 435–445. 8. Ellegren H (2004) Microsatellites: simple sequences with complex evolution. Nat Rev Genet 5: 435–445. 9. Sunnucks P (2000) Efficient genetic markers for population biology. Trends Ecol Evol 15: 199–203. 9. Sunnucks P (2000) Efficient genetic markers for population biology. Trends Ecol Evol 15: 199–203. 10. Choudhary S, Sethy NK, Shokeen B, Bhatia S (2009) Development of chickpea EST-SSR markers and analysis of allelic variation across related species. Theor Appl Genet 118: 591–608. 6. Tautz D, Renz M (1984) Simple sequence repeats are ubiquitous repetitive components of eukaryotic genomes. Nucleic Acids Res 12: 4127–4138. May 2014 \| Volume 9 \| Issue 5 \| e95458 9 PLOS ONE \| www.plosone.org Rapid Development of Microsatellite Markers Rapid Development of Microsatellite Markers ( g ) pp 32. Rozen S, Skaletsky H (1999) Primer3 on the WWW for general users and for biologist programmers. Bioinformatics methods and protocols. Humana Press. 365–386. 54. Ott J (1991) Analysis of Human Genetic Linkage (Revised Edition). Johns Hopkins University Press, Baltimore. May 2014 \| Volume 9 \| Issue 5 \| e95458 PLOS ONE \| www.plosone.org 10
W4390601682.txt	https://adabi.ejournaladabiah.org/index.php/adabi/article/download/2/2	en		The Role of the Head of the Education Service Technical Implementation Unit (UPTD) in Improving Employee Performance in Sangir Batang Hari District, South Solok Regency	Adabi	2,023	cc-by-sa	6,266	Adabi : Journal Of Public Administration And Business Volume 10 Number 2 2023 Adabi : Journal Of Public Administration And Business Journal Homepage: https://adabi.ejournaladabiah.org/index.php/adabi/ The Role of the Head of the Education Service Technical Implementation Unit (UPTD) in Improving Employee Performance in Sangir Batang Hari District, South Solok Regency Belli Yatra1, Susiyanti Meilina2 1,2 Public Administration Study Program, STISIP Imam Bonjol, Padang, Indonesia Coressponding: meilinasusi26@gmail.com ARTICLE INFO ABSTRACT Keywords: Role; Leadership; Discipline. This study aims to find out and examine how the role of the Head of the Sangir Batang Hari Education UPTD is whether it has been carried out properly or not. This study uses a descriptive qualitative method which explains in depth the role of the Head of Education UPTD Sangir Batang Hari District, South Solok Regency in improving employee performance. Data collection techniques using interview techniques, observation and documentation. Data analysis techniques, namely by triagulation, data reduction, data presentation and drawing conclusions. The result of this research is that the role of the Head of the Sangir Batang Hari Education UPTD has been very good in carrying out his role in improving employee performance and discipline. Meanwhile, the obstacle faced by the Head of the Education UPTD Sangir Batang Hari Subdistrict in improving employee performance at the Education UPTD itself, namely the most important obstacle was the lack of staff in the Education UPTD which was one of the most influential obstacles in improving performance at the Education UPTD. Furthermore, the constraints that exist in the UPTD itself also affect the human resources (HR) in the UPTD of Education who are not yet disciplined. Factors that influence the performance of UPTD Education staff in Sangir Batang Hari District, namely: Head of UPTD Education in Sangir Batang Hari District, South Solok Regency in order to increase the role that is more motivating for employees to carry out tasks according to the goals to be achieved so that employee performance can be even better for the sake of structure work and goal attainment. Work 1. Introduction Leadership is an inherent aspect of one's character and mentality. Humans possess the innate ability and authority to assume leadership roles, as it aligns with their natural disposition. Leadership is a dynamic process that imbues collaboration with purpose and is generated by the desire to guide others in attaining objectives (Greenleaf, 2002; Lambert et al., 2016). Leadership, in essence, refers to the capacity and expertise of an individual holding a position as a leader within a work unit to effectively influence others, particularly their subordinates, to think and behave in a manner that contributes significantly to the attainment of organizational objectives through positive actions. Leadership has a crucial function in an organization or agency, significantly impacting the level of work performance (Chiok Foong Loke, 2001). The growth and failures of an agency are heavily influenced by the leadership style of its CEO. This demonstrates that leadership plays a crucial role in attaining work efficiency. If a leader can effectively deploy suitable leadership strategies Adabi : Journal Of Public Administration And Business to the prevailing circumstances and conditions, people will be able to operate in a comfortable and highly motivated manner (Amanchukwu et al., 2015; Cameron, 2012). The efficacy and superiority of an organization are heavily contingent upon the calibre of its human capital (Chen & Wu, 2020; Taylor et al., 2015). In theory, the presence of high-calibre Human Resources (HR) within an organization is anticipated to enhance organizational performance. By assuming the appropriate leadership role, one can establish a work climate that is favourable and conducive to productivity. The job of a leader is crucial in every government agency or organization since they serve as the organization's intellectual centre. The leader consistently makes goaloriented decisions in alignment with the organization's objectives. To accomplish goals effectively and efficiently, it is necessary to have constructive teamwork inside an agency (Fiscella et al., 2017). In this scenario, employees and leaders must possess the ability to engage in effective collaboration to attain the desired objectives. Ki Hadjar Dewantara's conceptualization of leadership encompasses three philosophies that comprehensively address the multifaceted aspects essential to effective leadership. Hence, it might be referred to as plenary leadership or comprehensive leadership, encompassing all its facets. Ki Hadjar Dewantara's distinctive Indonesian leadership concept does not discriminate against individuals based on their status, but rather on their assigned responsibilities. The duties can vary, encompassing positions at the forefront, in the middle, and at the rear. Put simply, a leader must alternate between taking a leading position, a central position, and a supporting position at different times. Effective leadership is crucial for the advancement of an organization, as it plays a fundamental role in attaining organizational objectives. Leadership roles refer to the established standards and behaviours that an individual employs to exert influence on others (Newman et al., 2017). Leadership involves exerting influence over individuals to establish and achieve organizational objectives, while also motivating followers to act by these goals. To 14 exert influence and enhance the dynamics and ethos of the collective. An organization will be unable to realize its full potential and achieve success without good leadership. Leadership should be viewed as a chance to bring advantages to multiple parties, rather than being seen as a commodity or possession to exploit others for personal gain. A leader must possess a clear and perceptive understanding of the factors of success to effectively achieve organizational goals (Bryson, 2018). Leadership is the act of exerting influence over a group of individuals to accomplish a specific objective. Transformational leadership can be defined as a systematic approach to inspiring and motivating individuals to willingly undergo personal change and growth. This approach entails addressing their motivations, providing their needs, and showing respect for subordinates. In leadership, influence is the fundamental aspect that pertains to an individual's capacity to effectively modify the attitudes and behaviours of individuals or groups in a targeted manner. An effective leader must possess both authority and the ability to analyze the reciprocal influence processes that take place between the leader and their followers. Bass (2003) defines transformational leadership as the ability of a leader to exert influence over subordinates in specific ways. Through the implementation of transformational leadership, subordinates will experience a sense of trust, appreciation, loyalty, and respect for their leaders. Ultimately, subordinates will be incentivized to exceed expectations. Conversely, to ascertain the opinions of employees regarding the agency, a leader inside the organization must engage in proactive communication with their subordinates. The leader's demeanour will shape the team's progress inside the agency organization, impacting the accomplishments made and eventually affecting staff productivity. The proficiency of a leader in fostering team cohesion is crucial for the achievement of an agency's objectives. An employee's execution of an agency's obligations and tasks necessitates strong encouragement and inspiration from their supervisor. This is because nearly every action and policy implemented by leaders yields a beneficial influence on the individuals they supervise. An individual's motivation is 15 Belli Yatra, Susiyanti Meilina contingent upon the intensity of their underlying drive. Kartono (2018) asserts that a leader must possess the ability to offer effective incentives to their followers. regulations and policies at the Subdistrict level. The main goal is to improve the quality of the educational system, starting from the Kindergarten (TK) level. Given the crucial importance of leaders in motivating their subordinates, experts have proposed the idea of explicitly delineating the specific duties and responsibilities that leaders bear in influencing their subordinates. A role can be defined as the prescribed and anticipated conduct expected from an individual occupying a specific position. Based on the provided information, it can be inferred that the leadership role encompasses a collection of behaviours that are anticipated to be fulfilled by an individual in their capacity as a leader. Thoha (2017) define establishing a role as a sequence of consistent behaviours that result from occupying a specific position or the presence of something easily identifiable. The position's objective is to establish a framework of social values that governs the interactions between the individual performing the function and the individuals associated with or affected by that role. The Sangir Batang Hari Subdistrict Education UPT has a kindergarten program for children aged 4-7. There are a total of 12 kindergarten schools, consisting of 1 state kindergarten and 11 private kindergartens. The program has a total of 45 students and 11 teaching staff. Additionally, there are 6 playgroup (KB) schools with a total of 44 kids and 7 instructors. In Sangir Batang Hari District, there are a total of 20 public primary schools, accommodating 421 pupils and employing 238 instructors and staff members. Currently, there are a total of six Junior High Schools (SMP) with a student population of 468 and a teaching staff of 63. UPTD Education aims to establish a performance framework that can facilitate efficient and productive job tasks. Examining the current position of the head of the UPTD in Sangir Batang Hari District is intriguing for the author, as it pertains to enhancing the performance of their personnel. Based on observations, it is evident that the Sangir Batang Hari District Education UPTD office now has a shortage of staff. Consequently, it is worth considering whether the Head of the Education UPTD in Sangir Batang District, South Solok Regency, has an impact on this staff shortage at the UPTD. Employees in an educational institution have a crucial role as they are the driving force behind the successful execution and attainment of organizational objectives. Essentially, individuals possess varying degrees of proficiency in executing their assigned tasks and obligations. Performance refers to the work outcomes that individuals or groups can achieve within an organization, in line with their assigned authority and responsibilities. The aim is to legally and ethically accomplish the organization's goals, while adhering to moral and ethical standards. An Education Service Technical Implementation Unit (UPTD) is an institution that carries out government policies and acts as an extension of the Regency or City Education Office. Its main role is to implement regulations and policies in education at the subdistrict level. The UPTD possesses a capability to exert an impact on the performance of its employees in delivering educational services to schools and stakeholders. The Sangir Batang Hari Subdistrict Education UPTD in South Solok Regency is responsible for supporting the Regent in the implementation of educational government affairs. This includes assisting the Regency and City in implementing education 2. Literature Review Role In the comprehensive Indonesian dictionary, the term "role" refers to a collection of anticipated behaviors that individuals are required to possess while residing inside a particular culture. A role is a conceptual framework that defines the actions and responsibilities that individuals can undertake within a community or organization (Cohen, 2016; Langham et al., 2015). Roles can be defined as individual behaviors that play a crucial role in shaping the social structure of society (Soekanto, 2012). Each job seeks to establish a relationship between the person performing the role and the individuals connected to or affected by the role. This relationship is governed by mutually 16 Adabi : Journal Of Public Administration And Business recognized and followed societal ideals (Lin, 2017). role that needs to be fulfilled is fundamentally unchanged (Sweet et al., 2015). The role is a fluid component of one's position or status. This implies that an individual has fulfilled their rights and responsibilities in alignment with their position, so fulfilling a function. The two entities are inseparable due to their interdependence, signifying that there is no role without status and no status without role (Kitayama et al., 2022). Similar to one's position, individuals might assume multiple roles derived from their social tendencies. Consequently, this function dictates both his contributions to society and the opportunities society affords him (Sanford, 2017). Berry (Berry, 2016) defines roles as the expectations that are placed on individuals who hold specific social positions. These expectations are a reflection of social norms, indicating that the role is influenced by societal norms. Within this function, there are two specific expectations: the expectations that the role holder has towards society and the expectations that the role holder has towards themselves or their responsibilities. Based on this viewpoint, it may be inferred that a role refers to the conduct exhibited by an individual as a result of their responsibilities stemming from their position or occupation (Peng & Pierce, 2015). To summarise, individuals in higher positions within an organizational hierarchy necessitate fewer technical skills. According to Siswanto (2021), the importance of technical abilities increases as a person's position in an organizational hierarchy decreases. In his book "Sociology: An Introduction", Soekanto (2012) defines the role as a dynamic component of position or status. Engaging in the exercise of one's rights and fulfilling one's responsibilities by their designated position is the performance of a function. The differentiation between rank and role serves the purpose of advancing scientific knowledge. The two are inseparable due to their mutual interdependence. Every role is associated with a position, and every position is associated with a role. Similar to position, the term "role" also has two distinct definitions. Each individual possesses multiple roles derived from their social tendencies. Consequently, the job not only dictates an individual's contributions to society but also influences the opportunities bestowed upon them by society (Caza et al., 2018). A role is a collection of expected behaviors that individuals in society are anticipated to possess. Meanwhile, a role refers to the specific activity performed by an individual during an event (Kumar et al., 2022). A role is a set of consistent behaviors that result from holding a certain position or occupying a recognized office. An individual's personality can significantly impact the execution of a role (Bizzi, 2017). Roles emerge when individuals recognize that they are not operating in isolation. Being in a setting that requires constant interaction. The environment is vast and diverse, with each individual experiencing a unique experience. However, the Based on this perspective, it may be inferred that a role refers to the conduct exhibited by an individual as a result of their responsibilities stemming from their position or occupation. Rivai (2009) defines role as the regulated and expected behavior of an individual in a specific position. Moreover, as stated by Oc (2018), a role refers to the specific position that assumes leadership, particularly in the context of a certain situation or event. To summarise, individuals in higher positions within an organizational hierarchy necessitate fewer technical skills. According to Siswanto (2021), the technical abilities necessary become more crucial as a person's position in an organizational structure decreases. The role of the Head of Education UPTD The Head of UPTD assumes the responsibility of issuing warnings and acts as a leader by giving directions, overseeing operations, and providing warnings to enhance staff performance. Before implementing penalties, the Head of UPTD issues warnings, particularly to staff who have committed violations, to ensure effective and efficient work in the office. UPTD Education is an educational institution that carries out the policies of the local government in the field of education. It acts as a branch of the Education Office at the district 17 level, responsible for executing regulations and policies in education. The significance of roles lies in their ability to govern an individual's conduct. The influence of an individual can somewhat forecast the behaviors of others. The individual in question will possess the ability to adapt their conduct to align with the conduct of the individuals within their group. The social relationships within society are the interactions between individuals fulfilling their respective tasks in society. Roles are governed by relevant rules. The leadership role can be defined as the expected behaviors that a person in a desired leadership position is required to exhibit (Afsar et al., 2017). Organizations establish both the tasks to be performed and the expected behavior associated with those tasks, which highlights the significance of role expectations in governing subordinate conduct (Wang et al., 2017). The leader possesses the ability to exert influence over their subordinates, hence optimizing the work process to effectively accomplish the predetermined organizational objectives. Role theory is an interdisciplinary theory that combines numerous theories, orientations, and scientific disciplines in the fields of sociology, psychology, and anthropology. Role theory examines the concept of "role," frequently employed in the realm of theatre. In this context, an actor assumes the role of a certain character and is anticipated to exhibit behavior consistent with that character's position. An actor's role in the theatre can be compared to an individual's role in society, as they share certain characteristics. A role is the portrayal of a character by an actor in a theatrical performance. In a social context, a role refers to the specific function that a person fulfills when occupying a position in the social structure. An actor's role is a constraint imposed by another actor who is also performing in the same role. Based on the aforementioned explanation, it can be inferred that role theory is a theoretical framework that examines an individual's position and behavior, which are contingent upon the expectations placed upon them by those who are connected to that individual or actor. Actors possess an understanding of the social hierarchy they belong to hence they strive to consistently project Belli Yatra, Susiyanti Meilina an image of being "competent" and are regarded by their peers as "conforming" to the societal norms and expectations. Employee Performance Performance refers to the degree of accomplishment in executing specific duties. In the realm of human resource development, the performance of an employee is crucial for both the individual's work performance and the overall success of the firm. The term "performance" refers to the actual achievements and work results of an individual, specifically about their job responsibilities. It encompasses both the quality and quantity of the work produced by an employee (Syamsir & Saputra, 2022). Performance refers to the extent to which people efficiently and successfully meet job objectives (Mulia, 2021). Employee performance refers to the extent to which an employee's job achievements align with the predetermined work standards set by the organization. Subsequently, According to Robbins (2007), performance is the outcome that employees achieve in their work based on certain criteria that are relevant to their employment. Rivai (2009) defines performance as the willingness and ability of an individual or a group to effectively carry out their assigned tasks and achieve the desired outcomes. According to the aforementioned definition, performance is a strategic concept that aims to foster a collaborative relationship between management and employees to achieve favorable outcomes. The primary factor influencing performance is human resources. Even if planning is meticulously organized, the effectiveness of performance hinges on the individuals or personnel involved. If the execution lacks good quality and lacks sufficient work passion, the created plans will be rendered futile. 3. Methods This study uses qualitative research methodology. According to Bogdan and Taylor in Moleong (2016), qualitative approaches refer to research techniques that generate descriptive data in the form of written words uttered by individuals and observable behavior. 18 Adabi : Journal Of Public Administration And Business Sugiyono (2018) defines qualitative research methods as a research approach rooted in the postpositivist philosophy. It is employed to investigate the characteristics of natural phenomena, rather than conducting experiments. In this method, the researcher plays a central role as the primary instrument, employing triangulated data collection techniques. The analysis of data is conducted inductively and qualitatively, leading to the generation of results. Qualitative research prioritizes the interpretation of meaning over the pursuit of generalizable findings. Qualitative research descriptively presents data through reports and descriptions (Arikunto, 2018). Qualitative research is a type of research that aims to comprehensively uncover symptoms within their context. This is done through collecting data in natural settings, with the researchers themselves acting as a significant tool in the field. In the context of qualitative research, no numerical data was subjected to statistical analysis. However, researchers collected data using numerous methods including interviews, documentation, and direct observation. The research methodology employed in this work is a descriptive investigation. This study is part of field research, where research ideas were derived from observations and findings of previous studies regarding the role of the Head of the Education Service Technical Implementation Unit (UPTD) in enhancing employee performance in Sangir Batang Hari District, South Solok Regency. The research locus refers to the specific research location situated in the Education Service Unit (UPTD) in Sangir Batang Hari District, South Solok Regency. The author purposefully chose to do research at this specific site due to their employment there, which facilitated convenient access to the necessary information and data required for formulating this proposal. The primary objective of this research is to investigate the impact of the Head of the Education Service Technical Implementation Unit (UPTD) on enhancing staff performance in Sangir Batang Hari District, South Solok Regency. In qualitative research, the analysis process is conducted concurrently with the data collection procedure. The used model is a complex and dynamic analysis model that is implemented once the data has been acquired. Subsequently, the data is analyzed and conveyed through descriptions and explanations, enabling the derivation of conclusions and recommendations aligned with the research objectives. The research employs the Milles and Huberman (2018) analysis model for descriptive data analysis. This model involves an interactive approach with distinct stages, including data collection from diverse sources such as interviews and observations. Recorded in field notes. Data reduction involves the process of minimizing the amount of data that has been collected, examined, and analyzed. Data reduction is a method of analysis that enhances, classifies, and guides while eliminating unnecessary information and then arranging the data. Data presentation facilitates comprehension of past events and informs future analysis and decision-making processes. Conclusion: Data inference is performed once the data has been provided in a descriptive manner and with a comprehension of logical interpretation. 4. Result The Role of the Head of the Education Service Technical Implementation Unit (UPTD) in Improving Employee Performance in Sangir Batang Hari District, South Solok Regency Field research findings indicate that UPTD is an institution responsible for implementing district/city government policies in the field of education. It serves as an extension of the district or city education office, ensuring the implementation of regulations and policies in education at the sub-district level. The Head of the Sangir Batang Hari District Education UPTD exemplified the importance of punctuality and avoiding work delays to the employees. Additionally, the Head emphasized the significance of cooperation in achieving efficient performance. Moreover, the Head occasionally issued direct warnings to employees who violated performance standards. These actions were taken to establish organizational goals at the Sangir Batang Hari District Education UPTD. 19 The leader of the Sangir Batang Hari District Education UPTD fulfills his responsibilities by motivating and monitoring employees, as well as providing effective guidance. The Head of UPTD also has ethics in leading employees, because the leader's ethical behavior towards employees also greatly influences their performance. If the leader violates these ethics or displays inappropriate behavior towards employees, it can have the greatest influence on the work and the results of achieving the goals of the Sangir Batang Hari District education UPTD itself. Obstacles for the head of the Education UPTD in improving employee performance in Sangir Batang Hari District, South Solok Regency. The research findings indicate that the Sangir Batang Hari District Education UPTD faces a significant impediment, namely a shortage of workers. This obstacle has a substantial impact on the improvement of performance at the Education UPTD. Moreover, the author's interviews with the head of the UPTD, Sangir Batang Hari District, reveal that the limitations inside the UPTD have a direct impact on human resources (HR). The UPTD's leader issues warnings, particularly to staff who engage in infractions, prior to implementing fines to ensure the office's efficacy and efficiency. Optimal staff performance is a crucial determinant in an agency's endeavor to enhance productivity. The performance of an employee is an issue that pertains to the individual, as each employee have a certain level of proficiency in executing their responsibilities. 5. Discussion Research findings indicate that employees are a crucial asset for a company since their performance significantly impacts the company's success in fulfilling its vision and goal. Leaders have a crucial role in enhancing staff performance. The Head of the Sangir Batang Hari District Education UPTD implemented this action towards his employees. The mindset of employees will be positively influenced by effective leadership, leading to increased enthusiasm and improved performance in executing assigned duties under Belli Yatra, Susiyanti Meilina their superiors. To effectively lead an organization or company, leaders must demonstrate key leadership qualities. By exemplifying these characteristics, leaders can provide a clear roadmap for their subordinates, enabling them to fulfill their duties and responsibilities effectively. This, in turn, fosters an environment where employees are motivated to enhance their performance. Therefore, exhibiting such behavior not only shapes the role of a leader in enhancing employee performance but also serves as a source of motivation for employees to strive for excellence. The function of the Head of UPTD encompasses delivering instructions, supervision, and warnings to enhance employee performance. The UPTD Chief issues warnings, particularly to staff who commit violations, before implementing penalties to enhance the efficacy and efficiency of office operations. Following the instructions and supervision of the Head of the Sangir Batang Hari District UPTD, the employees demonstrated a significant improvement in their performance. They gained a better understanding of how to handle tasks and effectively address workrelated challenges. The direction and supervision I provide to these individuals enhance their ability to perform efficiently and contribute to the attainment of their job responsibilities and objectives. The leader of UPTD must also demonstrate ethical conduct in managing staff, as the leader's ethical behavior significantly impacts their performance. When the leader breaches these ethical standards or exhibits unethical conduct towards employees, it can exert a significant impact on the team's work and the attainment of its objectives. The UPTD education is located in Sangir Batang Hari District, under the South Solok Regency. Employee performance refers to the outcomes and productivity achieved by an employee in their assigned responsibilities within the organization during a specific timeframe. Optimal staff performance is a crucial determinant in an agency's endeavor to enhance productivity. The performance of an employee is an individualized concern, as each person Adabi : Journal Of Public Administration And Business possesses a distinct level of proficiency in executing their responsibilities. Enhancing employee performance at UPTD involves a structured and methodical approach whereby employee performance is elevated according to predetermined organizational standards in order to achieve preplanned objectives. Performance improvement is a metric used to evaluate the extent to which a person has met or exceeded the expectations and goals associated with their assigned work or position. Enhancing staff performance at UPTD can also serve as a means of holding them accountable for their actions. A company's achievement of a measurable outcome over a predetermined time period is the performance of its employees at UPTD. The job of UPTD employees directly affects how well they perform because it closely aligns with the strategic goals of an institution or organization. At the Education UPTD in Sangir Batang Hari District, South Solok Regency, there is a scarcity of personnel to fulfill the designated primary responsibilities of the job. Leaders must prioritize several factors while enhancing employee discipline, specifically: (1) facilitating the development of employees' behavioral patterns; (2) assisting employees in elevating their standards of behavior; and (3) utilizing rule enforcement as a strategic instrument. Leaders facilitate change by establishing a clear direction through vision and strategy, bringing together individuals who must collaborate to accomplish that vision, and encouraging and inspiring them to overcome challenges, therefore resulting in change. Work discipline in an institution refers to the set of rules, regulations, and practices that all employees must adhere to when doing their jobs. An exemplary employee demonstrates the ability to enforce work discipline inside an organization, such as an office, to ensure the seamless execution of personnel activities. Indeed, the implementation of employee discipline will enhance the organization's level of sophistication. Employee work is associated with satisfaction motivation theory, as indicated by (Umar and Husein 2010), which emphasizes the importance of corporations addressing employee 20 rights, including compensation, career planning, workplace safety and health, as well as training and development. This occurs due to the constrained workforce, which occasionally hinders productivity. It is highly regrettable for the head of education at UPTD. These limitations can sometimes hinder employees from exerting maximum effort and performing their duties with excellence. However, due to the awareness of the shortage of human resources in UPTD Education, employees are motivated to maintain a high level of performance and adhere to the established procedures in order to achieve their goals. In order to enhance performance, mere chats with colleagues are insufficient. At UPTD Education, staff actively engage in education and training programs, particularly resource development training, which is frequently conducted in the Sangir Hari area of South Solok Regency. Humans offer services to the community, including modest training in technological disciplines such as computer proficiency and other equipment. The research findings indicate that the Education UPTD in Sangir Batang Hari District, South Solok Regency, faces a significant obstacle in the form of insufficient staff. This obstacle has a substantial impact on the overall performance improvement of the Education UPTD. Moreover, the outcomes of the author's interview with the UPTD's director indicate that the challenges present within the UPTD have a direct impact on its human resources. Loose restrictions refer to limitations on resources that are not fully exploited. In the Education UPTD of Sangir Batang Hari District, South Solok Regency, a scarcity of staff, if not effectively utilized, will lead to a significant accumulation of tasks. The Education UPTD in Sangir Batang Hari District, South Solok Regency, faces significant difficulties that require immediate attention in order to implement employee development initiatives effectively. In order to accomplish the leadership duty of enhancing employee performance, it is essential to establish an efficient administration capable of executing the assigned tasks. A problem encountered by 21 the head of the Education UPTD is the insufficient number of personnel inside the organization. Consequently, in order to fulfill the necessary tasks, they have reached the conclusion that they should personally do the work. The head of the Education UPTD encounters a challenge in recruiting new personnel due to the arduous nature of the recruitment process. This difficulty arises primarily from the lack of interest in joining the Education UPTD, which in turn leads to the issue of inadequate remuneration for employees in relation to their job responsibilities. Without staff wages, there will be a decrease in motivation to join the Education UPTD. Interviews with leaders and employees indicate that employing a democratic and firm leadership style can enhance employees' capabilities. The primary objective is to oversee and assess the extent to which employees adhere to their responsibilities. This element also contributes to the enhancement of employee performance in Sangir Hari District Education, South Solok Regency. In order to achieve quality and professional performance, it is crucial for the head of UPTD to address the difficulties experienced by employees. To do so, it is necessary to have a clear understanding of the overall and specialized performance goals. The primary objective of performance management is to cultivate a culture in which individuals and teams take ownership of ongoing improvement in employee performance and the development of their own talents and contributions. In addition to that, the Head of UPTD endeavors to enhance staff performance by addressing the challenges they encounter. As a platform for enhancing employee knowledge and abilities in the workplace, particularly where UPTD employee performance is deficient. Consequently, the leadership will adopt a more stringent approach towards employees who violate regulations, including the termination of those who do so, in order to foster a culture of enhanced discipline in the workplace. Belli Yatra, Susiyanti Meilina 6. Conclusion Through the analysis of data and discussions presented in this thesis, the researcher can conclude that leadership plays a crucial role in enhancing the performance of UPTD employees in Sangir Batang Hari District, South Solok Regency. Specifically, the Head of the Education Service Technical Implementation Unit (UPTD) has a significant impact on improving employee performance. The director of the Education UPTD in Sangir Batang Hari District, South Solok Regency, makes sure that leaders who successfully carry out their primary duties and functions, exhibit moral leadership, and possess the necessary leadership traits have an impact on employee performance. The employees at the Education UPTD in Sangir Batang Hari District, South Solok Regency, demonstrate commendable performance by adhering to personnel standards, diligently executing assigned tasks, and striving to meet targets and achieve the objectives of the Education UPTD in Sangir Batang Hari District, South Solok Regency. The head of the Education UPTD in Sangir Batang Hari District, South Solok Regency, has the most difficulty because there aren't enough employees. Despite efforts to enhance employee performance to the fullest extent, the limited human resources available in the Education UPTD hinder the smooth operation of employee performance. References Afsar, B., Badir, Y. F., Saeed, B. Bin, & Hafeez, S. (2017). Transformational and transactional leadership and employee’s entrepreneurial behavior in knowledge–intensive industries. The International Journal of Human Resource Management, 28(2), 307– 332. Amanchukwu, R. N., Stanley, G. J., & Ololube, N. P. (2015). A review of leadership theories, principles and styles and their relevance to educational management. In Management. ed.fullerton.edu. https://ed.fullerton.edu/lift/_resources/pdfs /just_leadership/A REVIEW OF LEADERSHIP.pdf Arikunto, S. (2018). Prosedur Penelitian Suatu Pendekatan Praktek Cetakan I. Remaja 22 Adabi : Journal Of Public Administration And Business Rosdakarya. Bandung. (2022). Varieties of interdependence and the emergence of the Modern West: Toward the globalizing of psychology. American Psychologist, 77(9), 991. Berry, D. (2016). Journalism, ethics and society. Routledge. Bizzi, L. (2017). Network characteristics: When an individual’s job crafting depends on the jobs of others. Human Relations, 70(4), 436– 460. Kumar, S., Bhatt, R., & Ganguly, D. G. (2022). Organizational behaviour. Academic Guru Publishing House. Bryson, J. M. (2018). Strategic planning for public and nonprofit organizations: A guide to strengthening and sustaining organizational achievement. books.google.com. Lambert, L., Zimmerman, D. P., & Gardner, M. E. (2016). Liberating leadership capacity: Pathways to educational wisdom. Teachers College Press. Cameron, K. (2012). Positive leadership: Strategies for extraordinary performance. BerrettKoehler Publishers. Langham, E., Thorne, H., Browne, M., Donaldson, P., Rose, J., & Rockloff, M. (2015). Understanding gambling related harm: A proposed definition, conceptual framework, and taxonomy of harms. BMC Public Health, 16, 1–23. Caza, B. B., Vough, H., & Puranik, H. (2018). Identity work in organizations and occupations: Definitions, theories, and pathways forward. Journal of Organizational Behavior, 39(7), 889–910. Chen, T.-C., & Wu, Y. J. (2020). The influence of R&D intensity on financial performance: The mediating role of human capital in the semiconductor industry in Taiwan. Sustainability, 12(12), 5128. Chiok Foong Loke, J. (2001). Leadership behaviours: effects on job satisfaction, productivity and organizational commitment. Journal of Nursing Management, 9(4), 191–204. Cohen, A. (2016). Are they among us? A conceptual framework of the relationship between the dark triad personality and counterproductive work behaviors (CWBs). Human Resource Management Review, 26(1), 69–85. Fiscella, K., Mauksch, L., Bodenheimer, T., & Salas, E. (2017). Improving care Teams’ functioning: recommendations from team science. The Joint Commission Journal on Quality and Patient Safety, 43(7), 361–368. Greenleaf, R. K. (2002). Servant leadership: A journey into the nature of legitimate power and greatness. Paulist Press. Kartono, K. (2018). Pemimpin dan kepemimpinan. Rajawali Pers. Kitayama, S., Salvador, C. E., Nanakdewa, K., Rossmaier, A., San Martin, A., & Savani, K. Lin, N. (2017). Building a network theory of social capital. Social Capital, 3–28. Miles, M. B., Huberman, A. M., & Saldaña, J. (2018). Qualitative data analysis: A methods sourcebook. Sage publications. Moleong, L. J. (2016). Metodologi penelitian kualitatif. Remaja Rosdakarya. Mulia, R. A. (2021). Manajemen Sumber Daya Manusia: Teori dan Aplikasi Dalam Peningkatan Kinerja. CV. Eureka Media Aksara. Newman, A., Schwarz, G., Cooper, B., & Sendjaya, S. (2017). How servant leadership influences organizational citizenship behavior: The roles of LMX, empowerment, and proactive personality. Journal of Business Ethics, 145(1), 49–62. Oc, B. (2018). Contextual leadership: A systematic review of how contextual factors shape leadership and its outcomes. The Leadership Quarterly, 29(1), 218–235. Peng, H., & Pierce, J. (2015). Job-and organization-based psychological ownership: Relationship and outcomes. Journal of Managerial Psychology, 30(2), 151– 168. Rivai, V. (2009). Human resource management for companies: from theory to practice. Rajawali Pers. https://books.google.co.id/books?id=Rb7qZ 23 Belli Yatra, Susiyanti Meilina wEACAAJ Robbins, S. P., & Judge, T. (2007). Organizational Behavior. Pearson/Prentice Hall. Sanford, N. (2017). Self and society: Social change and individual development. Routledge. Siswanto, B. (2021). Pengantar books.google.com. manajemen. Soekanto, S. (2012). Sosiologi Suatu Pengantar (cetakan ke-44). In Jakarta: PT Raja Grafindo Persada. Sugiyono. (2018). Metode penelitian kuantitatif, kualitatif, dan R&D (edisi 2). Alfabeta. Sweet, J. J., Benson, L. M., Nelson, N. W., & Moberg, P. J. (2015). The American Academy of Clinical Neuropsychology, National Academy of Neuropsychology, and Society for Clinical Neuropsychology (APA Division 40) 2015 TCN professional practice and ‘salary survey’: Professional practices, beliefs, and incomes of US neuropsychologists. The Clinical Neuropsychologist, 29(8), 1069–1162. Syamsir, S., & Saputra, N. (2022). Administrasi Kepegawaian. CV. Eureka Media Aksara. Taylor, T., Doherty, A., & McGraw, P. (2015). Managing people in sport organizations: A strategic human resource management perspective. Routledge. Thoha, M. (2017). Contemporary administration. Kencana. public Wang, H., Feng, J., Prevellie, P., & Wu, K. (2017). Why do I contribute when I am an “insider”? A moderated mediation approach to perceived insider status and employee’s innovative behavior. Journal of Organizational Change Management, 30(7), 1184–1197.
https://openalex.org/W2790977446	https://europepmc.org/articles/pmc5849608?pdf=render	English	null	Size Exclusion Chromatography Method for Purification of Nicotinamide Mononucleotide (NMN) from Bacterial Cells	Scientific reports	2,018	cc-by	9,034	George Cătălin Marinescu 1,2, Roua-Gabriela Popescu1,2 & Anca Dinischiotu1 Over 12% of the world’s health resources are spent on treating diabetes, as high blood glucose is the third cause of mortality worldwide. Insulin resistance is the basis of the most common form of diabetes: type 2 diabetes. Recent animal studies report successful attempts at reversing type 2 diabetes by the administering of the NAD+ precursor nicotinamide mononucleotide (NMN). However, the current high price of this molecule urges for more efficient and cost-effective production methods. This work proposes a method for purifying NMN by Size Exclusion Chromatography (SEC) on silica with a covalently attached coating of poly(2-hydroxyethyl aspartamide) (PolyHEA) stationary phase using an isocratic elution with a denaturing mobile phase (50 mM formic acid) from a complex molecular mixture such as a fermentation broth. The eluted peaks were identified by UV-Vis analysis and confirmed with ESI+ mass spectrometry and a HPLC reversed-phase method. The proposed SEC method is simple, patent-free, directly applicable for industrial production with a minimum scale up effort. The need for multiple chromatographic steps is eliminated and the lysate filtration and clarification steps are simplified. Substantial reduction in NMN production costs and increased purity of NMN to the level suitable for usage in humans are expected. High blood glucose level is the third highest risk factor for premature mortality worldwide. The more common condition, type 2 diabetes, has increased with modern life, urbanisation, reduced physical activity and an ageing population. It affects the blood vessels, the heart, kidneys and nerves, eventually resulting in disability and pre- mature death1,2. Previously, an association between insulin resistance and type 2 diabetes has been recognised3. Moreover, abnormal mitochondrial function was correlated with type 2 diabetes4–8. In this context, gene expression and proteomics studies revealed a correlation between insulin resistance and a down-regulation of protein complexes involved in mitochondrial oxidative phosphorylation9–11. Lower expres- sion of mitochondrial complexes I-IV influences the redox state of the cell, diminishing ATP production and antioxidant capacity. In all the organs of 12 months old Wistar rats, an increase of p53 acetylation, a decreased activity of nicotinamide adenine dinucleotide (NAD+) dependent sirtuin-1 (SIRT1) and a mild over-expression of SIRT1 was correlated with diminished NAD+ level and NAD+/NADH ratio12. Although NAD+ is recycled by the salvage pathways13–15, it is consumed by poly(ADP-ribose) polymerases (PARPs) and glycohydrolases (CD38 and CD157)16 and therefore therapeutic interventions to restore NAD+ level are crucial. www.nature.com/scientificreports www.nature.com/scientificreports www.nature.com/scientificreports 1Department of Biochemistry and Molecular Biology, University of Bucharest, Bucharest, 050095, Romania. 2Independent Research Association, Bucharest, 012416, Romania. Correspondence and requests for materials should be addressed to G.C.M. (email: Catalin.Marinescu@independent-research.ro) Received: 24 July 2017 Accepted: 1 March 2018 Published: xx xx xxxx Size Exclusion Chromatography Method for Purification of Nicotinamide Mononucleotide (NMN) from Bacterial Cells Received: 24 July 2017 Accepted: 1 March 2018 Published: xx xx xxxx Received: 24 July 2017 Accepted: 1 March 2018 Published: xx xx xxxx George Cătălin Marinescu 1,2, Roua-Gabriela Popescu1,2 & Anca Dinischiotu1 Results Si l t Single step NMN chromatographic separation by SEC-HPLC. The SEC-HPLC protocol was per- formed as explained in the Methods section employing a clarified bacterial lysate sample and standard NMN (Fig. 1A,C and E). Analysing the Photodiode Array (PDA) data led to the identification of the retention time (RT) for NMN as 6 minutes. A clear separation of NMN from other molecules of lysate is shown (Fig. 1B and D). The most abundant contaminant in the lysate had a retention time between 3.5 and 4 min (Fig. 1F) which corresponds to the impurity peak identified in the NMN standard at the same RT (Fig. 1E).hi p y pi g The chromatogram obtained at 260 nm was extracted from the PDA data and used for quantification by peak area integration (Fig. 1E).i For accurate NMN quantification, the calibration method described in the Methods section was used. A sec- ond order polynomial equation was obtained with the coefficient of correlation of 0.99 (Fig. 1G), covering sam- ples with NMN concentrations from 0.03 to 3.34 g/L. Mass spectrometry identification of NMN. Although UV-Vis light absorption data recorded by the HPLC PDA detector provide spectrum information for NMN identification, to eliminate any doubt related to the correct peak identification, a mass detector was used. The Total Ion Current (TIC) chromatogram (Fig. 2A) resulting from running the protocol described in section Size Exclusion Chromatography High Performance Liquid Chromatography (SEC-HPLC) (NMN standard sample, SEC protocol on PolyHEA column) shows a peak at 6 minutes, thereby correlating with the HPLC data (Fig. 2B). The MS spectrum of the NMN separated from the bacterial lysate (Fig. 2H) is similar to the spectrum of the NMN standard (Fig. 2G). The specific m/z values (shown on both standard and lysate separations) are: 335 for NMN (molecular single protonated ion [M + H]+); 669 for double molecular NMN single protonated ion [2 M + H]+; 123 for the [M + H]+ nicotinamide (NAM) ion, resulting from ion-source fragmentation of NMN. Selecting only m/z values between 334.5 and 335.5 (Fig. 2E) from the full MS data recorded while running the bacterial lysate separation by SEC on a PolyHEA column, a clear peak having the same RT as 260 nm chromatograms of the NMN standard (Fig. 2B) and bacterial lysate sample (Fig. 2D) is shown. The most abundant contaminant from the bacterial lysate was simultaneously identified on: TIC chromatogram (Fig. 2C), 260 nm chromatogram (Fig. Results Si l t 2D) and the MS chromatogram of the ions having an m/z between 122.5 and 123.5 (Fig. 2F) at a retention time of 4 min. The molecular identity of col- lected NMN fractions was confirmed by comparison against standard NMN MS/MS spectra (Fig. 3). Purity evaluation of NMN separated by SEC-HPLC using reversed-phase HPLC. The NMN RT using the RP-HPLC protocol was 3 minutes and the purity was determined as a percentage of the NMN peak area relative to a total detected peaks area of 260 nm chromatogram (Table 1). On the recorded 3D PDA absorption spectrum (Fig. 4A) there was no detectable signal above 280 nm therefore a plane view of the interval between 200 and 300 nm was generated. This result highlights a contaminant eluting at RT 4 min (Fig. 4C) which is clearly rep- resented on the 260 nm chromatogram (Fig. 4E). The 3D PDA absorption spectrum of the NMN obtained from the bacterial lysate by the proposed SEC method (as shown in section Mass spectrometry (MS)) has no detectable impurity (Fig. 4B,D and F). Endotoxin content in NMN fractions. The endotoxin level in the NMN fractions purified by SEC method was 0.51 EU/mL. This is below the calculated acceptable parenteral administration limit (1.18 EU/mL) (Fig. 5). George Cătălin Marinescu 1,2, Roua-Gabriela Popescu1,2 & Anca Dinischiotu1 Nicotinamide mon- onucleotide (NMN), a polar molecule with a molecular weight of 334 Da17, was proven effective as an NAD+ precursor in glucose intolerance studies, restoring NAD+ levels in mice with type 2 diabetes induced by a high fat diet or aging18. The nuclear NAD+ level modulates mitochondrial encoded gene expression and mitochondrial homoeostasis through a new pathway regulated by SIRT119. NMN was also recently found useful in limiting brain injury following intracerebral hemorrhage20. j y g g In the past, NMN was prepared by incubation of diphosphopyridine nucleotide with potato pyrophosphatase21 or from nicotinamide by extracts of acetone-powered human erythrocytes22. These methods produced low quan- tities of NMN. Nowadays NMN is obtained by microbial biotechnology techniques. To reduce the high cost of NMN and to improve on the available purity requires innovation and optimisation of the current production 1Department of Biochemistry and Molecular Biology, University of Bucharest, Bucharest, 050095, Romania. 2Independent Research Association, Bucharest, 012416, Romania. Correspondence and requests for materials should be addressed to G.C.M. (email: Catalin.Marinescu@independent-research.ro) Scientific REPOrtS \| (2018) 8:4433 \| DOI:10.1038/s41598-018-22806-8 1 1 www.nature.com/scientificreports/ methods. In this study the focus was directed on the downstream part of the biotechnological process. Only one purification method for NMN, utilising of ion exchange chromatography (IEC) followed by precipitation in acetone23 was identified in the scientific literature. General methods for nucleotides and other small molecules purification from biological extracts proposed separation in formic acid followed by filtration, lyophilisation, and two-dimension high performance liquid chromatography (HPLC) methods: boronate-affinity and ion-pair24, or reversed-phase chromatography (RPC) or via high speed counter-current chromatography (HSCCC). Several analytical HPLC methods have been reported using RPC for NMN quantification on either C18 coated silica25 or porous graphitic carbon stationary phase19,26. A similar approach, but using a gradient elution on a RP XBP C18 column (100 mm × 2.1 mm, 5 μm) and positive Electrospray Ionisation (ESI+) mass spectrometry detector running in Selected Ion Monitoring mode was previously reported for quantification of a NMN related molecule, also a NAD+ precursor: Nicotinamide Riboside (NR)27.hfi p ( ) The goal of this study was to design a simple and cost effective NMN purification method from a complex molecular mixture, such as the lysate of E. coli strain BL21 (DE3) pLysS genotype E. coli B F–dcm ompT hsdS(rB– m B–) gal λ(DE3) [pLysS Camr] [pET28a-hdNadV Kmr]. Discussion (E,F) represent 260 nm chromatograms extracted from above PDA data. Mobile phase was 50 mM formic acid, with a flow rate 3 mL/min, the stationary phase was PolyHEA, column dimensions: 250 × 9.4 mm; 5 μm, 60-Å. Sample volume was 20 μL of 5 mM nicotinamide mononucleotide standard (Sigma Aldrich, however impure, as shown) (A,C,E), respective bacterial cell lysate (B,D,F). NMN elutes at 5.7 minutes and is clearly separated from other similar molecules (e.g. nicotinamide which has a similar light absorption spectrum, is more abundant in the bacterial lysate and elutes at 3.8 minutes). The second polynomial order calibration curve with a correlation coefficient of 0.99 was obtained for quantitative determination by integration of the peak area of 5 NMN samples with known concentration levels: 0.03; 0.16; 0.33; 1.67; 3.34 g/L (G). gu e . C S C sepa at o o cot a de o o uc eot de ( ).h e g t abso pt o spect a (200–600 nm) (A,B), (200–300 nm) (C,D) were recorded by the Jasco Photodiode Array (PDA) detector. (E,F) represent 260 nm chromatograms extracted from above PDA data. Mobile phase was 50 mM formic acid, with a flow rate 3 mL/min, the stationary phase was PolyHEA, column dimensions: 250 × 9.4 mm; 5 μm, 60-Å. Sample volume was 20 μL of 5 mM nicotinamide mononucleotide standard (Sigma Aldrich, however impure, as shown) (A,C,E), respective bacterial cell lysate (B,D,F). NMN elutes at 5.7 minutes and is clearly separated from other similar molecules (e.g. nicotinamide which has a similar light absorption spectrum, is more abundant in the bacterial lysate and elutes at 3.8 minutes). The second polynomial order calibration curve with a correlation coefficient of 0.99 was obtained for quantitative determination by integration of the peak area of 5 NMN samples with known concentration levels: 0.03; 0.16; 0.33; 1.67; 3.34 g/L (G). the results were obtained by RPC separation and selective data acquisition (multiple reaction monitoring) of the m/z values corresponding to the molecules of interest including NMN, but ignoring other possible co-eluting contaminants. This is adequate for analytical quantification but not suitable for preparative purposes. Selected-ion monitoring (SIM) filters out all information on the co-eluting molecules, and was therefore inadequate for pre- parative purposes. Although the further purification of the NMN standard by RPC was successful (Fig. Discussion Our attempt to reproduce the previously reported ion exchange method23 of NMN separation from the bacterial lysate did not produce the expected results, probably because of the high nicotinamide (NAM) content of our bacterial lysate, as growing media was supplemented with 1% NAM as substrate for Nicotinamide phosphoribosyl transferase (Nampt). The mass chromatogram of ions with an m/z between 122.5 and 123.5 (Fig. 2F) correspond- ing to the known m/z value for NAM [M + H] + ion correlates with the RT of the most abundant impurity from the lysate (Fig. 2D). NAM and NMN co-eluted in the Dowex Cl− Ion Exchange Chromatography (IEC) column and co-precipitated in acetone together with high amounts of salts (data not shown). As previously reported, IEC optimisation is a time consuming and expensive process28, thus efforts were spent on identification of a simpler method.t Often used for analytical purposes, RP-chromatography (RPC) was not the most desirable method for separat- ing hydrophilic molecules because they are not well retained by the stationary phase29. Previously, RPC used for NMN quantification by Triple Quadrupole LC/MS/MS25 provided excellent analytical quantitative data. However, Scientific REPOrtS \| (2018) 8:4433 \| DOI:10.1038/s41598-018-22806-8 2 www.nature.com/scientificreports/ Figure 1. HPLC SEC separation of nicotinamide mononucleotide (NMN). The light absorption spectra (200–600 nm) (A,B), (200–300 nm) (C,D) were recorded by the Jasco Photodiode Array (PDA) detector. (E,F) represent 260 nm chromatograms extracted from above PDA data. Mobile phase was 50 mM formic acid, with a flow rate 3 mL/min, the stationary phase was PolyHEA, column dimensions: 250 × 9.4 mm; 5 μm, 60-Å. Sample volume was 20 μL of 5 mM nicotinamide mononucleotide standard (Sigma Aldrich, however impure, as shown) (A,C,E), respective bacterial cell lysate (B,D,F). NMN elutes at 5.7 minutes and is clearly separated from other similar molecules (e.g. nicotinamide which has a similar light absorption spectrum, is more abundant in the bacterial lysate and elutes at 3.8 minutes). The second polynomial order calibration curve with a correlation coefficient of 0.99 was obtained for quantitative determination by integration of the peak area of 5 NMN samples with known concentration levels: 0.03; 0.16; 0.33; 1.67; 3.34 g/L (G). Figure 1. HPLC SEC separation of nicotinamide mononucleotide (NMN). The light absorption spectra Figure 1. HPLC SEC separation of nicotinamide mononucleotide (NMN). The light absorption spectra (200–600 nm) (A,B), (200–300 nm) (C,D) were recorded by the Jasco Photodiode Array (PDA) detector. www.nature.com/scientificreports/ Bacterial lysate: Total Ion Current (C); HPLC UV detector 260 nm, analogue input (D); m/z 335 mass chromatogram showing NMN peak at RT: 5.61 (E); m/z 123 mass chromatogram showing NAM peak at RT: 3.65 (F); NMN MS spectrum showing protonated molecular ion [M + H] + (335 Da), double protonated molecular ion [2 M + H] + (669 Da) and the nicotinamide protonated molecular ion (123 Da) known to result from NMN fragmentation inside the ionisation source (H). The most interesting and newest RPC methods providing good RT and resolution for NMN separation use porous graphitic carbon19,26. The main drawback of this technique is the high price tag of this material which makes its usage prohibitive for preparative applications. A previously reported method for separation of nucleo- tides from whole-cell extract by formic acid extraction, filtration followed by a two-step chromatographic process: boronate-affinity and ion-pair24, facilitated a good separation for NMN. This, however, had the disadvantage of the two-step chromatographic process, that are more expensive compared to our proposed single step SEC.h The key of the proposed SEC separation is the stationary phase: silica with a covalently attached coating of poly(2-hydroxyethyl aspartamide) (PolyHEA). When using a non-denaturing mobile phase, the hydrogen bonds between adjacent polymer chains of the coating make the coating impermeable. When a denaturing mobile phase (like 50 mM formic acid) is used, the chaotropic agent prefers the stationary phase instead of water to form hydro- gen bonds. The increased steric radius of the PolyHEA coating might occlude the 60 Å pore, but by disrupting the hydration layer the available pore volume increases and the space between polymer chains becomes accessible to the mobile phase. This distance now represents the effective pore diameter30. When a non-denaturing mobile phase is used with a 60 Å pore PolyHEA column, the separation range is 60–10000 Da (according to manufactur- er’s manual). Using the denaturing mobile phase, the separation range shifts to 20–600 Da, which makes it suitable for separation of NMN (334 Da) from other small solutes. This is the only SEC stationary phase (according to our knowledge) having a resolution in this range. g g g SEC was used for two reasons: to desalt the sample to use the electrospray ionisation mass spectrometer and to filter out the higher molecular weight compounds from the mixture. www.nature.com/scientificreports/ www.nature.com/scientificreports/ Figure 2. Mass spectrometry identification of nicotinamide mononucleotide (NMN). Thermo Velos Pro MS detector ESI + mode, Spray Voltage: 3 kV, Capillary temp: 375 °C; isocratic HPLC elution with 50 mM formic acid flow: 3 mL/min, PolyHEA column 250 × 9.4 mm, 5 μm, 60-Å. Sample volume: 20 μL. H-ESI-II ion source was connected parallel to the MD2015 UV-Vis detector using 1:10 post column flow splitter. Standard NMN 5 mM (Sigma Aldrich) sample elution: total Ion Current chromatogram (A) shows NMN eluting at 6 minutes. 260 nm Chromatogram (B). NMN is identified by protonated molecular ion [M + H] + (335 Da) on the MS spectrum at a retention time of 6.07 (G), double molecular protonated ions [2 M + H] + (669 Da) and nicotinamide (NAM) fragment protonated molecular ion (123 Da) resulted from NMN fragmentation inside the ionisation source. Bacterial lysate: Total Ion Current (C); HPLC UV detector 260 nm, analogue input (D); m/z 335 mass chromatogram showing NMN peak at RT: 5.61 (E); m/z 123 mass chromatogram showing NAM peak at RT: 3.65 (F); NMN MS spectrum showing protonated molecular ion [M + H] + (335 Da), double protonated molecular ion [2 M + H] + (669 Da) and the nicotinamide protonated molecular ion (123 Da) known to result from NMN fragmentation inside the ionisation source (H). Figure 2. Mass spectrometry identification of nicotinamide mononucleotide (NMN). Thermo Velos Pro MS detector ESI + mode, Spray Voltage: 3 kV, Capillary temp: 375 °C; isocratic HPLC elution with 50 mM formic acid flow: 3 mL/min, PolyHEA column 250 × 9.4 mm, 5 μm, 60-Å. Sample volume: 20 μL. H-ESI-II ion source was connected parallel to the MD2015 UV-Vis detector using 1:10 post column flow splitter. Standard NMN 5 mM (Sigma Aldrich) sample elution: total Ion Current chromatogram (A) shows NMN eluting at 6 minutes. 260 nm Chromatogram (B). NMN is identified by protonated molecular ion [M + H] + (335 Da) on the MS spectrum at a retention time of 6.07 (G), double molecular protonated ions [2 M + H] + (669 Da) and nicotinamide (NAM) fragment protonated molecular ion (123 Da) resulted from NMN fragmentation inside the ionisation source. Discussion 4C and E), the small RT difference between the two peaks limited the preparative usage of the method, while higher column load widened the peaks and decreased resolution. Since the most abundant impurity in our bacterial lysate was NAM (Figs 1F and 2D,F) and it had a similar RT with the main impurity from the NMN standard (Fig. 1E) on the proposed SEC method, it was thus concluded that the detected impurity in the NMN standard was also NAM. Therefore, the unknown peak (RT = 4 min) on the RP elution (Fig. 4C and E) of the NMN standard could be NAM. Both NAM and NMN peaks resulted by RP separation were wider than the peaks resulting from the proposed SEC method, generating a lower NMN concentration in the collected fraction and a higher likelihood of co-eluted contaminants. the results were obtained by RPC separation and selective data acquisition (multiple reaction monitoring) of the m/z values corresponding to the molecules of interest including NMN, but ignoring other possible co-eluting contaminants. This is adequate for analytical quantification but not suitable for preparative purposes. Selected-ion monitoring (SIM) filters out all information on the co-eluting molecules, and was therefore inadequate for pre- parative purposes. Although the further purification of the NMN standard by RPC was successful (Fig. 4C and E), the small RT difference between the two peaks limited the preparative usage of the method, while higher column load widened the peaks and decreased resolution. Since the most abundant impurity in our bacterial lysate was NAM (Figs 1F and 2D,F) and it had a similar RT with the main impurity from the NMN standard (Fig. 1E) on the proposed SEC method, it was thus concluded that the detected impurity in the NMN standard was also NAM. Therefore, the unknown peak (RT = 4 min) on the RP elution (Fig. 4C and E) of the NMN standard could be NAM. Both NAM and NMN peaks resulted by RP separation were wider than the peaks resulting from the proposed SEC method, generating a lower NMN concentration in the collected fraction and a higher likelihood of co-eluted contaminants. Scientific REPOrtS \| (2018) 8:4433 \| DOI:10.1038/s41598-018-22806-8 3 www.nature.com/scientificreports/ Our first intention was to collect the fraction containing NMN, to identify the co-eluting contaminants and further purify it by a second chromato- graphic method. It was a surprise to discover that this material made unnecessary any additional purification Scientific REPOrtS \| (2018) 8:4433 \| DOI:10.1038/s41598-018-22806-8 4 www.nature.com/scientificreports/ Figure 3. MS/MS Collision-Induced Dissociation (CID) spectral data of nicotinamide mononucleotide (NMN) from collected fractions (SEC and RPC) against standard NMN, using Thermo Velos Pro linear ion trap MS detector running in ESI+ mode, Spray Voltage: 3 kV, Capillary temperature: 375 °C, flow: 3 µL/min, H-ESI-II ion source. Figure 3. MS/MS Collision-Induced Dissociation (CID) spectral data of nicotinamide mononucleotide (NMN) from collected fractions (SEC and RPC) against standard NMN, using Thermo Velos Pro linear ion trap MS detector running in ESI+ mode, Spray Voltage: 3 kV, Capillary temperature: 375 °C, flow: 3 µL/min, H-ESI-II ion source. ( ) g , gh p detector running in ESI+ mode, Spray Voltage: 3 kV, Capillary temperature: 375 °C, flow: 3 µL/min, H-ESI-II ion source. Name Retention time [min] Quantity [%Area] Area [mAU Min] NMN 3.0 94.03 711.4 NAM 3.8 5.97 45.1 Table 1. Nicotinamide mononucleotide (NMN) standard (Sigma) Reversed-Phase HPLC peak detection and integration results. Table 1. Nicotinamide mononucleotide (NMN) standard (Sigma) Reversed-Phase HPLC peak detection an integration results. steps. Surprisingly, using the new method, further purification of the standard NMN (Sigma Aldrich, 95% adver- tised purity) was made possible as illustrated in Fig. 1E. Due to the known high NAM content of the bacterial lysate, the main contaminant in the lysate was identified as NAM, which has been determined to have a retention time of 3.5–4 minutes (Fig. 1F). Th d d l d h h l d b ( ) d ( ) steps. Surprisingly, using the new method, further purification of the standard NMN (Sigma Aldrich, 95% adver- tised purity) was made possible as illustrated in Fig. 1E. Due to the known high NAM content of the bacterial lysate, the main contaminant in the lysate was identified as NAM, which has been determined to have a retention time of 3.5–4 minutes (Fig. 1F).h The data generated revealed that the elution order between NAM (122 Da) and NMN (334 Da) was surpris- ingly reversed. According to the SEC principle, NMN should elute first, having a greater molecular weight, sug- gesting that some other physical interactions occur31. Scientific REPOrtS \| (2018) 8:4433 \| DOI:10.1038/s41598-018-22806-8 www.nature.com/scientificreports/ Electrostatic charge might be the cause of the observed reversed elution order, as the titration curve of the PolyHEA material showed that at a pH below 4.4 the material behaves as having a net positive charge, and above 4.4 it behaves as though having a negative charge30. With 50 mM formic acid as the mobile phase, the pH was 2.0, therefore the coating net positive charge might cause electrostatic repulsion of NAM (pKa = 3.63)32 which also has a net positive charge at pH 2.0. The NMN phosphate group might have formed an electrostatic attraction with the coating. To verify this hypothesis, the mobile phase was initially replaced with an uncharged chaotrope i.e. 1,1,1,3,3,3-Hexafluoro-2-propanol (HFIP) 50 mM. This did not change the elution order, but increased RT for both NAM and NMN (Fig. 6). To eliminate any electro- static effect, 200 mM NaCl was added to the mobile phase. In these conditions, RT of NAM was the same while NMN RT was slightly shorter, without changing the elution order (Fig. 6E). This suggested that although electro- static interactions determine coating repulsion for both NAM and (unexpectedly) NMN, they are not the cause Scientific REPOrtS \| (2018) 8:4433 \| DOI:10.1038/s41598-018-22806-8 5 www.nature.com/scientificreports/ Figure 4. Purity evaluation of nicotinamide mononucleotide (NMN) separated by SEC PolyH reversed-phase HPLC (C18). The light absorption spectrum recorded by the Jasco MD2015 U Array (PDA) detector during HPLC isocratic elution, NMN standard Sigma (however, impur versus NMN purified from bacterial lysate by SEC PolyHEA (B,D). HPLC Chromatogram 26 Figure 4. Purity evaluation of nicotinamide mononucleotide (NMN) separated by SEC PolyH reversed-phase HPLC (C18). The light absorption spectrum recorded by the Jasco MD2015 U Array (PDA) detector during HPLC isocratic elution, NMN standard Sigma (however, impur versus NMN purified from bacterial lysate by SEC PolyHEA (B,D). HPLC Chromatogram 26 elution of NMN standard Sigma 95% purity (E) versus NMN purified from the bacterial lysat (F). Mobile phase: 10% acetonitrile, flow rate 0.9 mL/min, column: Genesis C18 James Colum 4 µm, 120 Å, injected sample: 20 μL. 10 mM NMN, starts eluting at 3 minutes. Figure 4. Purity evaluation of nicotinamide mononucleotide (NMN) separated by SEC PolyHEA using reversed-phase HPLC (C18). The light absorption spectrum recorded by the Jasco MD2015 UV-Vis Photodiode Array (PDA) detector during HPLC isocratic elution, NMN standard Sigma (however, impure as shown) (A,C) versus NMN purified from bacterial lysate by SEC PolyHEA (B,D). www.nature.com/scientificreports/ Endotoxin detection in NMN fractions by the LAL chromogenic method. The concentration of NMN sample was 1 mM. Figure 6. Separation of nicotinamide mononucleotide (NMN) from (NAM) using uncharged chaotrope mobile phase (50 mM 1,1,1,3,3,3-Hexafluoro-2-propanol (HFIP)), isocratic elution, flow: 3 mL/min on PolyHEA column 250 × 9.4 mm, 5 μm, 60-Å. Jasco PU2089 HPLC pump. Detectors: Thermo Velos Pro linear ion trap MS running in ESI + mode, Spray Voltage: 3 kV, Capillary temp: 375 °C, H-ESI-II ion source, connected parallel to MD2015 UV-Vis detector using 1:10 post column flow splitter. Sample volume: 20 μL of standard NMN 10 mM (Sigma Aldrich). Total Ion Current (A); 260 nm absorption chromatogram (B); 335 Da ions (C); 123 Da ions (D); The same elution conditions but mobile phase supplemented with 200 mM NaCl, to eliminate the electrostatic repulsion effect, 260 nm chromatogram UV-Vis detector only (E). Figure 5. Endotoxin detection in NMN fractions by the LAL chromogenic method. The concentration of NMN sample was 1 mM. Figure 5. Endotoxin detection in NMN fractions by the LAL chromogenic method. The concentration of NMN sample was 1 mM. sample was 1 mM. Figure 6. Separation of nicotinamide mononucleotide (NMN) from (NAM) using uncharged c mobile phase (50 mM 1,1,1,3,3,3-Hexafluoro-2-propanol (HFIP)), isocratic elution, flow: 3 mL/ PolyHEA column 250 × 9.4 mm, 5 μm, 60-Å. Jasco PU2089 HPLC pump. Detectors: Thermo Ve d l k ll Figure 6. Separation of nicotinamide mononucleotide (NMN) from (NAM) using unchar mobile phase (50 mM 1,1,1,3,3,3-Hexafluoro-2-propanol (HFIP)), isocratic elution, flow: 3 PolyHEA column 250 × 9.4 mm, 5 μm, 60-Å. Jasco PU2089 HPLC pump. Detectors: Therm ion trap MS running in ESI + mode, Spray Voltage: 3 kV, Capillary temp: 375 °C, H-ESI-II ll l t MD2015 UV Vi d t t i 1 10 t l fl litt S l l Figure 6. Separation of nicotinamide mononucleotide (NMN) from (NAM) using uncharged chaotrope mobile phase (50 mM 1,1,1,3,3,3-Hexafluoro-2-propanol (HFIP)), isocratic elution, flow: 3 mL/min on PolyHEA column 250 × 9.4 mm, 5 μm, 60-Å. Jasco PU2089 HPLC pump. Detectors: Thermo Velos Pro linear ion trap MS running in ESI + mode, Spray Voltage: 3 kV, Capillary temp: 375 °C, H-ESI-II ion source, connected parallel to MD2015 UV-Vis detector using 1:10 post column flow splitter. Sample volume: 20 μL of standard NMN 10 mM (Sigma Aldrich). www.nature.com/scientificreports/ HPLC Chromatogram 260 nm isocratic elution of NMN standard Sigma 95% purity (E) versus NMN purified from the bacterial lysate by SEC PolyHEA (F). Mobile phase: 10% acetonitrile, flow rate 0.9 mL/min, column: Genesis C18 James Column 250 × 4.6 mm, 4 µm, 120 Å, injected sample: 20 μL. 10 mM NMN, starts eluting at 3 minutes. Figure 4. Purity evaluation of nicotinamide mononucleotide (NMN) separated by SEC PolyHEA using reversed-phase HPLC (C18). The light absorption spectrum recorded by the Jasco MD2015 UV-Vis Photodiode Array (PDA) detector during HPLC isocratic elution, NMN standard Sigma (however, impure as shown) (A,C) versus NMN purified from bacterial lysate by SEC PolyHEA (B,D). HPLC Chromatogram 260 nm isocratic elution of NMN standard Sigma 95% purity (E) versus NMN purified from the bacterial lysate by SEC PolyHEA (F). Mobile phase: 10% acetonitrile, flow rate 0.9 mL/min, column: Genesis C18 James Column 250 × 4.6 mm, 4 µm, 120 Å, injected sample: 20 μL. 10 mM NMN, starts eluting at 3 minutes. of the reversed elution order, as eliminating them does not restore the theoretical SEC elution order. No further investigations on the matter were performed, as it does not serve the purpose of this study. Although an ultracentrifugation step was performed, that was convenient for small scale processing during i f the reversed elution order, as eliminating them does not restore the theoretical SEC elution order. No further nvestigations on the matter were performed, as it does not serve the purpose of this study. Although an ultracentrifugation step was performed, that was convenient for small scale processing during the development of the method, for higher scale processing, filtration should be used instead. Moreover, the same filtration method and equipment could be used for both cell separation and lysate clarification. Another advantage of the SEC method presented in this study is that salt molecules (known to be harmful for the mass spectrometry electrospray ionisation source) have a longer RT so the flow can be diverted to waste after elution of the useful fractions, thereby protecting the equipment. Thus, the method applied also performs desalting without Scientific REPOrtS \| (2018) 8:4433 \| DOI:10.1038/s41598-018-22806-8 6 www.nature.com/scientificreports/ Figure 5. Endotoxin detection in NMN fractions by the LAL chromogenic method. The concentration of NMN sample was 1 mM. Figure 5. Endotoxin detection in NMN fractions by the LAL chromogenic method. The concentration of NMN sample was 1 mM. Figure 5. www.nature.com/scientificreports/ www.nature.com/scientificreports/ the need for any additional processing steps. As previously reported24 the formic acid from the collected NMN fraction is simply removed by lyophilisation.iii the need for any additional processing steps. As previously reported24 the formic acid from the collected NMN fraction is simply removed by lyophilisation.iii p y y y p Mass spectrometry confirmed NMN peaks identified by the UV-Vis spectrum. Specific MS (Fig. 2) and MS/ MS spectra (Fig. 3), consistent with previously reported data25–27 were identified in both standard and lysate, increasing certainty of the identity of the obtained product. Considering the molecular diversity of the bacterial lysate, the theoretical possibility to have an impurity with similar size but lower concentration compared to NMN still existed. This type of impurity could hypothetically have the same RT as NMN on the proposed SEC method. As NMN was more abundant, running the MS detector in Full Scan mode could also fail to detect this hypothet- ical impurity. A typical example is the lack of detection for a 123 m/z peak caused by the fragmentation of NMN in the ionisation source at RT = 6 min during the SEC lysate separation (Fig. 2F). Although the m/z = 123 signal is present (Fig. 2H), it is completely eclipsed on the m/z 122.5–123.5 ion chromatogram by the strong NAM peak at RT = 4. Therefore, the product resulting from SEC was evaluated by the second HPLC method based on a completely different separation principle (reversed-phase). During the RP separation, no impurity in the product collected from SEC was observed (Fig. 4B,D,F). Only the NMN UV absorption spectrum was observed which supported the results of this study that SEC alone separates NMN from the bacterial lysate. pp y p y Although the pharmaceutical formulation was beyond the scope of this study, the endotoxin level had to be controlled in case of parenteral administrated drugs. Its concentration was 0.51 EU/mL, below the accepted limit of 1.18 EU/mL. This resulted in a strong suspicion that the detected endotoxin level in the NMN fractions was caused by the HPLC solvents used, which were not labelled as endotoxin free. This hypothesis was supported by the determined endotoxin level in the laboratory water used in the HPLC solvents which was 1.52 EU/mL, three times higher than in the NMN sample. Methods NMN P ii NMN Purification Workflow. The process was designed following the classical biotechnological process workflow from the bacterial broth to the pure final product31,36. The bacterial cells were grown in a 10 L bioreactor and were separated from the media using a 10 L separatory funnel. The cell membranes were disrupted using a sonicator cell disruptor. The cell lysate was clarified by centrifugation and NMN was purified by a single step size exclusion chromatography technique (Fig. 7). Strains, Media, Growth Conditions and Separation of bacterial cells. An overnight culture of E. coli strain BL21(DE3)pLysS genotype E. coli B F–dcm ompT hsdS(rB– m B–) gal λ(DE3) [pLysS Camr] [pET28a-hdNadV Kmr] (a gift from Independent Research Association, Bucharest - 012416, Romania), 50 mL, grown in a shaking incubator at 250 rpm was used as inoculum for a 10 L bench-top bioreactor culture. This strain is carrying a plasmid (pET28a-hdNadV, Addgene ID #83362) expressing a gene (nadV) from Haemophilus ducreyi coding for Nampt catalysing the direct production of NMN from nicotinamide (NAM). LB growing media (NaCl 10 g/L, Tryptone 10 g/L, yeast extract 5 g/L) supplemented with 1% NAM and 1% glucose was sterilised in autoclave for 20 minutes at 120 °C. Once the sterilised media cooled down below 60 °C, kanamy- cin was added to a final concentration of 50 µg/mL. The temperature for bacterial culture was set to 37 °C. The bioreactor was continuously measuring the optical density of the culture at 600 nm (OD600), the stirring motor speed being set to 150 rpm. Once OD600 reached 0.45, the expression of nadV was induced by adding Isopropyl- 1-thio-β-D-galactopyranoside (IPTG) to a final concentration of 1 mM. When the measured OD600 of the cul- ture reached 0.75, the media containing bacterial cells was automatically pumped into a 10 L separatory funnel (supplied by Adrian Sistem SRL) and left at 10 °C for 24 h. The funnel tap was then slowly opened, and 100 mL of cells rich media were collected. Extraction and Clarification. The collected cells were evenly distributed in five 50 mL Falcon Conical Centrifuge Tubes and the cell membranes were disrupted by sonication at 20 kHz on ice in 3 cycles of 30 seconds separated by 30 seconds pause using a sonicator cell disruptor model W185F (Heat Systems-Ultrasonic Inc.). The lysate was ultracentrifuged using the Beckman Optima LE-80 K for 3 hours at 500,000 g at 4 °C. www.nature.com/scientificreports/ Having a molecular mass greater than 10 kDa, lipopolysaccharides (LPS) resulting from the biotechnological processes are large molecules, which sediment during the ultracentrifugation step. In the less probable situation when they are still present in the clarified lysate, SEC is a known method for their elimination33,34, and that is the reason for the lower endotoxin level detected in our NMN sample com- pared with the HPLC solvents used. Evidently, an even lower endotoxin level can be obtained by using endotoxin free solvents. In the case of oral administration, endotoxin contamination is not a concern, as they do not pass through an intestinal barrier35.f Growing evidence of anti-ageing and anti-diabetic effects of nicotinamide mononucleotide (NMN) and the current high price tag of this promising molecule were the two key facts motivating us to accomplish this work. The proposed method of NMN purification by Size Exclusion HPLC (SEC) on silica with a covalently attached coating of poly(2-hydroxyethyl aspartamide) (PolyHEA) stationary phase using isocratic elution with denaturing mobile phase (50 mM formic acid) is the first proposed single step chromatographic method for NMN purifi- cation from a complex molecular mixture such as a bacterial lysate. The method is simple, patent-free, directly applicable for industrial production with a minimum scale up effort, with bulk material being commercially available to prepare higher capacity columns. It not only eliminates the need for multiple chromatographic steps, but also simplifies the lysate filtration and clarification process, since SEC is not influenced by higher molecular mass compounds which are nonetheless not retained by the column. Therefore, the proposed purification method should substantially contribute to reduction of NMN production costs while also increasing the purity to the level suitable for usage in humans. www.nature.com/scientificreports/ Total Ion Current (A); 260 nm absorption chromatogram (B); 335 Da ions (C); 123 Da ions (D); The same elution conditions but mobile phase supplemented with 200 mM NaCl, to eliminate the electrostatic repulsion effect, 260 nm chromatogram UV-Vis detector only (E). Scientific REPOrtS \| (2018) 8:4433 \| DOI:10.1038/s41598-018-22806-8 7 www.nature.com/scientificreports/ www.nature.com/scientificreports/ Figure 7. The schematic diagram for the proposed nicotinamide mononucleotide (NMN) purification process. Bacterial cells are grown in the bioreactor, the media containing bacteria is pumped into a separating funnel and stored at 4 °C. After 24 hours, the sedimented bacterial cells are collected, cell membranes are disrupted by ultrasound on ice, cell debris and macromolecules are separated by ultracentrifugation and discarded, NMN is finally separated by a single step size exclusion chromatography (SEC) on a PolyHEA column eluted with 50 mM formic acid. Figure 7. The schematic diagram for the proposed nicotinamide mononucleotide (NMN) purification process. Bacterial cells are grown in the bioreactor, the media containing bacteria is pumped into a separating funnel and stored at 4 °C. After 24 hours, the sedimented bacterial cells are collected, cell membranes are disrupted by ultrasound on ice, cell debris and macromolecules are separated by ultracentrifugation and discarded, NMN is finally separated by a single step size exclusion chromatography (SEC) on a PolyHEA column eluted with 50 mM formic acid. High Performance Liquid Chromatography (HPLC). Size Exclusion Chromatography High Performance Liquid Chromatography (SEC-HPLC). Samples of 20 μL were consequently poured on the SEC column. The Jasco HPLC PU-2089 pump and Rheodyne 20 μL manual injection loop were connected to 250 × 9.4 mm; 5 μm, 60 Å PolyHEA column (silica with a covalently attached coating of poly(2-hydroxyethyl aspartamide) pur- chased from PolyLC INC (Item# 259HY05006). Post column, a 10:1 ratio flow splitter was used to connect a Jasco UV-Vis Photodiode Array (PDA) MD-2015 on the high flow side, respective of the Heated Electrospray Ionisation (H-ESI-II) from Thermo Scientific Velos Pro Linear Ion Trap Mass Spectrometer on the low flow side. Between the flow splitter and the ionisation source, the flow was passed through the mass spectrometer divert/ inject valve as specified in the Thermo Velos Pro user manual. Jasco LC-NETII-ADC controlled by Chrompass 1.8.6.1. software was used to run the LC System while the LTQ Tune Plus 2.7 and Xcalibur 2.2. were used for the mass spectrometry detector and acquisition process. The Jasco LC system was configured to generate the acqui- sition start signal for the MS to ensure reproducible retention times (RT) were achieved on both data systems. Analog channel 1 from the LC system was connected to the 1 V analogue input of the mass spectrometer. Methods NMN P ii The supernatant was carefully collected and used for NMN purification by High Performance Liquid Chromatography (HPLC). 8 Scientific REPOrtS \| (2018) 8:4433 \| DOI:10.1038/s41598-018-22806-8 www.nature.com/scientificreports/ Mass spectrometry grade water (39253 Fluka LC-MS CHROMASOLV) and formic acid (94318 Fluka) were purchased from Sigma-Aldrich as well as other chemicals and reagents, unless otherwise stated. For NMN SEC separation, the isocratic solvent A: 50 mM formic acid was used at a flow rate of 3 mL/min for 20 minutes. The UV-Vis light absorption spectrum (200–640 nm) was continuously recorded. The light absorption chromatogram at 260 nm was recorded by both Chrompass and Xcalibur software.f y pt External standard NMN (Sigma N3501-25MG) solutions were prepared for 5 different NMN concentrations (0.03, 0.16, 0.33, 1.67 and 3.34 g/L), of which 20 µL samples were consecutively eluted by SEC, in the ascending order of concentration. The NMN peak areas of the 260 nm chromatograms were then used to generate the cali- bration curve by regression37 for NMN quantification in Chrompass Software (Fig. 1G). y g qi pt g NMN fractions were collected each time, mixed together, lyophilised and stored at −80 °C. Reversed-Phase High Performance Liquid Chromatography (RP-HPLC). The NMN purified by SEC was hydrated in 200 µL deionised water, 20 µL of the resulting solution was injected in the HPLC sample loop and eluted follow- ing the same protocol, previously described.h Reversed-Phase High Performance Liquid Chromatography (RP-HPLC). The NMN purified by SEC was hydrated in 200 µL deionised water, 20 µL of the resulting solution was injected in the HPLC sample loop and eluted follow- ing the same protocol, previously described.h g p p y The Jasco HPLC PU-2089 pump and Rheodyne 20 µL manual injection loop were connected to the Genesis C18 James Column 250 × 4.6 mm; 4 µm; 120 Å column and the Jasco UV-Vis Photodiode Array (PDA) MD-2015 was connected to the column output. The HPLC system was controlled by Jasco LC-NETII-ADC and Chrompass 1.8.6.1. software. A volume of 20 µL of 10 mM NMN (Sigma) standard solution was also injected. Isocratic elu- tion with 10% acetonitrile at a flow rate of 0.9 mL/min for 5 minutes was used. The light absorption spectrum (220–640 nm) was recorded and a 260 nm chromatogram was generated. Scientific REPOrtS \| (2018) 8:4433 \| DOI:10.1038/s41598-018-22806-8 9 www.nature.com/scientificreports/ Mass spectrometry (MS). The mass spectrometer connected to the HPLC system as described in the Size Exclusion Chromatography High Performance Liquid Chromatography (SEC-HPLC) section was running in ESI + mode, Spray Voltage: 3 kV, Capillary Temperature was 375 °C. www.nature.com/scientificreports/ The full m/z range from 70 to 1000 Da was scanned. Prior to sample injection, the mass spectrometer divert valve was switched to the inject position, the MS data acquisition was started from the LTQ Tune software with the “waiting for contact closure” option. Thus, the start acquisition signal was generated by the HPLC injection. After 15 minutes of acquisition, the MS valve was switched to the divert position to protect the instrument from salts. Using Xcalibur software, the total ion current chromatogram and the signal intensity corresponding to m/z between 334.5 and 335.5 Da respective 122.5 and 123.5 Da were extracted from full MS raw data. The 1 V analogue input of the MS was continuously recording a 260 nm absorption data sent by the Jasco MD-2015 plus detector. MS/MS spectra (m/z: 90–340) for NMN containing fractions were obtained by Collision-Induced Dissociation (CID, energy values: 15 and 20) of single charged protonated ion (m/z = 335). Limulus amoebocyte lysate (LAL) assay. The lipopolysaccharides (LPS) content from the SEC-purified NMN fractions were determined by the Limulus amoebocyte assay chromogenic method following the kit man- ufacturer’s instructions (ToxinSensorTM Chromogenic LAL Endotoxin Assay Kit, GenScript, Piscataway, USA) which had a sensitivity between 0.005 and 1 endotoxin units per millilitre (EU/mL). The human equivalent dose (HED) of NMN was calculated as being 40.5 mg/kg by the methodology presented by Nair & Jacob38, starting from doses used without any observed adverse effect in animal studies18,39,40. The concentration of NMN sample was 1 mM which is above the minimum valid concentration (MVC) determined as 0.038 mg/mL41. At a daily perfusable dose of 2.4 g NMN in 250 mL isotonic solution the endotoxin calculated limit is 1.18 EU/mL42. References l Expression of Nuclear Encoded Genes Involved in Mitochondrial Oxidati Metabolism in Skeletal Muscle of Insulin-Resistant Women With Polycystic Ovary Syndrome. Diabetes 56, 2349–2355 (2007). 0 Ki J A W i Y & S J R R l f i h d i l d f i i i li i Ci R 102 401 414 (2008) y y y y 10. Kim, J.-A., Wei, Y. & Sowers, J. R. Role of mitochondrial dysfunction in insulin resistance. Circ. Res. 102, 401–414 (2008). y 11. Burkart, A. M. et al. Insulin Resistance in Human iPS Cells Reduces Mitochondrial Size and Function. Sci. Rep. 6, 22788 (201 2. Braidy, N. et al. Age related changes in NAD+ metabolism oxidative stress and sirt1 activity in wistar rats. PLoS One 6, 1–18 (2011) 12. Braidy, N. et al. Age related changes in NAD+ metabolism ox 3. Houtkooper, R. H., Cantó, C., Wanders, R. J. & Auwerx, J. The secret life of NAD+: An old metabolite controlling new metabolic signaling pathways. Endocr. Rev. 31, 194–223 (2010). g g p y ( ) 4. Sporty, J., Lin, S. J., Kato, M., Ognibene, T. & Stewart, B. Characterization of chromosomal integration sites for heterologous gene expression in Saccharomyces cerevisiae. Yeast 26, 545–551 (2009). p y ( ) 15. Aragonès, G. et al. Dietary proanthocyanidins boost hepatic NAD+ metabolism and SIRT1 expression and activity in a d dependent manner in healthy rats. Sci. Rep. 6, 24977, https://doi.org/10.1038/srep24977 (2016). y p g 6. Mouchiroud, L., Houtkooper, R. H. & Auwerx, J. NAD+ metabolism: a therapeutic target for age-related metabolic disease. Crit Rev. Biochem. Mol. Biol. 48, 397–408 (2013). 7. National Center for Biotechnology Information, PubChem Compound Database; CID = 14180. https://pubchem.ncbi.nlm.nih.gov compound/14180 (2017). compound/14180 (2017). 18. Yoshino, J., Mills, K. F., Yoon, M. J. & Imai, S. I. Nicotinamide mononucleotide, a key NAD+ intermediate, treats the pathophysiology of diet and age induced diabetes in mice Cell Metab 14 528 536 (2011) 8. Yoshino, J., Mills, K. F., Yoon, M. J. & Imai, S. I. Nicotinamide mononucleotide, a key NAD+ intermediate, treats the pathophysiology of diet- and age-induced diabetes in mice. Cell Metab. 14, 528–536 (2011). g ( ) 9. Zhang, T. et al. Enzymes in the NAD+ salvage pathway regulate SIRT1 activity at target gene promoters. J. Biol. Chem. 284 20408–20417 (2009).t 20. Wei, C., Kong, Y., Li, G., Guan, Y. & Wang, P. 1. International Diabetes Federation, IDF Diabetes Atlas. Available from http://www.diabetesatlas.org/ (2015). References l g p p 3. Taylor, R. Insulin resistance and type 2 diabetes. Diabetes 61, 778–779 (2012).h g p p 3. Taylor, R. Insulin resistance and type 2 diabetes. Diabetes 61, 778–779 (2012).h y yp ( ) 4. Martin, S. D. & McGee, S. L. The role of mitochondria in the aetiology of insulin resistance and type 2 diabetes. Biochim. Biophys Acta 1840, 1303–12 (2014). 5. Montgomery, M. K. & Turner, N. Mitochondrial dysfunction and insulin resistance: an update. Endocr. Connect. 4, R1–R15 (2015 5. Montgomery, M. K. & Turner, N. Mitochondrial dysfunction and insulin resistance: an update. Endocr. Connect. 4, R1–R15 (2015). 6 Petersen K F Dufour S Befroy D Garcia R & Shulman G I Impaired mitochondrial activity in the insulin resistant offspring 5. Montgomery, M. K. & Turner, N. Mitochondrial dysfunction and insulin resistance: an update. Endocr. Connect. 4, R1–R15 (2015). 6. Petersen, K. F., Dufour, S., Befroy, D., Garcia, R. & Shulman, G. I. Impaired mitochondrial activity in the insulin-resistant offspring of patients with type 2 diabetes N Engl J Med 7 664 671 (2004) g y y p ( ) 6. Petersen, K. F., Dufour, S., Befroy, D., Garcia, R. & Shulman, G. I. Impaired mitochondrial activity in the insulin-resistant offspring of patients with type 2 diabetes. N. Engl. J. Med. 7, 664–671 (2004).h 6. Petersen, K. F., Dufour, S., Befroy, D., Garcia, R. & Shulman, G. I. Imp of patients with type 2 diabetes. N. Engl. J. Med. 7, 664–671 (2004).h 6. Petersen, K. F., Dufour, S., Befroy, D., Garcia, R. & Shulman, G. I. Impaired mitochondrial activity in the insulin-resistant offsp of patients with type 2 diabetes. N. Engl. J. Med. 7, 664–671 (2004).h of patients with type 2 diabetes. N. Engl. J. Med. 7, 664–671 (2004).h y g 7. Szendroedi, J., Phielix, E. & Roden, M. The role of mitochondria in insulin resistance and type 2 diabetes mellitus. Nat. Rev. Endocrinol. 8, 92–103 (2011). 8. Ennequin, G. et al. Neuregulin 1 improves complex 2-mediated mitochondrial respiration in skeletal muscle of healthy and diabetic mice. Sci. Rep. 7, 1742, https://doi.org/10.1038/s41598-017-02029-z (2017). p p g 9. Skov, V., Glintborg, D., Knudsen, S. & Jensen, T. Expression of Nuclear-Encoded Genes Involved in Mitochondrial Oxidativ Metabolism in Skeletal Muscle of Insulin-Resistant Women With Polycystic Ovary Syndrome. Diabetes 56, 2349–2355 (2007). 9. Skov, V., Glintborg, D., Knudsen, S. & Jensen, T. Scientific REPOrtS \| (2018) 8:4433 \| DOI:10.1038/s41598-018-22806-8 p g ( ) 2. World Health Organization (WHO), Global Report on Diabetes. http://www.who.int/diabetes/en/ (2016). y yp ( ) Martin, S. D. & McGee, S. L. The role of mitochondria in the aetiology of insulin resistance and type 2 diabetes. Biochim. Biophys. cta 1840, 1303–12 (2014). www.nature.com/scientificreports/ www.nature.com/scientificreports/ ugbank, Nicotinamide https://www.drugbank.ca/drugs/DB02701 ( g g g 33. London, A. S., Mackay, K., Lihon, M., He, Y. & Alabi, B. R. Gel filtration chromatography as a method for removing bacterial endotoxin from antibody preparations. Biotechnol. Prog. 30, 1497–1501 (2014). y p p g ( ) 4. Ongkudon, C. M., Chew, J. H., Liu, B. & Danquah, M. K. Chromatographic Removal of Endotoxins: A Bioprocess Engineer’ Perspective. ISRN Chromatogr. 2012, 1–9 (2012).h p g ( ) 35. Dawson, M. E. Endotoxin testing in Pharmaceutical Dosage Forms - Parenteral Medications Third Edition: Volume 2: Facility Design, Sterilization and Processing (eds. Nema, S. & Ludwig, J. D.) 146–187 (CRC Press, 2010).t g g 6. Merchuk, J. C. & Gluz, M. Bioreactors, Air-Lift Reactors in Encyclopedia of Bioprocess Technology: Fermentation, Biocatalysis and Bioseparation (eds. Flickinger, M. C. & Drew, S. W.) 320–353 (John Wiley & Sons, Inc., 1999). p g y 7. Snyder, E. & Kirkland, J. J. Quantitation (including trace analysis) in Practical HPLC Method Development (ed. Kruse, H. P.) 655 (John Wiley & Sons, 1997). y 38. Nair, A. & Jacob, S. A simple practice guide for dose conversion between animals and human. J. Basic Clin. Pharm. 7, 27 (2016 y 38. Nair, A. & Jacob, S. A simple practice guide for dose conversion between animals and human. J. Basic Clin. Pharm. 7, 27 (2016). 9. Gomes, A. P. et al. Declining NAD+ induces a pseudohypoxic state disrupting nuclear-mitochondrial communication during aging Cell 155, 1624–1638 (2013). 40. Yamamoto, T. et al. Nicotinamide mononucleotide, an intermediate of NAD+ synthesis, protects the heart from ischemia and reperfusion. PLoS One 9 (2014). 0. Yamamoto, T. et al. Nicotinamide mononucleotide, an intermediate of NAD+ synthesis, protects the heart from ischemia and reperfusion. PLoS One 9 (2014). p 41. Dawson, M. E. Maximum Valid Dilution and Minimum Valid Concentration. LAL Updat. 13, 1–3 (1995). 42 Dawson, M E Endotoxin Limits LAL Updat 13, 1–3 (1995) 41. Dawson, M. E. Maximum Valid Dilution and Minimum Valid Concentration. LAL Updat. 13, 1–3 (1995). 42. Dawson, M. E. Endotoxin Limits. LAL Updat. 13, 1–3 (1995). Author Contributionsh The manuscript was written through contributions of all authors. G.C.M. and R.G.P. performed all experiments, analysed and interpreted the data and drafted the manuscript. A.D. and G.C.M. conceived and coordinated the study. All authors revised the manuscript for content and approved the final version for publication. g This study was supported by Independent Research Association, Bucharest - 012416, Romania. g This study was supported by Independent Research Association, Bucharest - 012416, Romania. g This study was supported by Independent Research Association, Bucharest - 012416, Romani g This study was supported by Independent Research Association, Bucharest - 012416, Romania. References l Nicotinamide mononucleotide attenuates brain injury after intracerebral hemorr by activating Nrf2/HO-1 signaling pathway. Sci. Rep. 1–13, https://doi.org/10.1038/s41598-017-00851-z (2017). 21. Plauf, G. W. E. & Plaut, K. A. A Method for the Purifkation of Nicotinamide Mononucleotide. Arch. Biochem. Biophys. 48, 189 (1953). ( ) 22. Preiss, J. & Handler, P. Enzymatic synthesis of nicotinamide mononucleotide. J. Biol. Chem. 225, 759–770 (1957).i 23. Armarego, W. L. F. & Chai, C. L. L. In Purification of Laboratory Chemicals (eds Armarego, W. L. F. & Chai, C. L. L.) 690 (Els 2009). 24. Payne, S. H. & Ames, B. N. A procedure for rapid extraction and high-pressure liquid chromatographic separation of the nucleo and other small molecules from bacterial cells. Anal. Biochem. 123, 151–161 (1982).h 25. Yamada, K., Hara, N., Shibata, T., Osago, H. & Tsuchiya, M. The simultaneous measurement of nicotinamide adenine dinucleotide and related compounds by liquid chromatography/electrospray ionization tandem mass spectrometry. Anal. Biochem. 352, 282–285 (2006). ( ) 6. Trammell, S. A. & Brenner, C. Targeted, LCMS-based Metabolomics for Quantitative Measurement of NAD(+) Metabolites Comput. Struct. Biotechnol. J. 4, e201301012 (2013). p 27. Dong, W.-R. et al. New function for Escherichia coli xanthosine phophorylase (xapA): genetic and biochemical evidences on its participation in NAD(+) salvage from nicotinamide. BMC Microbiol. 14, 29 (2014). g 28. Swadesh, J. K. Ion exchange chromatography in HPLC Practical and Industrial Applications (ed. Swadesh, J. K.), 213–287 (CRC Press LLC, 2001). ) 9. Larson, J. R., Tingstad, J. E., Swadesh, J. K. Chapter one Introduction in HPLC Practical and Industrial Applications (ed. Swadesh, J. K.), 2–49 (CRC Press LLC, 2001). 30. Alpert, A. J. Chapter 8. Size exclusion high-performance liquid chromatography of small solutes in Column Handbook for Size Exclusion Chromatography (ed. Wu, C. S.) 249–266 (Academic Press Inc., 1999). ll h h l f l d d l ( ) g p y ( ) ( ) 1. Cannell, R. How to approach the isolation of a natural product. Nat. Prod. Isol. 4, 1–51 (1998). g p y 31. Cannell, R. How to approach the isolation of a natural product. Nat. Prod. Isol. 4, 1–51 (1998). Scientific REPOrtS \| (2018) 8:4433 \| DOI:10.1038/s41598-018-22806-8 10 Scientific REPOrtS \| (2018) 8:4433 \| DOI:10.1038/s41598-018-22806-8 Additional Informationh Competing Interests: The authors declare no competing interests. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not per- mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. © The Author(s) 2018 Scientific REPOrtS \| (2018) 8:4433 \| DOI:10.1038/s41598-018-22806-8 11
https://openalex.org/W2794355545	https://hal.archives-ouvertes.fr/hal-01721265/document	English	null	A Systems Biology Study in Tomato Fruit Reveals Correlations between the Ascorbate Pool and Genes Involved in Ribosome Biogenesis, Translation, and the Heat-Shock Response	Frontiers in plant science	2,018	cc-by	16,320	A systems biology study in tomato fruit reveals correlations between the ascorbate pool and genes involved in ribosome biogenesis, translation, and the heat-shock response Rebecca Stevens, Pierre Baldet, Jean-Paul Bouchet, Mathilde Causse, Catherine Deborde, Claire Deschodt, Mireille Faurobert, Cecile Garchery, Virginie Garcia, Hélène Gautier, et al. To cite this version: Rebecca Stevens, Pierre Baldet, Jean-Paul Bouchet, Mathilde Causse, Catherine Deborde, et al.. A systems biology study in tomato fruit reveals correlations between the ascorbate pool and genes involved in ribosome biogenesis, translation, and the heat-shock response. Frontiers in Plant Science, 2018, 9 (137), 16 p. 10.3389/fpls.2018.00137. hal-01721265 Distributed under a Creative Commons Attribution 4.0 International License A Systems Biology Study in Tomato Fruit Reveals Correlations between the Ascorbate Pool and Genes Involved in Ribosome Biogenesis, Translation, and the Heat-Shock Response Rebecca G. Stevens 1, Pierre Baldet 2, Jean-Paul Bouchet 1, Mathilde Causse 1, Catherine Deborde 2,3, Claire Deschodt 1, Mireille Faurobert 1, Cécile Garchery 1, Virginie Garcia 2, Hélène Gautier 4, Barbara Gouble 5, Mickaël Maucourt 2,3, Annick Moing 2,3, David Page 5, Johann Petit 2, Jean-Luc Poëssel 1, Vincent Truffault 1 and Christophe Rothan 2 Keywords: ascorbate, cellular signaling, heat-shock response, redox, ribosome biogenesis, translation, tomato (Solanum lycopersicum) HAL Id: hal-01721265 L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. Distributed under a Creative Commons Attribution 4.0 International License ORIGINAL RESEARCH published: 14 February 2018 doi: 10.3389/fpls.2018.00137 Edited by: Atsushi Fukushima, RIKEN, Japan Edited by: Atsushi Fukushima, RIKEN, Japan Reviewed by: Victoriano Valpuesta, University of Málaga, Spain Jedrzej Jakub Szymanski, Weizmann Institute of Science, Israel Changing the balance between ascorbate, monodehydroascorbate, and dehydroascorbate in plant cells by manipulating the activity of enzymes involved in ascorbate synthesis or recycling of oxidized and reduced forms leads to multiple phenotypes. A systems biology approach including network analysis of the transcriptome, proteome and metabolites of RNAi lines for ascorbate oxidase, monodehydroascorbate reductase and galactonolactone dehydrogenase has been carried out in orange fruit pericarp of tomato (Solanum lycopersicum). The transcriptome of the RNAi ascorbate oxidase lines is inversed compared to the monodehydroascorbate reductase and galactonolactone dehydrogenase lines. Differentially expressed genes are involved in ribosome biogenesis and translation. This transcriptome inversion is also seen in response to different stresses in Arabidopsis. The transcriptome response is not well correlated with the proteome which, with the metabolites, are correlated to the activity of the ascorbate redox enzymes—ascorbate oxidase and monodehydroascorbate reductase. Differentially accumulated proteins include metacaspase, protein disulphide isomerase, chaperone DnaK and carbonic anhydrase and the metabolites chlorogenic acid, dehydroascorbate and alanine. The hub genes identiﬁed from the network analysis are involved in signaling, the heat-shock response and ribosome biogenesis. The results from this study therefore reveal one or several putative signals from the ascorbate pool which modify the transcriptional response and elements downstream. Changing the balance between ascorbate, monodehydroascorbate, and dehydroascorbate in plant cells by manipulating the activity of enzymes involved in ascorbate synthesis or recycling of oxidized and reduced forms leads to multiple phenotypes. A systems biology approach including network analysis of the transcriptome, proteome and metabolites of RNAi lines for ascorbate oxidase, monodehydroascorbate reductase and galactonolactone dehydrogenase has been carried out in orange fruit pericarp of tomato (Solanum lycopersicum). The transcriptome of the RNAi ascorbate oxidase lines is inversed compared to the monodehydroascorbate reductase and galactonolactone dehydrogenase lines. Differentially expressed genes are involved in ribosome biogenesis and translation. This transcriptome inversion is also seen in response to different stresses in Arabidopsis. The transcriptome response is not well correlated with the proteome which, with the metabolites, are correlated to the activity of the ascorbate redox enzymes—ascorbate oxidase and monodehydroascorbate reductase. Differentially accumulated proteins include metacaspase, protein disulphide isomerase, chaperone DnaK and carbonic anhydrase and the metabolites chlorogenic acid, dehydroascorbate and alanine. The hub genes identiﬁed from the network analysis are involved in signaling, the heat-shock response and ribosome biogenesis. Edited by: Atsushi Fukushima, RIKEN, Japan The results from this study therefore reveal one or several putative signals from the ascorbate pool which modify the transcriptional response and elements downstream. Correspondence: Rebecca G. Stevens Rebecca.stevens@inra.fr Specialty section: This article was submitted to Plant Systems and Synthetic Biology, a section of the journal Frontiers in Plant Science Specialty section: This article was submitted to Plant Systems and Synthetic Biology, a section of the journal Frontiers in Plant Science Specialty section: This article was submitted to Plant Systems and Synthetic Biology, a section of the journal Frontiers in Plant Science Received: 09 October 2017 Accepted: 24 January 2018 Published: 14 February 2018 Received: 09 October 2017 Accepted: 24 January 2018 Published: 14 February 2018 1 Institut National de la Recherche Agronomique, UR1052, Génétique et Amélioration des Fruits et Légumes, Montfavet, France, 2 Institut National de la Recherche Agronomique, Université de Bordeaux, UMR1332, Biologie du Fruit et Pathologie, Villenave d’Ornon, France, 3 Plateforme Métabolome du Centre de Génomique Fonctionnelle Bordeaux, Centre Institut National de la Recherche Agronomique de Bordeaux, Villenave d’Ornon, France, 4 Institut National de la Recherche Agronomique, UR1115, Plantes et Systèmes de culture Horticoles, Avignon, France, 5 Institut National de la Recherche Agronomique, Université d’Avignon et des Pays du Vaucluse, UMR408 Sécurité et Qualité des Produits d’Origine Végétale, Avignon, France INTRODUCTION A highly connected node (a hub) will share a number of edges with other nodes and perturbation of a regulatory hub will often give rise to highly pleiotropic phenotypes. This sort of approach is particularly useful to study a biological question as it can generate new hypotheses and identify regulatory genes. monodehydroascorbate radical, a relatively stable radical, which can regain an electron to regenerate ascorbate, or lose a second electron to become dehydroascorbate; these three forms are in equilibrium (Bors and Buettner, 1997). Plants also possess ascorbate oxidase enzymes which have roles in oxygen removal and signaling (De Tullio et al., 2007; De Tullio, 2012). Ascorbate oxidase activity has also been shown to have links with sugar metabolism and yield (Garchery et al., 2013), in a similar way to an isoform of monodehydroascorbate reductase (MDHAR), except that the phenotypes are inverted (Truﬀault et al., 2016): plants with reduced MDHAR activity show reduced growth and yield and lower levels of sugars (sucrose and hexose) particularly in leaves but counter-intuitively contain more ascorbate, particularly under light (Haroldsen et al., 2011; Gest et al., 2013a), whereas plants with reduced ascorbate oxidase (AO) activity have improved yield stability and more sucrose and hexose, particularly in leaves. A mitochondrial enzyme controls the ﬁnal stage of ascorbate synthesis, galactonolactone dehydrogenase (GLD) (Alhagdow et al., 2007; Leferink et al., 2009), this enzyme is an integral part of complex I and is required for complex I function independently of ascorbate levels (Schimmeyer et al., 2016). Similarly to AO and MDHAR, manipulation of GLD levels in tomato RNAi lines leads to phenotypes including diﬀerences in ascorbate redox state, sugar metabolism, plant growth and fruit yield (Alhagdow et al., 2007). Modiﬁcation of the ascorbate pool via manipulation of the activity of these enzymes has therefore generated many pleiotropic phenotypes at the physiological level. We hypothesize that the enzymes are aﬀecting the levels of one or several metabolites, closely linked to their activity, which can be seen as a hub, involved in many key cellular processes. Abbreviations: RNAi, RNA interference; GLD, galactonolactone dehydrogenase; AO, ascorbate oxidase; MDHAR, monodehydroascorbate reductase; GO, Gene Ontology; FDR, false discovery rate; WGCNA, Weighted gene correlation network analysis; WT, wild-type; TCA, tricarboxylic acid cycle; HSP70, heat-shock protein family; ATP, adenosine triphosphate. INTRODUCTION The conversion of sucrose into hexose- phosphates is the starting point for glycolysis, the TCA cycle and the oxidative pentose phosphate pathway which provide the tissue with its metabolic and energetic requirements. The fruit therefore represents a complex biological system built on multi-layered interactions between metabolites, proteins and genes which form a network characteristic of the tissue. Induced genetic modiﬁcations are a good way of perturbing the system by varying the level of a gene and its corresponding protein and/or metabolites (network nodes) to evaluate the new steady state of the network, thus pinpointing the interactions that a particular node will have (Bassel et al., 2012; Krouk et al., 2013). A highly connected node (a hub) will share a number of edges with other nodes and perturbation of a regulatory hub will often give rise to highly pleiotropic phenotypes. This sort of approach is particularly useful to study a biological question as it can generate new hypotheses and identify regulatory genes. Ascorbate is a major metabolite in plants including fruits with many functions often associated with its role in electron donation and as a cofactor (Gest et al., 2013b). Tomato has emerged as a model fruit for studying the relationships between ascorbate and fruit physiology. In addition to its nutritional value, ascorbate is a major antioxidant that contributes to fruit tolerance to biotic and abiotic stresses (Malacrida et al., 2006; Davey et al., 2007; Stevens et al., 2008). Studying the interactions between the ascorbate pool and fruit ripening is of particular interest because the vast majority of fresh market tomatoes produced in Europe are harvested at orange or light-red stages, before being fully ripe. At these stages, fruit undergo major modiﬁcations in color, ﬁrmness, sweetness, acidity, and aroma. These changes are under developmental and hormonal control and imply profound alterations in the transcriptome, proteome and metabolome (Klee and Giovannoni, 2011). In a previous study, we showed that chilling stress conditions (<10◦C) that can prevail during postharvest transport and storage aﬀect fruit ﬁrmness, and that this eﬀect is correlated with the ascorbate oxidation status (Stevens et al., 2008). However, the mechanisms by which ascorbate impinges on fruit physiology remain to be deciphered. INTRODUCTION Ascorbate’s functions are profoundly aﬀected by the oxidation state of the ascorbate pool: ascorbate is oxidized to the monodehydroascorbate radical, a relatively stable radical, which can regain an electron to regenerate ascorbate, or lose a second electron to become dehydroascorbate; these three forms are in equilibrium (Bors and Buettner, 1997). Plants also possess ascorbate oxidase enzymes which have roles in oxygen removal and signaling (De Tullio et al., 2007; De Tullio, 2012). Ascorbate oxidase activity has also been shown to have links with sugar metabolism and yield (Garchery et al., 2013), in a similar way to an isoform of monodehydroascorbate reductase (MDHAR), except that the phenotypes are inverted (Truﬀault et al., 2016): plants with reduced MDHAR activity show reduced growth and yield and lower levels of sugars (sucrose and hexose) particularly in leaves but counter-intuitively contain more ascorbate, particularly under light (Haroldsen et al., 2011; Gest et al., 2013a), whereas plants with reduced ascorbate oxidase (AO) activity have improved yield stability and more sucrose and hexose, particularly in leaves. A mitochondrial enzyme controls the ﬁnal stage of ascorbate synthesis, galactonolactone dehydrogenase (GLD) (Alhagdow et al., 2007; Leferink et al., 2009), this enzyme is an integral part of complex I and is required for complex I function independently of ascorbate levels (Schimmeyer et al., 2016). Similarly to AO and MDHAR, manipulation of GLD levels in tomato RNAi lines leads to phenotypes including diﬀerences in ascorbate redox state, sugar metabolism, plant growth and fruit yield (Alhagdow et al., 2007). Modiﬁcation of the ascorbate pool via manipulation of p g p y gy p Ripening fruit are non-photosynthetic sink tissues relying on import of sucrose for providing reducing power, ATP, and precursor molecules necessary for synthesis of fruit-speciﬁc metabolites or proteins. The conversion of sucrose into hexose- phosphates is the starting point for glycolysis, the TCA cycle and the oxidative pentose phosphate pathway which provide the tissue with its metabolic and energetic requirements. The fruit therefore represents a complex biological system built on multi-layered interactions between metabolites, proteins and genes which form a network characteristic of the tissue. Induced genetic modiﬁcations are a good way of perturbing the system by varying the level of a gene and its corresponding protein and/or metabolites (network nodes) to evaluate the new steady state of the network, thus pinpointing the interactions that a particular node will have (Bassel et al., 2012; Krouk et al., 2013). Citation: Stevens RG, Baldet P, Bouchet J-P, Causse M, Deborde C, Deschodt C, Faurobert M, Garchery C, Garcia V, Gautier H, Gouble B, Maucourt M, Moing A, Page D, Petit J, Poëssel J-L, Truffault V and Rothan C (2018) A Systems Biology Study in Tomato Fruit Reveals Correlations between the Ascorbate Pool and Genes Involved in Ribosome Biogenesis, Translation, and the Heat-Shock Response. Front. Plant Sci. 9:137. doi: 10.3389/fpls.2018.00137 February 2018 \| Volume 9 \| Article 137 Frontiers in Plant Science \| www.frontiersin.org 1 Ascorbate Signaling in Tomato Fruit Stevens et al. February 2018 \| Volume 9 \| Article 137 INTRODUCTION These enzymes are dioxygenases and therefore this is an example of a role for ascorbate as an electron donor to iron-dependent oxidoreductases. Ascorbate is also involved in oxidative protein folding in the endoplasmic reticulum and therefore plays a role in protein maturation (Banhegyi et al., 2003; Zito et al., 2012; Szarka and Lorincz, 2014). Dehydroascorbate has been shown to have a role in neuronal energy metabolism via activation of glucose-6-phosphate dehydrogenase thus increasing ﬂux through the pentose phosphate pathway (Puskas et al., 2000; Cisternas et al., 2014). Furthermore, ascorbate has been identiﬁed as a kinase inhibitor (Carcamo et al., 2004) as well as an inhibitor of hexokinase (Fiorani et al., 2000) and its oxidized form can inhibit glyceraldehyde 3-phosphate dehydrogenase leading to an energetic crisis and cell death in cancer cells which uptake more dehydroascorbate via glucose transporters (Yun et al., 2015). These examples have not been shown in plants: the study of ascorbate in non-photosynthetic tissues may be key to revealing additional functions for the molecule; fruit is an ideal tissue, particularly in advanced ripening stages when photosynthetic activity is low. Tomato has emerged as a model fruit for studying the relationships between ascorbate and fruit physiology. In addition to its nutritional value, ascorbate is a major antioxidant that contributes to fruit tolerance to biotic and abiotic stresses (Malacrida et al., 2006; Davey et al., 2007; Stevens et al., 2008). Studying the interactions between the ascorbate pool and fruit ripening is of particular interest because the vast majority of fresh market tomatoes produced in Europe are harvested at orange or light-red stages, before being fully ripe. At these stages, fruit undergo major modiﬁcations in color, ﬁrmness, sweetness, acidity, and aroma. These changes are under developmental and hormonal control and imply profound alterations in the transcriptome, proteome and metabolome (Klee and Giovannoni, 2011). In a previous study, we showed that chilling stress conditions (<10◦C) that can prevail during postharvest transport and storage aﬀect fruit ﬁrmness, and that this eﬀect is correlated with the ascorbate oxidation status (Stevens et al., 2008). However, the mechanisms by which ascorbate impinges on fruit physiology remain to be deciphered. Ripening fruit are non-photosynthetic sink tissues relying on import of sucrose for providing reducing power, ATP, and precursor molecules necessary for synthesis of fruit-speciﬁc metabolites or proteins. INTRODUCTION We have therefore carried out a large-scale study including phenotypic, metabolic, transcriptomic and proteomic data for orange fruit of RNAi lines for the three enzymes (AO, GLD, and MDHAR) shown in Figure 1 in order to identify their links with cellular yp y g y g Ascorbate is a major metabolite in plants including fruits with many functions often associated with its role in electron donation and as a cofactor (Gest et al., 2013b). Interestingly, although many studies in plants focus on the photo-protective or antioxidant functions of ascorbate, studies in animals have revealed further roles for ascorbate, for example the molecule functions as a cofactor for the ten-eleven translocation enzymes, which can convert 5-methylcytosine, to 5-hydroxymethylcytosine (5hmC) during demethylation of DNA in mouse embryonic stem cells (Blaschke et al., 2013). Modiﬁcation of the ascorbate pool via manipulation of the activity of these enzymes has therefore generated many pleiotropic phenotypes at the physiological level. We hypothesize that the enzymes are aﬀecting the levels of one or several metabolites, closely linked to their activity, which can be seen as a hub, involved in many key cellular processes. We have therefore carried out a large-scale study including phenotypic, metabolic, transcriptomic and proteomic data for orange fruit of RNAi lines for the three enzymes (AO, GLD, and MDHAR) shown in Figure 1 in order to identify their links with cellular February 2018 \| Volume 9 \| Article 137 Frontiers in Plant Science \| www.frontiersin.org 2 Ascorbate Signaling in Tomato Fruit Stevens et al. untransformed control. Plants homozygous for the transgene from the T2 generation were used. A total of 15 plants per line were used, 10 for fruit sampling and 5 for destructive phenotypic measurements. All 15 plants were used for non-destructive phenotypic measurements. FIGURE 1 \| Biological system showing the enzymes studied. Schematic diagram of the enzymes studied in tomato using RNAi. GLD, galactonolactone dehydrogenase, a mitochondrial enzyme; AO, ascorbate oxidase, probably apoplastic; MDHAR, monodehydroascorbate reductase, a cytosolic and peroxisomal isoform. Ascorbate is generated in the mitochondrion and is transported into the cell via an unknown mechanism. Ascorbate oxidase oxidizes ascorbate to monodehydroascorbate which will disproportionate into dehydroascorbate and ascorbate. Monodehydroascorbate is recycled to ascorbate by enzymes with monodehydroascorbate reductase activity. Pre- and Post-harvest Phenotypic Measurements Ascorbate is generated in the mitochondrion and is transported into the cell via an unknown mechanism. Ascorbate oxidase oxidizes ascorbate to monodehydroascorbate which will disproportionate into dehydroascorbate and ascorbate. Monodehydroascorbate is recycled to ascorbate by enzymes with monodehydroascorbate reductase activity. FIGURE 1 \| Biological system showing the enzymes studied. Schematic diagram of the enzymes studied in tomato using RNAi. GLD, galactonolactone dehydrogenase, a mitochondrial enzyme; AO, ascorbate oxidase, probably apoplastic; MDHAR, monodehydroascorbate reductase, a cytosolic and peroxisomal isoform. Ascorbate is generated in the mitochondrion and is transported into the cell via an unknown mechanism. Ascorbate oxidase oxidizes ascorbate to monodehydroascorbate which will disproportionate into dehydroascorbate and ascorbate. Monodehydroascorbate is recycled to ascorbate by enzymes with monodehydroascorbate reductase activity. metabolism and to correlate changes in gene expression, proteins and metabolites with the activity of these enzymes. A link is provided between the redox state of the ascorbate pool, cellular protein synthesis and stability, and ribosomal function. Statistical analysis on phenotypic characteristics was carried out using XLstat (Addinsoft, Paris, France) using a non-parametric comparison of means (Kruskal Wallis) with correction (Dunn). Diﬀerent letters in the tables indicate signiﬁcant diﬀerences (5% signiﬁcance level). Greenhouse Plant Culture West Virginia 106 cherry tomato plants were grown in a multi- span Venlo-type greenhouse in 5l pots (potting compost P3 Tref, Tref EGO substrates BV). Water and nutrients were supplied to the plants using a drip irrigation system to maintain 20– 30% drainage. Light intensities of 300–700 photosynthetically active radiation were obtained over the culture period. Flowers were mechanically pollinated three times a week and side shoots removed as they appeared. The 35S-RNAi lines used have been previously described [AO, Solyc04g054690 (Garchery et al., 2013), GLD, Solyc10g079470 (Alhagdow et al., 2007), MDHAR (or MD), Solyc09g009390 (Gest et al., 2013a)]. In each case the speciﬁc transcript levels and enzyme activity were checked and shown to decrease. The individual lines chosen were the following: GLD5.3, AO15.1, and MD5 with a wild-type (WT) Fruit were tagged at anthesis and harvested 20 days later (for ascorbate measurements and proteome analysis only) or when orange. Fruit were harvested during the growing season (from September to November) from trusses 2, 3, and 4 from 10 plants using three fruit/truss/plant to make 3 pools of 30 fruits per line. Fruit pericarp including the epidermis was retained, frozen in liquid nitrogen and stored at −80◦C before being ground to a ﬁne powder in liquid nitrogen for the following analyses. Experimental Design for Transcriptome and Metabolome Analyses Greenhouse Plant Culture Pre- and Post-harvest Phenotypic Measurements Plant height from the pot to the top of the apical meristem was measured weekly for 7 weeks on 10 plants per line. The growth rate was calculated as cm gained per day. The average leaf area of a fully expanded leaf (one leaf per plant, 10 plants) was calculated by scanning the leaf and calculating the area of the image using ImageJ software. Fruit characteristics (number of days between anthesis and orange maturity stage, and fruit weight) were calculated on at least six fruits per plant from 10 plants. Fruit weight, diameter, ﬁrmness measurements (compression and maximal force for skin rupture) and color were measured on a minimum of 20 fruits per line immediately after harvesting (ripe fruits). The weight loss of seven fruits per line, related to conservation properties, was measured after 5 days post- harvest at 20◦C. Texture measurements were performed using a multi-purpose texturometer (TaPlus, Lloyd). The global ﬁrmness of each fruit was measured by a uniaxial compression test with a 50 mm ﬂat disk (test-speed 10 cm.min−1) and expressed as the pressure (kPa) required for a 3% deformation of fruit height. The maximal force (N) corresponding to skin rupture was recorded using a ﬂat cylinder tip (2 mm diameter) for a 7 mm penetration into the equatorial part of the fruit (test- speed 20 mm.min−1). The color was determined using a CR-400 chromameter (Minolta Co. Ltd., Osaka, Japan) and expressed in the CIE 1976 L∗a∗b∗Color Space coordinates (illuminant D65, 0_ view angle, illumination area diameter 8 mm). The chroma a∗ axis is a good indicator for tomato ripening, varying from green (−60) to red (+60). FIGURE 1 \| Biological system showing the enzymes studied. Schematic diagram of the enzymes studied in tomato using RNAi. GLD, galactonolactone dehydrogenase, a mitochondrial enzyme; AO, ascorbate oxidase, probably apoplastic; MDHAR, monodehydroascorbate reductase, a cytosolic and peroxisomal isoform. Ascorbate is generated in the mitochondrion and is transported into the cell via an unknown mechanism. Ascorbate oxidase oxidizes ascorbate to monodehydroascorbate which will disproportionate into dehydroascorbate and ascorbate. Monodehydroascorbate is recycled to ascorbate by enzymes with monodehydroascorbate reductase activity. FIGURE 1 \| Biological system showing the enzymes studied. Schematic diagram of the enzymes studied in tomato using RNAi. GLD, galactonolactone dehydrogenase, a mitochondrial enzyme; AO, ascorbate oxidase, probably apoplastic; MDHAR, monodehydroascorbate reductase, a cytosolic and peroxisomal isoform. Microarray Analysis and Comparison of Transcriptome Data p Microarray experiments were performed on the three biological replicates (3 pools) and technical duplicates with the dyes reversed (dye swap). The TOM2 cDNA microarray [CGE, Boyce Thompson Institute, Cornell University, http://bti.cornell.edu/ CGEP/CGEP.html, (Fei et al., 2011)] containing 12,160 70-mer oligonucleotides corresponding to 11,862 unique and randomly selected transcripts of the tomato genome were used. Sample preparation and microarray hybridization and analysis were carried out as described (Mounet et al., 2009). Brieﬂy, one µg of DNA-free total RNA was ampliﬁed with the MessageAmpTM aRNA Ambion Kit (Ambion/Applied Biosystems), labeled using the CyScribe Post-Labeling Kit (GE Healthcare) and the cy3- and cy5-labeled cDNAs were hybridized to TOM2 microarrays using an automatic hybridization station HS 4800 (Tecan) as described. Microarray slides were scanned with a Genepix 4000 B ﬂuorescence reader (Axon Instruments) using Genepix 4.0 Pro image acquisition software. The photomultiplier tube voltage was adjusted to 640 V for Cy3 and 700 V for Cy5. Spot ﬂagging was carried out by Genepix (missing spots) and by visual inspection of the images in order to exclude saturated and heterogeneous spots. The raw data were then submitted to LIMMA R Package v2.12.0 (Ritchie et al., 2015) for data visualization, normalization as described (Alhagdow et al., 2007). Background was ﬁrst subtracted from the raw median intensity values using the normexp function. Two normalization algorithms were successively applied to the raw data (Yang et al., 2002): print-tip loess (default parameters) as the within-array normalization, followed by a quantile normalization (default parameters) as between-array normalization. The normalized data were then analyzed as described below. To convert to current Solyc identiﬁers, each of the TOM2 probes was blasted against the tomato genome (ITAG2.3, minimum alignment length 40 bp, minimum identity 84%), which had been sequenced in the meantime, and following this 7633 unique Solycs were retained. Individual Metabolites Ascorbate and dehydroascorbate concentrations in fruit powder were analyzed according to previous protocols based on the reduction of an iron-dipyridyl complex leading to a color change detectable spectrophotometrically with a ﬁlter at 525 nm (Stevens et al., 2006) with modiﬁcations (Garchery et al., 2013). y Individual polyphenols were analyzed using 200 mg of lyophilised powder from the orange fruit (3 pools), to which of 25 µL of taxifolin (1 mg.mL−1, Extrasynthèse, Genay France) was added as an internal standard. Three successive extractions with 70% ethanol were carried out before evaporation to dryness in a vacuum concentrator. The pellet was resuspended in 300 µL water, 700 µL methanol and ﬁltered with a PTFE syringe ﬁlter. Ten micro liters were analyzed by reversed-phase HPLC (Shimadzu Prominence system coupled with a Diode Array Detector (HPLC-DAD), Merck LichroCART R⃝250-4 column; phase Superspher R⃝100 RP-18 end capped 4 µm) using a gradient elution with ortho-phosphoric acidiﬁed water (pH 2.6) (A) and methanol (B) as solvents. The program was as follows: 0 min 15% B, 10 min 15% B, 20 min 21% B, 32 min 21% B, 42 min 23% B, 52 min 23% B, 107 min 60% B; ﬂow rate 0.5 mL.min−1 from 0 to 44 min, 0.5 mL.min−1 to 0.7 mL.min−1 from 44 to 46 min, 0.7 mL.min−1 from 46 to 60 min. Phenolic compounds were monitored at 280 nm. Chlorogenic acid and rutin were quantiﬁed using external calibration curves of standard compounds (Extrasynthèse, Genay, France); other compounds were expressed in arbitrary units. NMR Metabolomic Proﬁling Polar metabolites were extracted from lyophilized powder with a series of hot ethanol/water solutions and quantiﬁed by 1H- NMR as previously described (Mounet et al., 2007). For extract preparation and NMR acquisition parameters, special care was February 2018 \| Volume 9 \| Article 137 Frontiers in Plant Science \| www.frontiersin.org 3 Ascorbate Signaling in Tomato Fruit Stevens et al. taken to allow absolute quantiﬁcation of individual metabolites through addition of 5 mM ethylene diamine tetra-acetic acid disodium salt solution to improve the NMR signal resolution and quantiﬁcation of organic acids, adequate choice of the NMR acquisition parameters, and use through an electronic reference for quantiﬁcation (Moing et al., 2004). Unknown compounds were quantiﬁed in arbitrary units. The lyophilized extracts were titrated with KOD to apparent pH 6 in 400 mM potassium phosphate buﬀer solution in D2O and lyophilized again. The 1H-NMR spectra were recorded on each dried titrated extract solubilized in 0.5 mL D2O with the sodium salt of (trimethylsilyl)propionic-2,2,3,3-d4 acid (TSP) at a ﬁnal concentration of 0.01% for chemical shift calibration, at 500.162 MHz on an Avance spectrometer (Bruker BioSpin, Wissembourg, France) using a 5 mm inverse probe. Sixty-four scans of 32K data points were acquired with a spectral width of 6000 Hz, acquisition time of 2.73 s, 90◦pulse angle and recycle delays of 25 s. The raw 1H NMR spectral proﬁles have been deposited, with associated metadata, into the Metabolomics Repository of Bordeaux MeRy-B (http://services.cbib.u-bordeaux2.fr/MERYB/ public/PublicREF.php?REF=T07001). staining, image analysis was performed with Samespot software (version 4.1), and the normalized spot volumes were obtained. Of the 108 spots revealed, 49 were eliminated as not being variable. Protein identiﬁcation of the 59 variable spots was performed at the proteome platform of Le Moulon (Gif-sur- Yvette) using a nano-LC-MS/MS method following a previously described procedure (Xu et al., 2013). The database search was run against the tomato unigene database version 4 held by the Solanaceae Genome Network (http://solgenomics.net/; Tomato_200607_build_2) with X!Tandem software (http://www. thegpm.org/TANDEM/, version 2010.12.01.1). Sequences were subsequently converted to Solycs (ITAG version 2.4) by Blast analysis. This conversion was conﬁrmed by re-analyzing the mass spectra or by searching for similarity between the protein peptide sequences and the ITAG2.4 version of the genome by TBLASTN. Proteome Analysis Protein extraction, two-dimensional gel electrophoresis, and protein identiﬁcation and classiﬁcation were carried out as previously described (Xu et al., 2013). Brieﬂy, proteins were extracted using the phenol extraction method developed by Faurobert et al. (2007). Proteins were then separated by two-dimensional gel electrophoresis. After Coomassie colloidal Gene Ontology Gene ontology (GO) enrichment on the whole transcriptome signatures generated from pairwise comparisons of the lines was carried out using the R package “goseq” (Young et al., 2010) with the corrected p-value false detection rate set to 5%. The package requires the log of fold-change and p-value from the R/maanova analysis. No length correction was required as the data was from microarrays. Tomato gene ontology for the 182 diﬀerentially expressed genes was carried out using the online tools available from the Boyce Thompson Institute, Cornell University (http://bioinfo.bti.cornell.edu/cgi-bin/MetGenMAP/ home.cgi, Boyle et al., 2004). R/maanova Diﬀerential expression and transcriptome comparisons were carried out using the R/maanova package [https://www. bioconductor.org/packages/release/bioc/html/maanova.html (R February 2018 \| Volume 9 \| Article 137 Frontiers in Plant Science \| www.frontiersin.org 4 Ascorbate Signaling in Tomato Fruit Stevens et al. transcriptome data, chosen for its ROS signature, and the provided data of interest are calculated. The stored data are microarray-based data from wild type and Arabidopsis mutants for the following genes: ascorbate peroxidase, the ﬂu mutant, mitochondrial alternative oxidase, catalase 2 and superoxide dismutase subjected or not to the following stresses: ozone, high light, hydrogen peroxide, drought, rotenone, methyl viologen and 3-aminotriazole as described in the original research paper. The data from the RNAi lines normalized to wild-type (fold change and p-value) were compared with this series of transcriptomes arising from diﬀerent conditions or mutations in Arabidopsis thaliana. In order to compare the transcriptomes, the tomato Solycs were ﬁrst converted to their closest Arabidopsis homolog and submitted to the ROSMETER interface. In the heatmaps, red represents a positive correlation and green a negative one. Development-Core-Team, 2015)]. Means and standard errors were calculated for each gene in each line. A simple model taking into account the genotype (line) was used. A matrix of comparisons was used for the pairwise comparison of the transcriptome datasets ﬁrst of each transgenic line with WT and then with each other. For each comparison, a threshold for the FDR corrected p-value of 0.001 was applied. Following initial tests, we added a sub-hypothesis to the comparison matrix to test whether the AO-GLD transcriptome and the AO-MDHAR transcriptomes were inverted i.e., the wild-type expression was the average of the gene expression in the two transgenic lines being compared. Expression diﬀerences were expressed as log2 of the Fold Change between the two lines. The linear model was tested using 1,000 iterations and then an FDR adjustment of the p-values obtained. A threshold for the corrected p-value of 0.001 was set to identify the core gene set for the network visualization. Transgenic Lines Show Slight Differences in Growth and Development of Plants and Fruit Weighted gene correlation network analysis (WGCNA) was carried out using the normalized expression data for the 7633 genes with a “trait” ﬁle including the 22 metabolites and 40 proteins using the WGCNA package in “R” (Langfelder and Horvath, 2008). No outliers were detected after sample clustering so the whole dataset was used. A soft threshold of 14 and minimum module size of 25 were set. Networks were visualized using Cytoscape version 3.2.0 (Cline et al., 2007) (www.cytoscape.org) using the node and edge output ﬁle from the WGCNA package with a threshold set to 0.52 (to ﬁx a number of nodes and edges suitable for visualization). The dataset used contained the 182 diﬀerentially expressed genes (diﬀerentially expressed in the pairwise comparisons of the three types of transgenic lines), the 22 metabolites and 40 proteins. The number of genes was limited to facilitate network creation and to give weight to the proteins and metabolites. The layout was ﬁrst designated as degree sorted circle layout and then a degree layout was chosen to aid node identiﬁcation. Node size is proportional to the degree of connectivity. The three RNAi transgenic lines (AO: ascorbate oxidase; GLD: galactonolactone dehydrogenase and MDHAR: monodehydroascorbate reductase) and wild-type plants grown in the greenhouse were subjected to simple phenotypic analyses to measure plant height, growth rate and ﬁnal leaf area, the duration of fruit ripening (anthesis to orange fruit) and fruit weight were also measured. The results for the plant phenotypes are shown in Table 1: AO and GLD RNAi lines had a slower growth rate and a smaller ﬁnal plant height compared to wild-type plants. Fruit from the same two transgenic lines also took longer to reach maturity (Table 2). Fruit from the MDHAR silenced lines were signiﬁcantly smaller than fruit from the other lines, including wild-type (Tables 2, 3), as per previous results (Truﬀault et al., 2016). Fruits from the GLD RNAi line were also smaller in diameter (Table 3). Harvested fruits were analyzed for fruit ﬁrmness, color, and post-harvest weight loss (Table 3). No signiﬁcant diﬀerences were found between the lines in terms of compression or color but the AO and MDHAR lines required less force to penetrate the skin compared to wild type. MDHAR RNAi fruit also showed greater post-harvest weight-loss compared to the WT, AO, and GLD lines. Metabolic Differences Are Seen in Phenolic and Amino Acid Metabolism Metabolome analysis was carried out on the fruit pericarp samples using 1H-NMR proﬁling of polar extracts. Ascorbate (Table 4) and major fruit polyphenols (Table 5) were analyzed by speciﬁc methods. Ascorbate was signiﬁcantly increased in orange fruit pericarp of MDHAR lines as previously observed (Gest et al., 2013a) and dehydroascorbate was signiﬁcantly diﬀerent in AO and MDHAR lines, although this diﬀerence was the opposite of what would be expected given the activities of the enzymes; GLD and WT had intermediate values. Ascorbate data for fruit 20 days after anthesis is presented in Supplementary Table S1, unlike ROSMETER The ROSMETER bioinformatic tool allows evaluation of transcriptome signatures linked to reactive oxygen species (ROS) (Rosenwasser et al., 2013). The tool consists of a web-based interface whereby vector-based correlations between stored February 2018 \| Volume 9 \| Article 137 Frontiers in Plant Science \| www.frontiersin.org 5 Ascorbate Signaling in Tomato Fruit Stevens et al. TABLE 1 \| Plant phenotypes of transgenic lines and WT. Height after 7 days (cm) SE p Height after 7 weeks (cm) SE p Starting growth rate (cm/day) SE p Final growth rate (cm/day) SE p Average leaf area (cm2) SE p WT 26.2 3.7 a 193.0 9.0 b 2.7 0.2 b 5.84 0.32 b 313 46 a AO 23.7 1.8 a 181.8 4.9 ab 2.5 0.2 ab 5.74 0.27 ab 296 49 a GLD 24.5 2.3 a 171.3 6.3 a 2.4 0.1 a 5.22 0.31 a 348 10 a MDHAR 25.3 2.2 a 191.8 14.0 b 2.7 0.3 b 6.06 0.79 b 326 37 a Measurements were carried out on greenhouse-grown plants as described in the Materials and methods. Data show means from 10 plants per line with standard error (SE). A comparison of means was carried out using a Kruskal Wallis test with correction (Dunn). Different letters indicate signiﬁcant differences (5% signiﬁcance level). TABLE 1 \| Plant phenotypes of transgenic lines and WT. TABLE 2 \| Fruit phenotypes of transgenic lines and WT. Days to orange SE p Individual fruit weight (g) SE p % Dry weight (orange) SE p WT 40.3 2.1 a 6.3 1.2 a 7.1 0.1 a AO 41.9 1.5 b 6.1 0.8 a 7.0 0.1 a GLD 42.7 1.4 b 6.0 1.0 a 6.8 0.1 a MDHAR 40.7 1.8 a 3.8 0.9 b 7.0 0.1 a Measurements were carried out on greenhouse-grown plants as described in the Materials and methods. Data show means from at least 6 fruits per plant of 10 plants per line with standard error (SE). A comparison of means was carried out using a Kruskal Wallis test with correction (Dunn). Different letters indicate signiﬁcant differences (5% signiﬁcance level). The proteome of the GLD line did not show signiﬁcant diﬀerences to the wild-type, despite visually resembling the ascorbate oxidase proﬁle. ROSMETER A number of proteins showed opposite regulation in AO and MDHAR lines, these included chaperone DnaK (MDHAR down-regulated, AO up-regulated), a probable carbonic anhydrase, a protein disulphide isomerase and a metacaspase 7, a protein with cysteine endopeptidase activity and a positive mediator of cell death which is AO down-regulated and MDHAR up-regulated. One protein was up-regulated in the AO line which is an F1F0-ATPase inhibitor protein. The data for fruit 20 days after anthesis is shown in Supplementary Figure S1: only two proteins showed statistically signiﬁcant diﬀerences in spot volumes between wild-type and the MDHAR line. TABLE 2 \| Fruit phenotypes of transgenic lines and WT. Weighted-Gene Correlation Network Analysis Reveals Clusters of Genes Associated with Both Metabolites and Proteins orange fruit, no signiﬁcant diﬀerences are seen. The polyphenol that varied most in the transgenic lines was chlorogenic acid, which is also the most prevalent polyphenol, and its stereoisomer cis-chlorogenic acid: as for dehydroascorbate, levels were signiﬁcantly diﬀerent in AO and MDHAR lines, AO lines containing less chlorogenic acid than wild-type whereas the MDHAR lines contained more (Table 5 and Figure 2). Levels of these polyphenols in the GLD RNAi fruit were not signiﬁcantly diﬀerent to wild-type. The metabolites detected and quantiﬁed by 1H-NMR are shown in Figure 2 (and Supplementary Table S2). Metabolites that diﬀerentiate the lines included the amino acids alanine and tyrosine which were highest in MDHAR and lowest in AO lines (signiﬁcant for alanine) and the hexose sugars glucose and fructose which showed a tendency in contrast to the aforementioned amino acids to be highest in AO lines. The normalized transcriptome data for the four lines was subjected to a weighted-gene correlation network analysis to identify clusters of correlated genes and compare them with diﬀerent phenotypes (metabolite and protein levels). We identiﬁed 38 modules of between 26 and 851 genes (Supplementary Figure S2). The number of genes in each module was too small to give signiﬁcant diﬀerences following GO enrichment analysis but revealed co-regulation of metabolites or proteins with genes in modules or groups of modules. For example, the amino acids were positively correlated with the gene modules “dark red,” “blue,” “green-yellow” and “yellow.” The full data set with module membership is found in Supplementary Table S4. The Subset of 182 Genes Deﬁnes a Regulatory Network with Several Hub Genes Including Heat-Shock Transcription Factors, Chaperones, and Ribosomal Proteins expression between the lines and the associated p-value (−log10 transformed) are shown in Figure 4. The AO-MDHAR and AO- GLD comparisons generated the highest number of diﬀerentially expressed genes (comparisons 4; 290 genes and comparison 6; 330 genes). The comparisons WT = (transgenic X + transgenic Y)/2 are comparisons 5, 8, and 9. If the hypothesis is true then the p-values become non-signiﬁcant and the −log10 of the p- values gives low values which is the case for the comparisons 5 [WT = (AO + MDHAR)/2] and 8 [WT = (AO + GLD)/2] but not the case for the comparison 9 [WT = (GLD + MDHAR)/2]. The whole dataset including means, standard error, p and F- values and the fold-change with p-values for each comparison is available in Supplementary Table S5. expression between the lines and the associated p-value (−log10 transformed) are shown in Figure 4. The AO-MDHAR and AO- GLD comparisons generated the highest number of diﬀerentially expressed genes (comparisons 4; 290 genes and comparison 6; 330 genes). The comparisons WT = (transgenic X + transgenic Y)/2 are comparisons 5, 8, and 9. If the hypothesis is true then the p-values become non-signiﬁcant and the −log10 of the p- values gives low values which is the case for the comparisons 5 [WT = (AO + MDHAR)/2] and 8 [WT = (AO + GLD)/2] but not the case for the comparison 9 [WT = (GLD + MDHAR)/2]. The whole dataset including means, standard error, p and F- values and the fold-change with p-values for each comparison is available in Supplementary Table S5. The 182 core gene set was used in network analysis with the metabolites and the protein dataset. A topological overlap matrix using the WGCNA script was produced and an edge ﬁle for network visualization using Cytoscape. The network visualized with a degree-sorted circle layout, is shown in Figure 5 with the top eleven hub genes. The most highly connected genes included two or three ribosomal proteins, an observation which is compatible with those above, and two heat-shock proteins (including DnaJ) and a heat-shock transcription factor. Pairwise Comparison of Transcriptome Data Reveals the Same Set of Differentially Regulated Genes in the Comparison AO-MDHAR and AO-GLD Measurements show means with standard error (SE). A comparison of means was carried out using a Kruskal Wallis test with correction (Dunn). Different letters indicate signiﬁcant differences (5% signiﬁcance level). a* is the name of the color scale. TABLE 4 \| Ascorbate and dehydroascorbate levels in orange fruit pericarp of transgenic lines and WT. Reduced AsA (mg/100gfwt) SE p DHA (mg/100gfwt) SE p WT 15.46 1.30 a 2.71 1.39 ab AO 18.99 1.97 ab 4.73 1.32 a GLD 19.19 0.65 ab 2.83 0.97 ab MDHAR 21.72 1.05 b 0 0.68 b Ascorbate and dehydroascorbate were assayed in the pericarp tissue (3 biological replicates of 30 fruits per pool) of the lines by a spectrophotometric method. Measurements show means with standard error (SE). A comparison of means was carried out using a Kruskal Wallis test with correction (Dunn). Different letters indicate signiﬁcant differences (5% signiﬁcance level). biogenesis and translation (GO:00422 and GO:00064). These genes were up-regulated in the AO line and down-regulated in the MDHAR and GLD RNAi lines. The core set of 182 genes were analyzed using the Solanaceae Genome Network GO-enrichment tool and this analysis gave further categories associated with translation, ribosomes, unfolded protein binding, GTPase regulator activity and nucleotide binding. The categories “structural constituent of ribosome,” “small GTPase regulator activity” and “oxidoreductase activity acting on NADH or NADPH” were more than ten-fold enriched in the cluster compared to the genome (Table 7). TABLE 4 \| Ascorbate and dehydroascorbate levels in orange fruit pericarp of transgenic lines and WT. The Subset of 182 Genes Deﬁnes a Regulatory Network with Several Hub Genes Including Heat-Shock Transcription Factors, Chaperones, and Ribosomal Proteins The Subset of 182 Genes Deﬁnes a Regulatory Network with Several Hub Genes Including Heat-Shock Transcription Factors, Chaperones, and Ribosomal Proteins Pairwise Comparison of Transcriptome Data Reveals the Same Set of Differentially Regulated Genes in the Comparison AO-MDHAR and AO-GLD The diﬀerentially expressed protein spots led to the identiﬁcation of 34 proteins by mass spectrometry (Table 6). Diﬀerences between the lines are presented as a heatmap of the log2 of the ratio with the WT (Figure 3; Table 6, raw data in Supplementary Table S3). The line most aﬀected in terms of its fruit proteome is MDHAR, for which up-regulated proteins included a 60S ribosomal protein, a class I heat-shock protein and a cysteine protease inhibitor. Down-regulated proteins included a heat-shock protein, (hsp70 homolog), lactoylglutathione lyase and an amidase hydantoinase/carbamoylase family protein. Each of the three transgenic lines was compared with wild- type and then the lines were compared two by two. For the pairwise comparisons of the transgenic lines, the hypothesis that the gene expression in the wild-type was the average of the gene expression in each transgenic line was tested using the maanova model (see Materials and methods). Volcano plots for each pairwise comparison showing the log2 of the fold-change in gene February 2018 \| Volume 9 \| Article 137 Frontiers in Plant Science \| www.frontiersin.org 6 Ascorbate Signaling in Tomato Fruit Stevens et al. TABLE 3 \| Fruit physical characteristics of transgenic lines and WT at harvest. Fruit weight (g) SE p Diameter (mm) SE p Compression (kPa) SE p Max force—skin rupture (N) SE p Color (a) SE p Postharvest weight loss (%) SE p WT 9.3 0.4 a 25.7 0.4 a 49.5 2.5 a 5.7 0.2 a 10.3 1.0 a 0.043 0.001 a AO 8.3 0.4 a 24.4 0.5 ab 43.1 3.0 a 4.6 0.2 b 13.8 1.6 a 0.049 0.003 ab GLD 7.8 0.4 a 23.6 0.5 b 49.4 4.0 a 5.4 0.3 ab 13.1 1.7 a 0.046 0.001 a MDHAR 5.3 0.3 b 21.1 0.4 c 49.9 2.6 a 4.6 0.3 b 15.7 1.7 a 0.063 0.002 bc Fruit physical characteristics were measured on red-ripe fruit of at least 20 fruit per line as described in the Materials and methods. Measurements show means with standard error (SE). A comparison of means was carried out using a Kruskal Wallis test with correction (Dunn). Different letters indicate signiﬁcant differences (5% signiﬁcance level). a is the name of the color scale. TABLE 3 \| Fruit physical characteristics of transgenic lines and WT at harvest. Fruit physical characteristics were measured on red-ripe fruit of at least 20 fruit per line as described in the Materials and methods. The Differentially Expressed Genes Are Enriched for Functions Involving Translation and Ribosome-Biogenesis Our tomato transcriptome data was compared using ROSMETER, a bioinformatic tool which unites transcriptome data from diﬀerent stress experiments or mutants in Arabidopsis linked to speciﬁc ROS production in speciﬁc organelles (see Materials and methods for more information). Its use is based on the hypothesis that the chemical identity and sub-cellular localization of ROS leave a speciﬁc imprint on the transcriptomic response (Rosenwasser et al., 2013). We used the RNAi AO and RNAi MDHAR datasets to compare with the Arabidopsis transcriptomes. Our ﬁrst observation is that independently of the g The AO-MDHAR and AO-GLD comparisons generated the highest number of diﬀerentially expressed genes and the expression in the WT is always intermediate. Comparison of these two gene sets showed 182 genes common to each comparison, deﬁning a core set of genes for further analysis. Independently of the comparison and the direction of change, diﬀerentially expressed genes were part of one “molecular function” category: structural constituent of ribosome (GO:00037) and two “biological process” categories: ribosome February 2018 \| Volume 9 \| Article 137 Frontiers in Plant Science \| www.frontiersin.org 7 Ascorbate Signaling in Tomato Fruit Stevens et al. TABLE 5 \| Polyphenols in orange fruit pericarp of transgenic lines and WT. Chlorogenic acid (µg/gdwt) SE p Caffeic acid glucoside (arbitrary units) SE p Cis-chlorogenic acid (arbitrary units) SE p Quercetin derivative (arbitrary units) SE p Rutin (µg/gdwt) SE p WT 3245.2 31.2 ab 3.5 0.0 a 1.3 0.0 ab 0.4 0.00 a 435.5 2.9 a AO 2537.5 81.5 a 3.0 0.2 a 1.0 0.1 a 0.4 0.02 a 397.1 8.1 a GLD 3029.0 298.5 ab 3.2 0.5 a 1.1 0.1 ab 0.4 0.01 a 406.0 21.4 a MDHAR 3554.4 68.3 b 3.7 0.2 a 1.5 0.1 b 0.3 0.01 a 446.7 21.3 a Polyphenolic compounds were assayed in the pericarp tissue (3 biological replicates of 30 fruits per pool) of the lines by HPLC-DAD. When the puriﬁed compound was not available for exact quantiﬁcation, arbitrary units corresponding to peak surface were used. Measurements show means with standard error (SE). A comparison of means was carried out using a Kruskal Wallis test with correction (Dunn). Different letters indicate signiﬁcant differences (5% signiﬁcance level). TABLE 5 \| Polyphenols in orange fruit pericarp of transgenic lines and WT. Polyphenolic compounds were assayed in the pericarp tissue (3 biological replicates of 30 fruits per pool) of the lines by HPLC-DAD. The Differentially Expressed Genes Are Enriched for Functions Involving Translation and Ribosome-Biogenesis When the puriﬁed compound was not available for exact quantiﬁcation, arbitrary units corresponding to peak surface were used. Measurements show means with standard error (SE). A comparison of means was carried out using a Kruskal Wallis test with correction (Dunn). Different letters indicate signiﬁcant differences (5% signiﬁcance level). FIGURE 2 \| Accumulation of metabolites in orange pericarp fruit (3 biological replicates of 30 fruits) of the transgenic lines and wild-type. All metabolites except for ascorbate, dehydroascorbate and the polyphenols which were analyzed by speciﬁc methods, were analyzed by 1H-NMR. (A) Heatmap showing a log2 of the metabolite ratio with the wild-type for each transgenic line. The scale goes from green (metabolite decreased compared to wild-type) to purple (metabolite increased). A comparison of the means was carried out using a Kruskal Wallis test with correction (Dunn). Different letters indicate signiﬁcant differences (5% signiﬁcance level). (B) Boxplots showing raw data for three differentially accumulated metabolites in the transgenic lines and wild-type: alanine, dehydroascorbate and chlorogenic acid. A comparison of the means was carried out using a Kruskal Wallis test with correction (Dunn). Different letters indicate signiﬁcant differences (5% signiﬁcance level). FIGURE 2 \| Accumulation of metabolites in orange pericarp fruit (3 biological replicates of 30 fruits) of the transgenic lines and wild-type. All metabolites except for ascorbate, dehydroascorbate and the polyphenols which were analyzed by speciﬁc methods, were analyzed by 1H-NMR. (A) Heatmap showing a log2 of the metabolite ratio with the wild-type for each transgenic line. The scale goes from green (metabolite decreased compared to wild-type) to purple (metabolite increased). A comparison of the means was carried out using a Kruskal Wallis test with correction (Dunn). Different letters indicate signiﬁcant differences (5% signiﬁcance level). (B) Boxplots showing raw data for three differentially accumulated metabolites in the transgenic lines and wild-type: alanine, dehydroascorbate and chlorogenic acid. A comparison of the means was carried out using a Kruskal Wallis test with correction (Dunn). Different letters indicate signiﬁcant differences (5% signiﬁcance level). DISCUSSION stress and/or organelle, the transcriptome responses are inversed when these two lines are compared (Figure 6). The particular conditions to be highlighted include mitochondrial stress: in this case, the application of rotenone gives a strong negative correlation with the RNAi MDHAR and a positive correlation with RNAi AO 3 h after application. Negative correlations are also found for RNAi MDHAR with Arabidopsis catalase mutants after 3 and 8 h of high light, whereas the correlations are positive for RNAi AO. Positive correlations between RNAi MDHAR exist with 30 min of methyl viologen and generally the knockout ascorbate peroxidase mutant subjected to high light. Again for RNAi AO, in contrast to RNAi MDHAR, a negative correlation is found for 30 min of methyl viologen treatment. This study in fruit tissue has revealed some interesting correlations which might provide clues for further research on ascorbate function but some caveats should be pointed out: here we worked on ripening fruit, a non-photosynthetic tissue, where small diﬀerences in ascorbate content are observed between the various transgenic lines. The information provided will help decipher the relationships between ascorbate and fruit physiology, especially in fruit subject to biotic and abiotic stresses. However, it remains to be seen if the correlations are more general and an initial question would be to see if they are transferable to other tissues. As a 35S promoter was used it is February 2018 \| Volume 9 \| Article 137 Frontiers in Plant Science \| www.frontiersin.org 8 Ascorbate Signaling in Tomato Fruit Stevens et al. TABLE 6 \| Proteins in orange fruit pericarp of the lines as identiﬁed by mass spectrometry. Spot ID ITAG2.4 Annotated function Signiﬁcant difference in lines Fold change of transcript AO vs. DISCUSSION MDHAR p-value (FDR) Correlation transcript-p CM148 Solyc01g099190.2.1 Lipoxygenase No data No data MP25 Solyc01g103450.2.1 Chaperone DnaK MDHAR down, AO up MP26 Solyc01g104920.2.1 26S protease regulatory subunit 8 homolog −1.84 0.002 MP14 Solyc01g111300.2.1 Cold shock protein-1 MP31 Solyc02g070510.2.1 Proteasome subunit alpha type No data No data MB16 Solyc02g080630.2.1 Lactoylglutathione lyase MDHAR down MP18 Solyc02g085790.2.1 T-complex protein 1 subunit zeta 5.82 0.005 MP38 Solyc02g091100.2.1 Oxalyl-CoA decarboxylase MP42 Solyc02g092670.1.1 Subtilisin-like protease No data No data JC65 Solyc03g097270.2.1 Cysteine proteinase inhibitor MDHAR up −1.40 0.017 Yes MP24 Solyc04g045340.2.1 Phosphoglucomutase MP29 Solyc04g045340.2.1 Phosphoglucomutase CM57 Solyc05g005490.2.1 Carbonic anhydrase MDHAR up, AO down 1.98 0.008 No MP17 Solyc05g012580.1.1 Unknown Protein MP11 Solyc05g054580.2.1/ Solyc09g091520.1.1 60S acidic ribosomal protein P0 MDHAR up −1.34 0.023 Yes JC26 Solyc06g005160.2.1 Ascorbate peroxidase 1.25 0.036 MP13 Solyc06g005940.2.1 Protein disulﬁde isomerase MP33 Solyc06g005940.2.1 Protein disulﬁde isomerase JC60 Solyc06g059740.2.1 Alcohol dehydrogenase 2 1.49 0.004 TP17 Solyc06g059740.2.1 Alcohol dehydrogenase 2 1.49 0.004 MP20 Solyc06g060290.2.1 Protein disulﬁde isomerase MDHAR up, AO down No data No data MP36 Solyc06g065270.2.1 Adenylate kinase TP55 Solyc06g076520.1.1 Class I heat shock protein MDHAR up CM94 Solyc07g005820.2.1 Heat shock protein 70 MDHAR down MP23 Solyc07g066600.2.1 Phosphoglycerate kinase −1.36 0.033 MP22 Solyc08g014130.2.1 2-isopropylmalate synthase 1 2.05 0.000 JC79 Solyc08g014130.2.1 2-isopropylmalate synthase 1 2.05 0.000 MP21 Solyc08g062660.2.1 Ran GTPase binding protein No data No data MP41 Solyc09g082060.2.1 Cysteine synthase MP37 Solyc09g083410.2.1 Amidase hydantoinase/carbamoylase family protein MDHAR down No data No data MP19 Solyc09g090330.2.1 Harpin binding protein 1 No data No data MP12 Solyc09g098150.2.1 Metacaspase 7 MDHAR up, AO down 2.01 0.003 No MP16 Solyc09g098150.2.1 Metacaspase 7 2.01 0.003 MP32 Solyc09g098150.2.1 Metacaspase 7 2.01 0.003 MP34 Solyc10g081240.1.1 Protein grpE MP28 Solyc10g086580.1.1 Ribulose-1 5-bisphosphate carboxylase/oxygenase activase 1 MP27 Solyc11g006970.1.1 Unknown protein DS12 from 2D-PAGE of leaf, chloroplastic No data No data MP30 Solyc11g068510.1.1 F1F0-ATPase inhibitor protein AO up No data No data MP35 Solyc11g069000.1.1 T-complex protein 1 subunit beta −1.46 0.035 MP15 Solyc12g056230.1.1 Glutathione peroxidase No data No data The predicted Solyc (https://solgenomics.net/organism/Solanum_lycopersicum/genome; version 2.4 of the International Tomato Annotation Group, ITAG) is shown with the function and signiﬁcant difference between lines where applicable (5% signiﬁcance level). The fold-change and p-value of the corresponding transcript, when present on the m and for p-values of < 0.05, is also shown for the AO-MDHAR lines and whether this fold-change is correlated with the difference seen in protein levels. DISCUSSION l ibl th t th ﬀ t i f it th lt f 2017) Th f il bl f it iﬁ t (F TABLE 6 \| Proteins in orange fruit pericarp of the lines as identiﬁed by mass spectrometry. The predicted Solyc (https://solgenomics.net/organism/Solanum_lycopersicum/genome; version 2.4 of the International Tomato Annotation Group, ITAG) is shown with the annotated function and signiﬁcant difference between lines where applicable (5% signiﬁcance level). The fold-change and p-value of the corresponding transcript, when present on the microarray and for p-values of < 0.05, is also shown for the AO-MDHAR lines and whether this fold-change is correlated with the difference seen in protein levels. 2017). The use of available fruit-speciﬁc promoters (Fernandez et al., 2009) and indeed these genes are expressed in diﬀerent plant organs and fruit tissues (Supplementary Figures S3, S4) also possible that the eﬀects we see in fruit are the result of silencing the genes in another part of the plant (for an example of where this is true see our recent results by Truﬀault et al., also possible that the eﬀects we see in fruit are the result of silencing the genes in another part of the plant (for an example of where this is true see our recent results by Truﬀault et al., February 2018 \| Volume 9 \| Article 137 Frontiers in Plant Science \| www.frontiersin.org 9 Ascorbate Signaling in Tomato Fruit Stevens et al. FIGURE 3 \| Heatmap showing protein levels in orange pericarp fruit (3 biological replicates of 30 fruits) of the transgenic lines and wild-type. All proteins were separated by two-dimensional gel electrophoresis and identiﬁed by mass spectrometry. The log2 of the protein ratio with the wild-type for each transgenic line is presented. The scale goes from green (protein decreased compared to wild-type) to purple (protein increased). A comparison of the means was carried out using a Kruskal Wallis test with correction (Dunn). Different letters indicate signiﬁcant differences (5% signiﬁcance level). The proteins are identiﬁed by their Spot reference ID, the correspondence for which can be found in Table 6. genes to be up-regulated when ascorbate oxidase is down- regulated. The transcriptome inversion is also seen in datasets from Arabidopsis provided by the ROSMETER tool. Ascorbate Has Links with Translation and Protein Folding The transcriptional signature and network analysis suggest a role for ascorbate in control of translation and protein folding, at least in the non-photosynthetic tissue studied. Some evidence from the literature has established links between ascorbate and these cellular mechanisms: a good example is the posttranscriptional negative feedback control that ascorbate exerts over its synthesis via the GDP-l-galactose phosphorylase enzyme (Laing et al., 2015) which appears to aﬀect ribosome density on the 5′UTR of the mRNA. Ribosomes and translation are also regulated by stress and redox processes (Halbeisen and Gerber, 2009; Khandal et al., 2009; Mustroph et al., 2009; Gerashchenko et al., 2012; Gismondi et al., 2014; Moore et al., 2016). Protein folding is DISCUSSION The log2 of the protein ratio with the wild-type for each transgenic line is presented. The scale goes from green (protein decreased compared to wild-type) to purple (protein increased). A comparison of the means was carried out using a Kruskal Wallis test with correction (Dunn). Different letters indicate signiﬁcant differences (5% signiﬁcance level). The proteins are identiﬁed by their Spot reference ID, the correspondence for which can be found in Table 6. would aid in speciﬁcally attributing the eﬀects we observed to the fruit. A ﬁnal caveat is that the correlations we observed are based on microarray data: the 7600 gene tomato microarray is considerably enriched in genes expressed in ripening fruit (Alba et al., 2005; Fei et al., 2011) and therefore well adapted to the purpose of the study. However, global transcriptome analysis using RNA-seq would additionally allow the detection of weakly expressed genes and thus strengthen our conclusions. DISCUSSION The switch seen between AO and MDHAR RNAi lines is capable of distinguishing the conditions found in some of the ROSMETER transcriptome data, although there are limits to this analysis because correlations can be based on a small number of genes (a minimum of 45 according to the original paper). The ROSMETER analysis does not reveal links with hydrogen peroxide metabolism because correlations between our lines and the public data are inversed when comparing ascorbate peroxidase and catalase mutants, both of which in theory should contain more hydrogen peroxide as a result of decreased detoxiﬁcation. The strong response to rotenone is interesting because rotenone is an inhibitor of mitochondrial complex I, which is also where the galactonolactone dehydrogenase protein is located (Schimmeyer et al., 2016). Also in the paper by Rosenwasser et al. (2013), rotenone-induced stress is also positively correlated with developmental senescence, a process which has parallels with fruit ripening. The data seem to indicate the existence of a signal aﬀecting transcription closely linked to GLD activity and the redox state of the ascorbate pool. The nature of this signal is unknown but a candidate is monodehydroascorbate, as it is the molecule that links ascorbate oxidase and monodehydroascorbate reductase activity, which could act as a primary signal given (i) the short half-life of the molecule, a prerequisite for a role in signaling and (ii) the proton transfer that occurs when the radical is reduced back to ascorbate (Bielski et al., 1981) which gives potential for protein modiﬁcation. Most oxidation reactions with ascorbate are indeed single electron reactions (Heber et al., 1996) generating the radical monodehydroascorbate, this radical is considered to be common and a marker for stress. Another possibility is that redox changes in mitochondria act as a retrograde signal (Rhoads and Subbaiah, 2007); this could be a possibility given the involvement of GLD. Interestingly a dehydroascorbate reductase (DHAR) gene (Solyc05g054760) is one of the 182 diﬀerentially regulated genes and is down- regulated in GLD and MDHAR lines and up-regulated in the AO line. Compensatory eﬀects, either by enzymes such as DHAR, or other isoforms of the enzymes under study, cannot be ruled out. FIGURE 3 \| Heatmap showing protein levels in orange pericarp fruit (3 biological replicates of 30 fruits) of the transgenic lines and wild-type. All proteins were separated by two-dimensional gel electrophoresis and identiﬁed by mass spectrometry. A Transcriptional Signature Is Observed and Is Correlated to Silencing of the Ascorbate Metabolizing Enzymes The comparisons were carried out ﬁrst between the transgenic lines (GalLDH = GLD) and wild-type (AO-WT, comparison 1; GalLDH-WT, comparison 2; MDHAR-WT, comparison 3) before comparing the transgenic lines between themselves (AO-MDHAR, comparison 4, AO-GalLDH, comparison 6, and GalLDH-MDHAR, comparison 7). Afterward the hypothesis was tested that the WT expression was the average of the expression in two given transgenic lines (see Materials and methods; WT = average AO and MD, comparison 5, WT = average AO and GalLDH, comparison 8, WT = average GalLDH and MDHAR, comparison 9). The black line is drawn to represent a threshold p-value of 0.001, the red points represent differentially expressed genes. postharvest storage conditions that may result in protein mis- folding such as high CO2 (Rothan et al., 1997) or chilling temperatures (Stevens et al., 2008). also aﬀected by environmental challenges: the unfolded protein response results from a period of stress and leads to either up- regulation of proteins required for folding or degradation or in extreme cases cell death if prolonged (Ruberti et al., 2015; Wan and Jiang, 2016). Functional hypotheses such as these would need to be tested. For example, puriﬁcation of polysomes from the available transgenic lines could lead to information on the RNAs undergoing translation and/or associated proteins and ribosome subunits. Considering the fruit context, this experiment would be especially informative in ripening fruits subjected to high temperature stress or in ripe fruits subjected to stressful A Transcriptional Signature Is Observed and Is Correlated to Silencing of the Ascorbate Metabolizing Enzymes The transcriptomes revealed a switch linked to the activity of ascorbate oxidase on one side of the switch and the enzymes MDHAR/GLD acting in the other direction, always with wild- type gene expression in the middle. There is a tendency for February 2018 \| Volume 9 \| Article 137 Frontiers in Plant Science \| www.frontiersin.org 10 Stevens et al. Ascorbate Signaling in Tomato Fruit FIGURE 4 \| Pairwise transcriptome comparisons as illustrated using volcano plots (log2 of fold change plotted against –log10 of the p-value for the differential expression). The expression of the 7633 genes from the microarray analysis of RNA extracted from the orange fruit pericarp (3 biological replicates of 30 fruit) of the different lines was compared with technical duplicates (dyes swapped). The comparisons were carried out ﬁrst between the transgenic lines (GalLDH = GLD) and wild-type (AO-WT, comparison 1; GalLDH-WT, comparison 2; MDHAR-WT, comparison 3) before comparing the transgenic lines between themselves (AO-MDHAR comparison 4, AO-GalLDH, comparison 6, and GalLDH-MDHAR, comparison 7). Afterward the hypothesis was tested that the WT expression was the average of t expression in two given transgenic lines (see Materials and methods; WT = average AO and MD, comparison 5, WT = average AO and GalLDH, comparison 8, WT average GalLDH and MDHAR, comparison 9). The black line is drawn to represent a threshold p-value of 0.001, the red points represent differentially expressed gene FIGURE 4 \| Pairwise transcriptome comparisons as illustrated using volcano plots (log2 of fold change plotted against –log10 of the p-value for the differentia expression). The expression of the 7633 genes from the microarray analysis of RNA extracted from the orange fruit pericarp (3 biological replicates of 30 fruit) different lines was compared with technical duplicates (dyes swapped). The comparisons were carried out ﬁrst between the transgenic lines (GalLDH = GLD) FIGURE 4 \| Pairwise transcriptome comparisons as illustrated using volcano plots (log2 of fold change plotted against –log10 of the p-value for the differential expression). The expression of the 7633 genes from the microarray analysis of RNA extracted from the orange fruit pericarp (3 biological replicates of 30 fruit) of the different lines was compared with technical duplicates (dyes swapped). Frontiers in Plant Science \| www.frontiersin.org Hub Gene Identiﬁcation Points to Novel Regulatory Roles Linked to Cellular Energy Status and Signaling Heat-shock proteins also ﬁgure amongst the hub genes identiﬁed: three heat-shock proteins are identiﬁed: SolycHsfB2a (Solyc03g026020), a Class II small heat-shock protein February 2018 \| Volume 9 \| Article 137 Frontiers in Plant Science \| www.frontiersin.org 11 Ascorbate Signaling in Tomato Fruit Stevens et al. TABLE 7 \| Gene ontology enrichment on the 182 set of common genes using the tools available at Boyce Thompson Institute, Cornell University (http://bioinfo.bti.cornell. edu/cgi-bin/MetGenMAP/home.cgi). Gene ontology term Number of genes Cluster frequency (%) Genome frequency (%) Corrected p-value Unfolded protein binding 3 1.6 0.2 0.038 Structural molecule activity 11 6.0 0.7 0.000 Structural constituent of ribosome 10 5.5 0.5 0.000 Small molecule binding 16 8.8 3.9 0.028 Small GTPase regulator activity 3 1.6 0.1 0.032 Ribonucleotide binding 11 6.0 2.0 0.025 Purine ribonucleotide binding 11 6.0 2.0 0.023 Purine ribonucleoside triphosphate binding 11 6.0 2.0 0.025 Purine nucleotide binding 11 6.0 2.0 0.021 Protein binding 44 24.2 11.9 0.000 Oxidoreductase activity acting on NADH or NADPH 4 2.2 0.2 0.003 Oxidoreductase activity 16 8.8 2.8 0.000 Organic cyclic compound binding 16 8.8 3.3 0.013 Nucleotide binding 16 8.8 3.3 0.015 Nucleoside phosphate binding 16 8.8 3.3 0.012 Catalytic activity 53 29.1 17.9 0.000 Binding 79 43.4 25.1 0.000 ATP binding 9 4.9 1.4 0.020 Adenyl ribonucleotide binding 9 4.9 1.4 0.019 Adenyl nucleotide binding 9 4.9 1.4 0.020 The table shows the cluster frequency (the 182 genes) compared to the genome frequency for each category with the FDR-corrected p-value. The three categories (structural constituent of ribosomes, small GTPase regulator activity and oxidoreductase activity acting on NADH or NADPH) are where there is the biggest difference between the genome frequency and the cluster frequency. TABLE 7 \| Gene ontology enrichment on the 182 set of common genes using the tools available at Boyce Thompson Institute, Cornell University (http://bioinfo.bti.cornell. edu/cgi-bin/MetGenMAP/home.cgi). The table shows the cluster frequency (the 182 genes) compared to the genome frequency for each category with the FDR-corrected p-value. The three categories (structural constituent of ribosomes, small GTPase regulator activity and oxidoreductase activity acting on NADH or NADPH) are where there is the biggest difference between the genome frequency and the cluster frequency. Hub Gene Identiﬁcation Points to Novel Regulatory Roles Linked to Cellular Energy Status and Signaling to show improved yield stability (Garchery et al., 2013) whereas the opposite is true for GLD and MDHAR RNAi lines (Alhagdow et al., 2007; Truﬀault et al., 2016), a phenomenon which is correlated to the changes in the fruit transcriptomes and could be related to fruit growth. Explanations for the phenotypes might be found in the identities of the hub genes which are broadly involved in signaling translation and protein (re)folding. We also noted the presence of many ATP-dependent proteins. The yield phenotypes could therefore reﬂect diﬀerences in cellular energy levels. Le-HSP17.6 (Solyc08g062340) and a chaperone protein DnaJ (Solyc03g123560). Heat-shock proteins are known to be involved in numerous biological phenomena including stress responses, circadian clock regulation, ﬂowering time, and hypocotyl elongation (Charng et al., 2007; Mittler et al., 2012; Kolmos et al., 2014). The hub gene protein DnaJ (hsp40) stimulates the ATPase activity of DnaK (an hsp70 homolog which is interestingly diﬀerentially accumulated in the AO and MDHAR proteomes). These proteins bind to unfolded polypeptides to prevent their aggregation, for example under conditions of heat-shock. The DnaK protein is of interest because it has been detected in the proteome of the Arabidopsis mutant for GDP-L-galactose phosphorylase (vtc2), a major control point for ascorbate synthesis, as being up-regulated after 3 and 5 days of high light compared to the wild-type (Giacomelli et al., 2006). This parallels the situation in the AO line (up- regulation) and is opposite to the MDHAR line phenotype (down-regulation) meaning that ascorbate deﬁciency mirrors the RNAi AO phenotype. Another hub protein, Arabinogalactan, is involved in many biological processes including germination, cell extension, fertilization, binding, and release of calcium and is linked to the auxin response (Ellis et al., 2010). Only one of the hub genes has already been characterized from tomato (Table 7) which is the tft8 protein, a member of the 14-3-3 protein family involved in defense responses (Roberts and Bowles, 1999). Frontiers in Plant Science \| www.frontiersin.org Control The ITAG2.4 identiﬁer (Solyc) with the annotated function is shown with the degree of association as deﬁned by Cytoscape. The transcriptome data (fold change between AO and MDHAR RNAi lines with the FDR corrected p-value on the transcript levels) for each gene are also shown. FIGURE 5 \| Network analysis (Cytoscape) showing the most highly connected genes based on their expression correlations. A topological overlap matrix was generated using the WGCNA script for the 182 core genes, the 22 metabolites and the 40 proteins. The edge ﬁle produced was imported into Cytoscape for network visualization. The network was visualized with a degree-sorted circle layout. The node sizes and node label sizes are proportional to the degree of connectivity of each node as ﬁxed by Cytoscape. The color scale goes from blue (least connected nodes) to orange (most connected nodes). The hub genes are described in the table which gives the most highly connected genes following visualization of the topology overlap matrix, generated by the WGCNA “R” package, with Cytoscape. The ITAG2.4 identiﬁer (Solyc) with the annotated function is shown with the degree of association as deﬁned by Cytoscape. The transcriptome data (fold change between AO and MDHAR RNAi lines with the FDR corrected p-value on the transcript levels) for each gene are also shown. lines, this includes dehydroascorbate, chlorogenic acid, and alanine. The changes seen in dehydroascorbate are counter- intuitive when compared to what is expected based on enzyme activity. We, with other independent groups, have previously reported increases in ascorbate levels on reduction of MDHAR activity (a peroxisomal and cytosolic isoform; Haroldsen et al., 2011; Gest et al., 2013a) which implies feedback regulation. Links have been shown between polyphenols such as chlorogenic acid or anthocyanins and ascorbate (Page et al., 2012); also oxidation of chlorogenic acid is correlated with oxidation of ascorbate to dehydroascorbate (Takahama, 1998) and ascorbate can regenerate chlorogenic acid from its corresponding radicals (Yamasaki and Grace, 1998). Our results show less chlorogenic acid in lines with lower ascorbate oxidase. We should point out that the genotype used in this study (West Virginia 106) is particularly rich in chlorogenic acid, the wild-type contains 3.25 mg/g dwt of chlorogenic acid whereas reported concentrations range from 0.24 to 3.69 mg/100 g fresh weight (between ∼0.024 and 0.37 mg/g dwt; Slimestad and Verheul, 2005; Peng et al., 2008). Control As many studies have observed, RNA and protein abundance are often weakly correlated (Halbeisen and Gerber, 2009; Petricka et al., 2012) and our data corroborate this: the quantity of the diﬀerentially accumulated proteins is generally not linked to their transcriptional levels (Table 6). For the observable proteins, the GLD proteome is not signiﬁcantly aﬀected: the GLD proteome actually is remarkably visually similar to the AO proteome which is surprising as the transcriptome data showed contrasted diﬀerential expression of genes for these two lines. The metabolic proﬁle is similar to the proteome proﬁle in that several metabolites show opposite changes compared to wild-type between AO and MDHAR RNAi The major phenotypic diﬀerences seen in the transgenic lines are related to growth: in previous studies RNAi AO lines tend February 2018 \| Volume 9 \| Article 137 Frontiers in Plant Science \| www.frontiersin.org 12 Stevens et al. Ascorbate Signaling in Tomato Fruit FIGURE 5 \| Network analysis (Cytoscape) showing the most highly connected genes based on their expression correlations. A topological overlap matrix was generated using the WGCNA script for the 182 core genes, the 22 metabolites and the 40 proteins. The edge ﬁle produced was imported into Cytoscape for network visualization. The network was visualized with a degree-sorted circle layout. The node sizes and node label sizes are proportional to the degree of connectivity of each node as ﬁxed by Cytoscape. The color scale goes from blue (least connected nodes) to orange (most connected nodes). The hub genes are described in the table which gives the most highly connected genes following visualization of the topology overlap matrix, generated by the WGCNA “R” package, with Cytoscape. The ITAG2.4 identiﬁer (Solyc) with the annotated function is shown with the degree of association as deﬁned by Cytoscape. The transcriptome data (fold change between AO and MDHAR RNAi lines with the FDR corrected p-value on the transcript levels) for each gene are also shown. FIGURE 5 \| Network analysis (Cytoscape) showing the most highly connected genes based on their expression correlations. A topological overlap matrix was generated using the WGCNA script for the 182 core genes, the 22 metabolites and the 40 proteins. The edge ﬁle produced was imported into Cytoscape for network visualization. The network was visualized with a degree-sorted circle layout. The node sizes and node label sizes are proportional to the degree of connectivity of each node as ﬁxed by Cytoscape. Control The color scale goes from blue (least connected nodes) to orange (most connected nodes). The hub genes are described in the table which gives the most highly connected genes following visualization of the topology overlap matrix, generated by the WGCNA “R” package, with Cytoscape. The ITAG2.4 identiﬁer (Solyc) with the annotated function is shown with the degree of association as deﬁned by Cytoscape. The transcriptome data (fold change between AO and MDHAR RNAi lines with the FDR corrected p-value on the transcript levels) for each gene are also shown. FIGURE 5 \| Network analysis (Cytoscape) showing the most highly connected genes based on their expression correlations. A topological overlap matrix was generated using the WGCNA script for the 182 core genes, the 22 metabolites and the 40 proteins. The edge ﬁle produced was imported into Cytoscape for network visualization. The network was visualized with a degree-sorted circle layout. The node sizes and node label sizes are proportional to the degree of connectivity of each node as ﬁxed by Cytoscape. The color scale goes from blue (least connected nodes) to orange (most connected nodes). The hub genes are described in the table which gives the most highly connected genes following visualization of the topology overlap matrix, generated by the WGCNA “R” package, with Cytoscape. The ITAG2.4 identiﬁer (Solyc) with the annotated function is shown with the degree of association as deﬁned by Cytoscape. The transcriptome data (fold change between AO and MDHAR RNAi lines with the FDR corrected p-value on the transcript levels) for each gene are also shown. FIGURE 5 \| Network analysis (Cytoscape) showing the most highly connected genes based on their expression correlations. A topological overlap matrix was generated using the WGCNA script for the 182 core genes, the 22 metabolites and the 40 proteins. The edge ﬁle produced was imported into Cytoscape for network visualization. The network was visualized with a degree-sorted circle layout. The node sizes and node label sizes are proportional to the degree of connectivity of each node as ﬁxed by Cytoscape. The color scale goes from blue (least connected nodes) to orange (most connected nodes). The hub genes are described in the table which gives the most highly connected genes following visualization of the topology overlap matrix, generated by the WGCNA “R” package, with Cytoscape. AUTHOR CONTRIBUTIONS Supplementary Figure S3 \| Gene expression proﬁles for AO, Solyc04g054690; GLD, Solyc10g079470 and MDHAR, Solyc09g009390 as obtained from the TomExpress tool (http://tomexpress.toulouse.inra.fr). RS: supervised the experiments, carried out the bioinformatic analysis of the data and wrote the article with the contribution of all the authors; MF and J-PB: carried out the proteome study; PB, CD, MM, AM, and J-LP: carried out the metabolic analyses; VG and JP: carried out the microarrays and transcriptome data normalization; ClD: analyzed the ROSMETER results; CG: carried out the greenhouse measurements, sampling and provided technical assistance to RS; BG and DP: carried out the post-harvest study; MC, HG, and VT: helped with the writing; CR: conceived the project and obtained the funding. All authors approved the ﬁnal version of the manuscript. Supplementary Figure S4 \| Gene expression proﬁles for AO, Solyc04g054690; GLD, Solyc10g079470 and MDHAR, Solyc09g009390 as obtained from the Solgenomics expression atlas tool (http://tea.solgenomics.net). Supplementary Table S1 \| Ascorbate and dehydroascorbate levels in pericarp fruit 20 days after anthesis of transgenic lines and WT. Ascorbate and dehydroascorbate were assayed in the pericarp tissue (3 biological replicates of 30 fruits per pool) of the lines by a spectrophotometric method. Measurements show means with standard error (SE). A comparison of means was carried out using a Kruskal Wallis test with correction (Dunn). Different letters indicate signiﬁcant differences (5% signiﬁcance level). Supplementary Table S1 \| Ascorbate and dehydroascorbate levels in pericarp fruit 20 days after anthesis of transgenic lines and WT. Ascorbate and dehydroascorbate were assayed in the pericarp tissue (3 biological replicates of 30 fruits per pool) of the lines by a spectrophotometric method. Measurements show means with standard error (SE). A comparison of means was carried out using a Kruskal Wallis test with correction (Dunn). Different letters indicate signiﬁcant differences (5% signiﬁcance level). Supplementary Table S2 \| Fruit metabolite data (orange fruit pericarp). Quantitative data for the following metabolites: galactose, mannose, glutamate, glutamine, alanine, aspartate, tyrosine, fructose, glucose, sucrose, ascorbate, dehydroascorbate, GABA, citramalate, citrate, malate, chlorogenic acid, caffeic acid glucoside, cis-chlorogenic acid, quercetin derivative, and rutin as measured by speciﬁc methods (1H-NMR for polar compounds, HPLC-DAD for the major polyphenols, spectrophotometrically for ascorbate and dehydroascorbate) for each of the three pools for the wild type, AO, GLD and MDHAR RNAi lines. Supplementary Table S2 \| Fruit metabolite data (orange fruit pericarp). SUPPLEMENTARY MATERIAL In conclusion, this study in a non-photosynthetic tissue reveals that the activity of enzymes controlling the oxidation state of the ascorbate pool are correlated to a few metabolic changes and some changes in protein quantity, particularly those involved in protein folding. The relatively low number of metabolic changes compared to gene expression probably reﬂects buﬀering of the system. In addition to this post- transcriptional regulation, a transcriptional response is seen which is also linked to the activity of the ﬁnal biosynthesis enzyme, galactonolactone dehydrogenase. Plants silenced for this enzyme show a transcriptional response similar to that of MDHAR silenced plants and opposite to ascorbate oxidase silenced plants implying signaling from the ascorbate pool. Network analysis reveals hub genes involved in fundamental cellular mechanisms including cellular signaling, translation and protein synthesis as well as the heat-shock response. The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fpls.2018. 00137/full#supplementary-material The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fpls.2018. 00137/full#supplementary-material Supplementary Figure S1 \| Heatmap showing protein levels in pericarp of fruit 20 days after anthesis (3 pools of 30 fruits) of the transgenic lines and wild-type. All proteins were separated by two-dimensional gel electrophoresis and identiﬁed by mass spectrometry. The log2 of the protein ratio with the wild-type for each transgenic line is presented. The scale goes from green (protein decreased compared to wild-type) to purple (protein increased). A comparison of the means was carried out using a Kruskal Wallis test with correction (Dunn). Different letters indicate signiﬁcant differences (5% signiﬁcance level). SGN codes are found at https://solgenomics.net/. Supplementary Figure S2 \| WGCNA module identiﬁcation. A graphical representation of the data from Supplementary Table S4: module eigengenes are labeled by color and the module membership scale from red (high adjacency, positive correlation) to green (low adjacency, negative correlation) shown for each metabolite and protein. The 7633 genes were used to generate the topology overlap matrix. SGN codes are found at https://solgenomics.net/. FUNDING The authors are grateful to the INRA AgroBI programme for ﬁnancing the study, “VTCFruit” project. AUTHOR CONTRIBUTIONS Quantitative data for the following metabolites: galactose, mannose, glutamate, glutamine, alanine, aspartate, tyrosine, fructose, glucose, sucrose, ascorbate, dehydroascorbate, GABA, citramalate, citrate, malate, chlorogenic acid, caffeic acid glucoside, cis-chlorogenic acid, quercetin derivative, and rutin as measured by speciﬁc methods (1H-NMR for polar compounds, HPLC-DAD for the major polyphenols, spectrophotometrically for ascorbate and dehydroascorbate) for each of the three pools for the wild type, AO, GLD and MDHAR RNAi lines. Control The diﬀerences in alanine are also of interest as this metabolite is produced directly from pyruvate and therefore has close links with glycolysis, the oxidative pentose phosphate pathway and FIGURE 6 \| ROSMETER analysis of the WT vs. AO and WT vs. MDHAR transcriptome data from tomato fruit pericarp. Tomato genes were converted to the Arabidopsis homolog and the fold-changes and p-values were compared to the Arabidopsis transcriptomes previously described (Rosenwasser et al., 2013) which covered different stress conditions and/or mutant backgrounds in Arabidopsis as clusters as described in the text to the ﬁgure. The heatmap generated shows green for general transcriptome down-regulation (opposite fold-change) and red for up-regulation (similar fold-change) in the tomato lines compared to Arabidopsis. FIGURE 6 \| ROSMETER analysis of the WT vs. AO and WT vs. MDHAR transcriptome data from tomato fruit pericarp. Tomato genes were converted to the Arabidopsis homolog and the fold-changes and p-values were compared to the Arabidopsis transcriptomes previously described (Rosenwasser et al., 2013) which covered different stress conditions and/or mutant backgrounds in Arabidopsis as clusters as described in the text to the ﬁgure. The heatmap generated shows green for general transcriptome down-regulation (opposite fold-change) and red for up-regulation (similar fold-change) in the tomato lines compared to Arabidopsis. February 2018 \| Volume 9 \| Article 137 Frontiers in Plant Science \| www.frontiersin.org 13 Ascorbate Signaling in Tomato Fruit Stevens et al. the TCA cycle and cellular energy production. Alanine is a product of anaerobic metabolism resulting from high rates of amino acid interconversion (Rocha et al., 2010b). Unlike lactate and ethanol, alanine accumulation does not have detrimental side eﬀects but maintains the glycolytic ﬂux (Rocha et al., 2010a). version of the abstract was published at http://qualityfruit.inp- toulouse.fr/_attachments/preliminary-program-article/Book %2520of%2520Abstracts%2520COST%2520Porto2016%2520 %252814out16%2529.pdf?download=true, as part of the COST Action FA1106 “Quality Fruit.” Frontiers in Plant Science \| www.frontiersin.org REFERENCES Fernandez, A. I., Viron, N., Alhagdow, M., Karimi, M., Jones, M., Amsellem, Z., et al. (2009). Flexible tools for gene expression and silencing in tomato. Plant Physiol. 151, 1729–1740. doi: 10.1104/pp.109.147546 Alba, R., Payton, P., Fei, Z., McQuinn, R., Debbie, P., Martin, G. B., et al. (2005). Transcriptome and selected metabolite analyses reveal multiple points of ethylene control during tomato fruit development. Plant Cell 17, 2954–2965. doi: 10.1105/tpc.105.036053 Fiorani, M., De Sanctis, R., Scarlatti, F., Vallorani, L., De Bellis, R., Seraﬁni, G., et al. (2000). Dehydroascorbic acid irreversibly inhibits hexokinase activity. Mol. Cell. Biochem. 209, 145–153. doi: 10.1023/A:1007168032289 Alhagdow, M., Mounet, F., Gilbert, L., Nunes-Nesi, A., Garcia, V., Just, D., et al. (2007). Silencing of the mitochondrial ascorbate synthesizing enzyme L-galactono-1,4-lactone dehydrogenase aﬀects plant and fruit development in tomato. Plant Physiol. 145, 1408–1422. doi: 10.1104/pp.107. 106500 Garchery, C., Gest, N., Do, P. T., Alhagdow, M., Baldet, P., Menard, G., et al. (2013). A diminution in ascorbate oxidase activity aﬀects carbon allocation and improves yield in tomato under water deﬁcit. Plant Cell Environ. 36, 159–175. doi: 10.1111/j.1365-3040.2012.02564.x Gerashchenko, M. V., Lobanov, A. V., and Gladyshev, V. N. (2012). Genome- wide ribosome proﬁling reveals complex translational regulation in response to oxidative stress. Proc. Natl. Acad. Sci. U.S.A. 109, 17394–17399. doi: 10.1073/pnas.1120799109 Banhegyi, G., Csala, M., Szarka, A., Varsanyi, M., Benedetti, A., and Mandl, J. (2003). Role of ascorbate in oxidative protein folding. Biofactors 17, 37–46. doi: 10.1002/biof.5520170105 Gest, N., Garchery, C., Gautier, H., Jimenez, A., and Stevens, R. (2013a). Light- dependent regulation of ascorbate in tomato by a monodehydroascorbate reductase localized in peroxisomes and the cytosol. Plant Biotechnol. J. 11, 344–354. doi: 10.1111/pbi.12020 Bassel, G. W., Gaudinier, A., Brady, S. M., Hennig, L., Rhee, S. Y., and De Smet, I. (2012). Systems analysis of plant functional, transcriptional, physical interaction, and metabolic networks. Plant Cell 24, 3859–3875. doi: 10.1105/tpc.112.100776 Bielski, B. H. J., Allen, A. O., and Schwarz, H. A. (1981). Mechanisms of disproportionation of ascorbate radicals. J. Am. Chem. Soc. 103, 3516–3518. doi: 10.1021/ja00402a042 Gest, N., Gautier, H., and Stevens, R. (2013b). Ascorbate as seen through plant evolution: the rise of a successful molecule? J. Exp. Bot. 64, 33–53. doi: 10.1093/jxb/ers297 Giacomelli, L., Rudella, A., and van Wijk, K. J. (2006). High light response of the thylakoid proteome in arabidopsis wild type and the ascorbate-deﬁcient mutant vtc2-2. A comparative proteomics study. Plant Physiol. 141, 685–701. doi: 10.1104/pp.106.080150 Blaschke, K., Ebata, K. REFERENCES T., Karimi, M. M., Zepeda-Martinez, J. A., Goyal, P., Mahapatra, S., et al. (2013). Vitamin C induces Tet-dependent DNA demethylation and a blastocyst-like state in ES cells. Nature 500, 222–226. doi: 10.1038/nature12362 Gismondi, A., Caldarola, S., Lisi, G., Juli, G., Chellini, L., Iadevaia, V., et al. (2014). Ribosomal stress activates eEF2K-eEF2 pathway causing translation elongation inhibition and recruitment of terminal oligopyrimidine (TOP) mRNAs on polysomes. Nucleic Acids Res. 42, 12668–12680. doi: 10.1093/nar/gku996 Bors, W., and Buettner, G. R. (1997). “The vitamin C radical and its reactions,” in Vitamin C in Health and Disease, eds L. Packer and J. Fuchs (New York, NY: Marcel Dekker), 75–94. Boyle, E. I., Weng, S., Gollub, J., Jin, H., Botstein, D., Cherry, J. M., et al. (2004). GO::TermFinder–open source software for accessing Gene Ontology information and ﬁnding signiﬁcantly enriched Gene Ontology terms associated with a list of genes. Bioinformatics 20, 3710–3715. doi: 10.1093/bioinformatics/bth456 Halbeisen, R. E., and Gerber, A. P. (2009). Stress-dependent coordination of transcriptome and translatome in yeast. PLoS Biol. 7:e1000105. doi: 10.1371/journal.pbio.1000105 Haroldsen, V. M., Chi-Ham, C. L., Kulkarni, S., Lorence, A., and Bennett, A. B. (2011). Constitutively expressed DHAR and MDHAR inﬂuence fruit, but not foliar ascorbate levels in tomato. Plant Physiol. Biochem. 49, 1244–1249. doi: 10.1016/j.plaphy.2011.08.003 Carcamo, J. M., Pedraza, A., Borquez-Ojeda, O., Zhang, B., Sanchez, R., and Golde, D. W. (2004). Vitamin C is a kinase inhibitor: dehydroascorbic acid inhibits IkappaBα kinase β. Mol. Cell. Biol. 24, 6645–6652. doi: 10.1128/MCB.24.15.6645-6652.2004 Heber, U., Miyake, C., Mano, J., Ohno, C., and Asada, K. (1996). Monodehydroascorbate radical detected by electron paramagnetic resonance spectrometry is a sensitive probe of oxidative stress in intact leaves. Plant Cell Physiol. 37, 1066–1072. doi: 10.1093/oxfordjournals.pcp.a029055 Charng, Y. Y., Liu, H. C., Liu, N. Y., Chi, W. T., Wang, C. N., Chang, S. H., et al. (2007). A heat-inducible transcription factor, HsfA2, is required for extension of acquired thermotolerance in Arabidopsis. Plant Physiol. 143, 251–262. doi: 10.1104/pp.106.091322 Khandal, D., Samol, I., Buhr, F., Pollmann, S., Schmidt, H., Clemens, S., et al. (2009). Singlet oxygen-dependent translational control in the tigrina- d.12 mutant of barley. Proc. Natl. Acad. Sci. U.S.A. 106, 13112–13117. doi: 10.1073/pnas.0903522106 Cisternas, P., Silva-Alvarez, C., Martinez, F., Fernandez, E., Ferrada, L., Oyarce, K., et al. (2014). The oxidized form of vitamin C, dehydroascorbic acid, regulates neuronal energy metabolism. J. Neurochem. 129, 663–671. doi: 10.1111/jnc.12663 Klee, H. J., and Giovannoni, J. J. (2011). REFERENCES Genetics and control of tomato fruit ripening and quality attributes. Annu. Rev. Genet. 45, 41–59. doi: 10.1146/annurev-genet-110410-132507 Cline, M. S., Smoot, M., Cerami, E., Kuchinsky, A., Landys, N., Workman, C., et al. (2007). Integration of biological networks and gene expression data using Cytoscape. Nat. Protoc. 2, 2366–2382. doi: 10.1038/nprot.2007.324 Kolmos, E., Chow, B. Y., Pruneda-Paz, J. L., and Kay, S. A. (2014). HsfB2b-mediated repression of PRR7 directs abiotic stress responses of the circadian clock. Proc. Natl. Acad. Sci. U.S.A. 111, 16172–16177. doi: 10.1073/pnas.1418483111 Davey, M. W., Auwerkerken, A., and Keulemans, J. (2007). Relationship of apple vitamin C and antioxidant contents to harvest date and postharvest pathogen infection. J. Sci. Food Agric. 87, 802–813. doi: 10.1002/jsfa.2777 Krouk, G., Lingeman, J., Colon, A. M., Coruzzi, G., and Shasha, D. (2013). Gene regulatory networks in plants: learning causality from time and perturbation. Genome Biol. 14:123. doi: 10.1186/gb-2013-14-6-123 De Tullio, M. C. (2012). Beyond the antioxidant: the double life of vitamin C. Subcell. Biochem. 56, 49–65. doi: 10.1007/978-94-007-2199-9_4 De Tullio, M. C., Ciraci, S., Liso, R., and Arrigoni, O. (2007). Ascorbic acid oxidase is dynamically regulated by light and oxygen. A tool for oxygen management in plants? J. Plant Physiol. 164, 39–46. doi: 10.1016/j.jplph.2005. 09.016 Laing, W. A., Martinez-Sanchez, M., Wright, M. A., Bulley, S. M., Brewster, D., Dare, A. P., et al. (2015). An upstream open reading frame is essential for feedback regulation of ascorbate biosynthesis in Arabidopsis. Plant Cell 27, 772–786. doi: 10.1105/tpc.114.133777 Ellis, M., Egelund, J., Schultz, C. J., and Bacic, A. (2010). Arabinogalactan- proteins: key regulators at the cell surface? Plant Physiol. 153, 403–419. doi: 10.1104/pp.110.156000 Langfelder, P., and Horvath, S. (2008). WGCNA: an R package for weighted correlation network analysis. BMC Bioinformatics 9:559. doi: 10.1186/1471-2105-9-559 Faurobert, M., Mihr, C., Bertin, N., Pawlowski, T., Negroni, L., Sommerer, N., et al. (2007). Major proteome variations associated with cherry tomato pericarp development and ripening. Plant Physiol. 143, 1327–1346. doi: 10.1104/pp.106.092817 Leferink, N. G., van Duijn, E., Barendregt, A., Heck, A. J., and van Berkel, W. J. (2009). Galactonolactone dehydrogenase requires a redox-sensitive thiol for optimal production of vitamin C. Plant Physiol. 150, 596–605. doi: 10.1104/pp.109.136929 Fei, Z., Joung, J. G., Tang, X., Zheng, Y., Huang, M., Lee, J. M., et al. (2011). Tomato functional genomics database: a comprehensive resource and analysis package for tomato functional genomics. Nucleic Acids Res. 39, D1156–D1163. doi: 10.1093/nar/gkq991 Malacrida, C., Valle, E., and Boggio, S. (2006). ACKNOWLEDGMENTS Supplementary Table S3 \| Proteins. Raw data for quantiﬁable proteins (see Materials and methods) from orange pericarp with spot identiﬁer code for each of the three pools for the wild type, AO, GLD, and MDHAR RNAi lines. We acknowledge valuable help from Noé Fernandez from SGN for converting the TOM2 probes to Solyc identiﬁants and Shilo Rosenwasser for help with application of the ROSMETER tool to tomato data. We would also like to thank Manu Botton for help with the plant culture and pre-harvest phenotyping, Joël Chadoeuf for advice on statistical models in R ˜maanova, Camille Croset for polyphenol analysis, Marine Palau for contributing to the proteome analysis and Christopher Sauvage for advice on the R scripts. The metabolome analyses were performed at Bordeaux Metabolome facility-MetaboHUB. We thank Daniel Jacob for developing and maintaining the MeRy-B database. A Supplementary Table S4 \| WGCNA. Solyc identiﬁer, gene function, GO terms, module color, module membership (MM or eigengene-based connectivity) and associated p-value for the 38 modules and 7633 genes as generated by the WGCNA R package. Supplementary Table S5 \| Transcriptome raw data. For each of the 7633 genes mean expression (normalized) in each of the three transgenic lines and wild type with standard error (SE). F values, p-values, false discovery rate corrected p-values and fold change (FC) are also given for each comparison as follows: 1: AO vs. WT, 2: GLD vs. WT, 3: MDHAR vs. WT, 4: AO vs. MDHAR, 5: AO + MDHAR/(2 × WT) = myModel, 6: AO vs. GLD, 7: GLD vs. MDHAR, 8: AO + GLD/(2 × WT) = myModel2, 9: GLD + MDHAR/(2 × WT) = myModel3. Frontiers in Plant Science \| www.frontiersin.org February 2018 \| Volume 9 \| Article 137 14 Ascorbate Signaling in Tomato Fruit Stevens et al. REFERENCES Postharvest chilling induces oxidative stress response in the dwarf tomato cultivar Micro-Tom. Physiol. Plant. 127, 10–18. doi: 10.1111/j.1399-3054.2005.00636.x February 2018 \| Volume 9 \| Article 137 Frontiers in Plant Science \| www.frontiersin.org 15 Ascorbate Signaling in Tomato Fruit Stevens et al. Ruberti, C., Kim, S. J., Stefano, G., and Brandizzi, F. (2015). Unfolded protein response in plants: one master, many questions. Curr. Opin. Plant Biol. 27, 59–66. doi: 10.1016/j.pbi.2015.05.016 Mittler, R., Finka, A., and Goloubinoﬀ, P. (2012). How do plants feel the heat? Trends Biochem. Sci. 37, 118–125. doi: 10.1016/j.tibs.2011.11.007 Moing, A., Maucourt, M., Renaud, C., Gaudillere, M., Brouquisse, R., Leboutellier, B., et al. (2004). Quantitative metabolic proﬁling through one-dimensional 1H- NMR analyses: application to plant genetics and functional genomics. Funct. Plant Biol. 31, 889–902. doi: 10.1071/FP04066 Schimmeyer, J., Bock, R., and Meyer, E. H. (2016). L-Galactono-1,4- lactone dehydrogenase is an assembly factor of the membrane arm of mitochondrial complex I in Arabidopsis. Plant Mol. Biol. 90, 117–126. doi: 10.1007/s11103-015-0400-4 Moore, M., Gossmann, N., and Dietz, K. J. (2016). Redox regulation of cytosolic translation in plants. Trends Plant Sci. 21, 388–397. doi: 10.1016/j.tplants.2015.11.004 Slimestad, R., and Verheul, M. J. (2005). Seasonal variations in the level of plant constituents in greenhouse production of cherry tomatoes. J. Agric. Food Chem. 53, 3114–3119. doi: 10.1021/jf047864e Mounet, F., Lemaire-Chamley, M., Maucourt, M., Cabasson, C., Giraudel, J. L., Deborde, C., et al. (2007). Quantitative metabolic proﬁles of tomato ﬂesh and seeds during fruit development: complementary analysis with ANN and PCA. Metabolomics 3, 273–288. doi: 10.1007/s11306-007-0059-1 Stevens, R., Buret, M., Garchery, C., Carretero, Y., and Causse, M. (2006). Technique for rapid, small-scale analysis of vitamin C levels in fruit and application to a tomato mutant collection. J. Agric. Food Chem. 54, 6159–6165. doi: 10.1021/jf061241e Mounet, F., Moing, A., Garcia, V., Petit, J., Maucourt, M., Deborde, C., et al. (2009). Gene and metabolite regulatory network analysis of early developing fruit tissues highlights new candidate genes for the control of tomato fruit composition and development. Plant Physiol. 149, 1505–1528. doi: 10.1104/pp.108.133967 Stevens, R., Page, D., Gouble, B., Garchery, C., Zamir, D., and Causse, M. (2008). Tomato fruit ascorbic acid content is linked with monodehydroascorbate reductase activity and tolerance to chilling stress. Plant Cell Environ. 31, 1086–1096. doi: 10.1111/j.1365-3040.2008.01824.x Mustroph, A., Zanetti, M. E., Jang, C. J., Holtan, H. E., Repetti, P. P., Galbraith, D. W., et al. (2009). REFERENCES Proﬁling translatomes of discrete cell populations resolves altered cellular priorities during hypoxia in Arabidopsis. Proc. Natl. Acad. Sci. U.S.A. 106, 18843–18848. doi: 10.1073/pnas.0906131106 Szarka, A., and Lorincz, T. (2014). The role of ascorbate in protein folding. Protoplasma 251, 489–497. doi: 10.1007/s00709-013-0560-5 Takahama, U. (1998). Ascorbic acid-dependent regulation of redox levels of chlorogenic acid and its isomers in the apoplast of leaves of Nicotiana tabacum L. Plant Cell Physiol. 39, 681–689. doi: 10.1093/oxfordjournals.pcp.a029422 Page, M., Sultana, N., Paszkiewicz, K., Florance, H., and Smirnoﬀ, N. (2012). The inﬂuence of ascorbate on anthocyanin accumulation during high light acclimation in Arabidopsis thaliana: further evidence for redox control of anthocyanin synthesis. Plant Cell Environ. 35, 388–404. doi: 10.1111/j.1365-3040.2011.02369.x Truﬀault, V., Gest, N., Garchery, C., Florian, A., Fernie, A. R., Gautier, H., et al. (2016). Reduction of MDHAR activity in cherry tomato suppresses growth and yield and MDHAR activity is correlated with sugar levels under high light. Plant Cell Environ. 39, 1279–1292. doi: 10.1111/pce.12663 Peng, Y., Zhang, Y., and Ye, J. (2008). Determination of phenolic compounds and ascorbic acid in diﬀerent fractions of tomato by capillary electrophoresis with electrochemical detection. J. Agric. Food Chem. 56, 1838–1844. doi: 10.1021/jf0727544 Truﬀault, V., Riqueau, G., Garchery, C., Gautier, H., and Stevens, R.G. (2017). Is monodehydroascorbate reductase activity in leaf tissue critical for the maintenance of yield in tomato? J. Plant Physiol. 222, 1–8. doi: 10.1016/j.jplph.2017.12.012 Wan, S., and Jiang, L. (2016). Endoplasmic Reticulum (ER) stress and the Unfolded Protein Response (UPR) in plants. Protoplasma 253, 753–764. doi: 10.1007/s00709-015-0842-1 Petricka, J. J., Schauer, M. A., Megraw, M., Breakﬁeld, N. W., Thompson, J. W., Georgiev, S., et al. (2012). The protein expression landscape of the Arabidopsis root. Proc. Natl. Acad. Sci. U.S.A. 109, 6811–6818. doi: 10.1073/pnas.1202546109 Xu, J., Pascual, L., Aurand, R., Bouchet, J. P., Valot, B., Zivy, M., et al. (2013). An extensive proteome map of tomato (Solanum lycopersicum) fruit pericarp. Proteomics 13, 3059–3063. doi: 10.1002/pmic.201200438 Puskas, F., Gergely, P. Jr., Banki, K., and Perl, A. (2000). Stimulation of the pentose phosphate pathway and glutathione levels by dehydroascorbate, the oxidized form of vitamin C. FASEB J. 14, 1352–1361. doi: 10.1096/fasebj.14. 10.1352 Yamasaki, H., and Grace, S. C. (1998). EPR detection of phytophenoxyl radicals stabilized by zinc ions: evidence for the redox coupling of plant phenolics with ascorbate in the H2O2-peroxidase system. FEBS Lett. 422, 377–380. doi: 10.1016/S0014-5793(98)00048-9 R Development-Core-Team (2015). R: A Language and Environment for Statistical Computing. REFERENCES Vienna: R Foundation for Statistical Computing. Yang, Y. H., Dudoit, S., Luu, P., Lin, D. M., Peng, V., Ngai, J., et al. (2002). Normalization for cDNA microarray data: a robust composite method addressing single and multiple slide systematic variation. Nucleic Acids Res. 30:e15. doi: 10.1093/nar/30.4.e15 Rhoads, D. M., and Subbaiah, C. C. (2007). Mitochondrial retrograde regulation in plants. Mitochondrion 7, 177–194. doi: 10.1016/j.mito.2007. 01.002 Ritchie, M. E., Phipson, B., Wu, D., Hu, Y., Law, C. W., Shi, W., et al. (2015). Limma powers diﬀerential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 43, e47. doi: 10.1093/nar/gkv007 Young, M. D., Wakeﬁeld, M. J., Smyth, G. K., and Oshlack, A. (2010). Gene ontology analysis for RNA-seq: accounting for selection bias. Genome Biol. 11:R14. doi: 10.1186/gb-2010-11-2-r14 Roberts, M. R., and Bowles, D. J. (1999). Fusicoccin, 14-3-3 proteins, and defense responses in tomato plants. Plant Physiol. 119, 1243–1250. doi: 10.1104/pp.119.4.1243 Yun, J., Mullarky, E., Lu, C., Bosch, K. N., Kavalier, A., Rivera, K., et al. (2015). Vitamin C selectively kills KRAS and BRAF mutant colorectal cancer cells by targeting GAPDH. Science 350, 1391–1396. doi: 10.1126/science.aaa5004 Rocha, M., Licausi, F., Araujo, W. L., Nunes-Nesi, A., Sodek, L., Fernie, A. R., et al. (2010a). Glycolysis and the tricarboxylic acid cycle are linked by alanine aminotransferase during hypoxia induced by waterlogging of Lotus japonicus. Plant Physiol. 152, 1501–1513. doi: 10.1104/pp.109.150045 Zito, E., Hansen, H. G., Yeo, G. S., Fujii, J., and Ron, D. (2012). Endoplasmic reticulum thiol oxidase deﬁciency leads to ascorbic acid depletion and noncanonical scurvy in mice. Mol. Cell 48, 39–51. doi: 10.1016/j.molcel.2012.08.010 Rocha, M., Sodek, L., Licausi, F., Hameed, M. W., Dornelas, M. C., and van Dongen, J. T. (2010b). Analysis of alanine aminotransferase in various organs of soybean (Glycine max) and in dependence of diﬀerent nitrogen fertilisers during hypoxic stress. Amino Acids 39, 1043–1053. doi: 10.1007/s00726-010-0596-1 Conﬂict of Interest Statement: The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. Rosenwasser, S., Fluhr, R., Joshi, J. R., Leviatan, N., Sela, N., Hetzroni, A., et al. (2013). ROSMETER: a bioinformatic tool for the identiﬁcation of transcriptomic imprints related to reactive oxygen species type and origin provides new insights into stress responses. Plant Physiol. 163, 1071–1083. REFERENCES doi: 10.1104/pp.113.218206 Copyright © 2018 Stevens, Baldet, Bouchet, Causse, Deborde, Deschodt, Faurobert, Garchery, Garcia, Gautier, Gouble, Maucourt, Moing, Page, Petit, Poëssel, Truﬀault and Rothan. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Rothan, C., Duret, S., Chevalier, C., and Raymond, P. (1997). Suppression of ripening-associated gene expression in tomato fruits subjected to a high CO2 concentration. Plant Physiol. 114, 255–263. doi: 10.1104/pp. 114.1.255 February 2018 \| Volume 9 \| Article 137 Frontiers in Plant Science \| www.frontiersin.org 16
https://openalex.org/W4206083875	https://link.springer.com/content/pdf/10.1007/JHEP01(2022)085.pdf	English	null	Entanglement island, miracle operators and the firewall	The Journal of high energy physics/The journal of high energy physics	2,022	cc-by	11,941	Published for SISSA by Springer Published for SISSA by Springer Received: July 13, 2021 Revised: November 21, 2021 Accepted: November 22, 2021 Published: January 17, 2022 Received: July 13, 2021 Revised: November 21, 2021 Accepted: November 22, 2021 Published: January 17, 2022 Open Access, c⃝The Authors. Article funded by SCOAP3. Entanglement island, miracle operators and the ﬁrewall https://doi.org/10.1007/JHEP01(2022)085 Contents 1 Introduction 1 2 State reconstruction formula for an evaporating black hole 3 3 A paradox in state reconstruction 6 3.1 A no-go theorem 6 3.2 An apparent paradox 7 4 The miracle operators 9 5 The ﬁrewall 12 5.1 Entanglement breaking measurement 13 5.2 Entanglement checking 15 6 Further discussion and conclusion 16 A Deﬁnition of LOCC 18 B Positive conditional entropy for separable states 19 Contents 1 Introduction 1 2 State reconstruction formula for an evaporating black hole 3 3 A paradox in state reconstruction 6 3.1 A no-go theorem 6 3.2 An apparent paradox 7 4 The miracle operators 9 5 The ﬁrewall 12 5.1 Entanglement breaking measurement 13 5.2 Entanglement checking 15 6 Further discussion and conclusion 16 A Deﬁnition of LOCC 18 B Positive conditional entropy for separable states 19 2 State reconstruction formula for an evaporating black hole JHEP01(2022)085 6 Further discussion and conclusion A Deﬁnition of LOCC B Positive conditional entropy for separable states Entanglement island, miracle operators and the ﬁrewall JHEP01(2022)085 Xiao-Liang Qi Stanford Institute for Theoretical Physics, Stanford University, Stanford, CA 94305, U.S.A. Xiao-Liang Qi Stanford Institute for Theoretical Physics, Stanford University, Stanford, CA 94305, U.S.A. Abstract: In this paper, we obtain some general results on information retrieval from the black hole interior, based on the recent progress on quantum extremal surface formula and entanglement island. We study an AdS black hole coupled to a bath with generic dy- namics, and ask whether it is possible to retrieve information about a small perturbation in the interior from the bath system. We show that the one-norm distance between two reduced states in a bath region A is equal to the same quantity in the bulk quantum ﬁeld theory for region AI where I is the entanglement island of A. This is a straightforward generalization of bulk-boundary correspondence in AdS/CFT. However, we show that a contradiction arises if we apply this result to a special situation when the bath dynamics includes a unitary operation that carries a particular measurement to a region A and send the result to another region W. Physically, the contradiction arises between transferabil- ity of classical information during the measurement, and non-transferability of quantum information which determines the entanglement island. We propose that the resolution of the contradiction is to realize that the state re- construction formula does not apply to the special situation involving interior-information- retrieving measurements. This implies that the assumption of smooth replica AdS geometry with boundary condition set by the ﬂat space bath has to break down when the particular measurement operator is applied to the bath. Using replica trick, we introduce an explicitly construction of such operator, which we name as “miracle operators”. From this construc- tion we see that the smooth replica geometry assumption breaks down because we have to introduce extra replica wormholes connecting with the “simulated blackholes” introduced by the miracle operator. We study the implication of miracle operators in understanding the ﬁrewall paradox. Keywords: AdS-CFT Correspondence, Black Holes, Gauge-gravity correspondence, Mod- els of Quantum Gravity ArXiv ePrint: 2105.06579 Open Access, c⃝The Authors. Article funded by SCOAP3. Open Access, c⃝The Authors. Article funded by SCOAP3. 1 Introduction Since the discovery of Hawking radiation [1], the black hole information paradox remains an important open question in physics. The key question in black hole information paradox is the fate of information carried by objects falling into a black hole. If unitarity of quantum mechanics is preserved in the presence of gravity, one expects the Hawking radiation cannot be always thermal, and the entropy of an evaporating black hole has to decrease at late time, known as the Page curve. [2] An important recent progress towards solving the black hole information paradox is the discovery of entanglement island [3, 4]. The key idea of the entanglement island works is to generalize the Hubeny-Rangamani-Takayanagi (HRT) formula [5, 6] with quantum corrections [7] to the entanglement entropy of a region in the radiation. In standard holographic duality, the entanglement entropy of a boundary region A is dual to the area of an extremal surface γA which is homologous to A, with the location of γA determined by extremizing the generalized entropy \|γA\| 4GN + Sbulk(ΣA). Here ΣA is the region in the bulk with γA and A as its boundary, which is known as the entanglement wedge of A. Refs. [3, 4] and more recent works have generalized this formula to a coupled system of a holographic conformal ﬁeld theory (CFT) and a bath system. Sbulk(ΣA) is generalized to the entanglement entropy of a region ΣA in the quantum ﬁeld theory with ﬁxed geometry background, where ΣA includes the A region itself and possibly an additional region I – 1 – in the holographic bulk. Region I is determined by the same extremization procedure, which is called an “entanglement island” when it is nontrivial. The entanglement island formula can be derived using a replica calculation, where diﬀerent replicas are connected by a replica wormhole through the island region [8, 9]. For an evaporating black hole coupled with a ﬂat space bath, it was shown that for a large enough region in the radiation (i.e. the bath), the entanglement island becomes nontrivial after a time known as the Page time. Taking into account of the entanglement island, the entanglement entropy of the radiation becomes smaller than the thermal value in Hawking’s calculation, which correctly resolves the conﬂict between gravitational cal- culation and unitarity. 1 Introduction Following the entanglement wedge reconstruction in holographic duality, the island formula also suggests that (small perturbations in) the entanglement island region of the black hole interior can be reconstructed in the radiation. Retrieval of information from the entanglement island has been discussed using diﬀerent methods [8, 10]. JHEP01(2022)085 In this paper, we want to address the following questions: in general, how diﬃcult is the information retrieval from the entanglement island? Is it possible to carry a measurement in the radiation that measures the state of an interior qubit in the entanglement island? How will measurements in the radiation aﬀects the physics seen by an infalling observer? We begin by proving a general result on state reconstruction. Similar to the bulk-boundary correspondence of relative entropy shown in ref. [11], we use replica trick to show that the “boundary” one-norm distance between two (similar) states in a bath region A is equal to the “bulk” one-norm distance between corresponding states in the region AI, if A has an entanglement island I. This is a straightforward generalization of the AdS/CFT results where boundary region is replaced by a region in the bath, and the entanglement wedge is replaced by the entanglement island. As a consequence of this general state reconstruction formula, we show that a region in the bath that is only classically correlated with the rest of the system can never reconstruct information that is encoded to the system by applying a unitary in the interior, because the interior is space-like separated from the bath. However, this leads to a contradiction. If there is a region A with nontrivial entanglement island I, and another region W without entanglement island, we can carry a measurement on A that retrieves the information in the interior region I, and send the measurement result to W. On the other hand, after the measurement (which is a unitary acting on the bath system), we show that the state reconstruction formula suggests the state of W still has no corre- lation with the interior information, which is contradictory with the fact that W knows the measurement result from A. By contradiction this suggests that the replica wormhole calculation that leads to the state reconstruction formula cannot apply to both regions discussed above. We propose that whenever an interior-information-retrieval measurement occurs in the bath, the state reconstruction formula fails for the state after the measure- ment. 1 Introduction (Note that we always understand “measurement” as a unitary operator applying to A, W and other parts of the bath, so the problem is not caused by non-unitarity.) We name such measurement operators as “miracle operators” since they have to change the gravitational path integral in a nontrivial way. By explicitly constructing miracle oper- ators using a replica trick, we understand why the original state reconstruction formula fails when such an operator is inserted. Additional replica wormholes arise not between diﬀerent – 2 – replicas of the original black hole, but between the original black hole and “simulated black hole”1 used for deﬁning the miracle operator. Using this explicit construction, we obtain some new understanding to the ﬁrewall paradox [12, 13] in two diﬀerent setups. We show that a measurement in the radiation can measure the state of an interior qubit, and an infalling obsever will see that the entanglement of two qubits across the quantum extremal surface can be destroyed by such exterior measurements. Furthermore, we construct a diﬀerent measurement which applies to the radiation and checks that one of the radiation qubit is in a particular entangled state with an interior qubit in the entanglement island. In this setup, an infalling observer will still be able to verify the entanglement of the entangled pair, which appears to contradict with the expectation in ordinary quantum mechanics. This is allowed because the entanglement checking operator must come with its own grav- itational theory and provide additional copies of the universe including the infalling ob- server. Interestingly, in the replica calculation of this setup, one can see that the success of entanglement checking is related to the nontrivial homotopy group of the replica manifold. JHEP01(2022)085 The remainder of the paper is organized as follows. In section 2 we derive a state recon- struction formula for the evaporating black hole system based on the results of entanglement island and replica wormhole. In section 3 we show that there is an apparent paradox caused by applying the quantum extremal surface formula to general dynamics of the radiation. In section 4 we discuss the how the paradox is resolved by realizing that some operators can introduce extra copies of geometry and wormholes connecting with them, which we name as miracle operators. Section 5 applies the miracle operator to address the ﬁrewall paradox in two diﬀerent setups. 1 Introduction Finally section 6 contains a summary and some further discussion. 1We thank Ahmed Almheiri for suggesting this term. 2 State reconstruction formula for an evaporating black hole We begin by an overview of the island formula, from the point of view of replica calcula- tion. [8, 9] For concreteness we consider a single-sided AdS black hole formed from collapse, which is coupled with a ﬂat space bath. Taking a subsystem A of the bath, the compu- tation of tr (ρn A) involves taking n replicas of the entire system, and introducing a cyclic permutation of the diﬀerent replicas in region A. The geometry of the system is a union Bg ∪Rn, with Rn the ﬂat space with branch covering at the boundary of A, and Bg the geometry of the AdS black hole. The gravitational path integral is deﬁned for the metric of Bg, with the boundary condition ﬁxed by Rn. Zn(A) = Z Bg Dg Z Bg∪Rn Dφe−AEH[g]−AQF T [φ,g] (2.1) (2.1) Here φ represents all matter ﬁelds. (More precisely, we should introduce n+m replicas and introduce the twist only in the ﬁrst n copies, and in the end take m →−n, as is discussed in ref. [14]. The m replicas can be omitted if we assume no saddle point will contain wormhole connecting them with each other or with the ﬁrst n replicas.) Assuming the path integral is dominated by a classical saddle point which preserves the replica symmetry Zn, the QFT contribution is given by the n-th Renyi entropy of the matter ﬁelds in a region IA, with I – 3 – Figure 1. (a) Illustration of the replica manifold with branch-covering at A and possibly an island region I. (b) In the analytic continuation to n →1 limit, illustration of the quantum extremal surface (2.3) which is the boundary of I. Figure 1. (a) Illustration of the replica manifold with branch-covering at A and possibly an island region I. (b) In the analytic continuation to n →1 limit, illustration of the quantum extremal surface (2.3) which is the boundary of I. JHEP01(2022)085 in the gravitational bulk. The gravitational contribution AEH[gsaddle] can be computed by taking a quotient Bg/Zn and evaluate the Einstein-Hilbert action. If there is a nontrivial replica wormhole in region I, the boundary of I becomes a conical singularity with co- dimension 2 and conical angle 2π n in the quotient manifold Bg/Zn. 2Of course in a more realist system with dynamical gravity everywhere, the Hilbert space does not factorize. The states ρA, σA deﬁned below should be viewed as “eﬀective” quantum states [14] describing correlations of quantum ﬁeld theory degrees of freedom of the radiation qubits. It is unclear how to rigorously deﬁne this if gravity is dynamical in the bath. 2 State reconstruction formula for an evaporating black hole In the limit n →1, this leads to the quantum extremal surface formula −log Zn(A) ≃AEH[gsaddle] + (n −1)S(n) QFT (IA) (2.2) S(A) ≃extI \|∂I\| 4GN + SQFT (IA) (2.3) (2.2) (2.3) with extI referring to extremization over choice of spatial region I. If there are multiple extremal surfaces, the one with minimal S(A) should be chosen. The discussion here is entirely parallel with the proof of the HRT formula in ordinary AdS/CFT [7, 15–17], except that the gravitational theory now has a diﬀerent boundary condition set by the radiation. We would like to emphasize that the discussion seems independent from the detail of dynamics of the radiation. Whether the radiation is a ﬂat space CFT or a quantum computer, we expect the quantum extremal surface derivation above to hold, as long as the gravity in the AdS region remains semi-classical. In addition to the entropy formula, the replica calculation can also tell us more about how operator reconstruction works. We ﬁrst clarify some notation. Denote the Hilbert space of the holographic CFT as HB, and that of the bath as HR (R for “radiation”), the Hilbert space of the entire system is HQG = HB ⊗HR.2 We will refer to this Hilbert space as the quantum gravity Hilbert space. A classical saddle point is denoted by M, which is a manifold that includes the asympotic AdS part and the attached ﬂat space part. For each M (with Lorenzian signature) and a given Cauchy surface, the path integral deﬁnes a state of the quantum ﬁeld theory on that background. We denote the Hilbert space of the quantum ﬁeld theory as HQFT . HQG and HQFT are the analog of boundary Hilbert space – 4 – Figure 2. Illustration of two states ρ, σ which are only diﬀerent by a local unitary UP acting on a small region P in the island. For example UP can ﬂip the spin of a particle. The QFT states are deﬁned on a Cauchy surface that includes A and I. JHEP01(2022)085 Figure 2. Illustration of two states ρ, σ which are only diﬀerent by a local unitary UP acting on a small region P in the island. For example UP can ﬂip the spin of a particle. The QFT states are deﬁned on a Cauchy surface that includes A and I. and bulk Hilbert space in AdS/CFT without bath. 2 State reconstruction formula for an evaporating black hole In the rest of the paper, we will use bold font Greek letters ρ, σ to denote states in HQFT , and use regular font Greek letters ρ, σ to denote those in the boundary-and-bath Hilbert space HQG. Now we consider two states ρ, σ ∈HQFT . When these two states are close to each other, such that the energy momentum tensor diﬀerence is order O(G0 N), the back reac- tion to geometry is small. In the replica calculation, we can compute quantity such as tr ρn−m A σm A and expect that it is dominated by the same saddle point as tr (ρn A). In this case one obtains tr ρn−m A σm A ≃e−Antr ρn−m AI σm AI ; (2.4) (2.4) with An the contribution of the Einstein-Hilbert action at the saddle point manifold. Con- sequently, we obtain ∥ρA −σA∥2n = tr h (ρA −σA)2ni ≃e−A2n ∥ρAI −σAI∥2n (2.5) (2.5) for positive integer n. In the analytic continuation to n →1/2, A2n vanishes: A2n ≃(2n −1) \|∂I\| 4GN (2.6) (2.6) Therefore we obtain with I the entanglement island of A. As an example, consider ρ as the state of an evaporating black hole coupled with the bath, in which a region of the radiation A has an entanglement island I. σ = UP ρU† I is diﬀerent from ρ by a local unitary acting in a small spacetime region P ∈I, as is illustrated in ﬁgure 2. Eq. (2.7) tells us that the diﬀerence between ρ and σ is preserved in the radiation region A. For example, if UP rotates the spin of a qubit from z-direction eigenstates up to down, one can measure the spin z-component operator ZP of this qubit to distinguish ρAI and σAI. Then eq. (2.7) guarantees that there is a projection operator PA in A such that JHEP01(2022)085 \|tr (PAρA) −tr (PAσA)\| = 1 2 ∥ρA −σA∥1 = 1 2 ∥ρAI −σAI∥1 ≥1 2 \|tr (ZP ρAI) −tr (ZP σAI)\| (2.8) (2.8) Thus eq. (2.7) guarantees that ρA, σA can be distinguished by a measurement in region A, to the same degree that ρAI and σAI can be distinguished on AI. In the next section, we will discuss a more explicit setup for extracting such information, which leads to an apparent contradiction. Therefore we obtain ∥ρA −σA∥1 = ∥ρAI −σAI∥1 This discussion is a simple generalization of ref. [11] for relative entropy (see also [9, 10] for related discussions). Alternatively, we could also prove eq. (2.7) using the conclusion of ref. [11]: since the mapping C : ρAI →ρA preserves the relative entropy between two states ρ, σ in the code subspace, it is invertable and thus preserves the trace distance between two states. This result plays a central role in this paper, so we would like to summarize it in the following theorem: Theorem 1 State reconstruction formula. For two states ρ, σ ∈HQG, assume that the gravitational path integral in the calculation of tr (ρn A) and tr (σn A) are both dominated – 5 – by the same smooth manifold Mn preserving the cyclic permutation symmetry between replicas, up to corrections O(GN), then ∥ρA −σA∥1 = ∥ρAI −σAI∥1 (2.7) (2.7) 3.1 A no-go theorem If W has a nontrivial entanglement island I, the necessary condition is SQFT (WI) < \|∂I\| 4GN + SQFT (WI) < SQFT (W) (3.2) (3.2) However, in the situation we consider, W and the remainder of the system (R and black hole B) is coupled with LOCC, such that the state remains separable at all time: However, in the situation we consider, W and the remainder of the system (R and black hole B) is coupled with LOCC, such that the state remains separable at all time: ρQFT = X i piρi QFT (BR)ρi QFT (W) (3.3) (3.3) Therefore for any subsystem I, ρQFT (IW) is still separable. Consequently, the conditional entropy is always non-negative (see appendix B): HQFT (I\|W) ≡SQFT (WI) −SQFT (W) ≥0 (3.4) (3.4) Consequently, W always has a trivial island. We summarize this conclusion in the following theorem: Theorem 2 If i) a subsystem W of the radiation only couples with the remainder of the system (including the radiation and the black hole) through LOCC, and the initial state is separable between W and its complement, and ii) semiclassical approximation applies to the entropy calculation of W, then W has no entanglement island. 3.1 A no-go theorem As we discussed earlier, the island formula (2.3) only requires the geometry of the gravita- tional system to have small ﬂuctuation, independent from the dynamics of the radiation. In this subsection, we will consider a special class of radiation system, as is shown in ﬁg- ure 3. The radiation system R consists of two subsystems R1 and W. R1 is the “ordinary” radiation system, which is a ﬂat space QFT coupled with the black hole. W is an ancilla that only couples with R1 by local operations and classical communication (LOCC). The initial state of R is a direct product state ρ0R1 ⊗ρ0W . Physically, we can think W as a model of lab equipment of an observer we use to measure the radiation. We assume the observer only has access to the radiation system R1 through LOCC, which physically involves multiple rounds of carrying quantum measurements to R1 and applying unitaries on it. We include the deﬁnition of LOCC in appendix A. This setup is an example of an incoherent quantum algorithmic measurement (QUALM) deﬁned in ref. [18]. Now we study the entropy of the W subsystem. With the assumption that the island formula applies, we obtain S(W) = extI \|∂I\| 4GN + SQFT (WI) (3.1) (3.1) – 6 – Figure 3. (a) Illustration of the ancilla W which only couples with the rest of the bath R1 through LOCC. (b) A quantum circuit representation of the same setup, with black hole B couples with R1 through quantum gates, while R1 and W are coupled only by LOCC. Figure 3. (a) Illustration of the ancilla W which only couples with the rest of the bath R1 through LOCC. (b) A quantum circuit representation of the same setup, with black hole B couples with R1 through quantum gates, while R1 and W are coupled only by LOCC. JHEP01(2022)085 Empty I = ∅is always a saddle point, with contribution SQFT (W). If W has a nontrivial entanglement island I, the necessary condition is Empty I = ∅is always a saddle point, with contribution SQFT (W). 3.2 An apparent paradox Theorem 1 and Theorem 2 both seem to be general and are natural consequences of the island formula, but in the following we will show that there is a contradiction when we try to apply them to a particular setup. † Consider the setup we discussed earlier in ﬁgure 2, with two states ρ and σ = UP ρU† P that are only diﬀerent by a unitary applying to a small region in the island I of a radiation region A. In addition, we consider an ancilla W that is coupled with A through the following quantum channel: M (ρA ⊗\|0W ⟩⟨0W \|) = PAρAPA ⊗\|1W ⟩⟨1W \| + (IA −PA) ρA (IA −PA) ⊗\|0W ⟩⟨0W \| (3.5) CM (ρA ⊗\|0W ⟩⟨0W \|) = PAρAPA ⊗\|1W ⟩⟨1W \| + (IA −PA) ρA (IA −PA) ⊗\|0W ⟩⟨0W \| (3.5) (3.5) – 7 – Figure 4. Illustration of the particular setup in subsection (3.2) in Penrose diagram (a) and quantum circuit (b). We consider two states diﬀerent by an interior unitary UP , as was discussed in ﬁgure 2. In addition, a unitary UM (deﬁned by eq. (3.6) applied to region A and two ancilla qubits W, ˜W measures A and records the result on W. The green and red horizontal dashed lines indicate the time t1 before applying UM, and t2 right after applying UM. What is relevant to our discussion is the application of state reconstruction formula (2.7) for A at time t1 and W at time t2. JHEP01(2022)085 Figure 4. Illustration of the particular setup in subsection (3.2) in Penrose diagram (a) and quantum circuit (b). We consider two states diﬀerent by an interior unitary UP , as was discussed in ﬁgure 2. In addition, a unitary UM (deﬁned by eq. (3.6) applied to region A and two ancilla qubits W, ˜W measures A and records the result on W. The green and red horizontal dashed lines indicate the time t1 before applying UM, and t2 right after applying UM. What is relevant to our discussion is the application of state reconstruction formula (2.7) for A at time t1 and W at time t2. Here PA is the particular projection operator that distinguish ρA and σA optimally, such that tr (PAρA)−tr (PAσA) = 1 2 ∥ρA −σA∥1. The channel CM measures A with the projector PA, and save a copy of the measurement result on W. 3.2 An apparent paradox CM is an LOCC. We would like to emphasize that although CM is not a unitary operator, it could be realized by an unitary in a slightly bigger system. For example if we introduce another qubit ˜W, we can deﬁne the unitary on AW ˜W: UM = PA ⊗XW ⊗X ˜ W + (IA −PA) ⊗IW ⊗I ˜ W (3.6) (3.6) with XW and X ˜ W the Pauli x operator that ﬂips between \|0⟩and \|1⟩states. Applying UM to the state ρA ⊗\|0W ⟩⟨0W \| ⊗ 0 ˜ W 0 ˜ W and tracing over ˜W leads to CM (ρA ⊗\|0W ⟩⟨0W \|). We would like to emphasize this point to clarify that introducing CM in the discussion does not imply we have violated unitarity in R, as long as A, W, ˜W are all part of the bath system R. The setup is illustrated in ﬁgure 4. Now we apply the state reconstruction formula (Theorem 1) to two diﬀerent region 1. For the state before UM is applied, deﬁned at time t1 in ﬁgure 4, applying the state reconstruction formula to region A leads to ∥ρA −σA∥1 = ∥ρAI −σAI∥1, which is order 1. 2. For the state after UM is applied, deﬁned at time t2 in ﬁgure 4, applying the state reconstruction formula to W leads to 2. For the state after UM is applied, deﬁned at time t2 in ﬁgure 4, applying the state reconstruction formula to W leads to 2. For the state after UM is applied, deﬁned at time t2 in ﬁgure 4, applying the state reconstruction formula to W leads to ∥ρW −σW ∥1 = ∥ρW −σW ∥1 , (3.7) (3.7) since W still has only classical correlation with the rest of the system, and thus has no entanglement island according to Theorem 2. Furthermore, remember that ρW and σW are the states of W in QFT, which are obtained by ﬁxing the geometry and – 8 – deﬁning the QFT state on a Cauchy surface by a QFT path integral. Therefore a unitary UP that is acting in a spacetime region spacelike separated from W will have trivial eﬀect on the reduced state of W, leading to ρW = σW . Therefore eq. (3.7) implies ∥ρW −σW ∥1 = 0. 3.2 An apparent paradox However, the two conclusions above are in direct contradiction, because W has learned about the measurement result from A, and therefore “inherited” a nontrivial one-norm distance from that between ρA and σA. To see that, we write down the state of W after the coupling: JHEP01(2022)085 ρW = trA [CM (ρA ⊗\|0W ⟩⟨0W \|)] = pρ \|1W ⟩⟨1W \| + (1 −pρ) \|0W ⟩⟨0W \| (3.8) with pρ = tr (PAρA) (3.9) (3.8) (3.9) (3.8) (3.8) (3.9) (3.9) Similarly, when the state of A before measurement is σA, the state of W after measurement is σW = pσ \|1W ⟩⟨1W \|+(1−pσ) \|0W ⟩⟨0W \| of the same form, with pσ = tr(PAσA). Therefore we have ∥ρW −σW ∥1 = 2 \|pρ −pσ\| = ∥ρA −σA∥1 (3.10) (3.10) Physically, eq. (3.10) tells us that one-norm distance is transferable by classical commu- nication, because it is a measure of classical information one can learn by an optimal measurement. This transferability thus suggests ∥ρW −σW ∥1 should be order 1, in direct conﬂict with the result of state reconstruction formula applied to W. Therefore we con- clude that to avoid inconsistency, at least for in one of the two states we discussed (the state of A at t1 and the state of W at t2), the assumptions of Theorem 1 must fail. We will discuss more about the interpretation of this apparent paradox in next section. 4 The miracle operators Let us summarize the problem again. The replica calculation that leads to the QES formula also predicts eq. (2.7), which tells us that small perturbation in I can be reconstructed in A. On the other hand, the same formula suggests that an ancilla which only couples to the black hole and radiation by LOCC cannot probe such small perturbation anywhere space-like separated from A and the ancilla, because it does not have an entanglement island (Theorem 2). The only assumption we have used to achieve this paradox is the assumptions of Theorem 1, that the gravitational path integral is dominated by a smooth saddle point manifold in the replica calculation of both SA and SW . (Note that we do not need to assume the saddle point manifold to be the same one for these two calculations. The paradox remains as long as the manifolds involved are smooth and replica symmetric, even if we are allowed to consider a large back-reaction caused by UM.) Therefore by contradiction we have proved that this assumption must fail for at least one of the calculation. If we assume the SA calculation is correct (which occurs before W got involved), then we have to conclude that after applying UM, the calculation of tr (ρn W ) is not dominated by any smooth replica symmetric saddle point manifold. The problem remains if we replace a single saddle point with a sum over multiple saddle points. Since the contribution of each possible saddle – 9 – Figure 5. Illustration of the replica calculation for (a) tr(ρA ˆO)n for a regular operator ˆO, (b) tr (ρn A) and (c) tr ρAZ(k) A n with Z(k) A deﬁned in eq. (4.3). The illustration shows the case n = 2, k = 1. The part of Penrose diagram below a Cauchy surface represents the QFT state at that surface, prepared by the QFT path integral on the given background geometry. The red bridge connecting diﬀerent replicas in (b) and (c) is introduced by multiplying ρA with ρA or Z(k) A . The green bridge is the replica wormhole. Note that in (c) replica wormhole connects the physical copies (1, 1) and (2, 1) with simulated copies (1, 2) and (2, 2) in blue, which represent Z(k) A . JHEP01(2022)085 Figure 5. 4 The miracle operators Illustration of the replica calculation for (a) tr(ρA ˆO)n for a regular operator ˆO, (b) tr (ρn A) and (c) tr ρAZ(k) A n with Z(k) A deﬁned in eq. (4.3). The illustration shows the case n = 2, k = 1. The part of Penrose diagram below a Cauchy surface represents the QFT state at that surface, prepared by the QFT path integral on the given background geometry. The red bridge connecting diﬀerent replicas in (b) and (c) is introduced by multiplying ρA with ρA or Z(k) A . The green bridge is the replica wormhole. Note that in (c) replica wormhole connects the physical copies (1, 1) and (2, 1) with simulated copies (1, 2) and (2, 2) in blue, which represent Z(k) A . point (for the calculation of W Renyi entropy) with a nontrivial island I is suppressed by a factor e−∆S with ∆S ∼O (1/GN), the contribution remains suppressed even if we sum over polynomial (in 1/GN) number of saddle points. In other words, applying the particular unitary UM to the bath system and computing tr (ρn W ) (which can be interpreted as measuring the expectation value of a twist operator in the n-replica system) change the behavior of the gravitational path integral in a way that is much more dramatic than an ordinary back-reaction. From our discussion one can see that this problem occurs whenever we are carrying a measurement such as PA, which reveals nontrivial information about the interior. In the following we will refer to such interior-information-revealing measurement operators as “miracle operators”, due to their dramatic eﬀect to the spacetime geometry. To understand what happens when we apply miracle operators, we write down ex- plicitly the form of the optimal measurement PA. For any two states ρA, σA, if we write – 10 – ρA −σA in the diagonal basis ρA −σA = P n λn \|n⟩⟨n\|, then A −σA in the diagonal basis ρA −σA = P n λn \|n⟩⟨n\|, then PA = X n 1 2 (1 + sgn (λn)) \|n⟩⟨n\| (4.1) (4.1) with sgn(λn) = +1, 0, −1 for λn > 0, λn = 0, λn < 0, respectively.3 Therefore we can introduce the replica trick and express PA as: 2PA −IA = sgn (ρA −σA) = (ρA −σA)2n−1 n→1 2 (4.2) (4.2) from which we can see explicitly that PA involves gravitational path integral. 3If ρA −σA is not full rank, PA has eigenvalues 0, 1 and 1 2, so that it is not a projector. Correspondingly there are three possible measurement results W needs to record. However, when the input state is ρA or σA, only two measurement results will appear. All our discussion remains valid if we allow a general input state and deﬁne W to have Hilbert space dimension 3. 4 The miracle operators Consequently, the evaluation of PA can involve replica wormholes which connect the interior of the original black hole and those in PA. The operators reconstructed using Petz map [8] are also examples of miracle operators. The operator we consider is simpler than the Petz map case because we focus on a simpler task of distinguishing two particular states. JHEP01(2022)085 Using this replica trick, we can see the reason of the contradiction we ﬁnd in last section. For this purpose, deﬁne Z(k) A = (ρA −σA)2k−1 (4.3) (4.3) for integer k, and introduce a W that is coupled to A by the following LOCC channel: C(k) M (ρA ⊗\|0W ⟩⟨0W \|) = X(k) A+ρAX(k) A+ ⊗\|1W ⟩⟨1W \| + X(k) A−ρAX(k) A−⊗\|0W ⟩⟨0W \| (4.4) with X(k) A± = r1 2 IA ± Z(k) A (4.5) (4.5) Note that IA ± Z(k) A is positive, so that the square root operator X(k) A± is uniquely deﬁned and Hermitian. This channel carries a positive operator-valued measurement (POVM) to A and store the result in W. After applying the channel, the state of W is ρW = p(k) ρ \|1W ⟩⟨1W \| + 1 −p(k) ρ \|0W ⟩⟨0W \| (4.6) p(k) ρ = 1 2 1 + tr Z(k) A ρA (4.7) (4.6) (4.7) Now if we compute tr (ρn W ), we obtain tr (ρn W ) = 2−n h 1 + tr Z(k) A ρA n + 1 −tr Z(k) A ρA ni (4.8) (4.8) which involves tr Z(k) A ρA m = tr (ρA −σA)2k−1 ρA m for m = 0, 1, . . . , n. Each term in this trace involves 2km copies of the original system. Applying the replica wormhole calculation to this computation, it is easy to convince ourselves that replica wormhole will only appear between ρA and Z(k) A within each trace. We can label the 2km replica by (a, s) with a = 1, 2, . . . , m and s = 1, 2, . . . , 2k, with (a, 1) the m copies of ρA, and (a, s), s ≥2 – 11 – the 2k −1 copies of ρA or σA in the operator Z(k) A . 4 The miracle operators The replica wormhole will only connect (a, s) with (a, s + 1) (with s = 2k + 1 identiﬁed with s = 1). This is illustrated in ﬁgure 5. Compare this calculation with the “ordinary” case, such as the calculation of tr(ρn A), we see that the dominant saddle point conﬁguration contains wormholes connecting the “physical geometry” with “simulated geometry” in Z(k) A . This is the possibility that was missed in the derivation of the one norm correspondence in eq. (2.7). In summary, what we learned from miracle operators is that carrying a particular mea- surement can not only modify the bulk geometry of the black hole, but even modify the boundary condition of the geometry. Instead of being a smooth geometry with boundary condition set by the ﬂat space region, the geometry now contains wormholes that con- nect the system with the operator being measured. On comparison, if we view tr(ρn A) as an expectation value of cyclic permutation operator XA in n copies of the system, then this operator induces replica wormhole between diﬀerent replicas, but does not change the boundary condition of the geometry. If we take the point of view that gravity is an ensem- ble theory [19–33], then we believe that miracle operators diﬀer from ordinary operators by “knowing” the random parameters in the gravity theory. [34] For example, if we are consid- ering a Sachdev-Ye-Kitaev (SYK) model [35–38] coupled with bath, the miracle operators have to depend on the random coupling J in the SYK model. In more general cases, the natural of the random parameters remain an open question, but results in random tensor network models [39] suggest that such random parameters exist not only in black hole systems but also in spacetime geometry without black hole, such as the AdS vacuum [34]. JHEP01(2022)085 Another point we want to highlight is that such a dramatic eﬀect to spacetime geometry occurs even if we are only carrying a binary measurement that learns about a single bit of information in the interior. 5 The ﬁrewall We would like to investigate the implication of our result in understanding the ﬁrewall paradox [12, 13]. The ﬁrewall paradox points out that after Page time, a Hawking radiation mode b is entangled with the earlier radiation, so that if it is also entangled with its partner behind the horizon ˜b, this will violate strong subadditivity (SSA) of entanglement entropy. The entanglement island result points out that the SSA argument does not apply, since ˜b in the island is not independent from the earlier radiation. In ﬁgure 6 (a), if ˜b is in the entanglement island of A region, and b is outside A region, then it is ok to have b entangled with ˜b and also entangled with A, since ˜b is actually part of A. However, we would like to address more concretely whether an infalling observer will see a ﬁrewall. We will study two diﬀerent setups, illustrated in ﬁgure 6 (a) and ﬁgure 7 (a). In both cases there is an EPR pair of modes b,˜b and ˜b is in the entanglement island of a radiation region A. The diﬀerence is that in the ﬁrst setup the partner b is outside region A while in the second setup it’s in A. In both cases, an infalling observer X carries a Bell basis measurement to detect if b˜b is in one of the maximally entangled states. The ﬁrst setup addresses whether a measurement on A can act on ˜b and destroy the entanglement between b and ˜b, i.e. creates a ﬁrewall. The second setup addresses whether an entanglement checking experiment can be applied – 12 – Figure 6. (a) Illustration of a pair of qubits b,˜b, when ˜b ∈I and b /∈A. An infalling observer X carries a binary measurement that checks if b˜b is in the particular entangled state. (b) Illustration of the replica geometry that applies operator P ↑↓ A which distinguishes the two states with opposite spin of ˜b. The green and red lines represent the trajectory of b and ˜b in the two replicas respectively. (In general there are n replicas. We draw n = 2 here for simplicity.) Since only one of them crosses the branchcut line, X will not see the partner ˜b but see the copy of ˜b in a diﬀerent replica, leading to failure of entanglement checking. JHEP01(2022)085 Figure 6. 5 The ﬁrewall (a) Illustration of a pair of qubits b,˜b, when ˜b ∈I and b /∈A. An infalling observer X carries a binary measurement that checks if b˜b is in the particular entangled state. (b) Illustration of the replica geometry that applies operator P ↑↓ A which distinguishes the two states with opposite spin of ˜b. The green and red lines represent the trajectory of b and ˜b in the two replicas respectively. (In general there are n replicas. We draw n = 2 here for simplicity.) Since only one of them crosses the branchcut line, X will not see the partner ˜b but see the copy of ˜b in a diﬀerent replica, leading to failure of entanglement checking. to A, now including the information of both b and ˜b, and whether such an entanglement checking will create a ﬁrewall seen by the infalling observer X. In the following we will discuss these two setups in two subsections. 5.1 Entanglement breaking measurement Therefore in the saddle point geometry, there is a branchcut at A and I, where diﬀerent replicas are connected by cyclic permutation. In the ﬁrst replica there is an entangled pair of modes, while in the other 2n −1 replicas there is a pure state of ˜bb. One should remember that the infalling observer X is also part of the system, so it should also be present in each replica. Every term in the righthand side of this equation involves 2n replicas of the original system, with a branchcut at A in the same way as in Renyi 2n entropy calculation. In the limit when the back reaction caused by the modes ˜b, b is negligible, the 2n replica geometry contains the replica wormhole in the same way as in the Renyi entropy calculation, as is illustrated in ﬁgure 6 (b). Therefore in the saddle point geometry, there is a branchcut at A and I, where diﬀerent replicas are connected by cyclic permutation. In the ﬁrst replica there is an entangled pair of modes, while in the other 2n −1 replicas there is a pure state of ˜bb. One should remember that the infalling observer X is also part of the system, so it should also be present in each replica. JHEP01(2022)085 Now it is important to remember that the infalling observer is also part of the system, which should appear in each replica. The question is whether the entanglement checking measurement PX still succeeds in this replicated geometry. As is illustrated in ﬁgure 6 (b), ˜b goes across the branchcut line in I, but b does not, so that when the infalling observer in k-th replica brings b in that replica to the interior, it meets with the ˜b mode in the k +1-th replica. Consequently, none of the infalling observers will see an entangled pair of b˜b. In other words, the entanglement checking fails for all replica calculation with 2n replicas. The observer that was originally in replica 1 (the one with an entangled pair) will see a state ρb˜b = 1 2Ib ⊗(\|σ⟩⟨σ\|)˜b , σ =↑or ↓ (5.6) (5.6) such that ⟨ˆPX⟩= 1 4 (5.7) (5.7) Since this result applies to all replica number 2n, it is reasonable (although not rig- orous) to suggest that the same holds when taking analytic continuation n →1/2, which means the infalling observer will see the b˜b entanglement destroyed. 5.1 Entanglement breaking measurement To be concrete, let us assume b and ˜b are local wavepackets of bosons, such as photons. Denote ˆb† σ and ˆ˜b† σ as the creation operators of these two modes respectively, with σ =↑, ↓ two states of spin or other internal states. ˆb†,ˆ˜b† are deﬁned in the QFT Hilbert space. Denote the state without this pair as ρ, then we can create an EPR pair in one of the Bell basis states: σ = ˆ∆† ±ρ ˆ∆± ˆ∆± = 1 √ 2 ˆb↑ˆ˜b↓± ˆb↑ˆ˜b↓ (5.1) (5.1) In this subsection we only need one of them, which we will pick as σ−. In next subsection we will need both. In this subsection we only need one of them, which we will pick as σ−. In next subsection we will need both. Now when the system is in state σ−, an infalling observer X brings b into the interior and check whether b˜b is in the particular entangled state created by ˆ∆−. The measurement operator is PX = ˆ∆† −ˆ∆− (5.2) (5.2) Without other intervention to the system, PX will return eigenvalue 1 with probability 1. X records the result of the measurement. Now we would like to carry a measurement on A that measures the spin z component of ˜b. For that purpose let us consider two other states σ↑= ˆ˜b† ↑ˆb† ↓ρˆb↓ˆ˜b↑, σ↓= ˆ˜b† ↓ˆb† ↑ρˆb↑ˆ˜b↓ (5.3) (5.3) – 13 – – 13 – which contains Z eigenstates of ˜b qubit. Then we can deﬁne the optimal projection operator on A deﬁned in eq. (4.2) that distinguishes these two states: ˆZA ≡2P ↑↓ A −IA = sgn (σ↑,A −σ↓,A) (5.4) (5.4) We implement this operator by the replica trick in eq. (4.2): We implement this operator by the replica trick in eq. (4.2): We implement this operator by the replica trick in eq. (4.2): tr σ−,A ˆZA = tr h σ−,A (σ↑,A −σ↓,A)2n−1i n→1 2 (5.5) (5.5) Every term in the righthand side of this equation involves 2n replicas of the original system, with a branchcut at A in the same way as in Renyi 2n entropy calculation. In the limit when the back reaction caused by the modes ˜b, b is negligible, the 2n replica geometry contains the replica wormhole in the same way as in the Renyi entropy calculation, as is illustrated in ﬁgure 6 (b). 5.2 Entanglement checking Instead of measuring the state of ˜b, we would like to consider an entanglement checking measurement applied to the exterior. In the setup of ﬁgure 7 (a), we consider a situation when b has already entered region A. ˜b is in the entanglement island of A, so that it should be possible to check the entanglement of b˜b in region A. For example we can apply the Petz map to the operator PX applied by the infalling observer. Instead of Petz map, we prefer to make use of the optimal projector construction. For that purpose we consider two states σ±,QFT with two orthogonal maximally entangled states of b˜b, deﬁned in eq. (5.1). Then we can deﬁne the projection operator P +− A that distinguishes these two states JHEP01(2022)085 ˆFA ≡2P +− A −IA = sgn (σ+,A −σ−,A) (5.8) ˆFA ≡2P +− A −IA = sgn (σ+,A −σ−,A) (5.8) Now if we take the state σ−and measure ˆFA, by construction we obtain e take the state σ−and measure ˆFA, by construction we obtain tr σ−,A ˆFA = tr h σ−,A (σ+,A −σ−,A)2n−1i n→1 2 = −1 (5.9) tr σ−,A ˆFA = tr h σ−,A (σ+,A −σ−,A)2n−1i n→1 2 = − (5.9) This is because one term in the expansion −tr σ2n −,A →−1 when n →1 2, and all other terms are of the form tr σa +,Aσb −,A . . . ≃e−Agravtr σa +,AIσb −,AI . . . (5.10) tr σa +,Aσb −,A . . . ≃e−Agravtr σa +,AIσb −,AI . . . (5.10) (5.10) which contains a products of σ+,AI and σ−,AI. Since σ+ and σ−contains two orthogonal states of the b˜b pair, we have σ+,AIσ−,AI = 0, so that these terms all vanish. (The gravity contribution Agrav ≃(2n −1) \|∂I\| 4GN in the n →1 2 limit.) In parallel with the discussion in previous subsection, we can ask what happens to infalling observer X which carries the measurement ˆPX in eq. (5.2). Diﬀerent from the previous setup, now b goes through the branchcut at A and ˜b goes through that at I, so that they can still meet in the same replica. The observer X in each replica will observe a maximal entangled state \|−⟩since the only nonvanishing term contains σ2n −,A. 5.1 Entanglement breaking measurement In this sense the measurement P ↑↓ A creates a ﬁrewall at the quantum extremal surface, although it is not a ﬁrewall in the sense of energy. Degrees of freedom across the quantum extremal surface, as are seen by the infalling observer, are not entangled because they are entangled with partners in another replica, but this does not require a high energy excitation. It should be noted that this ﬁrewall creation occurs whether or not X jumps in be- fore or after the measurement on A is carried. This looks nonlocal but does not violate causality, because the action we take on A is a measurement. In ordinary quantum sys- tems, it is also allowed that two space-like separated measurements have correlated results. The calculation above shows that there is a ﬁrewall condition on a given output of the measurement P ↑↓ A . – 14 – 5.2 Entanglement checking Therefore we would argue that in the analytic continuation to n →1 2, the entanglement checking experiment by observer X is successful with probability 1. It is interesting to note that the trajectory of b and ˜b together form a noncontractable loop in the replica manifold (i.e. a nontrivial generator of the ﬁrst homotopy group). The diﬀerence between the two setups with b outside A (ﬁgure 6) and b inside A (ﬁgure 7) is topological. In short, we see that both the exterior and infalling entanglement checking can succeed, at least in this particular choice of measurements. Roughly speaking, this is possible because the measurement in A has to create other copies of the universe, including the infalling observer. Another implicit assumption (which seems physically reasonable) is that the infalling observer is deﬁned for each replica geometry separately, and there is no “super-observer” that can take diﬀerent replicas and apply a joint measurement. Our result seems to be consistent with the result of A. Almheiri [40] in the ﬁnal state projection model [41]. The gravitational calculation clariﬁes that entanglement checking experiment in A does not have to apply a projection to many qubits, but can extract only one bit of information about whether b˜b is in the particular entangled state \|−⟩. We realize that the argument about analytic continuation in this section is not rigorous. More careful analysis is a task for future work. – 15 – Figure 7. (a) Illustration of a pair of qubits b,˜b with ˜b ∈I, b ∈A. An infalling observer X measures whether b˜b is in a particular maximally entangled states such as \|−⟩. (b) Illustration of the replica geometry that applies operator P +− A (deﬁned in eq. (5.8)) which distinguishes the two states \|+⟩, \|−⟩in A. The green and red lines represent the trajectory of b and ˜b in the two replicas respectively, in the same way as in ﬁgure 6. We see that the entanglement checking of X will still succeed because both b and ˜b are acted by the same cyclic permutation operation. JHEP01(2022)085 Figure 7. (a) Illustration of a pair of qubits b,˜b with ˜b ∈I, b ∈A. An infalling observer X measures whether b˜b is in a particular maximally entangled states such as \|−⟩. (b) Illustration of the replica geometry that applies operator P +− A (deﬁned in eq. (5.8)) which distinguishes the two states \|+⟩, \|−⟩in A. 5.2 Entanglement checking The green and red lines represent the trajectory of b and ˜b in the two replicas respectively, in the same way as in ﬁgure 6. We see that the entanglement checking of X will still succeed because both b and ˜b are acted by the same cyclic permutation operation. 6 Further discussion and conclusion In summary, we have shown that there is a paradox if we assume the QES formula applies to an AdS black hole coupled with a radiation system with arbitrary dynamics. When the geometry is semiclassical, and the QES formula applies, information about the interior is not available to any observer who is only coupled with the radiation through LOCC (and does not have quantum entanglement with the radiation and black hole to start with). On the other hand, information in the entanglement island of a region A can be reconstructed in A, which suggests that there must exist miracle operators which induce qualitative change of bulk geometry when they are applied to A, even if only one bit of information is obtained in this measurement. We have explicitly constructed such an operator which distinguishes two given states optimally, and show that in a replica calculation, the large geometry ﬂuctuation corresponds to replica wormhole connecting the original universe with new copies of the universe contained in the projection operator itself. This construction helps us to address some questions in the ﬁrewall paradox. We show that a projective measurement in a radiation region A can break the entanglement between two modes across the QES surface, i.e. the boundary of the entanglement island. On the other hand, an entanglement checking measurement that veriﬁes the entanglement between a Hawking qubit b and the earlier radiation does not aﬀect the entanglement checking measurement of an infalling observer, at least in the particular setup we consider. There are many open questions related to this setup. One question is whether the diﬀerence between regular operators and miracle operators is a signature of gravity theory being an ensemble theory. [8, 19, 20, 23, 42] If gravity is an eﬀective description of an ensemble average over a family of boundary theories, then it is natural to distinguish between operators that does and does not depend on random parameters in the boundary theory. Only the latter will possibly detect the interior. A nontrivial question is whether wormholes connecting physical systems with simulated systems are still well-deﬁned if there is no ensemble averaging. – 16 – Another question is whether miracle operators can still be deﬁned when gravity in the bath is dynamical. For example, in a ﬂat space Schwarzchild black hole, can one deﬁne miracle operators that create simulated copies of spacetime and replica wormholes? Ref. 6 Further discussion and conclusion [14] proposed a generalization of the entropy calculation of a bath region to the dynamical gravity case, but in that generalization the entropy is a measure of “eﬀective” uncertainty in the state of low energy degrees of freedom. Roughly speaking, we can view the eﬀective entropy as the entropy of a “conditional state” of low energy degrees of freedom, with the condition that a classical background geometry is observed by a family of observers in the bath region. It is unclear to the author how this discussion will be modiﬁed if we more rigorously taken into account the nonlocality of quantum gravity Hilbert space [43– 46]. For example, we notice that ref. [43] proposed that entanglement islands can only be rigorously deﬁned in systems with massive gravitons. JHEP01(2022)085 Ref. [47] pointed out that in the ﬁnal state projection proposal [41] there is a problem with probability interpretation of the measurement carried by an infalling observer, be- cause of the failure of decoherence between diﬀerent histories. We have not addressed this problem in our discussion of infalling observer. For example consider an observer X made of a large number of spins with the initial state (\|↑⟩⟨↑\|)(⊗M). The measurement process is a unitary operator UM = PX ⊗IX + (I −PX) ⊗XX (6.1) UM = PX ⊗IX + (I −PX) ⊗XX (6.1) with XX ﬂips all the spins in X. In ordinary quantum measurement, we apply UM to obtain UMρb˜b ⊗(\|↑⟩⟨↑\|)(⊗M) U† M and then trace over any one of the M qubits in X. This removes the oﬀ-diagonal terms and leads to a measurement channel with XX ﬂips all the spins in X. In ordinary quantum measurement, we apply UM to obtain UMρb˜b ⊗(\|↑⟩⟨↑\|)(⊗M) U† M and then trace over any one of the M qubits in X. This removes the oﬀ-diagonal terms and leads to a measurement channel M ρb˜b⊗(\|↑⟩⟨↑\|)(⊗M) =PXρb˜bPX ⊗(\|↑⟩⟨↑\|)⊗(M−1)+(I−PX)ρb˜b(I−PX)⊗(\|↓⟩⟨↓\|)⊗(M−1) (6.2) (6.2) When X is behind the horizon, it is unclear whether we are still allowed to do the partial trace. On the other hand, considering that horizon is not a local concept, it is possible that we are crossing the horizon of a giant black hole now, which should not aﬀect the probabilis- tic interpretation of the quantum experiments that occur in a physics lab at this moment. 6 Further discussion and conclusion This suggests that we should be able to talk about what an infalling observer sees, in the same way as an outside observer. A more rigorous formalism is a task for future research. When X is behind the horizon, it is unclear whether we are still allowed to do the partial trace. On the other hand, considering that horizon is not a local concept, it is possible that we are crossing the horizon of a giant black hole now, which should not aﬀect the probabilis- tic interpretation of the quantum experiments that occur in a physics lab at this moment. This suggests that we should be able to talk about what an infalling observer sees, in the same way as an outside observer. A more rigorous formalism is a task for future research. Another question is how rare are the set of miracle operators. We know the special constructions discussed here. It will be interesting to have a more precise statement about how the miracle operators are very rare. In the entanglement checking setup in ﬁgure 7, if we deﬁne the optimal operator for two states σ1 and σ2 which are not orthogonal, all terms in eq. (5.9) are nonzero. This creates a linear superposition of diﬀerent “branches” in which infalling observer X can observe diﬀerent results. This is not immediately contradictory because the two states are not entirely distinguishable so the observer always has a chance to mistaken state 1 as state 2. However, it requires a more thorough investigation to understand whether all observations of the infalling observer are consistent with quantum mechanics. This question seems to be related to the discussion about state-dependence of the reconstruction map [48, 49]. – 17 – It is interesting to think this setup as an example of quantum algorithmic measurement (QUALM) [18]. A QUALM is a quantum algorithm that contains a known part controlled by the experimentalist, and an partially unknown part controlled by nature that contains some hidden parameters. The purpose of the algorithm is to ﬁnd out some message about the unknown parameters. In the black hole problem, as is illustrated in ﬁgure 3, the known part is the coupling between ancilla W and the radiation R, and the dynamics of R itself. The partially unknown part is the black hole B and its coupling with radiation R. 6 Further discussion and conclusion In the QUALM language, what we have shown is that the task of ﬁnding out the information in an interior qubit is very diﬃcult for an observer with incoherent access to the system, if the observer does not know the black hole microstate. It is reasonable to guess that the diﬃculty (i.e. QUALM complexity of the task) is exponential in 1/GN, since a saddle point with nontrivial island only has a contribution to the partition function that is exponential in 1/GN. It is interesting to relate this result to that of ref. [18] and see whether it is possible to make our result more rigorous in the QUALM framework. JHEP01(2022)085 Acknowledgments We would like to thank Ahmed Almheiri, Yiming Chen, Xi Dong, Daniel Harlow, Hong Liu, Donald Marolf, Daniel Ranard, Douglas Stanford, Zhenbin Yang and Pengfei Zhang for helpful discussions. This work is supported by the Simons Foundation, and in part by the DOE Oﬃce of Science, Oﬃce of High Energy Physics, the grant de-sc0019380. This work was developed in part at the Kavli Institute for Theoretical Physics during the workshop “Gravitational Holography”, which is supported in part by the National Science Foundation under Grant No. PHY-1748958. A Deﬁnition of LOCC In this appendix we give a precise deﬁnition of LOCC. For a useful reference, see ref. [50]. (In general LOCC is deﬁned for multiple parties, but here we will only discuss the two party case.) For two quantum systems A and W, with the Hilbert space HA ⊗HW , a one-way LOCC from A to W is a quantum channel of the following form: CA→W = M X n=1 MA n ⊗N W n (A.1) (A.1) Here MA n are completely positive (CP) maps with the additional condition that P n MS n is a completely positive and trace-preserving (CPTP) map. In other words, for any density operator ρ of subsystem A, P n MS n(ρ) is a density operator with unit trace. N W n for each n is a CPTP map. Physically, if we apply CA→W to a product state ρA ⊗σW , we obtain CA→W (ρA ⊗σW ) = M X n=1 MA n (ρA) ⊗N W n (σW ) (A.2) (A.2) – 18 – which is a separable state. Deﬁne which is a separable state. Deﬁne which is a separable state. Deﬁne pn = tr MA n (ρA) (A.3) ˜ρAn = p−1 n MA n (ρA) (A.4) (A.3) (A.4) (A.3) (A.4) the channel CA→W applies a weak measurement to A and if the measurement output is n, apply channel N W n to W. the channel CA→W applies a weak measurement to A and if the measurement output is n, apply channel N W n to W. A r-round LOCC between A and W contains r-round of back-and-forth communication between these two systems: JHEP01(2022)085 JHEP01(2022)085 C(r) AW = Cr W→A ◦Cr A→W ◦Cr−1 W→A ◦Cr−1 A→W ◦. . . ◦C1 W→A ◦C1 A→W (A.5) (A.5) B Positive conditional entropy for separable states For completeness, we include a proof that the conditional entropy is always positive for separable states. For a separable state ρWI = X i piρWi ⊗ρIi (B.1) (B.1) we consider an auxiliary state we consider an auxiliary state π ˜ WI = X i pi \|i⟩⟨i\| ⊗ρIi (B.2) (B.2) with \|i⟩an orthonormal basis in the ˜W system. The mutual information between ˜W and I is with \|i⟩an orthonormal basis in the ˜W system. The mutual information between ˜W and I is I( ˜W : I) = S X i piρIi ! − X i piS (ρIi) (B.3) (B.3) Which is also the Holevo information that measures the amount of information that can be read out from I about the classical message in ˜W. Therefore Which is also the Holevo information that measures the amount of information that can be read out from I about the classical message in ˜W. Therefore I ˜W : I π ≤SI ≡S X i piρIi ! (B.4) (B.4) Now we can apply a measure-and-prepare channel to map a state in ˜W to that in W: Now we can apply a measure-and-prepare channel to map a state in ˜W to that in W: C σ ˜ W = X i ⟨i\| σ ˜ W \|i⟩ρWi (B.5) (B.5) The channel maps the auxiliary state π ˜ WI to ρWI. Since mutual information cannot be increased by a quantum channel, we obtain that in ρWI The channel maps the auxiliary state π ˜ WI to ρWI. Since mutual information cannot be increased by a quantum channel, we obtain that in ρWI I(W : I)ρ ≤I ˜W : I π ≤SI (B.6) (B.6) Therefore S(W) ≤S(WI) (B.7) S(W) ≤S(WI) (B.7) – 19 – – 19 – Open Access. This article is distributed under the terms of the Creative Commons Attribution License (CC-BY 4.0), which permits any use, distribution and reproduction in any medium, provided the original author(s) and source are credited. References [1] S.W. Hawking, Particle creation by black holes, Commun. Math. Phys. 43 (1975) 199 [Erratum ibid. 46 (1976) 206] [INSPIRE]. [2] D.N. Page, Average entropy of a subsystem, Phys. Rev. Lett. 71 (1993) 1291 [gr-qc/9305007] [INSPIRE]. JHEP01(2022)085 JHEP01(2022)085 [3] G. Penington, Entanglement wedge reconstruction and the information paradox, JHEP 09 (2020) 002 [arXiv:1905.08255] [INSPIRE]. [3] G. Penington, Entanglement wedge reconstruction and the information paradox, JHEP 09 (2020) 002 [arXiv:1905.08255] [INSPIRE]. [4] G. Fejős and N. Yamamoto, Functional renormalization group approach to color superconducting phase transition, JHEP 12 (2019) 069 [arXiv:1908.03535] [INSPIRE]. [4] G. Fejős and N. Yamamoto, Functional renormalization group approach to color superconducting phase transition, JHEP 12 (2019) 069 [arXiv:1908.03535] [INSPIRE]. [5] S. Ryu and T. Takayanagi, Holographic derivation of entanglement entropy from AdS/CFT, Phys. Rev. Lett. 96 (2006) 181602 [hep-th/0603001] [INSPIRE]. [6] V.E. Hubeny, M. Rangamani and T. Takayanagi, A covariant holographic entanglement entropy proposal, JHEP 07 (2007) 062 [arXiv:0705.0016] [INSPIRE]. 6] V.E. Hubeny, M. Rangamani and T. Takayanagi, A covariant holographic entanglement entropy proposal, JHEP 07 (2007) 062 [arXiv:0705.0016] [INSPIRE]. [7] T. Faulkner, A. Lewkowycz and J. Maldacena, Quantum corrections to holographic entanglement entropy, JHEP 11 (2013) 074 [arXiv:1307.2892] [INSPIRE]. [8] G. Penington, S.H. Shenker, D. Stanford and Z. Yang, Replica wormholes and the black hole interior, arXiv:1911.11977 [INSPIRE]. [9] A. Almheiri, T. Hartman, J. Maldacena, E. Shaghoulian and A. Tajdini, Replica wormholes and the entropy of Hawking radiation, JHEP 05 (2020) 013 [arXiv:1911.12333] [INSPIRE]. [10] Y. Chen, Pulling out the island with modular ﬂow, JHEP 03 (2020) 033 [arXiv:1912.02210] [INSPIRE]. [10] Y. Chen, Pulling out the island with modular ﬂow, JHEP 03 (2020) 033 [arXiv:1912.02210] [INSPIRE]. [11] D.L. Jaﬀeris, A. Lewkowycz, J. Maldacena and S.J. Suh, Relative entropy equals bulk relative entropy, JHEP 06 (2016) 004 [arXiv:1512.06431] [INSPIRE]. [12] A. Almheiri, D. Marolf, J. Polchinski and J. Sully, Black holes: complementarity or ﬁrewalls?, JHEP 02 (2013) 062 [arXiv:1207.3123] [INSPIRE]. [12] A. Almheiri, D. Marolf, J. Polchinski and J. Sully, Black holes: complementarity or ﬁrewalls?, JHEP 02 (2013) 062 [arXiv:1207.3123] [INSPIRE]. [13] A. Almheiri, D. Marolf, J. Polchinski, D. Stanford and J. Sully, An apologia for ﬁrewalls, JHEP 09 (2013) 018 [arXiv:1304.6483] [INSPIRE]. [13] A. Almheiri, D. Marolf, J. Polchinski, D. Stanford and J. Sully, An apologia for ﬁrewalls, JHEP 09 (2013) 018 [arXiv:1304.6483] [INSPIRE]. [14] X. Dong, X.-L. Qi, Z. Shangnan and Z. References Yang, Eﬀective entropy of quantum ﬁelds coupled with gravity, JHEP 10 (2020) 052 [arXiv:2007.02987] [INSPIRE]. [14] X. Dong, X.-L. Qi, Z. Shangnan and Z. Yang, Eﬀective entropy of quantum ﬁelds coupled with gravity, JHEP 10 (2020) 052 [arXiv:2007.02987] [INSPIRE]. [15] A. Lewkowycz and J. Maldacena, Generalized gravitational entropy, JHEP 08 (2013) 090 [arXiv:1304.4926] [INSPIRE]. [16] X. Dong, A. Lewkowycz and M. Rangamani, Deriving covariant holographic entanglement, JHEP 11 (2016) 028 [arXiv:1607.07506] [INSPIRE]. [16] X. Dong, A. Lewkowycz and M. Rangamani, Deriving covariant holographic entanglement, JHEP 11 (2016) 028 [arXiv:1607.07506] [INSPIRE]. [17] X. Dong and A. Lewkowycz, Entropy, extremality, euclidean variations, and the equations of motion, JHEP 01 (2018) 081 [arXiv:1705.08453] [INSPIRE]. [17] X. Dong and A. Lewkowycz, Entropy, extremality, euclidean variations, and the equations of motion, JHEP 01 (2018) 081 [arXiv:1705.08453] [INSPIRE]. [18] D. Aharonov, J. Cotler and X.-L. Qi, Quantum algorithmic measurement, arXiv:2101.04634. [18] D. Aharonov, J. Cotler and X.-L. Qi, Quantum algorithmic measurement, arXiv:2101.04634. – 20 – [19] P. Saad, S.H. Shenker and D. Stanford, JT gravity as a matrix integral, arXiv:1903.11115 [INSPIRE]. [20] D. Stanford and E. Witten, JT gravity and the ensembles of random matrix theory, Adv. Theor. Math. Phys. 24 (2020) 1475 [arXiv:1907.03363] [INSPIRE]. [21] P. Saad, Late time correlation functions, baby universes, and ETH in JT gravity, arXiv:1910.10311 [INSPIRE]. [22] R. Bousso and E. Wildenhain, Gravity/ensemble duality, Phys. Rev. D 102 (2020) 066005 [arXiv:2006.16289] [INSPIRE]. [23] D. Marolf and H. Maxﬁeld, Transcending the ensemble: baby universes, spacetime wormholes, and the order and disorder of black hole information, JHEP 08 (2020) 044 [arXiv:2002.08950] [INSPIRE]. JHEP01(2022)085 [24] D. Stanford, More quantum noise from wormholes, arXiv:2008.08570 [INSPIRE]. [25] J. Pollack, M. Rozali, J. Sully and D. Wakeham, Eigenstate thermalization and disorder averaging in gravity, Phys. Rev. Lett. 125 (2020) 021601 [arXiv:2002.02971] [INSPIRE]. [26] N. Afkhami-Jeddi, H. Cohn, T. Hartman and A. Tajdini, Free partition functions and an averaged holographic duality, JHEP 01 (2021) 130 [arXiv:2006.04839] [INSPIRE]. [27] A. Maloney and E. Witten, Averaging over Narain moduli space, JHEP 10 (2020) 187 [arXiv:2006.04855] [INSPIRE]. [28] A. Belin and J. de Boer, Random statistics of OPE coeﬃcients and Euclidean wormholes, Class. Quant. Grav. 38 (2021) 164001 [arXiv:2006.05499] [INSPIRE]. [29] J. Cotler and K. Jensen, AdS3 gravity and random CFT, JHEP 04 (2021) 033 [arXiv:2006.08648] [INSPIRE]. [30] Y. Chen and H.W. Lin, Signatures of global symmetry violation in relative entropies and replica wormholes, JHEP 03 (2021) 040 [arXiv:2011.06005] [INSPIRE]. References [31] P.-S. Hsin, L.V. Iliesiu and Z. Yang, A violation of global symmetries from replica wormholes and the fate of black hole remnants, Class. Quant. Grav. 38 (2021) 194004 [arXiv:2011.09444] [INSPIRE]. [32] A. Belin, J. De Boer, P. Nayak and J. Sonner, Charged eigenstate thermalization, euclidean wormholes and global symmetries in quantum gravity, arXiv:2012.07875 [INSPIRE]. [33] A. Milekhin and A. Tajdini, Charge ﬂuctuation entropy of Hawking radiation: a replica-fre way to ﬁnd large entropy, arXiv:2109.03841 [INSPIRE]. [34] X.-L. Qi, Z. Shangnan and Z. Yang, Holevo information and ensemble theory of gravity, arXiv:2111.05355 [INSPIRE]. [35] S. Sachdev and J. Ye, Gapless spin ﬂuid ground state in a random, quantum Heisenberg magnet, Phys. Rev. Lett. 70 (1993) 3339 [cond-mat/9212030] [INSPIRE]. [36] A. Kitaev, Hidden correlations in the Hawking radiation and thermal noise, talk given at Fundamental physics prize symposium, (2014). [37] A. Kitaev, A simple model of quantum holography (part 1), talk given at the KITP program: Entanglement in strongly-correlated quantum matter, https://online.kitp.ucsb.edu/online/entangled15/kitaev/, University of California, Santa Barbara, CA, U.S.A., 7 April 2015. – 21 – [38] A. Kitaev, A simple model of quantum holography (part 2), talk given at the KITP program: Entanglement in strongly-correlated quantum matter, https://online.kitp.ucsb.edu/online/entangled15/kitaev2/, University of California, Santa Barbara, CA, U.S.A., 27 May 2015. [39] P. Hayden, S. Nezami, X.-L. Qi, N. Thomas, M. Walter and Z. Yang, Holographic duality from random tensor networks, JHEP 11 (2016) 009 [arXiv:1601.01694] [INSPIRE]. [40] A. Almheiri, Comments on the ﬁnal state proposal and the gravitational path integral, lecture at the YITP workshop Recent progress in theoretical physics based on quantum information theory, http://www2.yukawa.kyoto-u.ac.jp/∼qith2021/Almheiri.pdf, (2021). [41] G.T. Horowitz and J.M. Maldacena, The black hole ﬁnal state, JHEP 02 (2004) 008 [hep-th/0310281] [INSPIRE]. JHEP01(2022)085 JHEP01(2022)085 [42] J.M. Maldacena and L. Maoz, Wormholes in AdS, JHEP 02 (2004) 053 [hep-th/0401024] [INSPIRE]. [43] H. Geng et al., Inconsistency of islands in theories with long-range gravity, arXiv:2107.03390 [INSPIRE]. [44] H. Geng et al., Information transfer with a gravitating bath, SciPost Phys. 10 (2021) 103 [arXiv:2012.04671] [INSPIRE]. [45] C. Chowdhury, V. Godet, O. Papadoulaki and S. Raju, Holography from the Wheeler-DeWitt equation, arXiv:2107.14802 [INSPIRE]. [46] C. Chowdhury, O. Papadoulaki and S. Raju, A physical protocol for observers near the boundary to obtain bulk information in quantum gravity, SciPost Phys. 10 (2021) 106 [arXiv:2008.01740] [INSPIRE]. [47] R. Bousso and D. Stanford, Measurements without probabilities in the ﬁnal state proposal, Phys. Rev. D 89 (2014) 044038 [arXiv:1310.7457] [INSPIRE]. References [48] K. Papadodimas and S. Raju, An infalling observer in AdS/CFT, JHEP 10 (2013) 212 [arXiv:1211.6767] [INSPIRE]. [49] P. Hayden and G. Penington, Learning the alpha-bits of black holes, JHEP 12 (2019) 007 [arXiv:1807.06041] [INSPIRE]. [50] E. Chitambar, D. Leung, L. Mančinska, M. Ozols and A. Winter, Everything you always wanted to know about LOCC (but were afraid to ask), Commun. Math. Phys. 328 (2014) 303 [arXiv:1210.4583]. – 22 –
https://openalex.org/W4290988021	https://www.qeios.com/read/YE3NZW/pdf	English	null	Review of: "Clausius’ thermodynamics, engineering thermodynamics based on the entropy principle by discarding the energy premise"	null	2,022	cc-by	1,551	Review of: "Clausius’ thermodynamics, engineering thermodynamics based on the entropy principle by discarding the energy premise" Joseph O'Neill1 1 University of California, Los Angeles Potential competing interests: The author(s) declared that no potential competing interests exist. This paper reexamines some original works of Kelvin, Gibbs and (in English translation) Planck and Clausius out of which Classical Thermodynamics emerged. The familiar thermodynamic variables energy, entropy, and heat are reported to have evolved out of the earlier caloric concept. It is observed that Clausius-Kelvin thermodynamics was no single unified theory (e.g., the same doctrine derived twice independently). Rather, it was a “blend” of two separate theories. Of these two, the Clausius Theory fostered Gibbsian thermodynamics, while the Kelvin Theory underlies orthodox engineering thermodynamics to this day. The latter is said to rely on the “energy premise”. The energy premise means that an irreversible thermodynamic process is one that “dissipates energy”. The text points out, however, that an irreversible process more generally is one that “produces entropy”, since certain irreversible processes occur without dissipation of energy. Because it retains the energy premise and thereby is dominated by a “monolithic view of energy”, this paper asserts that “engineering thermodynamics is a defective theoretical system while Gibbsian thermodynamics is a successful one”. This aspect of engineering thermodynamics is, moreover, found to be pernicious and in need of reform. In particular, engineering thermodynamics places a premium on producing work. The Second Law, in the Kelvin view, implies that dissipating energy is inevitable when producing work. Dissipation of energy frequently implies discharging waste heat into the natural environment. The latter is often regarded by engineers, at times callously and cavalierly, as an infinite heat reservoir and heat sink. These discharges, occurring worldwide daily on massive scales as they do, disrupt the local and global ecological balance. Informed by its reading of the classic texts, the paper advocates an ideological approach to solving this socio-economic problem. The paper suggests that engineering thermodynamics replace the orthodox concept of energy-conversion with the concept of “transformations” in the sense of Clausius. Motivated by Poincaré, the author has previously introduced a thermodynamic quantity called the “entropy growth potential” (EGP). The EGP appears to mean the spontaneous change in the entropy of the universe (system+surroundings) that compensates for the artificial work production that occurs during an event. I.e., through intentional effort, humans extract (or input) a quantity W of work from a thermodynamic system; the sytem+environment react to this spontaneously by increasing their combined entropy by the amount EGP. Qeios, CC-BY 4.0 · Review, August 11, 2022 Qeios ID: YE3NZW · https://doi.org/10.32388/YE3NZW Review of: "Clausius’ thermodynamics, engineering thermodynamics based on the entropy principle by discarding the energy premise" Were engineers to start talking about Clausisan transformations and EGP rather than energy-conversions and efficiencies in a widespread manner, this paper proposes that this mindset would help curb the further degradation of the natural environment (through, e.g., burning of fossil fuel) and help to preserve it (through, e.g., reliance on renewable energy sources). That summarizes the major thesis. Qeios ID: YE3NZW · https://doi.org/10.32388/YE3NZW 1/3 Qeios, CC-BY 4.0 · Review, August 11, 2022 Let us begin with secondary, but important practical remarks. The manuscript be proofread by a native speaker of English. The author has a high familiarity with the classic source texts and inventive and well thought-out ideas. But the presentation is at times obscured by the syntactic and lexical nuances of the English language. Editing by a competent native-speaker would render it more accessible and unambiguously interpretable. Also, the typesetting of some of the equations is aesthetically off-putting. In particular, several terms are printed a line too high. Finally, some abbreviations, e.g., “NWCJ”, are left undefined in the text. All but the most common abbreviations should be defined upon first use. Now to primary matter. A tenet of this paper can be questioned. That is the proposition that orthodox engineering thermodynamics is a defective theoretical system, enthralled to the energy premise. Mainstream engineering training in thermodynamics is not defective in this way. On the contrary, it has long included scenarios of irreversible processes without thermal discharge. Go back, for example, to Adamson (1979), a standard text for teaching thermodynamics to engineers in its day. On pp. 318-19, this textbook derives the result that, while the entropy of mixing of an ideal solution is positive, its enthalpy (heat) of mixing is 0. The same result is arrived at on pp.453-54 of Van Wylen & Sonntag (1973), another textbook of that era. This is actually the very example cited in the worthy quotation from Planck in the manuscript. Further, on p. 438, Van Wylen & Sonntag remark profoundly: “…the increase in entropy depends only on the number of moles of component gases, and is independent of the composition of the gas. For example, when 1 mole of oxygen and 1 mole of nitrogen are mixed, the increase of entropy is the same as when 1 mole of hydrogen and 1 mole of nitrogen are mixed. Qeios ID: YE3NZW · https://doi.org/10.32388/YE3NZW Review of: "Clausius’ thermodynamics, engineering thermodynamics based on the entropy principle by discarding the energy premise" But we also know that if 1 mole of nitrogen is ‘mixed’ with another mole of nitrogen there is no increase in entropy. The question that arises is how dissimilar must the gases be in order to have an increase in entropy? The answer lies in our ability to distinguish between the two gases. The entropy increases whenever we can distinguish between the gases being mixed. When we cannot distinguish between the gases, there is no increase in entropy.” Thus, decades ago orthodox engineering training already not only integrated scenarios of entropy production without energy exchange was integrated into, but also understood their special significance to the entropy concept. Gibbs’s contributions, mentioned in the manuscript, have also long been part of regular course materials. The same is true for the Boltzmann entropy S=klnW, a highly general formulation unrestricted to thermal effects per se. Regarding another point in the manuscript, conventional engineering instruction does not depict the Second Law as an indirect consequence of the First Law. Rather, the Second Law is presented as an equally ranked, fully independent axiom. Overall, in principle, there is no reliquarian energy premise from Kelvin still misguiding engineers two centuries later. In practice, it is true that most engineers most of the time think of irreversible processes in terms of energy dissipation, as a shorthand. But that is because the most common and the most important applications involve energy dissipation. The 2/3 Qeios, CC-BY 4.0 · Review, August 11, 2022 perspective can easily be shifted should the situation demand it. But there is no need to change the formalism as the formalism already covers all possible scenarios. In particular, it is not essential to promulgate the EGP as a novel thermodynamic variable. Like the exergy, introduced as such in the 1950s, the EGP might lighten certain calculations and emphasize certain perspectives, but everything it does is already covered by other elements of standard theory. perspective can easily be shifted should the situation demand it. But there is no need to change the formalism as the formalism already covers all possible scenarios. In particular, it is not essential to promulgate the EGP as a novel thermodynamic variable. Like the exergy, introduced as such in the 1950s, the EGP might lighten certain calculations and emphasize certain perspectives, but everything it does is already covered by other elements of standard theory. Qeios ID: YE3NZW · https://doi.org/10.32388/YE3NZW Review of: "Clausius’ thermodynamics, engineering thermodynamics based on the entropy principle by discarding the energy premise" It is true that engineering training inculcates certain habits. These include assuming infinite heat sinks or reservoirs (or concentration sinks or reservoirs, etc.) This occurs both while working academic exercises and in handling professional problems in the field. It is plausible that the routine application of these habits has abetted innumerable untoward discharges of heat and toxins into the environment. But there are remedies to this short of modifying the basic theory. For example, engineering instruction could prescribe some exercises that take the environment into account explicitly. E.g., after performing a standard calculation of the waste heat generated by a thermodynamic process, a section could be added. Maybe, “You are discharging the heat at the rate dQ/dt into Lake X, which has volume V. Species Y of fish lives in this lake. People in the villages on the shore eat the fish. The species will not lay eggs if the water temperature rises above temperature T. Can you schedule the heat discharges so that the fish never stop breeding?” Or similar. The same engineering skills and habits can be redirected to advance alternative goals without changing the formalism. In the years since the 1970s, engineering education has only increased in theoretical depth. For example, material properties once looked-up in handbooks of empirical results are now routinely calculated from first principles. Along with theoretical sophistication, engineers on average have also risen in environmental sensitivity, owing in part to regulation and litigation. Secular changes have also occurred in the culture. Those, at least, are optimistic factors. Adamson AW. A Textbook of Physical Chemistry, 2nd ed. Academic Press, New York (1979). Van Wylen GJ, Sonntag RE. Fundamentals of Classical thermodynamics. Wiley, New York (1973). Qeios ID: YE3NZW · https://doi.org/10.32388/YE3NZW 3/3
https://openalex.org/W4248942648	https://journals.plos.org/plosmedicine/article/file?id=10.1371/journal.pmed.0050246&type=printable	English	null	Correction: A Prevalent Variant in PPP1R3A Impairs Glycogen Synthesis and Reduces Muscle Glycogen Content in Humans and Mice	PLoS medicine	2,008	cc-by	953	The author contributions section of this manuscript should be corrected as follows: The author contributions section of this manuscript should be corrected as follows: Mouse studies were designed and performed by DBS, LZ, CSC, ACG, S-EW, GM-G, SK, CBB, DMS, AV-P, GIS, BAP, PJR, and AADP-R. Human studies were designed and performed by DBS, BR, JES, BS, DD, NJW, FK, RT, and SOR. The manuscript was prepared by DBS, SOR, and AADP-R, and seen by all authors. Mouse studies were designed and performed by DBS, LZ, CSC, ACG, S-EW, GM-G, SK, CBB, DMS, AV-P, GIS, BAP, PJR, and AADP-R. Human studies were designed and performed by DBS, BR, JES, BS, DD, NJW, FK, RT, and SOR. The manuscript was prepared by DBS, SOR, and AADP-R, and seen by all authors. Citation: Savage DB, Zhai L, Ravikumar B, Choi CS, Snaar JE, et al. (2008) Correction: A Prevalent Variant in PPP1R3A Impairs Glycogen Synthesis and Reduces Muscle Glycogen Content in Humans and Mice. PLoS Med 5(12): e246. doi:10.1371/journal.pmed.0050246 Citation: Savage DB, Zhai L, Ravikumar B, Choi CS, Snaar JE, et al. (2008) Correction: A Prevalent Variant in PPP1R3A Impairs Glycogen Synthesis and Reduces Muscle Glycogen Content in Humans and Mice. PLoS Med 5(12): e246. doi:10.1371/journal.pmed.0050246 Received: October 9, 2008; Accepted: October 21, 2008; Published: December 23, 2008. Received: October 9, 2008; Accepted: October 21, 2008; Published: December 23, 2008. Copyright: © 2008 Savage et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Correction: Vitamin K Supplementation in Postmenopausal Women with Osteopenia (ECKO Trial): A Randomized Controlled Trial (ECKO Trial): A Randomized Controlled Trial Angela M. Cheung, Lianne Tile, Yuna Lee, George Tomlinson, Gillian Hawker, Judy Scher, Hanxian Hu, Reinhold Vieth, Lilian Thompson, Sophie Jamal, Robert Josse Correction for: Cheung AM, Tile L, Lee Y, Tomlinson G, Hawker G, et al. (2008) Vitamin K supplementation in postmenopausal women with osteopenia (ECKO Trial): A randomized controlled trial. PLoS Med 5(10): e196. doi:10.1371/journal.pmed.0050196 Three errors were made in the Funding and Acknowledgments sections of this article. In the Funding section, the “Wyeth Foundation” should be replaced by the “Centrum Foundation” and “Whitehall Robbins” should be spelled “Whitehall Robins.” In the Acknowledgments section, “Dr. Faten Al-Doori” should be spelled “Dr. Faten Al-Douri.” Citation: Cheung AM, Tile L, Lee Y, Tomlinson G, Hawker G, et al. (2008) Correction: Vitamin K Supplementation in Postmenopausal Women with Osteopenia (ECKO Trial): A Randomized Controlled Trial. PLoS Med 5(12): e247. doi:10.1371/journal.pmed.0050247 Received: October 17, 2008; Accepted: October 23, 2008; Published: December 23, 2008. Copyright: © 2008 Cheung et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Angela M. Cheung, Lianne Tile, Yuna Lee, George Tomlinson, Gillian Hawker, Judy Scher, Hanxian Hu, Reinhold Vieth, Lilian Thompson, Sophie Jamal, Robert Josse Angela M. Cheung, Lianne Tile, Yuna Lee, George Tomlinson, Gillian Hawker, Judy Scher, Hanxian Hu, Reinh hompson, Sophie Jamal, Robert Josse Correction for: Cheung AM, Tile L, Lee Y, Tomlinson G, Hawker G, et al. (2008) Vitamin K supplementation in postmenopausal women with osteopenia (ECKO Trial): A randomized controlled trial. PLoS Med 5(10): e196. doi:10.1371/journal.pmed.0050196 Three errors were made in the Funding and Acknowledgments sections of this article. In the Funding section, the “Wyeth Foundation” should be replaced by the “Centrum Foundation” and “Whitehall Robbins” should be spelled “Whitehall Robins.” In the Acknowledgments section, “Dr. Faten Al-Doori” should be spelled “Dr. Faten Al-Douri.” Citation: Cheung AM, Tile L, Lee Y, Tomlinson G, Hawker G, et al. (2008) Correction: Vitamin K Supplementation in Postmenopausal Women with Osteopenia (ECKO Trial): A Randomized Controlled Trial. PLoS Med 5(12): e247. doi:10.1371/journal.pmed.0050247 Received: October 17, 2008; Accepted: October 23, 2008; Published: December 23, 2008. Received: October 17, 2008; Accepted: October 23, 2008; Published: December 23, 2008. Copyright: © 2008 Cheung et al. Correction: A Prevalent Variant in PPP1R3A Impairs Glycogen Synthesis and Reduces Muscle Glycogen Content in Humans and Mice David B. Savage, Lanmin Zhai, Balasubramanian Ravikumar, Cheol Soo Choi, Johanna E. Snaar, Amanda C. McGuire, Sung-Eun Wou, Gemma Medina-Gomez, Sheene Kim, Cheryl B. Bock, Dyann M. Segvich, Bhavana Solanky, Dinesh Deelchand, Antonio Vidal-Puig, Nicholas J. Wareham, Gerald I. Shulman, Fredrik Karpe, Roy Taylor, Bartholomew A. Pederson, Peter J. Roach, Stephen O’Rahilly, Anna A. DePaoli-Roach Correction for: Correction for: Savage DB, Zhai L, Ravikumar B, Choi CS, Snaar JE, et al. (2008) A prevalent variant in PPP1R3A impairs glycogen synthesis and reduces muscle glycogen content in humans and mice. PLoS Med 5(1): e27. doi:10.1371/journal.pmed.0050027 One of the corresponding authors, Stephen O’Rahilly, has contacted us to request that two Ph.D. students at the time of the research—Dr. Bhavana Solanky and Dr. Dinesh Deelchand—be listed as authors on this paper for their contribution on the work on the human participants undertaken at the Sir Peter Mansfield Magnetic Resonance Centre in Nottingham, United Kingdom. All of the other authors listed on this paper agree to this change in the authorship. We have contacted both Dr. Bhavana Solanky and Dr. Dinesh Deelchand, who have confirmed that it is appropriate for them to be listed as authors on the paper and that they have no competing interests. Correction: Vitamin K Supplementation in Postmenopausal Women with Osteopenia (ECKO Trial): A Randomized Controlled Trial This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. December 2008 \| Volume 5 \| Issue 12 \| e246 \| e247 PLoS Medicine \| www.plosmedicine.org 1776
https://openalex.org/W4393375034	https://pubs.aip.org/aip/adv/article-pdf/doi/10.1063/9.0000706/19860763/045001_1_9.0000706.pdf	English	null	Broadband spin-filtered minimalistic magnetic tunnel junction	AIP advances	2,024	cc-by	3,812	RESEARCH ARTICLE \| APRIL 01 2024 RESEARCH ARTICLE \| APRIL 01 2024 Broadband spin-filtered minimalistic magnetic tunnel junction AIP Advances 14, 045001 (2024) https://doi.org/10.1063/9.0000706 AIP Advances 14, 045001 (2024) https://doi.org/10.1063/9.0000706 Articles You May Be Interested In Reduction in critical current of current induced switching in an inhomogeneous nanomagnet Appl. Phys. Lett. (March 2009) Characterization of the nontrivial and chaotic behavior that occurs in a simple city traffic model Chaos (March 2010) Spin transfer by nonuniform current injection into a nanomagnet 24 October 2024 05:50:34 AFFILIATIONS Department of Electrical Engineering, Indian Institute of Technology Ropar, Rupnagar, Punjab 140001, India Note: This paper was presented at the 68th Annual Conference on Magnetism and Magnetic Materials. a)Electronic mail: sabarna.chakraborti@gmail.com b)Author to whom correspondence should be addressed abhishek@iitrpr ac in Note: This paper was presented at the 68th Annual Conference on Magnetism and Magnetic Materials. a)Electronic mail: sabarna.chakraborti@gmail.com @g b)Author to whom correspondence should be addressed: abhishek@iitrpr.ac.in @g b)Author to whom correspondence should be addressed: a ABSTRACT The tri-layer magnetic tunnel junction (MTJ) has surfaced as a building block for engineering next-generation integrated circuits while combining the attributes of non-volatility and meager energy consumption. Nevertheless, the perceptible switching energy (≈20–50 fJ/bit) and sub-optimal tunnelmagnetoresistance (TMR) (≈200%–300%) have acted as major hindrances, concealing its potential to supersede the capabilities of static and dynamic random access memories. In this work, we introduce a novel device that features a minimalistic non-uniform heterostructure/superlattice instead of the oxide layer in a conventional MTJ and analyze it in the premise of the self-consistent coupling of the Non-Equilibrium-Green’s Function (NEGF) and the Landau-Liftshitz-Gilbert-Slonczewski (LLGS) equation. We ascertain that the coupling of the electrodes to the proposed heterostructure renders a highly spin-selective broadband transmittance, thereby enabling a towering TMR (%) of 3.7 × 104% along with a significant reduction in the spin transfer torque (STT) switching energy (≈1.96 fJ). Furthermore, the sizable slonczewski term (Is∥) originating from the heterostructure facilitates a swift STT-switching within the scale of a few hundred picoseconds (≈400 ps). 24 October 2024 05:50:34 © 2024 Author(s). All article content, except where otherwise noted, is licensed under a Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). https://doi.org/10.1063/9.0000706 © 2024 Author(s). All article content, except where otherwise noted, is licensed under a Creative Commo (https://creativecommons.org/licenses/by/4.0/). https://doi.org/10.1063/9.0000706 Broadband spin-filtered minimalistic magnetic tunnel junction Cite as: AIP Advances 14, 045001 (2024); doi: 10.1063/9.0000 Submitted: 2 October 2023 • Accepted: 15 March 2024 • Published Online: 1 April 2024 Cite as: AIP Advances 14, 045001 (2024); doi: 10.1063/9.0000706 Submitted: 2 October 2023 • Accepted: 15 March 2024 • Published Online: 1 April 2024 ARTICLE AIP Advances pubs.aip.org/aip/adv I. INTRODUCTION torque (STT) or spin–orbit torque (SOT), facilitating writing oper- ations via exclusively electrical currents.6 Perpendicular magnetic anisotropy (PMA) enables the magnetization of the pinned and the free layer to stand vertically against the plane of the film. A stan- dard PMA-MTJ (p-MTJ) achieves swift switching within the time frame of nanoseconds, coupled with reduced energy consumption and enhanced thermal stability in comparison to in-plane MTJs.7 Given that contemporary volatile memories such as SRAM and DRAM have the capability to achieve switching down to the fJ limit,8–10 enhancement in the TMR (≈200%–300%) and reduction of the switching energy (≈20–50 fJ/bit) of a conventional MTJ, con- strained by the principles of single barrier tunneling, are at the nexus of contemporary research.1,11–13 Spintronic devices have recently emerged as the pinnacle of storage technology, providing a harmonious blend of speed and energy efficacy. Owing to the union of CMOS integrability, high packing density, and non-volatility, magnetic tunnel junc- tions (MTJs) have lit up the horizon of spintronics with numerous possibilities ranging from magnetoresistive random access mem- ories (MRAMs),1 biomedical studies,2 spin transfer torque nano oscillators (STNOs)3 and wireless technology4 to a wide range of sensors.5 A canonical MTJ is realized by a nanopillar that constitutes a free and a pinned ferromagnet (FM) parted by an insulating material (e.g., MgO) as depicted in Fig. 1(a). The spin-dependent quantum transport through the device is gauged by the tunnel magnetoresistance (TMR), given by TMR (%) = (RPC −RAPC)/RPC × 100% where RAPC and RPC represents the resistances in anti- parallel (APC) and parallel (PC) configurations, respectively. The dynamics of the free FM can be modulated by either spin transfer In this work, we introduce a non-uniform heterostructure/superlattice-based MTJ (NHMTJ) with four barriers, conceived to markedly amplify the TMR (%) while concur- rently resulting in a significant reduction in power consumption. To systematically compare the switching bias of the NHMTJ against the standard MTJ, we instantiate a performance metric, designated as AIP Advances 14, 045001 (2024); doi: 10.1063/9.0000706 AIP Advances 14, 045001 (2024); doi: 10.1063/9.0000706 © Author(s) 2024 14, 045001-1 14, 045001-1 ARTICLE pubs.aip.org/aip/adv FIG. 2. Schematic of band diagram in a (a) conventional MTJ and the (b) NHMTJ. AIP Advances pubs.aip.org/aip/adv FIG. 1. Schematic of (a) conventional MTJ and (b) a non-uniform heterostructure/superlattice-based MTJ (NHMTJ). FIG. 2. Schematic of band diagram in a (a) conventional MTJ and the (b) NHMTJ. FIG. 1. I. INTRODUCTION Schematic of (a) conventional MTJ and (b) a non-uniform heterostructure/superlattice-based MTJ (NHMTJ). applied to the conventional MTJ (NHMTJ) for the transformations from PC to APC and vice versa, with SB denoting the switching bias. the suppression in the switching bias (SSB), defined by the following equation: AIP Advances 14, 045001 (2024); doi: 10.1063/9.0000706 II. DETAILS OF MODELLING AND SIMULATION SSB = ∣SBMTJ PC→APC∣+ ∣SBMTJ APC→PC∣ ∣SBNHMTJ PC→APC∣+ ∣SBNHMTJ APC→PC∣ (1) (1) To probe the intricacies of quantum transport within MTJs, we leverage the non-equilibrium Green’s function (NEGF) formalism to compute the spin current (IS) through the device while using the equation formulated below: 24 October 2024 05:50:34 where APC(PC) symbolizes the anti-parallel(parallel) configura- tion, SBMTJ(NHMTJ) PC→APC and SBMTJ(NHMTJ) APC→PC represent the switching biases FIG. 3. (a) I–V characteristics and (b) TMR of the MTJ, (c) I–V characteristics and (d) TMR of the NHMTJ in the PC(IP) and APC(IAP). FIG. 3. (a) I–V characteristics and (b) TMR of the MTJ, (c) I–V characteristics and (d) TMR of the NHMTJ in the PC(IP) and APC(IAP). AIP Advances 14, 045001 (2024); doi: 10.1063/9.0000706 14, 045001-2 14, 045001-2 14, 045001-2 © Author(s) 2024 AIP Advances FIG. 4. Transmittance profile: (a) Upspin and (b) downspin transmittance of the NHMTJ in the PC, (c) upspin and (d) downspin transmittance of the NHMTJ in the APC at 10 mV for various transverse modes. pin and (b) downspin transmittance of the NHMTJ in the PC, (c) upspin and (d) downspin transmittance of the NHMTJ in the APC at 10 FIG. 4. Transmittance profile: (a) Upspin and (b) downspin transmittance of the NHMTJ in the PC, (c) upspin and (d) downspin transmittance of the NHMTJ in the APC at 10 mV for various transverse modes. 24 October 2024 05:50:34 ISi = q ∫Real [ Trace(σi ⋅ˆIop)]dE (2) (2) constituents so that ⃗IS = Is,∥ˆM + Is,m ˆm + Is, ˆM × ˆm with ˆM and ˆm being the unit vectors along the magnetization of the pinned and free FM, respectively while the Is and Is∥denotes the field like and the Slonczewski term. where σi denotes the Pauli spin-matrices along ˆx, ˆy, ˆz directions and ˆIop13,14 is given by The threshold current needed for switching is given by Is∥,t = 2qMsVα ̵h Heff with Heff being HK + Hd/2, where Hd and HK denotes the demagnetizing field and the in-plane anisotropy, respectively.13,16 Given the fact that Is loses its requisite influence when superposed with Is∥, it doesn’t impact the threshold current required for switching13 the free FM. It is pertinent to observe that Hd customarily exhibits a value that is one order of magnitude higher than HK, culminating in a considerable uptick in power con- sumption. II. DETAILS OF MODELLING AND SIMULATION The adverse effects of Hd can be alleviated by the adoption of FMs with PMA,17 which aligns the magnetization perpendicularly to the thin film’s plane,18 with the threshold current being calculated as Is∥,t = 2qMsVα ̵h HK13 with HK being the perpendicular anisotropy. Iop,N−1,N = i ̵h(HN−1,NGn N,N−1 −(HN−1,NGn N,N−1)†) (3) (3) where HN−1,N represents the N −1th row, Nth column element of the device Hamiltonian and Gn N,N−1 represents the N −1th row, Nth column element of the Green’s function matrix.1 The dynamics of the free layer under the influence of spin current and an imposed magnetic field are modeled through a synergistic conjunction of the NEGF formalism with the Landau–Lifshitz–Gilbert–Slonczewski (LLGS) equation described by:15 (1 + α2)∂ˆm ∂t = −γ ˆm × ⃗Heff −γα( ˆm × ( ˆm × ⃗Heff )) − γ̵h 2qMSV (( ˆm × ( ˆm × ⃗IS)) −α( ˆm × ⃗IS)) (4) In this study, we utilized CoFeB13 as ferromagnetic (FM) con- tacts, characterized by a Fermi level E f = 2.25 eV and ferromagnetic exchange splitting of Δ = 2.15 eV.19 We assumed the effective masses in MgO barriers, NM quantum wells (Ru, Cu, Ti etc.13), and FM con- tacts to be mox = 0.18me, mnm = 0.9me, and m fm = 0.8me,20 respec- tively, where me signifies the free-electron mass. The barrier height from CoFeB toward MgO was marked at UB = 0.76 eV higher than the Fermi level,13 with the conduction band offset between the FM and NM layers being UBW = 0.5 eV. (4) In this equation, ˆm denotes the magnetization of the free FM, γ stands for the gyromagnetic ratio associated to electrons with α being the Gilbert damping parameter. Here Ms stands for the satu- ration magnetization, and V is the volume of the CoFeB contacts. The effective magnetic field, ⃗ Heff , is formulated as ⃗ Heff = ⃗ Happlied + ⃗ Hkmzˆz, with ⃗ Happlied representing the applied field, assumed zero in this work, and ⃗ Hk ⃗ mz symbolizing the uniaxial mag- netic anisotropy of the FM contacts with Perpendicular Magnetic Anisotropy (PMA). II. DETAILS OF MODELLING AND SIMULATION Besides, the spin current is bifurcated into three We meticulously kept the thickness of the NM q-wells to W = 0.35 nm in adherence to fabrication constraints.21 Further, we considered a minimum oxide width of 0.6 nm, which relies com- fortably within the reliable deposition limit of MgO.22 We keep AIP Advances 14, 045001 (2024); doi: 10.1063/9.0000706 AIP Advances 14, 045001 (2024); doi: 10.1063/9.0000706 © Author(s) 2024 14, 045001-3 14, 045001-3 ces ARTICLE pubs.aip.org/aip/adv FIG. 5. (a) Is∥of the MTJ and the (b) NHSLTJ with the applied voltage (at θ = 90○). Inset: Is∥−V characteristics at higher voltage. AIP Advances FIG. 5. (a) Is∥of the MTJ and the (b) NHSLTJ with the applied voltage (at θ = 90○). Inset: Is∥−V characteristics at higher voltage. FIG. 5. (a) Is∥of the MTJ and the (b) NHSLTJ with the applied voltage (at θ = 90○). Inset: Is∥−V characteristics at higher voltage. 0.0106 mA. For seamless switching, we adjust the bias to get the Is∥, 20% higher than the Is∥,t.13 the width of the free FM at 1.3 nm23 and establish the diameter of the CoFeB contacts at 30 nm, enabling the depiction of the free FM’s magnetization dynamics via a macro-spin model.24 We accounted for the saturation magnetization at MS = 1150 emu/cc and consider the perpendicular uni-axial magnetic anisotropy at Hk = 3.3 kOe.25 The thermal stability factor is calculated as ΔB = HKMsV 2kBT ≈42. Finally, we took the values for the gyromagnetic ratio and damping parameter as γ = 17.6 MHz/Oe and α = 0.01, respec- tively. Post evaluating all the parameters in the expression of the threshold current given by Is∥,t = 2qMsVα ̵h Hk, we obtain an Is∥,t of the width of the free FM at 1.3 nm23 and establish the diameter of the CoFeB contacts at 30 nm, enabling the depiction of the free FM’s magnetization dynamics via a macro-spin model.24 We accounted for the saturation magnetization at MS = 1150 emu/cc and consider the perpendicular uni-axial magnetic anisotropy at Hk = 3.3 kOe.25 The thermal stability factor is calculated as ΔB = HKMsV 2kBT ≈42. Finally, we took the values for the gyromagnetic ratio and damping parameter as γ = 17.6 MHz/Oe and α = 0.01, respec- tively. Post evaluating all the parameters in the expression of the threshold current given by Is∥,t = 2qMsVα ̵h Hk, we obtain an Is∥,t of IV. CONCLUSION In the next step, we design the NHMTJ [see Fig. 1(b)] with terminal barriers, q-wells, and the central barriers of thickness of 0.6, 0.35 and 0.9 nm respectively [see Fig. 2(b)]. We depict the I–V characteristics, and TMR of the device in Figs. 3(c) and 3(d). Leveraging the highly spin-selective nature of the box-cart-shaped transmittance see Fig. 4, the NHMTJ engenders a pronounced Is∥. Consequently, the free FM goes through the magnetization rever- sal [see Figs. 6(c) and 6(d)] at a nominal bias of 7.3 mV leading to an appreciable SSB of 9.17. Unlike the MTJ, the NHMTJ shows a symmetric Is∥-V characteristics yielding commensurate SBPC→APC and SBAPC→PC. Apart from that, the APC exhibits negligible trans- mittance while obstructing any conceivable flow of electrons (see Fig. 4). This precipitates a towering TMR of 3.7 ×104%. Concur- rently, owing to a maximum value of nearly 0.26 mA [see Fig. 5(b)], the Is∥expedites a swift switching within mere 400 ps, which is at par with the proficiency of typical spin-orbit-torque (SOT) assisted STT switching6 and yet provides a leeway to bypass complexity associated with SOT. Along with the insufficient TMR (%), the non-optimal switch- ing energy13 continues to remain the major hurdle hindering the opportunity for a conventional MTJ to become a premier substitute of sate-of-the-art storage devices such as SRAMs and DRAMs.8,9 In this study, we explore the broad-band spin filtering in a minimalistic NHMTJ that displays a sizable proliferation in the TMR (3.7 ×104%) with a noteworthy reduction in the switching energy (≈1.96 fJ). Since the NHMTJ facilitates a swift STT-based-switching in the time frame of a few hundred picoseconds, it adeptly counteract the deficiencies of unpredictable switching observed in SOT-based MTJs, maintain- ing synchrony with the speed.6,13 Furthermore, it is anticipated that the recent advancements in double metallic quantum well-based MTJs21 will intensify the exploration of heterostructure-based mem- ory devices. We hope that a seminal breakthrough in the periphery of non-volatile memories by the likes of energy-efficient MTJs as exemplified in this study might torch the avenues of the future with a scintillating prospectus in the beyond the Moore era. Given the fact that the MTJ falls short in yielding adequate Slonczewski (Is∥) at a lower voltage, it was imperative to provide an elevated bias to achieve the stipulated Is∥for the switching presented in Figs. 6(c) and 6(d). III. RESULTS We commence this section by first describing the attributes of a conventional MTJ depicted in Fig. 1(a), in conjunction with the STT-based switching of the free layer. The MTJ features a MgO barrier of thickness 1 nm sandwiched between the FM layers. The band diagram of the device is illustrated in Fig. 2(a) and the I–V 24 October 2024 05:50:34 FIG. 6. (a) STT-facilitated switching of the conentional MTJ from the APC to PC(V = 60 mV) and (b) PC to APC(V = −74 mV). STT-based switching of the NHMTJ from the (c) APC to PC(V = 7.3 mV) and the (d) PC to APC(V = −7.3 mV). FIG. 6. (a) STT-facilitated switching of the conentional MTJ from the APC to PC(V = 60 mV) and (b) PC to APC(V = −74 mV). STT-based switching of the NHMTJ from the (c) APC to PC(V = 7.3 mV) and the (d) PC to APC(V = −7.3 mV). AIP Advances 14, 045001 (2024); doi: 10.1063/9.0000706 AIP Advances 14, 045001 (2024); doi: 10.1063/9.0000706 © Author(s) 2024 © Author(s) 2024 AIP Advances ARTICLE pubs.aip.org/aip/adv FIG. 7. Variation of TMR and Is∥of the (a) conventional-MTJ with thickness of the insulating layer and the (b) NHMTJ with thickness of the central barriers while keeping the terminal barriers intact at 10 mV. AIP Advances FIG. 7. Variation of TMR and Is∥of the (a) conventional-MTJ with thickness of the insulating layer and the (b) NHMTJ with thickness of the central barriers while keeping the terminal barriers intact at 10 mV. FIG. 7. Variation of TMR and Is∥of the (a) conventional-MTJ with thickness of the insulating layer and the (b) NHMTJ with thickness o terminal barriers intact at 10 mV. of the (a) conventional-MTJ with thickness of the insulating layer and the (b) NHMTJ with thickness of the central barriers while keeping th characteristics for the PC and the APC is showcased in Fig. 3(a). The conventional MTJ delivers a peak TMR of 240%, as depicted in Fig. 3(b). Figures 6(a) and 6(b) describes the STT-facilitated switch- ing of the free FM from the PC to APC along with the APC to PC at switching biases −74 and 60 mV, respectively. The discrep- ancy in magnitude of the SBPC→APC and SBAPC→PC emanates from the asymmetrical Is∥−V characteristics of the MTJ presented in Fig. 5(a). the spin current but improves the spin selectivity thereby improv- ing the TMR. III. RESULTS Apart from that, the opposite behavior of the TMR and the Is∥of the NHMTJ with the increase in the thickness of the central barriers renders the designer a provision to modulate the per- formance matrices according to his need. These results also provide us with an estimate of how errors in fabricating the thickness of the NHMTJ shall impact its performance. g The spin current and TMR of the conventional MTJ can be dis- cerned from the spin selective transmittance of the PC and APC, underpinned by single barrier tunneling.13 24 October 2024 05:50:34 IV. CONCLUSION This incurred an average switching energy for the alterations between the PC to APC and APC to PC of 30 fJ. ACKNOWLEDGMENTS The author A.S. acknowledges the support by the Science and Engineering Research Board (SERB), Government of India, Grant No. SRG/2023/001327. Conversely, the NHMTJ offered an average switching energy of a meager 1.96 fJ under analogous conditions. In Fig. 7 we have shown the dependence of the MTJ and the NHMTJ with variations in the thickness of the insulating layers. In the case of the MTJ, we see that reducing the insulator thickness reduces both the TMR and the spin current. Whereas in the NHMTJ increasing the thickness of two central barriers reduces the transmission and hence decreases Author Contributions 5Z. Jin, M. A. I. Mohd Noor Sam, M. Oogane, and Y. Ando, Sensors 21, 668 (2021). Sabarna Chakraborti: Conceptualization (lead); Data curation (lead); Formal analysis (lead); Funding acquisition (lead); Inves- tigation (lead); Methodology (supporting); Project administra- tion (supporting); Resources (supporting); Software (supporting); Supervision (supporting); Validation (lead); Visualization (lead); Writing – original draft (lead); Writing – review & editing (lead). Korra Vamshi Krishna: Conceptualization (supporting); Data curation (supporting); Formal analysis (supporting); Funding acquisition (supporting); Investigation (supporting); Methodology (supporting); Project administration (supporting); Resources (sup- porting); Software (supporting); Supervision (supporting); Valida- tion (supporting); Visualization (supporting); Writing – original draft (supporting); Writing – review & editing (supporting). Viren- dra Singh: Software (supporting); Supervision (supporting); Vali- dation (supporting); Visualization (supporting). Abhishek Sharma: Conceptualization (lead); Data curation (supporting); Formal analy- sis (supporting); Funding acquisition (lead); Investigation (support- ing); Methodology (lead); Project administration (lead); Resources (lead); Software (lead); Supervision (lead); Validation (supporting); Visualization (supporting); Writing – original draft (supporting); Writing – review & editing (supporting). 6A. Meo, J. Chureemart, R. W. Chantrell, and P. Chureemart, Sci. Rep. 12, 3380 (2022). 7J. Slaughter, K. Nagel, R. Whig, S. Deshpande, S. Aggarwal, M. DeHerrera, J. Janesky, M. Lin, H.-J. Chia, M. Hossain et al., in 2016 IEEE International Electron Devices Meeting (IEDM) (IEEE, 2016), pp. 21–25. 8C.-C. Wang and C.-P. Kuo, in 2021 IEEE International Symposium on Circuits and Systems (ISCAS) (IEEE, 2021), pp. 1–4. 9 9J. Lee, D. Shin, Y. Kim, and H.-J. Yoo, in 2016 IEEE International Symposium on Circuits and Systems (ISCAS) (IEEE, 2016), pp. 1010–1013. 10E. Yu, S. Cho, H. Shin, and B.-G. Park, IEEE Electron Device Lett. 40, 562 (2019). 11W. Yang, Y. Cao, J. Han, X. Lin, X. Wang, G. Wei, C. Lv, A. Bournel, and W. Zhao, Nanoscale 13, 862 (2021). 12E. Balcı, U. O. Akkus, and S. Berber, ACS Appl. Mater. Interfaces 11, 3609 (2018). 13S. Chakraborti and A. Sharma, Nanotechnology 34, 185206 (2023). 14 14S. Datta, Electronic Transport in Mesoscopic Systems (Cambridge University Press, 1997). 15J. C. Slonczewski, J. Magn. Magn. Mater. 159, L1 (1996). 16D. Ralph, Y.-T. Cui, L. Liu, T. Moriyama, C. Wang, and R. Buhrman, Philos. Trans. R. Soc., A 369, 3617 (2011). 17B. Tudu and A. Tiwari, Vacuum 146, 329 (2017). 18 18R. Prakash, B. P. Singh Rathore, and D. Kaur, J. Alloys Compd. 726, 693 (2017). 19A. Sharma, A. A. Tulapurkar, and B. Muralidharan, Appl. Phys. Lett. 112, 192404 (2018). DATA AVAILABILITY The data that support the findings of this study are available from the corresponding author upon reasonable request. 20D. Datta, B. Behin-Aein, S. Datta, and S. Salahuddin, IEEE Trans. Nanotechnol. 11, 261 (2012). 24 October 2024 05:50:34 21B. Tao, C. Wan, P. Tang, J. Feng, H. Wei, X. Wang, S. Andrieu, H. Yang, M. Chshiev, X. Devaux et al., Nano Lett. 19, 3019 (2019). 22A. M. Deac, A. Fukushima, H. Kubota, H. Maehara, Y. Suzuki, S. Yuasa, Y. 22A. M. Deac, A. Fukushima, H. Kubota, H. Maehara, Y. Suzuki, S. Yuasa, Y. Nagamine, K. Tsunekawa, D. D. Djayaprawira, and N. Watanabe, Nat. Phys. 4, 803 (2008). 1A. Sharma, A. A. Tulapurkar, and B. Muralidharan, J. Appl. Phys. 129, 233901 (2021). 2D. Robbes, Sens. Actuators, A 129, 86 (2006). 3A. Sharma, A. A. Tulapurkar, and B. Muralidharan, Phys. Rev. Appl. 8, 064014 (2017). 4T. Chen, R. K. Dumas, A. Eklund, P. K. Muduli, A. Houshang, A. A. Awad, P. Dürrenfeld, B. G. Malm, A. Rusu, and J. Åkerman, Proc. IEEE 104, 1919 (2016). Conflict of Interest The authors have no conflicts to disclose. AIP Advances 14, 045001 (2024); doi: 10.1063/9.0000706 © Author(s) 2024 AIP Advances 14, 045001 (2024); doi: 10.1063/9.0000706 14, 045001-5 ARTICLE AIP Advances pubs.aip.org/aip/adv Author Contributions AIP Advances 14, 045001 (2024); doi: 10.1063/9.0000706 © Author(s) 2024 4T. Chen, R. K. Dumas, A. Eklund, P. K. Muduli, A. Houshang, A. A. Awad, P. Dürrenfeld, B. G. Malm, A. Rusu, and J. Åkerman, Proc. IEEE 104, 1919 (2016). 1A. Sharma, A. A. Tulapurkar, and B. Muralidharan, J. Appl. Phys. 129, 233901 (2021). 2 REFERENCES Nagamine, K. Tsunekawa, D. D. Djayaprawira, and N. Watanabe, Nat. Phys. 4, 803 (2008). 23S. Ikeda, K. Miura, H. Yamamoto, K. Mizunuma, H. Gan, M. Endo, S. Kanai, J. Hayakawa, F. Matsukura, and H. Ohno, Nat. Mater. 9, 721 (2010). 24 24H. Sato, E. Enobio, M. Yamanouchi, S. Ikeda, S. Fukami, S. Kanai, F. Matsukura, and H. Ohno, Appl. Phys. Lett. 105, 062403 (2014). 25M. Gajek, J. Nowak, J. Sun, P. Trouilloud, E. O’sullivan, D. Abraham, M. Gaidis, G. Hu, S. Brown, Y. Zhu et al., Appl. Phys. Lett. 100, 132408 (2012). AIP Advances 14, 045001 (2024); doi: 10.1063/9.0000706 © Author(s) 2024 14, 045001-6 © Author(s) 2024
https://openalex.org/W4300809525	https://aura.abdn.ac.uk/bitstream/2164/9178/1/1_s2.0_S0140673616316786_main.pdf	English	null	Global, regional, and national incidence, prevalence, and years lived with disability for 310 diseases and injuries, 1990–2015: a systematic analysis for the Global Burden of Disease Study 2015	Carolina Digital Repository (University of North Carolina at Chapel Hill)	2,016	cc-by	66,009	Lancet 2016; 388: 1545–602 Background Non-fatal outcomes of disease and injury increasingly detract from the ability of the world’s population to live in full health, a trend largely attributable to an epidemiological transition in many countries from causes aﬀ ecting children, to non-communicable diseases (NCDs) more common in adults. For the Global Burden of Diseases, Injuries, and Risk Factors Study 2015 (GBD 2015), we estimated the incidence, prevalence, and years lived with disability for diseases and injuries at the global, regional, and national scale over the period of 1990 to 2015. Methods We estimated incidence and prevalence by age, sex, cause, year, and geography with a wide range of updated and standardised analytical procedures. Improvements from GBD 2013 included the addition of new data sources, updates to literature reviews for 85 causes, and the identiﬁ cation and inclusion of additional studies published up to November, 2015, to expand the database used for estimation of non-fatal outcomes to 60 900 unique data sources. Prevalence and incidence by cause and sequelae were determined with DisMod-MR 2.1, an improved version of the DisMod-MR Bayesian meta-regression tool ﬁ rst developed for GBD 2010 and GBD 2013. For some causes, we used alternative modelling strategies where the complexity of the disease was not suited to DisMod-MR 2.1 or where incidence and prevalence needed to be determined from other data. For GBD 2015 we created a summary indicator that combines measures of income per capita, educational attainment, and fertility (the Socio-demographic Index [SDI]) and used it to compare observed patterns of health loss to the expected pattern for countries or locations with similar SDI scores. Findings We generated 9·3 billion estimates from the various combinations of prevalence, incidence, and YLDs for causes, sequelae, and impairments by age, sex, geography, and year. In 2015, two causes had acute incidences in excess of 1 billion: upper respiratory infections (17·2 billion, 95% uncertainty interval [UI] 15·4–19·2 billion) and diarrhoeal diseases (2·39 billion, 2·30–2·50 billion). Eight causes of chronic disease and injury each aﬀ ected more than 10% of the world’s population in 2015: permanent caries, tension-type headache, iron-deﬁ ciency anaemia, age-related and other hearing loss, migraine, genital herpes, refraction and accommodation disorders, and ascariasis. The impairment that aﬀ ected the greatest number of people in 2015 was anaemia, with 2·36 billion (2·35–2·37 billion) individuals aﬀ ected. Articles Articles yright © The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY license Lancet 2016; 388: 1545–602 The second and third leading impairments by number of individuals aﬀ ected were hearing loss and vision loss, respectively. Between 2005 and 2015, there was little change in the leading causes of years lived with disability (YLDs) on a global basis. NCDs accounted for 18 of the leading 20 causes of age-standardised YLDs on a global scale. Where rates were decreasing, the rate of decrease for YLDs was slower than that of years of life lost (YLLs) for nearly every cause included in our analysis. For low SDI geographies, Group 1 causes typically accounted for 20–30% of total disability, largely attributable to nutritional deﬁ ciencies, malaria, neglected tropical diseases, HIV/AIDS, and tuberculosis. Lower back and neck pain was the leading global cause of disability in 2015 in most countries. The leading cause was sense organ disorders in 22 countries in Asia and Africa and one in central Latin America; diabetes in four countries in Oceania; HIV/AIDS in three southern sub-Saharan African countries; collective violence and legal intervention in two north African and Middle Eastern countries; iron-deﬁ ciency anaemia in Somalia and Venezuela; depression in Uganda; onchoceriasis in Liberia; and other neglected tropical diseases in the Democratic Republic of the Congo. Interpretation Ageing of the world’s population is increasing the number of people living with sequelae of diseases and injuries. Shifts in the epidemiological proﬁ le driven by socioeconomic change also contribute to the continued increase in years lived with disability (YLDs) as well as the rate of increase in YLDs. Despite limitations imposed by gaps in data availability and the variable quality of the data available, the standardised and comprehensive approach of the GBD study provides opportunities to examine broad trends, compare those trends between countries or subnational geographies, benchmark against locations at similar stages of development, and gauge the strength or weakness of the estimates available. Global, regional, and national incidence, prevalence, and years lived with disability for 310 diseases and injuries, 1990–2015: a systematic analysis for the Global Burden of Disease Study 2015 GBD 2015 Disease and Injury Incidence and Prevalence Collaborators* This online publication has been corrected. The corrected version first appeared at thelancet.com on January 5, 2017 See Editorial page 1447 See Comment pages 1448 and 1450 Collaborators listed at the end of the Article Correspondence to: Prof Theo Vos, Institute for Health Metrics and Evaluation, Seattle, WA 98121, USA tvos@uw.edu Introduction Along with broad recognition that data from some regions were sparse and that more and higher quality data in general would probably improve estimation, useful debates on the GBD results have been published. These debates have focused on the analysis or presentation of individual diseases, such as changes over time in GBD estimates of dementia,5 the accuracy of HIV incidence estimates,6,7 the absence of sepsis as a disease,8,9 the quality of some cancer registry data,10 and the absence of mental disorders as sequelae of neglected tropical diseases.11 The GBD empirical approach to measuring the public’s view of health state severity has generated substantial interest with questions about the relative importance of diﬀ erent dimensions of health,12,13 the quantiﬁ cation of health loss,14,15 and discussions of the transferability of judgments about relative health to conventional notions of disability and dependence.5 In each cycle of the GBD, we seek to improve the estimates, reﬂ ecting published and unpublished critique through the acquisition of new data, expansion of the network of collaborators, changes in how data are corrected for bias, advances in modelling techniques, and the targeted expansion of the GBD cause list. Although substantial progress has been made toward reducing mortality and extending life expectancy throughout the world over the past few decades, the epidemiological transition is manifest in the growing importance of non-fatal diseases, outcomes, and injuries which pose, partly as a consequence of decreasing death rates, a rising challenge to the ability of the world’s population to live in full health. Complementing information on deaths by age, sex, cause, geography, and time with equally detailed information on disease incidence, prevalence, and severity is key to a balanced debate in health policy. For this reason, the Global Burden of Disease (GBD) Study uses the disability- adjusted life-year (DALY), combining years of life lost (YLLs) due to mortality and years lived with disability (YLDs) in a single metric. One DALY can be thought of as one lost year of healthy life. The sum of DALYs in a population can be thought of as the gap between the population’s present health status and an ideal situation where the entire population lives to an advanced age, free of disease. Overall approach We estimated incidence and prevalence by age, sex, cause, year, and geography using a wide range of updated and standardised analytical procedures. The overall logic of our analytical approach is shown for the entire non-fatal estimation process in ﬁ gure 1. The appendix provides a single source for detail of inputs, analytical processes, and outputs and methods speciﬁ c to each cause. This study complies with the Guidelines for Accurate and Transparent Health Estimates Reporting (GATHER) recommendations (methods appendix pp 1, 608–10).17 The estimates from GBD 2013 drew attention to large increases in the number of YLDs over the previous decade, whereas rates of YLDs for most causes remained stable or showed only small decreases.4 The GBD 2013 assessment largely attributed increases in the number of YLDs to musculoskeletal disorders, mental and substance use disorders, neurological disorders, and chronic respiratory diseases, as well as population growth and ageing. GBD 2013 also brought attention to increased diﬀ erences in trends between mortality and morbidity for many causes. YLDs as a proportion of DALYs increased globally, a manifestation of the continuing epidemiological transition in low-income and middle- income countries. Decreases in mortality from diseases such as pneumonia, diarrhoea, maternal and neonatal disorders, and an absence of progress in reducing YLD rates continued to drive a transition toward a greater global number of YLDs. Introduction Assessments of how diﬀ erent diseases lead to multimorbidity and reductions in functional health status are important for both health system planning1 and a broader range of social policy issues such as the appropriate age for retirement in some countries.2,3 Many challenges in making standardised estimates of non-fatal health outcomes are similar to those aﬀ ecting mortality estimates (including variations in case deﬁ nitions, data collection methods, variable quality of data collection, conﬂ icting data, and missing data) but are compounded by more sparse and varied data sources, the need to characterise each disease by its disabling sequelae or consequence(s), and the need to quantify the severity of these consequences. The standardised approach of the annual GBD updates addresses these measurement problems to enhance comparability between causes by geography and over time. The primary objective of this component of the GBD was to use all available data of suﬃ cient quality to generate reliable and valid assessments of disease and injury sequelae incidence, prevalence, and YLDs for all 310 causes in the GBD cause hierarchy for 591 locations in the GBD study during 1990–2015. We describe the change over time and between populations in relation to where countries fall on the development continuum.16 Continuing eﬀ orts to improve data and code transparency are an important part of the GBD cycle. These results thus supersede any previous publications about the GBD on disease incidence, prevalence, and YLDs. See Online for appendix Funding Bill & Melinda Gates Foundation. Funding Bill & Melinda Gates Foundation. Funding Bill & Melinda Gates Foundation. Copyright © The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY www.thelancet.com Vol 388 October 8, 2016 1545 Articles Articles See Online for appendix Geographies in GBD 2015 The geographies included in GBD 2015 have been arranged into a set of hierarchical categories composed of seven super-regions and a further nested set of 21 regions containing 195 countries and territories. Eight additional subnational assessments were done for Brazil, China, India, Japan, Kenya, Saudi Arabia, South Africa, Sweden, and the USA (methods appendix pp 611–24). For this study we present data at the national and territory level. Geographies in GBD 2015 1546 www.thelancet.com Vol 388 October 8, 2016 Articles Input data Process Results Database Non-fatal estimation process Raw data source Data adjustment Database Alternative disease modelling Impairment and underlying cause Post DisMod-MR DisMod-MR Injury modelling Disability weights Final burden estimates Colours of non-fatal estimation Case notiﬁcations 1 Data sources 2 Data adjustment 3b Alternative disease modelling strategies (details ﬁgure 1B) 3c Injury modelling strategy 5 Severity distribution 6 Disability weights 7 Comorbidity 8 YLDs 3a DisMod-MR 2.1 estimation 4 Impairment and underlying cause estimation Expansion factors for case notiﬁcations Population-at-risk data Seroprevalence data Disease registries Birth registries Active screening Intervention coverage Vital registration Surveillance Community surveys National surveys Outpatient hospital data Claims data: outpatient visits Inpatient hospital data SMR data from cohort studies Cohort study MEPS Expert estimates duration untreated injuries Sequelae mapped to health states Household surveys Open access web- based survey GBD collaborator advice Surveys with diagnostic information and SF-12 Opportunistic surveys by IHME to ﬁll SF-12 for 60 lay descriptions Cohort follow-up studies Claims data: inpatient visits Adjustment for under-reporting HIV/AIDS and TB Malaria Seroprevalence to incidence models Case fatality proportion and cause of death rate models Apply aetiology or severity proportions to disease or impairment morbidity estimates Scale impairment prevalence by underlying cause or severity to envelope Neonatal disorders Cancer Age–sex splitting Add study-level covariaties Pre DisMod bias correction Adjustment for multiple outpatient visits per prevalent or incident case based on claims data Adjustment from primary code to all code based on claims data for causes with long duration Adjustment for multiple admissions in same individual Generate cause–nature of injury matrices with negative binomial models Determine most severe nature of injury category in any individual Compute excess mortality before from available incidence or prevalence and CSMR data Adjusted input data Non-fatal database: prevalence, incidence, excess mortality rate, RR, SMR, duration, remission, severity proportions, and intermediary modelling variables CSMR from CoDCorrect Study covariates Disability weights for 235 health states Country covariates Apply cause–nature injury of matrices Regression to estimate disability weight by cause in survey respondents controlling for comorbidity YLD to YLL ratio for 12 residual causes without primary data Analysis of paired comparison and population health equivalence responses Incidence by cause of injury code Prevalence and incidence by disease or impairment Proportion of disease or impairment sequelae or causes Scaled proportion of disease or impairment sequelae or underlying causes Proportion by sequelae Prevalence and incidence of sequelae from alternative models Prevalence and incidence of sequelae Unadjusted YLD by sequelae Probability of long-term disability Estimate duration of short-term disability DisMod-MR 2.1 Meta-analysis Short-term incidence by cause–nature and inpatient or outpatient Short-term prevalence by cause–nature and inpatient or outpatient Lay descriptions for 235 health states Long-term incidence by cause–nature combination Long-term prevalence by cause–nature and inpatient or outpatient Map EQ5D to SF-12 DisMod-MR 2.1 Scale to 100% Scale to 100% Map SF-12 to GBD disability weights Meta-analysis proportion by severity level DisMod analysis proportion by severity level Comorbidity correction (COMO) Prevalence and incidence of sequelae of impairment or diseases (by severity or underlying cause) YLLs residual causes without primary data YLDs for each disease and injury by age, sex, year, and country Figure 1: Analytical ﬂ ow chart for the estimation of cause-speciﬁ c YLDs by location, age, sex, and year for GBD 2015 Ovals represent data inputs, square boxes represent analytical steps, cylinders represent databases, and parallelograms represent intermediate and ﬁ nal results. Geographies in GBD 2015 The ﬂ ow chart is colour-coded by major estimation component: raw data sources, in pink; data adjustments, in yellow; DisMod-MR 2.1 estimation, in purple; alternative modelling strategies, in light green; injury modelling strategy, in dark green; estimation of impairments and underlying causes, in brown; post-DisMod-MR and comorbidity correction, in blue; disability weights, in orange; and cause of death and demographic inputs, in grey. GBD=Global Burden of Disease. TB=tuberculosis. SF-12=Short Form 12 questions. MEPS=Medical Expenditure Panel Surveys. CSMR=cause-speciﬁ c mortality rate. SMR=standardised mortality ratio. YLDs=years lived with disability. YLLs=years of life lost. IHME=Institute for Health Metrics and Evaluation. Figure 1: Analytical ﬂ ow chart for the estimation of cause-speciﬁ c YLDs by location, age, sex, and year for GBD 2015 Figure 1: Analytical ﬂ ow chart for the estimation of cause-speciﬁ c YLDs by location, age, sex, and year for GBD 2015 Ovals represent data inputs, square boxes represent analytical steps, cylinders represent databases, and parallelograms represent intermediate and ﬁ nal results. The ﬂ ow chart is colour-coded by major estimation component: raw data sources, in pink; data adjustments, in yellow; DisMod-MR 2.1 estimation, in purple; alternative modelling strategies, in light green; injury modelling strategy, in dark green; estimation of impairments and underlying causes, in brown; post-DisMod-MR and comorbidity correction, in blue; disability weights, in orange; and cause of death and demographic inputs, in grey. GBD=Global Burden of Disease. TB=tuberculosis. SF-12=Short Form 12 questions. MEPS=Medical Expenditure Panel Surveys. CSMR=cause-speciﬁ c mortality rate. SMR=standardised mortality ratio. YLDs=years lived with disability. YLLs=years of life lost. IHME=Institute for Health Metrics and Evaluation. Figure 1: Analytical ﬂ ow chart for the estimation of cause-speciﬁ c YLDs by location, age, sex, and year for GBD 2015 Ovals represent data inputs, square boxes represent analytical steps, cylinders represent databases, and parallelograms represent intermediate and ﬁ nal results. The ﬂ ow chart is colour-coded by major estimation component: raw data sources, in pink; data adjustments, in yellow; DisMod-MR 2.1 estimation, in purple; alternative modelling strategies, in light green; injury modelling strategy, in dark green; estimation of impairments and underlying causes, in brown; post-DisMod-MR and comorbidity correction, in blue; disability weights, in orange; and cause of death and demographic inputs, in grey. GBD=Global Burden of Disease. TB=tuberculosis. SF-12=Short Form 12 questions. MEPS=Medical Expenditure Panel Surveys. CSMR=cause-speciﬁ c mortality rate. SMR=standardised mortality ratio. YLDs=years lived with disability. YLLs=years of life lost. IHME=Institute for Health Metrics and Evaluation. www.thelancet.com Vol 388 October 8, 2016 Period of analysis claims data, we generated several correction factors to account for bias in health service encounter data from elsewhere, which were largely available to us aggregated by ICD code and by primary diagnosis only. First, for chronic disorders, we estimated the ratio between prevalence from primary diagnoses and prevalence from all diagnoses associated with a claim. Second, we used the claims data to generate the average number of outpatient visits per disorder. Similarly, we generated per person discharge rates from hospital inpatient data in the USA and New Zealand, the only sources with unique patient identiﬁ ers available for GBD 2015. A complete set of age-speciﬁ c, sex-speciﬁ c, cause- speciﬁ c, and geography-speciﬁ c incidence and prevalence numbers and rates were computed for the years 1990, 1995, 2000, 2005, 2010, and 2015. In this study we focus on trends for main and national results over the past decade, from 2005 to 2015, together with more detailed results for 2015. Online data visualisations at vizhub provide access to results for all GBD metrics. Non-fatal modelling strategies vary substantially between causes. Figure 1 outlines the general process of non-fatal outcome estimation from data inputs to ﬁ nalisation of YLD burden results; step 3b of that process identiﬁ es alternative modelling approaches used for speciﬁ c causes (methods appendix pp 603, 604). The starting point for non-fatal estimation is the compilation of data sources identiﬁ ed through systematic analysis and extractions based on predetermined inclusion and exclusion criteria (methods appendix p 603). As part of the inclusion criteria, we deﬁ ned disease-speciﬁ c or injury-speciﬁ c reference case deﬁ nitions and study methods, as well as alternative allowable case deﬁ nitions and study methods which were adjusted for if we detected a systematic bias. We used 15 types of primary data sources representing disease prevalence, incidence, mortality risk, duration, remission, or severity in the estimation process (oval shapes in ﬁ gure 1). p In GBD 2013, we calculated a geographical and temporal data representativeness index (DRI) of non-fatal data sources for each cause or impairment. The DRI represents the fraction of countries for which any incidence, prevalence, remission, or mortality risk data were available for a cause. This metric quantiﬁ es data availability, not data quality. The overall DRI and period-speciﬁ c DRI measures for each cause and impairment are presented in the methods appendix (pp 662–68). For data visualisations at vizhub see http://vizhub. healthdata.org/gbd-compare For data in GBD 2015 see http:// ghdx.healthdata.org/global- burden-disease-study-2015 For Global Health Data Exchange see http://ghdx.healthdata.org Period of analysis DRI ranged from 90% for nine causes, including tuberculosis and measles, to less than 5% for acute hepatitis C and the category of other exposures to mechanical forces. Required case reporting resulted in high DRI values for notiﬁ able infectious diseases; the network of population-based registries for cancers resulted in a DRI of above 50%. DRI values ranged from 6·1% in North Korea to 91·3% in the USA. Many high-income countries, as well as Brazil, India, and China, had DRI values above 63%; data availability was low in several countries, including Equatorial Guinea, Djibouti, and South Sudan. www.thelancet.com Vol 388 October 8, 2016 Data sources For this iteration of the study, we updated data searches through systematic data and literature reviews for 85 causes published up to Oct 31, 2015. For other causes, input from GBD collaborators resulted in the identiﬁ cation and inclusion of a small number of additional studies published after January, 2013. Data were systematically screened from household surveys archived in the Global Health Data Exchange, sources suggested to us by in-country experts, and surveys identiﬁ ed in major multinational survey data catalogues and Ministry of Health and Central Statistical Oﬃ ce websites. Case notiﬁ cations reported to WHO were updated up to and including 2015. Citations for all data sources used for non- fatal estimation in GBD 2015 are provided in searchable form through a new web tool. A description of the search terms used for cause-speciﬁ c systematic reviews, inclusion and exclusion criteria, and the preferred and alternative case deﬁ nitions and study methods are detailed by cause in the methods appendix (pp 26–601). List of causes and sequelae The cause and sequelae list was expanded based upon feedback after the release of GBD 2013 and input from GBD 2015 collaborators. Nine causes for which non-fatal outcomes are estimated were added: Ebola virus disease, motor-neuron disease, environmental heat and cold exposure, four subtypes of leukaemia, and two subtypes of non-melanoma skin cancer (methods appendix pp 625–53). The incorporation of these changes expanded the cause list from the 301 causes with non-fatal estimates examined in GBD 2013, to 310 causes with non-fatal estimates and from 2337 to 2619 unique sequelae at Level 6 of the hierarchy. At the newly created Level 5 of the hierarchy there were 154 summary sequela categories. The methods appendix (pp 654–61) provides a list of International Classiﬁ cation of Diseases version 9 (ICD-9) and version 10 (ICD-10) codes used in the extraction of hospital and claims data, mapped to GBD 2015 non-fatal causes, impairments, and nature of injury categories. The GBD cause and sequelae list is organised hierarchically (methods appendix 625–53). At Level 1 there are three cause groups: communicable, maternal, neonatal, and nutritional diseases (Group 1 diseases); non-communicable diseases; and injuries. These Level 1 aggregates are subdivided at Level 2 of the hierarchy into 21 cause groupings. The disaggregation into Levels 3 and 4 contains the ﬁ nest level of detail for causes captured in GBD 2015. Sequelae of diseases and injuries are organised at Levels 5 and 6 of the hierarchy. The ﬁ nest detail for all sequelae estimated in GBD is at Level 6 and is aggregated into summary sequelae categories (Level 5) for causes with large numbers of sequelae. Sequelae in GBD are mutually exclusive and collectively exhaustive, and thus our YLD estimates at each level of the hierarchy sum to the total of the level above. Prevalence aggregations are estimated at the level of individuals who might have more than one sequela or disease and therefore are not additive. 1547 Articles Articles For DisMod-MR 2.1 engine and the code see http://ghdx. healthdata.org/global-burden- disease-study-2015 Non-fatal disease models In addition to the corrections applied to claims and hospital data, a number of other adjustments were applied including age–sex splitting, bias correction, adjustments for under-reporting of notiﬁ cation data, and computing expected values of excess mortality. In GBD 2013, we estimated expected values of excess mortality from prevalence or incidence and cause-speciﬁ c mortality rate data for a few causes only, including tuberculosis and chronic obstructive pulmonary disease. In order to achieve greater consistency between our cause of death and non-fatal data, we adopted this strategy systematically for GBD 2015. We matched every prevalence data point (or incidence datapoint for short duration disorders) with the cause-speciﬁ c mortality rate value corresponding to the age range, sex, year, and location of the datapoint. The ratio of cause-speciﬁ c mortality rate to prevalence is conceptually equivalent to an excess mortality rate. Hospital inpatient data were extracted from 284 country-year and 976 subnational-year combinations from 27 countries in North America, Latin America, Europe, and New Zealand. Outpatient encounter data were available from the USA, Norway, Sweden, and Canada for 48 country-years. For GBD 2015, we also accessed aggregate data derived from claims information in a database of US private and public insurance schemes for the years 2000, 2010, and 2012. From the linked To estimate non-fatal health outcomes in previous iterations of GBD, most diseases and impairments were modelled in DisMod-MR, a Bayesian meta-regression tool originally developed for GBD 2010 (step 3a in ﬁ gure 1).18 DisMod-MR was designed to address statistical challenges in estimation of non-fatal health outcomes, and for 1548 Articles synthesis of often sparse and heterogeneous epidemio- logical data. For GBD 2015, the computational engine of DisMod-MR 2.1 remained unchanged, but we substantially rewrote the code that organises the ﬂ ow of data and settings at each level of the analytical cascade. The sequence of estimation occurs at ﬁ ve levels: global, super- region, region, country, and where applicable, subnational locations (appendix pp 611–24). At each level of the cascade, the DisMod-MR 2.1 computational engine enforces consistency between all disease parameters. For GBD 2015, we generated ﬁ ts for the years 1990, 1995, 2000, 2005, 2010, and 2015. We log-linearly interpolated estimates for the intervening years in each 5-year period. Greater detail on DisMod-MR 2.1 is available at Global Health Data Exchange and the methods appendix (pp 7–11). same as the general cancer health states. Disability weights We used the same disability weights as in GBD 2013 (see methods appendix pp 669–94 for a complete listing of the lay descriptions and values for the 235 health states used in GBD 2015). Non-fatal disease models For motor- neuron disease we accessed the Pooled Resource Open- Access ALS Clinical Trials (PROACT) database containing detailed information on symptoms and impairments for more than 8500 patients who took part in the trials.23 Comorbidity In step 7, we estimated the co-occurrence of diﬀ erent diseases by simulating 40 000 individuals in each geography–age–sex–year combination as exposed to the independent probability of having any of the sequelae included in GBD 2015 based on disease prevalence. We tested the contribution of dependent and independent comorbidity in the US Medical Expenditure Panel Surveys (MEPS) data, and found that independent comorbidity was the dominant factor even though there are well known examples of dependent comorbidity. Age was the main predictor of comorbidity such that age-speciﬁ c microsimulations accommodated most of the required comorbidity correction. Taking dependent comorbidity into account changed the overall YLDs estimated in the MEPS data by only 2·5% (and ranging from 0·6% to 3·4% depending on age) in comparison to assuming independent comorbidity (methods appendix pp 18–20).24 In previous iterations of GBD, custom models were created for a short list of causes for which the compartment model underpinning DisMod (susceptible, diseased, and dead) was insuﬃ cient to capture the complexity of the disease or for which incidence and prevalence needed to be derived from other data. Step 3b of ﬁ gure 1 describes the development of custom models with greater detail shown in the methods appendix ﬁ gure 1B (p 604, and for associated write-ups pp 26–601) for HIV/AIDS, tuberculosis, malaria, cancer, neonatal disorders, infectious diseases for which we derived incidence from seroprevalence data, and infectious diseases for which we derived incidence from cause of death rates and pooled estimates of the case fatality proportion. In GBD 2013, we estimated the country–age–sex–year prevalence of nine impairments (step 4 of ﬁ gure 1). Impairments in GBD are disorders or speciﬁ c domains of functional health loss that are spread across many GBD causes as sequelae and for which there are better data to estimate the occurrence of the overall impairment than for each sequela based on the underlying cause. Overall impairment prevalence was estimated with DisMod-MR 2.1 except for anaemia, for which spatiotemporal Gaussian Process regression methods were applied. We constrained cause-speciﬁ c estimates of impairments, such as in the 19 causes of blindness, to sum to the total prevalence estimated for that impairment. Anaemia, epilepsy, hearing loss, heart failure, and intellectual disability were estimated at diﬀ erent levels of severity. www.thelancet.com Vol 388 October 8, 2016 YLD computation p We report 95% uncertainty intervals (95% UI) for each quantity in this analysis using 1000 samples from the posterior distribution of prevalence and 1000 samples of the disability weight to generate 1000 samples of the YLD distribution. The 95% UI is reported as the 25th and 975th values of the distribution. We report signiﬁ cant changes in disease estimates between countries or over time if the change was noted in more than 950 of the 1000 samples computed for each result. For GBD 2015, we computed age-standardised prevalence YLD rates from the updated world population age standard developed for GBD 2013.25 Less common diseases and their sequelae were included in 35 residual categories (methods appendix pp 695–97). For 22 of these residual categories, estimates were made from epidemiological data for incidence or prevalence. For 13 residual categories, we estimated YLDs by multiplying the residual YLL estimates by the ratio of YLDs to YLLs from the estimates for explicitly modelled Level 3 causes in the same disease category. Severity distributions y In step 5, sequelae were further deﬁned in terms of severity for 194 causes at Level 4 of the hierarchy (ﬁ gure 1A). We generally followed the same approach for estimating the distribution of severity as in GBD 2013. For Ebola virus disease, we created a health state for the infectious disease episode with duration derived from average hospital admission times, and a health state for ongoing postinfection malaise and joint problems based on four follow-up studies19–22 from which we derived an average duration. The health states for the subtypes of leukaemia and non-melanoma skin cancer were the Articles rate.26 For GBD 2015, we excluded mean age of the population because it is directly aﬀ ected by death rates. To improve interpretability for GBD 2015, we computed a Socio-demographic Index (SDI) similar to the computation of the human development index.27 In the SDI, each component was weighted equally and rescaled from zero (for the lowest value observed during 1980–2015) to one (for the highest value observed) for income per capita and average years of schooling, and the reverse for the total fertility rate. The ﬁ nal SDI score was computed as the geometric mean of each of the components. SDI ranged from 0·060 in Mozambique in 1987 to 0·978 in District of Columbia, USA, in 2015. Global incidence and prevalence p We generated over 9·3 billion outcomes of incidence, prevalence, and YLDs for 310 diseases, injuries, and aggregate categories; 2619 unique and aggregate sequelae; nine impairments; 63 age–sex groups; 591 geographies; and 26 individual years from 1990 to 2015. Each of these 9·3 billion estimates was calculated 1000 times to determine uncertainty intervals. Here, we present key summary ﬁ ndings on global incidence of short duration diseases, global prevalence of long-term disorders, global prevalence of impairments, global numbers and rates of YLDs and changes from 2005 to 2015, global YLL and YLD rates of change, patterns of comorbidity, the expected changes in the composition of YLDs with SDI, and country ﬁ ndings of leading causes of YLDs. y g g Disorders of less than 3 months duration and injuries with incidence of more than 1 million cases per year in 2015 are listed in table 1. There were two disorders with incidence greater than 1 billion per year: upper respiratory infections (17·2 billion [95% UI 15·4–19·2 billion]), and diarrhoeal diseases (2·39 billion [2·30–2·50 billion]). A further 13 diseases and injuries caused between 100 million and 1 billion incident cases a year and 16 diseases and injuries had incident cases of between 10 million and 100 million per year (table 1). The disease and injury sequelae with a duration of more than 3 months and a global prevalence of more than 1% in 2015 are presented in table 2, aggregated to the cause level. Prevalence for impairments is presented at the bottom of table 3. Eight out of 56 high-prevalence causes aﬀ ected more than 10% of the world’s population in 2015. A further 48 causes aﬀ ected between 1% and 10% of the world’s population (table 2). Although many of Disorders of less than 3 months duration and injuries with incidence of more than 1 million cases per year in 2015 are listed in table 1. There were two disorders with incidence greater than 1 billion per year: upper respiratory infections (17·2 billion [95% UI 15·4–19·2 billion]), and diarrhoeal diseases (2·39 billion [2·30–2·50 billion]). A further 13 diseases and injuries caused between 100 million and 1 billion incident cases a year and 16 diseases and injuries had incident cases of between 10 million and 100 million per year (table 1). Role of the funding source g The funder of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had ﬁ nal responsibility for the decision to submit for publication. Articles Incidence (thousands) Percentage change (%) 2005 2015 Upper respiratory infections 15 624 257 (13 851 237–17 411 199) 17 230 659 (15 351 516–19 172 439) 10·3 (9·0 to 11·6) Diarrhoeal diseases 2 235 739 (2 139 050–2 348 407) 2 392 517 (2 301 101––2 503 094) 7·0 (6·0 to 8·0)* Permanent caries 426 963 (371 216–484 332) 487 629 (423 507–552 895) 14·2 (13·1 to 15·4)* Otitis media 429 820 (353 915–525 059) 471 027 (386 606–577 286) 9·6 (7·9 to 11·2)* Lower respiratory infections 273 131 (257 286–288 078) 291 759 (276 244–307 004) 6·8 (5·6 to 8·1)* Malaria 339 275 (261 417–447 605) 286 859 (219 712–377 332) –15·4 (–23·0 to –8·1)* Gastritis and duodenitis 185 250 (167 471–204 974) 213 729 (192 486–236 339) 15·4 (10·8 to 17·2%)* Pyoderma 178 382 (172 199–184 147) 207 452 (200 498–213 936) 16·3 (15·6% to 17·0)* Gonococcal infection 138 220 (107 106–184 385) 172 676 (129 731–235 737) 24·9 (17·9 to 31·0)* Interstitial nephritis and urinary tract infections 127 380 (124 308–130 683) 152 295 (148 748–156 177) 19·6 (19·2 to 19·9)* Varicella and herpes zoster 128 678 (124 298–133 090) 142 413 (137 804–147 181) 10·7 (10·0 to 11·4)* Trichomoniasis 121 948 (104 825–141 284) 140 781 (121 207–163 163) 15·4 (14·5 to 16·5)* Acute hepatitis A 109 609 (101 813–117 648) 114 212 (103 349–124 776) 4·2 (–7·2 to 17·0) Hepatitis B 98 277 (86 507–112 527) 111 212 (97 410–126 251) 13·2 (–4·3 to 35·2) Gallbladder and biliary diseases 88 215 (79 276–96 495) 104 322 (93 074–114 430) 18·3 (16·0 to 20·7)* Peptic ulcer disease 83 388 (77 760–89 435) 87 410 (80 343–94 506) 4·8 (2·6 to 7·0)* Dengue 32 749 (18 879–68 335) 79 609 (53 784–169 704) 143·1 (–0·3 to 564·7) Other sense organ diseases 60 659 (58 721–62 579) 69 945 (67 856–72 080) 15·3 (14·6 to 16·0)* Chlamydial infection 56 976 (45 489–70 839) 61 173 (48 871–76 698) 7·4 (5·5 to 9·4)* Maternal abortion, miscarriage, and ectopic pregnancy 53 942 (43 630–67 168) 53 958 (43 224–67 417) 0·0 (–3·8 to 4·0) Syphilis 43 515 (37 479–51 054) 45 413 (37 787–54 921) 4·4 (–0·3 to 8·1) Deciduous caries 41 353 (28 723–58 131) 43 688 (29 630–62 543) 5·6 (2·1 to 8·1)* Genital herpes 29 112 (25 131–33 432) 39 791 (35 569–44 572) 36·7 (32·4 to 41·6)* Urolithiasis 17 875 (16 320–19 728) 22 080 (20 183–24 295) 23·5 (21·2 to 25·6)* Cellulitis 17 312 (15 988–18 739) 21 211 (19 582–22 985) 22·5 (21·0 to 24·1)* Maternal hypertensive disorders 20 416 (17 593–23 417) 20 731 (17 355–24 379) 1·5 (–3·0 to 6·3) Acute hepatitis E 18 869 (17 340–20 580) 19 525 (18 011–21 273) 3·5 (2·2% to 4·7)* Whooping cough 22 457 (17 322–28 268) 16 298 (12 599–20 445) –27·4 (–29·5 to –25·2)* (Table 1 continues on next page) www.thelancet.com Vol 388 October 8, 2016 Socio-demographic Index Socio-demographic Index In GBD 2013, a sociodemographic status variable was computed based on a principal components analysis of income per capita, educational attainment, average age of the population, and the total fertility 1549 Articles Articles Global incidence and prevalence The disease and injury sequelae with a duration of more than 3 months and a global prevalence of more than 1% in 2015 are presented in table 2, aggregated to the cause level. Prevalence for impairments is presented at the bottom of table 3. Eight out of 56 high-prevalence causes aﬀ ected more than 10% of the world’s population in 2015. A further 48 causes aﬀ ected between 1% and 10% of the world’s population (table 2). Although many of 1550 Articles these causes are not among the dominant causes of YLDs because of comparatively low average disability weights, some causes, such as headaches, gynaecological diseases, oral disorders, and skin diseases, put great demands on health system resources by their sheer numbers. Incidence (thousands) Percentage change (%) 2005 2015 (Continued from previous page) Typhoid fever 15 341 (13 454–17 417) 12 538 (10 887–14 283) –18·3 (–20·8 to –15·6)* Maternal sepsis and other maternal infections 11 938 (9457–15 125) 11 817 (9300–15 009) –1·0 (–4·6 to 2·8) Appendicitis 10 159 (9549 to 10 840) 11 619 (10 918 to 12 407) 14·4 (13·5 to 15·2)* Pancreatitis 7160 (6 660–7 694) 8 902 (8 212–9 643) 24·3 (21·9 to 26·6)* Maternal haemorrhage 8438 (6614–10 620) 8740 (6774–11 061) 3·6 (–0·3 to 7·6) Other sexually transmitted diseases 7569 (6511–8741) 7656 (6583–8853) 1·2 (–0·1 to 2·4) Ischaemic heart disease 6308 (5918–6716) 7287 (6798–7808) 15·5 (14·3 to 16·8)* Maternal obstructed labour and uterine rupture 6830 (5202–8933) 6521 (4977–8588) –4·5 (–9·1 to –0·3)* Hepatitis C 4609 (4316 to 4918) 5394 (5032 to 5778) 17·0 (15·7 to 18·4)* Ischaemic stroke 4651 (4397 to 4912) 5385 (5017 to 5726) 15·8 (13·4 to 18·4)* Measles 15 610 (6699–31 357) 4652 (2072–9206) –70·2 (–73·6 to –66·0)* Paratyphoid fever 5580 (4498–6896) 4587 (3738–5601) –17·8 (–22·6 to –12·7)* Haemorrhagic stroke 3144 (2973–3307) 3583 (3336–3822) 14·0 (11·4 to 16·2)* Paralytic ileus and intestinal obstruction 2664 (2494–2855) 3167 (2946–3406) 18·9 (16·9 to 20·7)* Encephalitis 1489 (1380–1608) 1603 (1472–1750) 7·7 (5·9 to 9·5)* Acute glomerulonephritis 1528 (1395–1672) 1534 (1394–1685) 0·4 (–2·6 to 3·1) Data in parentheses are 95% UIs. Percentage changes that are statistically signiﬁ cant. www.thelancet.com Vol 388 October 8, 2016 Global causes of disability Global trends in YLDs 2005 to 2015 GBD 2015 included the assessment of 2619 sequelae at Level 6 of the GBD cause hierarchy, including 1316 sequelae from injuries that contributed to the global burden of disability. Causes at Level 4 of the hierarchy that resulted in 30 million or more YLDs in 2015 included lower back pain, major depressive disorder, age-related and other hearing loss, and neck pain. Figure 2 compares the leading causes of global YLDs in 2005 and 2015, using the cause breakdowns at Level 3 of the GBD cause hierarchy. Among the ten leading causes of YLDs, iron-deﬁ ciency anaemia and depressive disorders switched ranks to positions three and four respectively, diabetes rose from the eighth to the sixth position, migraine dropped from position six to seven, and other musculoskeletal disorders dropped from rank seven to eight (ﬁ gure 2). Age-standardised YLD rates attributable to hepatitis A, B, and C increased, and decreased for hepatitis E. The number of individuals with chronic hepatitis C infection increased from 121 million (108–133 million) in 2005 to 142 million (127–157 million) in 2015. Estimates of prevalence and YLDs at the global level for 2005 and 2015 for each cause are presented in table 3 (see results appendix pp 717–40 for full detail at the sequelae level). Prevalence and age-standardised YLDs for 21 Group 1 diseases decreased signiﬁ cantly and by more than 10%, including measles, African trypanosomiasis, diphtheria, lymphatic ﬁ lariasis, and rabies. Global incidence and prevalence Table 1: Global incidence of short duration (less than 3 months) sequelae in 2005 and 2015 for all ages and both sexes combined, with percentage change between 2005 and 2015 for level 4 causes with incidence greater than 1 million cases per year y y Anaemia was the most common of our nine impairments, aﬀ ecting 2·36 billion (2·35–2·37 billion) people in 2015. The next most common impairments were hearing loss of greater than 20 dB (1·33 billion [1·26–1·40 billion]), vision loss (940 million [905–974 million]), developmental intellectual disability (153 million [114–191 million]), infertility (113 million [93·4–136 million]), heart failure (40·0 million [38·6–41·4 million]), and epilepsy (39·2 million [34·3–43·7 million]; table 3; see results appendix (pp 740–48) for the prevalence estimates of the underlying causes of these impairments). Iron deﬁ ciency was the cause of anaemia in more than half of all cases. Over 90% of hearing loss was classiﬁ ed as age-related or other hearing loss. The largest number of people with vision loss had uncorrected refraction error. Idiopathic developmental intellectual disability, idiopathic female infertility, and idiopathic epilepsy were the most common causes of their impairments. Ischaemic heart disease was the most common cause of heart failure. Articles Depressive disorders were the fourth leading cause of disability in 2005 and the third leading cause of disability in 2015, and age-standardised YLD rates associated with the disorder increased marginally (1·0% [0·5–1·5%]). Age-standardised rates of YLDs from alcohol use disorders decreased after 2005 (4·5% [2·3–6·4%]) whereas disability due to drug use disorders increased by 8·2% (6·2–10·2%). In contrast with global trends for NCDs, both relative ranks and age-standardised YLD rates decreased for most injuries. Falls, which were the 13th leading cause of disability in 2005, dropped to the 15th rank, and age-standardised YLD rates decreased (8·58% [5·23–12·2%]). Other unintentional injuries decreased in global rank from 27th in 2005 to 34th in 2015, and age-standardiSed YLD rates decreased by 16·7% (15·7–17·9%). The leading causes of disability varied considerably with age (ﬁ gure 3). The leading cause in children younger than 5 years was iron-deﬁ ciency anaemia followed by skin diseases, protein-energy malnutrition, and diarrhoea. In older children, iron-deﬁ ciency anaemia, skin diseases, asthma, and mental health disorders such as conduct, autistic spectrum, and anxiety disorders were top ten causes of disability. In adolescents and young adults (aged between 15 and 39 years), iron-deﬁ ciency anaemia, skin diseases, depression, lower back and neck pain, and migraine led the rankings. Other mental health disorders such as anxiety disorders and schizophrenia were in the top ten causes in this age group. In middle-aged adults, musculoskeletal disorders dominated the top rankings followed by mental health disorders, especially depression. Diabetes and sense organ disorders were more prominent causes of disability in middle-age. In older adults (older than 65 years), sense organ disorders were the top-ranked cause of disability. Musculoskeletal disorders remained a dominant source of disability, and chronic obstructive pulmonary disease entered the top ten. In the oldest age groups, ischaemic heart disease and Alzheimer’s and other dementias made their ﬁ rst appearance in the top ten. pp p We examined the trends in YLDs and YLLs in a scatterplot (ﬁ gure 4). YLLs decreased for the majority of causes. For Group 1 causes and injuries, the decrease in YLLs was accompanied by a decrease in YLDs, albeit at a slower pace. The exceptions were neonatal encephalopathy, haemolytic disease and other neonatal jaundice, leishmaniasis, meningitis, hepatitis, and sexually trans- mitted diseases with increasing YLD rates between 1990 and 2015. Articles Prevalence (thousands) Percentage change (%) 2005 2015 Permanent caries 2 045 859 (1 909 845–2 170 355) 2 344 628 (2 193 751–2 488 741) 14·6 (13·7 to 15·5) Tension-type headache 1 306 390 (1 160 423–1 463 889) 1 505 892 (1 337 310–1 681 575) 15·3 (14·0 to 16·6)* Iron-deﬁ ciency anaemia 1 456 387 (1 449 846–1 462 782) 1 477 531 (1 470 902–1 485 322) 1·5 (0·9 to 2·1)* Age-related and other hearing loss 943 504 (886 325–995 403) 1 210 055 (1 140 224–1 274 665) 28·3 (27·4 to 29·2)* Migraine 831 726 (755 991–918 968) 958 789 (872 109–1 055 631) 15·3 (14·0 to 16·6)* Genital herpes 716 115 (626 987–817 387) 845 826 (736 724–968 066) 18·1 (16·4 to 19·9)* Refraction and accommodation disorders 672 165 (649 969–694 024) 819 307 (789 917–848 059) 21·9 (20·5 to 23·2)* Ascariasis 854 489 (790 000–924 895) 761 894 (682 558–861 031) –10·8 (–22·5 to 2·5) G6PD trait 663 704 (625 032–703 799) 728 549 (676 735–781 534) 9·8 (7·9 to 11·4)* Acne vulgaris 605 008 (568 577–642 061) 632 741 (595 242–671 249) 4·6 (3·5 to 5·6)* Other skin and subcutaneous diseases 492 883 (480 852–505 426) 605 036 (589 500–619 676) 22·8 (22·1 to 23·4)* Deciduous caries 534 122 (449 525–635 866) 558 028 (462 649–669 027) 4·5 (2·3 to 6·1)* Low back pain 460 164 (444 680–477 119) 539 907 (521 449–559 556) 17·3 (16·5 to 18·2)* Periodontal diseases 428 784 (372 953–498 682) 537 506 (465 114–625 889) 25·4 (24·1 to 26·5)* Fungal skin diseases 434 604 (395 512–475 904) 492 373 (448 951–538 232) 13·3 (12·5 to 14·1)* Trichuriasis 473 399 (443 689–505 928) 463 652 (426 621–502 939) –2·1 (–12·0 to 9·0) Diabetes 333 325 (310 773–355 510) 435 328 (404 736–468 562) 30·6 (28·0 to 33·0)* Premenstrual syndrome 391 207 (375 009–407 896) 430 697 (410 841–450 494) 10·1 (7·9 to 11·8)* Hookworm disease 462 111 (430 885–495 596) 428 246 (394 486–468 292) –7·3 (–16·9 to 3·5) Sickle-cell trait 338 756 (318 736–378 582) 404 566 (381 223–448 155) 19·4 (18·3 to 20·3)* Asthma 327 097 (296 406–358 060) 358 198 (323 134–393 466) 9·5 (7·6 to 11·6)* Neck pain 295 532 (258 878–338 138) 358 007 (313 408–409 411) 21·1 (19·0 to 23·3)* Hepatitis B 293 745 (284 478–303 036) 343 251 (330 541–357 195) 16·9 (15·3 to 18·4)* Other musculoskeletal disorders 283 317 (254 135–315 519) 342 068 (305 431–385 147) 20·7 (17·5 to 24·0)* Urolithiasis 259 567 (238 100–281 892) 318 763 (290 695 349 154) 22·8 (20·8 to 24·9)* Thalassaemias trait 252 798 (247 008–259 972) 279 451 (272 819–287 357) 10·5 (10·1 to 11·0)* Malaria 207 773 (186 530–230 942) 278 961 (240 158–320 921) 34·3 (26·8 to 42·3)* Edentulism and severe tooth loss 216 473 (207 563–226 331) 275 619 (264 201–288 252) 27·3 (26·9 to 27·7)* Anxiety disorders 232 597 (204 165–264 445) 267 202 (234 064–306 318) 14·9 (13·0 to 16·8)* (Table 2 continues on next page) However, age-standardised rates of YLDs increased for osteoarthritis 3·90% (3·00–4·83%) by 2015. www.thelancet.com Vol 388 October 8, 2016 Changes in age-standardised YLDs and YLLs over time g g In 2005, non-communicable diseases (NCDs) accounted for 23 of the leading 25 causes of age-standardised YLDs worldwide and 23 of the 25 leading causes in 2015 (ﬁ gure 2). Although diabetes rose only two ranks in the list of leading cause of YLDs, from position eight to six between 2005 and 2015, the increase in age-standardised rate was 5·4% (3·2–7·5%). Musculoskeletal disorders occupied three of the leading 25 causes of disability in both 2005 and 2015; lower back and neck pain were the single largest cause with little change in their rates. Age-standardised YLD rates for all maternal causes and sequelae combined decreased between 2005 and 2015, whereas overall age-standardised YLDs for neonatal disorders increased by 3·11% (–0·32–6·79%) since 2005 to 144 YLDs per 100 000 (109–185 YLDs per 100 000) in 2015, and age-standardised YLD rates decreased and absolute numbers of cases increased for nutritional deﬁ ciencies. 1551 Articles Articles Articles A few Group 1 disorders and injuries had a faster decrease in YLDs than in YLLs: intestinal infections, obstructed labour, and neonatal sepsis. The only NCDs with a faster decrease in YLDs compared with YLLs were epilepsy and cervical cancer. Another small number of NCDs saw an increase in YLDs and YLLs, including drug use disorders, diabetes, and Parkinson’s disease. Articles NCDs with decreasing YLLs but increasing YLDs included www.thelancet.com Vol 388 October 8, 2016 1552 Articles Articles Prevalence (thousands) Percentage change (%) 2005 2015 (Continued from previous page) Other sense organ diseases 214 761 (207 401–222 622) 266 346 (257 047–276 383) 24·0 (23·1 to 24·9)* Schistosomiasis 329 773 (297 093–367 878) 252 340 (211 032–321 081) –23·5 (–32·5 to –4·4)* G6PD deﬁ ciency 231 109 (205 067–259 769) 247 074 (209 307–286 713) 6·9 (1·4 to 11·9)* Dermatitis 215 260 (199 590–230 536) 245 291 (227 283–262 752) 14·0 (13·2 to 14·8)* Osteoarthritis 178 665 (173 558–184 053) 237 369 (230 336–244 648) 32·9 (31·9 to 33·8)* Major depressive disorder 183 434 (163 947–206 420) 216 047 (192 863–243 319) 17·8 (16·6 to 19·0)* Scabies 191 482 (166 101–223 739) 204 152 (177 534–237 466) 6·6 (4·0 to 9·5)* Viral skin diseases 161 167 (152 218–170 651) 174 843 (165 156–185 072) 8·5 (8·0 to 9·0)* Chronic obstructive pulmonary disease 149 115 (137 380–160 739) 174 483 (160 205–188 952) 17·0 (15·1 to 19·0)* Genital prolapse 137 383 (121 623–154 875) 161 679 (142 335–182 566) 17·7 (15·2 to 20·0)* Gastritis and duodenitis 135 993 (134 420–137 351) 157 060 (154 055–160 141) 15·5 (12·9 to 17·9)* Peripheral vascular disease 115 109 (101 439–131 405) 154 651 (136 318–176 211) 34·4 (33·4 to 35·2)* Uterine ﬁ broids 126 797 (120 900–133 050) 151 115 (144 147–158 477) 19·2 (18·8 to 19·5)* Hepatitis C 120 457 (108 080–133 129) 142 123 (126 978–157 045) 18·0 (16·7 to 19·2)* Other mental and substance use disorders 107 895 (107 213–108 449) 128 178 (127 512–128 877) 18·8 (17·9 to 19·7)* Iodine deﬁ ciency 103 701 (93 441–116 438) 110 920 (100 337–125 253) 7·0 (4·7 to 9·4)* Ischaemic heart disease 87 511 (80 133–96 170) 110 193 (100 332–121 427) 25·9 (24·6 to 27·2)* Benign prostatic hyperplasia 80 684 (70 338–90 853) 104 625 (90 730–118 244) 29·7 (27·5 to 32·0)* Dysthymia 86 812 (74 974–99 103) 104 106 (90 398–118 969) 19·9 (18·4 to 21·5)* Chronic kidney disease due to diabetes 79 184 (68 481–90 737) 100 824 (86 923–115 652) 27·3 (24·9 to 29·9)* Chronic kidney disease due to other causes 74 917 (64 012–86 576) 94 553 (81 142–109 371) 26·2 (24·2 to 28·3)* Idiopathic developmental intellectual disability 82 996 (47 588–117 109) 92 074 (52 280–130 411) 10·9 (9·5 to 11·9)* Other cardiovascular and circulatory diseases 72 772 (68 090–77 763) 90 348 (84 711–96 659) 24·2 (22·7 to 25·5)* Otitis media 83 022 (73 657–93 266) 86 393 (76 374–97 104) 4·1 (2·7% to 5·5)* Psoriasis 67 753 (65 107–70 298) 79 700 (76 691–82 804) 17·6 (17·0 to 18·3)* Other haemoglobinopathies and haemolytic anaemias 73 045 (72 600–73 470) 74 385 (73 898–74 862) 1·8 (0·9 to 2·7%) Data in parentheses are 95% UIs. www.thelancet.com Vol 388 October 8, 2016 Table 2: Global prevalence of longer duration (more than 3 months) sequelae in 2005 and 2015 for all ages and both sexes combined, with percentage change between 2005 and 2015 for Level 4 causes with prevalence greater than 1% Articles Percentage changes that are statistically signiﬁ cant. cancers of the prostate, testis, uterus, kidney, colorectum, and pancreas, melanoma, and congenital disorders. Some cancers (stomach and Hodgkin’s lymphoma), rheumatic heart disease, asthma, chronic obstructive pulmonary disease, acute glomerulonephritis, peptic ulcer disease, gastritis, hernia, and gallbladder disease had decreasing YLLs and YLDs, but with a faster decrease in YLLs than in YLDs. The rate of change in YLDs for the main drivers of non-fatal health loss, musculoskeletal disorders, and mental and substance use disorders was small. Global distribution of disability weights across individuals Figure 5 shows the global distribution of individuals from our comorbidity microsimulation by six categories of disability, age, and sex for the highest and lowest SDI quintile. The six categories of disability are no disability, very mild disability (disability weight less than or equal to 0·01), mild disability (from 0·01 to 0·05 inclusive), moderate (from 0·05 to 0·1 inclusive), severe (from 0·1 to 0·3 inclusive), and profound (greater than 0·3). In 2015, most of the world’s population experienced mild or greater disability. Having no disability at all was most common in children. After age 25 years, the proportion of the population having no disability became progressively smaller, and by age 55 years in low SDI countries and age 75 years in high SDI countries, nearly everyone had some form of disability. Very mild to moderate disability (ie, individuals with a disability weight of 0·1 or less) was common in childhood and young adults, but was replaced by more severe disability with increasing age. The patterns were similar for both sexes, apart from a much larger amount of disability in women older than 80 years in the top SDI quintiles, reﬂ ecting the much higher average age of women in this age category. For policy considerations around the age of retirement in ageing populations, it is noteworthy that from age 60 years onwards more than half of the population had severe or worse disability. The extent to which this loss of health limits or precludes the ability to work depends on the nature of the impairments and the type of employment in older workers with that level of disability. Data in parentheses are 95% UIs. Percentage changes that are statistically signiﬁ cant. Articles African countries; depression in ﬁ ve eastern sub-Saharan Africa countries; other neglected tropical diseases in Angola, Democratic Republic of the Congo, and Gabon; sense organ disorders in Comoros and Myanmar; diabetes in Fiji and Marshall Islands; war in Lebanon and Syria; and onchocerciasis in Liberia. age-standardised YLD rates. Anaemia led to generally higher rates of age-standardised YLDs in women than in men across levels of SDI, but the largest imbalance occurred at SDI levels between 0·10 and 0·50. Disability from injuries exacted a larger burden for men than for women, particularly at lower levels of SDI. p y Without adjustments for population age structure (ﬁ gure 6B), the eﬀ ect of ageing populations and causes of disability that disproportionately aﬀ ect older individuals become prominent. At all levels of SDI, total YLDs per 100 000 did not notably diﬀ er by sex; instead, the cause composition of disability showed greater diﬀ erences. Below an SDI score of 0·25, communicable causes accounted for 30–45% of total disability, primarily due to nutritional deﬁ ciencies, malaria, and neglected tropical diseases. YLDs per 100 000 due to musculoskeletal disorders, particularly lower back and neck pain and other mus- culo skeletal disorders, increased substantially from low to high SDI, with a more pronounced increase beginning at an SDI score of 0·6. This rise was particularly evident in women. Regional, country, territory, and selected subnational results Lower back and neck pain was the leading cause of disability in all high-income countries in 2015. However, ratios of observed to expected YLDs from lower back and neck pain ranged from 0·60 for Singapore to more than 1·59 in Norway. Most high-income countries experienced higher than expected levels of disability due to depressive disorders. The USA and Australia were the only two high- income countries where drug use disorders were a top ten cause of disability, and observed levels of YLDs were much higher than expected. South Korea’s ratio of observed-to- expected levels of YLDs due to diabetes exceeded 1·30, whereas Japan’s diabetes-related disability was lower than expected on the basis of SDI. In the UK, observed disability due to asthma was well above expected levels. In 2015, lower back and neck pain was the leading cause of YLDs for all but two countries in Latin America and the Caribbean; Haiti and Venezuela were the exceptions, with iron-deﬁ ciency anaemia as the leading causes of disability for both countries. Articles Disability due to diabetes surpassed expected levels by at least a factor of two for six countries and territories: Antigua, Barbados, Dominica, Puerto Rico, Trinidad and Tobago, and the Virgin Islands. Peru, however, had a ratio of less than 0·6 for observed and expected YLDs for diabetes. Trends in age-standardised YLDs per capita g p p Globally, age-standardised YLDs per capita (an indicator of overall disability experienced per person in a given place) moderately decreased for both sexes between 1990 and 2015 (results appendix pp 714–16). Age-standardised YLDs per capita were consistently higher for women than for men at the global level. For both sexes, YLDs per capita were generally higher for lower levels of SDI. YLDs per capita were noticeably larger for low SDI and low-middle SDI groups than for other SDI levels (ie, high SDI, high-middle SDI, and middle SDI), which were more similar to each other. p In 2015, lower back and neck pain was the leading cause of YLDs for 24 out of 28 countries and territories of southeast Asia, east Asia, and Oceania. Other leading causes of disability were sensory disorders in Myanmar, diabetes in Fiji and the Marshall Islands, and iron-deﬁ ciency anaemia in Papua New Guinea. Although lower back and neck pain was the primary cause of disability, many countries had far lower levels of this than expected given their SDI, including Thailand (0·73), Indonesia (0·76), and Malaysia (0·75). Observed YLDs due to depressive disorders were often lower than expected, with 22 countries recording ratios below 0·80. Conversely, numerous geographies recorded YLD ratios exceeding 2·00 for diabetes (eg, 2·29 in Taiwan). Leading causes of YLDs and deviations from expected levels based on Socio-demographic Index www.thelancet.com Vol 388 October 8, 2016 Expected changes in disease proﬁ le with higher Socio-demographic Index Figure 6 depicts changes in patterns of disability by level of SDI for age-standardised and all-age YLD rates per 100 000. Age-standardised YLD rates gradually decreased with increasing SDI in both sexes (ﬁ gure 6A). The cause composition of YLDs somewhat varied across levels of SDI; these diﬀ erences were largely derived from absolute levels of disability due to communicable causes and nutritional deﬁ ciencies, and to a lesser extent, maternal and neonatal disorders. Age-standardised YLD rates due to NCDs and injuries were similar at all SDI levels. Across levels of SDI, mental and substance use disorders, musculoskeletal disorders, and other NCDs were consistently among the leading causes of 1553 Articles Articles Leading causes of YLDs and deviations from expected levels based on Socio-demographic Index Clear, though varied, patterns emerged across and within GBD regions in comparison of observed levels of YLDs due to leading causes of disability with levels expected on the basis of SDI. Figure 7 displays ratios of observed and expected YLDs for the leading ten causes at level 3 of the GBD hierarchy in 2015, colour coded by the magnitude of diﬀ erences between observed and expected YLDs. Globally, lower back and neck pain was the leading cause of disability in 2015. Two mental disorders, major depressive and anxiety disorders, were the third and ninth leading causes of global disability, and diabetes was the sixth leading driver of disability. Iron-deﬁ ciency anaemia was the only Group 1 cause among the leading ten causes for global YLDs (ranked fourth). Sensory disorders ranked second and skin diseases ranked ﬁ fth. Lower back and neck pain was the leading global cause of disability in 2015 in most countries. The leading cause of disability in 2015 was iron-deﬁ ciency anaemia in 27 countries; HIV/AIDS in all six southern sub-Saharan Beyond lower back and neck pain, which was the leading cause of YLDs for three of ﬁ ve countries in south Asia in 2015 (with India and Pakistan being the exception), a mixture of causes accounted for the region’s main causes of disability; this heterogeneity probably reﬂ ects the diversity of countries in the region and their places along the development spectrum. Leading causes of YLDs and deviations from expected levels based on Socio-demographic Index Iron-deﬁ ciency anaemia was the ﬁ rst leading cause and lower back and neck pain was the second leading cause of YLDs in both India and Pakistan, whereas sensory disorders, other musculoskeletal disorders, and iron-deﬁ ciency anaemia ranked second for 1554 Articles Articles Prevalence (thousands) YLDs (thousands) 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 All causes ·· ·· ·· 792 004·7 (588 538·7–1 019 955·2) 15·1 (14·5 to 15·6)* –2·1 (–2·5 to –1·7)* Communicable, maternal, neonatal, and nutritional diseases 4 133 822·3 (4 108 132·5 to 4 161 262·6) 8·0 (7·3 to 8·7)* –4·2 (–4·8 to –3·6)* 112 501·9 (80 347·4–156 270·9) –0·5 (–2·4 to 2·5) –9·7 (–11·4 to –7·0)* HIV/AIDS and tuberculosis 46 005·2 (44 985·8 to 47 146·6) 18·6 (15·2 to 22·1)* 2·9 (–0·0 to 5·9) 6702·3 (4798·6–8863·5) 2·1 (–3·5 to 7·2) –12·2 (–17·0 to –7·8)* Tuberculosis 8861·2 (8076·3 to 9707·2) 9·2 (6·3 to 12·2)* –7·2 (–9·5 to –4·7)* 2712·4 (1829·9–3744·0) 9·4 (6·4 to 12·5)* –6·8 (–9·2 to –4·2)* HIV/AIDS 37 277·5 (36 279·9 to 38 339·1) 21·1 (16·9 to 25·5)* 5·6 (2·0 to 9·6)* 3989·9 (2891·2–5159·5) –2·4 (–10·5 to 5·5) –15·6 (–23·0 to –8·5)* HIV/AIDS—tuberculosis 1258·0 (1142·5 to 1386·9) –17·9 (–20·7 to –14·9)* –28·3 (–30·7 to –25·7)* 471·9 (316·0–645·6) –17·8 (–20·7 to –14·7)* –28·1 (–30·6 to –25·5)* HIV/AIDS resulting in other diseases 37 543·3 (36 350·0 to 39 028·4) 23·6 (20·5 to 26·7)* 7·9 (5·1 to 10·6)* 3518·0 (2 555·4–4 558·9) 0·2 (–9·3 to 9·3) –13·6 (–22·1 to –5·2)* HIV aggregate 21 149·0 (20 260·3 to 22 292·3) –19·5 (–22·5 to –16·3)* –29·7 (–32·3 to –26·9)* 1543·6 (1 107·2–2 025·2) –20·4 (–23·4 to –17·2)* –30·4 (–33·1 to –27·6)* AIDS aggregate 16 394·3 (15 747·0 to 17 154·2) 299·5 (251·4 to 355·5)* 232·9 (193·5 to 279·0)* 1974·4 (1432·0–2559·9) 25·5 (0·9 to 56·1)* 6·1 (–14·6 to 32·4) Diarrhoea, lower respiratory infections, and other common infectious diseases 402 320·4 (393 746·5 to 408 994·2) 8·0 (7·2 to 8·8)* –2·9 (–3·7 to –2·3)* 14 865·0 (10 397·0–20 283·4) 6·2 (5·4 to 7·1)* –4·0 (–4·7 to –3·3)* Diarrhoeal diseases 35 820·6 (34 342·1 to 37 537·8) 6·4 (5·5 to 7·3)* –3·3 (–4·1 to –2·5)* 5731·7 (3 943·3–7 890·5) 6·4 (5·4 to 7·5)* –3·1 (–4·0 to –2·2)* Diarrhoea episodes aggregate 35 816·3 (34 337·5 to 37 534·1) 6·4 (5·5 to 7·3)* –3·3 (–4·1 to –2·5)* 5730·4 (3 942·2–7 889·0) 6·4 (5·4 to 7·5)* –3·1 (–4·0 to –2·2)* Guillain-Barré syndrome due to diarrhoeal diseases 4·2 (3·2 to 5·5) 16·9 (14·7 to 19·5)* –0·2 (–0·9 to 0·4) 1·3 (0·8–1·9) 16·9 (14·7 to 19·5)* –0·2 (–0·9 to 0·4) Intestinal infectious diseases 1666·8 (1594·4 to 1 732·8) –23·3 (–25·6 to –20·6)* –28·2 (–30·4 to –25·8)* 220·3 (149·3–306·5) –18·6 (–22·8 to –13·8)* –24·2 (–28·1 to –19·8)* Typhoid fever 1446·7 (1256·2 to 1 648·1) –18·3 (–20·8 to –15·6)* –23·9 (–26·3 to –21·3)* 191·0 (129·5–267·8) –17·7 (–22·1 to –12·6)* –23·2 (–27·4 to –18·5)* Typhoid fever episodes aggregate 1197·9 (1041·3 to 1366·6) –18·3 (–21·7 to –14·8)* –23·9 (–27·1 to –20·5)* 113·3 (75·8–162·5) –18·1 (–22·6 to –13·4)* –23·7 (–27·8 to –19·2)* Complications of typhoid fever aggregate 248·8 (210·8 to 292·2) –18·2 (–28·6 to –5·4)* –23·8 (–33·5 to –12·1)* 77·8 (51·6–110·8) –17·0 (–28·5 to –3·0)* –22·6 (–33·1 to –9·9)* Paratyphoid fever 529·3 (431·3 to 646·3) –17·8 (–22·6 to –12·7)* –23·8 (–28·2 to –19·2)* 27·5 (17·5–40·7) –17·5 (–24·7 to –9·2)* –23·5 (–30·1 to –16·1)* Paratyphoid fever episodes aggregate 502·2 (409·6 to 612·7) –17·8 (–22·6 to –12·4)* –23·8 (–28·2 to –19·0)* 24·4 (15·5–36·4) –17·5 (–24·9 to –8·7)* –23·5 (–30·7 to –15·5)* Intestinal perforation due to paratyphoid 27·1 (20·9 to 34·5) –17·6 (–33·0 to 1·4) –23·7 (–37·8 to –6·1)* 3·1 (1·9–4·6) –17·5 (–32·8 to 1·6) –23·6 (–37·7 to –5·9)* Other intestinal infectious diseases ·· ·· ·· 1·8 (0·6–4·1) –67·0 (–78·8 to –45·8)* –69·3 (–80·2 to –50·1)* Lower respiratory infections 8986·6 (8545·2 to 9393·7) 5·1 (4·0 to 6·2)* –8·2 (–9·1 to –7·3)* 540·4 (365·3–760·3) 4·9 (3·5 to 6·3)* –8·1 (–9·2 to –7·0)* Lower respiratory infection episodes aggregate 8982·2 (8539·8 to 9387·9) 5·1 (4·0 to 6·2)* –8·2 (–9·1 to –7·3)* 539·1 (364·5–759·5) 4·9 (3·5 to 6·3)* –8·1 (–9·2 to –7·0)* Guillain-Barré syndrome due to lower respiratory infections 4·4 (2·5 to 6·9) 16·9 (14·7 to 19·5)* –0·2 (–0·9 to 0·4) 1·3 (0·7–2·2) 16·9 (14·7 to 19·5)* –0·2 (–0·9 to 0·4) Upper respiratory infections 233 470·3 (208 009·7 to 259 076·8) 10·2 (8·9 to 11·5)* –1·4 (–2·0 to –0·8)* 2738·4 (1538·6–4644·3) 10·1 (8·9 to 11·5)* –1·3 (–1·9 to –0·6)* Upper respiratory infection episodes aggregate 233 458·2 (207 997·9 to 259 065·2) 10·2 (8·9 to 11·5)* –1·4 (–2·0 to –0·8)* 2734·8 (1 533·9–4 641·4) 10·1 (8·9 to 11·5)* –1·3 (–1·9 to –0·6)* Guillain-Barré syndrome due to upper respiratory infections 12·1 (9·6 to 15·2) 16·9 (14·7 to 19·5)* –0·2 (–0·9 to 0·4) 3·6 (2·2–5·4) 17·0 (14·7 to 19·5)* –0·2 (–0·9 to 0·4) (Table 3 continues on next page) Diarrhoea, lower respiratory infections, and other common infectious diseases 402 320·4 (393 746·5 to 408 994·2) Diarrhoeal diseases 35 820·6 (34 342·1 to 37 537·8) Diarrhoea episodes aggregate 35 816·3 (34 337·5 to 37 534·1) Guillain-Barré syndrome due to diarrhoeal diseases 4·2 (3·2 to 5·5) Intestinal infectious diseases 1666·8 (1594·4 to 1 732·8) Typhoid fever 1446·7 (1256·2 to 1 648·1) Typhoid fever episodes aggregate 1197·9 (1041·3 to 1366·6) Complications of typhoid fever aggregate 248·8 (210·8 to 292·2) Paratyphoid fever 529·3 (431·3 to 646·3) Paratyphoid fever episodes aggregate 502·2 (409·6 to 612·7) Intestinal perforation due to paratyphoid 27·1 (20·9 to 34·5) Other intestinal infectious diseases ·· Lower respiratory infections 8986·6 (8545·2 to 9393·7) Lower respiratory infection episodes aggregate 8982·2 (8539·8 to 9387·9) Guillain-Barré syndrome due to lower respiratory infections 4·4 (2·5 to 6·9) Upper respiratory infections 233 470·3 (208 009·7 to 259 076·8 Upper respiratory infection episodes aggregate 233 458·2 (207 997·9 to 259 065·2 Guillain-Barré syndrome due to upper respiratory infections 12·1 (9·6 to 15·2) –0·2 (–0·9 to 0·4) (Table 3 continues on next page) www.thelancet.com Vol 388 October 8, 2016 1555 Articles Articles Articles Prevalence (thousands) YLDs (thousands) 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 (Continued from previous page) Otitis media 112 089·1 (100 504·6 to 123 655·6) 5·2 (4·1 to 6·5)* –4·4 (–5·5 to –3·4)* 3321·5 (2 114·2–4 906·3) 3·3 (0·6 to 6·0)* –5·6 (–8·0 to –3·1)* Acute otitis media 25 630·8 (21 044·2 to 31 570·6) 9·4 (7·5 to 11·2)* 1·3 (–0·4 to 3·0) 334·9 (169·4–623·7) 9·4 (7·5 to 11·4)* 1·4 (–0·3 to 3·1) Chronic otitis media aggregate 86 458·3 (76 403·0 to 97 133·7) 4·1 (2·7 to 5·5)* –6·0 (–7·2 to –4·8)* 2986·6 (1904·2–4 371·1) 2·7 (–0·2 to 5·6) –6·3 (–8·9 to –3·6)* Meningitis 8733·5 (8320·6 to 9107·1) 16·9 (13·4 to 20·0)* 2·5 (–0·6 to 5·1) 1516·9 (1076·1–1986·0) 15·0 (13·7 to 16·5)* 1·9 (0·7 to 3·0)* Pneumococcal meningitis 7297·2 (6228·5 to 8635·3) 22·2 (19·7 to 25·1)* 6·9 (4·7 to 9·6)* 728·8 (513·1–959·9) 17·6 (15·5 to 19·7)* 3·9 (2·0 to 5·6)* Acute pneumococcal meningitis 40·3 (33·5 to 48·4) 25·3 (17·4 to 33·8)* 12·1 (5·2 to 19·5)* 5·3 (3·4–7·8) 25·1 (15·1 to 37·1)* 11·9 (3·2 to 22·3)* Complications of pneumococcal meningitis aggregate 7256·9 (6191·3 to 8591·7) 22·2 (19·7 to 25·1)* 6·9 (4·7 to 9·6)* 723·5 (509·2–952·4) 17·6 (15·4 to 19·6)* 3·8 (2·0 to 5·6)* H inﬂ uenzae type b meningitis 2455·6 (1966·7 to 3032·6) –8·3 (–12·2 to –3·9)* –18·0 (–21·5 to –13·9)* 302·1 (209·4–407·1) 6·9 (3·4 to 10·6)* –3·3 (–6·6 to 0·0) Acute H inﬂ uenzae type b meningitis 22·0 (17·2 to 28·7) –19·2 (–25·3 to –11·9)* –26·0 (–31·8 to –19·2)* 2·9 (1·8–4·5) –18·7 (–26·1 to –10·8)* –25·6 (–32·4 to –18·1)* Complications of H inﬂ uenzae type b meningitis aggregate 2433·5 (1947·4 to 3010·8) –8·2 (–12·1 to –3·7)* –17·9 (–21·5 to –13·7)* 299·1 (207·9–402·7) 7·3 (3·7 to 11·0)* –3·0 (–6·3 to 0·3) Meningococcal meningitis 1720·8 (1327·7 to 2 180·5) 22·0 (17·9 to 26·2)* 7·5 (3·8 to 11·3)* 161·6 (111·7–215·6) 19·7 (17·5 to 22·1)* 5·1 (3·2 to 7·4)* Acute meningococcal meningitis 22·8 (18·9 to 27·7) 24·4 (14·4 to 34·9)* 12·7 (3·7 to 22·4)* 3·0 (1·8–4·5) 24·7 (10·9 to 39·4)* 12·9 (0·5 to 25·9)* Complications of meningococcal meningitis aggregate 1698·0 (1305·2 to 2155·2) 22·0 (17·9 to 26·2)* 7·4 (3·8 to 11·2)* 158·6 (109·4–211·9) 19·6 (17·4 to 22·0)* 5·0 (3·1 to 7·3)* Other meningitis 3015·2 (2438·5 to 3618·0) 19·0 (16·6 to 21·6)* 3·5 (1·4 to 5·9)* 324·5 (225·1–428·1) 15·3 (13·3 to 17·4)* 0·9 (–0·8 to 2·7) Other meningitis episodes aggregate 134·5 (120·0 to 151·7) 16·3 (10·2 to 22·5)* 5·7 (0·1 to 11·4)* 17·8 (11·5–25·3) 16·6 (8·6 to 24·0)* 5·9 (–1·1 to 12·5) Complications of other meningitis aggregate 2880·4 (2300·7 to 3474·9) 19·1 (16·5 to 21·9)* 3·4 (1·2 to 5·8)* 306·7 (213·3–406·6) 15·2 (13·2 to 17·5)* 0·6 (–1·1 to 2·6) Encephalitis 4315·8 (3145·8 to 5875·6) 4·8 (2·8 to 7·4)* –9·3 (–11·3 to –6·7)* 457·0 (326·7–594·9) 6·7 (4·6 to 9·0)* –6·3 (–8·2 to –4·3)* Acute encephalitis 81·8 (74·8 to 89·5) 7·1 (5·2 to 9·0)* –4·1 (–5·7 to –2·4)* 10·8 (7·1–15·5) 7·5 (3·8 to 11·1)* –3·7 (–6·9 to –0·5)* Complications of encephalitis aggregate 4233·6 (3069·6 to 5787·5) 4·8 (2·7 to 7·3)* –9·4 (–11·4 to –6·8)* 446·2 (318·8–581·4) 6·7 (4·6 to 9·0)* –6·4 (–8·3 to –4·4)* Diphtheria 0·6 (0·3 to 1·2) –59·6 (–81·7 to –12·7)* –62·8 (–83·1 to –19·5)* 0·0 (0·0–0·1) –59·7 (–81·7 to –12·7)* –62·8 (–83·1 to –19·5)* Whooping cough 2232·6 (1725·9 to 2800·7) –27·4 (–29·5 to –25·2)* –32·4 (–34·3 to –30·3)* 110·3 (65·6–167·0) –27·3 (–29·6 to –24·8)* –32·3 (–34·5 to –30·0)* Tetanus 209·3 (204·8 to 214·7) 8·4 (6·1 to 9·8)* –2·7 (–5·0 to –1·4)* 9·1 (6·7–12·2) –8·9 (–14·5 to –3·7)* –18·1 (–23·3 to –13·3)* Severe tetanus 8·6 (6·5 to 13·0) –46·3 (–54·0 to –36·0)* –52·1 (–59·1 to –42·9)* 1·1 (0·7–1·9) –45·9 (–53·8 to –35·7)* –51·8 (–58·9 to –42·6)* Complications of tetanus aggregate 200·7 (196·7 to 204·5) 13·3 (12·8 to 13·9)* 1·8 (1·4 to 2·3)* 8·0 (5·8–10·6) 1·0 (–3·7 to 5·4) –8·8 (–13·0 to –4·9)* Measles 127·4 (56·8 to 252·2) –70·2 (–73·6 to –66·0)* –72·3 (–75·4 to –68·4)* 11·5 (4·3–25·3) –70·0 (–73·6 to –65·4)* –72·0 (–75·4 to –67·7)* Varicella and herpes zoster 5907·7 (5489·4 to 6344·6) 15·2 (14·1 to 16·5)* –0·4 (–0·8 to –0·0)* 207·8 (127·9–316·5) 18·5 (16·5 to 20·8)* –0·4 (–1·4 to 0·7) (Table 3 continues on next page Prevalence (thousands) Prevalence (thousand 2015 Continued from previous page) Otitis media 112 089·1 (100 504·6 to 123 655· Acute otitis media 25 630·8 (21 044·2 to 31 570·6) Chronic otitis media aggregate 86 458·3 (76 403·0 to 97 133·7) Meningitis 8733·5 (8320·6 to 9107·1) Pneumococcal meningitis 7297·2 (6228·5 to 8635·3) Acute pneumococcal meningitis 40·3 (33·5 to 48·4) Complications of pneumococcal meningitis aggregate 7256·9 (6191·3 to 8591·7) H inﬂ uenzae type b meningitis 2455·6 (1966·7 to 3032·6) Acute H inﬂ uenzae type b meningitis 22·0 (17·2 to 28·7) Complications of H inﬂ uenzae type b meningitis aggregate 2433·5 (1947·4 to 3010·8) Meningococcal meningitis 1720·8 (1327·7 to 2 180·5) Acute meningococcal meningitis 22·8 (18·9 to 27·7) Complications of meningococcal meningitis aggregate 1698·0 (1305·2 to 2155·2) Other meningitis 3015·2 (2438·5 to 3618·0) Other meningitis episodes aggregate 134·5 (120·0 to 151·7) Complications of other meningitis aggregate 2880·4 (2300·7 to 3474·9) Encephalitis 4315·8 (3145·8 to 5875·6) Acute encephalitis 81·8 (74·8 to 89·5) Complications of encephalitis aggregate 4233·6 (3069·6 to 5787·5) Diphtheria 0·6 (0·3 to 1·2) Whooping cough 2232·6 (1725·9 to 2800·7) Tetanus 209·3 (204·8 to 214·7) Severe tetanus 8·6 (6·5 to 13·0) Complications of tetanus aggregate 200·7 (196·7 to 204·5) Measles 127·4 (56·8 to 252·2) Varicella and herpes zoster 5907·7 (5489·4 to 6344·6) 1556 www.thelancet.com Vol 388 October 8, 2016 Articles Prevalence (thousands) YLDs (thousands) 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 (Continued from previous page) Neglected tropical diseases and malaria 1 900 062·6 (1 863 148·8 to 1 940 058·5) –0·4 (–2·5 to 1·9) –10·7 (–12·7 to –8·7)* 20 763·1 (13 382·4–32 174·5) –4·7 (–13·2 to 8·9) –14·5 (–22·2 to –2·4)* Malaria 295 717·3 (257 568·4 to 338 449·0) 29·9 (22·5 to 37·2)* 20·0 (12·9 to 27·2)* 3358·2 (2356·9–4703·9) 16·1 (12·8 to 19·3)* 8·5 (5·4 to 11·4)* Asymptomatic malaria parasitaemia (PfPR) 206 147·2 (168 670·7 to 246 888·7) 37·3 (27·0 to 48·1)* 26·4 (16·6 to 36·8)* 0·0 (0·0–0·0) 0·0 (0·0 to 0·0) 0·0 (0·0 to 0·0) Malaria episodes aggregate 16 756·8 (10 341·8 to 25 965·5) –15·9 (–23·6 to –8·4)* –21·1 (–28·4 to –14·1)* 346·6 (174·4–598·4) –15·8 (–23·6 to –8·2)* –20·9 (–28·2 to –13·7)* Complications of malaria aggregate 876·9 (798·1 to 957·9) 28·4 (26·5 to 30·5)* 18·4 (16·5 to 20·3)* 274·9 (209·7–344·8) 24·7 (20·5 to 29·1)* 15·0 (11·2 to 18·9)* Anaemia due to malaria parasitaemia (PfPR) aggregate 71 936·4 (70 137·5 to 73 504·2) 26·3 (24·4 to 28·0)* 17·0 (15·3 to 18·6)* 2736·7 (1848·0–3918·8) 21·1 (18·1 to 24·3)* 13·1 (10·5 to 16·1)* Chagas disease 6653·6 (5750·5 to 7575·6) –7·1 (–9·7 to –4·1)* –22·1 (–24·2 to –19·7)* 63·1 (42·1–90·8) –1·8 (–5·1 to 1·7) –21·0 (–23·6 to –18·3)* Chagas disease episodes aggregate 5634·7 (4850·5 to 6414·1) –7·8 (–10·3 to –4·9)* –22·2 (–24·3 to –19·8)* 0·0 (0·0–0·1) –18·7 (–22·9 to –14·4)* –27·6 (–31·1 to –24·1)* Complications of Chagas disease aggregate 666·0 (525·5 to 820·6) –5·2 (–7·9 to –1·9)* –22·0 (–24·3 to –19·6)* 44·8 (30·0–62·8) –2·5 (–6·0 to 1·2) –21·0 (–23·6 to –18·3)* Heart failure due to Chagas disease aggregate 353·0 (221·5 to 503·5) 0·3 (–2·9 to 4·0) –20·9 (–23·4 to –18·2)* 18·2 (9·7–29·4) 0·2 (–3·6 to 4·4) –20·9 (–23·8 to –17·8)* Leishmaniasis 3859·3 (3438·4 to 4570·4) 27·3 (5·9 to 55·0)* 11·4 (–7·0 to 35·1) 45·8 (22·8–86·7) 25·5 (21·9 to 28·9)* 10·5 (7·5 to 13·5)* Visceral leishmaniasis 60·8 (57·5 to 64·7) 10·6 (9·7 to 11·4)* 2·3 (1·6 to 3·0)* 4·3 (2·9–6·1) 10·8 (2·1 to 20·7)* 2·6 (–5·4 to 11·5) Cutaneous and mucocutaneous leishmaniasis 3895·9 (3 324·6 to 4 767·5) 27·0 (23·7 to 30·1)* 11·0 (8·3 to 13·8)* 41·5 (19·4–80·8) 27·3 (23·6 to 30·5)* 11·4 (8·3 to 14·5)* African trypanosomiasis 10·7 (6·0 to 17·0) –68·7 (–70·4 to –67·3)* –72·3 (–73·8 to –71·0)* 3·0 (1·4–5·3) –67·4 (–72·6 to –62·1)* –71·0 (–75·5 to –66·6)* Schistosomiasis 252 339·5 (211 032·5 to 321 081·3) –23·5 (–32·5 to –4·4)* –30·9 (–39·0 to –13·7)* 2472·6 (1275·0–4521·2) –21·8 (–29·3 to –4·0)* –29·1 (–36·0 to –13·1)* Schistosomiasis episodes aggregate 130 285·7 (114 433·9 to 156 900·0) –26·3 (–34·5 to –13·6)* –33·1 (–40·5 to –21·3)* 740·7 (292·3–1584·0) –27·3 (–35·2 to –14·1)* –34·0 (–41·0 to –22·0)* Complications of schistosomiasis aggregate 105 621·9 (76 641·0 to 160 339·4) –19·6 (–32·6 to 14·1) –27·7 (–39·4 to 3·0) 1188·8 (535·9–2 463·3) –18·0 (–30·5 to 15·6) –26·3 (–37·5 to 4·3) Anaemia due to schistosomiasis aggregate 16 331·3 (16 075·8 to 16 655·1) –19·6 (–21·1 to –17·7)* –27·0 (–28·3 to –25·3)* 543·1 (365·1–781·5) –21·5 (–24·0 to –18·6)* –27·8 (–30·0 to –25·1)* Cysticercosis 1931·0 (1597·8 to 2312·0) –6·2 (–10·2 to –2·5)* –20·8 (–24·3 to –17·6)* 286·7 (194·2–392·8) –16·3 (–21·3 to –11·8)* –29·2 (–33·3 to –25·3)* Cystic echinococcosis 1383·0 (1265·9 to 1498·6) 24·7 (22·6 to 27·0)* 6·7 (5·0 to 8·5)* 126·8 (86·7–174·6) 24·3 (21·1 to 27·5)* 6·6 (4·1 to 9·2)* Lymphatic ﬁ lariasis 38 464·1 (31 328·2 to 46 783·0) –44·9 (–49·9 to –39·9)* –51·6 (–56·0 to –47·2)* 2075·0 (1120·6–3311·2) –16·2 (–32·1 to –3·9)* –27·7 (–41·5 to –17·1)* Prevalence of detectable microﬁ liaria due to lymphatic ﬁ lariasis 19 707·9 (17 173·8 to 22 270·5) –58·8 (–61·8 to –55·3)* –63·5 (–66·1 to –60·3)* 0·0 (0·0–0·0) 0·0 (0·0 to 0·0) 0·0 (0·0 to 0·0) Complications of lymphatic ﬁ lariasis aggregate 18 756·3 (12 420·1 to 26 397·1) –14·8 (–29·9 to –3·5)* –26·7 (–40·1 to –16·7)* 2075·0 (1120·6–3311·2) –16·2 (–32·1 to –3·9)* –27·7 (–41·5 to –17·1)* Onchocerciasis 15 531·5 (11 963·5 to 19 993·8) –29·1 (–39·5 to –18·8)* –36·8 (–46·0 to –27·5)* 1135·7 (546·2–2005·4) –21·2 (–38·5 to –4·8)* –31·2 (–46·7 to –16·2)* Asymptomatic onchocerciasis 2280·4 (1525·0 to 3173·2) –62·4 (–66·1 to –59·6)* –66·2 (–69·5 to –63·7)* 0·0 (0·0–0·0) 0·0 (0·0 to 0·0) 0·0 (0·0 to 0·0) (Table 3 continues on next page) Prevalence (thousands) 2015 Percentage change counts between 20 and 2015 (Continued from previous page) Neglected tropical diseases and malaria 1 900 062·6 (1 863 148·8 to 1 940 058·5) –0·4 (–2·5 to 1·9) Malaria 295 717·3 (257 568·4 to 338 449·0) 29·9 (22·5 to 37·2)* Asymptomatic malaria parasitaemia (PfPR) 206 147·2 (168 670·7 to 246 888·7) 37·3 (27·0 to 48·1)* Malaria episodes aggregate 16 756·8 (10 341·8 to 25 965·5) –15·9 (–23·6 to –8·4)* Complications of malaria aggregate 876·9 (798·1 to 957·9) 28·4 (26·5 to 30·5)* Anaemia due to malaria parasitaemia (PfPR) aggregate 71 936·4 (70 137·5 to 73 504·2) 26·3 (24·4 to 28·0)* Chagas disease 6653·6 (5750·5 to 7575·6) –7·1 (–9·7 to –4·1)* Chagas disease episodes aggregate 5634·7 (4850·5 to 6414·1) –7·8 (–10·3 to –4·9)* Complications of Chagas disease aggregate 666·0 (525·5 to 820·6) –5·2 (–7·9 to –1·9)* Heart failure due to Chagas disease aggregate 353·0 (221·5 to 503·5) 0·3 (–2·9 to 4·0) Leishmaniasis 3859·3 (3438·4 to 4570·4) 27·3 (5·9 to 55·0)* Visceral leishmaniasis 60·8 (57·5 to 64·7) 10·6 (9·7 to 11·4)* Cutaneous and mucocutaneous leishmaniasis 3895·9 (3 324·6 to 4 767·5) 27·0 (23·7 to 30·1)* African trypanosomiasis 10·7 (6·0 to 17·0) –68·7 (–70·4 to –67·3)* Schistosomiasis 252 339·5 (211 032·5 to 321 081·3) –23·5 (–32·5 to –4·4)* Schistosomiasis episodes aggregate 130 285·7 (114 433·9 to 156 900·0) –26·3 (–34·5 to –13·6)* Complications of schistosomiasis aggregate 105 621·9 (76 641·0 to 160 339·4) –19·6 (–32·6 to 14·1) Anaemia due to schistosomiasis aggregate 16 331·3 (16 075·8 to 16 655·1) –19·6 (–21·1 to –17·7)* Cysticercosis 1931·0 (1597·8 to 2312·0) –6·2 (–10·2 to –2·5)* Cystic echinococcosis 1383·0 (1265·9 to 1498·6) 24·7 (22·6 to 27·0)* Lymphatic ﬁ lariasis 38 464·1 (31 328·2 to 46 783·0) –44·9 (–49·9 to –39·9)* Prevalence of detectable microﬁ liaria due to lymphatic ﬁ lariasis 19 707·9 (17 173·8 to 22 270·5) –58·8 (–61·8 to –55·3)* Complications of lymphatic ﬁ lariasis aggregate 18 756·3 (12 420·1 to 26 397·1) –14·8 (–29·9 to –3·5)* Onchocerciasis 15 531·5 (11 963·5 to 19 993·8) –29·1 (–39·5 to –18·8)* Asymptomatic onchocerciasis 2280·4 (1525·0 to 3173·2) –62·4 (–66·1 to –59·6)* Prevalence (thousands) Y 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 20 (Continued from previous page) Neglected tropical diseases and malaria 1 900 062·6 (1 863 148·8 to 1 940 058·5) –0·4 (–2·5 to 1·9) –10·7 (–12·7 to –8·7)* (1 Malaria 295 717·3 (257 568·4 to 338 449·0) 29·9 (22·5 to 37·2)* 20·0 (12·9 to 27·2)* (2 Asymptomatic malaria parasitaemia (PfPR) 206 147·2 (168 670·7 to 246 888·7) 37·3 (27·0 to 48·1)* 26·4 (16·6 to 36·8)* (0 Malaria episodes aggregate 16 756·8 (10 341·8 to 25 965·5) –15·9 (–23·6 to –8·4)* –21·1 (–28·4 to –14·1)* (1 Complications of malaria aggregate 876·9 (798·1 to 957·9) 28·4 (26·5 to 30·5)* 18·4 (16·5 to 20·3)* (2 Anaemia due to malaria parasitaemia (PfPR) aggregate 71 936·4 (70 137·5 to 73 504·2) 26·3 (24·4 to 28·0)* 17·0 (15·3 to 18·6)* (1 Chagas disease 6653·6 (5750·5 to 7575·6) –7·1 (–9·7 to –4·1)* –22·1 (–24·2 to –19·7)* (4 Chagas disease episodes aggregate 5634·7 (4850·5 to 6414·1) –7·8 (–10·3 to –4·9)* –22·2 (–24·3 to –19·8)* (0 Complications of Chagas disease aggregate 666·0 (525·5 to 820·6) –5·2 (–7·9 to –1·9)* –22·0 (–24·3 to –19·6)* (3 Heart failure due to Chagas disease aggregate 353·0 (221·5 to 503·5) 0·3 (–2·9 to 4·0) –20·9 (–23·4 to –18·2)* (9 Leishmaniasis 3859·3 (3438·4 to 4570·4) 27·3 (5·9 to 55·0)* 11·4 (–7·0 to 35·1) (2 Visceral leishmaniasis 60·8 (57·5 to 64·7) 10·6 (9·7 to 11·4)* 2·3 (1·6 to 3·0)* (2 Cutaneous and mucocutaneous leishmaniasis 3895·9 (3 324·6 to 4 767·5) 27·0 (23·7 to 30·1)* 11·0 (8·3 to 13·8)* (1 African trypanosomiasis 10·7 (6·0 to 17·0) –68·7 (–70·4 to –67·3)* –72·3 (–73·8 to –71·0)* (1 Schistosomiasis 252 339·5 (211 032·5 to 321 081·3) –23·5 (–32·5 to –4·4)* –30·9 (–39·0 to –13·7)* (1 Schistosomiasis episodes aggregate 130 285·7 (114 433·9 to 156 900·0) –26·3 (–34·5 to –13·6)* –33·1 (–40·5 to –21·3)* (2 Complications of schistosomiasis aggregate 105 621·9 (76 641·0 to 160 339·4) –19·6 (–32·6 to 14·1) –27·7 (–39·4 to 3·0) (5 Anaemia due to schistosomiasis aggregate 16 331·3 (16 075·8 to 16 655·1) –19·6 (–21·1 to –17·7)* –27·0 (–28·3 to –25·3)* (3 Cysticercosis 1931·0 (1597·8 to 2312·0) –6·2 (–10·2 to –2·5)* –20·8 (–24·3 to –17·6)* (1 Cystic echinococcosis 1383·0 (1265·9 to 1498·6) 24·7 (22·6 to 27·0)* 6·7 (5·0 to 8·5)* (8 Lymphatic ﬁ lariasis 38 464·1 (31 328·2 to 46 783·0) –44·9 (–49·9 to –39·9)* –51·6 (–56·0 to –47·2)* (1 Prevalence of detectable microﬁ liaria due to lymphatic ﬁ lariasis 19 707·9 (17 173·8 to 22 270·5) –58·8 (–61·8 to –55·3)* –63·5 (–66·1 to –60·3)* (0 Complications of lymphatic ﬁ lariasis aggregate 18 756·3 (12 420·1 to 26 397·1) –14·8 (–29·9 to –3·5)* –26·7 (–40·1 to –16·7)* (1 Onchocerciasis 15 531·5 (11 963·5 to 19 993·8) –29·1 (–39·5 to –18·8)* –36·8 (–46·0 to –27·5)* (5 Asymptomatic onchocerciasis 2280·4 (1525·0 to 3173·2) –62·4 (–66·1 to –59·6)* –66·2 (–69·5 to –63·7)* (0 (Continued from previous page) Neglected tropical diseases and malaria 1 900 062·6 (1 863 148·8 to 1 940 058·5) Malaria 295 717·3 (257 568·4 to 338 449·0 Asymptomatic malaria parasitaemia (PfPR) 206 147·2 (168 670·7 to 246 888·7 Malaria episodes aggregate 16 756·8 (10 341·8 to 25 965·5) Complications of malaria aggregate 876·9 (798·1 to 957·9) Anaemia due to malaria parasitaemia (PfPR) aggregate 71 936·4 (70 137·5 to 73 504·2) Chagas disease 6653·6 (5750·5 to 7575·6) Chagas disease episodes aggregate 5634·7 (4850·5 to 6414·1) Complications of Chagas disease aggregate 666·0 (525·5 to 820·6) Heart failure due to Chagas disease aggregate 353·0 (221·5 to 503·5) Leishmaniasis 3859·3 (3438·4 to 4570·4) Visceral leishmaniasis 60·8 (57·5 to 64·7) Cutaneous and mucocutaneous leishmaniasis 3895·9 (3 324·6 to 4 767·5) African trypanosomiasis 10·7 (6·0 to 17·0) Schistosomiasis 252 339·5 (211 032·5 to 321 081·3) Schistosomiasis episodes aggregate 130 285·7 (114 433·9 to 156 900·0) Complications of schistosomiasis aggregate 105 621·9 (76 641·0 to 160 339·4) Anaemia due to schistosomiasis aggregate 16 331·3 (16 075·8 to 16 655·1) Cysticercosis 1931·0 (1597·8 to 2312·0) Cystic echinococcosis 1383·0 (1265·9 to 1498·6) Lymphatic ﬁ lariasis 38 464·1 (31 328·2 to 46 783·0) Prevalence of detectable microﬁ liaria due to lymphatic ﬁ lariasis 19 707·9 (17 173·8 to 22 270·5) Complications of lymphatic ﬁ lariasis aggregate 18 756·3 (12 420·1 to 26 397·1) Onchocerciasis 15 531·5 (11 963·5 to 19 993·8) Asymptomatic onchocerciasis 2280·4 (1525·0 to 3173·2) www.thelancet.com Vol 388 October 8, 2016 1557 Articles Articles (Table 3 continues on next page) Articles Prevalence (thousands) YLDs (thousands) 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 (Continued from previous page) Skin disease due to onchocerciasis aggregate 12 223·9 (8 474·7 to 16 585·2) –17·4 (–31·0 to –2·7)* –26·5 (–38·6 to –13·1)* 1055·6 (467·4–1894·7) –22·2 (–41·8 to –3·2)* –31·6 (–49·0 to –14·8)* Vision loss due to onchocerciasis aggregate 1025·4 (724·8 to 1452·1) –2·3 (–15·0 to 11·8) –22·3 (–32·3 to –11·0)* 80·1 (51·3–117·5) –6·1 (–16·9 to 6·0) –25·4 (–33·9 to –15·9)* Trachoma 3557·1 (2940·5 to 4321·8) 4·6 (0·4 to 9·2)* –19·3 (–22·9 to –15·2)* 279·2 (192·8–396·0) –1·2 (–5·6 to 3·0) –23·9 (–27·5 to –20·3)* Dengue 4730·0 (2654·1 to 10 254·2) 143·6 (–0·3 to 564·7) 119·7 (–10·1 to 498·7) 764·1 (346·8–1 744·2) 140·8 (–0·1 to 558·4) 117·7 (–9·8 to 494·3) Dengue episodes aggregate 1409·2 (937·6 to 2943·8) 134·1 (–3·8 to 521·2) 110·8 (–13·2 to 459·0) 82·1 (45·0–183·9) 141·4 (0·4 to 549·2)* 117·9 (–9·4 to 486·7) Post-dengue chronic fatigue syndrome 3324·6 (1600·3 to 7347·9) 143·7 (–0·3 to 564·7) 119·8 (–10·1 to 498·7) 682·0 (295·2–1 608·9) 140·8 (–0·2 to 558·0) 117·6 (–9·8 to 494·1) Yellow fever 2·8 (0·8 to 7·7) –25·7 (–31·6 to –19·2)* –31·6 (–36·9 to –25·8)* 0·1 (0·0–0·3) –25·7 (–31·6 to –19·2)* –31·6 (–36·9 to –25·8)* Rabies 0·7 (0·6 to 0·8) –43·4 (–51·5 to –33·8)* –50·8 (–57·6 to –42·6)* 0·1 (0·1–0·1) –43·4 (–51·5 to –33·8)* –50·8 (–57·6 to –42·6)* Intestinal nematode infections 1 447 209·3 (1 414 995·0 to 1 481 332·3) –2·5 (–5·1 to 0·2) –12·8 (–15·2 to –10·4)* 3173·8 (1861·3–5090·8) –22·8 (–27·3 to –17·3)* –30·4 (–34·5 to –25·4)* Ascariasis 761 893·8 (682 557·5 to 861 031·5) –10·8 (–22·5 to 2·5) –20·3 (–30·7 to –8·2)* 871·0 (482·5–1 464·3) –37·7 (–45·3 to –29·4)* –43·9 (–50·8 to –36·4)* Complications of ascariasis aggregate 47 626·4 (44 087·5 to 51 711·9) –35·6 (–41·9 to –28·4)* –42·1 (–47·8 to –35·6)* 871·0 (482·5–1 464·3) –37·7 (–45·3 to –29·4)* –43·9 (–50·8 to –36·4)* Asymptomatic ascariasis 714 230·7 (634 201·9 to 814 009·5) –8·5 (–20·9 to 6·4) –18·2 (–29·3 to –4·9)* 0·0 (0·0–0·0) 0·0 (0·0 to 0·0) 0·0 (0·0 to 0·0) Trichuriasis 463 652·1 (426 621·2 to 502 939·4) –2·1 (–12·0 to 9·0) –12·2 (–21·1 to –2·1)* 544·1 (290·7–946·0) –16·7 (–30·6 to 3·6) –24·9 (–37·5 to –6·8)* Complications of trichuriasis aggregate 27 129·6 (24 188·7 to 31 890·8) –15·6 (–27·8 to 2·1) –24·1 (–35·0 to –8·2)* 544·1 (290·7–946·0) –16·7 (–30·6 to 3·6) –24·9 (–37·5 to –6·8)* Asymptomatic trichuriasis 436 525·4 (400 013·3 to 475 985·5) –1·1 (–11·8 to 11·0) –11·3 (–21·0 to –0·3)* 0·0 (0·0–0·0) 0·0 (0·0 to 0·0) 0·0 (0·0 to 0·0) Hookworm disease 428 245·9 (394 486·0 to 468 292·4) –7·3 (–16·9 to 3·5) –17·2 (–25·8 to –7·5)* 1758·8 (1088·5–2754·9) –14·6 (–19·9 to –9·4)* –22·9 (–27·7 to –18·2)* Complications of hookworm disease aggregate 51 012·9 (48 151·1 to 54 346·2) –14·0 (–21·0 to –6·2)* –23·2 (–29·5 to –16·3)* 1046·7 (572·8–1732·1) –15·5 (–24·0 to –6·3)* –24·4 (–32·1 to –16·3)* Anaemia due to hookworm disease aggregate 24 220·2 (24 078·2 to 24 365·2) –6·2 (–7·3 to –5·1) –15·5 (–16·5 to –14·6)* 712·1 (476·3–1031·1) –13·2 (–15·6 to –11·1)* –20·5 (–22·6 to –18·7)* Asymptomatic hookworm disease 352 948·6 (319 143·2 to 392 574·4) –6·4 (–18·3 to 7·5) –16·4 (–27·2 to –3·9)* 0·0 (0·0–0·0) 0·0 (0·0 to 0·0) 0·0 (0·0 to 0·0) Food-borne trematodiases 71 095·4 (67 365·5 to 75 246·9) 4·0 (1·5 to 6·6)* –9·3 (–11·4 to –7·1)* 1686·5 (857·1–3066·8) 3·7 (–0·4 to 10·1)* –10·0 (–13·4 to –5·2)* Asymptomatic food- borne trematodiases aggregate 56 131·7 (45 765·9 to 63 620·3) 3·8 (1·2 to 6·6)* –9·3 (–11·6 to –6·9)* 0·0 (0·0–0·0) 0·0 (0·0–0·0) 0·0 (0·0–0·0) Symptomatic food-borne trematodiases aggregate 14 963·7 (8 632·7 to 25 256·6) 4·8 (0·5 to 11·4)* –9·3 (–12·8 to –4·2)* 1686·5 (857·1–3066·8) 3·7 (–0·4 to 10·1) –10·0 (–13·4 to –5·2)* Leprosy 514·2 (487·0 to 545·7) –0·1 (–0·5 to 0·3) –19·7 (–20·2 to –19·3)* 31·0 (20·9–43·5) 0·6 (–1·5 to 2·9) –19·2 (–20·8 to –17·4)* Ebola 2·8 (1·2 to 5·2) 54 521·3 (46 386·2 to 68 346·9)* 49 817·9 (41 865·6 to 62 764·9)* 0·6 (0·2–1·1) 51 908·8 (42 739·0 to 66 884·3)* 47 453·0 (38 609·5 to 61 730·3)* Other neglected tropical diseases 61 452·6 (60 842·8 to 62 054·9) 0·1 (–1·0 to 1·3) –7·7 (–8·8 to –6·6)* 5261·1 (2558·5–12 199·0) 7·4 (–18·9 to 50·6) –1·2 (–25·5 to 38·0) (Table 3 continues on next page) Prevalence (thousands) 1558 www.thelancet.com Vol 388 October 8, 2016 Articles Articles Prevalence (thousands) YLDs (thousands) 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 (Continued from previous page) Anaemia due to other neglected tropical diseases aggregate 61 452·6 (60 842·8 to 62 054·9) 0·1 (–1·0 to 1·3) –7·7 (–8·8 to –6·6)* 2158·2 (1 446·3–3 073·1) –6·0 (–8·2 to –3·8)* –12·4 (–14·5 to –10·4)* Maternal disorders 11 372·9 (10 686·6 to 11 894·1) 2·1 (–4·2 to 9·2) –7·9 (–13·6 to –1·8)* 898·8 (642·8–1222·9) –5·1 (–18·7 to 11·2) –15·1 (–27·0 to –0·3)* Maternal haemorrhage 2304·8 (2231·6 to 2 393·5) 1·6 (–3·0 to 6·5) –7·8 (–12·0 to –3·4)* 84·7 (57·3–119·5) 0·1 (–13·5 to 17·2) –9·1 (–21·3 to 6·6) Maternal hemorrhage episodes aggregate 191·7 (127·8 to 277·5) 2·9 (–36·5 to 66·2) –6·4 (–42·1 to 49·9) 31·3 (18·1–49·0) 3·3 (–30·6 to 54·9) –6·0 (–36·3 to 41·1)* Anaemia due to maternal haemorrhage aggregate 2113·4 (2082·8 to 2141·5) 1·5 (–0·9 to 3·9) –7·9 (–10·1 to –5·8)* 53·5 (35·0–78·6) –1·8 (–5·9 to 2·6) –10·8 (–14·5 to –6·9)* Maternal sepsis and other maternal infections 3140·8 (2658·0 to 3727·7) 3·7 (–15·2 to 27·8) –7·8 (–24·2 to 13·1) 56·5 (29·5–102·4) –1·0 (–42·3 to 70·4) –9·7 (–47·1 to 55·2) Maternal sepsis and other maternal infections aggregate 900·0 (486·7 to 1462·3) –2·4 (–50·3 to 93·7) –10·2 (–54·4 to 78·1) 45·3 (22·2–82·7) –2·7 (–49·0 to 91·0) –10·5 (–53·2 to 73·6) Infertility due to puerperal sepsis 2242·1 (2 030·5 to 2483·0) 6·4 (4·4 to 8·9)* –6·8 (–8·6 to –4·7)* 11·2 (4·2–24·3) 6·7 (4·0 to 9·5)* –6·6 (–8·9 to –4·1)* Maternal hypertensive disorders 4624·4 (3030·8 to 6592·4) 2·7 (–37·9 to 69·8)* –5·8 (–43·0 to 55·4)* 222·4 (118·4–365·4) 2·6 (–37·3 to 68·5) –5·9 (–42·4 to 54·4) Maternal hypertensive disorders episodes aggregate 4629·2 (3031·5 to 6602·4) 2·8 (–37·8 to 70·0) –5·7 (–42·9 to 55·7) 219·1 (115·4–363·3) 2·7 (–37·0 to 68·8) –5·8 (–42·4 to 55·2) Eclampsia 5·4 (2·2 to 10·5) –5·8 (–69·0 to 182·2) –12·9 (–71·6 to 158·3) 3·3 (1·2–6·3) –5·7 (–69·0 to 182·2) –12·8 (–71·6 to 158·3) Maternal obstructed labour and uterine rupture 1077·2 (933·3 to 1255·2) –13·1 (–17·5 to –7·9)* –23·9 (–27·7 to –19·7)* 352·0 (234·6–499·9) –12·6 (–17·3 to –7·3)* –23·4 (–27·4 to –18·9)* Obstructed labour, acute event 89·6 (53·0 to 142·9) –3·6 (–45·9 to 74·7) –11·4 (–49·7 to 60·1) 27·7 (14·8–47·0) –3·1 (–44·9 to 72·1) –10·9 (–49·0 to 58·2) Fistula due to maternal obstructed labour and uterine rupture aggregate 987·6 (850·6 to 1156·0) –13·9 (–16·3 to –11·4)* –24·9 (–27·0 to –22·8)* 324·3 (216·2–457·4) –13·3 (–16·6 to –10·1)* –24·3 (–27·1 to –21·7)* Maternal abortion, miscarriage, and ectopic pregnancy 441·3 (285·5 to 630·6) –0·4 (–37·2 to 58·1) –9·4 (–42·4 to 44·2) 48·2 (27·2–78·5) –0·4 (–36·9 to 60·5) –9·3 (–42·4 to 45·6) Other maternal disorders ·· ·· ·· 135·0 (90·6–190·2) –2·2 (–19·6 to 17·5) –12·2 (–27·7 to 5·5) Neonatal disorders 52 961·9 (50 435·8 to 54 978·4)* 14·7 (11·9 to 17·8)* 3·9 (1·3 to 6·7)* 10 710·5 (8113·9–13 786·1) 13·3 (9·5 to 17·4)* 3·1 (–0·3 to 6·8) Neonatal preterm birth complications 41 855·4 (36 843·4 to 47 188·1)* 10·6 (6·2 to 15·2)* –0·4 (–4·3 to 3·8)* 5090·9 (3864·5–6555·6)* 4·4 (–1·1 to 10·1) –5·2 (–10·1 to 0·2) Vision loss due to retinopathy of prematurity aggregate 4245·1 (3386·2 to 5265·3) 11·1 (7·2 to 15·1)* –2·3 (–5·8 to 1·3) 135·5 (83·0–208·8) 13·9 (9·5 to 18·4) –0·8 (–4·5 to 3·2) Complications of preterm birth complications aggregate 37 609·0 (33 086·7 to 42 697·0) 10·5 (5·6 to 15·7)* –0·1 (–4·6 to 4·5) 4955·4 (3756·9–6394·8) 4·1 (–1·4 to 10·0) –5·3 (–10·4 to 0·2) Neonatal encephalopathy due to birth asphyxia and trauma 13 681·5 (9 797·3 to 19 759·3) 27·3 (24·6 to 30·0)* 16·2 (13·7 to 18·5)* 3769·4 (2451·7–5809·3) 25·1 (22·0 to 28·3)* 14·2 (11·5 to 17·0)* Neonatal sepsis and other neonatal infections 139·1 (79·9 to 216·0) –1·6 (–3·0 to –0·0)* –6·7 (–8·0 to –5·2)* 17·7 (9·2–31·0) –1·8 (–3·5 to 0·1) –6·9 (–8·5 to –5·1)* Severe infection due to neonatal sepsis and other neonatal infections 130·1 (71·1 to 207·6) –2·2 (–3·8 to –0·5)* –7·3 (–8·8 to –5·6)* 16·6 (8·1–30·0) –2·1 (–3·7 to –0·3)* –7·1 (–8·7 to –5·4)* (Table 3 continues on next page) Neonatal disorders 52 961·9 (50 435·8 to 54 978·4)* Neonatal preterm birth complications 41 855·4 (36 843·4 to 47 188·1)* Vision loss due to retinopathy of prematurity aggregate 4245·1 (3386·2 to 5265·3) Complications of preterm birth complications aggregate 37 609·0 (33 086·7 to 42 697·0) Neonatal encephalopathy due to birth asphyxia and trauma 13 681·5 (9 797·3 to 19 759·3) Neonatal sepsis and other neonatal infections 139·1 (79·9 to 216·0) Severe infection due to neonatal sepsis and other neonatal infections 130·1 (71·1 to 207·6) 14·7 (11·9 to 17·8)* 10·6 (6·2 to 15·2)* 11·1 (7·2 to 15·1)* 10·5 (5·6 to 15·7)* 27·3 (24·6 to 30·0)* –1·6 (–3·0 to –0·0)* –2·2 (–3·8 to –0·5)* –2·1 (–3·7 to –0·3)* –7·1 (–8·7 to –5·4)* –7·1 ·7 to –5·4)* www.thelancet.com Vol 388 October 8, 2016 1559 Articles Articles Articles Prevalence (thousands) 2015 Percenta counts b and 201 (Continued from previous page) Complications of neonatal sepsis and other neonatal infections aggregate 8·9 (7·5 to 10·4) 8·0 (4·3 to 1 Haemolytic disease and other neonatal jaundice 2042·8 (1825·0 to 2309·0) 20·0 (18·8 to Other neonatal disorders ·· ·· Nutritional deﬁ ciencies 1 482 655·3 (1 477 085·9 to 2·7 (2·2–3·2 Protein-energy malnutrition 22 834·3 (21 764·0 to 23 985·1) –4·8 (–10·2 to Iodine deﬁ ciency 110 919·5 (100 337·2 to 125 252·8) 7·0 (4·7 to 9 Visible goiter due to iodine deﬁ ciency aggregate 108 307·6 (97 705·7 to 122 260·0) 6·9 (4·5 to 9 Visible goiter with heart failure due to iodine deﬁ ciency aggregate 0·3 (0·3 to 0·4) 39·6 (38·3 to Intellectual disability due to iodine deﬁ ciency aggregate 2611·6 (1085·6 to 3621·3) 10·4 (6·1 to 1 Vitamin A deﬁ ciency 4901·1 (3975·6 to 5921·7) 9·3 (6·2 to 1 Iron-deﬁ ciency anaemia 1 477 530·5 (1 470 902·3 to 1 485 322·2) 1·5 (0·9 to 2 Iron-deﬁ ciency anaemia without heart failure aggregate 1 477 458·4 (1 470 824·4 to 39·7 (38·1–41 Iron-deﬁ ciency anaemia with heart failure aggregate 72·1 (64·6 to 80·7) 39·7 (38·1 to Other nutritional deﬁ ciencies ·· ·· Other communicable, maternal, neonatal, and nutritional diseases 1 604 919·9 (1 563 786·8 to 1 636 531·0) 16·1 (15·3 to Sexually transmitted diseases excluding HIV 1 145 527·5 (1 100 971·2 to 1 176 685·1) 16·9 (15·7 to Syphilis 43 604·9 (37 160·3 to 50 658·3) 3·7 (–0·4 to Early syphilis aggregate 42 350·9 (35 898·4 to 49 441·3) 3·3 (–0·9 to Adult tertiary syphilis 1254·0 (1127·1 to 1385·8) 15·8 (13·2 to Chlamydial infection 82 822·4 (66 426·4 to 104 328·5) 8·3 (6·1 to 1 Chlamydial infection episodes aggregate 81 072·5 (64 735·8 to 102 635·8) 8·3 (6·0 to 1 Pelvic inﬂ ammatory diseases due to chlamydial infection aggregate 1·73 (147·8 to 202·0) 1·4 (–0·8 to Prevalence (thousands) YLDs (thousands) 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 (Continued from previous page) Complications of neonatal sepsis and other neonatal infections aggregate 8·9 (7·5 to 10·4) 8·0 (4·3 to 12·7)* 2·4 (–1·1 to 6·9) 1·1 (0·6–1·6) 2·6 (–9·4 to 14·4) –2·7 (–14·1 to 8·5) Haemolytic disease and other neonatal jaundice 2042·8 (1825·0 to 2309·0) 20·0 (18·8 to 21·4)* 9·2 (8·0 to 10·3)* 603·0 (451·9–770·8) 17·0 (14·8 to 19·4)* 6·5 (4·4 to 8·6)* Other neonatal disorders ·· ·· ·· 1229·5 (836·5–1 647·2) 19·9 (0·8 to 41·1)* 8·5 (–8·5 to 27·6) Nutritional deﬁ ciencies 1 482 655·3 (1 477 085·9 to 2·7 (2·2–3·2)* –7·1 (–7·5––6·7)* 54 632·4 (36 772·8–77 894·3) –3·5 (–4·9––2·1)* –11·4 (–12·7––10·3)* Protein-energy malnutrition 22 834·3 (21 764·0 to 23 985·1) –4·8 (–10·2 to 0·8) –12·1 (–17·1 to –6·9)* 2823·3 (1820·3–3984·2) –4·5 (–9·8 to 1·1) –11·8 (–16·7 to –6·6)* Iodine deﬁ ciency 110 919·5 (100 337·2 to 125 252·8) 7·0 (4·7 to 9·4)* –6·0 (–8·0 to –3·8)* 2386·9 (1 508·3–3 564·2) 7·7 (5·7 to 9·8)* –4·4 (–6·4 to –2·4)* Visible goiter due to iodine deﬁ ciency aggregate 108 307·6 (97 705·7 to 122 260·0) 6·9 (4·5 to 9·4)* –6·2 (–8·2 to –4·0)* 1927·5 (1192·0–3040·1) 6·9 (4·4 to 9·3)* –6·0 (–8·2 to –3·9)* Visible goiter with heart failure due to iodine deﬁ ciency aggregate 0·3 (0·3 to 0·4) 39·6 (38·3 to 41·0)* 4·2 (3·3 to 5·2)* 0·0 (0·0–0·1) 39·6 (38·3 to 41·0)* 4·2 (3·3 to 5·2)* Intellectual disability due to iodine deﬁ ciency aggregate 2611·6 (1085·6 to 3621·3) 10·4 (6·1 to 12·6)* 2·0 (–2·0 to 4·3) 459·3 (163·4–752·8) 11·0 (6·4 to 13·6)* 2·8 (–1·6 to 5·2) Vitamin A deﬁ ciency 4901·1 (3975·6 to 5921·7) 9·3 (6·2 to 12·5)* 0·1 (–2·7 to 2·7) 232·4 (143·5–345·9) 10·9 (7·5 to 14·6)* 0·2 (–2·7 to 3·3) Iron-deﬁ ciency anaemia 1 477 530·5 (1 470 902·3 to 1 485 322·2) 1·5 (0·9 to 2·1)* –8·0 (–8·5 to –7·4)* 48 529·2 (32 560·8–69 725·2) –3·8 (–5·1 to –2·4)* –11·6 (–12·8 to –10·5)* Iron-deﬁ ciency anaemia without heart failure aggregate 1 477 458·4 (1 470 824·4 to 39·7 (38·1–41·3)* 6·1 (4·7–7·4)* 8·1 (5·5–11·4) 39·6 (35·2–44·3)* 6·0 (2·5–9·9)* Iron-deﬁ ciency anaemia with heart failure aggregate 72·1 (64·6 to 80·7) 39·7 (38·1 to 41·3)* 6·1 (4·7 to 7·4)* 8·1 (5·5–11·4) 39·6 (35·2 to 44·3)* 6·0 (2·5 to 9·9)* Other nutritional deﬁ ciencies ·· ·· ·· 660·7 (396·0–1026·9) –16·7 (–36·8 to 5·8) –22·9 (–41·5 to –2·1)* Other communicable, maternal, neonatal, and nutritional diseases 1 604 919·9 (1 563 786·8 to 1 636 531·0) 16·1 (15·3 to 17·0)* –0·4 (–1·1 to 0·4) 3929·7 (2566·9–5786·6) 6·5 (3·9 to 9·2)* –4·0 (–6·2 to –1·7)* Sexually transmitted diseases excluding HIV 1 145 527·5 (1 100 971·2 to 1 176 685·1) 16·9 (15·7 to 18·3)* –1·0 (–2·1 to 0·1) 1573·7 (974·0–2501·8) 16·9 (14·5 to 19·1)* 1·3 (–0·6 to 3·4) Syphilis 43 604·9 (37 160·3 to 50 658·3) 3·7 (–0·4 to 7·0) –9·8 (–13·4 to –6·9)* 242·8 (166·1–333·7) 15·3 (12·2 to 18·5)* –5·6 (–8·0 to –3·0)* Early syphilis aggregate 42 350·9 (35 898·4 to 49 441·3) 3·3 (–0·9 to 6·8) –9·9 (–13·6 to –6·9)* 10·3 (3·0–25·3) 3·4 (–1·1 to 7·4) –9·9 (–13·7 to –6·3)* Adult tertiary syphilis 1254·0 (1127·1 to 1385·8) 15·8 (13·2 to 18·3)* –5·6 (–7·8 to –3·6)* 232·6 (155·4–322·8) 15·9 (12·6 to 19·1)* –5·4 (–7·9 to –2·8)* Chlamydial infection 82 822·4 (66 426·4 to 104 328·5) 8·3 (6·1 to 10·7)* –2·9 (–4·8 to –0·9)* 364·6 (210·6–589·5) 10·1 (6·9 to 13·4)* –1·8 (–4·5 to 0·8) Chlamydial infection episodes aggregate 81 072·5 (64 735·8 to 102 635·8) 8·3 (6·0 to 10·7)* –2·9 (–4·8 to –0·9)* 332·7 (185·2–546·0) 10·7 (7·1 to 14·4)* –1·1 (–4·1 to 2·1) Pelvic inﬂ ammatory diseases due to chlamydial infection aggregate 1·73 (147·8 to 202·0) 1·4 (–0·8 to 3·4) –11·5 (–13·2 to –9·8)* 23·1 (15·5 to 32·5) 1·8 (–2·2 to 5·7) –11·1 (–14.5 to 7·8)* Prevalence (thousands) 1560 www.thelancet.com Vol 388 October 8, 2016 Articles Prevalence (thousands) YLDs (thousands) 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 (Continued from previous page) Infertility due to chlamydial infection aggregate 1576·6 (1450·8 to 1703·9) 11·1 (10·0 to 12·2)* –3·0 (–3·9 to –2·0) 8·8 (3·6 to 18·4) 11·1 (9·0 to 13·2)* –2·9 (–4·8 to –1·1)* Gonococcal infection 47 468·5 (35 848·1 to 62 099·7) 25·3 (18·3 to 31·2)* 12·2 (5·9 to 17·7)* 444·9 (260·8–691·8) 26·3 (20·0 to 31·4)* 12·6 (6·7 to 17·3)* Gonococcal infection episodes aggregate 46 561·5 (34 928·2 to 61 227·4) 25·6 (18·5 to 31·6)* 12·5 (6·1 to 18·1)* 430·3 (250·5–671·5) 27·3 (20·7 to 32·5)* 13·5 (7·4 to 18·3)* Pelvic inﬂ ammatory diseases due to gonococcal infection aggregate 907·0 (822·0 to 1 016·3) 10·0 (7·6 to 11·8)* –3·8 (–6·0 to –2·2)* 14·6 (9·6–21·4) 3·5 (–0·8 to 7·1) –8·7 (–12·4 to –5·7)* Trichomoniasis 167 618·6 (144 744·1 to 193 444·0) 16·2 (15·3 to 17·2)* 0·9 (0·4 to 1·3)* 194·3 (78·0–408·3) 16·1 (15·0 to 17·2)* 1·0 (0·3 to 1·8)* Genital herpes 885 169·4 (772 303·8 to 1 008 588·1) 18·1 (16·4 to 19·9)* –1·3 (–2·7 to 0·1) 236·4 (74·3–554·2) 19·5 (17·4 to 23·0)* 0·7 (–1·4 to 5·0) Moderate infection due to initial genital herpes episode 389·0 (88·0 to 909·7) 38·1 (33·6 to 43·2)* 29·5 (25·4 to 34·3)* 19·0 (4·3–45·8) 37·6 (30·9 to 45·6)* 29·1 (22·9 to 36·7)* Complications of genital herpes aggregate 884 780·3 (771 914·6 to 1 008 400·6) 18·1 (16·4 to 19·9)* –1·3 (–2·7 to 0·0) 217·3 (62·2–534·5) 18·1 (16·3 to 19·9)* –1·2 (–2·7 to 0·3) Other sexually transmitted diseases 4821·8 (4452·8 to 5174·6) 9·6 (8·1 to 11·0)* –4·7 (–5·9 to –3·6)* 90·7 (61·5–129·7) 3·7 (1·1 to 6·2)* –10·0 (–12·2 to –7·8)* Pelvic inﬂ ammatory diseases due to other sexually transmitted diseases aggregate 507·2 (437·2 to 591·9) 1·2 (–0·1 to 2·6) –12·2 (–13·1 to –11·2)* 66·8 (45·4–92·9) 1·3 (–1·2 to 4·2) –12·0 (–14·3 to –9·8)* Infertility due to other sexually transmitted diseases aggregate 4314·5 (3943·8 to 4657·1) 10·7 (9·1 to 12·2)* –3·8 (–5·1 to –2·4)* 23·9 (9·7–48·8) 10·8 (9·0 to 12·7)* –3·6 (–5·2 to –2·0)* Other sexually transmitted diseases ·· ·· ·· 329·4 (200·1–504·2) 19·5 (14·4 to 24·6)* 6·7 (2·2 to 11·3)* Hepatitis 497 901·2 (490 397·2 to 505 815·2) 16·7 (15·8 to 17·6)* 2·2 (1·4 to 2·9)* 406·4 (267·6–588·8) 12·6 (1·7 to 24·1)* 1·2 (–8·3 to 11·5) Acute hepatitis A 8785·5 (7 950·0 to 9 598·2) 4·2 (–7·2 to 17·0) –2·8 (–13·5 to 9·1) 172·5 (111·5–249·7) 9·5 (–0·1 to 20·4) 1·9 (–6·9 to 12·0) Hepatitis B 356 083·4 (342 942·4 to 370 231·5) 16·7 (15·0 to 18·4)* 2·8 (1·3 to 4·3)* 190·2 (121·1–282·2) 16·9 (–4·5 to 40·2) 1·2 (–16·9 to 20·7) Acute hepatitis B aggregate 12 832·2 (11 239·6 to 14 567·4) 13·2 (–4·3 to 35·2) 1·0 (–14·6 to 20·0) 190·2 (121·1–282·2) 16·9 (–4·5 to 40·2) 1·2 (–16·9 to 20·7) Chronic hepatitis B 343 251·2 (330 541·3 to 357 194·6) 16·9 (15·3 to 18·4)* 2·8 (1·5 to 4·3)* 0·0 (0·0–0·0) 0·0 (0·0 to 0·0) 0·0 (0·0 to 0·0) Hepatitis C 142 745·3 (127 559·8 to 157 704·1) 18·0 (16·7 to 19·2)* 1·2 (0·1 to 2·2)* 8·7 (4·3–16·5) 17·0 (12·2 to 21·7)* 3·3 (–0·5 to 7·3) Acute hepatitis C aggregate 622·3 (580·4 to 666·6) 17·0 (15·7 to 18·4)* 3·4 (2·5 to 4·3)* 8·7 (4·3–16·5) 17·0 (12·2 to 21·7)* 3·3 (–0·5 to 7·3) Chronic hepatitis C 142 122·9 (126 977·7 to 157 044·9) 18·0 (16·7 to 19·2)* 1·2 (0·1 to 2·2)* 0·0 (0·0–0·0) 0·0 (0·0 to 0·0) 0·0 (0·0 to 0·0) Acute hepatitis E 1501·9 (1385·4 to 1636·4) 3·5 (2·2 to 4·7)* –3·2 (–4·1 to –2·3)* 34·9 (22·4–51·7) 4·9 (–4·4 to 15·5) –2·6 (–10·9 to 6·8) Other infectious diseases 54 835·4 (54 385·8 to 55 273·8) 0·4 (–1·6 to 2·3) –8·0 (–9·8 to –6·2)* 1949·6 (1307·1–2797·1) –1·6 (–5·1 to 2·0) –8·7 (–12·0 to –5·5)* Non-communicable diseases 6 652 153·4 (6 633 099·8 to 6 668 805·2) 13·7 (13·6 to 13·8)* 0·0 (–0·0 to 0·1) 638 480·8 (478 716·6–819 498·9) 18·8 (18·4 to 19·2)* –0·1 (–0·4 to 0·2) (Table 3 continues on next page) www.thelancet.com Vol 388 October 8, 2016 1561 Articles Articles Articles YLDs (thousands) Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 2015 Percentage change in counts between 2005 and 2015 Percentage change i ASR between 2005 and 2015 44·4 (42·1 to 46·7)* 12·4 (10·6 to 14·2)* 8569·3 (6265·5–11 079·2) 36·6 (31·8 to 41·2)* 6·4 (2·7 to 9·7)* 38·6 (30·3 to 47·1)* 7·9 (1·5 to 14·3)* 209·1 (150·2–271·4) 36·7 (28·6 to 45·3)* 6·2 (–0·1 to 12·7) 18·0 (0·2 to 39·8)* –5·5 (–19·2 to 11·1) 69·7 (47·2–95·8) 16·8 (0·8 to 36·0)* –6·8 (–18·7 to 7·7) 32·7 (23·8 to 41·5)* 2·2 (–4·5 to 9·0) 79·5 (57·8–104·3) 30·7 (21·9 to 39·2)* 0·6 (–6·2 to 7·2) 19·6 (2·5 to 39·9)* –8·2 (–21·2 to 7·1) 128·6 (92·0–167·5) 9·5 (–3·1 to 24·1) –16·1 (–25·4 to –5·2)* 16·3 (8·9 to 24·4)* –9·9 (–15·5 to –3·7)* 396·5 (292·4–504·2) 12·0 (4·8 to 20·2)* –13·3 (–18·8 to –7·1)* 42·6 (38·1 to 47·2)* 8·7 (5·4 to 12·2)* 762·8 (564·2–979·6) 38·4 (33·7 to 43·3)* 5·3 (1·9 to 8·9)* 42·6 (38·1 to 47·3)* 8·7 (5·4 to 12·2)* 759·8 (561·9–975·3) 38·4 (33·7 to 43·3)* 5·4 (1·9 to 9·0)* 26·1 (25·0 to 27·2)* –4·9 (–5·7 to –4·1)* 3·0 (2·0–4·1)* 26·0 (20·7 to 31·6)* –4·7 (–9·4 to –0·0)* 59·8 (38·2 to 82·2)* 23·6 (6·8 to 40·6)* 188·2 (130·4–245·6) 25·2 (9·9 to 45·0)* –1·8 (–13·5 to 13·2) 20·6 (–9·1 to 63·2) –3·6 (–26·5 to 29·1) 59·9 (40·2–80·4) 9·5 (–8·3 to 34·1) –12·5 (–26·5 to 6·1) 80·1 (53·9 to 113·8)* 38·1 (17·1 to 65·1)* 44·3 (31·6–57·3) 47·7 (34·2 to 65·9)* 13·2 (2·7 to 27·7)* 58·8 (28·9 to 94·8)* 22·1 (–0·3 to 49·1) 53·5 (37·3–70·1) 39·3 (22·1 to 61·6)* 7·2 (–5·8 to 23·5) 24·9 (–5·6 to 63·5) –2·0 (–26·3 to 29·5) 30·5 (20·7–40·6) 12·1 (–4·5 to 32·8) –11·9 (–24·9 to 4·8) 22·0 (14·4 to 30·1)* –7·4 (–13·2 to –1·3)* 39·5 (27·7–51·5) 19·0 (11·8 to 26·8)* –9·5 (–15·1 to –3·7)* 46·2 (40·7 to 52·1)* 10·9 (6·7 to 15·4)* 86·6 (61·1–112·2) 38·4 (33·1 to 43·8)* 5·1 (1·1 to 9·3)* 26·2 (20·1 to 33·0)* –2·3 (–7·0 to 2·9) 147·3 (105·7–194·3) 24·0 (18·1 to 30·0)* –3·9 (–8·4 to 0·7) 24·4 (14·2 to 35·8)* –4·9 (–12·5 to 3·4) 93·6 (67·4–121·5) 21·5 (12·7 to 30·8)* –6·8 (–13·1 to 0·2) 28·7 (28·0 to 29·3)* 1·8 (1·3 to 2·4)* 53·7 (35·8–77·6) 28·6 (25·8 to 31·4)* 1·9 (–0·2 to 4·1) 37·7 (29·6 to 47·1)* 5·6 (–0·5 to 12·6) 513·6 (377·9–648·8) 31·1 (23·4 to 40·0)* 0·6 (–5·2 to 7·1) 59·0 (50·7 to 67·0)* 26·2 (19·5 to 32·6)* 180·5 (120·8–257·6) 56·3 (47·9 to 64·0)* 23·7 (16·9 to 29·9)* 76·8 (72·0 to 81·6)* 34·5 (30·8 to 38·2)* 148·0 (104·7–197·7) 82·9 (72·7 to 93·9)* 40·1 (31·9 to 48·7)* 93·5 (82·2 to 105·8)* 49·6 (40·5 to 59·2)* 142·0 (100·7–187·1) 86·4 (75·6 to 98·2)* 42·8 (34·2 to 52·1)* 27·0 (24·0 to 29·7)* –3·9 (-6·0 to –1·8)* 6·0 (2·9–11·1) 26·9 (23·5 to 30·1)* –3·8 (–6·4 to –1·2)* 41·5 (33·4 to 49·6)* 9·9 (4·1 to 15·7)* 1796·5 (1270·7–2411·6) 36·1 (29·2 to 42·7)* 5·3 (0·4 to 10·1)* 62·7 (42·9 to 83·3)* 32·0 (16·2 to 48·1)* 989·8 (719·8–1271·0) 46·6 (32·7 to 61·5)* 15·8 (5·3 to 27·5)* 25·1 (24·1 to 26·1)* –4·7 (–5·4 to –3·9)* 806·8 (520·3–1171·8) 25·1 (24·0 to 26·2)* –4·6 (–5·5 to –3·8)* (Table 3 continues on next pag Prevalence (thousands) YLDs (thousands) 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 (Continued from previous page) Neoplasms 90 497·5 (89 215·7 to 91 896·1) 44·4 (42·1 to 46·7)* 12·4 (10·6 to 14·2)* 8569·3 (6265·5–11 079·2) 36·6 (31·8 to 41·2)* 6·4 (2·7 to 9·7)* Lip and oral cavity cancer 2425·1 (2278·7 to 2582·3) 38·6 (30·3 to 47·1)* 7·9 (1·5 to 14·3)* 209·1 (150·2–271·4) 36·7 (28·6 to 45·3)* 6·2 (–0·1 to 12·7) Nasopharynx cancer 732·7 (580·5 to 883·9) 18·0 (0·2 to 39·8)* –5·5 (–19·2 to 11·1) 69·7 (47·2–95·8) 16·8 (0·8 to 36·0)* –6·8 (–18·7 to 7·7) Other pharynx cancer 945·5 (885·9 to 1015·0) 32·7 (23·8 to 41·5)* 2·2 (–4·5 to 9·0) 79·5 (57·8–104·3) 30·7 (21·9 to 39·2)* 0·6 (–6·2 to 7·2) Oesophageal cancer 746·0 (641·5 to 925·7) 19·6 (2·5 to 39·9)* –8·2 (–21·2 to 7·1) 128·6 (92·0–167·5) 9·5 (–3·1 to 24·1) –16·1 (–25·4 to –5·2)* Stomach cancer 3539·4 (3339·9 to 3776·2) 16·3 (8·9 to 24·4)* –9·9 (–15·5 to –3·7)* 396·5 (292·4–504·2) 12·0 (4·8 to 20·2)* –13·3 (–18·8 to –7·1)* Colon and rectum cancer 9399·0 (9059·3 to 9758·3) 42·6 (38·1 to 47·2)* 8·7 (5·4 to 12·2)* 762·8 (564·2–979·6) 38·4 (33·7 to 43·3)* 5·3 (1·9 to 8·9)* Colon and rectum cancer aggregate 9377·0 (9037·3 to 9736·4) 42·6 (38·1 to 47·3)* 8·7 (5·4 to 12·2)* 759·8 (561·9–975·3) 38·4 (33·7 to 43·3)* 5·4 (1·9 to 9·0)* Stoma due to colon and rectum cancer 22·0 (21·3 to 22·8) 26·1 (25·0 to 27·2)* –4·9 (–5·7 to –4·1)* 3·0 (2·0–4·1)* 26·0 (20·7 to 31·6)* –4·7 (–9·4 to –0·0)* Liver cancer 618·7 (550·7 to 688·0) 59·8 (38·2 to 82·2)* 23·6 (6·8 to 40·6)* 188·2 (130·4–245·6) 25·2 (9·9 to 45·0)* –1·8 (–13·5 to 13·2) Liver cancer due to hepatitis B 305·9 (235·8 to 418·9) 20·6 (–9·1 to 63·2) –3·6 (–26·5 to 29·1) 59·9 (40·2–80·4) 9·5 (–8·3 to 34·1) –12·5 (–26·5 to 6·1) Liver cancer due to hepatitis C 268·7 (228·4 to 320·8) 80·1 (53·9 to 113·8)* 38·1 (17·1 to 65·1)* 44·3 (31·6–57·3) 47·7 (34·2 to 65·9)* 13·2 (2·7 to 27·7)* Liver cancer due to alcohol use 273·6 (229·0 to 350·7) 58·8 (28·9 to 94·8)* 22·1 (–0·3 to 49·1) 53·5 (37·3–70·1) 39·3 (22·1 to 61·6)* 7·2 (–5·8 to 23·5) Liver cancer due to other causes 151·0 (118·8 to 201·0) 24·9 (–5·6 to 63·5) –2·0 (–26·3 to 29·5) 30·5 (20·7–40·6) 12·1 (–4·5 to 32·8) –11·9 (–24·9 to 4·8) Gallbladder and biliary tract cancer 149·4 (138·5 to 158·9) 22·0 (14·4 to 30·1)* –7·4 (–13·2 to –1·3)* 39·5 (27·7–51·5) 19·0 (11·8 to 26·8)* –9·5 (–15·1 to –3·7)* Pancreatic cancer 393·8 (372·9 to 417·1) 46·2 (40·7 to 52·1)* 10·9 (6·7 to 15·4)* 86·6 (61·1–112·2) 38·4 (33·1 to 43·8)* 5·1 (1·1 to 9·3)* Larynx cancer 1412·6 (1340·0 to 1499·9) 26·2 (20·1 to 33·0)* –2·3 (–7·0 to 2·9) 147·3 (105·7–194·3) 24·0 (18·1 to 30·0)* –3·9 (–8·4 to 0·7) Larynx cancer aggregate 820·9 (749·1 to 911·1) 24·4 (14·2 to 35·8)* –4·9 (–12·5 to 3·4) 93·6 (67·4–121·5) 21·5 (12·7 to 30·8)* –6·8 (–13·1 to 0·2) Laryngectomy due to larynx cancer 591·0 (576·5 to 605·4) 28·7 (28·0 to 29·3)* 1·8 (1·3 to 2·4)* 53·7 (35·8–77·6) 28·6 (25·8 to 31·4)* 1·9 (–0·2 to 4·1) Tracheal, bronchus, and lung cancer 3299·7 (3095·1 to 3536·4) 37·7 (29·6 to 47·1)* 5·6 (–0·5 to 12·6) 513·6 (377·9–648·8) 31·1 (23·4 to 40·0)* 0·6 (–5·2 to 7·1) Malignant skin melanoma 3082·3 (2475·8 to 3901·7) 59·0 (50·7 to 67·0)* 26·2 (19·5 to 32·6)* 180·5 (120·8–257·6) 56·3 (47·9 to 64·0)* 23·7 (16·9 to 29·9)* Non-melanoma skin cancer 2558·2 (2494·8 to 2625·6) 76·8 (72·0 to 81·6)* 34·5 (30·8 to 38·2)* 148·0 (104·7–197·7) 82·9 (72·7 to 93·9)* 40·1 (31·9 to 48·7)* Non-melanoma skin cancer (squamous-cell carcinoma) 2155·9 (2019·5 to 2312·2) 93·5 (82·2 to 105·8)* 49·6 (40·5 to 59·2)* 142·0 (100·7–187·1) 86·4 (75·6 to 98·2)* 42·8 (34·2 to 52·1)* N l ki 578 1 27 0 3 9 6 0 26 9 3 8 Prevalence (thousands) Prevalence (thousands) 2015 (Continued from previous page) Neoplasms 90 497·5 (89 215·7 to 91 896·1) Lip and oral cavity cancer 2425·1 (2278·7 to 2582·3) Nasopharynx cancer 732·7 (580·5 to 883·9) Other pharynx cancer 945·5 (885·9 to 1015·0) Oesophageal cancer 746·0 (641·5 to 925·7) Stomach cancer 3539·4 (3339·9 to 3776·2) Colon and rectum cancer 9399·0 (9059·3 to 9758·3) Colon and rectum cancer aggregate 9377·0 (9037·3 to 9736·4) Stoma due to colon and rectum cancer 22·0 (21·3 to 22·8) Liver cancer 618·7 (550·7 to 688·0) Liver cancer due to hepatitis B 305·9 (235·8 to 418·9) Liver cancer due to hepatitis C 268·7 (228·4 to 320·8) Liver cancer due to alcohol use 273·6 (229·0 to 350·7) Liver cancer due to other causes 151·0 (118·8 to 201·0) Gallbladder and biliary tract cancer 149·4 (138·5 to 158·9) Pancreatic cancer 393·8 (372·9 to 417·1) Larynx cancer 1412·6 (1340·0 to 1499·9) Larynx cancer aggregate 820·9 (749·1 to 911·1) Laryngectomy due to larynx cancer 591·0 (576·5 to 605·4) Tracheal, bronchus, and lung cancer 3299·7 (3095·1 to 3536·4) Malignant skin melanoma 3082·3 (2475·8 to 3901·7) Non-melanoma skin cancer 2558·2 (2494·8 to 2625·6) Non-melanoma skin cancer (squamous-cell carcinoma) 2155·9 (2019·5 to 2312·2) Non-melanoma skin cancer (basal-cell carcinoma) 578·1 (497·7 to 671·7) Breast cancer 21 361·8 (20 249·5 to 22 266·3) Breast cancer aggregate 10 718·5 (9581·7 to 11 667·2) Mastectomy due to breast cancer 10 638·5 (10 404·8 to 10 894·5) –4·6 (–5·5 to –3·8)* (Table 3 continues on next page) 1562 www.thelancet.com Vol 388 October 8, 2016 Articles Prevalence (thousands) YLDs (thousands) 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 (Continued from previous page) Cervical cancer 3442·4 (3129·9 to 3765·0) –1·6 (–11·4 to 9·8) –20·5 (–28·2 to –11·5)* 264·0 (187·9–342·6) –1·4 (–10·9 to 9·8) –20·5 (–28·0 to –11·6)* Uterine cancer 3827·7 (3425·6 to 4285·2) 39·5 (22·9 to 55·9)* 8·2 (–4·0 to 20·0) 250·2 (172·4–341·0) 37·1 (21·1 to 52·9)* 6·1 (–5·4 to 17·6) Ovarian cancer 1174·2 (1107·4 to 1250·1) 28·3 (20·8 to 36·3)* 0·3 (–5·1 to 6·3) 150·7 (110·5–192·3) 25·8 (17·8 to 33·9)* –1·7 (–7·4 to 4·3) Prostate cancer 14 434·4 (11 932·2 to 19 785·1) 70·6 (61·6 to 81·4)* 29·7 (22·5 to 38·4)* 1150·3 (804·9–1 643·3) 60·5 (51·7 to 70·6)* 21·5 (14·5 to 29·6)* Prostate cancer aggregate 13 492·3 (10 984·7 to 18 837·5) 74·6 (64·7 to 86·3)* 32·5 (24·6 to 42·1)* 1069·7 (739·5–1 543·7) 63·5 (53·7 to 74·9)* 23·5 (15·8 to 32·6)* Impotence and incontinence due to prostate cancer aggregate 942·0 (927·1 to 957·3) 29·0 (28·6 to 29·4)* –0·9 (–1·2 to –0·5)* 80·6 (55·2–111·8) 28·9 (27·1 to 30·5)* –1·0 (–2·4 to 0·2) Testicular cancer 685·8 (634·7 to 732·4) 41·9 (30·8 to 52·4)* 23·4 (13·9 to 32·2)* 42·1 (28·4–58·3) 40·3 (29·3 to 50·6)* 21·5 (11·9 to 30·4)* Kidney cancer 2870·3 (2728·0 to 3031·5) 57·9 (50·0 to 66·0)* 22·8 (16·8 to 29·1)* 202·7 (145·7–270·5) 54·4 (46·9 to 61·8)* 19·8 (14·0 to 25·6)* Bladder cancer 3407·9 (3240·0 to 3603·3) 35·7 (28·6 to 43·3)* 3·3 (–2·0 to 9·0) 267·0 (193·7–349·4) 32·1 (25·8 to 38·8)* 0·7 (–4·0 to 5·6) Bladder cancer episodes aggregate 3273·4 (3 105·3 to 3 470·7) 36·3 (28·9 to 44·3)* 3·7 (–1·8 to 9·6) 244·1 (176·2–317·4) 33·1 (26·1 to 40·4)* 1·2 (–3·9 to 6·7) Urinary incontinence due to bladder cancer 134·3 (129·7 to 139·4) 22·8 (22·0 to 23·7)* –5·7 (–6·2 to –5·1)* 22·9 (15·9–31·2) 23·0 (18·7 to 27·1)* –5·3 (–8·6 to –2·2)* Brain and nervous system cancer 1205·1 (1101·7 to 1322·9) 25·3 (14·9 to 37·6)* 4·2 (–3·9 to 13·8) 126·1 (90·3–164·5) 24·7 (15·0 to 36·3)* 2·8 (–4·7 to 12·0) Thyroid cancer 3166·5 (2936·7 to 3340·6) 101·1 (86·4 to 112·9)* 60·8 (48·9 to 70·2)* 190·0 (132·5–259·3) 96·8 (81·6 to 109·2)* 56·5 (44·3 to 66·3)* Mesothelioma 60·8 (58·1 to 63·6) 41·3 (35·1 to 47·7)* 9·5 (4·6 to 14·5)* 12·2 (8·7–15·8) 39·5 (32·1 to 46·8)* 7·9 (2·1 to 13·9)* Hodgkin’s lymphoma 574·4 (519·0 to 672·9) 17·8 (10·5 to 25·3)* 0·0 (–6·1 to 6·4) 48·5 (33·7–65·8) 14·6 (7·4 to 21·7)* –3·6 (–9·6 to 2·4)* Non-Hodgkin lymphoma 4292·3 (3741·0 to 4593·6) 63·5 (47·5 to 74·3)* 30·5 (19·1 to 38·4)* 312·3 (221·8–414·2) 57·9 (43·8 to 67·8)* 25·6 (15·1 to 33·1)* Multiple myeloma 488·2 (449·4 to 527·7) 54·6 (44·8 to 64·5)* 18·0 (10·7 to 25·4)* 104·0 (73·6–134·8) 49·0 (40·0 to 58·5)* 13·6 (6·9 to 21·0)* Leukaemia 2303·6 (2216·1 to 2384·5) 35·8 (29·7 to 41·8)* 10·2 (5·7 to 14·8)* 379·4 (276·2–488·9) 28·6 (21·2 to 36·9)* 4·3 (–1·1 to 10·4) Acute lymphoid leukaemia 875·5 (709·0 to 1072·3) 25·4 (7·3 to 46·4)* 10·2 (–5·0 to 27·8) 97·3 (68·0–129·8) 25·7 (11·5 to 40·8)* 8·1 (–3·1 to 20·0) Chronic lymphoid leukaemia 904·0 (833·2 to 974·4) 31·7 (22·2 to 42·9)* 4·5 (–2·2 to 12·4) 119·8 (87·2–152·9) 27·6 (18·9 to 37·4)* 0·5 (–6·0 to 7·7) Acute myeloid leukaemia 999·3 (882·5 to 1136·8) 38·4 (26·6 to 50·7)* 14·2 (5·7 to 23·5)* 120·9 (87·3–155·6) 36·5 (27·5 to 45·6)* 10·4 (3·7 to 17·4)* Chronic myeloid leukaemia 297·7 (271·8 to 324·7) 22·2 (14·3 to 31·0)* –3·4 (–9·3 to 3·4) 41·4 (29·5–53·6) 17·7 (10·2 to 26·0)* –7·5 (–13·3 to –1·2)* Other neoplasms 4577·5 (4145·5 to 4922·0) 36·1 (27·2 to 44·5)* 10·9 (4·1 to 17·6)* 323·2 (228·6–431·8) 34·4 (26·2 to 42·6)* 9·0 (2·4 to 15·4)* Cardiovascular diseases 422 738·4 (415 534·5 to 427 870·8) 24·8 (24·0 to 25·6)* –1·8 (–2·5 to –1·2)* 25 620·1 (18 401·6–33 656·6) 23·5 (21·7 to 24·5)* –2·2 (–3·2 to –1·7)* Rheumatic heart disease 33 438·8 (29 725·7 to 43 119·8) –4·7 (–11·0 to 0·1) –16·3 (–22·5 to –12·1)* 1654·7 (1041·4–2530·9) –3·4 (–9·9 to 1·4) –15·4 (–21·5 to –11·1)* Rheumatic heart disease, without heart failure 32 236·8 (28 520·8 to 41 912·7) –5·6 (–11·9 to –0·8)* –16·9 (–23·1 to –12·7)* 1518·6 (941·1–2356·4) –5·5 (–11·7 to –0·6)* –16·8 (–22·9 to –12·3)* (Table 3 continues on next page) (Continued from previous page) (Table 3 continues on next page) www.thelancet.com Vol 388 October 8, 2016 1563 Articles Articles (Table 3 continues on next page) Articles YLDs (thousands) nge in n 2005 Percentage change in ASR between 2005 and 2015 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 2·2 (1·3 to 3·2)* 136·1 (94·5–187·6) 29·6 (27·2 to 32·1)* 2·4 (0·6 to 4·2)* –3·4 (–4·2 to –2·6)* 7274·7 (4958·6–9940·3) 30·2 (29·1 to 31·3)* –0·3 (–1·0 to 0·3) –18·1 (–21·6 to –14·8)* 33·0 (23·0–44·9) 15·2 (12·6 to 17·9)* –13·1 (–15·1 to –11·1)* –0·3 (–1·2 to 0·5) 4794·5 (3100·7–6633·6) 29·5 (27·8 to 31·0)* –0·0 (–1·0 to 0·8) –0·6 (–1·2 to –0·1)* 2447·2 (1736·8–3342·6) 31·9 (31·0 to 32·8)* –0·5 (–1·1 to 0·1) –4·4 (–4·8 to –3·9)* 6455·2 (4487·0–8609·6) 20·7 (19·8 to 21·6)* –4·2 (–4·9 to –3·5)* –5·2 (–5·8 to –4·6)* 3660·0 (2559·3–4874·9) 22·0 (21·0 to 23·0)* –4·9 (–5·7 to –4·2)* –5·1 (–5·8 to –4·5)* 3605·1 (2512·6–4815·3) 22·1 (21·1 to 23·1)* –4·9 (–5·6 to –4·1)* –10·6 (–12·7 to –8·5)* 54·9 (36·3–74·5) 17·5 (13·8 to 20·9)* –10·0 (–12·9 to –7·2)* –3·3 (–3·8 to –2·7)* 2795·2 (1945·8–3749·6) 19·1 (18·1 to 20·2)* –3·1 (–3·9 to –2·3)* –3·2 (–3·8 to –2·6)* 2757·3 (1918·2–3700·6) 19·2 (18·1 to 20·3)* –3·0 (–3·8 to –2·2)* –9·1 (–11·1 to –7·3)* 37·9 (24·7–51·2) 14·6 (11·8 to 17·0)* –8·8 (–11·0 to –6·9)* 3·4 (2·9 to 3·9)* 670·0 (467·1–917·1) 37·2 (35·9 to 38·4)* 3·6 (2·8 to 4·4)* –1·3 (–2·0 to –0·7)* 275·1 (190·3–377·9) 26·3 (24·6 to 28·1)* –1·3 (–2·5 to –0·2)* 2·2 (1·3 to 3·1)* 7·8 (4·8–11·6) 21·1 (18·5 to 23·9)* 2·1 (0·3 to 3·8)* –1·5 (–2·2 to –0·9)* 267·3 (185·9–367·9) 26·5 (24·7 to 28·3)* –1·4 (-2·6 to -0·2)* –2·5 (–3·1 to –1·9)* 2634·6 (1782·6–3637·3) 28·2 (27·1 to 29·3)* –2·4 (–3·1 to –1·7)* 1·8 (1·3 to 2·3)* 572·8 (272·1–1056·2) 34·5 (32·4 to 36·8)* 2·0 (1·2 to 2·8)* –5·4 (–6·7 to –4·1)* 12·3 (8·5–17·1) 22·9 (18·6 to 27·6)* –5·2 (–8·8 to –1·2)* –0·7 (–2·7 to 1·2) 1·0 (0·6–1·4) 22·7 (20·0 to 25·3)* –0·9 (–3·1 to 1·2) –6·1 (–7·6 to –4·6)* 11·4 (7·9–15·9) 22·9 (18·3 to 28·0)* –5·5 (–9·3 to –1·2)* 1·2 (0·4 to 2·0)* 123·7 (86·0–173·3) 34·4 (32·5 to 36·5)* 1·3 (–0·2 to 2·7) 0·4 (–0·2 to 1·0)* 5947·0 (4074·6–8105·9) 23·7 (22·2 to 25·2)* 0·5 (–0·2 to 1·1) –3·3 (–4·1 to –2·4)* 30 465·9 (23 341·4–38 294·2) 11·8 (10·2 to 13·6)* –4·6 (–6·2 to –3·1)* –5·8 (–7·3 to –4·4)* 12 047·0 (10 206·8–13 725·4) 16·2 (13·4 to 18·8)* –5·9 (–8·0 to –3·9)* (Table 3 continues on next page Prevalence (thousands) YLDs (thousands) 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 (Continued from previous page) Heart failure due to rheumatic heart disease aggregate 1202·0 (1138·2 to 1269·0) 29·8 (28·5 to 31·2)* 2·2 (1·3 to 3·2)* 136·1 (94·5–187·6) 29·6 (27·2 to 32·1)* 2·4 (0·6 to 4·2)* Ischaemic heart disease 110 550·3 (100 675·9 to 121 798·7) 25·9 (24·6 to 27·2)* –3·4 (–4·2 to –2·6)* 7274·7 (4958·6–9940·3) 30·2 (29·1 to 31·3)* –0·3 (–1·0 to 0·3) Myocardial infarction episodes aggregate 15 930·1 (14 106·5 to 17 764·0) 6·4 (1·9 to 10·6)* –18·1 (–21·6 to –14·8)* 33·0 (23·0–44·9) 15·2 (12·6 to 17·9)* –13·1 (–15·1 to –11·1)* Angina due to ischaemic heart disease aggregate 72 344·6 (62 699·8 to 83 460·0) 29·3 (27·7 to 30·7)* –0·3 (–1·2 to 0·5) 4794·5 (3100·7–6633·6) 29·5 (27·8 to 31·0)* –0·0 (–1·0 to 0·8) Heart failure due to ischaemic heart disease aggregate 22 275·7 (21 272·0 to 23 320·4) 31·9 (31·1 to 32·7)* –0·6 (–1·2 to –0·1)* 2447·2 (1736·8–3342·6) 31·9 (31·0 to 32·8)* –0·5 (–1·1 to 0·1) Cerebrovascular disease 42 430·9 (42 068·2 to 42 767·1) 21·0 (20·4 to 21·5)* –4·4 (–4·8 to –3·9)* 6455·2 (4487·0–8609·6) 20·7 (19·8 to 21·6)* –4·2 (–4·9 to –3·5)* Ischaemic stroke 24 929·0 (24 362·2 to 25 610·0) 21·8 (21·0 to 22·6)* –5·2 (–5·8 to –4·6)* 3660·0 (2559·3–4874·9) 22·0 (21·0 to 23·0)* –4·9 (–5·7 to –4·2)* Chronic ischaemic stroke aggregate 24 663·2 (24 092·1 to 25 345·8) 21·9 (21·1 to 22·7)* –5·1 (–5·8 to –4·5)* 3605·1 (2512·6–4815·3) 22·1 (21·1 to 23·1)* –4·9 (–5·6 to –4·1)* Ischaemic stroke episodes aggregate 265·8 (246·6 to 285·1) 16·8 (14·2 to 19·7)* –10·6 (–12·7 to –8·5)* 54·9 (36·3–74·5) 17·5 (13·8 to 20·9)* –10·0 (–12·9 to –7·2)* Haemorrhagic stroke 18 669·6 (18 258·7 to 19 124·5) 19·2 (18·5 to 19·9)* –3·3 (–3·8 to –2·7)* 2795·2 (1945·8–3749·6) 19·1 (18·1 to 20·2)* –3·1 (–3·9 to –2·3)* Chronic haemorrhagic stroke aggregate 18 494·8 (18 088·5 to 18 959·5) 19·3 (18·6 to 19·9)* –3·2 (–3·8 to –2·6)* 2757·3 (1918·2–3700·6) 19·2 (18·1 to 20·3)* –3·0 (–3·8 to –2·2)* Acute haemorrhagic stroke aggregate 174·8 (162·3 to 187·8) 14·3 (11·6 to 16·6)* –9·1 (–11·1 to –7·3)* 37·9 (24·7–51·2) 14·6 (11·8 to 17·0)* –8·8 (–11·0 to –6·9)* Hypertensive heart disease 6086·2 (5732·7 to 6434·1) 37·1 (36·2 to 37·9)* 3·4 (2·9 to 3·9)* 670·0 (467·1–917·1) 37·2 (35·9 to 38·4)* 3·6 (2·8 to 4·4)* Cardiomyopathy and myocarditis 2536·8 (2404·8 to 2661·4) 26·7 (25·6 to 27·7)* –1·3 (–2·0 to –0·7)* 275·1 (190·3–377·9) 26·3 (24·6 to 28·1)* –1·3 (–2·5 to –0·2)* Acute myocarditis 156·3 (137·1 to 179·6) 21·4 (19·1 to 23·7)* 2·2 (1·3 to 3·1)* 7·8 (4·8–11·6) 21·1 (18·5 to 23·9)* 2·1 (0·3 to 3·8)* Heart failure due to cardiomyopathy aggregate 2380·4 (2254·3 to 2502·2) 27·0 (25·9 to 28·1)* –1·5 (–2·2 to –0·9)* 267·3 (185·9–367·9) 26·5 (24·7 to 28·3)* –1·4 (-2·6 to -0·2)* Atrial ﬁ brillation and ﬂ utter 33 294·3 (29 959·8 to 37 202·0) 28·2 (27·2 to 29·1)* –2·5 (–3·1 to –1·9)* 2634·6 (1782·6–3637·3) 28·2 (27·1 to 29·3)* –2·4 (–3·1 to –1·7)* Peripheral vascular disease 154 650·6 (136 318·0 to 176 210·9) 34·4 (33·4 to 35·2)* 1·8 (1·3 to 2·3)* 572·8 (272·1–1056·2) 34·5 (32·4 to 36·8)* 2·0 (1·2 to 2·8)* Endocarditis 115·7 (108·1 to 124·8) 22·7 (20·9 to 24·5)* –5·4 (–6·7 to –4·1)* 12·3 (8·5–17·1) 22·9 (18·6 to 27·6)* –5·2 (–8·8 to –1·2)* Endocarditis episodes aggregate 16·9 (15·7 to 18·2) 22·8 (20·3 to 25·3)* –0·7 (–2·7 to 1·2) 1·0 (0·6–1·4) 22·7 (20·0 to 25·3)* –0·9 (–3·1 to 1·2) Heart failure due to endocarditis aggregate 98·8 (91·1 to 107·6) 22·7 (20·7 to 24·7)* –6·1 (–7·6 to –4·6)* 11·4 (7·9–15·9) 22·9 (18·3 to 28·0)* –5·5 (–9·3 to –1·2)* Heart failure due to other cardiovascular diseases aggregate 1126·5 (1045·9 to 1206·1) 34·4 (33·3 to 35·6)* 1·2 (0·4 to 2·0)* 123·7 (86·0–173·3) 34·4 (32·5 to 36·5)* 1·3 (–0·2 to 2·7) Other cardiovascular diseases episodes aggregate 89 221·1 (83 635·8 to 95 470·4) 24·0 (22·6 to 25·4)* 0·4 (–0·2 to 1·0)* 5947·0 (4074·6–8105·9) 23·7 (22·2 to 25·2)* 0·5 (–0·2 to 1·1) Chronic respiratory diseases 514 625·8 (503 322·3 to 527 617·0) 12·1 (11·1 to 13·1)* –3·3 (–4·1 to –2·4)* 30 465·9 (23 341·4–38 294·2) 11·8 (10·2 to 13·6)* –4·6 (–6·2 to –3·1)* Chronic obstructive l di 174 483·1 (160204 9 t 188951 7) 17·0 (15 1 t 19 0)* –5·8 ( 7 3 t 4 4)* 12 047·0 (10206 8 13725 4) 16·2 (13 4 t 18 8)* –5·9 ( 8 0 t 3 9)* Prevalence (thousands) 1·3 (–0·2 to 2·7) 0·5 (–0·2 to 1·1) 1564 www.thelancet.com Vol 388 October 8, 2016 Articles Prevalence (thousands) YLDs (thousands) 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 (Continued from previous page) Chronic obstructive pulmonary disease without heart failure aggregate 168 694·6 (154 443·1 to 182 957·1) 16·5 (14·5 to 18·5)* –6·1 (–7·6 to –4·6)* 9671·6 (8183·6–11 080·9) 12·3 (9·2 to 15·3)* –8·0 (–10·5 to –5·6)* Severe chronic obstructive pulmonary disease with heart failure aggregate 5876·4 (5073·7 to 6548·3) 34·9 (31·1 to 37·8)* 2·4 (–0·3 to 4·2) 2375·5 (1890·3–2747·2) 35·1 (31·4 to 38·2)* 2·5 (–0·0 to 4·5) Pneumoconiosis 2405·8 (2317·2 to 2486·3) 30·3 (27·8 to 32·9)* 8·2 (6·2 to 10·2)* 474·0 (305·6–682·6) 26·6 (23·6 to 29·3)* 6·0 (3·6 to 8·4)* Silicosis 402·9 (364·9 to 442·2) 18·1 (15·7 to 20·3)* –4·3 (–6·1 to –2·6)* 65·7 (41·7–95·4) 18·5 (14·5 to 22·4)* –3·9 (–7·0 to –0·9)* Silicosis without heart failure aggregate 389·0 (351·0 to 426·8) 17·6 (15·2 to 19·9)* –4·6 (–6·4 to –2·8)* 59·2 (36·8–87·9) 17·1 (12·7 to 21·3)* –4·7 (–8·0 to –1·5)* Severe silicosis with heart failure aggregate 13·9 (12·0 to 15·6) 32·2 (29·1 to 34·8)* 1·5 (–0·5 to 3·3) 6·4 (4·5–8·5) 33·1 (26·9 to 40·5)* 2·3 (–2·7 to 7·9) Asbestosis 157·3 (144·8 to 170·9) 30·6 (28·8 to 32·6)* 2·7 (1·3 to 4·1)* 24·8 (16·0–35·6) 30·2 (27·0 to 33·6)* 2·5 (–0·1 to 5·2) Asbestosis without heart failure aggregate 154·5 (141·8 to 168·1) 30·6 (28·7 to 32·5)* 2·7 (1·3 to 4·1)* 23·5 (15·0–34·0) 29·9 (26·4 to 33·4)* 2·5 (–0·2 to 5·4) Severe asbestosis with heart failure aggregate 2·8 (2·6 to 3·0) 35·6 (34·2 to 37·1)* 1·1 (0·1 to 2·1)* 1·3 (0·9–1·7)* 35·7 (31·3 to 40·0)* 1·3 (–2·2 to 4·7) Coal workers pneumoconiosis 84·2 (76·7 to 93·6) 37·0 (34·1 to 39·9)* 7·6 (5·2 to 10·0)* 13·3 (8·4–19·1)* 36·3 (29·4 to 43·4)* 7·0 (1·3 to 12·9)* Coal workers pneumoconiosis without heart failure aggregate 82·3 (74·9 to 91·7) 36·9 (34·0 to 39·8)* 7·5 (5·2 to 10·0)* 12·4 (7·8–17·9)* 36·0 (28·6 to 43·5)* 6·9 (0·9 to 13·2)* Severe coal workers pneumoconiosis with heart failure aggregate 1·9 (1·6 to 2·1) 41·6 (37·8 to 45·1)* 8·2 (5·4 to 10·9)* 0·9 (0·6–1·2)* 41·6 (37·7 to 45·2)* 8·2 (5·2 to 10·9)* Other pneumoconiosis 2388·6 (2164·7 to 2629·1) 27·9 (25·3 to 30·7)* 8·3 (6·0 to 10·5)* 370·2 (239·2–533·9)* 27·5 (23·9 to 31·0)* 8·3 (5·3 to 11·2)* Other pneumoconiosis without heart failure aggregate 2374·2 (2150·3 to 2615·1) 27·8 (25·2 to 30·7)* 8·3 (6·0 to 10·6)* 363·6 (234·0–524·9)* 27·3 (23·6 to 30·8)* 8·4 (5·3 to 11·4)* Severe other pneumoconiosis with heart failure aggregate 14·4 (13·1 to 15·6) 41·1 (39·0 to 43·0)* 4·5 (3·4 to 5·5)* 6·5 (4·6–8·5) 40·5 (35·1 to 45·7)* 4·3 (0·3 to 8·4)* Asthma 358 197·9 (323 133·7 to 393 465·6) 9·5 (7·6 to 11·6)* –2·5 (–4·3 to –0·5)* 15 898·9 (10 371·0–22 344·1) 9·4 (7·4 to 11·5)* –2·3 (–4·3 to –0·3)* Asymptomatic asthma 108 461·1 (92 842·5 to 126 250·5) 9·5 (7·6 to 11·6)* –2·5 (–4·3 to –0·5)* 0·0 (0·0–0·0) 0·0 (0·0 to 0·0) 0·0 (0·0 to 0·0) Symptomatic asthma aggregate 249 736·8 (220 469·6 to 278 419·7) 9·5 (7·6 to 11·6)* –2·5 (–4·3 to –0·5)* 15 898·9 (10 371·0–22 344·1) 9·4 (7·4 to 11·5)* –2·3 (–4·3 to –0·3)* Interstitial lung disease and pulmonary sarcoidosis 1916·0 (1757·4 to 2075·5) 25·8 (23·8 to 27·7)* 0·6 (–0·9 to 2·1) 234·7 (147·5–334·2) 25·7 (23·5 to 28·1)* 0·7 (–1·1 to 2·6) Interstitial lung disease and pulmonary sarcoidosis without heart failure aggregate 1718·7 (1565·7 to 1881·0) 25·2 (23·1 to 27·2)* 0·5 (–1·1 to 2·1) 148·5 (88·8–221·4) 22·7 (19·6 to 25·9)* 0·2 (–2·2 to 2·6) Prevalence (thousan 2015 Continued from previous page) Chronic obstructive pulmonary disease without heart failure aggregate 168 694·6 (154 443·1 to 182 957 Severe chronic obstructive pulmonary disease with heart failure aggregate 5876·4 (5073·7 to 6548·3) Pneumoconiosis 2405·8 (2317·2 to 2486·3) Silicosis 402·9 (364·9 to 442·2) Silicosis without heart failure aggregate 389·0 (351·0 to 426·8) Severe silicosis with heart failure aggregate 13·9 (12·0 to 15·6) Asbestosis 157·3 (144·8 to 170·9) Asbestosis without heart failure aggregate 154·5 (141·8 to 168·1) Severe asbestosis with heart failure aggregate 2·8 (2·6 to 3·0) Coal workers pneumoconiosis 84·2 (76·7 to 93·6) Coal workers pneumoconiosis without heart failure aggregate 82·3 (74·9 to 91·7) Severe coal workers pneumoconiosis with heart failure aggregate 1·9 (1·6 to 2·1) Other pneumoconiosis 2388·6 (2164·7 to 2629·1) Other pneumoconiosis without heart failure aggregate 2374·2 (2150·3 to 2615·1) Severe other pneumoconiosis with heart failure aggregate 14·4 (13·1 to 15·6) Asthma 358 197·9 (323 133·7 to 393 465 Asymptomatic asthma 108 461·1 (92 842·5 to 126 250 Symptomatic asthma aggregate 249 736·8 (220 469·6 to 278 41 Interstitial lung disease and pulmonary sarcoidosis 1916·0 (1757·4 to 2075·5) Interstitial lung disease and pulmonary sarcoidosis without heart failure aggregate 1718·7 (1565·7 to 1881·0) 0·2 (–2·2 to 2·6) 0·2 6 (Table 3 continues on next page) www.thelancet.com Vol 388 October 8, 2016 1565 Articles Articles YLDs (thousands) hange in een 2005 Percentage change in ASR between 2005 and 2015 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 )* 1·3 (–0·1 to 2·9) 86·2 (54·8–117·6) 31·3 (28·6 to 34·1)* 1·3 (–1·0 to 3·7) ·· 1811·3 (1483·7–2123·4) 1·7 (–7·3 to 12·2) –15·8 (–23·5 to –6·9)* 8)* 8·2 (7·0 to 9·5)* 501·0 (352·9–683·1) 31·3 (29·8 to 32·9)* 7·3 (6·1 to 8·4)* 7)* 4·7 (3·2 to 6·2)* 130·6 (88·7–178·7) 30·5 (27·0 to 33·8)* 4·8 (2·2 to 7·3)* )* 7·6 (6·0 to 9·1)* 111·4 (78·2–152·6) 35·2 (31·6 to 38·6)* 7·6 (5·0 to 10·3)* 3)* 9·8 (8·5 to 11·2)* 133·9 (92·5–185·1) 37·2 (33·9 to 40·5)* 9·9 (7·5 to 12·4)* )* 6·8 (5·9 to 7·6)* 125·0 (87·5–172·2) 23·3 (20·1 to 26·8)* 6·8 (4·2 to 9·5)* 7)* –9·4 (–10·4 to –8·3)* 12 142·8 (8 492·0–16 592·5) 11·4 (9·8 to 13·2)* –9·3 (–10·6 to –7·9)* –17·0 (–17·8 to –16·0)* 2341·7 (1605·6–3293·6) 1·1 (–0·2 to 2·5) –19·5 (–20·6 to –18·4)* –16·7 (–18·4 to –15·0)* 518·7 (357·9–703·2) 4·5 (2·2 to 6·8)* –16·7 (–18·5 to –14·9)* –17·0 (–17·9 to –16·0)* 1823·1 (1225·8–2621·2) 0·2 (–1·4 to 1·9) –20·3 (–21·6 to –19·0)* )* –6·9 (–8·8 to –5·0)* 4913·0 (3347·3–6967·1) 11·2 (8·0 to 14·6)* –9·5 (–11·9 to –6·9)* )* –6·9 (–10·8 to –5·7)* 416·0 (277·1–567·9) 15·4 (11·2 to 17·5)* –6·7 (–10·6 to –5·2)* )* –6·9 (–8·9 to –5·0)* 4497·1 (3050·6–6393·0) 10·8 (7·4 to 14·5)* –9·8 (–12·4 to –7·0)* )* 2·8 (2·1 to 3·4)* 136·0 (92·0–184·6) 14·4 (10·6 to 18·0)* 3·0 (–0·2 to 6·1) 7)* –1·3 (–2·8 to –0·0)* 37·3 (25·1–51·1) 18·6 (15·9 to 21·2)* –1·1 (–3·4 to 1·2) –10·6 (–12·4 to –9·0)* 195·8 (97·0–363·1) 6·5 (4·3 to 8·7)* –10·4 (–12·2 to –8·8)* )* –3·7 (–4·1 to –3·4)* 2387·1 (1653·2–3 263·5) 15·6 (14·7 to 16·4)* –3·6 (–4·2 to –3·0)* )* –2·4 (–4·1 to –0·6)* 11·0 (7·4–14·8) 23·4 (19·7 to 27·2)* –2·0 (–5·2 to 1·2) )* –4·1 (–5·9 to –2·2)* 601·0 (409·4–826·4) 16·8 (14·2 to 19·3)* –3·9 (–5·9 to –1·9)* )* 1·9 (–0·0 to 3·7) 300·0 (203·4–406·7) 24·0 (20·3 to 27·4)* 2·0 (–0·8 to 4·6) ·· 1219·9 (838·5–1677·5) 21·9 (14·4 to 32·4)* –1·0 (–7·1 to 7·4) 4)* 0·5 (–0·3 to 1·2) 59 325·6 (40 845·0–81 083·3) 15·6 (14·3 to 16·9)* –1·4 (–2·4 to –0·4)* 0)* 1·1 (0·3 to 1·8)* 6851·2 (4883·9–9062·8) 38·8 (37·1 to 40·5)* 1·1 (0·2 to 1·8)* (Table 3 continues on next page) Prevalence (thousands) YLDs (thousands) 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 2015 Percentage change in counts between 2005 and 2015 Percentage change i ASR between 2005 and 2015 (Continued from previous page) Severe interstitial lung disease and pulmonary sarcoidosis with heart failure aggregate 197·4 (144·1 to 242·7) 31·5 (29·7 to 33·2)* 1·3 (–0·1 to 2·9) 86·2 (54·8–117·6) 31·3 (28·6 to 34·1)* 1·3 (–1·0 to 3·7) Other chronic respiratory diseases ·· ·· ·· 1811·3 (1483·7–2123·4) 1·7 (–7·3 to 12·2) –15·8 (–23·5 to –6·9)* Cirrhosis and other chronic liver diseases 2828·3 (2791·3 to 2862·3)* 34·0 (32·4 to 35·8)* 8·2 (7·0 to 9·5)* 501·0 (352·9–683·1) 31·3 (29·8 to 32·9)* 7·3 (6·1 to 8·4)* Cirrhosis and other chronic liver diseases due to hepatitis B 796·5 (737·6 to 857·3) 30·5 (28·4 to 32·7)* 4·7 (3·2 to 6·2)* 130·6 (88·7–178·7) 30·5 (27·0 to 33·8)* 4·8 (2·2 to 7·3)* Cirrhosis and other chronic liver diseases due to hepatitis C 680·8 (628·3 to 734·2) 35·3 (33·0 to 37·5)* 7·6 (6·0 to 9·1)* 111·4 (78·2–152·6) 35·2 (31·6 to 38·6)* 7·6 (5·0 to 10·3)* Cirrhosis and other chronic liver diseases due to alcohol use 815·6 (758·7 to 877·2) 37·3 (35·4 to 39·3)* 9·8 (8·5 to 11·2)* 133·9 (92·5–185·1) 37·2 (33·9 to 40·5)* 9·9 (7·5 to 12·4)* Cirrhosis and other chronic liver diseases due to other causes 750·3 (714·3 to 785·9) 23·8 (22·5 to 25·1)* 6·8 (5·9 to 7·6)* 125·0 (87·5–172·2) 23·3 (20·1 to 26·8)* 6·8 (4·2 to 9·5)* Digestive diseases 258 248·2 (255 386·4 to 260 843·9) 12·3 (10·9 to 13·7)* –9·4 (–10·4 to –8·3)* 12 142·8 (8 492·0–16 592·5) 11·4 (9·8 to 13·2)* –9·3 (–10·6 to –7·9)* Peptic ulcer disease 72 044·2 (71 302·7 to 72 763·7) 4·8 (3·7 to 6·0)* –17·0 (–17·8 to –16·0)* 2341·7 (1605·6–3293·6) 1·1 (–0·2 to 2·5) –19·5 (–20·6 to –18·4)* Peptic ulcer disease symptomatic episodes 4977·7 (4577·6 to 5379·5) 4·7 (2·4 to 6·9)* –16·7 (–18·4 to –15·0)* 518·7 (357·9–703·2) 4·5 (2·2 to 6·8)* –16·7 (–18·5 to –14·9)* Anaemia due to peptic ulcer disease aggregate 67 066·6 (66 465·4 to 67 616·9) 4·8 (3·7 to 6·1)* –17·0 (–17·9 to –16·0)* 1823·1 (1225·8–2621·2) 0·2 (–1·4 to 1·9) –20·3 (–21·6 to –19·0)* Gastritis and duodenitis 161 001·1 (157 928·3 to 164 071·1) 15·5 (13·0 to 17·9)* –6·9 (–8·8 to –5·0)* 4913·0 (3347·3–6967·1) 11·2 (8·0 to 14·6)* –9·5 (–11·9 to –6·9)* Gastritis and duodenitis, symptomatic episodes 3941·0 (3547·0 to 4358·9) 15·4 (10·9 to 17·3)* –6·9 (–10·8 to –5·7)* 416·0 (277·1–567·9) 15·4 (11·2 to 17·5)* –6·7 (–10·6 to –5·2)* Anaemia due to gastritis and duodenitis aggregate 157 060·1 (154 055·1 to 160 141·0) 15·5 (12·9 to 17·9)* –6·9 (–8·9 to –5·0)* 4497·1 (3050·6–6393·0) 10·8 (7·4 to 14·5)* –9·8 (–12·4 to –7·0)* Appendicitis 442·1 (414·6 to 472·3) 14·3 (13·4 to 15·2)* 2·8 (2·1 to 3·4)* 136·0 (92·0–184·6) 14·4 (10·6 to 18·0)* 3·0 (–0·2 to 6·1) Paralytic ileus and intestinal obstruction 119·7 (110·7 to 130·4) 18·9 (16·9 to 20·7)* –1·3 (–2·8 to –0·0)* 37·3 (25·1–51·1) 18·6 (15·9 to 21·2)* –1·1 (–3·4 to 1·2) Inguinal, femoral, and abdominal hernia 18 476·7 (16 577·2 to 20 339·8) 6·5 (4·2 to 8·6)* –10·6 (–12·4 to –9·0)* 195·8 (97·0–363·1) 6·5 (4·3 to 8·7)* –10·4 (–12·2 to –8·8)* Inﬂ ammatory bowel disease 11 223·5 (10 396·8 to 12 000·4) 15·7 (15·0 to 16·3)* –3·7 (–4·1 to –3·4)* 2387·1 (1653·2–3 263·5) 15·6 (14·7 to 16·4)* –3·6 (–4·2 to –3·0)* Vascular intestinal disorders 35·1 (32·4 to 38·1) 23·3 (21·0 to 25·8)* –2·4 (–4·1 to –0·6)* 11·0 (7·4–14·8) 23·4 (19·7 to 27·2)* –2·0 (–5·2 to 1·2) Gallbladder and biliary diseases 5656·5 (5199·3 to 6168·1) 16·9 (14·5 to 19·2)* –4·1 (–5·9 to –2·2)* 601·0 (409·4–826·4) 16·8 (14·2 to 19·3)* –3·9 (–5·9 to –1·9)* Pancreatitis 1018·2 (941·6 to 1 097·4) 24·2 (21·7 to 26·5)* 1·9 (–0·0 to 3·7) 300·0 (203·4–406·7) 24·0 (20·3 to 27·4)* 2·0 (–0·8 to 4·6) Other digestive diseases ·· ·· ·· 1219·9 (838·5–1677·5) 21·9 (14·4 to 32·4)* –1·0 (–7·1 to 7·4) Neurological disorders 2 261 316·4 (2 204 366·6 to 2 316 282·6) 15·5 (14·6 to 16·4)* 0·5 (–0·3 to 1·2) 59 325·6 (40 845·0–81 083·3) 15·6 (14·3 to 16·9)* –1·4 (–2·4 to –0·4)* Alzheimer’s disease and other dementias 45 956·3 (40 178·5 to 52 655·9) 37·7 (36·2 to 39·0)* 1·1 (0·3 to 1·8)* 6851·2 (4883·9–9062·8) 38·8 (37·1 to 40·5)* 1·1 (0·2 to 1·8)* (Table 3 continues on next page nce (thousands) Percentage change in counts between 2005 and 2015 Percentag ASR betw 2015 97·4 o 242·7) 31·5 (29·7 to 33·2)* 1·3 (–0·1 to 2 ·· ·· ·· 28·3 to 2862·3)* 34·0 (32·4 to 35·8)* 8·2 (7·0 to 9· 96·5 o 857·3) 30·5 (28·4 to 32·7)* 4·7 (3·2 to 6·2 80·8 o 734·2) 35·3 (33·0 to 37·5)* 7·6 (6·0 to 9· 15·6 o 877·2) 37·3 (35·4 to 39·3)* 9·8 (8·5 to 11 50·3 o 785·9) 23·8 (22·5 to 25·1)* 6·8 (5·9 to 7·6 48·2 6·4 to 3·9) 12·3 (10·9 to 13·7)* –9·4 (–10·4 to 44·2 7 to 72 763·7) 4·8 (3·7 to 6·0)* –17·0 (–17·8 to 77·7 to 5379·5) 4·7 (2·4 to 6·9)* –16·7 (–18·4 to 66·6 ·4 to 67 616·9) 4·8 (3·7 to 6·1)* –17·0 (–17·9 to 01·1 8·3 to 164 071·1) 15·5 (13·0 to 17·9)* –6·9 (–8·8 to – 41·0 to 4358·9) 15·4 (10·9 to 17·3)* –6·9 (–10·8 to 60·1 5·1 to 160 141·0) 15·5 (12·9 to 17·9)* –6·9 (–8·9 to – 42·1 o 472·3) 14·3 (13·4 to 15·2)* 2·8 (2·1 to 3·4 19·7 o 130·4) 18·9 (16·9 to 20·7)* –1·3 (–2·8 to –0 76·7 2 to 20 339·8) 6·5 (4·2 to 8·6)* –10·6 (–12·4 to 23·5 ·8 to 12 000·4) 15·7 (15·0 to 16·3)* –3·7 (–4·1 to –3 35·1 38·1) 23·3 (21·0 to 25·8)* –2·4 (–4·1 to –0 56·5 to 6168·1) 16·9 (14·5 to 19·2)* –4·1 (–5·9 to – 18·2 o 1 097·4) 24·2 (21·7 to 26·5)* 1·9 (–0·0 to 3 ·· ·· ·· 16·4 366·6 to 82·6) 15·5 (14·6 to 16·4)* 0·5 (–0·3 to 1 56·3 5 to 52 655·9) 37·7 (36·2 to 39·0)* 1·1 (0·3 to 1·8 1566 www.thelancet.com Vol 388 October 8, 2016 Articles Prevalence (thousands) YLDs (thousands) 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 (Continued from previous page) Parkinson’s disease 6193·3 (5725·7 to 6777·2) 31·6 (28·2 to 35·5)* 0·6 (–1·9 to 3·5) 737·8 (516·1–990·1) 31·4 (27·8 to 35·3)* 0·7 (–2·1 to 3·7) Epilepsy 23 414·5 (21 549·6 to 25 419·4) 11·3 (7·0 to 15·5)* –1·1 (–5·0 to 2·7) 6286·9 (4315·8–8250·2) –6·4 (–11·5 to –1·2)* –16·3 (–20·9 to –11·6)* Multiple sclerosis 2012·0 (1865·7 to 2166·9) 19·1 (17·0 to 21·3)* –2·0 (–3·9 to –0·3)* 667·5 (472·0–855·6) 18·8 (16·2 to 21·5)* –2·0 (–4·2 to 0·1) Motor neuron disease 202·4 (189·7 to 216·0) 22·3 (19·9 to 24·8)* –1·3 (–3·3 to 0·7) 42·3 (30·5–55·1) 22·1 (19·5 to 24·8)* –1·4 (–3·5 to 0·8) Migraine 958 789·2 (872 109·0 to 1 055 630·6) 15·3 (14·0 to 16·6)* 0·6 (–0·3 to 1·5) 32 898·8 (20 303·6–48 883·2) 15·3 (14·0 to 16·6)* 0·8 (–0·1 to 1·8) Tension-type headache 1 505 892·3 (1 337 310·0 to 1 681 575·0) 15·3 (14·0 to 16·6)* 0·5 (–0·1 to 1·1) 2260·5 (1058·1–4169·7) 15·3 (14·0 to 16·7)* 0·6 (–0·1 to 1·3) Medication overuse headache 58 454·5 (50 834·9 to 67 363·9) 19·0 (15·4 to 22·7)* 0·4 (–2·4 to 3·3) 9164·7 (6094·5–13 078·0) 18·9 (15·4 to 22·8)* 0·6 (–2·3 to 3·5) Other neurological disorders 11·5 (9·1 to 14·6) 17·0 (14·8 to 19·6)* –0·2 (–0·9 to 0·4) 415·9 (301·2–547·5) 32·2 (27·5 to 36·3)* –2·1 (–5·3 to 0·9) Mental and substance use disorders 1 058 903·8 (1 038 544·9 to 1 080 413·8) 14·3 (13·6 to 14·9)* 0·3 (–0·3 to 0·9) 149 977·9 (108 716·1–193 130·8) 16·1 (15·5 to 16·8)* 1·1 (0·7 to 1·4)* Schizophrenia 23 383·0 (20 608·0 to 26 418·4) 19·5 (18·6 to 20·4)* 0·1 (–0·4 to 0·6) 15 020·5 (10 816·1–18 623·2) 19·5 (18·5 to 20·4)* 0·3 (–0·4 to 0·9) Alcohol use disorders 63 469·5 (57 507·8 to 69 863·5) 11·1 (8·9 to 13·5)* –4·6 (–6·5 to –2·4)* 6321·3 (4205·6–8985·8)* 11·1 (9·0 to 13·6)* –4·5 (–6·4 to –2·3)* Alcohol dependence aggregate 62 575·3 (56 621·7 to 68 944·3) 11·1 (8·9 to 13·5)* –4·6 (–6·6 to –2·5)* 6276·2 (4167·7–8929·7) 11·1 (8·9 to 13·6)* –4·5 (–6·5 to –2·3)* Fetal alcohol syndrome aggregate 894·2 (833·3 to 951·9) 11·1 (10·3 to 11·7)* –0·5 (–1·1 to 0·2) 45·1 (29·2–65·1) 10·7 (8·2 to 13·1)* –0·6 (–2·7 to 1·6) Drug use disorders 46 388·6 (45 252·4 to 47 446·6) 16·3 (15·0 to 17·6)* 3·9 (2·8 to 5·0)* 9849·4 (6959·0–12775·2) 23·6 (21·5 to 25·9)* 8·2 (6·2 to 10·2)* Opioid use disorders 16 746·5 (14 659·3 to 19 107·5) 23·3 (20·7 to 26·0)* 6·4 (4·2 to 8·6)* 6969·6 (4842·5–8999·8) 23·3 (20·6 to 26·2)* 6·5 (4·2 to 9·0)* Asymptomatic opioid dependence 2716·8 (1979·9 to 3704·4) 23·3 (20·7 to 26·0)* 6·4 (4·2 to 8·6)* 0·0 (0·0–0·0) 0·0 (0·0 to 0·0) 0·0 (0·0 to 0·0)* Symptomatic opioid dependence aggregate 14 029·7 (12 042·2 to 16 122·8) 23·3 (20·7 to 26·0)* 6·4 (4·2 to 8·6)* 6969·6 (4842·5–8999·8) 23·3 (20·6 to 26·2)* 6·5 (4·2 to 9·0) Cocaine use disorders 3846·3 (3401·6 to 4309·8) 31·1 (28·0 to 34·3)* 14·1 (12·0 to 16·4)* 521·5 (329·0–733·9) 30·6 (27·2 to 34·4)* 14·0 (11·4 to 17·0)* Asymptomatic cocaine dependence 1927·8 (1544·3 to 2357·8) 31·1 (28·0 to 34·3)* 14·1 (12·0 to 16·4)* 0·0 (0·0–0·0) 0·0 (0·0 to 0·0) 0·0 (0·0 to 0·0) Symptomatic cocaine dependence aggregate 1918·5 (1542·1 to 2341·9) 31·1 (28·0 to 34·3)* 14·1 (12·0 to 16·4)* 521·5 (329·0–733·9) 30·6 (27·2 to 34·4)* 14·0 (11·4 to 17·0)* Amphetamine use disorders 6599·8 (5295·7 to 8024·4) 30·4 (27·0 to 34·2)* 19·2 (16·1 to 22·5)* 874·7 (513·4–1308·7) 30·3 (26·1 to 34·8)* 19·1 (15·5 to 23·0)* Asymptomatic amphetamine dependence 3613·7 (2722·6 to 4630·7) 30·4 (27·0 to 34·2)* 19·2 (16·1 to 22·5)* 0·0 (0·0–0·0) 0·0 (0·0 to 0·0) 0·0 (0·0 to 0·0) Symptomatic amphetamine dependence aggregate 2986·1 (2186·5 to 3 893·0) 30·4 (27·0 to 34·2)* 19·2 (16·1 to 22·5)* 874·7 (513·4–1 308·7) 30·3 (26·1 to 34·8)* 19·1 (15·5 to 23·0)* Cannabis use disorders 19 762·5 (17 982·4 to 21 770·2) 5·3 (4·0 to 6·5)* –3·8 (–4·7 to –2·9)* 577·2 (371·8–816·2) 5·3 (3·7 to 7·1)* –3·7 (–5·0 to –2·3)* (Table 3 continues on next page) (Continued from previous page) Parkinson’s disease 6193·3 (5725·7 to 6777·2) Epilepsy 23 414·5 (21 549·6 to 25 419·4) Multiple sclerosis 2012·0 (1865·7 to 2166·9) Motor neuron disease 202·4 (189·7 to 216·0) Migraine 958 789·2 (872 109·0 to 1 055 630·6) Tension-type headache 1 505 892·3 (1 337 310·0 to 1 681 575·0) Medication overuse headache 58 454·5 (50 834·9 to 67 363·9) Other neurological disorders 11·5 (9·1 to 14·6) Mental and substance use disorders 1 058 903·8 (1 038 544·9 to 1 080 413·8) Schizophrenia 23 383·0 (20 608·0 to 26 418·4) Alcohol use disorders 63 469·5 (57 507·8 to 69 863·5) Alcohol dependence aggregate 62 575·3 (56 621·7 to 68 944·3) Fetal alcohol syndrome aggregate 894·2 (833·3 to 951·9) Drug use disorders 46 388·6 (45 252·4 to 47 446·6) Opioid use disorders 16 746·5 (14 659·3 to 19 107·5) Asymptomatic opioid dependence 2716·8 (1979·9 to 3704·4) Symptomatic opioid dependence aggregate 14 029·7 (12 042·2 to 16 122·8) Cocaine use disorders 3846·3 (3401·6 to 4309·8) Asymptomatic cocaine dependence 1927·8 (1544·3 to 2357·8) Symptomatic cocaine dependence aggregate 1918·5 (1542·1 to 2341·9) Amphetamine use disorders 6599·8 (5295·7 to 8024·4) Asymptomatic amphetamine dependence 3613·7 (2722·6 to 4630·7) Symptomatic amphetamine dependence aggregate 2986·1 (2186·5 to 3 893·0) Cannabis use disorders 19 762·5 (17 982·4 to 21 770·2) 19·1 (15·5 to 23·0)* (Table 3 continues on next page) www.thelancet.com Vol 388 October 8, 2016 1567 Articles Articles Articles YLDs (thousands) ange in en 2005 Percentage change in ASR between 2005 and 2015 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 –3·8 (–4·7 to –2·9)* 0·0 (0·0–0·0) 0·0 (0·0 to 0·0) 0·0 (0·0 to 0·0) –3·8 (–4·7 to –2·9)* 577·2 (371·8–816·2) 5·3 (3·7 to 7·1)* –3·7 (–5·0 to –2·3)* ·· 906·4 (550·1–1 312·1) 30·5 (25·7 to 35·6)* 17·2 (12·9 to 21·7)* * 0·7 (–0·3 to 1·6) 54 255·4 (37 569·9–72 943·3) 18·2 (17·2 to 19·2)* 1·0 (0·5 to 1·5)* * 0·8 (0·2 to 1·4)* 44 224·4 (29 672·5–60 297·0) 17·8 (16·6 to 19·0)* 1·1 (0·5 to 1·7)* * 0·6 (–0·3 to 1·4) 10 031·0 (6 605·9–14 267·1) 19·8 (18·3 to 21·5)* 0·7 (–0·2 to 1·6) * 0·3 (0·1 to 0·5)* 9004·7 (5501·7–13 388·0) 14·9 (13·9 to 15·9)* 0·5 (–0·0 to 0·9) * 0·8 (–0·5 to 2·1) 24 643·0 (16 813·7–33 647·8) 14·8 (12·8 to 16·6)* 1·0 (–0·4 to 2·3) * 9·9 (7·5 to 12·4)* 1386·1 (915·9–1941·2) 19·0 (16·9 to 21·2)* 10·0 (8·1 to 11·9)* 5·1 (2·6 to 7·6)* 620·5 (411·3–897·7) 12·1 (9·0 to 15·1)* 5·2 (2·3 to 8·0)* * 14·2 (12·5 to 15·8)* 765·6 (490·0–1108·5) 25·3 (23·2 to 27·6)* 14·2 (12·2 to 16·2)* * 0·3 (–1·2 to 1·9) 10 051·5 (6740·1–13800·3) 12·3 (11·9 to 12·8)* 0·6 (0·2 to 0·9)* * 0·5 (0·3 to 0·7)* 6335·9 (4112·0–8916·1) 12·5 (11·9 to 13·1)* 0·7 (0·2 to 1·2)* * 0·2 (0·1 to 0·3)* 3715·6 (2479·6–5424·6) 12·0 (11·5 to 12·5)* 0·3 (–0·1 to 0·7) –3·6 (–4·2 to –3·0)* 620·1 (370·1–947·7) 0·4 (–0·6 to 1·2) –3·5 (–4·4 to –2·8)* 1·3 (0·9 to 1·6)* 5770·5 (3485·2–8955·7) 0·5 (–0·3 to 1·4) 1·4 (0·7 to 2·0)* 1·0 (0·0 to 1·8)* 3442·1 (1506·4–5996·9) 10·3 (8·2 to 11·4)* 0·2 (–1·5 to 1·2) * 0·2 (–0·6 to 0·9) 9613·4 (6699·7–12 951·9) 18·7 (17·7 to 19·8)* 0·3 (–0·6 to 1·1) )* –0·3 (–0·9 to 0·3) 66 092·7 (46 781·5–88 865·8) 22·3 (20·8 to 24·0)* 1·2 (0·1 to 2·3)* * 4·5 (2·4 to 6·4)* 33 360·8 (23 043·5–45 530·8) 32·5 (29·7 to 35·3)* 5·4 (3·2 to 7·5)* )* 3·2 (1·1 to 5·3)* 12 599·0 (7979·9–18 394·2) 27·7 (24·8 to 30·3)* 3·3 (1·2 to 5·4)* * 6·6 (4·1 to 9·0)* 20 459·6 (13 572·0–28 749·6) 35·6 (32·3 to 38·8)* 6·7 (4·2 to 9·0)* * 5·0 (3·6 to 6·3)* 302·3 (210·1–415·5) 36·1 (33·6 to 38·7)* 5·7 (4·1 to 7·4)* –14·5 (–17·3 to –12·6)* 5·0 (3·2–7·5) –0·6 (–4·6 to 2·8) –14·1 (–17·4 to –11·4)* * 0·9 (–0·0 to 1·8) 8172·8 (6118·6–10 229·3) 23·8 (22·9 to 24·8)* 0·1 (–0·4 to 0·6) )* 2·1 (0·5 to 3·7)* 2456·1 (1780·4–3151·2) 23·0 (20·7 to 25·4)* 0·6 (–0·7 to 2·0) (Table 3 continues on next page Prevalence (thousands) YLDs (thousands) 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 (Continued from previous page) Asymptomatic cannabis dependence 11 430·9 (10 159·4 to 12 892·9) 5·3 (4·0 to 6·5)* –3·8 (–4·7 to –2·9)* 0·0 (0·0–0·0) 0·0 (0·0 to 0·0) 0·0 (0·0 to 0·0) Symptomatic cannabis dependence aggregate 8331·6 (7244·3 to 9593·1) 5·3 (4·0 to 6·5)* –3·8 (–4·7 to –2·9)* 577·2 (371·8–816·2) 5·3 (3·7 to 7·1)* –3·7 (–5·0 to –2·3)* Other drug use disorders ·· ·· ·· 906·4 (550·1–1 312·1) 30·5 (25·7 to 35·6)* 17·2 (12·9 to 21·7)* Depressive disorders 311 147·6 (300 016·8 to 320 544·3) 18·4 (17·2 to 19·5)* 0·7 (–0·3 to 1·6) 54 255·4 (37 569·9–72 943·3) 18·2 (17·2 to 19·2)* 1·0 (0·5 to 1·5)* Major depressive disorder 216 047·0 (192 863·4 to 243 319·4) 17·8 (16·6 to 19·0)* 0·8 (0·2 to 1·4)* 44 224·4 (29 672·5–60 297·0) 17·8 (16·6 to 19·0)* 1·1 (0·5 to 1·7)* Dysthymia 104 106·3 (90 398·1 to 118 968·9) 19·9 (18·4 to 21·5)* 0·6 (–0·3 to 1·4) 10 031·0 (6 605·9–14 267·1) 19·8 (18·3 to 21·5)* 0·7 (–0·2 to 1·6) Bipolar disorder 44 015·8 (38 150·4 to 50 912·5) 14·9 (14·1 to 15·8)* 0·3 (0·1 to 0·5)* 9004·7 (5501·7–13 388·0) 14·9 (13·9 to 15·9)* 0·5 (–0·0 to 0·9) Anxiety disorders 267 202·4 (234 064·3 to 306 318·0) 14·9 (13·0 to 16·8)* 0·8 (–0·5 to 2·1) 24 643·0 (16 813·7–33 647·8) 14·8 (12·8 to 16·6)* 1·0 (–0·4 to 2·3) Eating disorders 6468·4 (6170·6 to 6753·6) 19·1 (16·5 to 21·8)* 9·9 (7·5 to 12·4)* 1386·1 (915·9–1941·2) 19·0 (16·9 to 21·2)* 10·0 (8·1 to 11·9)* Anorexia nervosa 2912·1 (2357·7 to 3596·2) 12·0 (9·3 to 14·8)* 5·1 (2·6 to 7·6)* 620·5 (411·3–897·7) 12·1 (9·0 to 15·1)* 5·2 (2·3 to 8·0)* Bulimia nervosa 3629·4 (2937·5 to 4406·8) 25·3 (23·6 to 27·1)* 14·2 (12·5 to 15·8)* 765·6 (490·0–1108·5) 25·3 (23·2 to 27·6)* 14·2 (12·2 to 16·2)* Autistic spectrum disorders 62 212·4 (58 979·0 to 64 744·3) 12·3 (10·5 to 14·1)* 0·3 (–1·2 to 1·9) 10 051·5 (6740·1–13800·3) 12·3 (11·9 to 12·8)* 0·6 (0·2 to 0·9)* Autism 24 790·1 (20 957·6 to 29 393·1) 12·6 (12·2 to 12·9)* 0·5 (0·3 to 0·7)* 6335·9 (4112·0–8916·1) 12·5 (11·9 to 13·1)* 0·7 (0·2 to 1·2)* Asperger syndrome and other autistic spectrum disorders 37 245·4 (31 510·5 to 44 883·1) 12·1 (11·8 to 12·4)* 0·2 (0·1 to 0·3)* 3715·6 (2479·6–5424·6) 12·0 (11·5 to 12·5)* 0·3 (–0·1 to 0·7) Attention-deﬁ cit or hyperactivity disorder 51 094·3 (46 162·4 to 57 267·7) 0·3 (–0·4 to 1·0) –3·6 (–4·2 to –3·0)* 620·1 (370·1–947·7) 0·4 (–0·6 to 1·2) –3·5 (–4·4 to –2·8)* Conduct disorder 48 135·5 (39 315·2 to 58 151·6) 0·4 (–0·1 to 1·1) 1·3 (0·9 to 1·6)* 5770·5 (3485·2–8955·7) 0·5 (–0·3 to 1·4) 1·4 (0·7 to 2·0)* Idiopathic developmental intellectual disability 92 074·0 (52 280·1 to 130 411·2) 10·9 (9·5 to 11·9)* 1·0 (0·0 to 1·8)* 3442·1 (1506·4–5996·9) 10·3 (8·2 to 11·4)* 0·2 (–1·5 to 1·2) Other mental and substance use disorders 128 178·3 (127 512·5 to 128 877·2) 18·8 (17·9 to 19·7)* 0·2 (–0·6 to 0·9) 9613·4 (6699·7–12 951·9) 18·7 (17·7 to 19·8)* 0·3 (–0·6 to 1·1) Diabetes, urogenital, blood, and endocrine diseases 2 944 626·8 (2 925 016·2 to 2 968 511·3) 15·5 (14·7 to 16·1)* –0·3 (–0·9 to 0·3) 66 092·7 (46 781·5–88 865·8) 22·3 (20·8 to 24·0)* 1·2 (0·1 to 2·3)* Diabetes 435 328·4 (404 736·1 to 468 562·4) 30·6 (28·0 to 33·0)* 4·5 (2·4 to 6·4)* 33 360·8 (23 043·5–45 530·8) 32·5 (29·7 to 35·3)* 5·4 (3·2 to 7·5)* Uncomplicated diabetes 271 713·6 (241 414·2 to 302 381·7) 27·8 (24·9 to 30·4)* 3·2 (1·1 to 5·3)* 12 599·0 (7979·9–18 394·2) 27·7 (24·8 to 30·3)* 3·3 (1·2 to 5·4)* Neuropathy and other complications of diabetes aggregate 159 067·7 (134 666·7 to 187 174·0) 35·6 (32·2 to 38·8)* 6·6 (4·1 to 9·0)* 20 459·6 (13 572·0–28 749·6) 35·6 (32·3 to 38·8)* 6·7 (4·2 to 9·0)* Vision loss due to diabetes aggregate 4547·1 (3823·8 to 5416·0) 34·7 (32·4 to 36·9)* 5·0 (3·6 to 6·3)* 302·3 (210·1–415·5) 36·1 (33·6 to 38·7)* 5·7 (4·1 to 7·4)* Acute glomerulonephritis 100·5 (92·1 to 110·1) –0·9 (–4·4 to 1·9) –14·5 (–17·3 to –12·6)* 5·0 (3·2–7·5) –0·6 (–4·6 to 2·8) –14·1 (–17·4 to –11·4)* Chronic kidney disease 322 510·6 (312 718·2 to 330 351·1) 26·9 (25·8 to 28·0)* 0·9 (–0·0 to 1·8) 8172·8 (6118·6–10 229·3) 23·8 (22·9 to 24·8)* 0·1 (–0·4 to 0·6) Chronic kidney disease due to diabetes 100 823·9 (86 922·7 to 115 652·3) 27·3 (24·9 to 29·9)* 2·1 (0·5 to 3·7)* 2456·1 (1780·4–3151·2) 23·0 (20·7 to 25·4)* 0·6 (–0·7 to 2·0) (Table 3 continues on next page Prevalence (thousands) 2015 Percentage chang counts between 2 and 2015 (Continued from previous page) Asymptomatic cannabis dependence 11 430·9 (10 159·4 to 12 892·9) 5·3 (4·0 to 6·5)* Symptomatic cannabis dependence aggregate 8331·6 (7244·3 to 9593·1) 5·3 (4·0 to 6·5)* Other drug use disorders ·· ·· Depressive disorders 311 147·6 (300 016·8 to 320 544·3) 18·4 (17·2 to 19·5)* Major depressive disorder 216 047·0 (192 863·4 to 243 319·4) 17·8 (16·6 to 19·0)* Dysthymia 104 106·3 (90 398·1 to 118 968·9) 19·9 (18·4 to 21·5)* Bipolar disorder 44 015·8 (38 150·4 to 50 912·5) 14·9 (14·1 to 15·8)* Anxiety disorders 267 202·4 (234 064·3 to 306 318·0) 14·9 (13·0 to 16·8)* Eating disorders 6468·4 (6170·6 to 6753·6) 19·1 (16·5 to 21·8)* Anorexia nervosa 2912·1 (2357·7 to 3596·2) 12·0 (9·3 to 14·8)* Bulimia nervosa 3629·4 (2937·5 to 4406·8) 25·3 (23·6 to 27·1)* Autistic spectrum disorders 62 212·4 (58 979·0 to 64 744·3) 12·3 (10·5 to 14·1)* Autism 24 790·1 (20 957·6 to 29 393·1) 12·6 (12·2 to 12·9)* Asperger syndrome and other autistic spectrum disorders 37 245·4 (31 510·5 to 44 883·1) 12·1 (11·8 to 12·4)* Attention-deﬁ cit or hyperactivity disorder 51 094·3 (46 162·4 to 57 267·7) 0·3 (–0·4 to 1·0) Conduct disorder 48 135·5 (39 315·2 to 58 151·6) 0·4 (–0·1 to 1·1) Idiopathic developmental intellectual disability 92 074·0 (52 280·1 to 130 411·2) 10·9 (9·5 to 11·9)* Other mental and substance use disorders 128 178·3 (127 512·5 to 128 877·2) 18·8 (17·9 to 19·7)* Diabetes, urogenital, blood, and endocrine diseases 2 944 626·8 (2 925 016·2 to 2 968 511·3) 15·5 (14·7 to 16·1)* Diabetes 435 328·4 (404 736·1 to 468 562·4) 30·6 (28·0 to 33·0)* Uncomplicated diabetes 271 713·6 (241 414·2 to 302 381·7) 27·8 (24·9 to 30·4)* Neuropathy and other complications of diabetes aggregate 159 067·7 (134 666·7 to 187 174·0) 35·6 (32·2 to 38·8)* Vision loss due to diabetes aggregate 4547·1 (3823·8 to 5416·0) 34·7 (32·4 to 36·9)* Acute glomerulonephritis 100·5 (92·1 to 110·1) –0·9 (–4·4 to 1·9) Chronic kidney disease 322 510·6 (312 718·2 to 330 351·1) 26·9 (25·8 to 28·0)* Chronic kidney disease due to diabetes 100 823·9 (86 922·7 to 115 652·3) 27·3 (24·9 to 29·9)* –3·5 (–4·4 to –2·8)* 1·4 (0·7 to 2·0)* 0·2 (–1·5 to 1·2) 0·3 (–0·6 to 1·1) 1·2 (0·1 to 2·3)* (Table 3 continues on next page) 1568 www.thelancet.com Vol 388 October 8, 2016 Articles Articles Prevalence (thousands) YLDs (thousands) 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 ontinued from previous page) Stage 3 chronic kidney disease due to diabetes aggregate 90 937·7 (77 906·5 to 104 617·0) 27·9 (25·4 to 30·5)* 2·3 (0·6 to 4·0)* 119·6 (79·7–175·5) 11·8 (6·5 to 16·7)* –6·3 (–10·0 to –2·8)* Stage 4 chronic kidney disease due to diabetes aggregate 6441·3 (5521·3 to 7401·3) 21·6 (19·1 to 24·2)* 0·9 (–0·4 to 2·2) 787·7 (541·8–1 106·1) 20·2 (17·6 to 22·7)* 0·4 (–1·0 to 1·8) Stage 5 chronic kidney disease untreated due to diabetes 2694·4 (2331·9 to 3091·5) 25·6 (23·0 to 28·2)* 1·3 (–0·0 to 2·8) 1359·3 (920·5–1791·4) 25·3 (22·4 to 28·4)* 1·4 (–0·3 to 3·2) End-stage chronic kidney disease due to diabetes aggregate 750·5 (643·3 to 864·2) 19·8 (16·4 to 23·4)* –3·0 (–5·2 to –0·7)* 189·5 (130·5–252·0) 26·6 (22·3 to 30·9)* 0·7 (–2·1 to 3·7) Chronic kidney disease due to hypertension 78 962·3 (67 846·7 to 91 021·2) 26·0 (23·6 to 28·6)* 0·2 (–1·4 to 1·9) 1508·0 (1108·7–1938·1) 28·5 (25·8 to 31·5)* 1·2 (–0·6 to 3·3) Stage 3 chronic kidney disease due to hypertension aggregate 72 772·4 (61 662·5 to 84 498·2) 25·9 (23·3 to 28·6)* 0·2 (–1·6 to 2·0) 80·8 (54·2–120·1) 22·1 (17·5 to 26·6)* –4·4 (–8·0 to –0·9)* Stage 4 chronic kidney disease due to hypertension aggregate 4102·2 (3503·5 to 4772·6) 25·9 (22·4 to 29·5)* 0·7 (–1·6 to 2·8) 489·8 (341·4–696·7) 25·2 (22·1 to 28·7)* 0·5 (–1·5 to 2·5) Stage 5 chronic kidney disease untreated due to hypertension 1652·2 (1404·5 to 1913·7) 30·7 (27·5 to 34·2)* 1·8 (–0·5 to 4·1) 816·8 (565·2–1060·3) 30·5 (27·0 to 34·5)* 2·0 (–0·5 to 4·7) End-stage chronic kidney disease due to hypertension aggregate 435·5 (363·3 to 509·1) 27·5 (22·0 to 33·6)* 0·6 (–3·1 to 4·9) 120·6 (83·1–162·3) 33·2 (27·4 to 39·8)* 3·8 (–0·3 to 8·9) Chronic kidney disease due to glomerulonephritis 67 348·7 (58 198·9 to 77 439·5) 28·9 (25·8 to 32·3)* 1·1 (–1·1 to 3·6) 1951·4 (1432·2–2479·2) 22·2 (20·1 to 24·4)* –0·7 (–2·1 to 0·7) Stage 3 chronic kidney disease due to glomerulonephritis aggregate 59 790·8 (50 844·6 to 69 692·7) 29·9 (26·4 to 33·6)* 1·3 (–1·1 to 4·1) 111·3 (73·5–165·2) 17·1 (13·1 to 20·7)* –5·6 (–8·8 to –2·7)* Stage 4 chronic kidney disease due to glomerulonephritis aggregate 4872·4 (4297·2 to 5506·6) 21·3 (18·8 to 23·9)* –0·5 (–2·2 to 1·6) 602·5 (418·4–845·6) 20·2 (17·7 to 22·9)* –0·5 (–2·3 to 1·5) Stage 5 chronic kidney disease untreated due to glomerulonephritis 2228·1 (1895·7 to 2596·9) 23·6 (21·3 to 25·9)* –0·6 (–2·0 to 0·7) 1122·6 (773·7–1461·3) 23·5 (20·5 to 26·3)* –0·4 (–2·4 to 1·5) End-stage chronic kidney disease due to glomerulonephritis aggregate 457·4 (377·1 to 543·0) 20·5 (16·7 to 24·5)* –2·6 (–5·1 to –0·0)* 115·0 (80·2–152·9) 26·0 (21·1 to 31·2)* 0·4 (–3·2 to 4·1) Chronic kidney disease due to other causes 94 553·5 (81 141·8 to 109 371·3) 26·2 (24·2 to 28·3)* 0·1 (–1·1 to 1·5) 2257·2 (1663·4–2892·2) 23·0 (21·1 to 25·1)* –0·5 (–1·6 to 0·6) Stage 3 chronic kidney disease due to other causes aggregate 85 459·8 (72 471·6 to 99 605·4) 26·6 (24·6 to 28·8)* 0·1 (–1·2 to 1·6)* 132·2 (88·0–194·5) 17·6 (14·0 to 20·8)* –6·2 (–8·8 to –3·8)* Stage 4 chronic kidney disease due to other causes aggregate 6058·4 (5257·9 to 6938·4) 21·3 (18·9 to 23·5)* –0·2 (–1·1 to 0·8) 739·7 (513·5–1 029·6) 19·4 (17·0 to 21·6)* –1·0 (–2·2 to 0·2) Stage 5 chronic kidney disease untreated due to other causes 2464·6 (2115·1 to 2835·9) 25·7 (23·8 to 27·5)* 0·1 (–0·9 to 1·1) 1233·0 (855·9–1604·1) 25·5 (23·0 to 28·0)* 0·3 (–1·1 to 1·8) (Table 3 continues on next page) Prevalence (thousands) 2015 Percentage chang counts between 2 and 2015 (Continued from previous page) Stage 3 chronic kidney disease due to diabetes aggregate 90 937·7 (77 906·5 to 104 617·0) 27·9 (25·4 to 30·5)* Stage 4 chronic kidney disease due to diabetes aggregate 6441·3 (5521·3 to 7401·3) 21·6 (19·1 to 24·2)* Stage 5 chronic kidney disease untreated due to diabetes 2694·4 (2331·9 to 3091·5) 25·6 (23·0 to 28·2)* End-stage chronic kidney disease due to diabetes aggregate 750·5 (643·3 to 864·2) 19·8 (16·4 to 23·4)* Chronic kidney disease due to hypertension 78 962·3 (67 846·7 to 91 021·2) 26·0 (23·6 to 28·6)* Stage 3 chronic kidney disease due to hypertension aggregate 72 772·4 (61 662·5 to 84 498·2) 25·9 (23·3 to 28·6)* Stage 4 chronic kidney disease due to hypertension aggregate 4102·2 (3503·5 to 4772·6) 25·9 (22·4 to 29·5)* Stage 5 chronic kidney disease untreated due to hypertension 1652·2 (1404·5 to 1913·7) 30·7 (27·5 to 34·2)* End-stage chronic kidney disease due to hypertension aggregate 435·5 (363·3 to 509·1) 27·5 (22·0 to 33·6)* Chronic kidney disease due to glomerulonephritis 67 348·7 (58 198·9 to 77 439·5) 28·9 (25·8 to 32·3)* Stage 3 chronic kidney disease due to glomerulonephritis aggregate 59 790·8 (50 844·6 to 69 692·7) 29·9 (26·4 to 33·6)* Stage 4 chronic kidney disease due to glomerulonephritis aggregate 4872·4 (4297·2 to 5506·6) 21·3 (18·8 to 23·9)* Stage 5 chronic kidney disease untreated due to glomerulonephritis 2228·1 (1895·7 to 2596·9) 23·6 (21·3 to 25·9)* End-stage chronic kidney disease due to glomerulonephritis aggregate 457·4 (377·1 to 543·0) 20·5 (16·7 to 24·5)* Chronic kidney disease due to other causes 94 553·5 (81 141·8 to 109 371·3) 26·2 (24·2 to 28·3)* Stage 3 chronic kidney disease due to other causes aggregate 85 459·8 (72 471·6 to 99 605·4) 26·6 (24·6 to 28·8)* Stage 4 chronic kidney disease due to other causes aggregate 6058·4 (5257·9 to 6938·4) 21·3 (18·9 to 23·5)* Stage 5 chronic kidney disease untreated due to other causes 2464·6 (2115·1 to 2835·9) 25·7 (23·8 to 27·5)* 0·1 (–0·9 to 1·1) 25·5 (23·0 to 28·0)* 0·3 (–1·1 to 1·8) (Table 3 continues on next page) www.thelancet.com Vol 388 October 8, 2016 1569 Articles Articles (Table 3 continues on next page) Articles Prevalence (thousands) YLDs (thousands) 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 (Continued from previous page) End-stage chronic kidney disease due to other causes aggregate 570·6 (485·6 to 668·1) 21·6 (18·0 to 25·0)* –2·2 (–4·2 to –0·2)* 152·4 (107·3–200·8) 26·3 (21·8 to 30·7)* 0·1 (–3·0 to 3·2) Urinary diseases and male infertility 436 081·7 (425 577·7 to 446 450·5) 24·2 (22·8 to 25·4)* 0·9 (–0·1 to 1·9) 4220·9 (2699·9–6045·2) 28·9 (26·8 to 31·0)* –0·7 (–2·1 to 0·7) Interstitial nephritis and urinary tract infections 2901·2 (2831·9 to 2980·4) 19·3 (19·0 to 19·7)* 3·1 (2·9 to 3·3) 96·6 (59·9–144·0) 19·2 (17·5 to 20·9)* 3·3 (1·8 to 4·7)* Urolithiasis 319 600·2 (291 519·2 to 349 966·1) 22·8 (20·8 to 24·9)* 0·9 (–0·7 to 2·5) 89·7 (60·9–122·5) 23·4 (20·0 to 27·0)* 2·8 (0·1 to 5·5)* Benign prostatic hyperplasia 104 625·3 (90 729·8 to 118 244·0) 29·7 (27·5 to 32·0)* –1·3 (–2·8 to 0·2) 3792·7 (2423·8–5437·3) 29·8 (27·6 to 32·2)* –1·3 (–2·8 to 0·3) Male infertility 29 033·2 (23 972·5 to 34 766·8) 22·6 (19·3 to 25·6)* 9·6 (6·8 to 12·1)* 173·9 (70·1–365·3)* 22·7 (19·3 to 25·9)* 9·8 (6·9 to 12·5)* Other urinary diseases ·· ·· ·· 67·9 (45·5–96·5) 17·2 (9·1 to 27·2)* –0·6 (–7·3 to 7·7)* Gynaecological diseases 794 303·2 (786 497·9 to 801 169·7) 13·8 (13·2 to 14·5)* –1·8 (–2·4 to –1·2)* 10 001·2 (6807·0–14 312·3) 10·7 (9·4 to 11·9)* –3·3 (–4·3 to –2·5)* Uterine ﬁ broids 151 115·0 (144 146·9 to 158 477·5) 19·2 (18·8 to 19·5)* –0·4 (–0·6 to –0·2)* 2383·5 (1450·2–3834·4) 11·1 (8·6 to 13·2)* –5·8 (–7·7 to –4·2)* Uterine ﬁ broids cases aggregate 108 335·0 (101 618·4 to 115 233·6) 23·0 (22·3 to 23·7)* 1·7 (1·2 to 2·1)* 645·9 (314·9–1193·3) 26·1 (25·0 to 27·3)* 3·4 (2·7 to 4·3)* Mild abdominal pain with anaemia due to uterine ﬁ broids aggregate 42 780·0 (41 634·5 to 43 881·4) 10·5 (9·5 to 11·5)* –5·5 (–6·3 to –4·6)* 1737·5 (1125·7–2625·4) 6·4 (4·5 to 8·1)* –8·9 (–10·5 to –7·4)* Polycystic ovarian syndrome 60 106·4 (46 139·3 to 75 963·4) 10·7 (10·1 to 11·4)* –0·5 (–0·9 to –0·0)* 532·6 (238·9–1 044·6) 10·8 (9·9 to 11·5)* –0·5 (–1·1 to 0·0) Polycystic ovarian syndrome cases aggregate 47 116·8 (35 880·1 to 60 919·9) 10·4 (9·8 to 11·1)* –0·8 (–1·2 to –0·3)* 408·0 (181·1–797·7) 10·5 (9·7 to 11·3)* –0·7 (–1·2 to –0·1)* Hirsutism and infertility due to polycystic ovarian syndrome aggregate 12 989·6 (8 635·6 to 18 975·6) 11·9 (10·9 to 12·7)* 0·4 (–0·4 to 1·1) 124·6 (47·2–278·0) 11·5 (9·6 to 12·9)* 0·1 (–1·5 to 1·3) Female infertility 61 300·1 (45 889·9 to 79 647·1) 24·6 (19·2 to 30·0)* 11·2 (6·5 to 15·8)* 344·5 (134·3–746·7) 25·0 (19·7 to 30·0)* 11·7 (7·2 to 16·2)* Endometriosis 10 758·2 (9 165·5 to 12 485·9) 11·6 (10·9 to 12·2)* –2·6 (–3·0 to –2·2)* 996·4 (647·1–1 384·7) 11·7 (10·4 to 12·9)* –2·4 (–3·5 to –1·5)* Endometriosis cases aggregate 10 167·5 (8 652·4 to 11 828·8) 11·5 (10·9 to 12·2)* –2·7 (–3·1 to –2·3)* 937·3 (610·7–1 312·6) 11·6 (10·3 to 12·9)* –2·5 (–3·6 to –1·6)* Abdominal pain and infertility due to endometriosis aggregate 590·7 (415·2 to 787·2) 12·3 (11·2 to 13·3)* –1·0 (–1·8 to –0·3)* 59·1 (34·2–90·8) 12·3 (9·5 to 15·2)* –0·9 (–3·4 to 1·5) Genital prolapse 161 679·1 (142 334·7 to 182 566·5) 17·7 (15·2 to 20·0)* –6·6 (–8·5 to –4·8)* 503·0 (242·5–904·3) 17·8 (15·4 to 20·3)* –6·5 (–8·4 to –4·6)* Premenstrual syndrome 430 696·9 (410 840·7 to 450 494·4) 10·1 (7·9 to 11·8)* –2·1 (–3·9 to –0·7)* 3621·6 (2249·5–5428·5) 10·2 (8·0 to 12·0)* –2·0 (–3·8 to –0·4)* Other gynaecological diseases 45 018·8 (44 057·3 to 46 031·2) 8·0 (7·3 to 8·6)* –4·2 (–4·8 to –3·7)* 1619·8 (1129·8–2210·0) 6·2 (4·9 to 7·6)* –5·7 (–6·8 to –4·6)* Other gynaecological diseases cases aggregate 23 908·1 (22 975·7 to 24 918·0) 13·8 (13·5 to 14·0)* –0·4 (–0·5 to –0·2)* 936·8 (635·4–1 305·9) 13·8 (12·6 to 15·1)* –0·2 (–1·2 to 0·8) Anaemia due to other gynaecological diseases aggregate 21 110·8 (20 900·6 to 21 328·2) 2·1 (0·9 to 3·4)* –8·2 (–9·3 to –7·1)* 682·9 (461·0–978·0) –2·7 (–4·6 to –0·7)* –12·3 (–14·0 to –10·6)* (Table 3 continues on next page) Prevalence (thousands) 2015 Percentage chang counts between 2 and 2015 (Continued from previous page) End-stage chronic kidney disease due to other causes aggregate 570·6 (485·6 to 668·1) 21·6 (18·0 to 25·0)* Urinary diseases and male infertility 436 081·7 (425 577·7 to 446 450·5) 24·2 (22·8 to 25·4)* Interstitial nephritis and urinary tract infections 2901·2 (2831·9 to 2980·4) 19·3 (19·0 to 19·7)* Urolithiasis 319 600·2 (291 519·2 to 349 966·1) 22·8 (20·8 to 24·9)* Benign prostatic hyperplasia 104 625·3 (90 729·8 to 118 244·0) 29·7 (27·5 to 32·0)* Male infertility 29 033·2 (23 972·5 to 34 766·8) 22·6 (19·3 to 25·6)* Other urinary diseases ·· ·· Gynaecological diseases 794 303·2 (786 497·9 to 801 169·7) 13·8 (13·2 to 14·5)* Uterine ﬁ broids 151 115·0 (144 146·9 to 158 477·5) 19·2 (18·8 to 19·5)* Uterine ﬁ broids cases aggregate 108 335·0 (101 618·4 to 115 233·6) 23·0 (22·3 to 23·7)* Mild abdominal pain with anaemia due to uterine ﬁ broids aggregate 42 780·0 (41 634·5 to 43 881·4) 10·5 (9·5 to 11·5)* Polycystic ovarian syndrome 60 106·4 (46 139·3 to 75 963·4) 10·7 (10·1 to 11·4)* Polycystic ovarian syndrome cases aggregate 47 116·8 (35 880·1 to 60 919·9) 10·4 (9·8 to 11·1)* Hirsutism and infertility due to polycystic ovarian syndrome aggregate 12 989·6 (8 635·6 to 18 975·6) 11·9 (10·9 to 12·7)* Female infertility 61 300·1 (45 889·9 to 79 647·1) 24·6 (19·2 to 30·0)* Endometriosis 10 758·2 (9 165·5 to 12 485·9) 11·6 (10·9 to 12·2)* Endometriosis cases aggregate 10 167·5 (8 652·4 to 11 828·8) 11·5 (10·9 to 12·2)* Abdominal pain and infertility due to endometriosis aggregate 590·7 (415·2 to 787·2) 12·3 (11·2 to 13·3)* Genital prolapse 161 679·1 (142 334·7 to 182 566·5) 17·7 (15·2 to 20·0)* Premenstrual syndrome 430 696·9 (410 840·7 to 450 494·4) 10·1 (7·9 to 11·8)* Other gynaecological diseases 45 018·8 (44 057·3 to 46 031·2) 8·0 (7·3 to 8·6)* Other gynaecological diseases cases aggregate 23 908·1 (22 975·7 to 24 918·0) 13·8 (13·5 to 14·0)* Anaemia due to other gynaecological diseases aggregate 21 110·8 (20 900·6 to 21 328·2) 2·1 (0·9 to 3·4)* 682·9 (461·0–978·0) –2·7 (–4·6 to –0·7)* –12·3 (–14·0 to –10·6)* (Table 3 continues on next page) 1570 www.thelancet.com Vol 388 October 8, 2016 Articles Prevalence (thousands) YLDs (thousands) 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 (Continued from previous page) Haemoglobinopathies and haemolytic anaemias 1 571 610·0 (1 545 204·5 to 1 605 099·8) 11·1 (9·5 to 12·6)* –1·1 (–2·5 to 0·3) 8221·5 (5528·6–11 766·6) 4·3 (2·9 to 5·5)* –4·9 (–6·0 to –3·9)* Thalassaemias 439·0 (405·7 to 496·3) 1·4 (–0·1 to 3·4) –6·5 (–7·8 to –4·6)* 31·3 (21·2–44·4) 1·0 (–3·8 to 6·1) –7·0 (–11·4 to –2·5)* β-thalassaemia major cases aggregate 229·2 (222·7 to 238·2) 4·4 (3·6 to 5·2)* –4·2 (–4·9 to –3·4)* 16·1 (11·0–22·9) 4·1 (–2·5 to 11·5) –4·7 (–10·6 to 1·9) Haemoglobin E or β-thalassaemia cases aggregate 67·4 (60·1 to 75·2) 5·4 (2·9 to 8·1)* –2·5 (–4·8 to 0·1) 5·1 (3·4–7·3) 1·8 (–10·0 to 14·2) –5·9 (–16·7 to 4·9) Haemoglobin H disease cases aggregate 139·0 (107·5 to 196·4) –5·3 (–10·0 to 0·8) –11·8 (–16·2 to –6·3)* 9·7 (6·4–14·1) –4·9 (–12·3 to 4·9) –11·4 (–18·2 to –2·4)* Heart failure due to thalassaemias aggregate 3·4 (3·2 to 3·6) 26·3 (24·3 to 28·4)* 0·1 (–1·1 to 1·4)* 0·4 (0·3–0·6) 26·4 (16·0 to 38·3)* 0·6 (–9·3 to 11·8) Thalassaemias trait 279 451·4 (272 818·9 to 287 357·4) 10·5 (10·1 to 11·0)* –2·1 (–2·5 to –1·7)* 3922·2 (2607·9–5653·6) 6·1 (4·8 to 7·4)* –3·6 (–4·6 to –2·6)* β-thalassaemia trait cases aggregate 226 029·6 (220 111·5 to 233 332·0) 10·1 (9·6 to 10·5)* –2·5 (–3·0 to –2·1)* 3661·7 (2437·6–5270·8) 5·9 (4·7 to 7·2)* –3·8 (–4·9 to –2·8)* Haemoglobin E trait cases aggregate 53 421·8 (51 122·3 to 55 882·5) 12·6 (11·5 to 13·7)* –0·1 (–1·0 to 0·8) 260·5 (174·7–382·8) 8·4 (5·2 to 11·3)* –0·5 (–3·2 to 2·1) Sickle-cell disorders 4449·9 (4293·7 to 4600·7) 8·2 (5·2 to 11·9)* 2·8 (–0·2 to 6·2) 371·4 (259·0–518·0) 7·7 (3·3 to 12·3)* 2·3 (–1·8 to 6·6) Homozygous sickle- cell and severe sickle- cell/β-thalassaemia cases aggregate 4035·1 (3876·0 to 4173·7) 6·5 (3·3 to 10·3)* 1·2 (–1·8 to 4·8) 339·4 (237·7–472·1) 6·5 (2·0 to 11·4)* 1·3 (–2·9 to 5·9) Haemoglobin SC disease cases aggregate 397·6 (368·2 to 427·4) 29·6 (18·3 to 45·6)* 21·5 (11·0 to 36·5)* 30·1 (21·1–42·1) 22·9 (9·6 to 37·4)* 15·2 (2·8 to 28·9)* Mild sickle-cell/β- thalassaemia cases aggregate 17·2 (15·9 to 18·6) 16·1 (10·1 to 21·8)* 6·4 (0·8 to 11·5)* 1·9 (1·3–2·6) 13·2 (4·7 to 21·9)* 2·7 (–4·9 to 10·5) Sickle-cell trait 404 565·9 (381 223·4 to 448 154·6) 19·4 (18·3 to 20·3)* 7·5 (6·5 to 8·3)* 1720·3 (1156·5–2459·0) 10·8 (7·3 to 13·4)* 1·4 (–1·8 to 3·6) G6PD deﬁ ciency 247 073·7 (209 306·7 to 286 712·7) 6·9 (1·4 to 11·9)* –4·4 (–9·3 to 0·1) 28·4 (19·3–39·6) 1·8 (–2·1 to 6·0) –7·7 (–11·2 to –4·0)* G6PD cases aggregate 247 068·5 (209 301·6 to 286 707·7) 6·9 (1·4 to 11·9)* –4·4 (–9·3 to 0·1) 27·7 (18·9–38·7) 1·2 (–2·8 to 5·5) –8·1 (–11·6 to –4·3)* Heart failure due to G6PD deﬁ ciency aggregate 5·2 (4·7 to 5·7) 38·1 (36·6 to 39·8)* 7·9 (6·8 to 9·1)* 0·6 (0·4–0·9) 38·1 (34·6 to 41·5)* 7·8 (5·0 to 10·6)* G6PD trait 728 549·2 (676 734·8 to 781 533·8) 9·8 (7·9 to 11·4)* –2·7 (–4·3 to –1·2)* 28·0 (19·3–38·9) 4·3 (0·1 to 8·8)* –5·3 (–9·0 to –1·2)* Other haemoglobinopathies and haemolytic anaemias 74 385·1 (73 897·6 to 74 861·6) 1·8 (0·9 to 2·7)* –9·0 (–9·7 to –8·2)* 2119·8 (1423·9–3029·9) –3·8 (–5·5 to –2·2)* –12·3 (–13·8 to –11·0)* Other haemoglobinopathies and haemolytic anaemias cases 74 273·3 (73 786·0 to 74 751·1) 1·8 (0·9 to 2·7)* –9·0 (–9·8 to –8·2)* 2067·4 (1384·3–2963·9) –4·1 (–5·8 to –2·4)* –12·6 (–14·1 to –11·2)* –12·6 (–14·1 to –11·2)* –4·1 (–5·8 to –2·4)* www.thelancet.com Vol 388 October 8, 2016 1571 Articles Articles (Table 3 continues on next page) Articles Prevalence (thousands) YLDs (thousands) 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 (Continued from previous page) Heart failure due to other haemoglobinopathies and haemolytic anaemias aggregate 111·8 (103·8 to 119·6) 34·9 (33·9 to 36·0)* 2·0 (1·3 to 2·7)* 13·1 (8·9–18·2) 34·9 (31·2 to 38·7)* 2·1 (–0·8 to 5·0) Endocrine, metabolic, blood, and immune disorders 66 128·9 (65 597·1 to 66 663·8) 4·0 (2·2 to 5·9)* –7·2 (–8·7 to –5·5)* 2110·5 (1431·5–2993·7) 1·3 (–1·1 to 3·8) –8·3 (–10·4 to –6·1)* Endocrine metabolic blood and immune disorders cases aggregate 7577·2 (7382·6 to 7766·6) 19·8 (19·0 to 20·6)* –0·6 (–1·2 to 0·0) 292·4 (200·2–401·8) 19·5 (18·1 to 21·0)* –0·5 (–1·6 to 0·6) 58 430·4 (57 936·0–58 946·7) 2·2 (0·3–4·3)* –8·0 (–9·7––6·2)* 1804·4 (1216·0–2582·3) –1·3 (–3·9–1·5) –9·5 (–11·9––7·0)* Anaemia due to endocrine metabolic blood and immune disorders aggregate 58 430·4 (57 936·0 to 58 946·7) 2·2 (0·3 to 4·3)* –8·0 (–9·7 to –6·2)* 1804·4 (1216·0–2582·3) –1·3 (–3·9 to 1·5) –9·5 (–11·9 to –7·0)* Heart failure due to endocrine metabolic blood and immune disorders aggregate 121·4 (113·3 to 130·0) 30·7 (29·6 to 31·9)* –1·9 (–2·6 to –1·2)* 13·6 (9·4–19·1) 30·7 (27·4 to 34·0)* –1·7 (–4·3 to 0·8) Musculoskeletal disorders 1 304 100·4 (1 288 602·8 to 1 316 641·4) 20·7 (20·2 to 21·2)* –0·7 (–1·1 to –0·3)* 146 783·8 (106 764·7–194 473·5) 20·5 (19·6 to 21·5)* –0·7 (–1·3 to –0·0)* Rheumatoid arthritis 24 491·2 (22 552·0 to 26 750·7) 23·8 (21·1 to 26·7)* 0·6 (–1·5 to 2·9) 5777·8 (4016·1–7769·6) 23·6 (20·9 to 26·7)* 0·7 (–1·4 to 3·1)* Osteoarthritis 237 368·6 (230 335·9 to 244 648·1) 32·9 (31·9 to 33·8)* 2·2 (1·6 to 2·9)* 12 886·2 (8 999·7–17 540·0) 34·8 (33·6 to 36·0)* 3·9 (3·0 to 4·8)* Osteoarthritis of the hip cases aggregate 35 629·2 (32 482·6 to 38 970·1) 33·5 (32·4 to 34·6)* 1·8 (1·0 to 2·6)* 1776·2 (1224·9–2477·0) 35·8 (34·4 to 37·3)* 3·7 (2·6 to 5·0)* Osteoarthritis of the knee cases aggregate 201 739·4 (195 205·3 to 208 276·6) 32·7 (31·7 to 33·9)* 2·3 (1·5 to 3·1)* 11 110·0 (7742·1–15 123·2) 34·6 (33·3 to 35·9)* 3·9 (3·0 to 4·9)* Low back and neck pain 820 689·8 (803 467·4 to 837 808·9) 18·7 (17·9 to 19·4)* –2·0 (–2·6 to –1·4)* 94 941·5 (67 825·3–128 035·0) 18·6 (17·6 to 19·6)* –2·1 (–2·7 to –1·4)* Low back pain 539 907·4 (521 448·6 to 559 556·0) 17·3 (16·5 to 18·2)* –2·8 (–3·4 to –2·2)* 60 074·8 (42 721·9–82 343·7) 17·2 (16·4 to 18·1)* –2·6 (–3·2 to –2·0)* Neck pain 358 006·6 (313 408·4 to 409 411·0) 21·1 (19·0 to 23·3)* –1·1 (–2·4 to 0·0) 34 866·7 (23 362·1–47 636·9) 21·0 (18·9 to 23·2)* –1·1 (–2·3 to 0·1) Gout 42 214·2 (37 688·2 to 47 495·5) 26·4 (25·2 to 27·7)* 0·5 (0·1 to 1·0)* 1342·8 (910·0–1843·8) 26·3 (24·6 to 27·9)* 0·6 (–0·3 to 1·5) Asymptomatic gout 37 712·9 (33 598·4 to 42 467·0) 26·4 (25·2 to 27·7)* 0·5 (0·1 to 1·0)* 0·0 (0·0–0·0) 0·0 (0·0 to 0·0) 0·0 (0·0 to 0·0) Other musculoskeletal disorders 342 067·7 (305 430·7 to 385 146·7) 20·7 (17·5 to 24·0)* 1·2 (–1·2 to 3·7) 31 835·4 (21 489·1–44 268·4) 20·5 (17·2 to 23·7)* 1·3 (–1·1 to 3·8) Other non-communicable diseases 5 316 342·1 (5 283 701·8 to 5 355 928·3) 14·6 (14·3 to 14·8)* 0·4 (0·2 to 0·6)* 139 001·8 (95 459·0–197 704·3) 20·4 (19·5 to 21·5)* 1·2 (0·6 to 2·0)* Congenital anomalies 95 706·4 (88 133·4 to 102 483·4) 22·5 (18·5 to 26·3)* 9·3 (5·8 to 12·8)* 8621·0 (5389·1–12 950·2) 28·5 (20·0 to 37·6)* 14·7 (7·1 to 22·7)* Neural tube defects 1449·6 (988·4 to 2091·8) 17·7 (14·5 to 20·3)* 7·2 (4·3 to 9·7)* 499·1 (285·3–823·7) 18·4 (14·1 to 22·2)* 8·1 (4·1 to 11·6)* Congenital heart anomalies 48 869·1 (34 325·7 to 72 829·2) 29·8 (25·3 to 34·1)* 15·4 (11·4 to 19·2)* 1688·2 (659·4–3116·9) 29·7 (25·6 to 33·7)* 15·6 (11·7 to 19·1)* Less severe heart anomalies cases aggregate 44 599·5 (29 800·6 to 68 498·5) 29·4 (24·5 to 34·0)* 15·2 (10·7 to 19·1)* 1431·0 (532·7–2747·7) 29·3 (24·2 to 34·0)* 15·2 (10·7 to 19·2)* (Table 3 continues on next page) 1·2 6 to 2·0)* 1·2 to 2·0 1·2 (0·6 to 2·0)* 1572 www.thelancet.com Vol 388 October 8, 2016 Articles Prevalence (thousands) YLDs (thousands) 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 (Continued from previous page) Severe congenital heart anomalies 4008·6 (2671·1 to 6011·2) 33·7 (27·4 to 39·0)* 18·5 (12·9 to 23·3)* 235·0 (93·2–467·8) 33·5 (27·4 to 38·7)* 18·5 (13·2 to 23·2)* Critical congenital heart anomalies 155·8 (90·3 to 252·4) 42·0 (27·5 to 52·9)* 28·6 (15·5 to 38·5)* 9·2 (3·2–18·7) 42·0 (26·4 to 54·2)* 28·9 (14·9 to 40·1)* Heart failure due to congenital heart anomalies aggregate 105·2 (99·2 to 111·4) 11·6 (10·9 to 12·4)* –1·0 (–1·5 to –0·5)* 12·9 (8·8–18·2) 11·6 (8·3 to 15·1)* –1·0 (–3·9 to 2·0) Cleft lip and cleft palate 6882·6 (4029·2 to 11 483·5) 19·5 (7·7 to 27·5)* 6·6 (–3·9 to 13·9) 79·6 (39·5–143·1) 12·2 (2·6 to 19·9)* 0·9 (–7·9 to 7·7) Down syndrome 5361·8 (3424·6 to 8192·1) 17·9 (13·7 to 21·8)* 6·4 (2·3 to 10·0)* 507·0 (286·2–837·2) 20·3 (15·8 to 24·2)* 6·5 (2·2 to 10·2)* Turner syndrome 372·4 (174·6 to 677·1) 11·9 (9·8 to 13·9)* 0·1 (–1·7 to 1·9) 6·6 (2·4–14·4) 11·4 (8·7 to 14·0)* –0·1 (–2·4 to 2·2) Klinefelter syndrome 220·5 (115·5 to 395·5) 12·8 (10·9 to 14·2)* 0·5 (–1·2 to 1·9) 1·3 (0·5–3·1) 13·5 (10·6 to 15·6)* 0·2 (–2·2 to 2·0) Other chromosomal abnormalities 4728·6 (2628·9 to 8303·2) 16·2 (11·8 to 19·8)* 4·3 (0·1 to 7·7)* 454·1 (233·5–852·1) 18·5 (13·6 to 22·3)* 4·2 (–0·5 to 7·9) Other congenital anomalies 32 363·0 (17 788·1 to 59 394·1) 14·5 (12·3 to 16·8)* 2·0 (–0·1 to 4·1) 5385·0 (3209·4–8562·1) 31·3 (18·8 to 45·9)* 17·1 (6·1 to 30·3)* Skin and subcutaneous diseases 2 239 493·4 (2 222 716·3 to 2 258 252·8) 12·5 (12·1 to 12·8)* 0·7 (0·4 to 1·1)* 44 896·0 (28 943·3–67 159·6) 11·7 (11·0 to 12·3)* 0·4 (0·1 to 0·7)* Dermatitis 245 290·6 (227 283·4 to 262 752·2) 14·0 (13·2 to 14·8)* 0·3 (0·1 to 0·7)* 8788·0 (5963·4–12 273·5) 13·6 (12·9 to 14·4)* 1·6 (1·0 to 2·2)* Eczema cases aggregate 85 585·6 (75 115·2 to 97 031·7) 13·0 (12·4 to 13·6)* 3·3 (2·8 to 3·8)* 5239·2 (3515·1 to 7348·6) 13·0 (12·2 13·9)* 3·4 (2·7 to 4·2)* Contact dermatitis cases aggregate 98 015·1 (85 097·0 to 110 396·7) 15·2 (13·3 to 16·9)* –1·0 (–1·2 to –0·9)* 2632·0 (1625·8–3852·4) 15·0 (13·1 to 16·8)* –0·9 (–1·4 to –0·5)* Seborrhoeic dermatitis cases aggregate 61 690·0 (54 642·8 to 68 788·8) 13·3 (12·0 to 14·7)* –1·3 (–1·7 to –1·0)* 916·8 (521·6–1454·1) 13·2 (11·8 to 14·7)* –1·3 (–1·8 to –0·8)* Psoriasis 79 699·7 (76 690·9 to 82 804·5) 17·6 (17·0 to 18·3)* 0·4 (–0·1 to 0·9) 6438·3 (4495·6–8734·5) 17·5 (16·6 to 18·4)* 0·6 (–0·0 to 1·2) Cellulitis 959·9 (887·6 to 1033·6) 24·0 (22·1 to 25·7)* 5·8 (4·3 to 7·2)* 69·0 (45·0–97·8) 23·6 (20·7 to 26·1)* 5·8 (3·5 to 8·1)* Pyoderma 5812·7 (5581·7 to 6019·0) 15·5 (14·7 to 16·3)* 1·4 (0·7 to 2·1)* 32·7 (13·2–68·4) 15·4 (14·0 to 16·8)* 1·5 (0·3 to 2·7)* Scabies 204 151·7 (177 533·7 to 237 466·2) 6·6 (4·0 to 9·5)* –2·6 (–4·5 to –0·6)* 5268·9 (2966·9–8605·6) 6·6 (3·8 to 9·5)* –2·5 (–4·5 to –0·5)* Fungal skin diseases 492 372·6 (448 950·9 to 538 232·5) 13·3 (12·5 to 14·1)* 1·6 (1·3 to 1·9)* 2783·3 (1105·6–5905·3) 13·3 (12·4 to 14·1)* 1·7 (1·4 to 2·0)* Viral skin diseases 174 843·1 (165 156·3 to 185 072·0) 8·5 (8·0 to 9·0)* –1·1 (–1·4 to –0·8)* 5396·9 (3417·1–7959·9) 8·4 (7·9 to 9·0)* –1·0 (–1·4 to –0·6)* Molluscum contagiosum cases aggregate 40 464·2 (35 685·6 to 46 233·4) 5·2 (4·7 to 5·8)* –0·9 (–1·3 to –0·5)* 1263·0 (783·0–1932·3) 5·3 (4·6 to 6·1)* –0·8 (–1·5 to –0·2)* Viral warts cases aggregate 134 378·9 (125 683·9 to 143 028·0) 9·5 (8·9 to 10·1)* –1·1 (–1·4 to –0·7)* 4133·9 (2628·4–6153·4) 9·4 (8·8 to 10·1)* –1·0 (–1·5 to –0·6)* Acne vulgaris 632 741·0 (595 241·9 to 671 248·9) 4·6 (3·5 to 5·6)* 0·8 (–0·1 to 1·7) 6854·0 (3275·1–12 713·9) 4·6 (3·5 to 5·6)* 0·8 (–0·1 to 1·8) Alopecia areata 20 594·9 (19 607·8 to 21 617·0) 14·0 (13·5 to 14·4)* –1·1 (–1·3 to –1·0)* 695·6 (428·5–1035·6) 13·9 (13·1 to 14·7)* –1·0 (–1·7 to –0·4)* Pruritus 69 582·5 (62 063·2 to 78 152·6) 17·6 (15·7 to 19·5)* 0·4 (–0·7 to 1·6) 741·6 (360·0–1348·6) 17·5 (15·6 to 19·4)* 0·5 (–0·7 to 1·7) Urticaria 67 749·8 (58 743·2 to 77 086·4) 10·7 (9·6 to 12·0)* –0·1 (–0·3 to 0·0) 4115·7 (2567·2–5833·6) 10·7 (9·6 to 12·0)* –0·0 (–0·5 to 0·4) (Table 3 continues on next page) Prevalence (thousands) 2015 Percentage change counts between 20 and 2015 (Continued from previous page) Severe congenital heart anomalies 4008·6 (2671·1 to 6011·2) 33·7 (27·4 to 39·0)* Critical congenital heart anomalies 155·8 (90·3 to 252·4) 42·0 (27·5 to 52·9)* Heart failure due to congenital heart anomalies aggregate 105·2 (99·2 to 111·4) 11·6 (10·9 to 12·4)* Cleft lip and cleft palate 6882·6 (4029·2 to 11 483·5) 19·5 (7·7 to 27·5)* Down syndrome 5361·8 (3424·6 to 8192·1) 17·9 (13·7 to 21·8)* Turner syndrome 372·4 (174·6 to 677·1) 11·9 (9·8 to 13·9)* Klinefelter syndrome 220·5 (115·5 to 395·5) 12·8 (10·9 to 14·2)* Other chromosomal abnormalities 4728·6 (2628·9 to 8303·2) 16·2 (11·8 to 19·8)* Other congenital anomalies 32 363·0 (17 788·1 to 59 394·1) 14·5 (12·3 to 16·8)* Skin and subcutaneous diseases 2 239 493·4 (2 222 716·3 to 2 258 252·8) 12·5 (12·1 to 12·8)* Dermatitis 245 290·6 (227 283·4 to 262 752·2) 14·0 (13·2 to 14·8)* Eczema cases aggregate 85 585·6 (75 115·2 to 97 031·7) 13·0 (12·4 to 13·6)* Contact dermatitis cases aggregate 98 015·1 (85 097·0 to 110 396·7) 15·2 (13·3 to 16·9)* Seborrhoeic dermatitis cases aggregate 61 690·0 (54 642·8 to 68 788·8) 13·3 (12·0 to 14·7)* Psoriasis 79 699·7 (76 690·9 to 82 804·5) 17·6 (17·0 to 18·3)* Cellulitis 959·9 (887·6 to 1033·6) 24·0 (22·1 to 25·7)* Pyoderma 5812·7 (5581·7 to 6019·0) 15·5 (14·7 to 16·3)* Scabies 204 151·7 (177 533·7 to 237 466·2) 6·6 (4·0 to 9·5)* Fungal skin diseases 492 372·6 (448 950·9 to 538 232·5) 13·3 (12·5 to 14·1)* Viral skin diseases 174 843·1 (165 156·3 to 185 072·0) 8·5 (8·0 to 9·0)* Molluscum contagiosum cases aggregate 40 464·2 (35 685·6 to 46 233·4) 5·2 (4·7 to 5·8)* Viral warts cases aggregate 134 378·9 (125 683·9 to 143 028·0) 9·5 (8·9 to 10·1)* Acne vulgaris 632 741·0 (595 241·9 to 671 248·9) 4·6 (3·5 to 5·6)* Alopecia areata 20 594·9 (19 607·8 to 21 617·0) 14·0 (13·5 to 14·4)* Pruritus 69 582·5 (62 063·2 to 78 152·6) 17·6 (15·7 to 19·5)* Urticaria 67 749·8 (58 743·2 to 77 086·4) 10·7 (9·6 to 12·0)* (Table 3 continues on next page) www.thelancet.com Vol 388 October 8, 2016 1573 Articles Articles Articles Prevalence (thousands) YLDs (thousands) 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 (Continued from previous page) Forces of nature, war, and legal intervention 88 991·7 (65 131·0 to 116 786·1) 28·8 (18·4 to 40·5)* 11·7 (2·9 to 21·7)* 2820·2 (1921·3 to 3850·1) 23·4 (1·7 to 47·8)* 7·2 (–11·4 to 28·3) Exposure to forces of nature 25 293·7 (16 193·6 to 37 658·7) 13·5 (–2·1 to 32·2) –1·3 (–14·5 to 14·1) 796·9 (541·4 to 1091·4) –27·4 (–43·2 to –6·2)* –35·9 (–49·6 to –17·8)* Collective violence and legal intervention 63 697·9 (45 248·7–83 186·4) 36·1 (20·9–54·9)* 18·1 (5·1–34·4)* 2023·3 (1303·8–2903·0) 70·5 (33·0–118·3)* 47·1 (15·3–88·8)* Impairments ·· ·· ·· ·· ·· ·· Anaemia 2 359 107·9 (2 349 938·3 to 2 369 065·8) 4·0 (3·5 to 4·5)* –7·1 (–7·6 to –6·7)* 77 876·7 (52 436·9 to 111 360·0) –0·7 (–2·0 to 0·4) –10·2 (–11·2 to –9·2)* Developmental intellectual disability 152 664·0 (113 511·5 to 190 764·1) 12·6 (12·2 to 13·0)* 2·0 (1·6 to 2·4)* 16 875·1 (12 072·6 to 22 373·1) 15·0 (13·4 to 17·0)* 3·8 (2·2 to 5·7)* Epilepsy 39 160·5 (34 270·8 to 43 685·8) 15·8 (12·6 to 18·9)* 3·5 (0·5 to 6·4)* 11 770·0 (8 578·8 to 15 276·3) 5·8 (0·5 to 11·0)* –4·9 (–9·8 to –0·1)* Guillain-Barré syndrome 35·0 (28·6 to 42·1) 17·0 (14·7 to 19·5)* –0·2 (–0·9 to 0·4) 10·4 (6·6 to 15·0) 17·0 (14·7 to 19·5) –0·2 (–0·9 to 0·4) Hearing loss 1 330 902·9 (1 261 401·0 to 1 397 100·6) 26·0 (25·2 to 26·7)* 0·9 (0·5 to 1·3)* 46 183·1 (31 566·7 to 63 209·9) 23·6 (22·1 to 25·1)* 0·6 (–0·4 to 1·5) Heart failure 40 048·6 (38 613·2 to 41 418·2) 32·3 (31·6 to 33·1)* 0·3 (–0·2 to 0·8) 6199·4 (4700·7 to 7785·5) 33·2 (31·7 to 34·6)* 1·1 (0·1 to 2·0)* Infertility 113 113·8 (93 429·4–136 497·5) 21·1 (17·6–24·5)* 8·0 (4·9–10·9)* 752·4 (327·7–1543·8) 19·9 (16·8–22·7)* 7·0 (4·3–9·4)* Pelvic inﬂ ammatory disease 754·1 (652·8 to 878·7) 1·1 (0·1 to 2·1)* –12·0 (–12·7 to –11·2)* 99·9 (67·8 to 138·6) 1·3 (–0·7 to 3·4) –11·7 (–13·4 to –10·0)* Vision loss 939 580·2 (905 009·5 to 974 112·3) 22·4 (21·2 to 23·6)* –0·4 (–1·3 to 0·4) 24 462·7 (16 954·9 to 34 455·1) 22·4 (21·6 to 23·3)* –0·2 (–0·8 to 0·3) Data in parentheses are 95% UIs. Articles Percentage changes that are statistically signiﬁ cant. Table 3: Global prevalence and YLDs for 2015, percentage change of counts, and percentage change of age-standardised rates between 2005 and 2015 for all causes, Level 5 sequelae, and nine impairments Prevalence (thousands) Data in parentheses are 95% UIs. Percentage changes that are statistically signiﬁ cant. Table 3: Global prevalence and YLDs for 2015, percentage change of counts, and percentage change of age-standardised rates between 2005 and 2015 for all causes, Level 5 sequelae, and nine impairments region. Observed YLDs due to diabetes consistently exceeded expected levels, with two countries posting ratios higher than 3·00 (ie, Qatar [3·12], and the United Arab Emirates [3·52]). Iran and Morocco recorded much higher disability due to drug use disorders than expected based on SDI. disability in Nepal, Bangladesh, and Bhutan, respectively. YLDs due to anxiety were lower than expected in all countries in the region except for Bangladesh. In the GBD super-region of central Europe, eastern Europe, and central Asia, lower back and neck pain was the leading cause of disability in every geographical region in 2015; Although all countries in eastern and central Europe recorded higher than expected YLD ratios for most of their top ten causes, central Asia mostly had lower than expected or as expected YLD ratios. Sense organ diseases were the second leading causes of disability and depressive disorders were the third leading causes in this super-region. In comparison with the rest of the world, ranks of cause- speciﬁ c disability, and their ratios of observed to expected values, were vastly diﬀ erent in sub-Saharan Africa. Of the 46 countries within the super-regions, nine had lower back and neck pain as the leading cause of YLDs in 2015. In southern Sub-Saharan Africa, HIV/AIDS was the leading cause of disability for all countries. Iron-deﬁ ciency anaemia ranked as the leading cause of YLDs for 11 countries in western sub-Saharan Africa. For the remaining countries, lower back and neck pain were primarily the leading cause of YLDs; Liberia, the exception, had onchocerciasis as its leading cause of disability. Malaria and various neglected tropical diseases caused more YLDs than expected in most west African countries. www.thelancet.com Vol 388 October 8, 2016 Articles YLDs (thousands) 5 Percentage change in ASR between 2005 and 2015 2015 Percentage change in counts between 2005 and 2015 Percentage change i ASR between 2005 and 2015 4·7 (3·3 to 6·0)* 161·3 (112·0–218·7) 34·5 (31·9 to 37·3)* 4·7 (2·6 to 6·9)* 2·4 (1·9 to 2·9)* 3550·5 (1706·2–6514·4) 22·6 (22·0 to 23·3)* 2·5 (2·0 to 2·9)* 0·6 (0·3 to 0·9)* 68 515·2 (47 798·2–93 894·8) 25·2 (24·2 to 26·4)* 0·6 (–0·0 to 1·3) 4·2 (2·8 to 5·5)* 541·3 (370·0–747·9) 39·4 (37·0 to 41·9)* 4·3 (2·6 to 5·8)* –3·2 (–4·5 to –2·0)* 3879·7 (2766·9–5229·2) 25·5 (24·1 to 26·9)* –4·2 (–5·2 to –3·2)* 8·2 (6·9 to 9·8)* 462·4 (327·4–633·0) 47·7 (45·2 to 49·9)* 6·9 (5·3 to 8·6)* –0·4 (–1·3 to 0·6) 14 593·8 (9392·9–22 901·1) 21·1 (20·2 to 22·1)* –0·3 (–0·9 to 0·4) 1·4 (1·0 to 1·8)* 40 596·8 (27 898·4–56 075·0) 26·4 (24·6 to 28·3)* 1·1 (0·0 to 2·1)* 5·5 (4·2 to 6·7)* 1 756·4 (1248·9–2392·7) 29·9 (27·8 to 32·0)* 5·6 (4·3 to 6·7)* 0·8 (0·1 to 1·4)* 6684·7 (4185·0–9714·7) 23·8 (22·9 to 24·8)* 0·9 (0·2 to 1·5)* 0·9 (0·4 to 1·4)* 35·4 (21·6–51·6) 14·8 (12·7 to 16·9)* 0·9 (–0·9 to 2·7) 0·8 (0·1 to 1·4)* 6649·3 (4162·3–9662·4) 23·9 (23·0 to 24·9)* 0·9 (0·2 to 1·5)* 0·4 (–0·1 to 0·9) 16 969·6 (10 296·5–26 044·5) 22·4 (21·6 to 23·2)* –0·2 (–0·5 to 0·1) –2·3 (–4·2 to –0·7)* 147·2 (63·0–292·1) 4·1 (1·6 to 5·9)* –2·7 (–4·9 to –0·8)* 0·8 (0·3 to 1·4)* 1743·4 (776·7–3320·9) 13·5 (12·6 to 14·4)* –0·4 (–1·2 to 0·4) 1·1 (0·5 to 1·7)* 3520·7 (1359·3–7253·7) 25·4 (24·1 to 26·5)* 1·2 (0·6 to 1·8)* –0·9 (–1·1 to –0·7)* 7640·3 (5096·6–10 562·3) 27·3 (26·9 to 27·7)* –0·8 (–1·1 to –0·6)* –0·1 (–0·2 to 0·1) 3918·1 (2432·7–5879·4) 15·8 (15·3 to 16·4)* 0·0 (–0·2 to 0·3) –3·4 (–4·1 to –2·6)* 41 022·0 (29 290·4–55 288·4) 8·0 (4·6 to 11·1)* –9·9 (–12·5 to –7·5)* –0·0 (–0·4 to 0·4) 6444·8 (4500·6–8733·0) 12·2 (8·7 to 15·8)* –8·0 (–10·7 to –5·3)* 0·3 (–0·1 to 0·7) 5956·1 (4130·8–8076·3) 13·3 (9·9 to 16·7)* –7·1 (–9·7 to –4·6)* –0·7 (–1·4 to 0·0) 770·8 (532·9–1060·8) 9·6 (5·5 to 13·6)* –10·6 (–13·7 to –7·6)* –0·4 (–1·2 to 0·4) 968·2 (655·2–1348·5) 9·9 (5·5 to 14·2)* –9·0 (–12·5 to –5·9)* 7·3 (6·5 to 8·1)* 1397·9 (962·0–1919·8) 17·6 (13·0 to 22·1)* –3·3 (–7·0 to 0·2) –4·1 (–4·7 to –3·4)* 2585·4 (1811·2–3556·8) 12·2 (9·6 to 14·8)* –8·4 (–10·4 to –6·6)* (Table 3 continues on next pag Prevalence (thousands) YLDs (thousands) 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 (Continued from previous page) Decubitus ulcer 1087·5 (991·2 to 1189·7) 35·4 (33·4 to 37·4)* 4·7 (3·3 to 6·0)* 161·3 (112·0–218·7) 34·5 (31·9 to 37·3)* 4·7 (2·6 to 6·9)* Other skin and subcutaneous diseases 605 036·3 (589 500·0 to 619 675·6) 22·8 (22·1 to 23·4)* 2·4 (1·9 to 2·9)* 3550·5 (1706·2–6514·4) 22·6 (22·0 to 23·3)* 2·5 (2·0 to 2·9)* Sense organ diseases 1 788 125·6 (1 771 367·0 to 1 808 438·8) 23·7 (23·3 to 24·1)* 0·6 (0·3 to 0·9)* 68 515·2 (47 798·2–93 894·8) 25·2 (24·2 to 26·4)* 0·6 (–0·0 to 1·3) Glaucoma 5954·5 (5077·5 to 6905·8) 39·1 (36·9 to 41·4)* 4·2 (2·8 to 5·5)* 541·3 (370·0–747·9) 39·4 (37·0 to 41·9)* 4·3 (2·6 to 5·8)* Cataract 59 727·0 (53 633·2 to 66 879·2) 26·9 (25·2 to 28·5)* –3·2 (–4·5 to –2·0)* 3879·7 (2766·9–5229·2) 25·5 (24·1 to 26·9)* –4·2 (–5·2 to –3·2)* Macular degeneration 6188·4 (5250·3 to 7273·2) 48·5 (46·1 to 50·8)* 8·2 (6·9 to 9·8)* 462·4 (327·4–633·0) 47·7 (45·2 to 49·9)* 6·9 (5·3 to 8·6)* Refraction and accommodation disorders 819 307·4 (789 917·4 to 848 059·2) 21·9 (20·5 to 23·2)* –0·4 (–1·3 to 0·6) 14 593·8 (9392·9–22 901·1) 21·1 (20·2 to 22·1)* –0·3 (–0·9 to 0·4) Age-related and other hearing loss 1 210 055·2 (1 140 224·2 to 1 274 664·8) 28·3 (27·4 to 29·2)* 1·4 (1·0 to 1·8)* 40 596·8 (27 898·4–56 075·0) 26·4 (24·6 to 28·3)* 1·1 (0·0 to 2·1)* Other vision loss 25 797·4 (22 675·6 to 28 504·6) 27·1 (24·6 to 29·5)* 5·5 (4·2 to 6·7)* 1 756·4 (1248·9–2392·7) 29·9 (27·8 to 32·0)* 5·6 (4·3 to 6·7)* Other sense organ diseases 267 689·7 (258 379·8 to 277 711·4) 24·0 (23·1 to 24·9)* 0·8 (0·1 to 1·4)* 6684·7 (4185·0–9714·7) 23·8 (22·9 to 24·8)* 0·9 (0·2 to 1·5)* Acute other sense organ diseases 1343·8 (1303·5 to 1385·9) 14·9 (14·2 to 15·7)* 0·9 (0·4 to 1·4)* 35·4 (21·6–51·6) 14·8 (12·7 to 16·9)* 0·9 (–0·9 to 2·7) Chronic other sense organ diseases 266 345·9 (257 047·4 to 276 383·0) 24·0 (23·1 to 24·9)* 0·8 (0·1 to 1·4)* 6649·3 (4162·3–9662·4) 23·9 (23·0 to 24·9)* 0·9 (0·2 to 1·5)* Oral disorders 3 521 901·1 (3 468 088·2 to 3 575 854·4) 14·5 (13·8 to 15·0)* 0·4 (–0·1 to 0·9) 16 969·6 (10 296·5–26 044·5) 22·4 (21·6 to 23·2)* –0·2 (–0·5 to 0·1) Deciduous caries 572 694·1 (475 089·8 to 686 991·2) 4·5 (2·3 to 6·1)* –2·3 (–4·2 to –0·7)* 147·2 (63·0–292·1) 4·1 (1·6 to 5·9)* –2·7 (–4·9 to –0·8)* Permanent caries 2 521 197·8 (2 361 418·3 to 2 679 668·7) 14·5 (13·7 to 15·4)* 0·8 (0·3 to 1·4)* 1743·4 (776·7–3320·9) 13·5 (12·6 to 14·4)* –0·4 (–1·2 to 0·4) Periodontal diseases 537 506·0 (465 113·9 to 625 888·8) 25·4 (24·1 to 26·5)* 1·1 (0·5 to 1·7)* 3520·7 (1359·3–7253·7) 25·4 (24·1 to 26·5)* 1·2 (0·6 to 1·8)* Edentulism and severe tooth loss 275 619·1 (264 200·8 to 288 252·3) 27·3 (26·9 to 27·7)* –0·9 (–1·1 to –0·7)* 7640·3 (5096·6–10 562·3) 27·3 (26·9 to 27·7)* –0·8 (–1·1 to –0·6)* Other oral disorders 133 719·0 (127 499·7 to 139 776·2) 15·9 (15·4 to 16·4)* –0·1 (–0·2 to 0·1) 3918·1 (2432·7–5879·4) 15·8 (15·3 to 16·4)* 0·0 (–0·2 to 0·3) Injuries 1 019 157·4 (977 405·7 to 1 064 704·2) 16·4 (15·5 to 17·4)* –3·4 (–4·1 to –2·6)* 41 022·0 (29 290·4–55 288·4) 8·0 (4·6 to 11·1)* –9·9 (–12·5 to –7·5)* Transport injuries 115 981·0 (110 575·9 to 123 047·1) 23·1 (22·5 to 23·7)* –0·0 (–0·4 to 0·4) 6444·8 (4500·6–8733·0) 12·2 (8·7 to 15·8)* –8·0 (–10·7 to –5·3)* Road injuries 107 722·4 (102 214·5 to 114 850·0) 23·4 (22·8 to 24·1)* 0·3 (–0·1 to 0·7) 5956·1 (4130·8–8076·3) 13·3 (9·9 to 16·7)* –7·1 (–9·7 to –4·6)* Pedestrian road injuries 14 157·8 (12 664·5 to 16 095·9) 22·9 (21·7 to 24·1)* –0·7 (–1·4 to 0·0) 770·8 (532·9–1060·8) 9·6 (5·5 to 13·6)* –10·6 (–13·7 to –7·6)* Cyclist road injuries 18 700·6 (16 545·4 to 20 951·1) 21·4 (19·9 to 22·7)* –0·4 (–1·2 to 0·4) 968·2 (655·2–1348·5) 9·9 (5·5 to 14·2)* –9·0 (–12·5 to –5·9)* Motorcyclist road injuries 26 417·2 (23 496·7 to 29 722·0) 31·7 (30·3 to 33·0)* 7·3 (6·5 to 8·1)* 1397·9 (962·0–1919·8) 17·6 (13·0 to 22·1)* –3·3 (–7·0 to 0·2) Motor vehicle road 44311 4 18 7 4 1 2585 4 12 2 8 4 Prevalence (thousands) 2015 Percentage chang counts between 2 and 2015 (Continued from previous page) Decubitus ulcer 1087·5 (991·2 to 1189·7) 35·4 (33·4 to 37·4)* Other skin and subcutaneous diseases 605 036·3 (589 500·0 to 619 675·6) 22·8 (22·1 to 23·4)* Sense organ diseases 1 788 125·6 (1 771 367·0 to 1 808 438·8) 23·7 (23·3 to 24·1)* Glaucoma 5954·5 (5077·5 to 6905·8) 39·1 (36·9 to 41·4)* Cataract 59 727·0 (53 633·2 to 66 879·2) 26·9 (25·2 to 28·5)* Macular degeneration 6188·4 (5250·3 to 7273·2) 48·5 (46·1 to 50·8)* Refraction and accommodation disorders 819 307·4 (789 917·4 to 848 059·2) 21·9 (20·5 to 23·2)* Age-related and other hearing loss 1 210 055·2 (1 140 224·2 to 1 274 664·8) 28·3 (27·4 to 29·2)* Other vision loss 25 797·4 (22 675·6 to 28 504·6) 27·1 (24·6 to 29·5)* Other sense organ diseases 267 689·7 (258 379·8 to 277 711·4) 24·0 (23·1 to 24·9)* Acute other sense organ diseases 1343·8 (1303·5 to 1385·9) 14·9 (14·2 to 15·7)* Chronic other sense organ diseases 266 345·9 (257 047·4 to 276 383·0) 24·0 (23·1 to 24·9)* Oral disorders 3 521 901·1 (3 468 088·2 to 3 575 854·4) 14·5 (13·8 to 15·0)* Deciduous caries 572 694·1 (475 089·8 to 686 991·2) 4·5 (2·3 to 6·1)* Permanent caries 2 521 197·8 (2 361 418·3 to 2 679 668·7) 14·5 (13·7 to 15·4)* Periodontal diseases 537 506·0 (465 113·9 to 625 888·8) 25·4 (24·1 to 26·5)* Edentulism and severe tooth loss 275 619·1 (264 200·8 to 288 252·3) 27·3 (26·9 to 27·7)* Other oral disorders 133 719·0 (127 499·7 to 139 776·2) 15·9 (15·4 to 16·4)* Injuries 1 019 157·4 (977 405·7 to 1 064 704·2) 16·4 (15·5 to 17·4)* Transport injuries 115 981·0 (110 575·9 to 123 047·1) 23·1 (22·5 to 23·7)* Road injuries 107 722·4 (102 214·5 to 114 850·0) 23·4 (22·8 to 24·1)* Pedestrian road injuries 14 157·8 (12 664·5 to 16 095·9) 22·9 (21·7 to 24·1)* Cyclist road injuries 18 700·6 (16 545·4 to 20 951·1) 21·4 (19·9 to 22·7)* Motorcyclist road injuries 26 417·2 (23 496·7 to 29 722·0) 31·7 (30·3 to 33·0)* Motor vehicle road injuries 44 311·4 (40 212·6 to 49 496·3) 18·7 (17·7 to 19·8)* Prevalence (thousands) www.thelancet.com Vol 388 October 8, 2016 1574 Articles Prevalence (thousands) YLDs (thousands) 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 2015 Percentage change in counts between 2005 and 2015 Percentage change in ASR between 2005 and 2015 (Continued from previous page) Other road injuries 4135·3 (3603·7 to 4692·7) 40·2 (38·6 to 42·1)* 15·2 (14·4 to 16·1)* 233·7 (158·3–320·9) 32·1 (28·3 to 35·5)* 9·5 (6·8 to 11·9)* Other transport injuries 8258·5 (7469·0 to 9141·6) 18·8 (17·9 to 19·7)* –3·2 (–3·6 to –2·7)* 488·7 (348·6–660·5) 0·3 (–4·0 to 5·2) –17·1 (–20·5 to –13·3)* Unintentional injuries 790 101·8 (759 587·4 to 824 995·0) 14·3 (13·8 to 14·8)* –5·2 (–5·5 to –4·9)* 30 679·5 (21 602·9–41 953·7) 6·2 (3·1 to 9·0)* –11·4 (–13·6 to –9·2)* Falls 225 736·9 (207 914·3 to 245 787·2) 25·1 (24·4 to 25·8)* 1·7 (1·3 to 2·1)* 11 770·2 (8257·6–16 166·4) 11·3 (6·7 to 15·7)* –8·6 (–12·1 to –5·2)* Drowning 4455·5 (4041·2 to 4919·4) 7·0 (5·8 to 8·1)* –13·1 (–13·5 to –12·7)* 247·3 (173·4–335·8) –6·7 (–10·3 to –2·4)* –23·4 (–26·2 to –20·2)* Fire, heat, and hot substances 69 666·6 (60 988·1 to 78 550·1) 10·8 (10·1 to 11·7)* –8·3 (–8·8 to –7·9)* 2269·4 (1571·4–3131·8) 1·9 (–1·8 to 5·5) –13·5 (–16·1 to –11·2)* Poisonings 5722·4 (4829·3 to 6854·1) 13·8 (12·7 to 14·9)* –2·0 (–3·2 to –0·6)* 549·4 (367·8–791·1) 8·8 (6·7 to 10·8)* –4·5 (–6·6 to –2·6)* Exposure to mechanical forces 183 582·4 (169 658·6 to 197 859·6) 13·9 (13·0 to 14·7)* –3·5 (–4·0 to –2·9)* 3580·8 (2485·7–4922·9) 5·9 (2·9 to 8·9)* –9·9 (–12·1 to –7·6)* Unintentional ﬁ rearm injuries 2333·0 (2074·5 to 2625·1) 13·4 (12·4 to 14·4)* –6·3 (–7·1 to –5·6)* 93·2 (65·4–126·2) 4·1 (0·6 to 7·4)* –12·6 (–15·0 to –10·1)* Unintentional suﬀ ocation 9759·3 (8142·5 to 11 916·9) 9·7 (5·5 to 13·9)* –7·7 (–10·3 to –5·1)* 353·7 (245·8–480·3) 11·4 (8·6 to 14·0)* –6·8 (–8·6 to –5·1)* Other exposure to mechanical forces 171 490·1 (157 821·0 to 185 364·4) 14·1 (13·2 to 14·9)* –3·2 (–3·7 to –2·6)* 3133·9 (2172·1–4356·6) 5·4 (2·2 to 8·5)* –10·1 (–12·6 to –7·7)* Adverse eﬀ ects of medical treatment 11 158·6 (8773·5 to 13 610·2) 3·1 (1·3 to 4·9)* –13·9 (–15·4 to –12·4)* 1487·4 (926·9–2223·0) 3·1 (1·3 to 4·9)* –13·9 (–15·4 to –12·4)* Animal contact 34 049·7 (30 774·2 to 37 569·0) 9·4 (8·0 to 10·6)* –6·6 (–7·5 to –5·9)* 1100·9 (763·8–1486·9) 4·6 (3·0 to 6·2)* –8·8 (–10·0 to –7·5)* Venomous animal contact 14 680·7 (13 173·4 to 16 375·1) 13·2 (12·2 to 14·2)* –3·6 (–4·4 to –2·8)* 820·7 (558·1–1117·5) 5·3 (3·7 to 7·1)* –7·8 (–9·2 to –6·2)* Non-venomous animal contact 19 369·1 (16 794·5 to 22 306·5) 6·7 (4·8 to 8·2)* –8·8 (–10·1 to –7·7)* 280·3 (187·2–410·5) 2·5 (0·2 to 4·7)* –11·4 (–13·1 to –9·8)* Foreign body 32 345·9 (28 877·3 to 36 026·4) 15·7 (14·1 to 17·1)* –1·9 (–2·9 to –1·1)* 1263·3 (899·4–1722·5) 3·7 (0·2 to 7·5)* –10·9 (–13·6 to –8·1)* Pulmonary aspiration and foreign body in airway 13 351·2 (11 069·0 to 16 798·7) 15·5 (12·5 to 18·4)* –1·6 (–3·6 to 0·2) 690·2 (477·9–973·1) –0·2 (–4·1 to 4·5) –14·0 (–17·2 to –10·4)* Foreign body in eyes 1007·2 (444·0 to 1 593·4) 16·3 (15·0 to 18·2)* 1·4 (0·8 to 1·9)* 56·2 (24·0–100·6) 15·2 (13·8 to 16·7)* 0·5 (–0·9 to 1·3) Foreign body in other body part 17 987·4 (15 874·1 to 20 237·9) 15·9 (14·9 to 17·0)* –2·3 (–3·0 to –1·6)* 517·0 (364·4–713·3) 8·0 (5·5 to 10·6)* –7·7 (–9·6 to –5·8)* Environmental heat and cold exposure 61 942·3 (56 211·9 to 68 761·8) 15·9 (15·4 to 16·5)* –4·2 (–4·6 to –3·8)* 2657·1 (1861·7–3636·2) 6·4 (3·7 to 9·4)* –11·0 (–13·0 to –8·9)* Other unintentional injuries 161 441·6 (147 971·3 to 176 204·3) 4·7 (4·0 to 5·3)* –13·7 (–14·1 to –13·3)* 5753·6 (3913·1–8024·1) 0·4 (–1·4 to 2·1) –16·7 (–17·8 to –15·7)* Self-harm and interpersonal violence 24 083·0 (22 279·7 to 25 744·4) 15·2 (14·6 to 15·7)* –5·5 (–5·9 to –5·2)* 1077·5 (762·7–1 448·2) 1·9 (–1·6 to 5·8) –15·1 (–17·8 to –12·1)* Self-harm 6358·6 (5682·1 to 7154·5) 15·2 (14·5 to 15·8)* –6·2 (–6·7 to –5·8)* 332·7 (232·4–446·9) 0·9 (–2·1 to 4·5) –15·9 (–18·4 to –13·2)* Interpersonal violence 17 724·4 (16 186·6 to 19 268·4) 15·2 (14·5 to 15·9)* –5·3 (–5·7 to –4·8)* 744·8 (521·9–1007·3) 2·4 (–1·4 to 6·4) –14·6 (–17·5 to –11·6)* Assault by ﬁ rearm 1043·6 (900·8 to 1213·2) 16·1 (14·9 to 17·3)* –4·9 (–5·8 to –4·2)* 46·5 (32·4–64·9) 4·6 (0·7 to 8·1)* –13·2 (–16·1 to –10·5)* Assault by sharp object 4019·5 (3330·4 to 4668·1) 12·5 (11·5 to 13·6)* –7·0 (–7·7 to –6·3)* 123·4 (84·8–168·7) –6·4 (–11·4 to –0·9)* –21·1 (–24·8 to –16·8)* Assault by other means 12 661·3 (11 425·5 to 14 018·9) 16·0 (15·2 to 16·8)* –4·7 (–5·3 to –4·3)* 574·8 (404·0–782·2) 4·3 (0·8 to 8·0)* –13·3 (–15·9 to –10·4)* (Table 3 continues on next page) Prevalence (thousands) 2015 Percentage change counts between 20 and 2015 (Continued from previous page) Other road injuries 4135·3 (3603·7 to 4692·7) 40·2 (38·6 to 42·1)* Other transport injuries 8258·5 (7469·0 to 9141·6) 18·8 (17·9 to 19·7)* Unintentional injuries 790 101·8 (759 587·4 to 824 995·0) 14·3 (13·8 to 14·8)* Falls 225 736·9 (207 914·3 to 245 787·2) 25·1 (24·4 to 25·8)* Drowning 4455·5 (4041·2 to 4919·4) 7·0 (5·8 to 8·1)* Fire, heat, and hot substances 69 666·6 (60 988·1 to 78 550·1) 10·8 (10·1 to 11·7)* Poisonings 5722·4 (4829·3 to 6854·1) 13·8 (12·7 to 14·9)* Exposure to mechanical forces 183 582·4 (169 658·6 to 197 859·6) 13·9 (13·0 to 14·7)* Unintentional ﬁ rearm injuries 2333·0 (2074·5 to 2625·1) 13·4 (12·4 to 14·4)* Unintentional suﬀ ocation 9759·3 (8142·5 to 11 916·9) 9·7 (5·5 to 13·9)* Other exposure to mechanical forces 171 490·1 (157 821·0 to 185 364·4) 14·1 (13·2 to 14·9)* Adverse eﬀ ects of medical treatment 11 158·6 (8773·5 to 13 610·2) 3·1 (1·3 to 4·9)* Animal contact 34 049·7 (30 774·2 to 37 569·0) 9·4 (8·0 to 10·6)* Venomous animal contact 14 680·7 (13 173·4 to 16 375·1) 13·2 (12·2 to 14·2)* Non-venomous animal contact 19 369·1 (16 794·5 to 22 306·5) 6·7 (4·8 to 8·2)* Foreign body 32 345·9 (28 877·3 to 36 026·4) 15·7 (14·1 to 17·1)* Pulmonary aspiration and foreign body in airway 13 351·2 (11 069·0 to 16 798·7) 15·5 (12·5 to 18·4)* Foreign body in eyes 1007·2 (444·0 to 1 593·4) 16·3 (15·0 to 18·2)* Foreign body in other body part 17 987·4 (15 874·1 to 20 237·9) 15·9 (14·9 to 17·0)* Environmental heat and cold exposure 61 942·3 (56 211·9 to 68 761·8) 15·9 (15·4 to 16·5)* Other unintentional injuries 161 441·6 (147 971·3 to 176 204·3) 4·7 (4·0 to 5·3)* Self-harm and interpersonal violence 24 083·0 (22 279·7 to 25 744·4) 15·2 (14·6 to 15·7)* Self-harm 6358·6 (5682·1 to 7154·5) 15·2 (14·5 to 15·8)* Interpersonal violence 17 724·4 (16 186·6 to 19 268·4) 15·2 (14·5 to 15·9)* Assault by ﬁ rearm 1043·6 (900·8 to 1213·2) 16·1 (14·9 to 17·3)* Assault by sharp object 4019·5 (3330·4 to 4668·1) 12·5 (11·5 to 13·6)* Assault by other means 12 661·3 (11 425·5 to 14 018·9) 16·0 (15·2 to 16·8)* www.thelancet.com Vol 388 October 8, 2016 1575 Articles Articles Discussion We used 60 900 data sources to estimate the incidence and prevalence of 2619 sequelae of 310 causes for 591 geographical regions for the years 1990, 1995, 2000, 2005, 2010, and 2015. After accounting for variation over time and across regions in case deﬁ nitions, disease assays, survey instruments, and coding practice, and using standardised modelling approaches to deal with missing data, conﬂ icting data, and a range of sampling and non- sampling errors, we characterised the global and national patterns in non-fatal health outcomes. We found that global age-standardised YLDs per capita decreased slightly in the last 25 years from 0·114 (0·085–0·147) in 1990 to 0·110 (0·082–0·141) per capita in 2015, a total decrease of 3·69% (3·12–4·25%) over a generation (appendix pp 714–16). The ageing of the world’s population and the general increase in YLDs per capita with age resulted in an increase in global YLDs per capita. Within this overall pattern, 133 out of 310 causes had statistically signiﬁ cant decreases in YLDs per capita between 2005 and 2015 compared with 82 causes with statistically signiﬁ cant increases in YLDs per capita over the same time period. The most important contributors to global YLDs were musculoskeletal disorders (18·5% [16·4–20·9%] of all YLDs in 2015), mental and substance use disorders (18·4% [15·6–21·2%]), and the category of other NCDs (17·9% [15·0–21·7%]), dominated by hearing loss, vision loss, and skin diseases. Disability and retirement age As life expectancy has steadily increased in most countries, there have been calls in some high SDI countries to extend retirement ages to reﬂ ect these changes in survival.28–30 A crucial factor in these debates is whether individuals are living for longer and have higher average levels of functional health at each age or not. In this study, we found that age-standardised YLDs per capita decreased, but rather slowly, over the period 2005 to 2015. Comparison of trends in age-standardised YLDs and YLLs by cause shows that the main causes of YLDs are not decreasing and in some settings might be increasing. Other than a somewhat slower rate of improvement for YLDs at oldest age (≥80 years), age discontinuities are not observable in these patterns. Ultimately, the debate on retirement age will hinge on the skill sets required for diﬀ erent types of work, whether work contributes to diminished or improved functional health status, and societal expectations for retirement. Articles For eastern sub-Saharan Africa, sense organ diseases, iron deﬁ ciency, depressive disorders, and lower back and neck pain were leading causes of YLDs g p g Sizeable discrepancies occurred for observed and expected YLDs based on SDI throughout north Africa and the Middle East, probably reﬂ ecting the uneven achievements in development found in this region. Lower back and neck pain was the primary driver of YLDs in the region. Exceptions to this were Lebanon and Syria, for which war caused the most disability in 2015, and Afghanistan for which iron-deﬁ ciency was the leading cause. Depression, sense organ diseases, diabetes ranked second, third, and fourth, respectively, for the 1576 Articles in 2015. In central sub-Saharan Africa, skin diseases were consistently among the leading three-to-four causes of disability in this region. HIV/AIDS resulted in far more YLDs than expected based on SDI, with Equatorial Guinea, and Central African Republic recording ratios of observed versus expected YLDs higher than 2·00. much faster for YLLs; this pattern is to be expected if access to eﬀ ective treatment is decreasing the number of deaths but not reducing prevalence or incidence. The last category includes disorders where decreases are equal, if not slightly higher, for YLDs including ischaemic heart disease, falls, cervical cancer, and other injuries. Examination of diﬀ erential trends by region or country might provide insights into the role of access to treatment in reducing YLLs and, conversely, the role of increasing or decreasing risks and social determinants in driving changes in disease incidence and prevalence. More detailed exploration of these diﬀ erential trends is warranted in future work. Discussion Our ﬁ ndings suggest that the burden from mental and substance use disorders, and musculoskeletal disorders, which are frequent causes of early retirement31,32 in terms of age-speciﬁ c prevalence, has not declined much over time; the continued prevalence of these and other disorders associated with increasing age might limit the capacities of an older workforce. www.thelancet.com Vol 388 October 8, 2016 Typology of diseases based on YLDs and YLLs Typology of diseases based on YLDs and YLLs The comparison of trends in age-standardised YLDs and YLLs provides insight into the drivers of diﬀ erences in rates of increases and decreases for diseases and injuries. Diseases and injuries can be parsed into four groups. The ﬁ rst of these are a small set of disorders including causes such as drug use disorders and skin cancer for which age- standardised rates of both YLDs and YLLs increased at a rate of more than 0·4% per year from 2005 to 2015. A second category of diseases includes those in which YLDs are increasing but YLLs are decreasing, including disorders such as thyroid cancer, cirrhosis, and neonatal encephalopathy. One explanation for this pattern is that risk factors, broadly deﬁ ned, are causing an increase in incidence, whereas access to eﬀ ective treatment has improved enough to continue reducing mortality. For a third category of disorders, YLDs are essentially constant, but YLLs are decreasing. This grouping is quite large and includes many of the major causes of disability such as musculoskeletal disorders, various neurological disorders, and a number of cancers. For many infectious causes, decreased are occurring for both YLLs and YLDs but are Epidemiological transition The analysis of YLD rates by Level 2 causes as a function of SDI shows a very diﬀ erent picture than that reported for YLLs. At the lower end of the SDI spectrum, incremental increases in SDI are associated with reductions in both age-standardised YLDs and all-age YLD rates spurred by decreases in disability from neglected tropical diseases and anaemia as well as decreases in tuberculosis, diarrhoea, pneumonia, meningitis, and other infectious diseases. At SDI levels above 0·8 (see methods appendix pp 698–711 for SDI values for each country), age-standardised rates of health loss increase slightly, attributable to higher rates of mental and substance use disorders, musculoskeletal disorders, and neurological disorders. In this phase of the epidemiological transition, this increase in rates, although small, combined with population ageing, results in increases in YLDs per 100 000. Typology of diseases based on YLDs and YLLs The rising average YLD per capita rates have implications for health systems; a larger 1577 Articles Articles % change number of YLDs 1990–2005 % change all-age YLD rate 1990–2005 % change age- standardised YLD rate 1990–2005 % change number of YLDs 2005–15 % change all-age YLD rate 2005–15 % change age- standardised YLD rate 2005–15 34·5 9·4 –1·8 39·4 13·4 2·1 14·8 –6·6 –0·6 32·9 8·0 0·6 21·9 –0·8 0·5 29·7 5·5 –0·3 51·8 23·4 13·5 69·2 37·6 20·7 26·1 2·6 –1·5 2·6 –16·5 –15·5 33·9 8·9 –1·6 36·1 10·7 0·7 13·4 –7·8 –13·9 22·2 –0·6 –9·8 53·0 24·4 6·3 29·1 5·0 –3·4 23·2 0·2 0·5 32·5 7·8 0·2 42·1 15·6 11·6 10·8 –9·9 –5·3 29·4 5·2 0·1 32·6 7·9 –1·5 10·9 –9·8 –7·9 48·9 21·1 22·4 35·3 10·1 –2·4 15·8 –5·8 0·7 0·7 –18·1 –23·6 28·2 4·2 –2·5 40·7 14·4 –2·7 –2·2 –20·5 –9·9 18·6 4·9 –2·1 25·2 10·8 0·6 18·2 4·5 1·0 –3·8 –14·9 –11·6 11·7 –1·2 0·4 32·5 17·2 5·4 15·3 2·0 0·8 20·5 6·6 1·3 14·8 1·5 1·0 22·4 8·2 –0·2 9·4 –3·3 –2·3 19·5 5·7 0·3 34·8 19·2 3·9 16·2 2·8 –5·9 11·3 –1·5 –8·6 12·3 –0·7 0·6 10·7 –2·1 –3·3 23·6 9·4 8·2 18·7 5·0 0·3 18·9 5·2 0·6 14·9 1·6 0·5 28·5 13·7 14·7 4·3 –7·7 –4·9 23·8 9·5 0·1 30·2 15·2 –0·3 38·8 22·8 1·1 20·7 6·8 –4·2 11·1 –1·7 –4·5 –6·4 –17·2 –16·3 23·9 9·6 0·5 Leading causes 2005 1 Lower back and neck pain 2 Sense organ diseases 3 Iron-deﬁciency anaemia 4 Depressive disorders 5 Skin diseases 6 Migraine 7 Other musculoskeletal disorders 8 Diabetes 9 Anxiety disorders 10 Asthma 11 Oral disorders 12 Schizophrenia 13 Falls 14 COPD 15 Osteoarthritis 16 Gynaecological diseases 17 Autistic spectrum 18 Other mental and substance 19 Drug use disorders 20 Haemoglobinopathies 21 Bipolar disorder 22 Medication overuse headache 23 Epilepsy 24 Congenital anomalies 25 Chronic kidney disease 26 Conduct disorder 27 Other unintentional 28 Alcohol use disorders 29 Ischaemic heart disease 30 Diarrhoeal diseases 31 Cerebrovascular disease 33 Alzheimer’s disease 34 Other cardiovascular 39 Intestinal nematode Leading causes 1990 1 Lower back and neck pain 2 Iron-deﬁciency anaemia 3 Sense organ diseases 4 Depressive disorders 5 Skin diseases 6 Migraine 7 Other musculoskeletal disorders 8 Anxiety disorders 9 Diabetes 10 Asthma 11 Oral disorders 12 Falls 13 Schizophrenia 14 COPD 15 Autistic spectrum 16 Haemoglobinopathies 17 Gynaecological diseases 18 Intestinal nematode 19 Osteoarthritis 20 Other mental and substance 21 Bipolar disorder 22 Epilepsy 23 Medication overuse headache 24 Other unintentional 25 Drug use disorders 26 Diarrhoeal diseases 27 Conduct disorder 28 Chronic kidney disease 29 Congenital anomalies 30 Alcohol use disorders 33 Cerebrovascular disease 34 Ischaemic heart disease 36 Other cardiovascular 40 Alzheimer’s disease Leading causes 2015 1 Lower back and neck pain 2 Sense organ diseases 3 Depressive disorders 4 Iron-deﬁciency anaemia 5 Skin diseases 6 Diabetes 7 Migraine 8 Other musculoskeletal disorders 9 Anxiety disorders 10 Oral disorders 11 Asthma 12 Schizophrenia 13 Osteoarthritis 14 COPD 15 Falls 16 Autistic spectrum 17 Gynaecological diseases 18 Drug use disorders 19 Other mental and substance 20 Medication overuse headache 21 Bipolar disorder 22 Congenital anomalies 23 Haemoglobinopathies 24 Chronic kidney disease 25 Ischaemic heart disease 26 Alzheimer’s disease 27 Cerebrovascular disease 28 Alcohol use disorders 29 Epilepsy 30 Other cardiovascular 33 Conduct disorder 34 Other unintentional 35 Diarrhoeal diseases 46 Intestinal nematode Communicable, maternal, neonatal, and nutritional Non-communicable Injuries Figure 2: Leading 30 Level 3 causes of global YLDs for both sexes combined, 1990, 2005, and 2015, with percentage change in number of YLDs, and all-age and age-standardised rates Causes are connected by lines between time periods. Typology of diseases based on YLDs and YLLs For the time period of 1990 to 2005 and for 2005 to 2015, three measures of change are shown: percent change in the number of YLDs, percent change in the all-age YLD rate, and percent change in the age-standardised YLD rate. YLD=years lived with disability. COPD=chronic obstructive pulmonary disease. Typology of diseases based on YLDs and YLLs % change number of YLDs 1990–2005 % change all-age YLD rate 1990–2005 % change age- standardised YLD rate 1990–2005 % change number of YLDs 2005–15 % change all-age YLD rate 2005–15 % change age- standardised YLD rate 2005–15 34·5 9·4 –1·8 39·4 13·4 2·1 14·8 –6·6 –0·6 32·9 8·0 0·6 21·9 –0·8 0·5 29·7 5·5 –0·3 51·8 23·4 13·5 69·2 37·6 20·7 26·1 2·6 –1·5 2·6 –16·5 –15·5 33·9 8·9 –1·6 36·1 10·7 0·7 13·4 –7·8 –13·9 22·2 –0·6 –9·8 53·0 24·4 6·3 29·1 5·0 –3·4 23·2 0·2 0·5 32·5 7·8 0·2 42·1 15·6 11·6 10·8 –9·9 –5·3 29·4 5·2 0·1 32·6 7·9 –1·5 10·9 –9·8 –7·9 48·9 21·1 22·4 35·3 10·1 –2·4 15·8 –5·8 0·7 0·7 –18·1 –23·6 28·2 4·2 –2·5 40·7 14·4 –2·7 –2·2 –20·5 –9·9 18·6 4·9 –2·1 25·2 10·8 0·6 18·2 4·5 1·0 –3·8 –14·9 –11·6 11·7 –1·2 0·4 32·5 17·2 5·4 15·3 2·0 0·8 20·5 6·6 1·3 14·8 1·5 1·0 22·4 8·2 –0·2 9·4 –3·3 –2·3 19·5 5·7 0·3 34·8 19·2 3·9 16·2 2·8 –5·9 11·3 –1·5 –8·6 12·3 –0·7 0·6 10·7 –2·1 –3·3 23·6 9·4 8·2 18·7 5·0 0·3 18·9 5·2 0·6 14·9 1·6 0·5 28·5 13·7 14·7 4·3 –7·7 –4·9 23·8 9·5 0·1 30·2 15·2 –0·3 38·8 22·8 1·1 20·7 6·8 –4·2 11·1 –1·7 –4·5 –6·4 –17·2 –16·3 23·9 9·6 0·5 Leading causes 2005 1 Lower back and neck pain 2 Sense organ diseases 3 Iron-deﬁciency anaemia 4 Depressive disorders 5 Skin diseases 6 Migraine 7 Other musculoskeletal disorders 8 Diabetes 9 Anxiety disorders 10 Asthma 11 Oral disorders 12 Schizophrenia 13 Falls 14 COPD 15 Osteoarthritis 16 Gynaecological diseases 17 Autistic spectrum 18 Other mental and substance 19 Drug use disorders 20 Haemoglobinopathies 21 Bipolar disorder 22 Medication overuse headache 23 Epilepsy 24 Congenital anomalies 25 Chronic kidney disease 26 Conduct disorder 27 Other unintentional 28 Alcohol use disorders 29 Ischaemic heart disease 30 Diarrhoeal diseases 31 Cerebrovascular disease 33 Alzheimer’s disease 34 Other cardiovascular 39 Intestinal nematode Leading causes 1990 1 Lower back and neck pain 2 Iron-deﬁciency anaemia 3 Sense organ diseases 4 Depressive disorders 5 Skin diseases 6 Migraine 7 Other musculoskeletal disorders 8 Anxiety disorders 9 Diabetes 10 Asthma 11 Oral disorders 12 Falls 13 Schizophrenia 14 COPD 15 Autistic spectrum 16 Haemoglobinopathies 17 Gynaecological diseases 18 Intestinal nematode 19 Osteoarthritis 20 Other mental and substance 21 Bipolar disorder 22 Epilepsy 23 Medication overuse headache 24 Other unintentional 25 Drug use disorders 26 Diarrhoeal diseases 27 Conduct disorder 28 Chronic kidney disease 29 Congenital anomalies 30 Alcohol use disorders 33 Cerebrovascular disease 34 Ischaemic heart disease 36 Other cardiovascular 40 Alzheimer’s disease Leading causes 2015 1 Lower back and neck pain 2 Sense organ diseases 3 Depressive disorders 4 Iron-deﬁciency anaemia 5 Skin diseases 6 Diabetes 7 Migraine 8 Other musculoskeletal disorders 9 Anxiety disorders 10 Oral disorders 11 Asthma 12 Schizophrenia 13 Osteoarthritis 14 COPD 15 Falls 16 Autistic spectrum 17 Gynaecological diseases 18 Drug use disorders 19 Other mental and substance 20 Medication overuse headache 21 Bipolar disorder 22 Congenital anomalies 23 Haemoglobinopathies 24 Chronic kidney disease 25 Ischaemic heart disease 26 Alzheimer’s disease 27 Cerebrovascular disease 28 Alcohol use disorders 29 Epilepsy 30 Other cardiovascular 33 Conduct disorder 34 Other unintentional 35 Diarrhoeal diseases 46 Intestinal nematode Communicable, maternal, neonatal, and nutritional Non-communicable Injuries Figure 2: Leading 30 Level 3 causes of global YLDs for both sexes combined, 1990, 2005, and 2015, with percentage change in number of YLDs, and all-age and age-standardised rates Causes are connected by lines between time periods. Typology of diseases based on YLDs and YLLs For the time period of 1990 to 2005 and for 2005 to 2015, three measures of change are shown: percent change in the number of YLDs, percent change in the all-age YLD rate, and percent change in the age-standardised YLD rate. YLD=years lived with disability. COPD=chronic obstructive pulmonary disease. Figure 2: Leading 30 Level 3 causes of global YLDs for both sexes combined, 1990, 2005, and 2015, with percentage change in number of YLDs, and all-age and age-standardised rates Causes are connected by lines between time periods. For the time period of 1990 to 2005 and for 2005 to 2015, three measures of change are shown: percent change in the number of YLDs, percent change in the all-age YLD rate, and percent change in the age-standardised YLD rate. YLD=years lived with disability. COPD=chronic obstructive pulmonary disease. Mental and substance use disorders fraction of the population is likely to need care for many disorders. Some of these disorders are currently costly to manage.33 The increase in health-care costs per individual in the population that occurs once an SDI of 0·8 is exceeded is a predictable component of the epidemiological transition; these costs should be anticipated during the health planning processes of countries in that stage of the transition. Consistent with the ﬁ ndings of earlier GBD studies, GBD 2015 conﬁ rms the large contribution of mental and substance use disorders to global disability. For the ﬁ rst time, in GBD 2015 we found a positive association between conﬂ ict and depression and anxiety, albeit with wide uncertainty. This uncertainty is due to sparse and low-quality data for these disorders in post-conﬂ ict countries. Going forward, we plan to make separate estimates for post-traumatic stress disorder in GBD, which might add considerably more data from conﬂ ict settings and show a stronger association. GBD results have provided an evidence base to support global action such as a stated commitment by the World Bank and WHO to make mental health a global development priority39 and the consideration of mental health and substance use disorders in shaping the Sustainable Development Goals.40 Cost-eﬀ ective interventions are able to reduce the burden imposed by mental and substance use disorders, including in low-income and middle-income countries.41 www.thelancet.com Vol 388 October 8, 2016 p Musculoskeletal disorders Musculoskeletal disorders continue to be a leading cause of disability worldwide and more so when taking into account that additional musculoskeletal burden from long-term sequelae of fractures and dislocations is classiﬁ ed under injuries in GBD. A key component of healthy ageing is to maintain mobility, and a key public health intervention recommended for improving health outcomes for all chronic diseases is physical activity. Painful musculoskeletal disorders increase with age and are a great threat to mobility, compromising health more broadly.34–36 Even if cures for musculoskeletal disorders are not yet available, the clinical goal of preventing disability is attainable.37,38 Increases in deaths attributable to drug use in the USA have resulted in considerable policy and media attention. p Musculoskeletal disorders In our assessment, the ten countries with the 1578 Articles ≥80 years Early neonatal 1 2 3 4 5 6 7 8 9 10 Late neonatal Post neonatal 1–4 years 5–9 years 10–14 years 15–19 years 20–24 years 25–29 years 30–34 years 35–39 years 40–44 years 45–49 years 50–54 years 55–59 years 60–64 years 65–69 years 70–74 years 75–79 years Rate of change 2005–15 Iron Iron Iron Iron Iron Iron Skin Depression Back & neck Back & neck Back & neck Back & neck Back & neck Back & neck Back & neck Back & neck Sense Sense Sense Sense NN sepsis PEM Diarrhoea Skin Skin Skin Depression Back & neck Depression Depression Depression Depression Depression Sense Sense Sense Back & neck Back & neck Back & neck Alzheimer’s PEM Diarrhoea PEM PEM Asthma Conduct Iron Skin Migraine Migraine Migraine Sense Sense Depression Diabetes Diabetes Diabetes Diabetes Diabetes Back & neck Haemog Congenital Haemog Diarrhoea Sense Anxiety Back & neck Migraine Skin Skin Sense Migraine Diabetes Diabetes Depression Depression Depression Depression Alzheimer's Diabetes Other inf Haemog Skin Sense Haemog Asthma Migraine Iron Iron Iron Other MSK Other MSK Other MSK Other MSK Other MSK Other MSK Other MSK Oral Depression Falls Diarrhoea NN preterm Other NTD Asthma Other NTD Migraine Anxiety Other MSK Other MSK Sense Skin Diabetes Migraine Migraine Migraine Osteoarth Osteoarth Other MSK Oral IHD NN preterm Other nutr Congenital Haemog Conduct Sense Sense Anxiety Anxiety Other MSK Iron Skin Skin Skin Osteoarth Oral Oral Osteoarth Osteoarth Osteoarth Congenital NN enceph Other inf Other NTD Malaria Depression Conduct Sense Sense Anxiety Anxiety Iron Iron Osteoarth Skin Skin COPD COPD Other MSK Depression Endocrine Other inf NN preterm Congenital ASD Back & neck Other MSK Other mental Drugs Schiz Diabetes Anxiety Anxiety Anxiety Oral Migraine Skin IHD COPD COPD NN enceph Epilepsy Endocrine Otitis Anxiety Haemog Asthma Drugs Schiz Gynae Schiz Schiz Schiz Schiz Anxiety COPD IHD Skin IHD Oral –0·19 to –0·03 0·15 to 0·19 0·19 to 0·24 0·24 to 0·29 0·29 to 0·42 0·42 to 0·63 –0·03 to 0·01 0·01 to 0·06 0·06 to 0·09 0·09 to 0·15 Figure 3: Leading ten Level 3 causes of global age-speciﬁ c years lived with disability in 2015 Each cause is coloured by the percentage change in age-speciﬁ c years lived with disability from 2005 to 2015. Alzheimer’s=Alzheimer disease and other dementias. ASD=autism. Back & neck=low back and neck pain. p Musculoskeletal disorders Conduct=conduct disorders. Congenital=congenital anomalies. COPD=chronic obstructive pulmonary disease. Drugs=drug use disorders. Endocrine=endocrine, metabolic, blood, and immune disorders. Gyne=gynaecological disorders. Haemog=haemoglobinopathies and haemolytic anaemias. IHD=ischaemic heart disease. Iron=iron-deﬁ ciency anaemia. NN enceph=neonatal encephalopathy due to birth asphyxia and trauma. NN preterm=neonatal preterm birth complications. NN sepsis=neonatal sepsis and other neonatal infections. Other NTD=other neglected tropical diseases. Oral=oral disorders. Osteoarth=osteoarthritis. Other inf=other infectious diseases. Other mental=other mental and substance use disorders. Other MSK=other musculoskeletal disorders. Other nutr=other nutritional deﬁ ciencies. PEM=protein-energy malnutrition. Schiz=schizophrenia. Sense=sense organ disease. Skin=skin and subcutaneous diseases. Figure 3: Leading ten Level 3 causes of global age-speciﬁ c years lived with disability in 2015 Each cause is coloured by the percentage change in age-speciﬁ c years lived with disability from 2005 to 2015. Alzheimer’s=Alzheimer disease and other dementias. ASD=autism. Back & neck=low back and neck pain. Conduct=conduct disorders. Congenital=congenital anomalies. COPD=chronic obstructive pulmonary disease. Drugs=drug use disorders. Endocrine=endocrine, metabolic, blood, and immune disorders. Gyne=gynaecological disorders. Haemog=haemoglobinopathies and haemolytic anaemias. IHD=ischaemic heart disease. Iron=iron-deﬁ ciency anaemia. NN enceph=neonatal encephalopathy due to birth asphyxia and trauma. NN preterm=neonatal preterm birth complications. NN sepsis=neonatal sepsis and other neonatal infections. Other NTD=other neglected tropical diseases. Oral=oral disorders. Osteoarth=osteoarthritis. Other inf=other infectious diseases. Other mental=other mental and substance use disorders. Other MSK=other musculoskeletal disorders. Other nutr=other nutritional deﬁ ciencies. PEM=protein-energy malnutrition. Schiz=schizophrenia. Sense=sense organ disease. Skin=skin and subcutaneous diseases. for diabetes. We also included, wherever possible, studies reporting on mean FPG but not diabetes prevalence, using a regression between mean FPG and diabetes prevalence from studies reporting on both. We have also made the assessment of diabetes prevalence more consistent with cause of death data for diabetes. Our improved eﬀ orts at measuring the prevalence of diabetes conﬁ rms the increase in global age-standardised incidence, prevalence, and YLDs. Of note, the increase in YLDs at the global scale is greater than the increase in YLLs, which reﬂ ects improved access to treatments that lower case fatality. The rise of diabetes prevalence, related to the global increase in body-mass index,46 given the costs of treating the disease47 and the related increases in cardiovascular risks, poses one of the more important challenges to health systems in the coming years. This is particularly the case in regions with high prevalence of diabetes such as central America, north Africa and the Middle East, and Oceania. p Musculoskeletal disorders highest prevalence of opioid dependence in decreasing order in 2015 were Iran, United Arab Emirates, Russia, Morocco, the USA, Australia, Ukraine, Tunisia, Belarus, Canada, and Iraq. Among these countries, we estimated the highest excess mortality from opioid dependence in the eastern European countries at about twice the level of that in the USA, Canada, and the north African and Middle Eastern countries; the lowest rate among these highly prevalent countries was in Australia. Within the USA, prescription opioids have been estimated to account for 37% of drug overdose deaths in 2013.42,43 The availability of overdose response treatments in the form of naloxone kits to laypersons has also accelerated.44 An alternative approach is opioid substitution treatment to reduce the risk of overdose. The intensity at which countries use harm-reduction strategies such as needle exchange and opioid substitution programmes follows an inverse pattern,45 with the most intense programmes in Australia, but very rare use of such strategies in eastern Europe, suggesting that embracing harm reduction is an eﬀ ective means of reducing drug deaths. Ezzati and colleagues48 and the International Diabetes Federation (IDF)49 have estimated diabetes prevalence for many countries; the intra-class correlation coeﬃ cient for the estimates from Ezzati and colleagues and the GBD for 2015 is 0·74, and for the IDF estimates is 0·65. These large diﬀ erences appear to stem from the inclusion of cause of death data in the modelling for GBD and also from the inclusion of self-reported diabetes prevalence in the study by Ezzati and colleagues.48 In GBD, these sources were excluded from our analyses because of changing patterns in the prevalence of known and Articles Melanoma=malignant skin melanoma. Skin C=non-melanoma skin cancer. Cervix C=cervical cancer. Uterus C=uterine cancer. Prostate C=prostate cancer. Testis C=testicular cancer. Kidney C=kidney cancer. Thyroid C=thyroid cancer. Hodgkin=Hodgkin lymphoma. Lymphoma=non-Hodgkin lymphoma. Myeloma=multiple myeloma. Oth neopla=Other neoplasms. RHD=rheumatic heart disease. Stroke=cerebrovascular disease. HTN HD=hypertensive heart disease. PVD=peripheral vascular disease. COPD=chronic obstructive pulmonary disease. Asthma=asthma. ILD=interstitial lung disease and pulmonary sarcoidosis. Oth resp=other chronic respiratory diseases. PUD=peptic ulcer disease. Gastritis=gastritis and duodenitis. Hernia=inguinal, femoral, and abdominal hernia. Oth digest=other digestive diseases. Parkinson=Parkinson’s disease. Schiz=schizophrenia. Drugs=drug use disorders. AGN=acute glomerulonephritis. Congenital=congenital anomalies. Skin=skin and subcutaneous diseases. Road inj=road injuries. Oth trans=other transport injuries. Drown=drowning. Fire=ﬁ re, heat, and hot substances. Poison=poisonings. Mech=exposure to mechanical forces. Med treat=adverse eﬀ ects of medical treatment. Animal=animal contact. F body=foreign body. Heat & cold=environmental heat and cold exposure. Oth unint=other unintentional injuries. Violence=interpersonal violence. War=collective violence and legal intervention. 99 5 TB=tuberculosis. HIV=HIV/AIDS. Diarrhoea=diarrhoeal diseases. Intest Inf=intestinal infectious diseases. LRI=lower respiratory infections. URI=upper respiratory infections. Otitis=otitis media. Whooping=whooping cough. Varicella=varicella and herpes zoster. Chagas=chagas disease. Cysticer=cysticercosis. LF=lymphatic ﬁ lariasis. Oncho=onchocerciasis. Trachoma=trachoma. Dengue=dengue. Yellow Fev=yellow fever.Nematode=intestinal nematode infections. Mat hem=maternal haemorrhage. Mat sepsis=maternal sepsis and other maternal infections. Mat HTN=maternal hypertensive disorders. Obst labour=maternal obstructed labour and uterine rupture. Comp abort=maternal abortion, miscarriage, and ectopic pregnancy. Oth mat=other maternal disorders. NN preterm=neonatal preterm birth complications. NN enceph=neonatal encephalopathy due to birth asphyxia and trauma. NN sepsis=neonatal sepsis and other neonatal infections. NN haemol=haemolytic disease and other neonatal jaundice. Oth NN=other neonatal disorders. PEM=protein-energy malnutrition. Iodine=iodine deﬁ ciency. Oth nutr=other nutritional deﬁ ciencies. STD=sexually transmitted diseases excluding HIV. Hep=hepatitis. Stomach C=stomach cancer. Melanoma=malignant skin melanoma. Skin C=non-melanoma skin cancer. Cervix C=cervical cancer. Uterus C=uterine cancer. Prostate C=prostate cancer. Testis C=testicular cancer. Kidney C=kidney cancer. Thyroid C=thyroid cancer. Hodgkin=Hodgkin lymphoma. Lymphoma=non-Hodgkin lymphoma. Myeloma=multiple myeloma. Oth neopla=Other neoplasms. RHD=rheumatic heart disease. Stroke=cerebrovascular disease. HTN HD=hypertensive heart disease. PVD=peripheral vascular disease. COPD=chronic obstructive pulmonary disease. Asthma=asthma. ILD=interstitial lung disease and pulmonary sarcoidosis. Oth resp=other chronic respiratory diseases. PUD=peptic ulcer disease. Gastritis=gastritis and duodenitis. Hernia=inguinal, femoral, and abdominal hernia. Oth digest=other digestive diseases. Parkinson=Parkinson’s disease. Schiz=schizophrenia. Drugs=drug use disorders. AGN=acute glomerulonephritis. Congenital=congenital anomalies. Skin=skin and subcutaneous diseases. Road inj=road injuries. Oth trans=other transport injuries. Drown=drowning. Fire=ﬁ re, heat, and hot substances. Poison=poisonings. Mech=exposure to mechanical forces. Med treat=adverse eﬀ ects of medical treatment. Animal=animal contact. F body=foreign body. Heat & cold=environmental heat and cold exposure. Oth unint=other unintentional injuries. Violence=interpersonal violence. Articles Cervix C Stomach C Mech RHD –5 –4 –3 –2 –1 0 1 2 3 4 5 –5 –4 –3 –2 –1 0 1 2 3 4 5 YLDs (%) YLLs (%) NN sepsis LF Mat HTN Trachoma Obst labour Mat sepsis Comp abort Mat hem Cysticer Intest inf Yellow fev Nematode Whooping Other mat Oncho NN enceph NN hemol Otitis PEM LRI Iodine NN preterm Varicella STD Diarrhoea URI Chagas TB Hep Other NN Dengue HIV Other unint Self-harm F body Heat and cold Other trans Violence Road injuries Animal Poison Fire Drowning Med treat War Schiz Asthma Hodgkin HTN HD Stroke COPD Gastritis PUD AGN Other digest Other resp Epilepsy Hernia Parkinson’s ILD PVD Skin Drugs Testis C Uterus C Congenital Melanoma skin C Lymphoma Myeloma Diabetes Other neopla Kidney C Prostate C Thyroid C Key Non-communicable diseases Injuries Communicable, maternal, neonatal, and nutritional diseases Figure 4: Global annualised rate of change in age-standardised years of life lost (YLLs) and years lived with disability (YLDs) for Level 3 causes between 1990 and 2015 TB b l i HIV HIV/AIDS Di h di h l di I I f i i l i f i di LRI l i i f i URI i Thyroid C Testis C Figure 4: Global annualised rate of change in age-standardised years of life lost (YLLs) and years lived with disability (YLDs) for Level 3 causes between 1990 and 2015 b l / h d h l d f l f d l f Figure 4: Global annualised rate of change in age-standardised years of life lost (YLLs) and years lived with disability (YLDs) for Level 3 causes betw d 1990 and 2015 TB=tuberculosis. HIV=HIV/AIDS. Diarrhoea=diarrhoeal diseases. Intest Inf=intestinal infectious diseases. LRI=lower respiratory infections. URI=upper respiratory infections. Otitis=otitis media. Whooping=whooping cough. Varicella=varicella and herpes zoster. Chagas=chagas disease. Cysticer=cysticercosis. LF=lymphatic ﬁ lariasis. Oncho=onchocerciasis. Trachoma=trachoma. Dengue=dengue. Yellow Fev=yellow fever.Nematode=intestinal nematode infections. Mat hem=maternal haemorrhage. Mat sepsis=maternal sepsis and other maternal infections. Mat HTN=maternal hypertensive disorders. Obst labour=maternal obstructed labour and uterine rupture. Comp abort=maternal abortion, miscarriage, and ectopic pregnancy. Oth mat=other maternal disorders. NN preterm=neonatal preterm birth complications. NN enceph=neonatal encephalopathy due to birth asphyxia and trauma. NN sepsis=neonatal sepsis and other neonatal infections. NN haemol=haemolytic disease and other neonatal jaundice. Oth NN=other neonatal disorders. PEM=protein-energy malnutrition. Iodine=iodine deﬁ ciency. Oth nutr=other nutritional deﬁ ciencies. STD=sexually transmitted diseases excluding HIV. Hep=hepatitis. Stomach C=stomach cancer. Diabetes To obtain standard estimates for health loss due to diabetes for GBD 2015 using both fasting plasma glucose (FPG) means and standard deviations as well as diabetes prevalence, we re-extracted all available data from 1990 to 2015. Across studies, we found 20 diﬀ erent case deﬁ nitions 1579 www.thelancet.com Vol 388 October 8, 2016 Articles Articles Dementia 50 39 28 17 6 0 6 17 28 39 50 0–4 5–9 10–14 15–19 20–24 25–29 30–34 35–39 40–44 45–49 50–54 55–59 60–64 65–69 70–74 75–79 ≥80 Age (years) A Female Male Brayne and colleagues50,51 reported that age-speciﬁ c prevalence of Alzheimer’s was decreasing in the UK. Of the four studies in the USA with similar measurements over time that were reviewed, one showed a decrease, whereas no change in prevalence was seen in the remaining studies.52,53 Our assessment of age-standardised prevalence showed that Alzheimer’s and other dementias decreased slowly in the UK, by 3·69% (2·53–4·85%) from 2005 to 2015, but remained constant in the rest of the world. The decrease might be due to reductions in vascular dementias rather than reductions in Alzheimer’s and other dementias. Our assessment also suggests that the number of individuals with Alzheimer’s disease increased from 21·7 million (18·9–24·8 million) worldwide in 1990 to 46·0 million (40·2–52·7 million) in 2015. Although it is useful to know that age-standardised rates might be starting to decrease, if only slightly, the rapid increase in the absolute number of cases points to the major challenge that dementia presents to societies with increasing life expectancy. This number is similar to the most recent estimates from the World Alzheimer Report 2015 that estimated 46·8 million people living with dementia in 2015. However, there is less agreement about new cases per year: The World Alzheimer Report estimated 9·9 million new cases of dementia per year in 2015 compared with 6·44 million (5·52–7·45 million) new cases in 2015 estimated in GBD 2015. The World Alzheimer Report separately analysed prevalence and incidence whereas we use DisMod-MR 2.1 to produce internally consistent prevalence and incidence estimates. We observed a fundamental disagreement in the underlying incidence and prevalence data in many countries and decided to exclude incidence data from our modelling approach, putting greater trust in prevalence studies than incidence studies, arguing that the point of incidence in dementia is more diﬃ cult to establish than prevalence when the diagnosis over time becomes more established. Articles War=collective violence and legal intervention. 1580 www.thelancet.com Vol 388 October 8, 2016 Articles unknown diabetes in surveys that vary with geography, making it diﬃ cult to crosswalk these studies. 50 39 28 17 6 0 6 17 28 39 50 0–4 5–9 10–14 15–19 20–24 25–29 30–34 35–39 40–44 45–49 50–54 55–59 60–64 65–69 70–74 75–79 ≥80 Age (years) A 46 36 26 15 5 0 5 15 26 36 46 0–4 5–9 10–14 15–19 20–24 25–29 30–34 35–39 40–44 45–49 50–54 55–59 60–64 65–69 70–74 75–79 ≥80 Age (years) Population (millions) B Female Male No disability Very mild Mild Moderate Severe Very severe Disability level Hepatitis C highly eﬀ ective but costly.54 Some programmes have been launched in Egypt and other lower-SDI countries that oﬀ er treatment at lower cost. Good data on the number of courses of treatment that have been delivered are not widely available. Given the number of individuals with chronic infection and the potential to be cured, tracking the fraction of people treated each year should be undertaken. Even in high-income countries such as the USA, the fraction of those treated among those who would beneﬁ t from treatment is not yet available. The availability of new medical treatments for chronic hepatitis C has driven considerable interest in the prevalence data for this disease. Available treatments are 15 000 15 000 13 000 13 000 11 000 11 000 9000 9000 7000 7000 5000 5000 3000 3000 1000 1000 0·25 0·50 0·75 Males Females Socio-demographic Index Age-standardised YLD rate per 100 000 people A 0·25 0·50 0·75 Males Females Socio-demographic Index A Figure 6: Expected relationship between age-standardised years lived with disability (YLD) rates per 100 000 people for the 21 GBD Level 2 causes and SDI (A) and the expected relationship between all-age years lived with disability (YLD) rates per 100 000 people for the 21 GBD Level 2 causes and SDI (B) by sex These stacked curves represent the average relationship between SDI and each cause of YLDs observed across all geographies over the time period 1990 to 2015. In each ﬁ gure, the y-axis spans from lowest SDI up to highest SDI. To the left of the midline are male rates, and the female rates are to the right; higher rates are further from the midline. SDI=Socio-demographic Index. Articles Hepatitis C The availability of new medical treatments for chronic hepatitis C has driven considerable interest in the prevalence data for this disease. Available treatments are 15 000 15 000 13 000 13 000 11 000 11 000 9000 9000 7000 7000 5000 5000 3000 3000 1000 1000 0·25 0·50 0·75 Males Females Socio-demographic Index Age-standardised YLD rate per 100 000 people A 15 000 15 000 13 000 13 000 11 000 11 000 9000 9000 7000 7000 5000 5000 3000 3000 1000 1000 0·25 0·50 0·75 Socio-demographic Index All-age YLD rate per 100 000 people B Forces of nature, war, and legal intervention Self-harm and interpersonal violence Unintentional injuries Transport injuries Other non-communicable diseases Musculoskeletal disorders Diabetes, urogenital, blood, and endocrine diseases Mental and substance use disorders Neurological disorders Digestive diseases Cirrhosis and other chronic liver diseases Chronic respiratory diseases Cardiovascular diseases Neoplasms Other communicable, maternal, neonatal, and nutritional diseases Nutritional deﬁciences Neonatal disorders Maternal disorders Neglected tropical diseases and malaria Diarrhoea, lower respiratory, and other common infectious diseases HIV/AIDS and tuberculosis Malaria Our assessment of malaria prevalence and incidence in high burden sub-Saharan Africa was based on the Malaria Atlas Project spatial analysis of prevalence surveys done across Africa between 2000 and 2015.55,56 The dynamic map of malaria prevalence was updated for GBD 2015 and extended back to 1980. Children younger than 5 years in sub-Saharan Africa showed a remarkable 33·5% (24·8–46·0%) decrease in incidence from the 2005 peak of 0·78 (0·56–0·96) cases per person per year to 0·52 (0·33–0·71) in 2015. Yet there were 81·7 million (52·0–111·7 million) incident cases in this age group in 2015. The decrease in children was more rapid than that recorded for older age groups. In adolescents and adults, incidence dropped by 30·0% (17·2–39·2%) and 21·1% (15·3–26·9%), respectively, and resulted in 58·7 million (37·7–91·5 million) and 53·3 million (40·7–67·5 million) incident cases, respectively, in 2015. The observed decreases in malaria incidence underscore the remarkable eﬀ ect of the scale-up of antimalarial interventions57 across Africa in the past decade.55 Sustaining the levels of these interventions is crucial to continue to reduce malaria morbidity on the continent, avoid resurgence, and improve on these levels that are vital to global aspirations for malaria eradication.58 Overall, we estimate there were 286·9 million (219·7–377·3 million) cases of malaria worldwide in 2015, whereas the World Malaria Report (WMR) 2015 reports 214 million (149–303 million) cases in 2015.59 Of these cases, the WMR estimates that 88% of cases occurred in the WHO African region (188 million) whereas we estimate 189·4 million (151·7–239·4 million; 66·1%) cases for GBD 2015. Outside of Africa, the WMR reports 26 million cases versus our estimate of 97·3 million (50·2–181·8 million), but both sets of estimates feature identical regional rankings among the six WHO regions with malaria. Dementia B 46 36 26 15 5 0 5 15 26 36 46 0–4 5–9 10–14 15–19 20–24 25–29 30–34 35–39 40–44 45–49 50–54 55–59 60–64 65–69 70–74 75–79 ≥80 Age (years) Population (millions) B No disability Very mild Mild Moderate Severe Very severe Disability level Figure 5: Population pyramids with the number of individuals, by age and sex, grouped by severity of their disability weight (DW) for all comorbid conditions combined into no disability, very mild disability (DW 0–0·01), mild disability (DW 0·01–0·05), moderate disability (DW 0·05–0·1), severe disability (DW 0·1–0·3), and very severe disability (DW >0·3) for geographies of high (A) and low (B) quintiles of Socio-demographic Index in 2015 Disability weights are combined multiplicatively as 1−(1−DW1)(1− DW2)... (1−DWn) for n comorbid sequelae. Socio-demographic Index (SDI) is calculated for each geography as a function of lag-dependent income per capita, average educational attainment in the population aged over 15 years, and the total fertility rate. SDI units are interpretable; a zero represents the lowest level of income per capita, educational attainment, and highest total fertility rate (TFR) observed from 1980 to 2015 and a one represents the highest income per capita, educational attainment, and lowest TFR observed in the same period. Cutoﬀ s on the SDI scale for the quintiles have been selected based on examination of the entire distribution of geographies between 1980 and 2015. www.thelancet.com Vol 388 October 8, 2016 1581 Articles Articles (Figure 7 continues on next page) Malaria Increasing convergence in estimates is expected as dynamic maps of 15 000 15 000 13 000 13 000 11 000 11 000 9000 9000 7000 7000 5000 5000 3000 3000 1000 1000 Age-standardised YLD rate per 100 000 people 15 000 13 000 11 000 9000 0 7000 5000 5000 3000 3000 1000 1000 Age-standardised YLD rate per 100 000 people 15 000 15 000 13 000 13 000 11 000 11 000 9000 9000 7000 7000 5000 5000 3000 3000 1000 1000 0·25 0·50 0·75 Socio-demographic Index All-age YLD rate per 100 000 people B Forces of nature, war, and legal intervention Self-harm and interpersonal violence Unintentional injuries Transport injuries Other non-communicable diseases Musculoskeletal disorders Diabetes, urogenital, blood, and endocrine diseases Mental and substance use disorders Neurological disorders Digestive diseases Cirrhosis and other chronic liver diseases Chronic respiratory diseases Cardiovascular diseases Neoplasms Other communicable, maternal, neonatal, and nutritional diseases Nutritional deﬁciences Neonatal disorders Maternal disorders Neglected tropical diseases and malaria Diarrhoea, lower respiratory, and other common infectious diseases HIV/AIDS and tuberculosis 0·25 0·50 0·75 Socio-demographic Index B 15 000 15 000 13 000 13 000 11 000 11 000 9000 9000 7000 7000 5000 5000 3000 3000 1000 1000 All-age YLD rate per 100 000 people 15 000 15 000 13 000 13 000 11 000 11 000 9000 9000 7000 7000 5000 5000 3000 3000 1000 1000 All-age YLD rate per 100 000 people Figure 6: Expected relationship between age-standardised years lived with disability (YLD) rates per 100 000 people for the 21 GBD Level 2 causes and SDI (A) and the expected relationship between all-age years lived with disability (YLD) rates per 100 000 people for the 21 GBD Level 2 causes and SDI (B) by sex These stacked curves represent the average relationship between SDI and each cause of YLDs observed across all geographies over the time period 1990 to 2015. In each ﬁ gure, the y-axis spans from lowest SDI up to highest SDI. To the left of the midline are male rates, and the female rates are to the right; higher rates are further from the midline. SDI=Socio-demographic Index. www.thelancet.com Vol 388 October 8, 2016 (Figure 7 continues on next page) www.thelancet.com Vol 388 October 8, 2016 Malaria 1582 www.thelancet.com Vol 388 October 8, 2016 Articles Global High income High-income North America Canada Greenland USA Australasia Australia New Zealand High-income Asia Paciﬁc Brunei Japan Singapore South Korea Western Europe Andorra Austria Belgium Cyprus Denmark Finland France Germany Greece Iceland Ireland Israel Italy Luxembourg Malta Netherlands Norway Portugal Spain Sweden Back & neck (0·96) Back & neck (1·11) Back & neck (1·08) Back & neck (1·31) Back & neck (1·1) Back & neck (1·05) Back & neck (1·12) Back & neck (1·12) Back & neck (1·15) Back & neck (0·84) Back & neck (0·82) Back & neck (0·82) Back & neck (0·6) Back & neck (0·9) Back & neck (1·27) Back & neck (1·41) Back & neck (1·22) Back & neck (1·39) Back & neck (1·4) Back & neck (1·45) Back & neck (1·38) Back & neck (1·15) Back & neck (1·48) Back & neck (1·28) Back & neck (1·38) Back & neck (1·25) Back & neck (1·2) Back & neck (1·31) Back & neck (1·42) Back & neck (1·31) Back & neck (1·5) Back & neck (1·59) Back & neck (1·34) Back & neck (0·97) Back & neck (1·25) Sense (1·0) Sense (0·94) Depression (1·31) Sense (0·91) Depression (0·94) Depression (1·34) Depression (1·3) Depression (1·33) Depression (1·14) Sense (1·04) Depression (0·9) Sense (1·05) Sense (1·01) Sense (1·02) Sense (0·94) Sense (0·91) Sense (0·93) Sense (1·06) Sense (0·96) Sense (0·79) Sense (0·85) Sense (0·92) Sense (0·98) Sense (0·95) Sense (0·93) Sense (0·92) Sense (0·94) Sense (1·09) Sense (0·92) Sense (0·89) Sense (0·9) Sense (0·81) Sense (0·83) Sense (0·99) Sense (0·78) Depression (0·93) Depression (1·11) Diabetes (2·3) Depression (0·98) Other MSK (2·18) Diabetes (2·39) Sense (0·84) Other MSK (2·0) Sense (0·85) Depression (0·86) Skin (0·93) Depression (0·87) Depression (0·95) Depression (0·83) Depression (1·05) Depression (1·12) Depression (1·07) Depression (1·01) Depression (1·07) Depression (1·1) Depression (1·23) Depression (1·03) Depression (1·08) Depression (1·21) Depression (0·89) Depression (1·07) Depression (1·1) Depression (1·06) Depression (1·11) Depression (1·02) Depression (0·99) Depression (1·03) Depression (1·09) Depression (1·09) Depression (1·05) Iron (1·01) Skin (0·92) Sense (0·89) Skin (1·02) Skin (1·03) Sense (0·89) Other MSK (1·93) Sense (0·84) Anxiety (1·77) Skin (0·98) Sense (1·05) Skin (1·0) Skin (0·93) Diabetes (1·37) Migraine (1·17) Skin (0·88) Migraine (1·29) Skin (0·87) Migraine (1·14) Skin (0·94) Skin (0·92) Skin (0·86) Migraine (1·23) Migraine (1·11) Migraine (1·15) Skin (0·91) Skin (0·86) Migraine (1·39) Migraine (1·28) Diabetes (1·02) Migraine (1·21) Anxiety (1·87) Migraine (1·13) Migraine (1·15) Diabetes (1·36) Skin (0·93) Diabetes (1·29) Other MSK (1·83) Other MSK (1·62) Sense (0·68) Other MSK (1·86) Skin (0·93) Skin (0·93) Skin (0·92) Diabetes (0·96) Diabetes (2·39) Oral (0·97) Migraine (0·75) Skin (0·92) Skin (0·87) Migraine (1·15) Skin (0·88) Migraine (1·14) Skin (0·88) Migraine (1·09) Falls (1·58) Anxiety (1·54) Skin (0·86) Skin (0·83) Skin (0·79) Migraine (1·14) Migraine (1·14) Diabetes (1·01) Skin (0·87) Migraine (1·13) Skin (0·87) Skin (0·85) Skin (0·88) Skin (0·86) Skin (0·9) Diabetes (0·98) Migraine (1·04) Skin (0·94) Diabetes (1·57) Migraine (1·03) Skin (0·93) Anxiety (1·72) Migraine (1·21) Diabetes (1·51) Migraine (0·89) Iron (1·06) Diabetes (0·83) Iron (1·05) Migraine (0·83) Anxiety (1·28) Oral (1·13) Oral (1·23) Diabetes (1·1) Diabetes (1·27) Other MSK (1·23) Migraine (1·14) Migraine (1·04) Anxiety (1·47) Oral (1·13) Anxiety (1·23) Anxiety (1·58) Iron (1·11) Skin (0·86) Diabetes (1·3) Skin (0·9) Anxiety (1·59) Migraine (1·07) Diabetes (0·78) Diabetes (0·78) Migraine (0·91) Migraine (0·93) Other MSK (1·27) Anxiety (1·54) Migraine (1·13) Anxiety (1·28) Anxiety (1·57) Migraine (1·17) Anxiety (1·72) Other MSK (1·53) Oral (0·96) Migraine (0·84) Migraine (0·91) Oral (0·95) Iron (1·28) Diabetes (0·88) Falls (1·23) Anxiety (1·22) Oral (1·17) Anxiety (1·22) Diabetes (1·12) Diabetes (1·06) Falls (1·23) Diabetes (1·02) Anxiety (1·19) Oral (1·18) Iron (1·0) Diabetes (1·2) Anxiety (1·24) Oral (1·21) Oral (1·15) Diabetes (1·23) Diabetes (1·22) Oral (1·09) Iron (1·25) Anxiety (1·22) Other MSK (1·29) Anxiety (1·32) Migraine (0·97) Iron (1·17) Iron (0·99) Migraine (0·95) Diabetes (1·27) Diabetes (1·22) Migraine (1·0) Other MSK (0·92) Falls (1·35) Other MSK (0·93) Other MSK (0·83) Other MSK (0·9) Oral (0·99) Anxiety (1·24) Diabetes (0·9) Falls (1·25) Oral (1·09) Oral (1·07) Oral (1·12) Oral (1·0) Oral (1·05) Diabetes (0·71) Iron (0·83) Other MSK (1·09) Oral (1·15) Other MSK (1·0) Anxiety (1·23) Anxiety (1·18) Oral (1·32) Oral (1·15) Anxiety (1·13) Falls (1·02) Other MSK (1·02) Anxiety (0·93) Oral (0·95) Drugs (2·94) Anxiety (1·23) Diabetes (0·72) Drugs (3·02) Asthma (1·59) Drugs (3·05) Asthma (1·63) Iron (1·12) Anxiety (0·85) Alzheimer's (1·05) Anxiety (0·85) Oral (0·94) Other MSK (0·94) Diabetes (0·94) Falls (1·19) Anxiety (1·17) Asthma (1·24) Anxiety (1·25) Asthma (1·35) Other MSK (1·02) Falls (1·02) Falls (0·86) Diabetes (0·81) Oral (1·17) Other MSK (0·95) Alzheimer's (1·22) Falls (1·15) Falls (1·13) Other MSK (1·13) Other MSK (0·87) Other MSK (0·94) Anxiety (1·01) Oral (0·9) Oral (0·86) Iron (0·92) Iron (0·9) Oral (0·95) Falls (1·17) Iron (0·87) Drugs (2·91) Asthma (1·58) Oral (1·09) Falls (0·89) Asthma (1·13) Iron (1·06) Falls (0·92) Falls (1·06) Falls (1·04) Iron (0·82) Other MSK (0·86) Other MSK (0·83) Other MSK (0·79) Iron (0·79) Iron (1·0) Diabetes (0·69) Other MSK (0·82) Alzheimer's (1·01) Falls (0·96) Asthma (1·36) War (6977·72) Falls (0·94) Asthma (1·42) Asthma (1·12) Falls (0·94) Falls (1·07) Asthma (1·18) Oral (0·91) Iron (0·89) 1 2 3 4 5 6 7 8 9 10 (Figure 7 continues on next page www.thelancet.com Vol 388 October 8, 2016 1583 Articles Back & neck (1·48) Back & neck (1·17) Back & neck (1·17) Back & neck (1·11) Back & neck (1·23) Back & neck (1·15) Back & neck (0·88) Back & neck (0·87) Back & neck (0·88) Back & neck (0·89) Back & neck (1·2) Back & neck (1·21) Back & neck (1·19) Back & neck (1·16) Back & neck (1·24) Back & neck (1·23) Back & neck (1·2) Back & neck (1·2) Back & neck (1·26) Back & neck (1·29) Back & neck (1·37) Back & neck (1·16) Back & neck (1·28) Back & neck (1·37) Back & neck (1·38) Back & neck (1·38) Back & neck (1·26) Back & neck (1·27) Back & neck (1·21) Back & neck (1·33) Back & neck (1·31) Back & neck (1·34) Back & neck (1·35) Back & neck (1·01) Back & neck (1·0) Back & neck (0·98) Back & neck (0·93) Sense (0·9) Sense (0·79) Sense (0·79) Sense (0·73) Migraine (1·75) Sense (0·79) Depression (1·07) Depression (1·07) Depression (1·1) Sense (0·8) Sense (1·16) Sense (1·25) Sense (1·18) Sense (1·08) Sense (1·22) Sense (1·17) Sense (0·96) Sense (1·3) Sense (1·17) Sense (1·1) Sense (0·94) Sense (0·98) Sense (1·04) Sense (0·97) Sense (1·09) Sense (1·05) Sense (0·99) Sense (1·04) Sense (1·27) Sense (1·03) Sense (1·0) Sense (1·06) Sense (0·99) Sense (0·98) Sense (1·03) Sense (1·08) Sense (1·05) Depression (1·08) Depression (0·96) Depression (0·95) Depression (0·97) Sense (0·79) Depression (1·07) Sense (0·84) Sense (0·84) Sense (0·86) Depression (1·03) Depression (1·1) Depression (1·17) Depression (1·15) Depression (1·25) Depression (0·97) Depression (1·16) Depression (1·0) Depression (1·14) Depression (1·29) Depression (1·06) Depression (0·98) Depression (0·98) Depression (1·06) Depression (1·04) Depression (1·13) Depression (1·07) Depression (1·01) Depression (1·04) Depression (1·07) Depression (1·04) Depression (1·02) Depression (1·08) Depression (1·09) Depression (0·97) Depression (1·01) Depression (1·02) Depression (1·01) Skin (0·88) Skin (0·88) Skin (0·87) Skin (0·88) Depression (0·97) Asthma (2·59) Skin (0·93) Skin (0·92) Skin (0·96) Skin (0·93) Skin (0·89) Iron (1·81) Skin (1·02) Skin (0·88) Diabetes (1·32) Skin (0·86) Skin (0·95) Iron (2·22) Skin (0·89) Diabetes (1·14) Skin (0·83) Diabetes (1·01) Diabetes (1·18) Skin (0·9) Diabetes (1·46) Diabetes (1·62) Diabetes (1·14) Diabetes (1·15) Diabetes (1·25) Skin (0·9) Diabetes (1·22) Skin (0·89) Skin (0·9) Iron (1·36) Diabetes (1·4) Iron (1·55) Diabetes (1·19) Migraine (1·08) Migraine (1·1) Migraine (1·04) Migraine (1·03) Asthma (2·35) Skin (0·9) Anxiety (1·53) Anxiety (1·52) Anxiety (1·57) Anxiety (1·49) Iron (1·45) Skin (0·89) Migraine (1·02) Migraine (1·03) Skin (0·85) Diabetes (1·08) Migraine (1·02) Skin (0·88) Migraine (1·03) Skin (0·9) Iron (1·25) Skin (0·9) Skin (0·9) Diabetes (0·83) Skin (0·88) Skin (0·96) Skin (0·89) Skin (0·91) Skin (0·88) Migraine (0·98) Skin (0·92) Diabetes (1·25) Diabetes (1·07) Skin (0·87) Skin (0·87) Skin (0·88) Skin (0·85) Oral (1·22) Asthma (1·86) Other MSK (1·22) Asthma (1·41) Skin (0·89) Migraine (1·03) Migraine (0·93) Iron (1·12) Migraine (0·95) Iron (1·09) Migraine (0·99) Migraine (1·01) Diabetes (0·97) Diabetes (1·05) Migraine (1·03) Migraine (1·04) Iron (1·24) Migraine (1·0) Diabetes (0·86) Migraine (0·98) Migraine (0·98) Migraine (0·98) Migraine (0·98) Migraine (0·97) Migraine (0·98) Migraine (0·98) Migraine (0·98) Migraine (0·98) Migraine (0·97) Diabetes (0·75) Migraine (0·98) Migraine (0·97) Migraine (0·98) Migraine (0·98) Migraine (0·97) Diabetes (1·36) Migraine (0·98) Falls (1·41) Other MSK (1·18) Asthma (1·79) Falls (1·21) Diabetes (1·07) Diabetes (0·95) Iron (1·02) Migraine (0·91) Diabetes (0·9) Migraine (0·94) Diabetes (1·0) Diabetes (0·93) Iron (1·22) Iron (1·16) Iron (1·18) Iron (1·18) Diabetes (0·67) Diabetes (0·95) Oral (0·92) Oral (1·01) Anxiety (0·88) War (1887·24) Oral (1·01) Oral (1·07) Oral (1·04) Oral (0·94) Oral (0·97) Iron (1·12) Oral (1·03) Oral (1·04) Oral (0·98) Iron (1·09) Oral (1·16) Diabetes (0·98) Iron (1·11) Migraine (0·96) Iron (1·25) Anxiety (1·25) Anxiety (1·07) Anxiety (1·08) Oral (0·91) Other MSK (1·02) Other MSK (1·08) Diabetes (0·83) Diabetes (0·82) Other MSK (1·05) Diabetes (0·6) Oral (0·99) Oral (0·96) Falls (1·13) Oral (0·93) Oral (0·89) Oral (0·95) Oral (0·94) Drugs (3·39) Osteoarth (1·24) Anxiety (0·93) Oral (0·98) Oral (1·06) Anxiety (0·96) Iron (1·1) Iron (1·05) Anxiety (0·95) Anxiety (0·89) Oral (0·98) Anxiety (0·95) Iron (1·18) Anxiety (0·91) Anxiety (0·94) Iron (1·18) Anxiety (0·8) Oral (1·06) Anxiety (0·82) Oral (1·02) Diabetes (0·89) Oral (0·91) Oral (0·91) Anxiety (0·89) Falls (1·21) Falls (1·07) Other MSK (1·07) Other MSK (1·09) Oral (0·98) Asthma (1·11) Anxiety (0·81) Drugs (3·26) Oral (0·89) Osteoarth (1·38) Osteoarth (1·35) Osteoarth (1·32) Anxiety (0·7) Oral (0·98) Drugs (3·26) Osteoarth (1·29) Diabetes (0·51) Iron (1·01) Osteoarth (1·21) Osteoarth (1·2) Anxiety (0·96) Osteoarth (1·35) Osteoarth (1·21) Anxiety (0·91) Osteoarth (1·34) Osteoarth (1·28) Osteoarth (1·23) Oral (0·85) Osteoarth (1·35) Oral (1·06) Anxiety (0·81) Oral (1·03) Anxiety (0·82) Other MSK (0·74) Iron (0·92) Iron (0·95) Diabetes (0·63) Anxiety (1·12) Anxiety (1·09) Asthma (1·03) Asthma (1·01) Asthma (1·08) Other MSK (0·92) Osteoarth (1·29) Osteoarth (1·33) Drugs (2·84) Alcohol (5·28) Anxiety (0·77) IHD (1·94) Alcohol (3·29) Alcohol (6·05) IHD (1·76) Iron (0·91) Osteoarth (0·91) Anxiety (0·88) Iron (1·12) Anxiety (0·92) Osteoarth (1·43) Iron (0·85) Iron (0·58) Osteoarth (1·26) IHD (2·09) Anxiety (0·89) Iron (0·8) Osteoarth (1·35) Anxiety (0·95) Asthma (0·69) Disaster (138·88) Osteoarth (1·26) Osteoarth (1·21) 1 2 3 4 5 6 7 8 9 10 Switzerland UK England Northern Ireland Scotland Wales Southern Latin America Argentina Chile Uruguay Central Europe, eastern Europe, and central Asia Eastern Europe Belarus Estonia Latvia Lithuania Moldova Russia Ukraine Central Europe Albania Bosnia Bulgaria Croatia Czech Republic Hungary Macedonia Montenegro Poland Romania Serbia Slovakia Slovenia Central Asia Armenia Azerbaijan Georgia (Figure 7 continues on next pag 1584 Articles Articles Back & neck (1·04) Back & neck (1·04) Back & neck (0·98) Back & neck (1·06) Back & neck (0·94) Back & neck (1·0) Back & neck (0·87) Back & neck (0·86) Back & neck (0·94) Back & neck (0·89) Back & neck (0·94) Back & neck (0·96) Back & neck (0·98) Back & neck (0·79) Back & neck (0·96) Back & neck (0·91) Iron (3·06) Back & neck (0·93) Back & neck (1·0) Back & neck (0·91) Back & neck (0·92) Back & neck (0·89) Back & neck (0·82) Back & neck (0·82) Back & neck (0·81) Back & neck (0·84) Back & neck (0·89) Back & neck (0·82) Back & neck (0·83) Back & neck (0·82) Back & neck (0·8) Back & neck (0·84) Iron (1·01) Back & neck (0·83) Back & neck (0·88) Back & neck (0·83) Back & neck (0·82) Iron (2·4) Iron (1·3) Depression (0·94) Iron (1·03) Depression (1·0) Depression (0·94) Sense (1·06) Sense (1·12) Sense (1·14) Sense (1·16) Sense (0·96) Iron (1·06) Sense (0·95) Sense (1·17) Sense (0·95) Sense (1·19) Back & neck (0·9) Sense (1·07) Iron (1·52) Sense (1·02) Sense (1·14) Sense (1·06) Depression (1·2) Sense (1·19) Sense (1·07) Depression (1·04) Depression (1·25) Sense (1·11) Diabetes (2·12) Sense (0·98) Depression (1·11) Iron (1·41) Back & neck (0·96) Sense (1·0) Sense (1·2) Sense (1·05) Sense (1·05) Sense (1·1) Depression (0·91) Sense (0·94) Depression (0·87) Sense (1·06) Sense (0·93) Depression (1·08) Depression (0·94) Depression (1·02) Depression (1·01) Iron (1·59) Sense (0·94) Skin (1·06) Diabetes (1·9) Depression (0·87) Depression (0·99) Sense (1·12) Depression (1·09) Sense (0·97) Depression (1·08) Depression (1·11) Depression (1·11) Sense (1·21) Depression (1·18) Diabetes (2·05) Iron (1·19) Sense (1·19) Depression (1·14) Sense (1·07) Depression (1·07) Sense (1·06) Depression (1·02) Depression (0·89) Depression (1·1) Diabetes (2·37) Depression (1·11) Depression (1·1) Depression (1·03) Sense (0·85) Skin (0·87) Sense (0·8) Iron (1·31) Iron (1·08) Skin (0·99) Diabetes (1·44) Skin (1·02) Skin (0·91) Depression (0·9) Skin (0·96) Depression (0·88) Depression (0·91) Skin (0·96) Skin (0·95) Depression (0·96) Iron (1·34) Depression (1·03) Iron (1·33) Skin (1·03) Iron (1·2) Diabetes (2·2) Diabetes (1·94) Depression (1·15) Skin (1·05) Skin (1·02) Skin (1·08) Depression (1·12) Skin (1·05) Iron (1·52) Sense (0·91) Disaster (1227·89) Diabetes (1·43) Depression (1·23) Diabetes (1·48) Diabetes (1·84) Skin (0·9) Skin (0·86) Iron (0·97) Skin (0·85) Skin (0·88) Skin (0·87) Anxiety (1·52) Skin (0·95) Anxiety (1·35) Iron (1·01) Skin (0·94) Depression (0·85) Iron (0·79) Skin (0·91) Iron (0·82) Diabetes (1·35) Skin (0·98) Skin (1·02) Skin (1·02) Skin (0·99) Iron (1·29) Skin (1·09) Skin (1·05) Skin (1·05) Skin (0·96) Sense (0·92) Iron (1·32) Diabetes (0·91) Skin (1·05) Iron (1·19) Skin (1·05) Diabetes (1·36) Skin (1·02) Skin (1·05) Skin (1·02) Skin (1·05) Skin (1·05) Migraine (0·96) Migraine (1·0) Migraine (0·98) Migraine (1·0) Migraine (0·96) Migraine (0·98) Iron (1·02) Iron (0·93) Migraine (0·89) Migraine (0·88) Diabetes (0·94) Diabetes (1·67) Asthma (1·54) Migraine (0·91) Migraine (0·91) Iron (1·2) Diabetes (1·17) Anxiety (1·37) Anxiety (1·33) Diabetes (1·15) Migraine (1·07) Diabetes (1·24) Iron (1·43) Iron (1·54) Other MSK (1·57) Anxiety (1·35) Anxiety (1·53) Anxiety (1·46) Iron (1·29) Anxiety (1·37) Diabetes (1·78) Skin (1·04) Sense (0·8) Iron (1·25) Anxiety (1·53) Iron (1·26) Iron (1·23) Diabetes (1·17) Anxiety (0·76) Diabetes (0·79) Anxiety (0·78) Diabetes (1·33) Diabetes (0·84) Diabetes (1·12) Migraine (0·9) Diabetes (0·81) Anxiety (1·08) Migraine (0·91) Migraine (0·9) Migraine (0·91) Anxiety (0·85) Anxiety (1·05) Migraine (0·88) Migraine (0·86) Migraine (1·0) Migraine (0·99) Anxiety (1·37) Anxiety (1·38) Anxiety (1·49) Anxiety (1·49) Anxiety (1·49) Iron (1·49) Diabetes (1·14) Diabetes (1·87) Iron (1·36) Anxiety (1·41) Migraine (0·97) Other MSK (1·73) Anxiety (1·33) Asthma (1·97) Anxiety (1·38) Iron (1·38) Anxiety (1·41) Anxiety (1·41) Oral (1·14) Diabetes (0·52) Anxiety (0·78) Diabetes (0·66) Anxiety (0·82) Anxiety (0·78) Migraine (1·0) Anxiety (1·01) Asthma (1·23) Diabetes (0·72) Anxiety (1·03) Anxiety (1·05) Anxiety (1·05) Other MSK (0·94) Diabetes (0·85) Asthma (1·37) Anxiety (1·06) Diabetes (0·76) Diabetes (0·73) Migraine (0·87) Asthma (1·22) Migraine (1·0) Migraine (0·94) Migraine (0·94) Anxiety (1·46) Migraine (0·97) Migraine (0·96) Migraine (0·95) Migraine (0·95) Asthma (1·5) Anxiety (1·41) Migraine (0·97) Anxiety (1·45) Migraine (0·95) Other MSK (1·3) Migraine (0·95) Migraine (0·95) Anxiety (0·82) Oral (1·05) Oral (1·02) Oral (1·06) Oral (1·02) Oral (1·03) Asthma (1·18) Other MSK (0·89) Other MSK (0·93) Asthma (1·18) Asthma (1·24) Asthma (1·02) Diabetes (0·89) Oral (1·08) Asthma (0·96) Anxiety (1·09) Asthma (1·01) Asthma (1·1) Asthma (1·11) Oral (1·09) Oral (1·09) Asthma (1·62) Other MSK (1·37) Other MSK (1·47) Migraine (0·94) Asthma (1·58) Other MSK (1·41) Other MSK (1·11) Asthma (1·52) Diabetes (0·71) Migraine (0·94) Asthma (1·58) Migraine (1·0) Other MSK (1·39) Migraine (0·95) Asthma (1·52) Asthma (1·45) Osteoarth (1·31) Asthma (0·69) Falls (1·37) Epilepsy (1·54) Asthma (0·65) Asthma (0·79) Other MSK (1·01) Oral (1·08) Oral (1·07) Other MSK (0·98) Oral (1·07) Oral (1·18) Oral (1·07) Iron (0·5) Oral (1·08) Oral (1·03) Other MSK (0·9) Oral (1·09) Oral (1·12) Asthma (0·86) Diabetes (0·57) Other MSK (1·05) Asthma (1·31) Asthma (1·38) Asthma (1·26) Other MSK (1·13) Asthma (1·46) Asthma (1·31) Other MSK (1·25) Other MSK (0·91) Asthma (1·43) Other MSK (0·85) Diabetes (1·38) Asthma (1·48) Asthma (1·41) Other MSK (1·25) Other MSK (1·21) 1 2 3 4 5 6 7 8 9 10 Kazakhstan Kyrgyzstan Mongolia Tajikistan Turkmenistan Uzbekistan Latin America and Caribbean Central Latin America Colombia Costa Rica El Salvador Guatemala Honduras Mexico Nicaragua Panama Venezuela Andean Latin America Bolivia Ecuador Peru Caribbean Antigua Bahamas Barbados Belize Bermuda Cuba Dominica Dominican Republic Grenada Guyana Haiti Jamaica Puerto Rico Saint Lucia Saint Vincent and the Grenadines (Fi 7 ti t www.thelancet.com Vol 388 October 8, 2016 1585 Articles Back & neck (0·82) Back & neck (0·81) Back & neck (0·89) Back & neck (0·87) Back & neck (0·86) Back & neck (0·97) Back & neck (0·86) Back & neck (0·89) Back & neck (0·88) Back & neck (1·05) Back & neck (0·99) Back & neck (0·81) Back & neck (0·94) Back & neck (0·76) Back & neck (0·9) Back & neck (0·75) Back & neck (0·82) Back & neck (0·82) Sense (1·13) Back & neck (0·95) Back & neck (0·81) Back & neck (0·79) Back & neck (0·73) Back & neck (0·87) Back & neck (0·85) Back & neck (1·01) Back & neck (0·9) Back & neck (0·91) Diabetes (3·07) Back & neck (0·94) Back & neck (0·93) Diabetes (3·31) Back & neck (0·86) Iron (0·97) Back & neck (0·9) Back & neck (0·95) Back & neck (0·92) Sense (0·98) Diabetes (3·11) Sense (1·19) Depression (1·23) Depression (1·23) Depression (1·24) Sense (0·94) Sense (0·91) Sense (0·91) Sense (0·77) Sense (1·08) Sense (1·01) Iron (0·94) Sense (0·99) Skin (1·05) Sense (1·0) Sense (0·98) Diabetes (2·04) Back & neck (0·79) Sense (1·05) Sense (1·14) Sense (1·18) Sense (1·04) Sense (0·97) Sense (0·92) Iron (0·98) Diabetes (2·78) Skin (1·31) Back & neck (0·84) Diabetes (3·17) Diabetes (3·07) Back & neck (0·88) Skin (1·25) Back & neck (0·95) Skin (1·33) Skin (1·25) Skin (1·34) Depression (1·07) Sense (1·17) Diabetes (2·14) Sense (0·99) Sense (0·99) Sense (0·97) Depression (0·74) Depression (0·73) Depression (0·74) Skin (0·93) Diabetes (2·29) Skin (1·06) Sense (0·95) Skin (1·05) Iron (0·7) Skin (1·05) Iron (1·35) Sense (1·15) Skin (1·04) Skin (1·07) Diabetes (1·44) Skin (1·07) Depression (0·82) Iron (0·66) Skin (1·07) Skin (1·36) Skin (1·34) Diabetes (1·68) Skin (1·27) Sense (1·27) Nematode (7·02) Skin (1·29) Diabetes (3·3) Skin (1·29) Sense (0·92) Diabetes (2·69) Diabetes (1·49) Skin (1·05) Depression (1·18) Depression (1·24) Anxiety (2·14) Anxiety (2·15) Skin (1·14) Skin (1·0) Skin (0·96) Skin (0·96) Iron (0·92) Other MSK (1·59) Depression (0·76) Skin (1·05) Diabetes (1·17) Sense (0·84) Depression (0·84) Skin (1·08) Depression (0·84) Depression (0·69) Diabetes (1·49) Skin (1·06) Depression (0·84) Skin (1·06) Skin (1·07) Depression (0·77) Diabetes (2·69) Sense (1·03) Sense (0·91) Sense (1·03) Skin (1·24) Skin (1·27) Iron (1·23) Depression (0·82) Sense (0·84) Diabetes (1·5) Sense (0·81) Sense (0·83) Iron (1·25) Skin (1·11) Skin (1·02) Skin (1·0) Skin (1·0) Anxiety (1·81) Diabetes (0·76) Diabetes (0·64) Diabetes (0·62) Depression (0·61) Depression (0·73) Diabetes (1·1) Depression (0·67) Depression (0·77) Depression (0·68) Diabetes (1·51) Depression (0·75) Skin (1·07) Diabetes (1·23) Iron (0·92) Depression (0·81) Diabetes (1·34) Other MSK (1·63) Depression (0·7) Migraine (0·85) Sense (0·96) Iron (0·91) Depression (0·72) Iron (1·1) Other MSK (1·96) Sense (0·81) Sense (0·87) Sense (1·28) Diabetes (2·35) Iron (0·84) Iron (0·53) Iron (0·97) Diabetes (1·21) Iron (1·44) Other MSK (1·44) Migraine (1·12) Migraine (1·12) Iron (1·06) Other MSK (1·22) Other MSK (1·1) Other MSK (1·09) Other MSK (1·15) Skin (0·89) Other MSK (1·53) Migraine (0·82) Other MSK (1·67) Diabetes (1·36) Other MSK (1·82) Migraine (0·84) Other MSK (1·77) Iron (0·69) Depression (0·75) Iron (1·04) Migraine (0·82) Diabetes (0·83) Migraine (0·82) Other MSK (1·18) Depression (0·7) Depression (0·77) Other MSK (1·65) Depression (0·75) Depression (0·89) Iron (0·54) Nematode (7·21) Other MSK (1·61) Depression (0·63) Depression (0·74) Depression (0·65) Depression (0·73) Anxiety (1·38) Anxiety (1·5) Iron (1·38) Iron (1·0) Iron (1·0) Migraine (1·06) Iron (0·8) Iron (0·79) Iron (0·79) Diabetes (0·59) Anxiety (1·03) Iron (0·81) Other MSK (1·11) Iron (0·94) Other MSK (1·27) Nematode (4405·93) Other MSK (1·17) Migraine (0·82) Other MSK (1·38) Other MSK (1·69) Migraine (0·82) Other MSK (1·15) Migraine (0·84) Asthma (1·15) Diabetes (0·6) Asthma (1·76) Other MSK (1·87) Asthma (1·45) Asthma (1·64) Iron (0·97) Depression (0·65) Depression (0·69) Iron (1·08) Asthma (1·7) Other MSK (1·6) Other MSK (1·56) Other MSK (1·69) Migraine (0·96) Other MSK (1·49) Anxiety (1·54) Diabetes (0·86) Diabetes (0·86) Diabetes (0·92) Migraine (0·65) Schiz (1·38) Schiz (1·39) COPD (2·1) Iron (1·05) Migraine (0·83) Asthma (1·14) Migraine (0·83) Migraine (0·82) Anxiety (1·2) Anxiety (0·76) Anxiety (0·8) Migraine (0·82) Migraine (0·82) Asthma (1·14) Anxiety (0·81) Iron (0·82) Other MSK (1·03) Schiz (1·2) Other MSK (1·58) Asthma (1·31) Iron (0·66) Other MSK (1·45) Asthma (1·47) Asthma (1·68) Other MSK (1·7) Migraine (0·71) Nematode (3·34) Asthma (1·16) Asthma (1·45) Asthma (1·68) Asthma (1·53) Migraine (0·94) Migraine (0·96) Other MSK (1·21) Other MSK (1·22) Other MSK (0·9) Schiz (1·31) Migraine (0·57) Migraine (0·57) Anxiety (0·83) Migraine (0·63) Anxiety (0·75) Anxiety (0·81) Anxiety (0·77) Asthma (1·28) Migraine (0·79) Haemog (3·61) Asthma (1·03) Asthma (1·08) Asthma (1·29) Anxiety (0·79) Asthma (0·91) Anxiety (0·79) Diabetes (1·05) Other nutr (72·36) Nematode (3·75) Migraine (0·72) Migraine (0·73) Nematode (58·08) Other NTD (234 964·29) Migraine (0·72) Asthma (1·41) Anxiety (0·85) Other MSK (1·43) Migraine (0·73) Migraine (0·73) Migraine (0·73) Other MSK (1·19) Asthma (1·58) Asthma (1·46) Asthma (1·41) Asthma (1·43) Asthma (1·01) Anxiety (0·72) Anxiety (0·71) Anxiety (0·7) Migraine (0·59) Osteoarth (1·44) Asthma (1·02) Diabetes (0·82) Asthma (0·93) Anxiety (0·8) Schiz (1·18) Diabetes (0·56) Oral (0·84) Anxiety (0·79) Anxiety (0·77) Other MSK (0·86) Schiz (1·15) Asthma (1·08) Anxiety (0·84) Asthma (0·86) Migraine (0·77) Anxiety (0·83) Anxiety (0·82) Migraine (0·72) Migraine (0·72) Other MSK (1·03) Migraine (0·72) Asthma (1·16) LF (100·35) Anxiety (0·82) Anxiety (0·88) Osteoarth (2·01) 1 2 3 4 5 6 7 8 9 10 Suriname Trinidad and Tobago Virgin Islands, USA Tropical Latin America Brazil Paraguay Southeast Asia, east Asia, and Oceania East Asia China North Korea Taiwan (province of China) Southeast Asia Cambodia Indonesia Laos Malaysia Maldives Mauritius Myanmar Philippines Sri Lanka Seychelles Thailand Timor-Leste Vietnam Oceania American Samoa Micronesia Fiji Guam Kiribati Marshall Northern Mariana Islands Papua New Guinea Samoa Solomon Islands Tonga (Figure 7 continues on next page (Figure 7 continues on next page) 1586 www.thelancet.com Vol 388 October 8, 2016 Articles Articles Back & neck (0·91) Back & neck (1·07) Back & neck (1·07) Iron (0·8) Back & neck (1·04) Back & neck (0·95) Back & neck (1·08) Back & neck (1·03) Back & neck (1·08) Back & neck (1·0) Back & neck (0·88) War (10 689·82) Back & neck (1·04) Back & neck (1·14) Back & neck (1·06) Back & neck (0·95) Back & neck (0·94) Back & neck (0·99) Back & neck (1·07) War (1612·15) Back & neck (1·02) Back & neck (1·13) Back & neck (0·99) Back & neck (1·11) Iron (1·37) Iron (1·37) Back & neck (1·06) Back & neck (1·05) Iron (1·58) Back & neck (1·13) Iron (1·02) Iron (0·79) HIV (111·59) HIV (140·61) HIV (73·48) HIV (77·97) HIV (165·51) Skin (1·31) Depression (0·93) Depression (0·93) Back & neck (1·15) Diabetes (1·47) Diabetes (2·9) Iron (1·53) Depression (1·05) Diabetes (2·19) Depression (1·01) Depression (1·18) Back & neck (0·83) Depression (1·0) Diabetes (2·21) Depression (1·29) Depression (1·07) Depression (1·15) Depression (1·11) Iron (0·81) Back & neck (1·05) Diabetes (1·35) Sense (0·85) Depression (1·18) War (248·22) Back & neck (0·91) Back & neck (0·91) Other MSK (2·7) Iron (1·48) Back & neck (0·89) Sense (0·89) Back & neck (0·87) Back & neck (0·98) Back & neck (1·03) Back & neck (0·98) Back & neck (1·08) Back & neck (0·99) Back & neck (1·04) Diabetes (1·79) Sense (0·86) Sense (0·86) War (163·45) Depression (0·95) Depression (1·11) Diabetes (1·5) Sense (0·94) Depression (0·93) Diabetes (1·99) Migraine (1·06) Diabetes (2·0) Diabetes (1·3) Sense (0·79) Skin (0·83) Diabetes (2·42) Diabetes (3·12) Migraine (1·27) Depression (0·84) Depression (0·94) Depression (1·0) Depression (0·92) Diabetes (3·52) Depression (0·84) Sense (1·11) Sense (1·11) Sense (0·95) Sense (0·98) Sense (1·16) Skin (0·97) Sense (0·93) Depression (1·04) Depression (1·05) Iron (1·51) Depression (1·14) Iron (1·28) Sense (1·12) Sense (0·82) Diabetes (1·62) Diabetes (1·62) Skin (0·84) Sense (0·82) Sense (1·02) Sense (0·93) Diabetes (1·48) War (395·43) Skin (0·92) Skin (0·89) Depression (1·02) Sense (0·81) Depression (0·88) Sense (0·72) Other Cardio (19·2) Migraine (1·12) Skin (0·93) Skin (0·84) Sense (0·78) Sense (0·82) Diabetes (1·3) Other MSK (1·92) Skin (0·84) Depression (0·92) Depression (0·92) Depression (0·82) Depression (0·83) Depression (0·94) Migraine (1·33) Depression (0·9) Sense (1·02) Sense (1·06) Depression (1·09) Sense (0·91) Depression (1·09) Depression (1·04) Depression (0·67) Iron (0·82) Iron (0·82) Depression (0·8) Skin (0·83) Other MSK (1·96) Depression (0·77) Migraine (0·98) Sense (0·79) Iron (0·97) Diabetes (2·46) Sense (0·9) Skin (0·87) Skin (0·92) Iron (0·59) Sense (1·01) Skin (0·94) Sense (1·01) Sense (0·75) Skin (0·8) Skin (0·89) Other MSK (1·67) Migraine (1·14) Sense (0·77) Other MSK (1·96) Other MSK (1·96) Iron (0·67) Skin (0·95) Other MSK (1·93) Other MSK (1·93) Skin (1·0) Skin (0·93) Skin (0·95) Sense (0·98) Skin (0·93) Sense (0·95) Diabetes (1·65) Other MSK (1·55) Skin (0·84) Skin (0·84) Sense (0·73) Migraine (1·08) Skin (0·9) Skin (0·84) Skin (0·8) Skin (0·75) Sense (0·92) Sense (1·09) Skin (0·93) Migraine (1·07) Migraine (1·05) Migraine (1·07) Migraine (1·1) Sense (0·96) Diabetes (1·51) Migraine (1·08) Iron (0·73) Other MSK (1·58) Skin (0·83) Sense (1·15) Iron (0·44) Skin (0·89) Skin (0·89) Skin (0·94) Other MSK (1·84) Migraine (1·2) Iron (0·48) Migraine (1·13) Other NTD (1·24) Iron (0·96) Skin (0·95) Iron (0·64) Skin (0·96) Skin (0·95) Asthma (1·45) Migraine (1·07) Migraine (1·07) Diabetes (2·72) Other MSK (1·52) Migraine (1·07) Migraine (1·07) Other MSK (1·44) Iron (0·65) Migraine (1·05) Iron (1·18) Iron (1·35) War (464·77) Iron (0·56) Anxiety (1·07) Skin (0·85) Anxiety (1·05) Anxiety (1·08) Diabetes (1·9) Migraine (1·07) Migraine (1·02) Migraine (1·01) Skin (0·91) Migraine (1·11) Migraine (1·19) Migraine (1·19) Migraine (1·2) Migraine (1·19) Skin (0·86) Depression (0·63) Other MSK (1·56) Migraine (0·8) Diabetes (1·54) Diabetes (1·33) Diabetes (1·63) Diabetes (1·07) Iron (1·01) Migraine (0·73) Other MSK (1·34) Other MSK (1·34) Other unint (5·28) Iron (0·7) Iron (1·43) Other MSK (1·19) Anxiety (1·07) Migraine (1·06) Other MSK (1·69) Anxiety (1·12) Migraine (1·03) Iron (0·72) Other MSK (1·36) Diabetes (1·14) Iron (1·15) Other MSK (1·26) Other MSK (1·33) Anxiety (1·15) Diabetes (1·07) Anxiety (1·04) Iron (0·93) Drugs (2·16) Diabetes (1·76) Diabetes (0·9) Diabetes (0·9) Anxiety (1·11) Diabetes (1·17) Diabetes (0·83) Disaster (661·96) Diabetes (1·47) HIV (4·74) Migraine (0·78) Asthma (1·49) Asthma (1·68) Asthma (1·42) Migraine (0·78) Nematode (2·91) Anxiety (1·06) Anxiety (1·06) Migraine (1·18) Anxiety (1·04) Anxiety (1·07) Anxiety (1·0) Drugs (3·41) Anxiety (1·14) Anxiety (1·07) Other MSK (1·32) Anxiety (1·27) Anxiety (1·04) Anxiety (1·09) War (198·72) Other MSK (1·43) Iron (1·14) Iron (0·52) Asthma (1·18) Anxiety (1·06) Iron (0·66) Anxiety (0·92) Anxiety (1·04) Anxiety (1·17) Anxiety (0·8) Anxiety (0·8) Diabetes (1·0) Diarrhoea (3·33) Anxiety (0·74) Anxiety (0·94) Anxiety (0·92) Asthma (0·98) Asthma (1·29) Migraine (0·77) Migraine (0·77) Migraine (0·78) Asthma (1·27) Anxiety (0·83) Asthma (0·94) Asthma (0·94) Anxiety (1·23) Asthma (1·0) Drugs (1·78) Asthma (0·96) Other cardio (6·21) Other MSK (0·92) Asthma (1·03) Asthma (1·01) Asthma (1·14) Other MSK (1·19) Drugs (4·01) Other MSK (1·08) Anxiety (1·05) Heat & cold (17·26) Asthma (0·8) Other MSK (1·16) Asthma (0·99) Asthma (1·05) Oral (1·18) Iron (1·21) Other MSK (1·34) COPD (1·77) COPD (1·77) Epilepsy (2·97) Anxiety (0·88) COPD (2·01) Asthma (1·07) Asthma (0·91) Malaria (2·25) Anxiety (0·79) TB (15·89) Anxiety (0·78) Anxiety (0·76) Anxiety (0·81) 1 2 3 4 5 6 7 8 9 10 Vanuatu North Africa and Middle East North Africa and Middle East Afghanistan Algeria Bahrain Egypt Iran Iraq Jordan Kuwait Lebanon Libya Morocco Palestine Oman Qatar Saudi Arabia Sudan Syria Tunisia Turkey United Arab Emirates Yemen South Asia South Asia Bangladesh Bhutan India Nepal Pakistan Sub-Saharan Africa Southern sub-Saharan Africa Botswana Lesotho Namibia South Africa (Fi 7 ti t (Figure 7 continues on next page) www.thelancet.com Vol 388 October 8, 2016 1587 Articles Articles HIV (172·35) HIV (46·19) Iron (0·99) Iron (0·76) Iron (0·73) Back & neck (1·05) Back & neck (1·01) Iron (1·31) Iron (0·75) Iron (0·73) Back & neck (0·92) Back & neck (1·1) Iron (0·77) Oncho (92 161·85) Iron (1·11) Iron (1·24) Iron (0·78) Iron (1·18) Back & neck (1·04) Iron (1·04) Back & neck (1·08) Back & neck (1·1) Depression (1·11) Depression (1·05) Sense (1·09) Iron (0·88) Iron (0·89) Depression (1·13) Iron (1·05) Iron (0·7) Depression (1·05) Iron (0·71) Back & neck (0·8) Iron (1·23) Iron (0·81) Iron (0·85) Depression (1·31) Back & neck (0·94) Back & neck (1·04) Back & neck (1·11) Back & neck (1·08) Back & neck (1·08) Depression (0·99) Depression (1·05) Back & neck (1·18) Back & neck (1·06) Back & neck (1·08) Iron (0·87) Iron (0·68) Back & neck (1·11) Back & neck (1·1) Back & neck (0·93) Back & neck (1·07) Back & neck (1·16) Back & neck (1·18) Iron (0·65) Back & neck (0·98) Iron (0·59) Iron (0·71) Iron (0·62) Back & neck (0·85) Depression (1·06) Depression (1·12) Depression (1·06) Sense (1·0) Other NTD (4·38) Sense (1·14) Iron (0·53) Depression (1·04) Sense (0·94) Depression (1·03) Depression (1·05) Depression (1·08) Skin (0·91) Depression (1·09) Iron (0·82) Depression (0·99) Depression (0·94) Depression (0·91) Iron (0·57) Iron (1·08) Skin (0·91) Skin (1·08) Depression (1·01) Depression (1·0) Skin (1·06) Depression (0·92) Iron (0·73) Sense (0·94) Sense (1·06) Sense (0·94) Depression (1·05) Depression (0·99) Depression (0·95) Depression (0·93) Depression (0·95) Sense (0·99) Skin (0·92) Back & neck (0·86) Sense (1·02) Sense (1·0) Skin (1·02) Depression (1·13) Depression (1·09) Back & neck (0·85) Sense (0·99) Depression (0·88) Sense (0·99) Sense (1·02) Skin (0·93) Sense (0·96) Skin (0·95) Depression (1·04) Sense (0·98) Sense (0·91) Skin (0·92) Skin (0·92) Sense (0·98) Depression (0·92) Sense (0·97) Skin (0·91) Sense (0·99) Sense (0·93) Sense (0·92) Sense (0·91) Depression (0·92) Depression (0·96) Skin (0·94) Sense (1·03) Sense (0·97) Sense (0·93) Skin (0·91) Skin (0·97) Skin (0·95) Sense (0·9) Iron (0·6) Back & neck (0·73) Skin (0·91) Back & neck (0·8) Sense (1·09) Back & neck (0·83) Sense (0·93) HIV (7·43) Iron (0·5) Skin (0·93) Oncho (45 415·75) Sense (0·86) Iron (0·45) Sense (0·97) Skin (0·94) Skin (0·85) Skin (0·91) Sense (0·87) Sense (0·93) Skin (0·93) Sense (0·92) Depression (0·94) Sense (0·84) Skin (0·69) Depression (0·92) Skin (0·91) Skin (0·91) Skin (0·83) Skin (0·91) Depression (0·92) Skin (0·78) Skin (0·92) Skin (0·91) Sense (0·87) Sense (0·94) Back & neck (0·78) Iron (0·26) Skin (0·91) Skin (0·91) Back & neck (0·7) Iron (0·32) Skin (1·01) Skin (0·9) Skin (0·91) Back & neck (0·83) War (198·09) Back & neck (0·81) Back & neck (0·81) Back & neck (0·65) Back & neck (0·79) Iron (0·92) Sense (0·88) Malaria (3·39) Malaria (5·86) Malaria (1·56) Oncho (654 332·73) Migraine (0·76) Migraine (0·86) Malaria (15·32) Malaria (3·43) Malaria (150·78) Malaria (2·14) Migraine (0·85) Depression (0·82) Malaria (1·11) Migraine (0·79) Migraine (0·98) Schisto (2926·68) Other NTD (2·54) Migraine (0·83) Malaria (5·16) Malaria (8·85) Migraine (0·79) Migraine (0·88) Malaria (7·6) Other NTD (2·76) War (102·99) Schisto (26·64) Back & neck (0·81) Asthma (1·28) HIV (7·57) Skin (0·89) Skin (0·77) Schisto (1·65) Skin (0·91) Migraine (0·72) Malaria (9·75) Diabetes (1·51) Asthma (1·18) Migraine (0·79) Migraine (0·85) Migraine (0·9) HIV (9·44) Diabetes (0·83) Diarrhoea (1·0) Migraine (0·8) Migraine (0·84) Migraine (0·76) Schisto (13·79) Malaria (1·77) Schisto (18·67) Migraine (0·9) Asthma (0·94) Heat & cold (3·9) Malaria (58·71) Migraine (0·78) Anxiety (0·84) Migraine (0·84) Migraine (0·82) Other NTD (0·88) Asthma (1·0) Migraine (0·8) Heat & cold (9·43) Schisto (42·57) Anxiety (1·0) HIV (9·68) Schisto (145·77) Malaria (2·62) Malaria (1·88) Asthma (1·25) Iodine (3·5) Schisto (11·22) Asthma (0·97) HIV (6·86) Asthma (1·95) Migraine (0·78) Schisto (10·95) Anxiety (0·83) Haemog (1·72) Malaria (45·65) Other cardio (9·52) Asthma (0·85) HIV (4·5) Anxiety (0·85) Schisto (6627·12) Migraine (0·87) HIV (2·78) Malaria (1·85) Anxiety (0·9) Anxiety (0·8) Malaria (0·32) Migraine (0·73) Malaria (24·3) Diarrhoea (0·96) Oncho (46 362·69) Asthma (0·93) Anxiety (0·94) Diarrhoea (0·99) Asthma (1·06) Migraine (0·75) Migraine (0·81) Migraine (0·76) Migraine (0·8) Migraine (0·79) Migraine (0·83) Migraine (0·84) Migraine (0·79) Asthma (1·03) Heat & cold (5·89) Anxiety (0·9) Asthma (1·01) Nematode (5·73) Diabetes (1·12) Anxiety (0·82) Asthma (0·81) Anxiety (0·9) Migraine (0·77) Anxiety (0·73) Anxiety (0·87) Asthma (0·91) Diarrhoea (0·94) Anxiety (0·75) Asthma (0·95) Schisto (15·56) Migraine (0·85) Heat & cold (3·29) Heat & cold (5·99) Diarrhoea (0·77) Anxiety (0·8) Asthma (0·88) Asthma (0·72) Asthma (0·86) Anxiety (0·82) HIV (3·82) Anxiety (1·0) Anxiety (0·92) Diabetes (1·16) Asthma (0·97) Other NTD (0·66) Anxiety (0·87) Anxiety (0·92) Anxiety (0·96) Asthma (0·96) Anxiety (0·92) Migraine (0·93) Asthma (1·1) HIV (5·38) Migraine (0·79) Migraine (0·77) Anxiety (0·75) Asthma (0·82) Schisto (43·81) Diarrhoea (0·81) Anxiety (0·78) Asthma (0·76) Heat & cold (4·25) Anxiety (0·81) Asthma (0·74) Diabetes (0·66) Anxiety (0·86) Asthma (0·9) Anxiety (0·85) Diarrhoea (0·66) Schisto (237·73) Anxiety (0·97) Asthma (0·86) Anxiety (0·79) Diabetes (1·01) Anxiety (0·84) Diarrhoea (1·04) Asthma (0·94) Malaria (0·79) Diabetes (1·11) Anxiety (0·87) Anxiety (0·94) Asthma (0·79) Schisto (17·62) Diarrhoea (1·07) Asthma (0·9) Anxiety (0·97) Other NTD (0·82) Other NTD (0·84) Migraine (0·84) Malaria (12·32) Anxiety (0·89) 1 2 3 4 5 6 7 8 9 10 (Figure 7 continues on next page) 1588 Articles Depression (1·11) Other NTD (4·36) Other NTD (4·05) Iron (0·84) Iron (1·03) Other NTD (4·56) Back & neck (0·95) Other NTD (24·82) HIV (20·66) Iron (0·7) Skin (1·13) Back & neck (1·12) Sense (1·27) Iron (0·76) Sense (1·23) Iron (1·58) Skin (0·91) Back & neck (1·1) Sense (1·25) Sense (1·17) Back & neck (1·0) Back & neck (1·15) Skin (1·14) Sense (1·31) Sense (0·96) Sense (1·22) Back & neck (1·0) Skin (1·1) Skin (1·16) Sense (1·21) Depression (1·11) Back & neck (0·94) Iron (0·52) Skin (1·11) Depression (1·0) Depression (0·95) Other NTD (5·85) Skin (1·1) Diabetes (1·22) Depression (1·13) Back & neck (0·77) Depression (0·97) Iron (0·38) Oncho (37 527·86) Depression (1·04) Oncho (31 562·22) Iron (0·75) Skin (1·1) Migraine (0·88) Oncho (31 767·28) Asthma (1·59) Asthma (1·46) Asthma (1·25) Depression (0·96) Asthma (1·72) Schisto (55 308·55) Schisto (1984·48) Asthma (1·44) Schisto (567·16) Other NTD (0·95) Diabetes (1·2) Asthma (1·4) Malaria (6683·41) Diabetes (1·14) Asthma (0·93) Malaria (1·43) Malaria (15·61) HIV (3·4) Schisto (10 909·69) Malaria (1·11) Migraine (0·78) Asthma (1·62) Anxiety (0·88) Migraine (0·87) Migraine (0·8) Migraine (0·87) Migraine (0·77) Migraine (0·9) HIV (21·87) Migraine (0·77) 1 2 3 4 5 6 7 8 9 10 Zambia Central sub-Saharan Africa Angola Central African Republic Congo (Brazzaville) DR Congo Equatorial Guinea Gabon Colour key 0·0–0·81 1·05–1·13 1·13–1·26 1·26–1·58 ≥1·58 0·81–0·88 0·88–0·94 0·94–0·99 0·99–1·05 Figure 7: Leading ten causes of years lived with disability (YLDs) with the ratio of observed years lived with disability (YLDs) to years lived with disability (YLDs) expected on the basis of SDI in 2015, by location Shades of blue represent much lower observed YLDs than expected levels based on SDI, whereas red shows that observed YLDs exceed expected levels. Articles Other MSK=other musculoskeletal disorders. Other nutr=other nutritional deﬁ ciencies. Other unint=other unintentional injuries. PEM=protein-energy malnutrition. Schitso=schistosomiasis. Schiz=schizophrenia. Sense=sense organ diseases. Skin=skin and subcutaneous diseases. TB=tuberculosis. y Shades of blue represent much lower observed YLDs than expected levels based on SDI, whereas red shows that observed YLDs exceed expected levels. Alzheimer’s=Alzheimer’s disease and other dementias. Back & neck=low back and neck pain. COPD=chronic obstructive pulmonary disease. Drugs=drug use disorders. Haemog=haemoglobinopathies and haemolytic anaemias. Heat & cold=environmental heat and cold exposure. IHD=ischaemic heart disease. Iron=iron-deﬁ ciency anaemia. NTD=neglected tropical diseases. Oncho=onchocerciasis. Oral=oral disorders. Osteoarth=osteoarthritis. Other cardio=other cardiovascular and circulatory diseases. Other MSK=other musculoskeletal disorders. Other nutr=other nutritional deﬁ ciencies. Other unint=other unintentional injuries. PEM=protein-energy malnutrition. Schitso=schistosomiasis. Schiz=schizophrenia. Sense=sense organ diseases. Skin=skin and subcutaneous diseases. TB=tuberculosis. Shades of blue represent much lower observed YLDs than expected levels based on SDI, whereas red shows that observed YLDs exceed expected levels. Alzheimer’s=Alzheimer’s disease and other dementias. Back & neck=low back and neck pain. COPD=chronic obstructive pulmonary disease. Drugs=drug use disorders. Haemog=haemoglobinopathies and haemolytic anaemias. Heat & cold=environmental heat and cold exposure. IHD=ischaemic heart disease. Iron=iron-deﬁ ciency anaemia. NTD=neglected tropical diseases. Oncho=onchocerciasis. Oral=oral disorders. Osteoarth=osteoarthritis. Other cardio=other cardiovascular and circulatory diseases. Other MSK=other musculoskeletal disorders. Other nutr=other nutritional deﬁ ciencies. Other unint=other unintentional injuries. PEM=protein-energy malnutrition. Schitso=schistosomiasis. Schiz=schizophrenia. Sense=sense organ diseases. Skin=skin and subcutaneous diseases. TB=tuberculosis. malaria prevalence and incidence become available for all malaria-endemic countries outside of Africa. than WHO (12·0%). Our list of countries with high burden of tuberculosis is consistent with that of WHO, with a few exceptions. Afghanistan and Cambodia are lower in our estimates, and surpassed by Ukraine and Angola. Tuberculosis We added new data from tuberculosis prevalence surveys done in Indonesia, Ghana, and several subnational locations in India. Our analysis of tuberculosis relied on prevalence surveys, case notiﬁ cations and cause-speciﬁ c mortality estimates. We used the expert judgment values for the case-detection rates (CDRs) from WHO as an initial guide of how much notiﬁ cations need to be increased to reﬂ ect incidence of all tuberculosis, but relied on DisMod-MR 2.1 to ﬁ nd an estimate that is consistent with available prevalence and cause-speciﬁ c mortality rates. We found that, particularly in older age groups, our estimated CDR often falls below the all-age CDR of WHO. In GBD 2013, we used a relative risk approach to predict the proportion of HIV-infected individuals with tuberculosis. In GBD 2015, we improved our modelling strategy by making use of more abundantly available data on the proportions of HIV infected cases among all tuberculosis cases from the WHO case notiﬁ cations to separate out combined HIV and tuberculosis from all forms of tuberculosis. In our modelling of tuberculosis, we have not separately estimated the incidence and prevalence related to multidrug-resistant tuberculosis. Given the policy interest in multidrug-resistant tuberculosis, we plan to include estimates for multidrug-resistant tuberculosis in future rounds of the GBD. Our global tuberculosis (all- forms) incidence estimate (10·2 million [9·2–11·5 million] cases in 2015) is slightly higher than that of WHO (9·6 million cases in 2014), and we estimate a slightly larger fraction of combined HIV and tuberculosis (13·0%) Cardiovascular diseases Ageing and population growth have led to a growing number of people living with atherosclerotic vascular disease worldwide, despite the decrease in incident myocardial infarction and ischaemic stroke in high- income regions. Rising levels of obesity and diabetes in many countries makes this an area of major concern for global health. Increased attention will need to be directed toward this highly treatable set of disorders. Health systems will need to improve the delivery of cost-eﬀ ective treatments such as blood pressure and cholesterol- lowering drugs while eﬀ orts continue to focus on decreasing tobacco smoking, improving diet, and increasing physical activity. Investments in universal primary care and public health campaigns need to be balanced against the need to improve access to emergency care, including the strengthening of pre-hospital systems and expanded access to revascularisation treatments. www.thelancet.com Vol 388 October 8, 2016 Articles Alzheimer’s=Alzheimer’s disease and other dementias. Back & neck=low back and neck pain. COPD=chronic obstructive pulmonary disease. Drugs=drug use disorders. Haemog=haemoglobinopathies and haemolytic anaemias. Heat & cold=environmental heat and cold exposure. IHD=ischaemic heart disease. Iron=iron-deﬁ ciency anaemia. NTD=neglected tropical diseases. Oncho=onchocerciasis. Oral=oral disorders. Osteoarth=osteoarthritis. Other cardio=other cardiovascular and circulatory diseases. Other MSK=other musculoskeletal disorders. Other nutr=other nutritional deﬁ ciencies. Other unint=other unintentional injuries. PEM=protein-energy malnutrition. Schitso=schistosomiasis. Schiz=schizophrenia. Sense=sense organ diseases. Skin=skin and subcutaneous diseases. TB=tuberculosis. Figure 7: Leading ten causes of years lived with disability (YLDs) with the ratio of observed years lived with disability (YLDs) to years lived with disability (YLDs) expected on the basis of SD in 2015, by location Shades of blue represent much lower observed YLDs than expected levels based on SDI, whereas red shows that observed YLDs exceed expected levels. Alzheimer’s=Alzheimer’s disease and other dementias. Back & neck=low back and neck pain. COPD=chronic obstructive pulmonary disease. Drugs=drug use disorders. Haemog=haemoglobinopathies and haemolytic anaemias. Heat & cold=environmental heat and cold exposure. IHD=ischaemic heart disease. Iron=iron-deﬁ ciency anaemia. NTD=neglected tropical diseases. Oncho=onchocerciasis. Oral=oral disorders. Osteoarth=osteoarthritis. Other cardio=other cardiovascular and circulatory diseases. Other MSK=other musculoskeletal disorders. Other nutr=other nutritional deﬁ ciencies. Other unint=other Figure 7: Leading ten causes of years lived with disability (YLDs) with the ratio of observed years lived with disability (YLDs) to years lived with disability (YLDs) expected on the basis of SD in 2015, by location Shades of blue represent much lower observed YLDs than expected levels based on SDI, whereas red shows that observed YLDs exceed expected levels. Alzheimer’s=Alzheimer’s disease and Figure 7: Leading ten causes of years lived with disability (YLDs) with the ratio of observed years lived with disability (YLDs) to years lived with disability (YLDs) expected on the basis of SD in 2015, by location Figure 7: Leading ten causes of years lived with disability (YLDs) with the ratio of observed years lived with disability (YLDs) to years lived with d n 2015, by location 5, y Shades of blue represent much lower observed YLDs than expected levels based on SDI, whereas red shows that observed YLDs exceed expected levels. Alzheimer’s=Alzheimer’s disease and other dementias. Back & neck=low back and neck pain. COPD=chronic obstructive pulmonary disease. Drugs=drug use disorders. Haemog=haemoglobinopathies and haemolytic anaemias. Heat & cold=environmental heat and cold exposure. IHD=ischaemic heart disease. Iron=iron-deﬁ ciency anaemia. NTD=neglected tropical diseases. Oncho=onchocerciasis. Oral=oral disorders. Osteoarth=osteoarthritis. Other cardio=other cardiovascular and circulatory diseases. Vision loss Globally, 34·3 million (30·7–38·0 million) people are blind, an additional 24·3 million (21·6–27·6 million) have severe vision impairment, 214 million (193–237 million) have moderate vision impairment, and 663 million (638–690 million) have near vision impairment in 2015. Combined, vision loss accounts for 24·5 million (17·0–34·5 million) YLDs and is the third largest impairment after anaemia (77·9 million [52·4–111·4 million] YLDs) and hearing loss (46·2 million [31·6–63·2 million] YLDs). YLDs from vision loss are slightly lower than those for anxiety disorders, which rank ninth worldwide. Largely due to ageing, substantial increases occurred from 2005 to 2015 in the number of people with blindness (increasing by 23·3% [21·9–24·6%]), severe visual impairment (24·3% [22·4–26·0%]), moderate visual impairment (22·0% [20·1–23·7%]), and near vision impairment (22·5% [20·9–24·0%]). Most causes of vision loss can be prevented or cured using cost-eﬀ ective interventions.61,62 Advances in data and analysis DisMod-MR For GBD 2015, we improved the estimation for subnational units in DisMod-MR 2.1 compared to GBD 2013 and made substantive eﬀ orts to standardise the modelling approach across diseases. Three changes had an important eﬀ ect on our estimates. First, the cascade was implemented such that national estimates were used to inform subnational estimation (in GBD 2013, each subnational unit was treated as a country within a GBD region). Second, wherever possible, we improved the linkage in the estimation between cause-speciﬁ c mortality and disease incidence and prevalence by the systematic inclusion of estimated excess mortality data to provide more informative priors for each location in the geographical hierarchy. Third, in GBD 2010, and to a lesser degree in GBD 2013, many DisMod-MR models did not include ﬁ xed eﬀ ects of country covariates for incidence or prevalence. Spatial heterogeneity was largely estimated from the random eﬀ ects. For GBD 2015, we systematically tested the inclusion of more ﬁ xed eﬀ ects in DisMod-MR 2.1 models including, where appropriate, variables related to the risk factor exposures we estimate for each disease. Emerging infectious diseases Emerging infectious diseases methods to estimate incidence, which precludes the calculation of uncertainty for those estimates. The GBD relies on the strength of the mortality estimates and transformation of the mortality estimates to incidence, taking into account uncertainty associated with both the mortality estimates and with the mortality to incidence ratios. Until high-quality cancer incidence is available in all countries, the GBD approach, which maximises the amount of data used to determine incidence estimates, is highly beneﬁ cial. Emerging and re-emerging infectious diseases pose a persistent yet dynamic challenge to global health63 and to the GBD study. GBD 2015 has included Ebola virus disease in response to the devastating outbreak in west Africa starting in 2013,64 whereas we have not captured the recent expansions of chikungunya65 and Zika virus across the Americas,66 for instance. Quantifying the eﬀ ect of newly emerging epidemic diseases is diﬃ cult because the formal reporting of the disease events might considerably lag the ﬁ rst presentation of cases. As a result, contemporary events such as the substantial deviation from baseline levels in reported congenital abnormalities in Brazil in 2015,67 now established to be associated with Zika virus,68 will need to be incorporated into GBD 2016. Chronic kidney disease y Compared with our GBD 2013 estimates, we estimated global prevalence of chronic kidney disease to be lower by a third. Data for chronic kidney disease at younger ages are sparse. We increased the number of datapoints below the age of 30 years from 20 to 51, and that led to considerably lower prevalence estimates for these ages. We have used diﬀ erent strategies for estimating deaths versus YLDs secondary to chronic kidney disease due to other causes. In our cause of death analysis, deaths secondary to many of the causes typically categorised within the chronic kidney disease other category, such as cystic disease and other congenital renal diseases, were already counted under the primary disease following the principles of the ICD, and thus could not be recounted within the chronic kidney disease other category. In comparison, for the YLD analysis we counted chronic kidney disease from these causes in the chronic kidney disease other category. This diﬀ erence in the classiﬁ cation of morbidity and mortality from chronic kidney disease other is driven by the underlying nature of the data on causes of death and prevalence. In future GBD versions, we plan to make explicit estimates of death and YLDs for polycystic kidney disease, an important component of the other category. www.thelancet.com Vol 388 October 8, 2016 Cancer The International Agency for Research on Cancer last produced cancer estimates by country, age, sex, and cancer site for 2012 for the GLOBOCAN project.60 The total estimated cancer incidence from GLOBOCAN for 2012 was 14·1 million individuals. By comparison, GBD 2015 estimated 18·6 million (18·0–19·4 million) new cases of cancer in 2015, which includes cases of non- melanoma skin cancer. GLOBOCAN used nine diﬀ erent 1589 Articles Articles Crosswalks Because of the diversity of data sources available, we estimated the statistical association between measurements taken using diﬀ erent case deﬁ nitions, diﬀ erent assays, diﬀ erent survey items, or diﬀ erent ascertainment methods to a reference category. We used these statistical associations to crosswalk each type of data to the reference category. Crosswalks are a crucial dimension in the GBD analysis; in most cases we estimate these crosswalks from within DisMod-MR 2.1. If crosswalks were believed to vary substantially by age, sex, or location, these crosswalks were estimated outside of DisMod-MR 2.1 and adjusted data were used as inputs. Where data are sparse, the estimated crosswalks can shift substantially when new studies are identiﬁ ed that inform the crosswalk. With each iteration of the GBD, we have recognised the crucial nature of the crosswalk analysis for data processing. We believe that more research is needed in future iterations of GBD to standardise the approaches used for estimating crosswalks across disorders and risk factors, and for propagating uncertainty in crosswalks into the ﬁ nal results. g For the USA we used claims data for selected disorders. These data were particularly useful in the estimation of state-by-state prevalence for some disorders. Claims data, however, have important potential biases. Because of exclusion of some disadvantaged groups from health insurance, disease rates might be diﬀ erent for individuals that are covered compared with those from the general population.69,70 As the claims dataset included information from Medicaid and Medicare, the coverage schemes for low-income citizens and older adults (older than 65 years) in the USA, the bias toward underestimation might not be so great. Indeed, we found that for many diseases, such as asthma, diabetes, or rheumatoid arthritis, the claims data were consistent with high-quality survey estimates. A potential problem of overestimation exists because visits to rule out a diagnosis can be coded to the diagnosis. Others working with claims have counted diagnoses only if they had appeared at least twice during a year. Applying such a restriction would have led to rejection of 44–68% of skin disorders, such as acne, psoriasis, or scabies, where one would expect that even after a single visit a diagnosis can be reliably made. Crosswalks As we tend to adjust the claims data used in DisMod-MR 2.1 if a systematic bias is detected from population-based sources, such as the National Health and Nutrition Examination Survey or other quality surveys, we believe our estimates have not been inﬂ uenced much by this problem. The potential for greater inclusion of claims data in the GBD analysis is large; in countries with universal health access, claims data might overcome the problems of non-responder bias in survey data, which tend to lead to underestimates of the true population prevalence or incidence. Disability weights Disability weights For GBD 2015, we have not incorporated any new population-based data on disability weights; we have used the same disability weights as for GBD 2013. The present set of disability weights were based on population surveys in nine countries and an open internet survey. So far, these studies have not found systematic variation in disability weights across populations or within populations as a function of education, income per capita, or other variables. In future, we hope that more population surveys on disability weights will continue to be done to enrich the database for disability weight estimation. Given that we have not been able to collect disability weights in every location included in the GBD, it is possible that disability weights might vary systematically across communities. Even if this were to be conﬁ rmed empirically, for global standardised comparisons we would still use the global 1590 Articles Self-reported functional health status average disability weights; however, country analyses might use local disability weights. Until there is evidence of systematic variation across communities, this remains a theoretical and not a practical consideration. Given the close correlation between the internet survey and the population-based surveys, we are considering the implementation of an open rolling internet survey to collect more data for disability weights including new or revised health-state descriptions. f p f We found only modest progress in high-SDI countries in reducing age-standardised YLDs per capita over the last 25 years. There are, however, reports in the literature of substantial improvements based on self-rated health.71,72 In the USA, these improvements seem to be survey programme-speciﬁ c and have not been reported in all data systems.71 Nevertheless, more work is needed to understand the diﬀ erent trends between our assessment of YLDs per capita and self-rated health. Item response theory has been used in some ﬁ elds to identify changing response patterns for single items compared with the pool of all items.73 Item response theory cannot capture systematic changes that aﬀ ect a set of items, and anchoring vignettes have been proposed as a strategy to deal with these challenges.72,74 In the era of increased interest in measuring functional health status, more exploration of the reasons for the divergence from a sequelae-based approach to measuring functional health status used in GBD versus an overall self-rating will be an important avenue of future research. Using claims data Using claims data www.thelancet.com Vol 388 October 8, 2016 GATHER compliance d ll d Providing all documentation for data sources, establishing access to the datasets used in modelling, and posting the code has been a labour-intensive activity; GBD 2015 is compliant with the new GATHER guidelines as a result.17 Posting code, we believe, will stimulate other researchers to explore the methods used in GBD estimation of non- fatal outcomes and hopefully lead to suggestions for improved estimation. With a steadily growing set of co- investigators and a widening community interested in global health estimation, the enhanced transparency of GBD will improve the eﬃ cacy of the overall eﬀ ort. Limitations The GBD 2015 study has some key limitations. First, although we have sought to capture and include in our estimations many sources of uncertainty, we have not captured all sources. We have not been able to routinely capture uncertainty due to diﬀ erent model speciﬁ cations in our results. We do not have the ability to reﬂ ect the uncertainty of data sources that exist but of which we are not aware. Subnational collaborations in China and Mexico revealed many data sources not captured in previous iterations of the GBD study. Inclusion of these data sources can lead to shifts in estimates that are outside the previously estimated 95% UIs. Sixth, we have emphasised in the reporting of results the changes from calendar year 2005 to calendar year 2015, although we have estimated results for 1990, 1995, 2000, 2005, 2010, and 2015 and interpolated for the years in between. For some disorders, reporting the change from 2005 to 2015 (such as for Ebola virus disease) does not capture the major epidemic in west Africa in 2014. Likewise, disasters and wars are often concentrated in a single year; comparisons of any two years can provide misleading inferences about trends. Seventh, for a few disorders (eg, tetanus, neonatal sepsis, rabies, and diphtheria) we estimate disease incidence from estimates of mortality and the inverse of the case-fatality rate estimated from available studies. When the case- fatality rate is very low, these estimates of incidence are highly sensitive to very small changes in the estimated case-fatality rate and have large uncertainty intervals. Second, within DisMod-MR 2.1 we estimate the average association between diﬀ erent types of studies for reporting on a speciﬁ c outcome such as the diﬀ erence between 12-month recall and point prevalence in a survey. These estimated associations from within the DisMod-MR 2.1 likelihood estimation are used to adjust the data to a reference case deﬁ nition or study design. These estimated adjustment factors are themselves uncertain, which increases the overall uncertainty in our estimation. More standardisation in data collection would be highly desirable and would reduce our dependence on the relatively challenging estimation process involved in crosswalks. Finally, although extraordinary eﬀ orts have gone into vetting the results by GBD researchers and the collaborator network, there are probably some ﬁ ndings that have not been scrutinised carefully enough. Future directions for GBD With each cycle of the GBD we expect, given the present interest in subnational assessments, to report increasingly granular results. These subnational ﬁ ndings will be of important national policy interest, but the discipline of examining the evidence for each community and modelling at the more granular level will, we believe, improve the quality of the national estimation as well. Additions to the cause list will continue to be driven by policy interest, such as the need to incorporate Zika virus, to split diabetes into separate estimates for type 1 and type 2, and by adding diseases to our cause list that are main contributors to large residual categories, such as other cardiovascular disease, other musculoskeletal disorders, and other neurological disorders. We plan to continue to expand our networks of topic-speciﬁ c and country collaborators to enhance the quality of our estimates through their feedback and by enhancing the amount of data we can bring to bear on GBD estimation. Our country collaborations to generate subnational estimates have shown us that these eﬀ orts greatly enhance access to valuable data sources, particularly administrative datasets on inpatient and outpatient episodes and unpublished surveys. p p y Fifth, we estimated separate DisMod-MR 2.1 models for 1990, 1995, 2000, 2005, 2010, and 2015. Independent estimation of each time period implies that the uncertainty intervals for each period are also independent. Furthermore, compositional bias in the data available in diﬀ erent time periods might lead to spurious time trends. A more appealing strategy would be to simultaneously estimate the trends in incidence, excess mortality, and remission by age and sex consistent with all the available data for a geography on prevalence, incidence, remission, excess mortality, and cause-speciﬁ c mortality by age and sex. DisMod-MR 2.1 does not allow for time-varying trends in incidence, excess mortality, and remission, but a prototype DisMod-MR AT (age and time) has been developed and is being tested. Allowing for all rates to change at diﬀ erent paces over age and time increases the number of parameters that need to be estimated by orders of magnitude. Articles Third, for some disorders the data available to estimate excess mortality by age and sex and to capture how excess mortality changes with development status are very limited. There are probably many unpublished sources of information in some countries that could be usefully brought to bear on the challenge of estimating the age–sex–year–location levels of excess mortality. incidence and prevalence, GBD provides a framework to assist countries in prioritising the collection of new data to inform better monitoring of functional health status. By setting clear reference case deﬁ nitions and data collection methods (methods appendix pp 4–7), GBD can also provide guidance on how to collect information most relevant to population health measurement. Fourth, the microsimulation step of this study assumes that prevalence within an age–sex–location–year group of diﬀ erent sequelae is independent. Although this is clearly known not to be the case for some pairs of disorders, the independence assumption provides reasonable overall adjustments for comorbidity. Progress in incorporating dependent comorbidity has been diﬃ cult because information on the correlation structure of prevalence is extremely limited and only available for a minor fraction of all possible pairs of conditions in the GBD study. www.thelancet.com Vol 388 October 8, 2016 Data gaps We have described the availability of data by cause and geography over diﬀ erent time periods, highlighting very large gaps in information availability. The data availability by country ranged from 6·1% in North Korea to 91·3% in the USA (methods appendix p 606). Among the top ﬁ ve causes of YLDs, data availability was rather poor. Lower back pain, iron-deﬁ ciency anaemia, major depressive disorder, other hearing loss, and neck pain all have data representativeness indices below 50%. Because of the broad overview provided of major data sources for disease www.thelancet.com Vol 388 October 8, 2016 1591 Articles Articles Conclusion The GBD studies seek to quantify the prevalence and incidence of the major sequelae for a comprehensive list of diseases and injuries; given the diversity of data sources, the range of biases in these sources, and the gaps in availability, this is a challenging task. Despite the limitations, the standardised and comprehensive approach of the GBD studies study provides useful insights. We believe users of the ﬁ ndings, including the wealth of country-speciﬁ c and cause-speciﬁ c detail available in the appendix, can usefully examine the broad trends for their country or subnational geography, benchmark against geographies at a similar level of development, and understand the strength or weakness of these estimates. Regular quantiﬁ cation of health is particularly important as the new health-related targets of the Sustainable Development Goals have broadened the health agenda throughout the world. At the same time, development and transformations in the risks to health experienced by diﬀ erent groups in the world are leading to some broad transformations in health. Everyone needs to have access to the timeliest, valid, reliable, and local information possible to enrich debates on how to accelerate health progress in all communities. Limitations Making all results and underlying data available and having an increasing number of visualisation tools are eﬀ ective 1592 Articles Euripide Frinel G Arthur Avokpaho, Ashish Awasthi, Beatriz Paulina Ayala Quintanilla, Peter Azzopardi, Umar Bacha, Alaa Badawi, Kalpana Balakrishnan, Amitava Banerjee, Aleksandra Barac, Suzanne L Barker-Collo, Till Bärnighausen, Lars Barregard, Lope H Barrero, Arindam Basu, Shahrzad Bazargan-Hejazi, Ettore Beghi, Brent Bell, Michelle L Bell, Derrick A Bennett, Isabela M Bensenor, Habib Benzian, Adugnaw Berhane, Eduardo Bernabé, Balem Demtsu Betsu, Addisu Shunu Beyene, Neeraj Bhala, Samir Bhatt, Sibhatu Biadgilign, Kelly Bienhoﬀ , Boris Bikbov, Stan Biryukov, Donal Bisanzio, Espen Bjertness, Jed Blore, Rohan Borschmann, Souﬁ ane Boufous, Michael Brainin, Alexandra Brazinova, Nicholas J K Breitborde, Jonathan Brown, Rachelle Buchbinder, Geoﬀ rey Colin Buckle, Zahid A Butt, Bianca Calabria, Ismael Ricardo Campos-Nonato, Julio Cesar Campuzano, Hélène Carabin, Rosario Cárdenas, David O Carpenter, Juan Jesus Carrero, Carlos A Castañeda-Orjuela, Jacqueline Castillo Rivas, Ferrán Catalá-López, Jung-Chen Chang, Peggy Pei-Chia Chiang, Chioma Ezinne Chibueze, Vesper Hichilombwe Chisumpa, Jee-Young Jasmine Choi, Rajiv Chowdhury, Hanne Christensen, Devasahayam Jesudas Christopher, Liliana G Ciobanu, Massimo Cirillo, Matthew M Coates, Samantha M Colquhoun, Cyrus Cooper, Monica Cortinovis, John A Crump, Solomon Abrha Damtew, Rakhi Dandona, Farah Daoud, Paul I Dargan, José das Neves, Gail Davey, Adrian C Davis, Diego De Leo, Louisa Degenhardt, Liana C Del Gobbo, Robert P Dellavalle, Kebede Deribe, Amare Deribew, Sarah Derrett, Don C Des Jarlais, Samath D Dharmaratne, Preet K Dhillon, Cesar Diaz-Torné, Eric L Ding, Tim R Driscoll, Leilei Duan, Manisha Dubey, Bruce Bartholow Duncan, Hedyeh Ebrahimi, Richard G Ellenbogen, Iqbal Elyazar, Matthias Endres, Aman Yesuf Endries, Sergey Petrovich Ermakov, Babak Eshrati, Kara Estep, Talha A Farid, Carla Soﬁ a e Sa Farinha, André Faro, Maryam S Farvid, Farshad Farzadfar, Valery L Feigin, David T Felson, Seyed-Mohammad Fereshtehnejad, Jeﬀ erson G Fernandes, Joao C Fernandes, Florian Fischer, Joseph R A Fitchett, Kyle Foreman, F Gerry R Fowkes, Jordan Fox, Richard C Franklin, Joseph Friedman, Joseph Frostad, Thomas Fürst, Neal D Futran, Belinda Gabbe, Parthasarathi Ganguly, Fortuné Gbètoho Gankpé, Teshome Gebre, Tsegaye Tewelde Gebrehiwot, Amanuel Tesfay Gebremedhin, Johanna M Geleijnse, Bradford D Gessner, Katherine B Gibney, Ibrahim Abdelmageem Mohamed Ginawi, Ababi Zergaw Giref, Maurice Giroud, Melkamu Dedefo Gishu, Giorgia Giussani, Elizabeth Glaser, William W Godwin, Hector Gomez-Dantes, Philimon Gona, Amador Goodridge, Sameer Vali Gopalani, Carolyn C Gotay, Atsushi Goto, Hebe N Gouda, Rebecca Grainger, Felix Greaves, Francis Guillemin, Yuming Guo, Rahul Gupta, Rajeev Gupta, Vipin Gupta, Reyna A Gutiérrez, Demewoz Haile, Alemayehu Desalegne Hailu, Gessessew Bugssa Hailu, Yara A Halasa, Randah Ribhi Hamadeh, Samer Hamidi, Mouhanad Hammami, Jamie Hancock, Alexis J Handal, Graeme J Hankey, Yuantao Hao, Hilda L Harb, Sivadasanpillai Harikrishnan, Josep Maria Haro, Rasmus Havmoeller, Roderick J Hay, Ileana Beatriz Heredia-Pi, Pouria Heydarpour, Hans W Hoek, Masako Horino, Nobuyuki Horita, H Dean Hosgood, Damian G Hoy, Aung Soe Htet, Hsiang Huang, John J Huang, Chantal Huynh, Marissa Iannarone, Kim Moesgaard Iburg, Kaire Innos, Manami Inoue, Veena J Iyer, Kathryn H Jacobsen, Nader Jahanmehr, Mihajlo B Jakovljevic, Mehdi Javanbakht, Sudha P Jayaraman, Achala Upendra Jayatilleke, Sun Ha Jee, Panniyammakal Jeemon, Paul N Jensen, Ying Jiang, Tariku Jibat, Aida Jimenez-Corona, Ye Jin, Jost B Jonas, Zubair Kabir, Yogeshwar Kalkonde, Ritul Kamal, Haidong Kan, André Karch, Corine Kakizi Karema, Chante Karimkhani, Amir Kasaeian, Anil Kaul, Norito Kawakami, Peter Njenga Keiyoro, Andrew Haddon Kemp, Andre Keren, Chandrasekharan Nair Kesavachandran, Yousef Saleh Khader, Abdur Rahman Khan Ejaz Ahmad Khan* Young-Ho Khang* strategies that we will keep expanding to meet our goal of producing the highest quality global health data to our growing audience of policy makers, researchers, the media, and the general population. www.thelancet.com Vol 388 October 8, 2016 Articles Sahil Khera, Tawﬁ k Ahmed Muthafer Khoja, Jagdish Khubchandani, Christian Kieling, Pauline Kim, Cho-il Kim, Daniel Kim, Yun Jin Kim, Niranjan Kissoon, Luke D Knibbs, Ann Kristin Knudsen, Yoshihiro Kokubo, Dhaval Kolte, Jacek A Kopec, Soewarta Kosen, Georgios A Kotsakis, Parvaiz A Koul, Ai Koyanagi, Michael Kravchenko, Barthelemy Kuate Defo, Burcu Kucuk Bicer, Andreas A Kudom, Ernst J Kuipers, G Anil Kumar, Michael Kutz, Gene F Kwan, Aparna Lal, Ratilal Lalloo, Tea Lallukka, Hilton Lam, Jennifer O Lam, Sinead M Langan, Anders Larsson, Pablo M Lavados, Janet L Leasher, James Leigh, Ricky Leung, Miriam Levi, Yichong Li, Yongmei Li, Juan Liang, Shiwei Liu, Yang Liu, Belinda K Lloyd, Warren D Lo, Giancarlo Logroscino, Katharine J Looker, Paulo A Lotufo, Raimundas Lunevicius, Ronan A Lyons, Mark T Mackay, Mohammed Magdy Abd El Razek, Mahdi Mahdavi, Marek Majdan, Azeem Majeed, Reza Malekzadeh, Wagner Marcenes, David Joel Margolis, Jose Martinez-Raga, Felix Masiye, João Massano, Stephen Theodore McGarvey, John J McGrath, Martin McKee, Brian J McMahon, Peter A Meaney, Alem Mehari, Fabiola Mejia-Rodriguez, Alemayehu B Mekonnen, Yohannes Adama Melaku, Peter Memiah, Ziad A Memish, Walter Mendoza, Atte Meretoja, Tuomo J Meretoja, Francis Apolinary Mhimbira, Anoushka Millear, Ted R Miller, Edward J Mills, Mojde Mirareﬁ n, Philip B Mitchell, Charles N Mock, Alireza Mohammadi, Shaﬁ u Mohammed, Lorenzo Monasta, Julio Cesar Montañez Hernandez, Marcella Montico, Meghan D Mooney, Maziar Moradi-Lakeh, Lidia Morawska, Ulrich O Mueller, Erin Mullany, John Everett Mumford, Michele E Murdoch, Jean B Nachega, Gabriele Nagel, Aliya Naheed, Luigi Naldi, Vinay Nangia, John N Newton, Marie Ng, Frida Namnyak Ngalesoni, Quyen Le Nguyen, Muhammad Imran Nisar, Patrick Martial Nkamedjie Pete, Joan M Nolla, Ole F Norheim, Rosana E Norman, Bo Norrving, Bruno P Nunes, Felix Akpojene Ogbo, In-Hwan Oh, Takayoshi Ohkubo, Pedro R Olivares, Bolajoko Olubukunola Olusanya, Jacob Olusegun Olusanya, Alberto Ortiz, Majdi Osman, Erika Ota, Mahesh PA, Eun-Kee Park, Mahboubeh Parsaeian, Valéria Maria de Azeredo Passos, Angel J Paternina Caicedo, Scott B Patten, George C Patton, David M Pereira, Rogelio Perez-Padilla, Norberto Perico, Konrad Pesudovs, Max Petzold, Michael Robert Phillips, Frédéric B Piel, Julian David Pillay, Farhad Pishgar, Dietrich Plass, James A Platts-Mills, Suzanne Polinder, Constance D Pond, Svetlana Popova, Richie G Poulton, Farshad Pourmalek, Dorairaj Prabhakaran, Noela M Prasad, Mostafa Qorbani, Rynaz H S Rabiee, Amir Radfar, Anwar Rafay, Kazem Rahimi, Vafa Rahimi-Movaghar, Mahfuzar Rahman, Mohammad Hifz Ur Rahman, Sajjad Ur Rahman, Rajesh Kumar Rai, Sasa Rajsic, Usha Ram, Puja Rao, Amany H Refaat, Marissa B Reitsma, Giuseppe Remuzzi, Serge Resnikoﬀ , Alex Reynolds, Antonio L Ribeiro, Maria Jesus Rios Blancas, Hirbo Shore Roba, David Rojas-Rueda, Luca Ronfani, Gholamreza Roshandel, Gregory A Roth, Dietrich Rothenbacher, Ambuj Roy, Rajesh Sagar, Ramesh Sahathevan, Juan R Sanabria, Maria Dolores Sanchez-Niño, Itamar S Santos, João Vasco Santos, Rodrigo Sarmiento-Suarez, Benn Sartorius, Maheswar Satpathy, Miloje Savic, Monika Sawhney, Michael P Schaub, Maria Inês Schmidt, Ione J C Schneider, Ben Schöttker, David C Schwebel, James G Scott, Soraya Seedat, Sadaf G Sepanlou, Edson E Servan-Mori, Katya A Shackelford, Amira Shaheen, Masood Ali Shaikh, Rajesh Sharma, Upasana Sharma, Jiabin Shen, Donald S Shepard, Kevin N Sheth, Kenji Shibuya, Min-Jeong Shin, Rahman Shiri, Ivy Shiue, Mark G Shrime, Inga Dora Sigfusdottir, Diego Augusto Santos Silva, Dayane Gabriele Alves Silveira, Abhishek Singh, Jasvinder A Singh, Om Prakash Singh, Prashant Kumar Singh, Anna Sivonda, Vegard Skirbekk, Jens Christoﬀ er Skogen, Amber Sligar, Karen Sliwa, Michael Soljak, Kjetil Søreide, Reed J D Sorensen, Joan B Soriano, Luciano A Sposato, Chandrashekhar T Sreeramareddy, Vasiliki Stathopoulou, Nicholas Steel, Dan J Stein, Sahil Khera, Tawﬁ k Ahmed Muthafer Khoja, Jagdish Khubchandani, Christian Kieling, Pauline Kim, Cho-il Kim, Daniel Kim, Yun Jin Kim, Niranjan Kissoon, Luke D Knibbs, Ann Kristin Knudsen, Yoshihiro Kokubo, Dhaval Kolte, Jacek A Kopec, Soewarta Kosen, Georgios A Kotsakis, Parvaiz A Koul, Ai Koyanagi, Michael Kravchenko, Barthelemy Kuate Defo, Burcu Kucuk Bicer, Andreas A Kudom, Ernst J Kuipers, G Anil Kumar, Michael Kutz, Gene F Kwan, Aparna Lal, Ratilal Lalloo, Tea Lallukka, Hilton Lam, Jennifer O Lam, Sinead M Langan, Anders Larsson, Pablo M Lavados, Janet L Leasher, James Leigh, Ricky Leung, Miriam Levi, Yichong Li, Yongmei Li, Juan Liang, Shiwei Liu, Yang Liu, Belinda K Lloyd, Warren D Lo, Giancarlo Logroscino, Katharine J Looker, Paulo A Lotufo, Raimundas Lunevicius, Ronan A Lyons, Mark T Mackay, Mohammed Magdy Abd El Razek, Mahdi Mahdavi, Marek Majdan, Azeem Majeed, Reza Malekzadeh, Wagner Marcenes, David Joel Margolis, Jose Martinez-Raga, Felix Masiye, João Massano, Stephen Theodore McGarvey, John J McGrath, Martin McKee, Brian J McMahon, Peter A Meaney, Alem Mehari, Fabiola Mejia-Rodriguez, Alemayehu B Mekonnen, Yohannes Adama Melaku, Peter Memiah, Ziad A Memish, Walter Mendoza, Atte Meretoja, Tuomo J Meretoja, Francis Apolinary Mhimbira, Anoushka Millear, Ted R Miller, Edward J Mills, Mojde Mirareﬁ n, Philip B Mitchell, Charles N Mock, Alireza Mohammadi, Shaﬁ u Mohammed, Lorenzo Monasta, Julio Cesar Montañez Hernandez, Marcella Montico, Meghan D Mooney, Maziar Moradi-Lakeh, Lidia Morawska, Ulrich O Mueller, Erin Mullany, John Everett Mumford, Michele E Murdoch, Jean B Nachega, Gabriele Nagel, Aliya Naheed, Luigi Naldi, Vinay Nangia, John N Newton, Marie Ng, Frida Namnyak Ngalesoni, Quyen Le Nguyen, Muhammad Imran Nisar, Patrick Martial Nkamedjie Pete, Joan M Nolla, Ole F Norheim, Rosana E Norman, Bo Norrving, Bruno P Nunes, Felix Akpojene Ogbo, In-Hwan Oh, Takayoshi Ohkubo, Pedro R Olivares, Bolajoko Olubukunola Olusanya, Jacob Olusegun Olusanya, Alberto Ortiz, Majdi Osman, Erika Ota, Mahesh PA, Eun-Kee Park, Mahboubeh Parsaeian, Valéria Maria de Azeredo Passos, Angel J Paternina Caicedo, Scott B Patten, George C Patton, David M Pereira, Rogelio Perez-Padilla, Norberto Perico, Konrad Pesudovs, Max Petzold, Michael Robert Phillips, Frédéric B Piel, Julian David Pillay, Farhad Pishgar, Dietrich Plass, James A Platts-Mills, Suzanne Polinder, Constance D Pond, Svetlana Popova, Richie G Poulton, Farshad Pourmalek, Dorairaj Prabhakaran, Noela M Prasad, Mostafa Qorbani, Rynaz H S Rabiee, Amir Radfar, Anwar Rafay, Kazem Rahimi, Vafa Rahimi-Movaghar, Mahfuzar Rahman, Mohammad Hifz Ur Rahman, Sajjad Ur Rahman, Rajesh Kumar Rai, Sasa Rajsic, Usha Ram, Puja Rao, Amany H Refaat, Marissa B Reitsma, Giuseppe Remuzzi, Serge Resnikoﬀ , Alex Reynolds, Antonio L Ribeiro, Maria Jesus Rios Blancas, Hirbo Shore Roba, David Rojas-Rueda, Luca Ronfani, Gholamreza Roshandel, Gregory A Roth, Dietrich Rothenbacher, Ambuj Roy, Rajesh Sagar, Ramesh Sahathevan, Juan R Sanabria, Maria Dolores Sanchez-Niño, Itamar S Santos, João Vasco Santos, Rodrigo Sarmiento-Suarez, Benn Sartorius, Maheswar Satpathy, Miloje Savic, Monika Sawhney, Michael P Schaub, Maria Inês Schmidt, Ione J C Schneider, Ben Schöttker, David C Schwebel, James G Scott, Soraya Seedat, Sadaf G Sepanlou, Edson E Servan-Mori, Katya A Shackelford, Amira Shaheen, Masood Ali Shaikh, Rajesh Sharma, Upasana Sharma, Jiabin Shen, Donald S Shepard, Kevin N Sheth, Kenji Shibuya, Min-Jeong Shin, Rahman Shiri, Ivy Shiue, Mark G Shrime, Inga Dora Sigfusdottir, Diego Augusto Santos Silva, Dayane Gabriele Alves Silveira, Abhishek Singh, Jasvinder A Singh, Om Prakash Singh, Prashant Kumar Singh Anna Sivonda* Vegard Skirbekk* Timothy J Steiner, Sabine Steinke, Lars Stovner, Konstantinos Stroumpoulis, Bruno F Sunguya, Patrick Sur, Soumya Swaminathan, Bryan L Sykes, Cassandra E I Szoeke, Rafael Tabarés-Seisdedos, Jukka S Takala, Nikhil Tandon, David Tanne, Mohammad Tavakkoli, Bineyam Taye, Hugh R Taylor, Braden J Te Ao, Bemnet Amare Tedla, Abdullah Sulieman Terkawi, Alan J Thomson, Andrew L Thorne-Lyman, Amanda G Thrift, George D Thurston, Ruoyan Tobe-Gai, Marcello Tonelli, Roman Topor-Madry, Fotis Topouzis, Bach Xuan Tran, Thomas Truelsen , Zacharie Tsala Dimbuene, Miltiadis Tsilimbaris, Abera Kenay Tura, Emin Murat Tuzcu, Stefanos Tyrovolas, Kingsley N Ukwaja, Eduardo A Undurraga, Chigozie Jesse Uneke, Olalekan A Uthman, Coen H van Gool, Yuri Y Varakin, Tommi Vasankari, Narayanaswamy Venketasubramanian, Raj Kumar Verma, Francesco S Violante, Sergey K Vladimirov, Vasiliy Victorovich Vlassov, Stein Emil Vollset, Gregory R Wagner, Stephen G Waller, Linhong Wang, David A Watkins, Scott Weichenthal, Elisabete Weiderpass, Robert G Weintraub, Andrea Werdecker, Ronny Westerman, Richard A White, Hywel C Williams, Charles Shey Wiysonge, Charles D A Wolfe, Sungho Won, Rachel Woodbrook, Mamo Wubshet, Denis Xavier, Gelin Xu, Ajit Kumar Yadav, Lijing L Yan, Yuichiro Yano, Mehdi Yaseri, Pengpeng Ye, Henock Gebremedhin Yebyo, Paul Yip, Naohiro Yonemoto, Seok-Jun Yoon, Mustafa Z Younis, Chuanhua Yu, Zoubida Zaidi, Maysaa El Sayed Zaki, Hajo Zeeb, Maigeng Zhou, Sanjay Zodpey, Liesl Joanna Zuhlke, Christopher J L Murray. Articles Authors listed alphabetically. †Corresponding author. Timothy J Steiner, Sabine Steinke, Lars Stovner, Chuanhua Yu, Zoubida Zaidi, Maysaa El Sayed Zaki, Hajo Zeeb, Maigeng Zhou, Sanjay Zodpey, Liesl Joanna Zuhlke, Christopher J L Murray. Authors listed alphabetically. †Corresponding author. Conclusion GBD 2015 Disease and Injury Incidence and Prevalence Collaborators Theo Vos†, Christine Allen, Megha Arora, Ryan M Barber, Zulﬁ qar A Bhutta, Alexandria Brown, Austin Carter, Daniel C Casey, Fiona J Charlson, Alan Z Chen, Megan Coggeshall, Leslie Cornaby, Lalit Dandona, Daniel J Dicker, Tina Dilegge, Holly E Erskine, Alize J Ferrari, Christina Fitzmaurice, Tom Fleming, Mohammad H Forouzanfar, Nancy Fullman, Peter W Gething, Ellen M Goldberg, Nicholas Graetz, Juanita A Haagsma, Simon I Hay, Catherine O Johnson, Nicholas J Kassebaum, Toana Kawashima, Laura Kemmer, Ibrahim A Khalil, Yohannes Kinfu, Hmwe H Kyu, Janni Leung, Xiaofeng Liang, Stephen S Lim, Alan D Lopez, Rafael Lozano, Laurie Marczak, George A Mensah, Ali H Mokdad, Mohsen Naghavi, Grant Nguyen, Elaine Nsoesie, Helen Olsen, David M Pigott, Christine Pinho, Zane Rankin, Nikolas Reinig, Joshua A Salomon, Logan Sandar, Alison Smith, Jeﬀ rey Stanaway, Caitlyn Steiner, Stephanie Teeple, Bernadette A Thomas, Christopher Troeger, Joseph A Wagner, Haidong Wang, Valentine Wanga, Harvey A Whiteford, Leo Zoeckler, Amanuel Alemu Abajobir, Kalkidan Hassen Abate, Cristiana Abbafati, Kaja M Abbas, Foad Abd-Allah, Biju Abraham, Ibrahim Abubakar, Laith J Abu-Raddad, Niveen M E Abu-Rmeileh, Ilana N Ackerman, Akindele Olupelumi Adebiyi, Zanﬁ na Ademi, Arsène Kouablan Adou, Kossivi Agbelenko Afanvi, Emilie Elisabet Agardh, Arnav Agarwal, Aliasghar Ahmad Kiadaliri, Hamid Ahmadieh, Oluremi N Ajala, Rufus Olusola Akinyemi, Nadia Akseer, Ziyad Al-Aly, Khurshid Alam, Noore K M Alam, Saleh Fahed Aldhahri, Miguel Angel Alegretti, Zewdie Aderaw Alemu, Lily T Alexander, Samia Alhabib, Raghib Ali, Ala’a Alkerwi, François Alla, Peter Allebeck, Rajaa Al-Raddadi, Ubai Alsharif, Khalid A Altirkawi, Nelson Alvis-Guzman, Azmeraw T Amare, Alemayehu Amberbir, Heresh Amini, Walid Ammar, Stephen Marc Amrock, Hjalte H Andersen, Gregory M Anderson, Benjamin O Anderson, Carl Abelardo T Antonio, Atsede Fantahun Aregay, Johan Ärnlöv, Al Artaman, Hamid Asayesh, Reza Assadi, Suleman Atique, 1593 www.thelancet.com Vol 388 October 8, 2016 Articles Articles Aﬁ lliations Institute for Health Metrics and Evaluation (Prof T Vos PhD, C Allen BA, M Arora BSA, R M Barber BS, A Brown MA, A Carter BS, D C Casey BA, F J Charlson PhD, A Z Chen BS, M Coggeshall BA, L Cornaby BS, Prof L Dandona MD, D J Dicker BS, T Dilegge BS, H E Erskine PhD, A J Ferrari PhD, C Fitzmaurice MD, T Fleming BS, M H Forouzanfar MD, N Fullman MPH, E M Goldberg BS, N Graetz MPH, J A Haagsma PhD, Prof S I Hay DSc, C O Johnson PhD, N J Kassebaum MD, T Kawashima MS, L Kemmer PhD, I A Khalil MD, H H Kyu PhD, J Leung PhD, Prof S S Lim PhD, Prof A D Lopez PhD, L Marczak PhD, Prof A H Mokdad PhD, Prof M Naghavi PhD, G Nguyen MPH, E Nsoesie PhD, H Olsen MAIS, D M Pigott DPhil, C Pinho BA, Z Rankin BS, N Reinig BS, L Sandar BS, A Smith BA, J Stanaway PhD, C Steiner MPH, S Teeple BA, B A Thomas MD, C Troeger MPH, J A Wagner BS, H Wang PhD, V Wanga MS, Prof H A Whiteford PhD, L Zoeckler BA, L T Alexander BA, G M Anderson MSEE, B Bell MLIS, K Bienhoﬀ MA, S Biryukov BS, J Blore PhD, J Brown MAIS, M M Coates MPH, F Daoud BS, K Estep MPA, K Foreman PhD, J Fox BS, J Friedman BA, J Frostad MPH, W W Godwin BS, J Hancock MLS, C Huynh BA, M Iannarone MSc, P Kim BA, M Kutz BS, F Masiye PhD, A Millear BA, M Mirareﬁ n MPH, M D Mooney BS, M Moradi-Lakeh MD, E Mullany BA, J E Mumford BA, M Ng PhD, P Rao MPH, M B Reitsma BS, A Reynolds BA, G A Roth MD, K A Shackelford BA, A Sivonda MHA, A Sligar MPH, R J D Sorensen MPH, P Sur BA, Prof S E Vollset DrPH, R Woodbrook MLIS/MA, Prof M Zhou PhD, Prof C J L Murray DPhil), Harborview/UW Medicine (R G Ellenbogen MD), School of Dentistry (G A Kotsakis DDS), Harborview Injury Prevention and Research Center (C N Mock PhD), University of Washington, Seattle, WA, USA (Prof B O Anderson MD, N D Futran MD, P N Jensen PhD, D A Watkins MD); Centre of Excellence in Women and Child Health (Prof Z A Bhutta PhD), Aga Khan University, Karachi, Pakistan (M I Nisar MSc); Centre for Global Child Health, The Hospital for Sick Children, Toronto, ON, Canada (Prof Z A Bhutta PhD, N Akseer MSc); School of Public Health (F J Charlson PhD, H E Erskine PhD, A J Ferrari PhD, J Leung PhD, Prof H A Whiteford PhD, A A Abajobir MPH, L D Knibbs PhD), School of Dentistry (Prof R Lalloo PhD), Centre for Clinical Research (J G Scott PhD), The University of Queensland, Brisbane, QLD, Australia (N K M Alam MPH, H N Gouda PhD, Y Guo PhD, Prof J J McGrath MD); Queensland Centre for Mental Health Research, Brisbane, QLD, Australia (F J Charlson PhD, H E Erskine PhD, A J Ferrari PhD, J Leung PhD, Prof H A Whiteford PhD); Centre for Control of Chronic Institute for Health Metrics and Evaluation (Prof T Vos PhD, C Allen BA, M Arora BSA, R M Barber BS, A Brown MA, A Carter BS, D C Casey BA, F J Charlson PhD, A Z Chen BS, M Coggeshall BA, L Cornaby BS, Prof L Dandona MD, D J Dicker BS, T Dilegge BS, H E Erskine PhD, A J Ferrari PhD, C Fitzmaurice MD, T Fleming BS, M H Forouzanfar MD, N Fullman MPH, E M Goldberg BS, N Graetz MPH, J A Haagsma PhD, Prof S I Hay DSc, C O Johnson PhD, N J Kassebaum MD, T Kawashima MS, L Kemmer PhD, I A Khalil MD, H H Kyu PhD, J Leung PhD, Prof S S Lim PhD, Prof A D Lopez PhD, L Marczak PhD, Prof A H Mokdad PhD, Prof M Naghavi PhD, G Nguyen MPH, E Nsoesie PhD, H Olsen MAIS, D M Pigott DPhil, C Pinho BA, Z Rankin BS, N Reinig BS, L Sandar BS, A Smith BA, J Stanaway PhD, C Steiner MPH, S Teeple BA, B A Thomas MD, C Troeger MPH, J A Wagner BS, H Wang PhD, V Wanga MS, Prof H A Whiteford PhD, L Zoeckler BA, L T Alexander BA, G M Anderson MSEE, B Bell MLIS, K Bienhoﬀ MA, S Biryukov BS, J Blore PhD, J Brown MAIS, M M Coates MPH, F Daoud BS, K Estep MPA, K Foreman PhD, J Fox BS, J Friedman BA, J Frostad MPH, W W Godwin BS, J Hancock MLS, C Huynh BA, M Iannarone MSc, P Kim BA, M Kutz BS, F Masiye PhD, A Millear BA, M Mirareﬁ n MPH, M D Mooney BS, M Moradi-Lakeh MD, E Mullany BA, J E Mumford BA, M Ng PhD, P Rao MPH, M B Reitsma BS, A Reynolds BA, G A Roth MD, K A Shackelford BA, A Sivonda MHA, A Sligar MPH, R J D Sorensen MPH, P Sur BA, Prof S E Vollset DrPH, R Woodbrook MLIS/MA, Prof M Zhou PhD, Prof C J L Murray DPhil), Harborview/UW Medicine (R G Ellenbogen MD), School of Dentistry (G A Kotsakis DDS), Harborview Injury Prevention and Research Center (C N Mock PhD), University of Washington, Seattle, WA, USA (Prof B O Anderson MD, N D Futran MD, P N Jensen PhD, D A Watkins MD); Centre of Excellence in Women and Child Health (Prof Z A Bhutta PhD), Aga Khan University, Karachi, Pakistan (M I Nisar MSc); Centre for Global Child Health, The Hospital for Sick Children, Toronto, ON, Canada (Prof Z A Bhutta PhD, N Akseer MSc); School of Public Health (F J Charlson PhD, H E Erskine PhD, A J Ferrari PhD, J Leung PhD, Prof H A Whiteford PhD, A A Abajobir MPH, L D Knibbs PhD), School of Dentistry (Prof R Lalloo PhD), Centre for Clinical Research (J G Scott PhD), The University of Queensland, Brisbane, QLD, Australia (N K M Alam MPH, H N Gouda PhD, Y Guo PhD, Prof J J McGrath MD); Queensland Centre for Mental Health Research, Brisbane, QLD, Australia (F J Charlson PhD, H E Erskine PhD, A J Ferrari PhD, J Leung PhD, Prof H A Whiteford PhD); Centre for Control of Chronic Quyen Le Nguyen, Muhammad Imran Nisar, Rosana E Norman, Bo Norrving, Bruno P Nunes, Felix Akpojene Ogbo, In-Hwan Oh, Takayoshi Ohkubo, Pedro R Olivares, Bolajoko Olubukunola Olusanya, N Graetz MPH, J A Haagsma PhD, Prof S I Hay DSc, C O Johnson PhD, N J Kassebaum MD, T Kawashima MS, L Kemmer PhD, I A Khalil M H H Kyu PhD, J Leung PhD, Prof S S Lim PhD, Prof A D Lopez PhD Valéria Maria de Azeredo Passos, Angel J Paternina Caicedo, L Marczak PhD, Prof A H Mokdad PhD, Prof M Naghavi PhD, Scott B Patten, George C Patton, David M Pereira, G Nguyen MPH, E Nsoesie PhD, H Olsen MAIS, D M Pigott DPhi C Pinho BA, Z Rankin BS, N Reinig BS, L Sandar BS, A Smith BA, J Stanaway PhD, C Steiner MPH, S Teeple BA, A Thomas MD, C Troeger MPH, J A Wagner BS, H Wang PhD V Wanga MS, Prof H A Whiteford PhD, L Zoeckler BA, Svetlana Popova, Richie G Poulton, Farshad Pourmalek, T Alexander BA, G M Anderson MSEE, B Bell MLIS, K Bienhoﬀ MA Biryukov BS, J Blore PhD, J Brown MAIS, M M Coates MPH, Daoud BS, K Estep MPA, K Foreman PhD, J Fox BS, J Friedman BA Frostad MPH, W W Godwin BS, J Hancock MLS, C Huynh BA, M Iannarone MSc, P Kim BA, M Kutz BS, F Masiye PhD, A Millea M Mirareﬁ n MPH, M D Mooney BS, M Moradi-Lakeh MD, E Mullany BA, J E Mumford BA, M Ng PhD, P Rao MPH, 1594 www.thelancet.com Vol 388 October 8, 2016 Articles Sweden; Health Services Management Research Center, Institute for Futures Studies in Health, Kerman University of Medical Sciences, Kerman, Iran (A Ahmad Kiadaliri PhD); Ophthalmic Research Center (H Ahmadieh MD, M Yaseri PhD), School of Public Health (N Jahanmehr PhD), Shahid Beheshti University of Medical Sciences, Tehran, Iran; Department of Ophthalmology, Labbaﬁ nejad Medical Center, Tehran, Iran (H Ahmadieh MD); University of Pittsburgh Medical Center, McKeesport, PA, USA (O N Ajala MD); Newcastle University, Newcastle upon Tyne, UK (R O Akinyemi PhD); Washington University in Saint Louis, St Louis, MO, USA (Z Al-Aly MD); Sydney School of Public Health (Prof T R Driscoll PhD), The University of Sydney, Sydney, NSW, Australia (K Alam PhD, Prof A H Kemp PhD, J Leigh PhD, A B Mekonnen MS); Queensland Health, Brisbane, QLD, Australia (N K M Alam MPH); King Saud University, Riyadh, Saudi Arabia (S F Aldhahri MD, K A Altirkawi MD); King Fahad Medical City, Riyadh, Saudi Arabia (S F Aldhahri MD); Department of Preventive and Social Medicine, Faculty of Medicine, University of the Republic, Montevideo, Uruguay (M A Alegretti MD); Debre Markos University, Debre Markos, Ethiopia (Z A Alemu MPH); King Abdullah Bin Abdulaziz University Hospital, Riyadh, Saudi Arabia (S Alhabib PhD); Luxembourg Institute of Health (LIH), Strassen, Luxembourg (A Alkerwi PhD); School of Public Health, University of Lorraine, Nancy, France (Prof F Alla PhD, F Guillemin PhD); Department of Public Health Sciences (P Allebeck PhD, R H S Rabiee MPH), Department of Clinical Science, Intervention and Technology (Prof J J Carrero PhD), Department of Neurobiology, Care Sciences and Society (NVS) (S M Fereshtehnejad PhD), Department of Medical Epidemiology and Biostatistics (E Weiderpass PhD), Karolinska Institutet, Stockholm, Sweden (R Havmoeller PhD); Ministry of Health, Jeddah, Saudi Arabia (R Al-Raddadi PhD); Charité Universitätsmedizin, Berlin, Germany (U Alsharif MPH); Universidad de Cartagena, Cartagena de Indias, Colombia (Prof N Alvis-Guzman PhD); School of Medicine (A T Amare MPH, Y A Melaku MPH), University of Adelaide, Adelaide, SA, Australia (L G Ciobanu MS); College of Medicine and Health Sciences, Bahir Dar University, Bahir Dar, Ethiopia (A T Amare MPH); Dignitas International, Zomba, Malawi (A Amberbir PhD); Environmental Health Research Center, Kurdistan University of Medical Sciences, Sanandaj, Iran (H Amini MSPH); Department of Epidemiology and Public Health (H Amini MSPH, T Fürst PhD), Swiss Tropical and Public Health Institute, Basel, Switzerland (C K Karema MSc); Ministry of Public Health, Beirut, Lebanon (W Ammar PhD, H L Harb MPH); Oregon Health & Science University, Portland, OR, USA (S M Amrock MD); Center for Sensory-Motor Interaction, Department of Health Science and Technology, Faculty of Medicine, Aalborg University, Aalborg, Denmark (H H Andersen MSc); Department of Health Policy and Administration, College of Public Health, University of the Philippines Manila, Manila, Philippines (C A T Antonio MD); School of Public Health (Y A Melaku MPH), Mekelle University, Mekelle, Ethiopia (A F Aregay MS, B D Betsu MS, G B Hailu MSc, H G Yebyo MS); Department of Medical Sciences, Uppsala University, Uppsala, Sweden (Prof J Ärnlöv PhD, Prof A Larsson PhD); Dalarna University, Falun, Sweden (Prof J Ärnlöv PhD); Consultant, Windsor, ON, Canada (A Artaman PhD); Department of Medical Emergency, School of Paramedic, Qom University of Medical Sciences, Qom, Iran (H Asayesh PhD); Mashhad University of Medical Sciences, Mashhad, Iran (R Assadi PhD); Graduate Institute of Biomedical Informatics, Taipei Medical University, Taipei, Taiwan (S Atique MS); Institut de Recherche Clinique du Bénin, Cotonou, Benin Republic (E F G A Avokpaho MPH); Laboratoire d’Etudes et de Recherche-Action en Santé (LERAS Afrique), Parakou, Benin Republic (E F G A Avokpaho MPH); Sanjay Gandhi Postgraduate Institute of Medical Sciences, Lucknow, India (A Awasthi MSc); The Judith Lumley Centre for Mother, Infant and Family Health Research, La Trobe University, Melbourne, VIC, Australia (B P Ayala Quintanilla PhD); Peruvian National Institute of Health, Lima, Peru (B P Ayala Quintanilla PhD); Wardliparingga Aboriginal Research Unit, South Australian Health and Medical Research Institute, Adelaide, SA, Australia (P Azzopardi MEpi); School of Health Sciences, University of Management and Technology, Lahore, Pakistan (U Bacha PhD); Public Health Agency of Canada Toronto ON Canada (A Badawi PhD); Conditions (P Jeemon PhD), Public Health Foundation of India, New Delhi, India (Prof L Dandona MD, R Dandona PhD, G A Kumar PhD); Department of Zoology (P W Gething PhD), Nuﬃ eld Department of Medicine (D Bisanzio PhD, A Deribew PhD), NIHR Musculoskeletal Biomedical Research Centre (Prof C Cooper FMedSci), Oxford Big Data Institute, Li Ka Shing Centre for Health Information and Discovery (Prof S I Hay DSc), University of Oxford, Oxford, UK (R Ali FRCP, D A Bennett PhD, K Rahimi DM); Centre for Research & Action in Public Health, Faculty of Health, University of Canberra, Canberra, ACT, Australia (Y Kinfu PhD); National Center for Chronic and Noncommunicable Disease Control and Prevention (L Duan MD, Y Li MPH, S Liu PhD, Y Jin MS, Prof L Wang MD, P Ye MPH, Prof M Zhou PhD), Chinese Center for Disease Control and Prevention (Prof X Liang MD), Beijing, China; Melbourne School of Population and Global Health (Prof A D Lopez PhD), Department of Paediatrics (P Azzopardi MEpi), The Peter Doherty Institute for Infection and Immunity (K B Gibney FRACP), Department of Medicine (A Meretoja PhD), Murdoch Childrens Research Institute (K Alam PhD, P Azzopardi MEpi, R Borschmann PhD, S M Colquhoun PhD, Prof G C Patton MD, R G Weintraub MBBS), Institute of Health and Ageing (Prof C E I Szoeke PhD), The University of Melbourne, Melbourne, VIC, Australia (Z Ademi PhD, K Alam PhD, R Borschmann PhD, S M Colquhoun PhD, Prof H R Taylor AC, R G Weintraub MBBS); National Institute of Public Health, Cuernavaca, Mexico (R Lozano MD, I R Campos-Nonato PhD, J C Campuzano PhD, H Gomez-Dantes MSc, I B Heredia-Pi PhD, F Mejia-Rodriguez MD, J C Montañez Hernandez MSc, M J Rios Blancas MPH, Prof E E Servan-Mori MSc); Center for Translation Research and Implementation Science, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD, USA (G A Mensah MD); Department of Global Health and Population (Prof J A Salomon PhD), Department of Nutrition (A L Thorne-Lyman ScD), Harvard T H Chan School of Public Health (O N Ajala MD, Prof T Bärnighausen MD, I R Campos-Nonato PhD, E L Ding ScD, M S Farvid PhD, G R Wagner MD), Harvard Medical School (M Osman MD, M G Shrime MD), Harvard University, Boston, MA, USA (J R A Fitchett MD); Jimma University, Jimma, Ethiopia (K H Abate MS, T T Gebrehiwot MPH, A T Gebremedhin MPH); La Sapienza, University of Rome, Rome, Italy (C Abbafati PhD); Virginia Tech, Blacksburg, VA, USA (Prof K M Abbas PhD); Department of Neurology, Cairo University, Cairo, Egypt (Prof F Abd-Allah MD); NMSM Government College Kalpetta, Kerala, India (Prof B Abraham MPhil); Institute for Global Health (Prof I Abubakar PhD), Farr Institute of Health Informatics Research (A Banerjee DPhil), Department of Epidemiology and Public Health (H Benzian PhD), University College London, London, UK; Infectious Disease Epidemiology Group, Weill Cornell Medical College in Qatar, Doha, Qatar (L J Abu-Raddad PhD); Institute of Community and Public Health, Birzeit University, Ramallah, Palestine (N M Abu-Rmeileh PhD); Department of Epidemiology and Preventive Medicine, School of Public Health and Preventive Medicine (I N Ackerman PhD, Prof R Buchbinder PhD), School of Public Health and Preventive Medicine (Prof B Gabbe PhD), Department of Medicine, School of Clinical Sciences at Monash Health (Prof A G Thrift PhD), Monash University, Melbourne, VIC, Australia; College of Medicine (A O Adebiyi MD), University of Ibadan, Ibadan, Nigeria (R O Akinyemi PhD); University College Hospital, Ibadan, Nigeria (A O Adebiyi MD); University of Basel, Basel, Switzerland (Z Ademi PhD, T Fürst PhD); Association Ivoirienne pour le Bien-Être Familial, Abidjan, Côte d’Ivoire (A K Adou MD); Direction du District Sanitaire de Haho, Notse, Togo (K A Afanvi MD); Faculte des Sciences de Sante, Universite de Lome, Lome, Togo (K A Afanvi MD); Institution of Public Health Sciences, Stockholm, Sweden (E E Agardh PhD); Dalla Lana School of Public Health (N Akseer MSc), Department of Nutritional Sciences, Faculty of Medicine (A Badawi PhD), Centre for Addiction and Mental Health (S Popova PhD), University of Toronto, Toronto, ON, Canada (A Agarwal BHSc); McMaster University, Hamilton ON Canada (A Agarwal BHSc); Department of Clinical Conditions (P Jeemon PhD), Public Health Foundation of India, New Delhi, India (Prof L Dandona MD, R Dandona PhD, G A Kumar PhD); Department of Zoology (P W Gething PhD), Nuﬃ eld Department of Medicine (D Bisanzio PhD, A Deribew PhD), NIHR Musculoskeletal Biomedical Research Centre (Prof C Cooper FMedSci), Oxford Big Data Institute, Li Ka Shing Centre for Health Information and Discovery (Prof S I Hay DSc), University of Oxford, Oxford, UK (R Ali FRCP, D A Bennett PhD, K Rahimi DM); Centre for Research & Action in Public Health, Faculty of Health, University of Canberra, Canberra, ACT, l f h l f h d I B Heredia-Pi PhD, F Mejia-Rodriguez MD, J C Montañez Hernandez MSc, M J Rios Blancas MPH, I R Campos-Nonato PhD, E L Ding ScD, M S Farvid PhD, G R Wagner MD), Harvard Medical School (M Osman MD, M G Shrime MD), Harvard University, Boston, MA, USA (J R A Fitchett MD); Jimma University, Jimma, Ethiopia (K H Abate MS, T T Gebrehiwot MPH, A T Gebremedhin MPH); La Sapienza, University of Rome, Rome, Italy (C Abbafati PhD); Virginia Tech, Blacksburg, VA, USA (Prof K M Abbas PhD); Department of Neurology, Cairo University, Cairo, Egypt 1595 www.thelancet.com Vol 388 October 8, 2016 Articles Articles www.thelancet.com Vol 388 October 8, 2016 Articles Department of Medicine, University of Valencia/INCLIVA Health Research Institute and CIBERSAM, Valencia, Spain (F Catalá-López PhD); Clinical Epidemiology Program, Ottawa Hospital Research Institute, Ottawa, ON, Canada (F Catalá-López PhD); School of Nursing, College of Medicine, National Taiwan University, Taipei, Taiwan (Prof J Chang PhD); Clinical Governance Unit, Gold Coast Health, Southport, QLD, Australia (P P Chiang PhD); National Center for Child Health and Development, Setagaya, Japan (C E Chibueze PhD); University of Zambia, Lusaka, Zambia (V H Chisumpa MPhil, F Masiye PhD); University of Witwatersrand, Johannesburg, South Africa (V H Chisumpa MPhil); Seoul National University Medical Library, Seoul, South Korea (J J Choi PhD); Department of Public Health and Primary Care, University of Cambridge, Cambridge, UK (R Chowdhury PhD); Bispebjerg University Hospital, Copenhagen, Denmark (Prof H Christensen DMSCi); Christian Medical College, Vellore, India (Prof D J Christopher MD); University of Salerno, Baronissi, Italy (Prof M Cirillo MD); MRC Lifecourse Epidemiology Unit, University of Southampton, Southampton, UK (Prof C Cooper FMedSci); NIHR Biomedical Research Centre, University of Southampton and University Hospital Southampton NHS Foundation Trust, Southampton, UK (Prof C Cooper FMedSci); IRCCS-Istituto di Ricerche Farmacologiche Mario Negri, Bergamo, Italy (M Cortinovis Biotech D, G Giussani Biol, D N Perico MD, Prof G Remuzzi MD, ); Centre for International Health, Dunedin School of Medicine (Prof J A Crump MD), Injury Prevention Research Unit, Department of Preventive and Social Medicine, Dunedin School of Medicine (Prof S Derrett PhD), University of Otago, Dunedin, New Zealand (Prof R G Poulton PhD); Wolaita Sodo University, Wolaita Sodo, Ethiopia (S A Damtew MPH); School of Public Health (K Deribe MPH, A D Hailu MPH), Addis Ababa University, Addis Ababa, Ethiopia (S A Damtew MPH, A Z Giref PhD, D Haile MPH, T Jibat MS, B Taye PhD); Guy’s and St Thomas’ NHS Foundation Trust, London, UK (Prof P I Dargan FRCP); i3S - Instituto de Investigação e Inovação em Saúde and INEB - Instituto de Engenharia Biomédica (J das Neves PhD), Faculty of Medicine (J Massano MD, J V Santos BHlthSc); University of Porto, Porto, Portugal; Wellcome Trust Brighton & Sussex Centre for Global Health Research, Brighton, UK (Prof G Davey MD); Public Health England, London, UK (Prof A C Davis PhD, F Greaves PhD, Prof J N Newton FRCP, Prof N Steel PhD); Griﬃ th University, Brisbane, QLD, Australia (Prof D De Leo DSc); Stanford University, Stanford, CA, USA (L C Del Gobbo PhD); University of Colorado School of Medicine and the Colorado School of Public Health, Aurora, CO, USA (R P Dellavalle MD); Brighton and Sussex Medical School, Brighton, UK (K Deribe MPH); KEMRI-Wellcome Trust Research Programme, Kiliﬁ , Kenya (A Deribew PhD); Mount Sinai Beth Israel, New York, NY, USA (Prof D C Des Jarlais PhD); Icahn School of Medicine at Mount Sinai, New York, NY, USA (Prof D C Des Jarlais PhD); Department of Community Medicine, Faculty of Medicine, University of Peradeniya, Peradeniya, Sri Lanka (S D Dharmaratne MD); Centre for Control of Chronic Conditions (P Jeemon PhD), Public Health Foundation of India, Gurgaon, India (P K Dhillon PhD, P Ganguly MD, Prof S Zodpey PhD); Hospital de la Santa Creu i Sant Pau, Barcelona, Spain (C Diaz-Torné MD); International Institute for Population Sciences, Mumbai, India (M Dubey MPhil, M H U Rahman MPhil, Prof U Ram PhD, A Singh PhD, R K Verma MPhil, A K Yadav MPhil); Federal University of Rio Grande do Sul, Porto Alegre, Brazil (B B Duncan PhD, C Kieling MD, Prof M I Schmidt MD); University of North Carolina, Chapel Hill, NC, USA (B B Duncan PhD); Non-communicable Diseases Research Center, Endocrinology and Metabolism Population Sciences Institute (H Ebrahimi MD, F Pishgar MD, F Farzadfar MD, A Kasaeian PhD, M Parsaeian PhD), Liver and Pancreaticobiliary Diseases Research Center, Digestive Disease Research Institute, Shariati Hospital (H Ebrahimi MD), Multiple Sclerosis Research Center, Neuroscience Institute (P Heydarpour MD), Hematology-Oncology and Stem Cell Transplantation Research Center (A Kasaeian PhD), Digestive Diseases Research Institute (Prof R Malekzadeh MD, G Roshandel PhD, S G Sepanlou PhD), Department of Epidemiology and Biostatistics, School of Public Health (M Parsaeian PhD), Uro-Oncology Research Center (F Pishgar MD), Sina Trauma and Surgery Research Center (Prof V Rahimi-Movaghar MD), T h U i i f M di l S i T h I (M Y i PhD) Department of Environmental Health Engineering, Sri Ramachandra University, Chennai, India (K Balakrishnan PhD); Faculty of Medicine, University of Belgrade, Belgrade, Serbia (A Barac PhD); School of Psychology, University of Auckland, Auckland, New Zealand (S L Barker-Collo PhD); Africa Health Research Institute, Mtubatuba, South Africa (Prof T Bärnighausen MD); Institute of Public Health, Heidelberg University, Heidelberg, Germany (Prof T Bärnighausen MD, S Mohammed PhD); Department of Occupational and Environmental Health (Prof L Barregard MD), Health Metrics Unit (Prof M Petzold PhD), University of Gothenburg, Gothenburg, Sweden; Department of Industrial Engineering, School of Engineering, Pontiﬁ cia Universidad Javeriana, Bogotá, Colombia (L H Barrero ScD); School of Health Sciences, University of Canterbury, Christchurch, New Zealand (A Basu PhD); College of Medicine, Charles R Drew University of Medicine and Science, Los Angeles, CA, USA Medicine and Science, Los Angeles, CA, USA (Prof S Bazargan-Hejazi PhD); David Geﬀ en School of Medicine, University of California, Los Angeles, Los Angeles, CA, USA (Prof S Bazargan-Hejazi PhD); Kermanshah University of Medical Science, Kermanshah, Iran (Prof S Bazargan-Hejazi PhD); IRCCS-Istituto di Ricerche Farmacologiche Mario Negri, Milan, Italy (E Beghi MD); School of Medicine (K N Sheth MD), Yale University, New Haven, CT, USA (Prof M L Bell PhD, J J Huang MD); Internal Medicine Department (Prof I S Santos PhD), University of São Paulo, São Paulo, Brazil (I M Bensenor PhD, Prof P A Lotufo DrPH); Department of Epidemiology and Health Promotion, College of Dentistry (H Benzian PhD), New York University, New York, NY, USA; Debre Berhane University, Debre Berhan, Ethiopia (A Berhane PhD); Division of Health and Social Care Research (Prof C D Wolfe MD), King’s College London, London, UK (E Bernabé PhD, Prof R J Hay DM); College of Health and Medical Sciences (H S Roba MPH), Haramaya University, Harar, Ethiopia (A S Beyene MPH); Queen Elizabeth Hospital Birmingham, Birmingham, UK (N Bhala DPhil); University of Otago Medical School, Wellington, New Zealand (N Bhala DPhil); Department of Infectious Disease Epidemiology (T Fürst PhD), Department of Epidemiology and Biostatistics (F B Piel PhD), Division of Brain Sciences (Prof T J Steiner PhD), Imperial College London, London, UK (S Bhatt DPhil, F Greaves PhD, Prof A Majeed MD, M Medicine and Science, Los Angeles, CA, USA (Prof S Bazargan-Hejazi PhD); David Geﬀ en School of Medicine, University of California, Los Angeles, Los Angeles, CA, USA (Prof S Bazargan-Hejazi PhD); Kermanshah University of Medical Science, Kermanshah, Iran (Prof S Bazargan-Hejazi PhD); IRCCS-Istituto di Ricerche Farmacologiche Mario Negri, Milan, Italy (E Beghi MD); School of Medicine (K N Sheth MD), Yale University, New Haven, CT, USA (Prof M L Bell PhD, J J Huang MD); Internal Medicine Department (Prof I S Santos PhD), University of São Paulo, São Paulo, Brazil (I M Bensenor PhD, Prof P A Lotufo DrPH); Department of Epidemiology and Health Promotion, College of Dentistry (H Benzian PhD), New York University, New York, NY, USA; Debre Berhane University, Debre Berhan, Ethiopia (A Berhane PhD); Division of Health and Social Care Research (Prof C D Wolfe MD), King’s College London, London, UK (E Bernabé PhD, Prof R J Hay DM); College of Health and Medical Sciences (H S Roba MPH), Haramaya University, Harar, Ethiopia (A S Beyene MPH); Queen Elizabeth Hospital Birmingham, Birmingham, UK (N Bhala DPhil); University of Otago Medical School, Wellington, New Zealand (N Bhala DPhil); Department of Infectious Disease Epidemiology (T Fürst PhD), Department of Epidemiology and Biostatistics (F B Piel PhD), Division of Brain Sciences (Prof T J Steiner PhD), Imperial College London, London, UK (S Bhatt DPhil, F Greaves PhD, Prof A Majeed MD, M Soljak PhD); Independent Public Health Consultants, Addis Ababa, Ethiopia (S Biadgilign MPH); Department of Nephrology Issues of Transplanted Kidney, Academician V I Shumakov Federal Research Center of Transplantology and Artiﬁ cial Organs, Moscow, Russia (B Bikbov MD); Department of Community Medicine (Prof E Bjertness PhD), University of Oslo, Oslo, Norway (A S Htet MPhil); Transport and Road Safety (TARS) Research (S Boufous PhD), National Drug and Alcohol Research Centre (Prof L Degenhardt PhD), Brien Holden Vision Institute (Prof S Resnikoﬀ MD), University of New South Wales, Kensington, NSW, Australia (B Calabria PhD, Prof P B Mitchell MD); Danube-University Krems, Krems, Austria (Prof M Brainin PhD); Faculty of Health Sciences and Social Work, Department of Public Health, Trnava University, Trnava, Slovakia (A Brazinova PhD, M Majdan PhD); International Neurotrauma Research Organization, Vienna, Austria (A Brazinova PhD); Departments of Pediatrics and Neurology (W D Lo MD), College of Medicine (J Shen PhD), The Ohio State University, Columbus, OH, USA (Prof N J K Breitborde PhD); Monash Department of Clinical Epidemiology, Cabrini Institute, Melbourne, VIC, Australia (Prof R Buchbinder PhD); University of California, San Francisco, San Francisco, CA, USA (G C Buckle MD); Al Shifa Trust Eye Hospital, Rawalpindi, Pakistan (Z A Butt PhD); National Centre for Epidemiology and Population Health, Australian National University, Canberra, ACT, Australia (B Calabria PhD, A Lal PhD); Department of Biostatistics and Epidemiology, University of Oklahoma Health Sciences Center, Oklahoma City, OK, USA (H Carabin PhD); Metropolitan Autonomous University, Mexico City, Mexico (R Cárdenas ScD); University at Albany, Rensselaer, NY, USA (Prof D O Carpenter MD); Colombian National Health Observatory, Instituto Nacional de Salud, Bogotá, Colombia (C A Castañeda-Orjuela MSc); Epidemiology and Public Health Evaluation Group, Public Health Department, Universidad Nacional de Colombia, Bogotá, Colombia (C A Castañeda-Orjuela MSc); Caja Costarricense de Seguro Social, San Jose, Costa Rica (Prof J Castillo Rivas MPH); Universidad de Costa Rica, San Pedro, Montes de Oca, Costa Rica (Prof J Castillo Rivas MPH); 1596 www.thelancet.com Vol 388 October 8, 2016 Articles Bergen, Norway (A D Hailu MPH, Prof O F Norheim PhD); Kilte Awlaelo Health and Demographic Surveillance System, Mekelle, Ethiopia (G B Hailu MSc); Arabian Gulf University, Manama, Bahrain (Prof R R Hamadeh DPhil); Hamdan Bin Mohammed Smart University, Dubai, United Arab Emirates (S Hamidi PhD); Wayne County Department of Health and Human Services, Detroit, MI, USA (M Hammami MD); University of New Mexico, Albuquerque, NM, USA (A J Handal PhD); School of Medicine and Pharmacology, University of Western Australia, Perth, WA, Australia (Prof G J Hankey MD); Harry Perkins Institute of Medical Research, Nedlands, WA, Australia (Prof G J Hankey MD); Western Australian Neuroscience Research Institute, Nedlands, WA, Australia (Prof G J Hankey MD); School of Public Health, Sun Yat-sen University, Guangzhou, China (Prof Y Hao PhD); Sree Chitra Tirunal Institute for Medical Sciences and Technology, Trivandrum, India (S Harikrishnan DM); Parc Sanitari Sant Joan de Déu - CIBERSAM, Sant Boi de Llobregat (Barcelona), Spain (J M Haro MD); Universitat de Barcelona, Barcelona, Spain (J M Haro MD); International Foundation for Dermatology, London, UK (Prof R J Hay DM); Department of Psychiatry, University Medical Center Groningen (Prof H W Hoek MD), University of Groningen, Groningen, Netherlands (A K Tura MPH); Department of Epidemiology, Mailman School of Public Health (Prof H W Hoek MD), Columbia University, New York, NY, USA (Prof V Skirbekk PhD); Nevada Division of Public and Behavioral Health, Department of Health and Human Services, Carson City, NV, USA (M Horino MPH); Department of Pulmonology, Yokohama City University Graduate School of Medicine, Yokohama, Japan (N Horita MD); Albert Einstein College of Medicine, Bronx, NY, USA (Prof H D Hosgood PhD); Public Health Division, The Paciﬁ c Community, Noumea, New Caledonia (D G Hoy PhD); International Relations Division, Ministry of Health, Nay Pyi Taw, Myanmar (A S Htet MPhil); Cambridge Health Alliance, Cambridge, MA, USA (H Huang MD); Aarhus University, Aarhus, Denmark (K M Iburg PhD); National Institute for Health Development, Tallinn, Estonia (K Innos PhD); Graduate School of Medicine (M Inoue MD), School of Public Health (Prof N Kawakami MD), University of Tokyo, Tokyo, Japan (K Shibuya MD); Department of Global and Community Health, George Mason University, Fairfax, VA, USA (K H Jacobsen PhD); Faculty of Medical Sciences, University of Kragujevac, Kragujevac, Serbia (Prof M B Jakovljevic PhD); University of Aberdeen, Aberdeen, UK (M Javanbakht PhD); Department of Surgery, Virginia Commonwealth University, Richmond, VA, USA (S P Jayaraman MD); Postgraduate Institute of Medicine, Colombo, Sri Lanka (A U Jayatilleke PhD); Institute of Violence and Injury Prevention, Colombo, Sri Lanka (A U Jayatilleke PhD); Graduate School of Public Health, Yonsei University, Seoul, South Korea (Prof S H Jee PhD); Centre for Chronic Disease Control, New Delhi, India (P Jeemon PhD, D Prabhakaran DM); Department of Health Development, Institute of Industrial Ecological Sciences, University of Occupational and Environmental Health, Kitakyushu, Japan (Y Jiang PhD); Department of Ocular Epidemiology and Visual Health, Institute of Ophthalmology Conde de Valencia, Mexico City, Mexico (A Jimenez-Corona PhD); General Directorate of Epidemiology, Ministry of Health, Mexico City, Mexico (A Jimenez- Corona PhD); Department of Ophthalmology, Medical Faculty Mannheim, Ruprecht-Karls-University Heidelberg, Mannheim, Germany (Prof J B Jonas MD); University College Cork, Cork, Ireland (Z Kabir PhD); Society for Education, Action and Research in Community Health, Gadchiroli, India (Y Kalkonde MD); CSIR-Indian Institute of Toxicology Research, Lucknow, India (R Kamal MSc, C N Kesavachandran PhD); Fudan University, Shanghai, China (H Kan MD); Epidemiological and Statistical Methods Research Group, Helmholtz Centre for Infection Research, Braunschweig, Germany (A Karch MD); Hannover-Braunschweig Site, German Center for Infection Research, Braunschweig, Germany (A Karch MD); Quality and Equity Health Care, Kigali, Rwanda (C K Karema MSc); Case Western University Hospitals, Cleveland, OH, USA (C Karimkhani MD); Oklahoma State University, Tulsa, OK, USA (A Kaul MD); Institute of Tropical and Infectious Diseases, Nairobi, Kenya (P N Keiyoro PhD); School of Continuing and Distance Education, Nairobi, Kenya (P N Keiyoro PhD); Farr Institute (Prof R A Lyons MD), Swansea University, Swansea, UK (Prof A H Kemp PhD); Assuta Hospitals, Assuta Hashalom Tel Aviv Israel (Prof A Keren MD); Jordan University Eijkman-Oxford Clinical Research Unit, Jakarta, Indonesia (I Elyazar PhD); Charité University Medicine Berlin, Berlin, Germany (Prof M Endres MD); Arba Minch University, Arba Minch, Ethiopia (A Y Endries MPH); The Institute of Social and Economic Studies of Population, Russian Academy of Sciences, Moscow, Russia (Prof S P Ermakov DSc); Federal Research Institute for Health Organization and Informatics, Ministry of Health of the Russian Federation, Moscow, Russia (Prof S P Ermakov DSc); Ministry of Health and Medical Education, Tehran, Iran (B Eshrati PhD); Arak University of Medical Sciences, Arak, Iran (B Eshrati PhD); University of Louisville, Louisville, KY, USA (T A Farid MD, A R Khan MD); DGS Directorate General of Health, Lisboa, Portugal (C S E S Farinha MSc); Universidade Aberta, Lisboa, Portugal (C S E S Farinha MSc); Federal University of Sergipe, Aracaju, Brazil (Prof A Faro PhD); Harvard/MGH Center on Genomics, Vulnerable Populations, and Health Disparities, Mongan Institute for Health Policy, Massachusetts General Hospital, Boston, MA, USA (M S Farvid PhD); National Institute for Stroke and Applied Neurosciences (V L Feigin PhD), Auckland University of Technology, Auckland, New Zealand (B J Te Ao MPH); School of Medicine (G F Kwan MD), Boston University, Boston, MA, USA (Prof D T Felson MD); Institute of Education and Sciences, German Hospital Oswaldo Cruz, São Paulo, Brazil (Prof J G Fernandes PhD); Centre for Experimental Medicine & Rheumatology, William Harvey Research Institute, Barts and The London School of Medicine & Dentistry, Queen Mary University of London, London, UK (J C Fernandes PhD); Bielefeld University, Bielefeld, Germany (F Fischer MPH); Alzheimer Scotland Dementia Research Centre (I Shiue PhD), University of Edinburgh, Edinburgh, UK (Prof F G R Fowkes PhD); James Cook University, Townsville, QLD, Australia (R C Franklin PhD); Department of Infectious Disease Epidemiology (T Fürst PhD), Department of Epidemiology and Biostatistics (F B Piel PhD), Division of Brain Sciences (Prof T J Steiner PhD), Imperial College London, London, UK (F Greaves PhD, Prof A Majeed MD, M Soljak PhD); Indian Institute of Public Health Gandhinagar, Ahmedabad, India (P Ganguly MD, V J Iyer MPH); Leras Afrique, Cotonou, Benin (F G Gankpé MD); CHU Hassan II, Fès, Morocco (F G Gankpé MD); The Task Force for Global Health, Decatur, GA, USA (T Gebre PhD); Ludwig Maximilians University, Munich, Germany (A T Gebremedhin MPH); Division of Human Nutrition (J M Geleijnse PhD), Wageningen University, Wageningen, Netherlands (T Jibat MS); Agence de Médecine Préventive, Paris, France (B D Gessner MD); The Royal Melbourne Hospital, Melbourne, VIC, Australia (K B Gibney FRACP); College of Medicine, University of Hail, Hail, Saudi Arabia (I A Ginawi MD); University Hospital of Dijon, Dijon, France (Prof M Giroud MD); College of Health and Medical Sciences (H S Roba MPH), Haramaya University, Dire Dawa, Ethiopia (M D Gishu MS, A K Tura MPH); Kersa Health and Demographic Surveillance System, Harar, Ethiopia (M D Gishu MS); Heller School for Social Policy and Management (E Glaser PhD), Brandeis University, Waltham, MA, USA (Y A Halasa MS, Prof D S Shepard PhD, E A Undurraga PhD); University of Massachusetts Boston, Boston, MA, USA (Prof P Gona PhD); Instituto de Investigaciones Cientiﬁ cas y Servicios de Alta Tecnologia- INDICASAT-AIP, Ciudad del Saber, Panamá (A Goodridge PhD); Department of Health and Social Aﬀ airs, Government of the Federated States of Micronesia, Palikir, Federated States of Micronesia (S V Gopalani MPH); University of British Columbia, Vancouver, BC, Canada (C C Gotay PhD, Prof N Kissoon MD, J A Kopec PhD, F Pourmalek PhD); Division of Epidemiology, Center for Public Health Sciences (A Goto PhD), National Cancer Center, Tokyo, Japan (M Inoue MD); Centre for International Health, Dunedin School of Medicine (Prof J A Crump MD), Injury Prevention Research Unit, Department of Preventive and Social Medicine, Dunedin School of Medicine (Prof S Derrett PhD), University of Otago, Wellington, New Zealand (R Grainger PhD); West Virginia Bureau for Public Health, Charleston, WV, USA (R Gupta MD); Eternal Heart Care Centre and R h I i J i I di (R G PhD) D f 1597 www.thelancet.com Vol 388 October 8, 2016 Articles Articles www.thelancet.com Vol 388 October 8, 2016 Articles Manami Inoue is the beneﬁ ciary of a ﬁ nancial contribution from the AXA Research fund as chair holder of the AXA Department of Health and Human Security, Graduate School of Medicine, The University of Tokyo from Nov 1, 2012; the AXA Research Fund has no role in this work. Sinead M Langan holds an NIHR Clinician Scientist Fellowship (NIHR/CS/010/014); the views expressed in this publication are those of the authors and not necessarily those of the NHS, the NIHR, or the UK Department of Health. Scott Weichenthal acknowledges ﬁ nancial support from the Cancer Research Society of Canada. Yogeshwar Kalkonde is a Wellcome Trust/ DBT India Alliance Intermediate Fellow in Public Health. John J McGrath received a NHMRC John Cade Fellowship (APP1056929). Sarah Derrett reports grants, personal fees, non-ﬁ nancial support and other from EuroQol Research Foundation, outside the submitted work. Dan J Stein reports personal fees from Lundbeck, personal fees from Novartis, personal fees from AMBRF, grants from NRGF, personal fees from Biocodex, personal fees from Sevier, grants from MRC, personal fees from SUN, and personal fees from CIPLA, outside the submitted work. Tea Lallukka reports funding from The Academy of Finland, grant #287488. Charles D A Wolfe’s research was funded and supported by the National Institute for Health Research (NIHR) Biomedical Research Centre based at Guy’s and St Thomas’ NHS Foundation Trust and King’s College London. The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR, or the Department of Health. The other authors declare no competing interests. Declaration of interests Bruce Bartholow Duncan and Maria Inês Schmidt have received additional funding from the Brazilian Ministry of Health (Process No 25000192049/2014-14). Itamar S Santos reports grants from FAPESP (Brazilian public agency), outside the submitted work. Carl Abelardo T Antonio reports grants, personal fees and non-ﬁ nancial support from Johnson & Johnson (Philippines), Inc, outside the submitted work. Cyrus Cooper reports other from Alliance for Better Bone Health, other from Amgen, other from Eli Lilly, other from GSK, other from Medtronic, other from Merck, other from Novartis, other from Pﬁ zer, other from Roche, other from Servier, outside the submitted work. Walter Mendoza is currently employed by the Peru Country Oﬃ ce of the United Nations Population Fund, an institution which does not necessarily endorse this study. Donald S Shepard would like to acknowledge grant support from Sanoﬁ Pasteur. Rafael Tabarés-Seisdedos and Ferrán Catalá-López are supported in part by grant PROMETEOII/2015/021 from Generalitat Valenciana, and Rafael Tabarés-Seisdedos is supported by the national grant PI14/00894 from ISCIII-FEDER. Walter Mendoza is currently employed by the Peru Country Oﬃ ce of the United Nations Population Fund, an institution which does not necessarily endorse this study. Veena J Iyer has received a Public Health Research Initiative Fellowship (2014-17) from the Department of Science and Technology (DST) for a project titled “Relationship between Enteric Fever incidence and Climate in the city of Ahmedabad, 1990–2014”. Pablo M Lavados reports grants, personal fees and non-ﬁ nancial support from BAYER, non-ﬁ nancial support from Boehringer Ingelheim, grants and personal fees from AstraZeneca, grants from CONICYT, and grants from The George Institute for Global Health, outside the submitted work. Noela M Prasad reports and is an employee of an international NGO that raises funds from the Australian Public, and holds a honorary position at CERA, which receives Operational Infrastructure Support from the Victorian Government. Contributors h h Christopher J L Murray and Theo Vos prepared the ﬁ rst draft. Alan D Lopez and Christopher J L Murray conceived the study and provided overall guidance. All other authors provided data, developed models, reviewed results, initiated modelling infrastructure, and/or reviewed and contributed to the report. Articles Scott Weichenthal acknowledges ﬁ nancial support from the Cancer Research Society of Canada. Yogeshwar Kalkonde is a Wellcome Trust/ DBT India Alliance Intermediate Fellow in Public Health. John J McGrath received a NHMRC John Cade Fellowship (APP1056929). Sarah Derrett reports grants, personal fees, non-ﬁ nancial support and other from EuroQol Research Foundation, outside the submitted work. Dan J Stein reports personal fees from Lundbeck, personal fees from Novartis, personal fees from AMBRF, grants from NRGF, personal fees from Biocodex, personal fees from Sevier, grants from MRC, personal fees from SUN, and personal fees from CIPLA, outside the submitted work. Tea Lallukka reports funding from The Academy of Finland, grant #287488. Charles D A Wolfe’s research was funded and supported by the National Institute for Health Research (NIHR) Biomedical Research Centre based at Guy’s and St Thomas’ NHS Foundation Trust and King’s College London. The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR, or the Department of Health. The other authors declare no competing interests. Acknowledgments We would like to thank the countless individuals who have contributed to th Gl b l B d f Di St d 2015 i i iti Thi Articles Bradford D Gessner reports grants from Crucell, GSK, Hilleman Labs, Novartis, Pﬁ zer, Merck, and Sanoﬁ Pasteur, outside the submitted work. Ai Koyanagi’s work is supported by the Miguel Servet contract ﬁ nanced by the CP13/00150 and PI15/00862 projects, integrated into the National R + D + I and funded by the ISCIII - General Branch Evaluation and Promotion of Health Research - and the European Regional Development Fund (ERDF-FEDER). Dorairaj Prabhakaran reports grants from the Bill & Melinda Gates Foundation. Matthias Endres reports that The Center for Stroke Research Berlin has received institutional funding from the German Ministry for Research and Education (BMBF). Katharine J Looker has received funding from the World Health Organization for the HSV-2 seroprevalence review which informs this work; during the study, KJL also received separate funding from the World Health Organization, USAID/PATH, Sexual Health 24 and the National Institute for Health Research (NIHR) Health Protection Research Unit (HPRU) in Evaluation of Interventions at the University of Bristol; these funders had no role in the writing of the manuscript nor the decision to submit it for publication, and the views expressed in this Article do not necessarily represent the views, decisions or policies of the World Health Organization, the NHS, the NIHR, the Department of Health or Public Health England. Donal Bisanzio is supported by Bill and Melinda Gates Foundation (#OPP1068048). Thomas Fürst has received ﬁ nancial support from the Swiss National Science Foundation (SNSF; project no P300P3-154634). Rodrigo Sarmiento-Suarez receives institutional support from Universidad de Ciencias Aplicadas y Ambientales, UDCA, Bogotá Colombia. Ronan A Lyons is supported by two grants: The Farr Institute of Health Informatics Research: Arthritis Research UK, the British Heart Foundation, Cancer Research UK, the Economic and Social Research Council, the Engineering and Physical Sciences Research Council, the Medical Research Council, the National Institute of Health Research, the National Institute for Social Care and Health Research (Welsh Assembly Government), the Chief Scientist Oﬃ ce (Scottish Government Health Directorates), and The Wellcome Trust. Grant No MR/K006525/1, and the National Centre for Population Health and Wellbeing Research. Health and Care Research Wales. Stefanos Tyrovolas’s work is supported by the Foundation for Education and European Culture (IPEP), the Sara Borrell postdoctoral programme (reference no CD15/00019 from the Instituto de Salud Carlos III (ISCIII - Spain) and the Fondos Europeo de Desarrollo Regional (FEDER). Articles of Science and Technology, Irbid, Jordan (Prof Y S Khader ScD); Health Services Academy, Islamabad, Pakistan (E A Khan MPH); College of Medicine (Prof Y H Khang MD), Graduate School of Public Health (Prof S Won PhD), Seoul National University, Seoul, South Korea; New York Medical College, Valhalla, NY, USA (S Khera MD, M Tavakkoli MD); Executive Board of the Health Ministers’ Council for Cooperation Council States, Riyadh, Saudi Arabia (Prof T A M Khoja FRCP); Ball State University, Muncie, IN, USA (J Khubchandani PhD); Hospital de Clínicas de Porto Alegre, Porto Alegre, Brazil (C Kieling MD); Korea Health Industry Development Institute, Cheongju-si, South Korea (C Kim PhD); Department of Health Sciences, Northeastern University, Boston, MA, USA (Prof D Kim DrPH); Southern University College, Skudai, Malaysia (Y J Kim PhD); Centre for Disease Burden (A K Knudsen PhD, Prof S E Vollset DrPH), Department of Health Promotion (J C Skogen PhD), Norwegian Institute of Public Health, Oslo, Norway (M Savic PhD, Prof V Skirbekk PhD); Department of Preventive Cardiology, National Cerebral and Cardiovascular Center, Suita, Japan (Y Kokubo PhD); Division of Cardiology (D Kolte MD), Brown University, Providence, RI, USA (Prof S T McGarvey PhD); Center for Community Empowerment, Health Policy and Humanities, NIHRD, Jakarta, Indonesia (S Kosen MD); Sher-i-Kashmir Institue of Medical Sciences, Srinagar, India (Prof P A Koul MD); Research and Development Unit, Parc Sanitari Sant Joan de Deu (CIBERSAM), Barcelona, Spain (A Koyanagi MD); Research Center of Neurology, Moscow, Russia (M Kravchenko PhD, Prof Y Y Varakin MD); Department of Demography and Public Health Research Institute (Prof B Kuate Defo PhD), Department of Social and Preventive Medicine, School of Public Health (Prof B Kuate Defo PhD), University of Montreal, Montreal, QC, Canada; Institute of Public Health, Hacettepe University, Ankara, Turkey (B Kucuk Bicer PhD); University of Cape Coast, Cape Coast, Ghana (A A Kudom PhD); Department of Public Health (S Polinder PhD), Erasmus MC, University Medical Center Rotterdam, Rotterdam, Netherlands (Prof E J Kuipers PhD); Work Organizations, Work Disability Prevention, The Finnish Institute of Occupational Health, Helsinki, Finland (T Lallukka PhD, R Shiri PhD); Department of Public Health, Faculty of Medicine (T Lallukka PhD), University of Helsinki, Helsinki, Finland (T J Meretoja PhD); Institute of Health Policy and Development Studies, National Institutes of Health, Manila, Philippines (Prof H Lam PhD); Johns Hopkins Bloomberg School of Public Health (J O Lam PhD, Prof J B Nachega PhD), Johns Hopkins University, Baltimore, MD, USA (B X Tran PhD); London School of Hygiene & Tropical Medicine, London, UK (S M Langan PhD, Prof M McKee DSc); Servicio de Neurologia, Clinica Alemana, Universidad del Desarrollo, Santiago, Chile (P M Lavados MD); College of Optometry, Nova Southeastern University, Fort Lauderdale, FL, USA (J L Leasher OD); State University of New York, Albany, Rensselaer, NY, USA (R Leung PhD); Tuscany Regional Centre for Occupational Injuries and Diseases, Florence, Italy (M Levi PhD); San Francisco VA Medical Center, San Francisco, CA, USA (Y Li PhD); National Oﬃ ce for Maternal and Child Health Surveillance, West China Second University Hospital, Sichuan University, Chengdu, China (Prof J Liang MD); Emory University, Atlanta, GA, USA (Prof Y Liu PhD, Prof M R Phillips MD); Eastern Health Clinical School (B K Lloyd PhD), Monash University, Fitzroy, VIC, Australia; Turning Point, Eastern Health, Melbourne, VIC, Australia (B K Lloyd PhD); Nationwide Children’s Hospital, Columbus, OH, USA (W D Lo MD); University of Bari, Bari, Italy (Prof G Logroscino PhD); University of Bristol, Bristol, UK (K J Looker PhD); Aintree University Hospital National Health Service Foundation Trust, Liverpool, UK (Prof R Lunevicius PhD); School of Medicine, University of Liverpool, Liverpool, UK (Prof R Lunevicius PhD); Royal Children’s Hospital, Melbourne, VIC, Australia (M T Mackay MBBS, R G Weintraub MBBS); Aswan University Hospital, Aswan Faculty of Medicine, Aswan, Egypt (M Magdy Abd El Razek MBBCh); Social Security Organization Research Institute, Tehran, Iran (M Mahdavi PhD); Institute of Health Policy and Management, Erasmus University Rotterdam, Rotterdam, Netherlands (M Mahdavi PhD); Division of Population and Patient Health, King’s College London Dental Institute, London, UK (Prof W Marcenes PhD); Pennsylvania, Philadelphia, PA, USA (D J Margolis PhD); University Hospital Doctor Peset, University of Valencia, Valencia, Spain (J Martinez-Raga PhD); CEU Cardenal Herrera University, Moncada (Valencia), Spain (J Martinez-Raga PhD); Hospital Pedro Hispano/ULS Matosinhos, Matosinhos, Portugal (J Massano MD); Alaska Native Tribal Health Consortium, Anchorage, AK, USA (B J McMahon MD); Children’s Hospital of Philadelphia, Philadelphia, PA, USA (P A Meaney MD); College of Medicine, Howard University, Washington, DC, USA (A Mehari MD); University of Gondar, Gondar, Ethiopia (A B Mekonnen MS, B A Tedla BS); University of West Florida, Pensacola, FL, USA (P Memiah PhD); Saudi Ministry of Health, Riyadh, Saudi Arabia (Prof Z A Memish MD); College of Medicine, Alfaisal University, Riyadh, Saudi Arabia (Prof Z A Memish MD); United Nations Population Fund, Lima, Peru (W Mendoza MD); Department of Neurology, Helsinki University Hospital, Helsinki, Finland (A Meretoja PhD); Helsinki University Hospital, Comprehensive Cancer Center, Breast Surgery Unit, Helsinki, Finland (T J Meretoja PhD); Ifakara Health Institute, Bagamoyo, Tanzania (F A Mhimbira MS); Paciﬁ c Institute for Research & Evaluation, Calverton, MD, USA (T R Miller PhD); Centre for Population Health, Curtin University, Perth, WA, Australia (T R Miller PhD); University of Ottawa, Ottawa, ON, Canada (E J Mills PhD); Neuroscience Research Center, Baqiyatallah University of Medical Sciences, Tehran, Iran (A Mohammadi PhD); Health Systems and Policy Research Unit, Ahmadu Bello University, Zaria, Nigeria (S Mohammed PhD); Institute for Maternal and Child Health, IRCCS “Burlo Garofolo”, Trieste, Italy (L Monasta DSc, M Montico MSc, L Ronfani PhD); Department of Community Medicine, Gastrointestinal and Liver Disease Research Center, Preventive Medicine and Public Health Research Center, Iran University of Medical Sciences, Tehran, Iran (M Moradi-Lakeh MD); International Laboratory for Air Quality and Health (L Morawska PhD), Institute of Health and Biomedical Innovation (R E Norman PhD), Queensland University of Technology, Brisbane, QLD, Australia; Competence Center Mortality- Follow-Up of the German National Cohort (A Werdecker PhD), Federal Institute for Population Research, Wiesbaden, Germany (Prof U O Mueller PhD, R Westerman PhD); Graduate School of Public Health (Prof J B Nachega PhD), Public Health Dynamics Laboratory (A J Paternina Caicedo MD), University of Pittsburgh, Pittsburgh, PA, USA; West Herts Hospitals NHS Trust, Watford, Hertfordshire, UK (M E Murdoch FRCP); Stellenbosch University, Cape Town, South Africa (Prof J B Nachega PhD, Prof S Seedat PhD, Prof C S Wiysonge PhD); Institute of Epidemiology and Medical Biometry, Ulm University, Ulm, Germany (Prof G Nagel PhD, Prof D Rothenbacher MD); International Centre for Diarrhoeal Disease Research, Bangladesh (icddr,b), Dhaka, Bangladesh (A Naheed PhD); Azienda Ospedaliera Papa Giovanni XXIII, Bergamo, Italy (Prof L Naldi MD, Prof G Remuzzi MD); Suraj Eye Institute, Nagpur, India (V Nangia MD); Ministry of Health and Social Welfare, Dar es Salaam, Tanzania (F N Ngalesoni MSc); Institute for Global Health Innovations, Duy Tan University, Da Nang, Vietnam (Q L Nguyen MD); Institute For Research, Socio-Economic Development and Communication, Yaoundé, Cameroon (P M Nkamedjie Pete MS); Hospital Universitari de Bellvitge, L’Hospitalet, Spain (J M Nolla PhD); Federal University of Pelotas, Pelotas, Brazil (Prof B P Nunes PhD); Centre for Health Research, Western Sydney University, Sydney, NSW, Australia (F A Ogbo MPH); Department of Preventive Medicine, School of Medicine, Kyung Hee University, Seoul, South Korea (Prof I Oh PhD); Teikyo University School of Medicine, Tokyo, Japan (Prof T Ohkubo MD); Universidad Autonoma de Chile, Talca, Chile (Prof P R Olivares PhD); Center for Healthy Start Initiative, Lagos, Nigeria (B O Olusanya PhD, J O Olusanya MBA); IIS-Fundacion Jimenez Diaz-UAM, Madrid, Spain (Prof A Ortiz PhD); YBank, Cambridge, MA, USA (M Osman MD); St Luke’s International University, Tokyo, Japan (E Ota PhD); JSS Medical College, JSS University, Mysore, India (Prof Mahesh PA DNB); Department of Medical Humanities and Social Medicine, College of Medicine, Kosin University, Busan, South Korea (E Park PhD); Universidade Federal de Minas Gerais, Belo Horizonte, Brazil (Prof V M D A Passos PhD); Universidad de Cartagena, Cartagena, Colombia (A J Paternina Caicedo MD); Department of Community Health Sciences (Prof S B Patten PhD), University of Calgary, Calgary, AB, Pennsylvania, Philadelphia, PA, USA (D J Margolis PhD); University Hospital Doctor Peset, University of Valencia, Valencia, Spain (J Martinez-Raga PhD); CEU Cardenal Herrera University, Moncada (Valencia), Spain (J Martinez-Raga PhD); Hospital Pedro Hispano/ULS Matosinhos, Matosinhos, Portugal (J Massano MD); Alaska Native Tribal Health Consortium, Anchorage, AK, USA (B J McMahon MD); Children’s Hospital of Philadelphia, Philadelphia, PA, USA (P A Meaney MD); College of Medicine, Howard University, Washington, DC, USA (A Mehari MD); University of Gondar, Gondar, Ethiopia (A B Mekonnen MS, B A Tedla BS); University of West Florida, Pensacola, FL, USA (P Memiah PhD); Saudi Ministry of Health, Riyadh, Saudi Arabia (Prof Z A Memish MD); College of Medicine, Alfaisal University, Riyadh, Saudi Arabia (Prof Z A Memish MD); United Nations Population Fund, Lima, Peru (W Mendoza MD); Department of Neurology, Helsinki University Hospital, Helsinki, Finland (A Meretoja PhD); Helsinki University Hospital, Comprehensive Cancer Center, Breast Surgery Unit, Helsinki, Finland (T J Meretoja PhD); Ifakara Health Institute, Bagamoyo, Tanzania (F A Mhimbira MS); Paciﬁ c Institute for Research & Evaluation, Calverton, MD, USA (T R Miller PhD); Centre for Population Health, Curtin University, Perth, WA, Australia (T R Miller PhD); University of Ottawa, Ottawa, ON, Canada (E J Mills PhD); Neuroscience Research Center, Baqiyatallah University of Medical Sciences, Tehran, Iran (A Mohammadi PhD); Health Systems and Policy Research Unit, Ahmadu Bello University, Zaria, Nigeria (S Mohammed PhD); Institute for Maternal and Child Health, IRCCS “Burlo Garofolo”, Trieste, Italy (L Monasta DSc, M Montico MSc, L Ronfani PhD); Department of Community Medicine, Gastrointestinal and Liver Disease Research Center, Preventive Medicine and Public Health Research Center, Iran University of Medical Sciences, Tehran, Iran (M Moradi-Lakeh MD); International Laboratory for Air Quality and Health (L Morawska PhD), Institute of Health and Biomedical Innovation (R E Norman PhD), Queensland University of Technology, Brisbane, QLD, Australia; Competence Center Mortality- Follow-Up of the German National Cohort (A Werdecker PhD), Federal I tit t f P l ti R h Wi b d G of Science and Technology, Irbid, Jordan (Prof Y S Khader ScD); Health Services Academy, Islamabad, Pakistan (E A Khan MPH); College of Medicine (Prof Y H Khang MD), Graduate School of Public Health (Prof S Won PhD), Seoul National University, Seoul, South Korea; New York Medical College, Valhalla, NY, USA (S Khera MD, M Tavakkoli MD); Executive Board of the Health Ministers’ Council for Cooperation Council States, Riyadh, Saudi Arabia (Prof T A M Khoja FRCP); Ball State University, Muncie, IN, USA (J Khubchandani PhD); Hospital de Clínicas de Porto Alegre, Porto Alegre, Brazil (C Kieling MD); Korea Health Industry Development Institute, Cheongju-si, South Korea (C Kim PhD); Department of Health Sciences, Northeastern University, Boston, MA, USA (Prof D Kim DrPH); Southern University College, Skudai, Malaysia (Y J Kim PhD); Centre for Disease Burden (A K Knudsen PhD, Prof S E Vollset DrPH), Department of Health Promotion (J C Skogen PhD), Norwegian Institute of Public Health, Oslo, Norway (M Savic PhD, Prof V Skirbekk PhD); Department of Preventive Cardiology, National Cerebral and Cardiovascular Center, Suita, Japan (Y Kokubo PhD); Division of Cardiology (D Kolte MD), Brown University, Providence, RI, USA (Prof S T McGarvey PhD); Center for Community Empowerment, Health Policy and Humanities, NIHRD, Jakarta, Indonesia (S Kosen MD); Sher-i-Kashmir Institue of Medical Sciences, Srinagar, India (Prof P A Koul MD); Research and Development Unit, Parc Sanitari Sant Joan de Deu (CIBERSAM), Barcelona, Spain (A Koyanagi MD); Research Center of Neurology, Moscow, Russia (M Kravchenko PhD, Prof Y Y Varakin MD); Department of Demography and Public Health Research Institute (P f B K t D f PhD) D t t f S i l d P ti Demography and Public Health Research Institute (Prof B Kuate Defo PhD), Department of Social and Preventive Medicine, School of Public Health (Prof B Kuate Defo PhD), University of Montreal, Montreal, QC, Canada; Institute of Public Health, Hacettepe University, Ankara, Turkey (B Kucuk Bicer PhD); University of Cape Coast, Cape Coast, Ghana (A A Kudom PhD); Department of Public Health (S Polinder PhD), Erasmus MC, University Medical Center Rotterdam, Rotterdam, Netherlands (Prof E J Kuipers PhD); Work Organizations, Work Disability Prevention, The Finnish Institute of Occupational Health, Helsinki, Finland (T Lallukka PhD, R Shiri PhD); Department of Public Health, Faculty of Medicine (T Lallukka PhD), University of Helsinki, Helsinki, Finland (T J Meretoja PhD); Institute of Health Policy and Development Studies, National Institutes of Health, Manila, Philippines (Prof H Lam PhD); Johns Hopkins Bloomberg School of Public Health (J O Lam PhD, Prof J B Nachega PhD), Johns Hopkins University, Baltimore, MD, USA (B X Tran PhD); London School of Hygiene & Tropical Medicine, London, UK (S M Langan PhD, Prof M McKee DSc); Servicio de Neurologia, Clinica Alemana, Universidad del Desarrollo, Santiago, Chile (P M Lavados MD); College of Optometry, Nova Southeastern University, Fort Lauderdale, FL, USA (J L Leasher OD); State University of New York, Albany, Rensselaer, NY, USA (R Leung PhD); Tuscany Regional Centre for Occupational Injuries and Diseases, Florence, Italy (M Levi PhD); San Francisco VA Medical Center, San Francisco, CA, USA (Y Li PhD); National Oﬃ ce for Maternal and Child Health Surveillance, West China Second University Hospital, Sichuan University, Chengdu, China (Prof J Liang MD); Emory University, Atlanta, GA, USA (Prof Y Liu PhD, Prof M R Phillips MD); Eastern Health Clinical School (B K Lloyd PhD), Monash University, Fitzroy, VIC, Australia; Turning Point, Eastern Health, Melbourne, VIC, Australia (B K Lloyd PhD); Nationwide Children’s Hospital, Columbus, OH, USA (W D Lo MD); University of Bari, Bari, Italy (Prof G Logroscino PhD); University of Bristol, Bristol, UK (K J Looker PhD); Aintree University Hospital National Health Service Foundation Trust, Liverpool, UK (Prof R Lunevicius PhD); School of Medicine, University of Liverpool, Liverpool, UK (Prof R Lunevicius PhD); Royal Children’s Hospital, Melbourne, VIC, Australia (M T Mackay MBBS, R G Weintraub MBBS); Aswan University Hospital, Aswan Faculty of Medicine, Aswan, Egypt (M Magdy Abd El Razek MBBCh); Social Security Organization Research Institute, Tehran, Iran (M Mahdavi PhD); Institute of Health Policy and Management, Erasmus University Rotterdam, Rotterdam, Netherlands (M Mahdavi PhD); Division of Population and Patient Health, King’s College London Dental Institute, London, UK (Prof W Marcenes PhD); Perelman School of Medicine (P A Meaney MD) University of 1598 www.thelancet.com Vol 388 October 8, 2016 Articles la Princesa, Universidad Autónoma de Madrid, Cátedra UAM-Linde, Palma de Mallorca, Spain (Prof J B Soriano PhD); Department of Clinical Neurological Sciences, Western University, London, ON, Canada (L A Sposato MD); Department of Community Medicine, International Medical University, Kuala Lumpur, Malaysia (C T Sreeramareddy MD); Attikon University Hospital, Athens, Greece (V Stathopoulou PhD); University of East Anglia, Norwich, UK (Prof N Steel PhD); South African Medical Research Council Unit on Anxiety & Stress Disorders, Cape Town, South Africa (Prof D J Stein PhD); Department of Neuroscience, Norwegian University of Science and Technology, Trondheim, Norway (Prof T J Steiner PhD, Prof L J Stovner PhD); Department of Dermatology, University Hospital Muenster, Muenster, NRW, Germany (S Steinke DrMed); Norwegian Advisory Unit on Headache, St Olavs Hospital, Trondheim, Norway (Prof L J Stovner PhD); Alexandra General Hospital of Athens, Athens, Greece (K Stroumpoulis PhD); Centre Hospitalier Public du Cotentin, Cherbourg, France (K Stroumpoulis PhD); Muhimbili University of Health and Allied Sciences, Dar es Salaam, Tanzania (B F Sunguya PhD); Indian Council of Medical Research, New Delhi, India (S Swaminathan MD); Departments of Criminology, Law & Society, Sociology, and Public Health, University of California, Irvine, Irvine, CA, USA (Prof B L Sykes PhD); Department of Medicine, University of Valencia, INCLIVA Health Research Institute and CIBERSAM, Valencia, Spain (Prof R Tabarés-Seisdedos PhD); WSH Institute, Ministry of Manpower, Singapore, Singapore (J S Takala DSc); Tampere University of Technology, Tampere, Finland (J S Takala DSc); Chaim Sheba Medical Center, Tel Hashomer, Israel (Prof D Tanne MD); Tel Aviv University, Tel Aviv, Israel (Prof D Tanne MD); James Cook University, Cairns, QLD, Australia (B A Tedla BS); Outcomes Research Consortium (A S Terkawi MD), Cleveland Clinic, Cleveland, OH, USA (Prof E M Tuzcu MD); Department of Anesthesiology, King Fahad Medical City, Riyadh, Saudi Arabia (A S Terkawi MD); Adaptive Knowledge Management, Victoria, BC, Canada (A J Thomson PhD); WorldFish, Penang, Malaysia (A L Thorne-Lyman ScD); Nelson Institute of Environmental Medicine, School of Medicine (Prof G D Thurston ScD), New York University, Tuxedo, NY, USA; National Center for Child Health and Development, Tokyo, Japan (R Tobe-Gai PhD); Institute of Public Health, Faculty of Health Sciences, Jagiellonian University Medical College, Kraków, Poland (R Topor-Madry PhD); Faculty of Health Sciences, Wroclaw Medical University, Wroclaw, Poland (R Topor-Madry PhD); Aristotle University of Thessaloniki, Thessaloniki, Greece (Prof F Topouzis PhD); Hanoi Medical University, Hanoi, Vietnam (B X Tran PhD); Department of Neurology, Rigshospitalet, University of Copenhagen, Denmark (T Truelsen DMSc); Department of Population Sciences and Development, Faculty of Economics and Management, University of Kinshasa, Kinshasa, Democratic Republic of the Congo (Z Tsala Dimbuene PhD); African Population and Health Research Center, Nairobi, Kenya (Z Tsala Dimbuene PhD); Department of Medicine, University of Crete, Heraklion, Greece (Prof M Tsilimbaris PhD); Parc Sanitari Sant Joan de Déu, Fundació Sant Joan de Déu, Universitat de Barcelona, CIBERSAM, Barcelona, Spain (S Tyrovolas PhD); Department of Internal Medicine, Federal Teaching Hospital, Abakaliki, Nigeria (K N Ukwaja MD); Ebonyi State University, Abakaliki, Nigeria (C J Uneke PhD); Warwick Medical School, University of Warwick, Coventry, UK (O A Uthman PhD); National Institute for Public Health and the Environment, Bilthoven, Netherlands (C H van Gool PhD); UKK Institute for Health Promotion Research, Tampere, Finland (Prof T Vasankari PhD); Raﬄ es Neuroscience Centre, Raﬄ es Hospital, Singapore, Singapore (N Venketasubramanian FRCP); University of Bologna, Bologna, Italy (Prof F S Violante MD); Federal Research Institute for Health Organization and Informatics, Moscow, Russia (S K Vladimirov PhD); National Research University Higher School of Economics, Moscow, Russia (Prof V V Vlassov MD); National Institute for Occupational Safety and Health, Washington, DC, USA (G R Wagner MD); Uniformed Services University of Health Sciences, Bethesda, MD, USA (Prof S G Waller MD); McGill University, Montreal, QC, Canada (S Weichenthal PhD); Department of Research, Cancer Registry of Norway, Institute of Population-Based Cancer Research, Oslo, Farmacognosia, Departamento de Química, Faculdade de Farmácia, Universidade do Porto, Porto, Portugal (Prof D M Pereira PhD); National Institute of Respiratory Diseases, Mexico City, Mexico (Prof R Perez-Padilla MD); Flinders University, Adelaide, SA, Australia (Prof K Pesudovs PhD); University of the Witwatersrand, Johannesburg, South Africa (Prof M Petzold PhD); Shanghai Jiao Tong University School of Medicine, Shanghai, China (Prof M R Phillips MD); Durban University of Technology, Durban, South Africa (J D Pillay PhD); Exposure Assessment and Environmental Health Indicators, German Environment Agency, Berlin, Germany (D Plass DrPH); Department of Anesthesiology (A S Terkawi MD), University of Virginia, Charlottesville, VA, USA (J A Platts-Mills MD); University of Newcastle, Callaghan, NSW, Australia (Prof C D Pond PhD); The Fred Hollows Foundation, Sydney, NSW, Australia (N M Prasad DO); Centre for Eye Research Australia, Melbourne, VIC, Australia (N M Prasad DO); Department of Community Medicine, School of Medicine, Alborz University of Medical Sciences, Karaj, Iran (M Qorbani PhD); A T Still University, Kirksville, MO, USA (A Radfar MD); Contech School of Public Health, Lahore, Pakistan (A Rafay MS); Research and Evaluation Division, BRAC, Dhaka, Bangladesh (M Rahman PhD); Hamad Medical Corporation, Doha, Qatar (S U Rahman FCPS); Society for Health and Demographic Surveillance, Suri, India (R K Rai MPH); ERAWEB Program, University for Health Sciences, Medical Informatics and Technology, Hall in Tirol, Austria (S Rajsic MD); Walden University, Minneapolis, MN, USA (Prof A H Refaat PhD); Suez Canal University, Ismailia, Egypt (Prof A H Refaat PhD); Department of Biomedical and Clinical Sciences L Sacco, University of Milan, Milan, Italy (Prof G Remuzzi MD); Hospital das Clinicas da Universidade Federal de Minas Gerais, Belo Horizonte, Brazil (Prof A L Ribeiro MD); (ISGlobal) Instituto de Salud Global de Barcelona, Barcelona, Spain (D Rojas-Rueda PhD); Golestan Research Center of Gastroenterology and Hepatology, Golestan University of Medical Sciences, Gorgan, Iran (G Roshandel PhD); All India Institute of Medical Sciences, New Delhi, India (A Roy DM, R Sagar MD, M Satpathy PhD, Prof N Tandon PhD); Ballarat Health Service, Ballarat, VIC, Australia (R Sahathevan PhD); Universiti Kebangsaan Malaysia Medical Centre, Kuala Lumpur, Malaysia (R Sahathevan PhD); Marshall University J Edwards School of Medicine, Huntington, WV, USA (J R Sanabria MD); Case Western Reserve University, Cleveland, OH, USA (J R Sanabria MD); IIS-Fundacion Jimenez Diaz, Madrid, Spain (M D Sanchez-Niño PhD); Universidad Ciencias Aplicadas y Ambientales, Bogotá, Colombia (A L Thorne Lyman ScD); Nelson Institute of Environmental Medicine, School of Medicine (Prof G D Thurston ScD), New York University, Tuxedo, NY, USA; National Center for Child Health and Development, Tokyo, Japan (R Tobe-Gai PhD); Institute of Public Health, Faculty of Health Sciences, Jagiellonian University Medical College, Kraków, Poland (R Topor-Madry PhD); Faculty of Health Sciences, Wroclaw Medical University, Wroclaw, Poland (R Topor-Madry PhD); Aristotle University of Thessaloniki, Thessaloniki, Greece (Prof F Topouzis PhD); Hanoi Medical University, Hanoi, Vietnam (B X Tran PhD); Department of Neurology, Rigshospitalet, University of Copenhagen, Denmark (T Truelsen DMSc); Department of Population Sciences and Development, Faculty of Economics and Management, University of Kinshasa, Kinshasa, Democratic Republic of the Congo (Z Tsala Dimbuene PhD); African Population and Health Research Center, Nairobi, Kenya (Z Tsala Dimbuene PhD); Department of Medicine, University of Crete, Heraklion, Greece (Prof M Tsilimbaris PhD); Parc Sanitari Sant Joan de Déu, Fundació Sant Joan de Déu, Universitat de Barcelona, CIBERSAM, Barcelona, Spain (S Tyrovolas PhD); Department of Internal Medicine, Federal Teaching Hospital, Abakaliki, Nigeria (K N Ukwaja MD); Ebonyi State University, Abakaliki, Nigeria (C J Uneke PhD); Warwick Medical School, University of Warwick, Coventry, UK (O A Uthman PhD); National Institute for Public Health and the Environment, Bilthoven, Netherlands (C H van Gool PhD); UKK Institute for Health Promotion Research, Tampere, Finland (Prof T Vasankari PhD); Raﬄ es Neuroscience Centre, Raﬄ es Hospital, Singapore, Singapore (N Venketasubramanian FRCP); University of Bologna, Bologna, Italy (Prof F S Violante MD); Federal Research Institute for Health Organization and Informatics, Moscow, Russia (S K Vladimirov PhD); National Research University Higher School of Economics, Moscow, Russia (Prof V V Vlassov MD); National Institute for Occupational Safety and Health, Washington, DC, USA (G R Wagner MD); Uniformed Services University of Health Sciences, Bethesda, MD, USA (Prof S G Waller MD); McGill University, Montreal, QC, Canada (S Weichenthal PhD); Department of Research, Cancer Registry of Norway, Institute of Population-Based Cancer Research, Oslo, Norway (E Weiderpass PhD); Department of Community Medicine, (R Sarmiento-Suarez MPH); University of KwaZulu-Natal, Durban, South Africa (Prof B Sartorius PhD); Marshall University, Huntington, WV, USA (M Sawhney PhD); Swiss Research Institute of Public Health and Addiction (M P Schaub PhD), University of Zurich, Zurich, Switzerland (H G Yebyo MS); Federal University of Santa Catarina, Florianópolis, Brazil (I J C Schneider PhD, D A S Silva PhD); German Cancer Research Center, Heidelberg, Germany (B Schöttker MPH); Institute of Health Care and Social Sciences, FOM University, Essen, Germany (B Schöttker MPH); University of Alabama at Birmingham, Birmingham, AL, USA (D C Schwebel PhD, J A Singh MD); Department of Public Health, An-Najah University, Nablus, Palestine (A Shaheen PhD); Independent Consultant, Karachi, Pakistan (M A Shaikh MD); Indian Institute of Technology Ropar, Rupnagar, India (R Sharma MA); ICMR National Institute of Epidemiology, Chennai, India (U Sharma MPH); Research Institute at Nationwide Children’s Hospital, Columbus, OH, USA (J Shen PhD); Department of Public Health Science, Graduate School (Prof M Shin PhD), Department of Preventive Medicine, College of Medicine (S Yoon PhD), Korea University, Seoul, South Korea; Faculty of Health and Life Sciences, Northumbria University, Newcastle upon Tyne, UK (I Shiue PhD); Reykjavik University, Reykjavik, Iceland (I D Sigfusdottir PhD); Brasília University, Brasília, Brazil (D G A Silveira MD); Department of Medicine, Institute of Medical Sciences, Banaras Hindu University, Varanasi, India (O P Singh PhD); Institute for Human Development, New Delhi, India (P K Singh PhD); Alcohol and Drug Research Western Norway (J C Skogen PhD), Stavanger University Hospital, Stavanger, Norway (K Søreide PhD); Faculty of Health Sciences, Hatter Institute for Cardiovascular Research in Africa (Prof K Sliwa PhD), Department of Psychiatry (Prof D J Stein PhD), University of Cape Town, Cape Town, South Africa (D A Watkins MD); Instituto de Investigación Hospital Universitario de 1599 www.thelancet.com Vol 388 October 8, 2016 Articles Articles Articles Acknowledgments Faculty of Health Sciences, University of Tromsø, The Arctic University of Norway, Tromsø, Norway (E Weiderpass PhD); Genetic Epidemiology Group, Folkhälsan Research Center, Helsinki, Finland (E Weiderpass PhD); German National Cohort Consortium, Heidelberg, Germany (R Westerman PhD); Department of Infectious Disease Epidemiology and Modelling (R A White PhD), Norwegian Institute of Public Health, Oslo, Norway (M Savic PhD, Prof V Skirbekk PhD); Centre of Evidence- based Dermatology, University of Nottingham, Nottingham, UK (Prof H C Williams DSc); South African Medical Research Council, Cape Town, South Africa (Prof C S Wiysonge PhD); National Institute for Health Research Comprehensive Biomedical Research Centre, Guy’s & St Thomas’ NHS Foundation Trust and King’s College London, London, UK (Prof C D Wolfe MD); St Paul’s Hospital, Millennium Medical College, Addis Ababa, Ethiopia (M Wubshet PhD); St John’s Medical College and Research Institute, Bangalore, India (Prof D Xavier MD); Department of Neurology, Jinling Hospital, Nanjing University School of Medicine, Nanjing, China (Prof G Xu PhD); Global Health Research Center, Duke Kunshan University, Kunshan, China (Prof L L Yan PhD); Department of Preventive Medicine, Northwestern University, Chicago, IL, USA (Y Yano MD); Social Work and Social Administration Department and The Hong Kong Jockey Club Centre for Suicide Research and Prevention, University of Hong Kong, Hong Kong, China (Prof P Yip PhD); Department of Biostatistics, School of Public Health, Kyoto University, Kyoto, Japan (N Yonemoto MPH); Jackson State University, Jackson, MS, USA (Prof M Z Younis DrPH); Department of Epidemiology and Biostatistics, School of Public Health (Prof C Yu PhD), Global Health Institute (Prof C Yu PhD), Wuhan University, Wuhan, China; University Hospital, Setif, Algeria (Prof Z Zaidi PhD); Faculty of Medicine, Mansoura University, Mansoura, Egypt (Prof M E Zaki PhD); Leibniz Institute for Prevention Research and Epidemiology, Bremen, Germany (Prof H Zeeb PhD); Red Cross War Memorial Children’s Hospital, Cape Town, South Africa (L J Zuhlke PhD). Bradford D Gessner reports grants from Crucell, GSK, Hilleman Labs, Novartis, Pﬁ zer, Merck, and Sanoﬁ Pasteur, outside the submitted work. Ai Koyanagi’s work is supported by the Miguel Servet contract ﬁ nanced by the CP13/00150 and PI15/00862 projects, integrated into the National R + D + I and funded by the ISCIII - General Branch Evaluation and Promotion of Health Research - and the European Regional Development Fund (ERDF-FEDER). Dorairaj Prabhakaran reports grants from the Bill & Melinda Gates Foundation. Articles Matthias Endres reports that The Center for Stroke Research Berlin has received institutional funding from the German Ministry for Research and Education (BMBF). Katharine J Looker has received funding from the World Health Organization for the HSV-2 seroprevalence review which informs this work; during the study, KJL also received separate funding from the World Health Organization, USAID/PATH, Sexual Health 24 and the National Institute for Health Research (NIHR) Health Protection Research Unit (HPRU) in Evaluation of Interventions at the University of Bristol; these funders had no role in the writing of the manuscript nor the decision to submit it for publication, and the views expressed in this Article do not necessarily represent the views, decisions or policies of the World Health Organization, the NHS, the NIHR, the Department of Health or Public Health England. Donal Bisanzio is supported by Bill and Melinda Gates Foundation (#OPP1068048). Thomas Fürst has received ﬁ nancial support from the Swiss National Science Foundation (SNSF; project no P300P3-154634). Rodrigo Sarmiento-Suarez receives institutional support from Universidad de Ciencias Aplicadas y Ambientales, UDCA, Bogotá Colombia. Ronan A Lyons is supported by two grants: The Farr Institute of Health Informatics Research: Arthritis Research UK, the British Heart Foundation, Cancer Research UK, the Economic and Social Research Council, the Engineering and Physical Sciences Research Council, the Medical Research Council, the National Institute of Health Research, the National Institute for Social Care and Health Research (Welsh Assembly Government), the Chief Scientist Oﬃ ce (Scottish Government Health Directorates), and The Wellcome Trust. Grant No MR/K006525/1, and the National Centre for Population Health and Wellbeing Research. Health and Care Research Wales. Stefanos Tyrovolas’s work is supported by the Foundation for Education and European Culture (IPEP), the Sara Borrell postdoctoral programme (reference no CD15/00019 from the Instituto de Salud Carlos III (ISCIII - Spain) and the Fondos Europeo de Desarrollo Regional (FEDER). Manami Inoue is the beneﬁ ciary of a ﬁ nancial contribution from the AXA Research fund as chair holder of the AXA Department of Health and Human Security, Graduate School of Medicine, The University of Tokyo from Nov 1, 2012; the AXA Research Fund has no role in this work. Sinead M Langan holds an NIHR Clinician Scientist Fellowship (NIHR/CS/010/014); the views expressed in this publication are those of the authors and not necessarily those of the NHS, the NIHR, or the UK Department of Health. Acknowledgments Lijing L Yan is supported by the National Natural Sciences Foundation of China grants (71233001 and 71490732). Miriam Levi would like to acknowledge the institutional support received from CeRIMP, Regional Centre for Occupational Diseases and Injuries, Tuscany Region, Florence, Italy. Simon I Hay is funded by a Senior Research Fellowship from the Wellcome Trust (#095066), and grants from the Bill & Melinda Gates Foundation (OPP1119467, OPP1093011, OPP1106023 and OPP1132415). No individuals acknowledged received additional compensation for their eﬀorts 11 Ton TGN, Mackenzie C, Molyneux DH. The burden of mental health in lymphatic ﬁ lariasis. Infect Dis Poverty 2015; 4: 34. 12 Wagner RG, Ibinda F, Tollman S, Lindholm L, Newton CR, Bertram MY. Diﬀ ering methods and deﬁ nitions inﬂ uence DALY estimates: using population-based data to calculate the burden of convulsive epilepsy in rural South Africa. PLoS One 2015; 10: e0145300. 13 Alzheimer’s Disease International. World Alzheimer Report, 2015. https://www.alz.co.uk/research/WorldAlzheimerReport2015.pdf (accessed May 26, 2016). 14 Hausman DM. Health, well-being, and measuring the burden of disease. Popul Health Metr 2012; 10: 13. 15 Nord E. Uncertainties about disability weights for the Global Burden of Disease study. Lancet Glob Health 2015; 3: e661–62. 16 GBD Mortality and Causes of Death Collaborators. Global, regional, and national life expectancy, all-cause and cause-speciﬁ c mortality for 249 causes of death, 1980–2015: a systematic analysis for the Global Burden of Disease Study 2015. Lancet 388: 1459–544. 17 Stevens GA, Alkema L, Black RE, et al. Guidelines for Accurate and Transparent Health Estimates Reporting: The GATHER statement. Lancet 2016; published online June 28. http://dx.doi.org/10.1016/ S0140-6736(16)30388-9. 18 Kiyono P. An integrative metaregression framework for descriptive epidemiology, 1 edn. Seattle: University of Washington Press, 2015. 19 Clark DV, Kibuuka H, Millard M, et al. Long-term sequelae after Ebola virus disease in Bundibugyo, Uganda: a retrospective cohort study. Lancet Infect Dis 2015; 15: 905–12. 20 Qureshi AI, Chughtai M, Loua TO, et al. Study of Ebola virus disease survivors in Guinea. Clin Infect Dis 2015; 61: 1035–42. 21 Rowe AK, Bertolli J, Khan AS, et al. Clinical, virologic, and immunologic follow-up of convalescent Ebola hemorrhagic fever patients and their household contacts, Kikwit, Democratic Republic of the Congo. J Infect Dis 1999; 179: S28–35. 22 Bwaka MA, Bonnet M-J, Calain P, et al. Ebola Hemorrhagic Fever in Kikwit, Democratic Republic of the Congo: clinical observations in 103 patients. J Infect Dis 1999; 179: S1–S7. 23 PRO-ACT. References 1 Atun R, de Jongh T, Secci F, Ohiri K, Adeyi O. Integration of targeted health interventions into health systems: a conceptual framework for analysis. Health Policy Plan 2010; 25: 104–11. 25 Global, regional, and national age-sex speciﬁ c all-cause and cause-speciﬁ c mortality for 240 causes of death, 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet 2015; 385: 117–71. 2 Park J-H, Eum J-H, Bold B, Cheong H-K. Burden of disease due to dementia in the elderly population of Korea: present and future. BMC Public Health 2013; 13: 293. 26 Murray CJL, Barber RM, Foreman KJ, et al. Global, regional, and national disability-adjusted life years (DALYs) for 306 diseases and injuries and healthy life expectancy (HALE) for 188 countries, 1990–2013: quantifying the epidemiological transition. Lancet 2015; 386: 2145–91. 3 George-Carey R, Adeloye D, Chan KY, et al. An estimate of the prevalence of dementia in Africa: a systematic analysis. J Glob Health 2012; 2: 020401. 4 Global Burden of Disease 2013 Collaborators. Global, regional, and national incidence, prevalence, and years lived with disability for 301 acute and chronic diseases and injuries in 188 countries, 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet 2015; 386: 743–800. 27 United Nations Development Programme. Human development report 2015. http://hdr.undp.org/sites/default/ﬁ les/2015_human_ development_report.pdf (accessed May 26, 2016). 28 Goldman DP, Cutler D, Rowe JW, et al. Substantial health and economic returns from delayed aging may warrant a mew focus for medical research. Health Aﬀ (Millwood) 2013; 32: 1698–705. 5 Prince MJ, Wu F, Guo Y, et al. The burden of disease in older people and implications for health policy and practice. Lancet 2015; 385: 549–62. 29 Dufouil C, Pereira E, Chêne G, et al. Older age at retirement is associated with decreased risk of dementia. Eur J Epidemiol 2014; 29: 353–61. 6 Hallett TB, Zaba B, Stover J, et al. Embracing diﬀ erent approaches to estimating HIV incidence, prevalence and mortality. AIDS 2014; 28: S523–32. 30 Belbase A, Sanzenbacher G, Gillis CM. Does age-related decline in ability correspond with retirement age? Rochester, NY: Social Science Research Network, 2015 http://papers.ssrn.com/ abstract=2665830 (accessed June 24, 2016). 7 Supervie V, Archibald CP, Costagliola D, et al. GBD 2013 and HIV incidence in high income countries. Lancet 2015; 385: 1177. 8 Cohen J, Vincent J-L, Adhikari NKJ, et al. Sepsis: a roadmap for future research. Acknowledgments https://nctu.partners.org/ProACT/Data/Index?Length=0 &LongLength=0&Rank=1&SyncRoot=System.Type%5B%5D&IsRea dOnly=False&IsFixedSize=True&IsSynchronized=False (accessed May 26, 2016). 24 Vos T, Flaxman AD, Naghavi M, et al. Years lived with disability (YLDs) for 1160 sequelae of 289 diseases and injuries 1990–2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet 2012; 380: 2163–96. Acknowledgments We would like to thank the countless individuals who have contributed to the Global Burden of Disease Study 2015 in various capacities. This paper uses data from SHARE Waves 1, 2, 3 (SHARELIFE), 4, and 5 (DOIs: 10.6103/SHARE.w1.500, 10.6103/SHARE.w2.500, 10.6103/SHARE.w3.500, 10.6103/SHARE.w4.500, 10.6103/SHARE.w5.500), see Börsch-Supan et al (2013) for methodological details. The SHARE data collection has been 1600 www.thelancet.com Vol 388 October 8, 2016 Articles primarily funded by the European Commission through FP5 (QLK6-CT-2001-00360), FP6 (SHARE-I3: RII-CT-2006-062193, COMPARE: CIT5-CT-2005-028857, SHARELIFE: CIT4-CT-2006-028812), and FP7 (SHARE-PREP: N°211909, SHARE-LEAP: N°227822, SHARE M4: N°261982). Additional funding from the German Ministry of Education and Research, the US National Institute on Aging (U01_AG09740-13S2, P01_ AG005842, P01_AG08291, P30_AG12815, R21_AG025169, Y1-AG-4553-01, IAG_BSR06-11, OGHA_04-064) and from various national funding sources is gratefully acknowledged (see www.share-project.org). The data reported here have been supplied by the United States Renal Data System (USRDS). The interpretation and reporting of these data are the responsibility of the author(s) and in no way should be seen as an oﬃ cial policy or interpretation of the US Government. The Palestinian Central Bureau of Statistics granted the researchers access to relevant data in accordance with license no. SLN2014-3-170, after subjecting data to processing aiming to preserve the conﬁ dentiality of individual data in accordance with the General Statistics Law—2000. The researchers are solely responsible for the conclusions and inferences drawn upon available data. This study has been realised using the data collectved by the Swiss Household Panel (SHP), which is based at the Swiss Centre of Expertise in the Social Sciences FORS. The project is ﬁ nanced by the Swiss National Science Foundation. The following individuals would like to acknowledge various forms of institutional support: Amanda G Thrift is supported by a fellowship from the National Health and Medical Research Council (GNT1042600). Panniyammakal Jeemon is supported by the Wellcome Trust-DBT India Alliance, Clinical and Public Health, Intermediate Fellowship (2015–20). Amador Goodridge would like to acknowledge funding for me from Sistema Nacional de Investigadores de Panamá-SNI. José das Neves was supported in his contribution to this work by a Fellowship from Fundação para a Ciência e a Tecnologia, Portugal (SFRH/BPD/92934/2013). Boris Bikbov, Monica Cortinovis, Giuseppe Remuzzi, and Norberto Perico would like to acknowledge that their contribution to this paper has been on behalf of the International Society of Nephrology (ISN) as a follow-up of the activities of the GBD 2010 Genitourinary Diseases Expert Group. Articles Time for mental health to come out of the shadows. Lancet 2016; 387: 2274–75. 60 Globocan 2012. http://globocan.iarc.fr/Default.aspx (accessed June 24, 2016). 40 Izutsu T, Tsutsumi A, Minas H, Thornicroft G, Patel V, Ito A. Mental health and wellbeing in the Sustainable Development Goals. Lancet Psychiatry 2015; 2: 1052–54. 61 Pizzarello L, Abiose A, Ffytche T, et al. VISION 2020: The Right to Sight: a global initiative to eliminate avoidable blindness. Arch Ophthalmol 1960 2004; 122: 615–20. 41 Patel V, Chisholm D, Parikh R, et al. Addressing the burden of mental, neurological, and substance use disorders: key messages from Disease Control Priorities, 3rd edn. Lancet 2016; 387: 1672–85. 62 Baltussen R, Smith A. Cost eﬀ ectiveness of strategies to combat vision and hearing loss in sub-Saharan Africa and South East Asia: mathematical modelling study. BMJ 2012; 344: e615. 42 Centers for Disease Control and Prevention. Drug-poisoning deaths involving heroin: United States, 2000–2013. http://www.cdc.gov/ nchs/products/databriefs/db190.htm (accessed June 24, 2016). 63 Currie J, Grenfell B, Farrar J. Beyond Ebola. Science 2016; 351: 815–16. 64 Baize S, Pannetier D, Oestereich L, et al. Emergence of Zaire Ebola virus disease in Guinea. N Engl J Med 2014; 371: 1418–25. 43 Centers for Disease Control and Prevention. Morbidity and Mortality Weekly Report (MMWR). https://www.cdc.gov/mmwr/ preview/mmwrhtml/mm6401a10.htm (accessed June 24, 2016). 65 Weaver SC. Arrival of Chikungunya virus in the new world: prospects for spread and impact on public health. PLoS Negl Trop Dis 2014; 8: e2921. 44 Centers for Disease Control and Prevention. Opioid overdose prevention programs providing naloxone to laypersons—United States, 2014. http://origin.glb.cdc.gov/mmwr/preview/mmwrhtml/ mm6423a2.htm?s_cid=mm6423a2_w (accessed June 24, 2016). 66 Fauci AS, Morens DM. Zika Virus in the Americas—yet another arbovirus threat. N Engl J Med 2016; 374: 601–04. 67 Microcephaly in Infants, Pernambuco State, Brazil, 2015. Emerg Infect Dis 2016; 22: 1090–93. 45 Mathers BM, Degenhardt L, Ali H, et al. HIV prevention, treatment, and care services for people who inject drugs: a systematic review of global, regional, and national coverage. Lancet 2010; 375: 1014–28. 68 Rasmussen SA, Jamieson DJ, Honein MA, Petersen LR. Zika Virus and birth defects—reviewing the evidence for causality. N Engl J Med 2016; 374: 1981–87. 46 GBD 2015 Risk Factors Collaborators. Global, regional, and national comparative risk assessment of 79 behavioural, environmental and occupational, and metabolic risks or clusters of risks, 1990–2015: a systematic analysis for the Global Burden of Disease Study 2015. Lancet 2016; 388: 1659–724. Articles 33 Begg S, Vos T, Goss J, Mann N. An alternative approach to projecting health expenditure in Australia. Aust Health Rev 2008; 32: 148–55. 52 Rocca WA, Petersen RC, Knopman DS, et al. Trends in the incidence and prevalence of Alzheimer’s disease, dementia, and cognitive impairment in the United States. Alzheimers Dement 2011; 7: 80–93. 34 Desveaux L, Beauchamp M, Goldstein R, Brooks D. Community-based exercise programs as a strategy to optimize function in chronic disease: a systematic review. Med Care 2014; 52: 216–26. 53 Brookmeyer R, Evans DA, Hebert L, et al. National estimates of the prevalence of Alzheimer’s disease in the United States. Alzheimers Dement 2011; 7: 61–73. 54 Helsper CW, Hellinga HL, van Essen GA, et al. Real-life costs of hepatitis C treatment. Neth J Med 2012; 70: 145–53. 35 Jahanbin I, Hoseini Moghadam M, Nazarinia MA, Ghodsbin F, Bagheri Z, Ashraf AR. The eﬀ ect of conditioning exercise on the health status and pain in patients with rheumatoid arthritis: a randomized controlled clinical trial. Int J Community Based Nurs Midwifery 2014; 2: 169–76. 55 Bhatt S, Weiss DJ, Cameron E, et al. The eﬀ ect of malaria control on Plasmodium falciparum in Africa between 2000 and 2015. Nature 2015; 526: 207–11. 36 Gay C, Chabaud A, Guilley E, Coudeyre E. Educating patients about the beneﬁ ts of physical activity and exercise for their hip and knee osteoarthritis. Systematic literature review. Ann Phys Rehabil Med 2016; 59: 174–83. 56 Cameron E, Battle KE, Bhatt S, et al. Deﬁ ning the relationship between infection prevalence and clinical incidence of Plasmodium falciparum malaria. Nat Commun 2015; 6: 8170. 57 Bhatt S, Weiss DJ, Mappin B, et al. Coverage and system eﬃ ciencies of insecticide-treated nets in Africa from 2000 to 2017. eLife 2015; 4: e09672. 37 McAlindon TE, Bannuru RR, Sullivan MC, et al. OARSI guidelines for the non-surgical management of knee osteoarthritis. Osteoarthris Cartilage 2014; 22: 363–88. 58 Aspiration to action. What will it take to end malaria? http:// endmalaria2040.org/assets/Aspiration-to-Action.pdf (accessed May 26, 2016). 38 Heimans L, Akdemir G, Boer KVCW, et al. Two-year results of disease activity score (DAS)-remission-steered treatment strategies aiming at drug-free remission in early arthritis patients (the IMPROVED-study). Arthritis Res Ther 2016; 18: 23. 59 WHO. World Malaria Report 2015. http://www.who.int/malaria/ publications/world-malaria-report-2015/report/en/ (accessed Aug 5, 2016). 39 Kleinman A, Estrin GL, Usmani S, et al. References Lancet Infect Dis 2015; 15: 581–614. 31 van Rijn RM, Robroek SJW, Brouwer S, Burdorf A. Inﬂ uence of poor health on exit from paid employment: a systematic review. Occup Environ Med 2014; 71: 295–301. 9 Marshall JC. Understanding the global burden of pediatric sepsis. Am J Respir Crit Care Med 2015; 191: 1096–98. 10 Anderson BO, Flanigan J. Novel methods for measuring global cancer burden: implications for global cancer control. JAMA Oncol 2015; 1: 425–27. 32 Lahelma E, Pietiläinen O, Rahkonen O, Lallukka T. Common mental disorders and cause-speciﬁ c disability retirement. Occup Environ Med 2015; 72: 181–87. 1601 www.thelancet.com Vol 388 October 8, 2016 Articles Articles Articles 69 Jensen ET, Cook SF, Allen JK, et al. Enrollment factors and bias of disease prevalence estimates in administrative claims data. Ann Epidemiol 2015; 25: 519–25. 70 Kottke TE, Baechler CJ, Parker ED. Accuracy of heart disease prevalence estimated from claims data compared with an electronic health record. Prev Chronic Dis 2012; 9: E141. 47 Association AD. Economic costs of diabetes in the US in 2012. Diabetes Care 2013; 36: 1033–46. 71 Salomon JA, Nordhagen S, Oza S, Murray CJL. Are Americans feeling less healthy? The puzzle of trends in self-rated health. Am J Epidemiol 2009; 170: 343–51. 48 NCD Risk Factor Collaboration. Worldwide trends in diabetes since 1980: a pooled analysis of 751 population-based studies with 4·4 million participants. Lancet 2016; 387: 1513–30. 72 Salomon JA, Tandon A, Murray CJL. Comparability of self rated health: cross sectional multi-country survey using anchoring vignettes. BMJ 2004; 328: 258. 49 International Diabetes Federation. IDF diabetes atlas. http://www.diabetesatlas.org/ (accessed June 25, 2016). 50 Langa KM. Is the risk of Alzheimer’s disease and dementia declining? Alzheimers Res Ther 2015; 7: 34. 73 Meijer RR. Diagnosing item score patterns on a test using item response theory-based person-ﬁ t statistics. Psychol Methods 2003; 8: 72–87. 51 Matthews FE, Arthur A, Barnes LE, et al. A two-decade comparison of prevalence of dementia in individuals aged 65 years and older from three geographical areas of England: results of the Cognitive Function and Ageing Study I and II. Lancet 2013; 382: 1405–12. 74 King G, Wand J. Comparing incomparable survey responses: new tools for anchoring vignettes. Polit Anal 2007; 15: 46–66. 1602 www.thelancet.com Vol 388 October 8, 2016
https://openalex.org/W2594135875	https://europepmc.org/articles/pmc4849087?pdf=render	English	null	Correction: Corrigendum: TET-catalyzed 5-hydroxymethylcytosine regulates gene expression in differentiating colonocytes and colon cancer	Scientific reports	2,016	cc-by	175	www.nature.com/scientificreports www.nature.com/scientificreports www.nature.com/scientificreports Scientific Reports \| 6:24963 \| DOI: 10.1038/srep24963 Corrigendum: TET-catalyzed 5-hydroxymethylcytosine regulates gene expression in differentiating colonocytes and colon cancer Christopher G. Chapman, Christopher J. Mariani, Feng Wu, Katherine Meckel, Fatma Butun, Alice Chuang, Jozef Madzo, Marc B. Bissonnette, John H. Kwon & Lucy A. Godley Scientific Reports 5:17568; doi: 10.1038/srep17568; published online 03 December 2015; updated on 28 April 2016 The original version of this Article contained a typographical error in the spelling of the author Marc B. Bissonnette, which was incorrectly given as Marc B. Bissonette. This has now been corrected in the PDF and HTML versions of the Article. This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/
https://openalex.org/W4386149673	https://revista.uepb.edu.br/REDISC/article/download/1012/1147	Portuguese	null	Atrás do pensamento: a escrita sem logos de Objeto gritantee Água viva, de Clarice Lispector	Discursividades	2,022	cc-by	8,569	Renata Coutinho Villon. Bacharel em letras português-francês; UFRJ / CNPq; renatavillon@gmail.com DISCURSIVIDADES e-ISSN 2594-6269 DISCURSIVIDADES e-ISSN 2594-6269 Behind the thought: the irrational writing of Objeto Gritante and Água viva, by Clarice Lispector Based on issues of the animal raised by Derrida (2006), the concept of perspectivism given by Viveiros de Castro (1996) and the analyses of the writer’s work made by Sammer (2020) and Roncador (2002), the article will also analyze what in Clarice’s writing seems to glimpse at a particular relationship between wo men and animals, passing also through the concept of écriture féminine as coined by Helène Cixous (1976). language, however, remains, and there appears to be in the description of an unnamed “behind the thought” a critique of logocentrism. Thus, the present article seeks to analyze what in the work seems to allude to a way of thinking about life, language, and animal beings that escapes from the rationalization principle and gives a new meaning to the term “animality”. Based on issues of the animal raised by Derrida (2006), the concept of perspectivism given by Viveiros de Castro (1996) and the analyses of the writer’s work made by Sammer (2020) and Roncador (2002), the article will also analyze what in Clarice’s writing seems to glimpse at a particular relationship between wo men and animals, passing also through the concept of écriture féminine as coined by Helène Cixous (1976). Keywords: Clarice Lispector; Helène Cixous; animal studies; female authorship. Behind the thought: the irrational writing of Objeto Gritante and Água viva, by Clarice Lispector Renata Coutinho Villon Renata Coutinho Villon Renata Coutinho Villon Renata Coutinho Villon Resumo: A obra Água viva, originalmente chamada de Objeto gritante, foi muitas vezes modificada até sua definitiva versão ser publicada. A preo- cupação de Clarice com a linguagem, no entanto, se mantém, haver na des- crição de um “atrás do pensamento” não-nomeado uma crítica ao logocen- trismo. Dessa forma, o presente artigo busca analisar o que na obra parece aludir a uma forma de pensar a vida, a linguagem e os seres animais que foge do princípio de racionalização e dá um novo sentido ao termo “animalidade”. A partir de questões do animal levantadas por Derrida (2006), o conceito de perspectivismo dado por Viveiros de Castro (1996) e as análises da obra da escritora feitas por Sammer (2020) e Roncador (2002), o artigo também analisará o que na escrita de Clarice parece entrever uma relação particular entre mulheres e animais, passando também pelo conceito de écriture fémi- nine tal como cunhado por Helène Cixous (1976). Resumo: A obra Água viva, originalmente chamada de Objeto gritante, foi muitas vezes modificada até sua definitiva versão ser publicada. A preo- cupação de Clarice com a linguagem, no entanto, se mantém, haver na des- crição de um “atrás do pensamento” não-nomeado uma crítica ao logocen- trismo. Dessa forma, o presente artigo busca analisar o que na obra parece aludir a uma forma de pensar a vida, a linguagem e os seres animais que foge do princípio de racionalização e dá um novo sentido ao termo “animalidade”. Abstract: The work Água viva, originally called Objeto gritante, was thorou ghly modified until its final version was published. Clarice’s concern with Palavras-chave: Clarice Lispector; Helène Cixous; animalidade; escrita feminin DISCURSIVIDADES s vol. 11, n. 2, e-1122207, jul-dez. 2022 DISCURSIVIDADES e-ISSN 2594-6269 language, however, remains, and there appears to be in the description of an unnamed “behind the thought” a critique of logocentrism. Thus, the present article seeks to analyze what in the work seems to allude to a way of thinking about life, language, and animal beings that escapes from the rationalization principle and gives a new meaning to the term “animality”. Introdução C larice Lispector iniciou a empreitada de escrita, que resultou na pu blicação de Água viva em 1973, com um manuscrito intitulado Atrás do pensamento: monólogo com a vida, mas cujo nome logo mudou para Objeto gritante. A oscilação de títulos é significativa por exprimir o conteúdo do próprio livro, isto é, a falta de palavras: a existência de coisas — de um it — que não podem ser nomeadas, ou que não se que rem nomeadas; enfim, esse atrás do pensamento. C l b Mas não foi apenas o título que passou por mudanças radicais, já que o livro inteiro foi revisitado, revisado, editado e reescrito de forma meticulosa. Talvez Clarice tenha pensado que, ao mandar o manuscri to a Alexandre E. Severino em 1971 para ser traduzido para o inglês, ele estivesse acabado; mas logo essa certeza se desfez, e Clarice voltou atrás em sua decisão de publicá-lo, mergulhando novamente em seu DISCURSIVIDADES s vol. 11, n. 2, e-1122207, jul-dez. 2022 2 DISCURSIVIDADES e-ISSN 2594-6269 texto vivo1. Tal decisão foi tomada após consultar um punhado de pes soas próximas, entre elas José Américo Motta Pessanha, que lhe res pondeu em formato de carta em 5 de março de 1972, questionando, de certa forma, a carga autobiográfica contida em Objeto gritante, e lhe deixando uma pergunta que parece ter sido recebida e processada com muita inquietação: Tento me explicar melhor: você se transcendia e se “resolvia” em termos de criação literária; agora a “literatura” desce a você e fica (ou aparece) como imanente ao seu cotidiano: você é seu pró prio tema como num divã de psicanalista, em que se fala, fala, sem texto previamente ensaiado. Esse encontro de você-Clarice com você-escritora certamente resulta de um processo pessoal que a levou até aí. Pergunto – se é que é justo e oportuno fazê-lo – e então, o que virá depois? Você continuará a ser seu próprio tema, diretamente apresentado, face desnuda sem as máscaras das personagens? Ou voltará a falar de si através de outras vozes, multiplicando seu mistério e sua perplexidade no jogo de espe lhos das personagens? (2019, p. 142)2. Essa inclinação e essa mistura entre Clarice e Escritora foi pontua da por diversos críticos, principalmente ao observar as últimas obras escritas pela autora. 1. Os detalhes dessa retomada ao livro e a relação com Alexandre Severino são abordados por Sônia Roncador no ensaio “Clarice Lispector esconde um objeto gritante: notas sobre um pro jeto abandonado”, presente na edição de Água viva publicada pela Rocco em 2019 e que traz manuscritos do Objeto gritante e ensaios inéditos. 2. Carta também presente na edição de Água viva anteriormente citada. 3. Outro ensaio presente na edição da Rocco de 2019, denominado “As duas versões de Água Viva”. Introdução Sônia Roncador (2002), por exemplo, fala sobre o “impulso autobiográfico” presente nos seus escritos “derradeiros”, mais ou menos a partir da escrita de Objeto. Roncador então aponta para o marco que foi a escrita desse livro para Clarice, para uma espé DISCURSIVIDADES s vol. 11, n. 2, e-1122207, jul-dez. 2022 3 DISCURSIVIDADES e-ISSN 2594-6269 cie de questionamento interno acerca da vida e da literatura — ou da vida na literatura, ou vice-versa — e de um certo embaralhamento das definições de arte que aparecem: DISCURSIVIDADES e-ISSN 2594-6269 DISCURSIVIDADES e-ISSN 2594-6269 cie de questionamento interno acerca da vida e da literatura — ou da vida na literatura, ou vice-versa — e de um certo embaralhamento das definições de arte que aparecem: cie de questionamento interno acerca da vida e da literatura — ou da vida na literatura, ou vice-versa — e de um certo embaralhamento das definições de arte que aparecem: “Objeto gritante” é, então, o registro de um momento crucial na trajetória artística de Clarice Lispector — momento esse em que a autora não só decide romper com certos aspectos da sua escrita anteriores ao manuscrito, como também questionar determina dos valores artísticos que seguramente grande parte da crítica contemporânea a “Objeto gritante” defendia. Ao violar certas regras de composição artística e ao incorporar em “Objeto gri tante” termos [...] que são um afrontamento ao decoro literário, Clarice questiona a instituição literária em geral e, em particular, sua própria obra. Além disso, o conteúdo autobiográfico de “Ob jeto gritante”, bem como a inclusão nesse texto dos vestígios das circunstâncias externas de sua produção, problematizam outros pressupostos artísticos, como, por exemplo, o ideal de sublima ção ou distanciamento do autor que nos fala Pessanha em sua carta à autora (RONCADOR, 2002, p. 90). Roncador conclui que Objeto Gritante é “resultado de um desejo de Clarice de se autoexpor” (2019, p. 160), enquanto Água Viva seria este ticamente uma obra mais próxima de seus outros escritos. Alexandre Severino atribuiu a eliminação das mais de cem páginas que estavam no manuscrito original ao fato de Clarice querer que a carga impessoal do texto, o “reflexo do ser humano”, se sobressaísse: “Foi precisamente para reduzir o mais possível o pessoal, dando maior relevo aos aspec tos impessoais do texto, que a primeira versão foi completamente mo dificada e mais tarde substituída pela versão atual [...]” (2019, p. 149)3. DISCURSIVIDADES s vol. 11, n. 3. Outro ensaio presente na edição da Rocco de 2019, denominado “As duas versões de Água Viva” Introdução 2, e-1122207, jul-dez. 2022 DISCURSIVIDADES e-ISSN 2594-6269 Tais trechos eram também recortes de crônicas de jornal que Clarice publicou em diferentes momentos, revelando assim nelas o impulso autobiográfico que não se concretizou numa publicação de Objeto. Enfim, em Água viva a narradora se torna menos Clarice e mais per sonagem, sendo a narradora uma pintora que se aventura na escrita de algo que nem ela mesma compreende totalmente, tentando mirar e acertar na palavra com todo o seu corpo: “Ao escrever não posso fa bricar como na pintura, quando fabrico artesanalmente uma cor. Mas estou tentando escrever-te com o corpo todo, enviando uma seta que se finca no ponto tenro e nevrálgico da palavra”. (2019, p. 24). Interes sa-nos observar o que nessa escrita que é feita “corporalmente”, e não “racionalmente”, pode resultar em uma crítica ao logocentrismo e o que ele implica socialmente. A visão de Clarice e o desnudamento da essência humana Esse “impulso autobiográfico” anteriormente referido também pode ser encarado como um impulso de desnudamento, tanto de si quanto literário. É também essa a impressão de Pessanha ao escrever a carta a Clarice: Tive a impressão de que você quis escrever espontaneamente, lu dicamente, a-literariamente. Verdade? Parece que, depois de re cusar os artifícios e as artimanhas da razão (melhor talvez — das racionalizações), você parece querer rejeitar os artifícios da arte. E despojar-se, ser você mesma, menos indisfarçada aos próprios olhos e aos olhos do leitor. Daí o despudor com que se mostra em seu cotidiano (mental e das circunstâncias), não se incomodando em justapor trechos de diversos níveis e sem temer o trivial. Falar de Deus e de qualquer coisa, sem selecionar tema, sem rebuscar DISCURSIVIDADES s vol. 11, n. 2, e-1122207, jul-dez. 2022 5 DISCURSIVIDADES e-ISSN 2594-6269 DISCURSIVIDADES e-ISSN 2594-6269 forma. Sem ser “escritora”. Ser apenas mulher-que-escreve-o -que-(pré)pensa-ou-pensa-sentindo? (2019, p. 139). forma. Sem ser “escritora”. Ser apenas mulher-que-escreve-o -que-(pré)pensa-ou-pensa-sentindo? (2019, p. 139). Esse “despojamento” ou “desnudamento” capaz de fazer entrever um outro eu, liberto das convenções literárias ou demasiadamente racionais (presa aos “artifícios” e “artimanhas da razão”), é algo com que Clarice parece ter flertado em algumas de suas obras, tratando de uma vivência entre suas personagens e o mundo sempre num limite de nudez. Efetivamente, é em uma de suas últimas obras, A hora da estrela (1977), que o narrador Rodrigo S.M fala abertamente de um desnudamento literário: “Escrevo sobre o mínimo parco enfeitando-o com púrpura, joias e esplendor. É assim que se escreve? Não, não é acumulando e sim desnudando. Mas tenho medo da nudez, pois ela é a palavra final.” (p. 84). Haveria então uma preocupação em “florear” a palavra, torná-la menos ela mesma, de certa forma, menos “assustadora” e mais supor tável aos próprios olhos e aos olhos do leitor; mas haveria, ao mesmo tempo, um enfrentamento desse medo da palavra pura, nua. E mesmo que isso se evidencie em muitos de seus escritos, esse enfrentamento é ainda mais significativo no Objeto e também em Água viva, pela sua constante preocupação com o “atrás do pensamento”, o lugar no qual as palavras conhecidas, demasiadamente racionalizadas, são incapa zes de descrever o que quer que seja: “Bem sei que terei que parar. A visão de Clarice e o desnudamento da essência humana Não por falta de palavras, mas porque estas coisas e sobretudo as que só penso e não escrevi – não se dizem” (2019, p. 92). Mas há também, além da impossibilidade, uma recusa a nomear, como se o próprio ato de nomeação dessas coisas que simplesmente são, sem precisarem ser explicadas, fosse, de certa forma, uma volta às racionalizações: DISCURSIVIDADES s vol. 11, n. 2, e-1122207, jul-dez. 2022 6 DISCURSIVIDADES e-ISSN 2594-6269 SCURSIVIDADES N 2594-6269 Há muita coisa a dizer que não sei como dizer. Faltam as palavras. Mas recuso-me a inventar novas: as que existem já devem dizer o que se consegue dizer e o que é proibido. E o que é proibido eu adivinho. Se houver força. Atrás do pensamento não há palavras: é-se. Minha pintura não tem palavras: fica atrás do pensamento. Nesse terreno do é-se sou puro êxtase cristalino. É-se. Sou-me. Tu te és (2019, p. 40). Há muita coisa a dizer que não sei como dizer. Faltam as palavras. Mas recuso-me a inventar novas: as que existem já devem dizer o que se consegue dizer e o que é proibido. E o que é proibido eu adivinho. Se houver força. Atrás do pensamento não há palavras: é-se. Minha pintura não tem palavras: fica atrás do pensamento. Nesse terreno do é-se sou puro êxtase cristalino. É-se. Sou-me. Tu te és (2019, p. 40). Parece que, quanto mais nuas as palavras, mais próximas a esse es tado indizível do atrás do pensamento se aproximam. E na escrita de Água viva, sem rumo, cheia de interrupções e entrecortes, essa pre ocupação com as palavras se torna evidente, mas ainda assim com o medo do inominável e do que ocorre nessa espécie de pré-linguagem que existe antes do pensamento: Atrás do pensamento atinjo um estado. Recuso-me a dividi-lo em palavras – e o que não posso e não quero exprimir fica sendo o mais secreto dos meus segredos. Sei que tenho medo de momen tos nos quais não uso o pensamento e é um momentâneo estado difícil de ser alcançado, e que, todo secreto, não usa mais as pa lavras com que se produzem pensamentos. Não usar palavras é perder a identidade? é se perder nas essenciais trevas daninhas? (2019, p. 79). 4 ç p 5. Derrida destrincha cada um desses componentes que teriam atestado a passividade animal diante da sabedoria humana: a perpétua nudez do animal diante do humano, vestido; a visão animal ignorada, por mais penetrante que seja; e a falta de linguagem dos animais, junto ao fato do homem (ligado à figura de Adão na tradição judaico-cristã) tê-los nomeado, eximindo -os da possibilidade de autoidentificação e resposta diante da figura do homem. 6. A nudez da linguagem também entra em questão nesta obra de Derrida, que abre o semi nário assim: “No princípio — gostaria de me confiar a palavras que sejam, se possível fosse, nuas”. Ele evoca o Gênese e o nu na filosofia, além de evocar um desnudamento que seria típico dos animais e do qual ele gostaria de se aproximar, ou do qual não conseguiria não se aproximar, ao ser encarado nu pelo seu gato. . Os extratos retirados dessa edição francesa de 2006 são todos traduzidos por mim 7. “Sobrenatureza, Água viva: o conceito à luz da ficção de Clarice Lispector”. In: Remate de Males, V. 40, n.1. Campinas: 2020. 8. In: Revista Mana, vol. 2, n. 2. 1996. A visão de Clarice e o desnudamento da essência humana Parece haver um retorno à noção cartesiana: se o “penso, logo exis to” é verdade, então o estado indizível do pré-pensamento seria um tipo de inexistência ou falta de identidade — e, consequentemente, de perda da humanidade. Apesar do abandono das racionalizações em curso na obra clariceana, ainda há o reconhecimento do medo de per der totalmente essa identidade humana. No entanto, esse medo so mente se aplica ao se pensar no abandono dessa identidade tal como DISCURSIVIDADES s vol. 11, n. 2, e-1122207, jul-dez. 2022 DISCURSIVIDADES e-ISSN 2594-6269 ela é tida na sociedade ocidental antropocêntrica. Jacques Derrida, em L’animal que donc je suis (2006)4, questiona esse constructo em tor no da superioridade humana, que necessariamente coloca seres não -humanos numa posição de assujeitamento e até de subjugamento. A explicação cartesiana para distinguir animais racionais de irracionais é posta em dúvida por Derrida, trazendo questionamentos do próprio Descartes acerca de sua célebre frase: “mas não conheço ainda bastan te claramente o que sou, eu que estou certo de que sou” (1983, p. 7). De todos os temas abordados por Derrida a respeito da questão animal, três parecem descrever precisamente aquilo que separaria o animal do humano: a nudez, a linguagem e a visão5. E são também esses os temas que parecem entrar em questão na obra clariceana, e em especial em Objeto e em Água Viva: o desnudamento de si e da linguagem6, mas também a visão de algo não totalmente compreendi do; o contato, por meio dos sentidos, com o próprio âmago desnudo da vida — o it, que também seria um embaralhamento dos conceitos de humanidade e desumanidade: “Estou terrivelmente lúcida e parece que alcanço um plano mais alto de humanidade. Ou da desumanidade – o it.” (2019, p. 63). Assim, a obra clariceana parece abordar e se aproximar dos con ceitos de perspectivismo ameríndio e sobrenatureza, como aponta a 6. A nudez da linguagem também entra em questão nesta obra de Derrida, que abre o semi nário assim: “No princípio — gostaria de me confiar a palavras que sejam, se possível fosse, nuas”. Ele evoca o Gênese e o nu na filosofia, além de evocar um desnudamento que seria típico dos animais e do qual ele gostaria de se aproximar, ou do qual não conseguiria não se aproximar, ao ser encarado nu pelo seu gato. DISCURSIVIDADES s vol. 11, n. 2, e-1122207, jul-dez. tureza, Água viva: o conceito à luz da ficção de Clarice Lispector”. In: Remate de 0, n.1. Campinas: 2020. A visão de Clarice e o desnudamento da essência humana 2022 8 DISCURSIVIDADES e-ISSN 2594-6269 historiadora Renata Sammer (2020)7. O termo “sobrenatureza”, como ela explica, é retomado a partir de questões caras à antropologia, prin cipalmente acerca do embate entre natureza e cultura — que está no centro da formulação antropocêntrica, colocando o homem como aquele capaz de dominar a natureza para construir cultura, enquanto os outros seres seriam desprovidos da última — e o que desse embate é posto em dúvida, por exemplo, por Eduardo Viveiros de Castro a partir de sua contribuição à antropologia. Em “Os pronomes cosmológicos e o perspectivismo ameríndio”8, Viveiros de Castro (1996) lista algumas oposições que se enquadram nos rótulos de natureza e cultura; entre eles se encontra “animalidade e humanidade”, “corpo e espírito”. E conclui que essas distinções clássicas não se aplicam quando se trata de cosmologias não-ocidentais como a por ele referida, onde “a condi ção original comum aos humanos e animais não é a animalidade, mas a humanidade” (p. 119). Basicamente, o perspectivismo parte do princípio de que os animais possuem uma condição humana, e se veem como humanos, sendo sua aparência exterior apenas uma máscara animal. Isso não significa di zer que todos os seres na configuração ameríndia se veem como iguais, mas sim que todos os seres são vistos como sujeitos portadores de pon tos de vista, apesar das diferenças exteriores e motivações divergentes: É sujeito quem tem alma, e tem alma quem é capaz de um ponto de vista. As “almas” ameríndias, humanas ou animais, são assim categorias perspectivas, dêiticos cosmológicos cuja análise pede menos uma psicologia animista ou uma ontologia substancialista que uma teoria do signo ou uma pragmática epistemológica. É sujeito quem tem alma, e tem alma quem é capaz de um ponto de vista. As “almas” ameríndias, humanas ou animais, são assim categorias perspectivas, dêiticos cosmológicos cuja análise pede menos uma psicologia animista ou uma ontologia substancialista que uma teoria do signo ou uma pragmática epistemológica. DISCURSIVIDADES s vol. 11, n. 2, e-1122207, jul-dez. 2022 9 DISCURSIVIDADES e-ISSN 2594-6269 Todo ser a que se atribui um ponto de vista será assim sujeito, espírito; ou melhor, ali onde estiver o ponto de vista, também estará a posição de sujeito (p. 126). A visão de Clarice e o desnudamento da essência humana DISCURSIVIDADES e-ISSN 2594-6269 DISCURSIVIDADES e-ISSN 2594-6269 Todo ser a que se atribui um ponto de vista será assim sujeito, espírito; ou melhor, ali onde estiver o ponto de vista, também estará a posição de sujeito (p. 126). Essa visão perspectivista é capaz de chamar a atenção da sociedade ocidentalizada ao animal, além de fazer com que neles se reconheça uma possível alteridade e subjetividade, tal como seria também o con selho de Derrida ao falar sobre a experiência de ser olhado por seu gato e não conseguir ignorar a potência daquele olhar, assim como tantos teóricos antropocêntricos fizeram: Haveria, em primeiro lugar, os textos assinados por pessoas que sem dúvida viram, observaram, analisaram, refletiram o animal mas nunca se viram vistas pelo animal; jamais cruzaram o olhar de um animal pousado sobre elas (para não dizer sobre sua nu dez); mas mesmo que se tenham visto vistas, um dia, furtivamen te, pelo animal, elas absolutamente não o levaram em considera ção (temática, teórica, filosófica); não puderam ou quiseram tirar nenhuma consequência sistemática do fato de que um animal pudesse, encarando-as, olhá-las, vestidas ou nuas, e, em uma pa lavra, sem palavras dirigir-se a elas, absolutamente não levaram em consideração o fato de que o que chamam de “animal” pudes se olhá-las e dirigir-se a elas de lá de baixo, com base em uma origem completamente outra. (2002, p. 32). A partir deste movimento de aceitação de subjetividade num ser animal, surge também o questionamento em relação à própria posição central dos seres humanos no mundo, com a tendência de excluir, de menosprezar ou até mesmo de eliminar alteridades que não sejam suficientemente vistas à nossa semelhança. Assim, todo o debate a res peito da aceitação de outros seres apenas mostra a potência da obra DISCURSIVIDADES s vol. 11, n. 2, e-1122207, jul-dez. 2022 10 DISCURSIVIDADES e-ISSN 2594-6269 clariceana neste sentido, que está a todo momento incluindo de forma central seres não humanos em suas narrativas, com personagens que se indagam a respeito da natureza ao seu redor e buscam verdadeira mente vê-la, se deixando também serem afetadas por ela9. Sammer (2020), a partir do conceito de sobrenatureza10, confirma o que seria a compreensão de mundo de Clarice: “o mundo humano é apenas mais um mundo entre outros, um modo de ser entre outras — múltiplas — naturezas” (p. 214). 9. Para nos restringirmos a um exemplo, temos G.H que, repensando o momento de epifania que se dá no decorrer da obra, admite ter “humanizado demais a vida”; ao olhar nos olhos da barata e provar sua essência, disse: “Eu, corpo neutro de barata, eu com uma vida que finalmente não me escapa pois enfim a vejo fora de mim — eu sou a barata [...]” (1964, p. 50). 10. Em suas palavras: “A sobrenatureza é o domínio do sobrenatural. Onde falam os animais, onde ocorrem as metamorfoses, os fenômenos que as ciências da natureza não podem expli car. Sobrenatural é tudo aquilo que, por não respeitar as leis da causalidade postuladas pelas ciências modernas da natureza, foi legado ao mundo da cultura” (p. 208). A visão de Clarice e o desnudamento da essência humana Assim, a escrita de Clarice está cons tantemente fluindo por entre esses múltiplos mundos e formas de ser no mundo, adentrando os conceitos de animalidade e humanidade, habitando o it que é território das metamorfoses: “A sobrenatureza é o que torna possíveis as transformações e o contato entre mundos distintos. Bichos, humanos e plantas partilham um mesmo e único it” (SAMMER, 2020, p. 214). Clarice parece não se conformar com apenas a sua forma de existência, adentrando assim nas mais diferentes vidas e temporalidades, indo até o pré-histórico em busca do it essencial: Dentro da caverna obscura tremeluzem pendurados os ratos com asas em forma de cruz dos morcegos. Vejo aranhas penugentas e negras. Ratos e ratazanas correm espantados pelo chão e pelas paredes. Entre as pedras o escorpião. Caranguejos, iguais a eles mesmos desde a pré-história, através de mortes e nascimentos, pareceriam bestas ameaçadoras se fossem do tamanho de um DISCURSIVIDADES s vol. 11, n. 2, e-1122207, jul-dez. 2022 11 DISCURSIVIDADES e-ISSN 2594-6269 DISCURSIVIDADES e-ISSN 2594-6269 homem. Baratas velhas se arrastam na penumbra. E tudo isso sou eu (LISPECTOR, 2019, p. 27). homem. Baratas velhas se arrastam na penumbra. E tudo isso sou eu (LISPECTOR, 2019, p. 27). Esse é tanto o trabalho de visão quanto o de desnudamento que se dá na obra de Clarice: buscar, por meio da escrita, modos de vida que não são exclusivamente humanos. Apesar do apagamento da pessoa lidade — da “autobiografia” — que se dá entre Objeto gritante e Água viva, a carga do pessoal ainda está lá, nos desejos e identificações que não parecem se aplicar somente às personagens, mas também à própria Clarice — uma experiência que transborda os limites de si mesma, abar cando outras existências e sensações. Como diz ainda Sammer: “[…] a escritora não ambiciona apreender a vida que é fluida como água, mas desdobrá-la ao ritmo do ‘it’, para assim circundar e cultivar o inefável, o que não está contido na palavra” (2020, p. 218). E na busca pelo “it”, a observação da vida animal se torna essencial nessa apreensão do atrás do pensamento, do desnudamento inevitável de si e da palavra: Preciso sentir de novo o it dos animais. Há muito tempo não en tro em contato com a vida primitiva animálica. Estou precisan do estudar bichos. A visão de Clarice e o desnudamento da essência humana Quero captar o it para poder pintar não uma águia e um cavalo, mas um cavalo com asas abertas de grande águia. Arrepio-me toda ao entrar em contato físico com bichos ou com a simples visão deles. Os bichos me fantasticam. Eles são o tempo que não se conta. Pareço ter certo horror daquela criatura viva que não é humana e que tem meus próprios instintos em bora livres e indomáveis. Animal nunca substitui uma coisa por outra (LISPECTOR, 2019, p. 58). Tanto no manuscrito de Objeto gritante quanto na crônica “Bichos (1)”, Clarice, falando em termos muito próximos, acrescenta ainda 12 DISCURSIVIDADES e-ISSN 2594-6269 um trecho a este extrato: “E move-se, esta coisa viva. Move-se inde pendente e por força desta coisa sem nome que é Vida”11. Ao invés de encontrar superioridades na vida humana, Clarice reconhece nos ani mais uma vida mais em contato com o it puro, parecendo ver na falta de palavras dessas criaturas o tão procurado estado do atrás do pensa mento, desejando assim atender ao chamado ancestral e animalesco: “Não ter nascido bicho é uma minha secreta nostalgia. Eles às vezes clamam do longe de muitas gerações e eu não posso responder senão ficando inquieta. É o chamado” (2019, p. 61). Esse trecho aparece tan to no Objeto quanto no desfecho da crônica “Bichos (2)”, substituindo apenas a palavra “inquieta” por “desassossegada”12. Se isso não fosse o suficiente para ligar essas palavras a um desejo autêntico — “autobio gráfico” — de Clarice, o fato de ela estar se referindo ao seu cachorro Dilermando e sua relação única de puro amor no momento do extrato anterior seria suficiente para sanar qualquer dúvida13. Esse “chamado” que se apresenta a Clarice aparece também em ou tro ponto da obra no formato de um “grito ancestral”: Às vezes eletrizo-me ao ver bicho. Estou agora ouvindo o grito ancestral dentro de mim: parece que não sei quem é mais a cria tura, se eu ou o bicho. E confundo-me toda. Fico ao que parece com medo de encarar instintos abafados que diante do bicho sou 13. No Objeto, o nome Dilermando está rabiscado e substituído por “meu cão”, indício do subsequente apagamento de tão exacerbada pessoalidade da obra que seria de fato publicada. 11. Presente na página 23 do manuscrito do Objeto, e na página 201 do livro Todas as crônicas (2018), que compila as crônicas publicadas por Clarice ao longo de sua vida. “Bichos (1)” foi publicada inicialmente em 13 de março de 1971. 11. Presente na página 23 do manuscrito do Objeto, e na página 201 do livro Todas as crônicas (2018), que compila as crônicas publicadas por Clarice ao longo de sua vida. “Bichos (1)” foi publicada inicialmente em 13 de março de 1971. 12. Na página 32 do manuscrito e 204 do livro de crônicas (2018). 13. No Objeto, o nome Dilermando está rabiscado e substituído por “meu cão”, indício do subsequente apagamento de tão exacerbada pessoalidade da obra que seria de fato publicada. Nas crônicas, Clarice indica que o amor que ela e Dilermando partilharam chegou a exacerbar aquele das relações entre humanos: “Nenhum ser humano me deu jamais a sensação de ser totalmente amada como fui amada sem restrições por esse cão” (2018, p. 202). A visão de Clarice e o desnudamento da essência humana Nas crônicas, Clarice indica que o amor que ela e Dilermando partilharam chegou a exacerbar aquele das relações entre humanos: “Nenhum ser humano me deu jamais a sensação de ser totalmente amada como fui amada sem restrições por esse cão” (2018, p. 202). DISCURSIVIDADES s vol. 11, n. 2, e-1122207, jul-dez. 2022 13 DISCURSIVIDADES e-ISSN 2594-6269 RSIVIDADES 6269 obrigada a assumir. [...] Não humanizo bicho porque é ofensa há de respeitar-lhe a natureza — eu é que me animalizo. Não é difícil e vem simplesmente. É só não lutar contra e é só entregar-se. Nada existe de mais difícil do que entregar-se ao instante. Esta dificuldade é dor humana. É nossa. Eu me entrego em palavras e me entrego quando pinto. (2019, p. 59) obrigada a assumir. [...] Não humanizo bicho porque é ofensa há de respeitar-lhe a natureza — eu é que me animalizo. Não é difícil e vem simplesmente. É só não lutar contra e é só entregar-se. Nada existe de mais difícil do que entregar-se ao instante. Esta dificuldade é dor humana. É nossa. Eu me entrego em palavras e me entrego quando pinto. (2019, p. 59) Esse trecho também se repete, com algumas diferenças, tanto no Objeto quanto na crônica já citada “Bichos (1)”, complementando al gumas noções que aparecem no extrato de Água Viva e comprovando, novamente, a inevitável pessoalidade e intimidade na forma como Cla rice escreve sobre os bichos: Ter bicho é uma experiência vital. E a quem não conviveu com um animal falta um certo tipo de intuição do mundo vivo. Quem se recusa à visão de um bicho está com medo de si próprio. Mas às vezes me arrepio vendo um bicho. Sim, às vezes sinto o mudo grito ancestral dentro de mim quando estou com eles: pa rece que não sei mais quem é o animal, se eu ou o bicho, e me confundo toda, fico ao que parece com medo de encarar meus próprios instintos abafados que diante do bicho, sou obrigada a assumir, exigentes como são, que se há de fazer, pobre de nós. [...] Não humanizo bicho porque é ofensa, há de respeitar-lhes a natura — eu é que me animalizo. Não é difícil e vem simplesmen te. É só não lutar contra e é só entregar-se. Mas indo bem mais fundo atinjo a conclusão de que nada existe de mais difícil que entregar-se totalmente. A visão de Clarice e o desnudamento da essência humana Esta dificuldade é dor humana14. 14. Página 27 no manuscrito e 202 em Todas as crônicas (2018). DISCURSIVIDADES s vol. 11, n. 2, e-1122207, jul-dez. 2022 14 DISCURSIVIDADES e-ISSN 2594-6269 Como dito anteriormente, a narradora de Água viva tem medo des se abandono das racionalidades, assim como a Clarice das crônicas, que mesmo assim não abre mão desse contato consigo mesma que vem com o conhecimento dos animais. O “arrepio”, o “eletrificar-se”, é também comum entre Clarice e personagem, assim como a confusão entre ser humano e ser animal que ocorre a partir dessa visão. Diante de um animal (voltando, novamente, a Derrida), não se pode ignorar o olhar que inevitavelmente diz algo: há então o medo da iden tificação com uma vida não-humana, desse mudo grito ancestral — é interessante notar que no extrato da crônica e do Objeto há o enfoque na mudez desse grito —, desses exigentes instintos que, sem palavras, impulsionam, por trás do pensar. Há também a animalização voluntária de si, esse entregar-se aos instintos e ao chamado sem palavras, diante do qual a humanidade — ou o constructo humano que existe na socieda de — ainda luta para manter-se, causando a dor inevitável da vida “hu mana”. Mas em Água viva a narradora continua: esse entregar-se vem através da arte, do produzir por meio de palavras ou da pintura. 15. Essa identificação parece ser atestada por diversos pensadores e escritores, entre eles a própria Clarice. G.H, ao olhar a barata que está esmagada pela cintura, admite só conseguir reconhecê-la como fêmea por conta desse mesmo esmagamento; no conto “Uma galinha”, publicado em Laços de família (1960), a galinha pode muito bem ser vista como uma analogia animal do feminino, partindo dessa mesma identificação entre o animal feito para o consumo e a mulher feita para servir ao lar, numa outra espécie de consumo, atestada também por Ca rol J. Adams em Políticas sexuais da carne (2018). Além disso, o próprio Derrida postula que o ódio ao animal é estendido, de certa forma, ao ódio da feminidade, e afirma que “o mal do animal é o macho” (2006, p. 144), animal este que no contexto judaico-cristão é subjugado ao homem assim como a mulher de certa forma também o é. Escrita feminina como combate ao logocentrismo Essa arte que inverte os papéis e as formas do que se pensa ser o humano, o animal, a própria natureza, revolve a escrita e as palavras, além da própria forma como o mundo é visto e a incapacidade de des crevê-lo, seria então uma arte nascida do contato com esse “chamado” para além das línguas, mas por trás do pensar, e também, de certa DISCURSIVIDADES s vol. 11, n. 2, e-1122207, jul-dez. 2022 15 DISCURSIVIDADES e-ISSN 2594-6269 forma, de uma certa identificação entre o feminino e o animal15. Isso nos remeteria também ao conceito de écriture féminine, postulado por Helène Cixous em “The laugh of the Medusa”16 (1976), e que diz res peito a uma escrita que, feita ou não por mulheres, parte de uma “ine vitável luta com o homem clássico” (p. 875). Assim, ela não pretende se referir a uma sexualidade feminina totalmente definida e uniforme, mas sim a todos aqueles que têm uma luta comum contra as postula ções falocêntricas que impediram mulheres de escrever — ou de escre ver em seus próprios termos — por tanto tempo. Na introdução de outro livro de Cixous onde ela analisa a obra de Clarice (Reading with Clarice, (1990)), Verena Andermatt Conley ex plica que a écriture féminine “sugere uma escrita baseada no encontro com o outro — seja um corpo, um escrito, um dilema social, um, mo mento de paixão — que leva à supressão das hierarquias e oposições que determinam os limites da vida mais consciente” (p. vii). Ou seja: é uma escrita capaz de reescrever as práticas sociais. Ainda nessa in trodução, Conley explica que Cixous identifica na obra de Clarice uma clara prática de écriture féminine, como explicita também Roncador em sua tese anteriormente citada: 16. As citações da Cixous no presente trabalho são traduzidas por mim, inclusive aquelas pre sentes no livro que aparecerá mais adiante, Reading with Clarice (1990). DISCURSIVIDADES s vol. 11, n. 2, e-1122207, jul-dez. 2022 16 DISCURSIVIDADES e-ISSN 2594-6269 De acordo com Cixous, o corpo que dita a escrita de Água viva é “feminino” (ela não diz “de uma mulher”), e, como tal, tem um modo particular de lidar com o próprio prazer: permitindo-se ser afetado pelo outro a ponto de quase perder a própria integridade. Escrita feminina como combate ao logocentrismo Ela escreve que em Água viva o encontro da narradora com o ou tro (ela menciona os animais, as flores, seu ex-amante), ou seja, a simples presença do outro determina a regra geral da narração [...] (RONCADOR, 2002, p. 70). Já Conley também fala que “para Cixous, em Água viva Lispector insiste no apagamento do sujeito, o que resulta na abertura de uma perspectiva sem limites e na dissolução da moldura da experiência hu mana comum ou representável” (1990, p. viii). Todas essas definições apontam para a busca de novas formas de ser, fugindo inclusive da lógica e do antropocentrismo, como resistência também à cultura falo cêntrica e como forma de fazer uma escritura feminina. Como Cixous diz em “The laugh of the Medusa” (1976): “Quase toda a história da escrita se confunde com a história da razão de que é o efeito, suporte e um dos álibis privilegiados. Ela foi homogênea à tradição falocêntrica. Ela é, em si, o falocentrismo que se vê, que goza de si mesmo e se re gozija” (p. 879); e complementa mais tarde que se deve “escrever para forjar a arma anti-logos” (p. 880). Haveria então, como constatado pela própria Cixous, esse mecanis mo “anti-logos” na obra de Clarice, e também uma recusa de se confor mar a lógicas dominantes inclusive no que diz respeito à estruturação de suas narrativas, o que é elevado ao extremo no Objeto. Cixous refle te também sobre isso: Se ela hesitou em publicar Água Viva por três anos, é porque o texto suscita a questão da lei. Dada sua economia libidinal, Clari DISCURSIVIDADES s vol. 11, n. 2, e-1122207, jul-dez. 2022 17 DISCURSIVIDADES e-ISSN 2594-6269 ce não gosta de narrativa e tudo o que ela implica na literatura. O gênero com suas leis não corresponde ao seu prazer. Ao mesmo tempo, a literatura produziu narrativas, estruturou coisas. Então talvez, tardiamente, ela se preocupou com a ideia de uma não -narrativa levada ao extremo, o que quer dizer um texto sem a lei narrativa, como Água viva (1990, p. 14). ce não gosta de narrativa e tudo o que ela implica na literatura. O gênero com suas leis não corresponde ao seu prazer. Ao mesmo tempo, a literatura produziu narrativas, estruturou coisas. Escrita feminina como combate ao logocentrismo Então talvez, tardiamente, ela se preocupou com a ideia de uma não -narrativa levada ao extremo, o que quer dizer um texto sem a lei narrativa, como Água viva (1990, p. 14). Apesar dessa preocupação que a levou a reestruturar o livro antes de sua publicação, Clarice toma novos caminhos, seguindo ainda mais o seu prazer do que a “lei” narrativa da qual fala Cixous. Isso fica claro também num trecho de Água viva em que a narradora-pintora reflete sobre seu abandono da lógica e sobre sua escrita que se afasta de qual quer concepção de gênero literário: Que mal porém tem eu me afastar da lógica? Estou lidando com a matéria-prima. Estou atrás do que fica atrás do pensamento. Inútil querer me classificar: eu simplesmente escapulo não dei xando, gênero não me pega mais. Estou num estado muito novo e verdadeiro, curioso de si mesmo, tão atraente e pessoal a ponto de não poder pintá-lo ou escrevê-lo (LISPECTOR, 2019, p. 25). Ela se apraz de sua escrita inclassificável, de sua fuga graciosa de qualquer coisa que não seja esse estado “novo e verdadeiro”, do it vivo e não facilmente nomeável. Ela segue — assim como Derrida “segue” os animais, como implica o duplo sentido da palavra “suis” em francês, que aparece no título de L’animal que donc je suis (2006) e pode signi ficar tanto “o animal que logo sou” quanto “o animal que logo sigo” —, persegue, sem, porém, efetuar qualquer tipo de captura, esse estado do atrás do pensamento. Novamente, esse prazer e essa fuga das lógicas não se aplicam somente à personagem de Água viva, mas também à DISCURSIVIDADES s vol. 11, n. 2, e-1122207, jul-dez. 2022 18 DISCURSIVIDADES e-ISSN 2594-6269 Clarice-escritora, que fala sobre sua fuga dos gêneros na crônica “Má quina escrevendo”: DISCURSIVIDADES e-ISSN 2594-6269 Clarice-escritora, que fala sobre sua fuga dos gêneros na crônica “Má quina escrevendo”: Clarice-escritora, que fala sobre sua fuga dos gêneros na crônica “Má quina escrevendo”: Sinto que já cheguei à quase liberdade. A ponto de não precisar mais escrever. Se eu pudesse, deixava meu lugar nesta página em branco: cheio de maior silêncio. E cada um que olhasse o espaço em branco, o encheria com seus próprios desejos. Vamos falar a verdade: isto aqui não é crônica coisa nenhuma. Isto é apenas. Não entra em gênero. Gêneros não me interessam mais. Interessa-me o mistério. Escrita feminina como combate ao logocentrismo Preciso ter um ritual para o mistério? Acho que sim. Para me prender à matemática das coisas. No entanto, já estou de algum modo presa à terra: sou uma filha da natureza: que ro pegar, sentir, tocar, ser. E tudo isso já faz parte de um todo, de um mistério. Sou uma só. Antes havia uma diferença entre mim e o escrever (ou não havia? não sei). Agora mais não. Sou um ser. E deixo que você seja. Isso o assusta? Creio que sim. Mas vale a pena. Mesmo que doa. Dói só no começo (2018, pp. 223-224). Clarice admite abertamente seu desinteresse pelos gêneros, se li gando ao mistério e à natureza. Sua escrita se torna um abandono e uma recusa aos saberes, se ligando mais ao silêncio das coisas que não podem ser plenamente explicadas, mas que podem ser sentidas, que podem simplesmente ser. Essa seria a liberdade tão ansiada por Clari ce: habitar o sobrenatural, o it, se fundir a todas as coisas, ser sem ter que pensar para ser. E é também por isso que ela escreve, na quebra de silêncio contraditório que nasce do próprio amor pelo silêncio, como ela diz na crônica “Amor”: “Mas estilhaçar o silêncio em palavras é um dos meus modos desajeitados de amar o silêncio. E é quebrando o silêncio que muitas vezes tenho matado o que compreendo. Se bem que — glória a Deus — sei mais silêncio que palavras” (2018, p. 243). DISCURSIVIDADES s vol. 11, n. 2, e-1122207, jul-dez. 2022 19 DISCURSIVIDADES e-ISSN 2594-6269 Escrever para ela é também uma entrada na incerteza, deixando que as palavras se dispam das compreensões e adentrem no mistério com ela. Escrever para ela é também uma entrada na incerteza, deixando que as palavras se dispam das compreensões e adentrem no mistério com ela. No atrás do pensamento, as palavras não são as que entram em con formidade com a realidade já marcada por convenções. Em Água viva, o fluir que ocorre vai, então, na direção reversa do esperado, na dire ção reversa de onde desembocam os saberes, indo em direção a novas realidades, ao sobrenatural: Eis que de repente vejo que há muito não estou entendendo. O gume de minha faca está ficando cego? 17. No ensaio “Refusing to name the animals”, David Weiss diz que “a nomeação das criaturas feita por Adão é conectada ao seu direito de nascença de dominação sobre eles […]. Esse é o perigo: nomear é enjaular; preservar é matar” (1990, tradução minha). 18. Há uma curta narrativa de Ursula K. Le Guin chamada “She unnames them”, publicada no “The New Yorker” em 1985, em que a personagem que aparenta ser Eva “desnomeia” (neolo gismo que segue a mesma linha de inovação que a palavra “unname” do original) os animais antes de se libertar de seu próprio nome e sair em direção a uma nova vida, terminando em termos que aqui parecem muito clariceanos, numa busca por novas palavras: “In fact, I had only just then realized how hard it would have been to explain myself. I could not chatter away as I used to do, taking it all for granted. My words must be as slow, as new, as single, as tenta tive as the steps I took going down the path away from the house, between the dark--branched, tall dancers motionless against the winter shining” (n.p.). Escrita feminina como combate ao logocentrismo Parece-me que o mais provável é que não entendo porque o que vejo agora é difícil: estou entrando sorrateiramente em contato com uma realidade nova para mim que ainda não tem pensamentos correspondentes e muito menos ainda alguma palavra que a signifique: é uma sen sação atrás do pensamento (LISPECTOR, 2019, p. 57). Um pouco mais adiante é dito: “eu me ultrapasso abdicando de meu nome, e então sou o mundo” (p. 58). Abdicar do nome é, no contexto derridiano, abdicar da nomeação imposta, como ocorre com os ani mais pela ação de Adão. O “próprio do homem” seria a linguagem, o direito de nomear e de responder, enquanto o animal só poderia se assujeitar sem direito de resposta. Mas esse direito de nomear não dei xa de ser uma morte, tendo em vista a necessidade humana de que rer sempre nomear o mundo, sabê-lo e compreendê-lo por completo. Nessa necessidade de saber absoluto, o homem comprova sua própria mortalidade, além de muitas vezes causar a morte daqueles a quem nomeia e, subsequentemente, aprisiona17, pois “ser chamado, se ouvir DISCURSIVIDADES s vol. 11, n. 2, e-1122207, jul-dez. 2022 20 DISCURSIVIDADES e-ISSN 2594-6269 nomear, receber um nome pela primeira vez talvez seja se saber mor tal e até se sentir morrer” (p. 39), como é dito ainda em L’animal que donc je suis (2006). Assim como em Água viva há a rejeição da nome ação, há também a rejeição da própria condição humana, num mo mento de pura revolta em que ocorre a vontade de não mais ser gente, de não saber mais a respeito da vida e da morte, invejando os animais “inconscientes de sua condição” (2019, p. 99). Essa libertação e esse despojamento do próprio nome imposto18 aqui é muito significativo, já que, nos próprios termos de Cixous (1976), a tradição logocêntrica e antropocêntrica é também falocêntrica, numa lógica que não deixou de aprisionar mulheres assim como aprisionou todos aqueles que não se incluíam na racionalização dominante. Considerações finais Ainda em relação ao extrato da página 57 de Água viva anterior mente citado, Clarice repete o mesmo trecho, com quase exatamente as mesmas palavras, num ponto mais adiante: Eis que de repente vejo que não sei nada. O gume de minha faca está ficando cego? Parece-me que o mais provável é que não entendo porque o que vejo agora é difícil: estou entrando sor rateiramente em contato com uma realidade nova para mim e DISCURSIVIDADES s vol. 11, n. 2, e-1122207, jul-dez. 2022 21 DISCURSIVIDADES e-ISSN 2594-6269 que ainda não tem pensamentos correspondentes, e muito me nos ainda alguma palavra que a signifique. É mais uma sensação atrás do pensamento (2019, p. 76). DISCURSIVIDADES e-ISSN 2594-6269 DISCURSIVIDADES e-ISSN 2594-6269 que ainda não tem pensamentos correspondentes, e muito me nos ainda alguma palavra que a signifique. É mais uma sensação atrás do pensamento (2019, p. 76). O que surge após o segundo proferimento desse trecho não é, como na primeira vez em que ele aparece, a abdicação do nome; é, ao invés disso, um elogio à vida como liberdade e leveza, às riquezas da terra, ao viver simplesmente de prazeres que podem ser vistos como inúteis, “como saber arrumar flores num jarro” (p. 77). O que parece ocorrer é um movimento decorrente da adquirida liberdade de se auto-determi nar, ou de viver sem determinar-se, e de não determinar nenhum dos seres existentes. Da liberdade que é saber silêncios, habitar nas so brenaturezas, no âmago do it vivo, e de se colocar no mesmo nível dos animais, plantas e coisas tão diversos que habitam o mundo, e assim habitar o mundo com eles. A perda da “formação humana” parece se dar assim como um ato de amor, que é um dos temas desse livro que tem tantos nomes e, ao mesmo tempo, nenhum nome. “Atrás do pensamento — mais atrás ainda — está o teto que eu olhava enquanto infante. De repente chora va. Já era amor” (2019, p. 52). Clarice-escritora, assim como Clarice -personagem, diz tomar conta do mundo, abarcar suas dores; diz ter amor em abundância apesar de não saber usá-lo, e por isso clama ao Deus (p. 65). A pergunta se torna: quem de nós sabe amar? Como Ci xous aponta, a sociedade “clássica” e racionalizante “fabricou a infame lógica do anti-amor” (1976, p. 878). Contra esta lógica, Clarice busca o amor. Considerações finais E contra o anti-amor, o silêncio parece ser sua maior arma. DISCURSIVIDADES s vol. 11, n. 2, e-1122207, jul-dez. 2022 22 DISCURSIVIDADES e-ISSN 2594-6269 Referências ADAMS, Carol J. Políticas sexuais da carne: uma teoria feminista-vegeta riana. Tradução de Cristina Cupertino. Alaúde, 2018. Arquivo Kobo. CIXOUS, Helène. “The laugh of the Medusa”. Tradução de Keith Cohen e Paula Cohen. In: Signs, Vol. 1, No. 4, pp. 875-893. Chicago: The University of Chicago press, 1976. Disponível em: <http://www.jstor.org/ stable/3173239?origin=JSTOR-pdf>. Acesso em: 15 de fevereiro de 2022. CIXOUS, Helène. Reading with Clarice. Tradução de Verena Andermatt Conley. Minneapolis: University of Minnesota Press, 1990. DERRIDA, Jacques. O animal que logo sou. Tradução de Fábio Landa. São Paulo: Editora UNESP, 2002. DERRIDA, Jacques. L’animal que donc je suis. Paris: Galilée, 2006. DESCARTES, René. Meditações. (Os Pensadores). São Paulo: Abril Cultural, 1983. LE GUIN, Ursula. “She unnames them”. In: The New Yorker, January 21, 1985, p. 27. Disponível em: <https://www.newyorker.com/magazine/1985/01/21/ she-unnames-them>. Acesso em: 15 de fevereiro de 2022. LISPECTOR, Clarice. Água viva: edição com manuscritos e ensaios inéditos. Rocco digital. 2019. LISPECTOR, Clarice. A hora da estrela. Rocco, 1977. Disponível em: <https://lelivros.love/book/download-a-hora-da-estrela-clarice-lispector- epub-mobi-e-pdf/>. Acesso em: 15 de fevereiro de 2022. LISPECTOR, Clarice. A paixão segundo G.H. Rocco, 1964. Disponível em: <https://kbook.com.br/wp-content/files_mf/paix%C3%A3osegundogh. pdf>. Acesso em: 15 de fevereiro de 2022. DISCURSIVIDADES s vol. 11, n. 2, e-1122207, jul-dez. 2022 23 DISCURSIVIDADES e-ISSN 2594-6269 LISPECTOR, Clarice. Laços de família. Rocco, 1960. Disponível em: <https:// cs.ufgd.edu.br/download/Lacos%20de%20Familia%20-%20Clarice%20 Lispector.pdf>. Acesso em: 15 de fevereiro de 2022. LISPECTOR, Clarice. Laços de família. Rocco, 1960. Disponível em: <https:// cs.ufgd.edu.br/download/Lacos%20de%20Familia%20-%20Clarice%20 Lispector.pdf>. Acesso em: 15 de fevereiro de 2022. LISPECTOR, Clarice. Arquivo de Clarice Lispector. In: Arquivo-Museu de Literatura da Fundação Casa Rui Barbosa, Rio de Janeiro. LISPECTOR, Clarice. Todas as crônicas. Rocco digital, 2018. RONCADOR, Sônia. Poéticas do empobrecimento: a escrita derradeira de Clarice. São Paulo: Annablume, 2002. SAMMER, Renata. Sobrenatureza, Água viva: o conceito à luz da ficção de Clarice Lispector. In: Remate de Males, V. 40, n.1. Campinas: 2020. DOI: 10.20396/remate.v40i1.8656951 VIVEIROS DE CASTRO, Eduardo. “Os pronomes cosmológicos e o perspectivismo ameríndio”. In: Revista Mana, vol. 2, n. 2. 1996. Disponível em: <https:// www.scielo.br/j/mana/a/F5BtW5NF3KVT4NRnfM93pSs/?lang=pt>. Acesso em: 15 de fevereiro de 2022. VIVEIROS DE CASTRO, Eduardo. “Os pronomes cosmológicos e o perspectivismo ameríndio”. In: Revista Mana, vol. 2, n. 2. 1996. Disponível em: <https:// www.scielo.br/j/mana/a/F5BtW5NF3KVT4NRnfM93pSs/?lang=pt>. Acesso em: 15 de fevereiro de 2022. WEISS, David. Refusing to name the animals. Considerações finais Disponível em: <https:// www.gettysburgreview.com/selections/detail.dot?inode=c76f42cd-c64f- 4614-9ca9-2fb7807c7f46>. Acesso em: 15 de fevereiro de 2022. WEISS, David. Refusing to name the animals. Disponível em: <https:// www.gettysburgreview.com/selections/detail.dot?inode=c76f42cd-c64f- 4614-9ca9-2fb7807c7f46>. Acesso em: 15 de fevereiro de 2022. Recebido em: 22/02/2022 Aprovado em: 26/07/2022 Licenciado por DISCURSIVIDADES s vol. 11, n. 2, e-1122207, jul-dez. 2022 24
https://openalex.org/W3195938518	https://jurnalius.ac.id/ojs/index.php/jurnalIUS/article/download/771/pdf_206	Indonesian	null	Perbandingan Hukum Terhadap Tindak Pidana Peredaran Pangan Berbahaya Antara Indonesia Dengan Tiongkok	Jurnal ius	2,021	cc-by-sa	9,693	Abstract The public must be protected from dangerous food. This research aims to find out and analyze the legal protection, prevention of crime against dangerous food at this time. This research is useful theoretically and practically. The type of research is normative juridical, the analysis of legal materials is described in analytical and prescriptive descriptive. The legal form of protection against the crime of distributing dangerous food in Indonesia is regulated in the Consumer Protection Act in general and the Food Law in particular. In China, it is regulated in the Food Safety legislation of the People’s Republic of China. China’s State Food and Drug Administration SFDA is responsible for overseeing and coordinating health, food, and drug agencies. Each country needs to pay attention to the legal protection system that is applied and give severe punishments to the perpetrators so that they have a deterrent effect. Efforts to Combat the Crime of Dangerous Food Circulation in Indonesia are carried out with Pre-Entif, Preventive and Repressive Efforts. China imposes prison sentences and fines, and there is oversight by the SFDA. Efforts are needed to provide an understanding of how food is harmful to the community to reduce the risk of hazardous food. Keywords: Legal Protection; Criminal Acts; Hazardous Food Keywords: Legal Protection; Criminal Acts; Hazardous Food Jurnal IUS Kajian Hukum dan Keadilan Volume 9, Issue 2, August 2021, E-ISSN 2477-815X, P-ISSN 2303-3827 Nationally Accredited Journal, Decree No. 30/E/KPT/2018 open access at : http://jurnalius.ac.id/ojs/index.php/jurnalIUS Jurnal IUS Kajian Hukum dan Keadilan Volume 9, Issue 2, August 2021, E-ISSN 2477-815X, P-ISSN 2303-3827 Nationally Accredited Journal, Decree No. 30/E/KPT/2018 open access at : http://jurnalius.ac.id/ojs/index.php/jurnalIUS Jurnal IUS Kajian Hukum dan Keadilan \| Vol. 9 \| Issue 2 \| August 2021 \| hlm, 438~45 tindak pidana peredaran pangan berbahaya di Indonesia diatur dalam Undang-Undang Perlindungan Konsumen secara umum dan Undang-Undang Pangan secara khusus. Di Tiongkok diatur dalam peraturan perundang-undangan Keamanan Pangan Republik Rakyat Tiongkok. Administrasi Makanan dan Obat-Obatan Negara China SFDA bertanggung jawab mengawasi dan mengoordinasikan lembaga kesehatan, makanan, dan obat-obatan. Tiap negara perlu memperhatikan sistem perlindungan hukum yang diterapkan dan memberi hukuman yang berat kepada pelaku sehingga memberikan efek jera. Upaya Penanggulangan Tindak Pidana Peredaran Pangan Berbahaya di Indonesia dilakukan dengan Upaya Pre- Entif, Preventif dan Represif. Ditiongkok menerapkan hukuman pidana penjara dan denda, dan adanya pengawasan oleh SFDA. Diperlukan upaya memberi pemahaman bagaimana pangan yang berbahaya kepada masyarakat agar dapat mengurangi resiko pangan berbahaya. Kata Kunci: Perlindungan Hukum; Tindak Pindana; Pangan Berbahaya Kata Kunci: Perlindungan Hukum; Tindak Pindana; Pangan Berbahaya y ( ) g g g g 2 Parliamentary Center, Kronologi Kasus Susu Melamin http://kebebasaninformasi.org/2011/02/17/kronologi-kasus- %E2%80%9Csusu-formula%E2%80%9D/, tanggal akses tanggal 20 Desember 2019. 1 Pasal 1 ayat (2) Undang-Undang No. 8 Tahun 1999 Tentang Perlindungan Konsumen. Abstrak Mayarakat harus dilindungi dari pangan berbahaya Tujuan Penelitian untuk mengetahui dan menganalisis perlindungan hukum, penanggulangan tindak pidana pangan berbahaya saat ini. Penelitian ini bermanfaat secara teoritis dan Praktis. Jenis penelitian Yuridis Normatif, Analisis bahan hukum secara kualitatif diuraikan secara deskriptif analitis dan preskriptif. Bentuk perlindungan hukum terhadap DOI: http://dx.doi.org/10.29303/ius.v9i2.771 1 Pasal 1 ayat (2) Undang-Undang No. 8 Tahun 1999 Tentang Perlindungan Konsumen. 2 Parliamentary Center Kronologi Kasus Susu Melamin http://kebebasaninformasi org/2011/02/17/kronologi-kasus- 1 Pasal 1 ayat (2) Undang-Undang No. 8 Tahun 1999 Tentang Perlindungan Konsumen. PENDAHULUAN Masalah perlindungan konsumen terutama dalam hal pangan perlu diperhatikan lebih lanjut dikarenakan perkembangan industri pangan yang pesat sehingga tuntutan masyarakat terhadap jenis dan kualitas produk semakin tinggi yang harus dapat dipenuhi. Bila konsumen tidak teliti dalam memilih produk pangan yang akan dibeli, konsumen hanya akan menjadi objek eksploitasi dari produsen yang tidak bertanggung jawab. Untuk dapat dipenuhinya tuntutan masyarakat, maka diberlakukan Undang-Undang Perlindungan Konsumen agar konsumen mendapatkan jaminan atas setiap produk makanan yang akan dikonsumsi. Menurut pasal 1 angka 1 Undang-Undang Nomor 8 Tahun 1999 tentang Perlindungan konsumen (UUPK) dituliskan, “Perlindungan konsumen adalah segala upaya yang menjamin adanya kepastian hukum untuk memberi perlindungan kepada konsumen”. Pengertian konsumen sendiri tertera pada pasal 1 angka 2 UUPK, “Konsumen adalah setiap orang pemakai barang dan/atau jasa yang tersedia dalam masyarakat, baik bagi kepentingan diri sendiri, orang lain, maupun makhluk hidup lain dan tidak untuk diperdagangkan.1 Institut Pertanian Bogor pernah melakukan sebuah penelitian di Tahun 2003. Pada saat itu terdapat sampling susu yang beredar dan sekarang produknya sudah tidak ada lagi menjadi objek penelitian IPB. Juru bicara IPB mengatakan IPB tidak bermaksud untuk membuat masyarakat panik dikarenakan penelitian susu formula dan makanan bayi. Tetapi penelitian ini bertujuan supaya membuktikan bahwa masih ada kasus yang terjadi setelah diberlakukannya UUPK dan dapat disimpulkan bahwa Indonesia terdapat susu formula dan makanan bayi yang terkontaminasi bakteri. Dalam penelitian falkutas kedokteran hewan IPB menyatakan bahwa sebanyak 22,73% susu formula (dari 22 sampel) dan 40% makanan bayi (dari 15 sampel) yang dijual antara bulan April-Juni tahun 2006 sudah terkontaminasi bakteri Enterobacter Sakazakii, dan semua produknya merupakan produk lokal.2. Pada tahun 2011, BPOM juga melakukan sampling dan pengujian laboratorium terhadap pangan jajanan anak sekolah (PJAS) yang diambil dari 866 sekolah dasar/madrasah ibtidaiyah yang tersebar di 30 kota di Indonesia. Sampel pangan jajanan yang diambil sebanyak 4.808 sampel, dan 1.705 (35,46%) sampel di antaranya tidak memenuhi persyaratan (TMS) keamanan dan atau mutu pangan. Dari hasil pengujian terhadap parameter uji bahan tambahan pangan 438 Jurnal IUS Kajian Hukum dan Keadilan P-ISSN: 2303-3827, E-ISSN: 2477-815X yang dilarang, yaitu boraks dan formalin yang dilakukan terhadap 3.206 sampel produk PJAS yang terdiri dari mie basah, bakso, kudapan dan makanan ringan, diketahui bahwa 94 (2,93%) sampel mengandung boraks dan 43 (1,34%) sampel mengandung formalin. 3 Yhona Paratmanitya , Veriani Aprilia (2016), Kandungan bahan tambahan pangan berbahaya pada makanan jajanan anak sekolah di Kabupaten Bantul, Vol. 4 No. 1 edisi Januari 2016 https://ejournal.almaata.ac.id/index.php/ IJND/ article/view/329, diakses tanggal 15 Agustus 2021. 4 Detik Finance, Skandal Makanan di Tiongkok, dari Daging Busuk Sampai Susu Melamin, https://finance.detik. com/berita-ekonomi-bisnis/d-2646629/skandal-makanan-di-tiongkok-dari-dari-daging-busuk-sampai-susu-melamin, di akses pada tanggal 20 Desember 2019. 5 BBC News (2017), Cina selidiki pusat produksi saus dan bumbu penyedap makanan palsu, https://www.bbc.com/ indonesia/majalah-38650024, di akses pada tanggal 15 Agustus 2021. 6 H. Ishaq, 2017, Metode Penelitian Hukum, Alfabeta, Bandung, hlm.251. j p 6 H. Ishaq, 2017, Metode Penelitian Hukum, Alfabeta, Bandung, hlm.251. 3 Yhona Paratmanitya , Veriani Aprilia (2016), Kandungan bahan tambahan pangan berbahaya pada makanan jajanan anak sekolah di Kabupaten Bantul, Vol. 4 No. 1 edisi Januari 2016 https://ejournal.almaata.ac.id/index.php/ IJND/ article/view/329, diakses tanggal 15 Agustus 2021. 7 Setiono, Rule Of Law (supremasi hukum), Magister Ilmu Hukum Program Pascasarjana Universitas Sebelas Maret, Surakarta, 2004, hlm.3. 8 Asri, (2016), “Perlindungan Hukum Bagi Konsumen Terhadap Pangan”, Jurnal IUS, Vol 4, No. 2. 9 Sovia Hasanah, (2016)“Pidana Bagi Penjual Makanan Berbahaya”, Hukum Online.Com. 10 Pasal 1(1) Peraturan Pemerintah Nomor 28 Tahun 2004 Tentang Keamanan, Mutu dan Gizi Pangan.. 11 Pasal 1(5) Undang-Undang Nomor 18 Tahun 2012 Tentang Pangan. , , , 8 Asri, (2016), “Perlindungan Hukum Bagi Konsumen Terhadap Pangan”, Jurnal IUS, Vol 4, No. 2. 9 Sovia Hasanah, (2016)“Pidana Bagi Penjual Makanan Berbahaya”, Hukum Online.Com. 10 Pasal 1(1) Peraturan Pemerintah Nomor 28 Tahun 2004 Tentang Keamanan, Mutu dan Gizi Pangan.. 11 Pasal 1(5) Undang-Undang Nomor 18 Tahun 2012 Tentang Pangan. PEMBAHASAN Bentuk Perlindungan Hukum Terhadap Tindak Pidana Peredaran Pangan Ber bahaya di Indonesia 7 Setiono, Rule Of Law (supremasi hukum), Magister Ilmu Hukum Program Pascasarjana Universitas Sebelas Maret, Surakarta, 2004, hlm.3. 8 Asri (2016) “Perlindungan Hukum Bagi Konsumen Terhadap Pangan” Jurnal IUS Vol 4 No 2 10 Pasal 1(1) Peraturan Pemerintah Nomor 28 Tahun 2004 Tentang Keamanan, Mutu dan Gizi Pangan.. 11 Pasal 1(5) Undang-Undang Nomor 18 Tahun 2012 Tentang Pangan. , , , Asri, (2016), “Perlindungan Hukum Bagi Konsumen Terhadap Pangan”, Jurnal IUS, Vol 4, No. 2. Sovia Hasanah (2016)“Pidana Bagi Penjual Makanan Berbahaya” Hukum Online Com PENDAHULUAN Hasil pengujian terhadap parameter uji pewarna bukan untuk pangan (rhodamin B) yang dilakukan terhadap 3.925 sampel produk PJAS yang terdiri dari es (mambo, loli), minuman berwarna merah, sirup, jeli/agar-agar, kudapan dan makanan ringan diketahui bahwa 40 (1,02%) sampel mengandung rhodamin B.3 Bahkan bukan hanya di negara Indonesia bisa terjadi kasus pangan berbahaya, bahkan di negara lebih maju seperti Tiongkok sekalipun juga terdapat kecurangan dalam produksi bahan pangan. Negara Tiongkok juga terdapat kasus makanan yang pernah menyita perhatian dunia. Shanghai Husi Food Co yang dimiliki oleh lllinois OSI Group diduga telah memproduksi daging busuk yang dipasarkan ke beberapa perusahaan waralaba seperti Mc Donald, Starbucks, KFC dan Pizza Hut.4 Pada tahun 2017, Badan pengawasan obat dan makanan di China mengirim tim untuk menyelediki 50 pabrik di daerah Tianjin. Pabrik-pabrik tersebut memproduksi bumbu, penyedap, dan saus palsu dimana ditemukan adanya penggunaan garam industri dalam proses produksi.5 Diperlukan upaya untuk penanggulangan pangan berbahaya. Bagaimana bentuk perlindungan hukum bagi konsumen atas peredaran pangan berbahaya di Indonesia, memberikan pandangan yang berbeda dilihat dari bagaimana bentuk perlindungan hukum bagi konsumen atas peredaran pangan berbahaya di Tiongkok. Dan Bagaimana upaya penanggulangan tindak pidana peredaran pangan berbahaya di Indonesia dan di Tiongkok, sehingga masyarakat atau konsumen mendapatkan haknya untuk terbebas dari pangan berbahaya sebagaimana yang telah diatur dalam kebijakan Pemerintah. Tujuan Penelitian ini menganalisis bagaimana bentuk perlindungan hukum terhadap tindak pidana peredaran pangan berbahaya di Indonesia. Bagaimana bentuk perlindungan hukum terhadap tindak pidana peredaran pangan berbahaya di Tiongkok dan menganalisis upaya penanggulangan tindak pidana peredaran pangan berbahaya di Indonesia dan di Tiongkok. Penelitian ini bermanfaat secara teoritis dan Praktis berkaitan dengan perlindungan konsumen dalam bidang pangan. Penelitian ini menggunakan Penelitian Yuridis Normatif dan komparatif dengan pendekatan terhadap asas dan aturan hukum Perlindungan Konsumen, Pangan di Indonesia dan di Tiongkok. Penelitian komparatif terarah pada perbedaan variabel suatu aspek penelitian. Sumber hukum yaitu bahan hukum primer Undang-Undang Nomor 18 Tahun 2012 Tentang Pangan dan Undang-Undang Perlindungan Konsumen No. 8 Tahun 1999. Bahan hukum sekunder rancangan hasil penelitian, hasil karya dari kalangan hukum. Menggunakan Bahan hukum terrier yaitu kamus, ensiklopedia, dan indeks komulatif.6 Metode pengumpulan bahan hukum studi kepustakaan. Analisis bahan hukum dalam penelitian ini menggunakan analisis secara kualitatif, menguraikan secara deskriptif analitis dan preskriptif. Jurnal IUS Kajian Hukum dan Keadilan 439 urnal IUS Kajian Hukum dan Keadilan \| Vol. 9 \| Issue 2 \| August 2021 \| hlm, 440~457 , ( ), g g p g , J , , Sovia Hasanah, (2016)“Pidana Bagi Penjual Makanan Berbahaya”, Hukum Online.Com. l ( ) i h h d i i Bentuk Perlindungan Hukum Terhadap Tindak Pidana Peredaran Pangan Ber bahaya di Indonesia Perlindungan hukum harus disertai di dalam kehidupan bermasyarakat karena manusia merupakan makhluk sosial yang di dalam kehidupannya memiliki hak-hak dasar yang dikenal dengan hak asasi manusia yang harus dihormati dan dihargai oleh pihak lain. Pada dasarnya hak-hak pribadi subjek hukum dalam hal ini adalah konsumen dalam hukum keamanan pangan adalah hak untuk hidup, hak untuk mengonsumsi makanan yang layak, hak untuk pangan yang aman, dan lain-lain. Perlindungan hukum adalah tindakan atau upaya untuk melindungi masyarakat dari perbuatan sewenang-wenang oleh penguasa yang tidak sesuai dengan aturan hukum, untuk mewujudkan ketertiban dan ketentraman sehingga memungkinkan manusia untuk menikmati martabatnya sebagai manusia.7 Teori perlindungan hukum juga digunakan pada penelitian ini karena bertujuan untuk mengintegrasikan dan mengkoordinasikan berbagai keperluan dalam masyarakat karena dalam suatu lalu lintas kepentingan, perlindungan terhadap kepentingan tertentu dapat dilakukan dengan cara membuat perbatasan berbagai kepentingan di lain pihak. Menurut Satjipto Rahardjo, Perlindungan hukum yaitu memberi pengayoman terhadap hak asasi manusia (HAM) yang telah dirugikan pihak lain dan perlindungan itu diberikan supaya masyarakat dapat memiliki semua hak- hak yang diberikan oleh hukum. Selanjutnya Phillipus M. Hadjon memberi penjelasan bahwa perlindungan hukum bagi rakyat sebagai tindakan pemerintah yang bersifat preventif dan respresif.8 Dalam Pasal 8 UU Nomor 8 Tahun 1999 Tentang Perlindungan Konsumen mengatur bahwa pelaku usaha dilarang memproduksi dan/atau memperdagangkan barang dan/ atau jasa yang tidak memenuhi standar yang dipersyaratkan dan ketentuan perundang- undangan yang mana makanan dan minuman sudah ada standar keamaman pangan dan mutu pangan yang ditetapkan oleh pemerintah. Jadi jika penjual menjual makanan yang tidak sesuai dengan stardar kemanan pangan, maka ia juga telah melanggar ketentuan dalam UU perlindungan konsumen.9 Pangan artinya semua yang berasal dari sumber hayati dan air, baik yang diolah maupun yang tidak diolah, yang diperuntukkan sebagai makanan atau minuman bagi konsumsi manusia, termasuk bahan tambahan pangan, bahan baku pangan dan bahan lain yang digunakan dalam proses penyiapan, pengolahan, dan/atau pembuatan makanan atauminuman.10 Dalam Peraturan Pemerintah Pasal 1 Nomor 86 Tahun 2019 menjelaskan bahwa keamanan pangan adalah kondisi dan upaya yang diperlukan untuk mencegah pangan dari kemungkinan cemaran biologis, kimia, dan benda lain yang dapat mengganggu, merugikan, dan membayarkan kesehatan manusia serta tidak bertentangan dengan agama, keyakinan, dan budaya masyarakat sehingga aman untuk dikonsumsi.11 Setiap makanan yang diproduksi dan hendak diperdagangkan harus memenuhi standar mutu makanan yang telah ditetapkan di Indonesia. 12 Ibid 13 Zenzen Zaunidhin, (2015) “Standar Kebutuhan Gizi Masyarakat” Agrotani.Com https://www.agrotani.com/ standar-kebutuhan-gizi-masyarakat-di-indonesia-2016-2/. 14 Dewi Nurchayati, (2015), “Upaya Meningkatkan Pengetahuan Makanan Sehat”, Jurnal Penelitian PAUDIA Vol 3 No.2 , http://journal.upgris.ac.id/index.php/paudia/article/view/510 , diakses pada tanggal 03 Maret 2021. 15 Elvira Syamsir, (2016) “Apa sih Tanggal Kadaluarsa itu?” , Personal Blog Ilmu Pangan. 16 Christine F. Mamuja, 2016, “Pengawasan Mutu dan Keamanan Pangan”, Unsrat Press, Manado, hlm.47. Jurnal IUS Kajian Hukum dan Keadilan \| Vol. 9 \| Issue 2 \| August 2021 \| hlm, 442~457 4) Keruskan Kimia : Kerusakan kimia merupakan jenis kerusakan bahan pangan yang dikarenkan adanya kegiatan bersifat fisiologis dan adanya hama yang menyerang seperti rodentia dan serangga. 4) Keruskan Kimia : Kerusakan kimia merupakan jenis kerusakan bahan pangan yang dikarenkan adanya kegiatan bersifat fisiologis dan adanya hama yang menyerang seperti rodentia dan serangga. Penggunaan BTP merupakan hal yang diperbolehkan dan wajar-wajar saja, namum pada sebagian produsen, bahan tambahan pangan ini dapat menjadi ajang pelanggaran dalam memasarkan produk makanan di masyarakat. Produsen pangan yang tidak beritikad baik dan hanya ingin memperbanyak keuntungan pribadi ini sering kali menggunakan bahan tambahan yang dilarang penggunaannya untuk makanan, atau menggunakan bahan tambahan pangan dalam dosis yang tidak sesuai dengan seharusnya sehingga tidak bagus untuk kesehatan tubuh manusia17 Dengan demikian, hal tersebut harus segera diatasi dengan berbagai cara di negara Indonesia seperti diterapkannya hukum pidana karena jika terus dibiarkan akan berakibat banyak masyarakat yang akan dirugikan dalam hal keuangan dan kesehatannya. Sanksi pidana untuk pengamanan peredaran makanan dan minuman diatur dalam Undang- Undang No. 18 Tahun 2012 Tentang Pangan Pasal 138. Menurut ketentuan Pasal 23 PP Nomor 28 Tahun 2004 Tentang Keamanan Mutu dan Gizi Makanan, setiap pelaku usaha tentunya dilarang mengedarkan dan memproduksi pangan yang berbahaya diantaranya sebagai berikut :18 1. Yang sudah kedaluwarsa; 2. Yang mengandung bahan yang busuk, kotor, tengik, atau mengandung bahan nabati atau hewani yang sudah berpenyakit atau berasal dari bangkai sehingga pangan tidak layak dikonsumsi oleh manusia; 3. Yang mengandung bahan yang dilarang digunakan dalam kegiatan atau proses produksi pangan; 4. Yang mengandung cemaran yang telah melampaui ambang batas maksimal yang ditetapkan; 5. Yang mengandung bahan-bahan beracun yang bisa menyebabkan kesehatan atau jiwa manusia. Fakta kondisi yang menjadi salah satu latar belakang terbentuknya Undang-Undang tentang pangan, adalah karena pangan merupakan kebutuhan dasar manusia yang paling penting dalam mempertahankan hidupnya, yang pemenuhnya juga merupakan bagian dari hak asasi manusia dalam UUD Negara RI tahun 1945, sebagai komponen dasar untuk mewujudkan sumber daya manusia yang berkualitas. Seperti kasus yang pernah terjadi di kota Jakarta pada tahun 2015, pihak BPOM menemukan 3.499 item pangan senilai Rp. 4,8 miliar yang tidak memenuhi standar yang ditentukan pemerintah. BPOM telah melakukan tindak lanjut atas kasus ini dengan membawanya ke ranah hukum dan menjatuhkan sanksi yang sesuai. 18 Pasal 23 Undang-Undang Nomor 28 Tahun 2004 Tentang Keamanan, Mutu dan Gizi Makanan. 19 Kartika Sari Tarigan, “BPOM : Jual Pangan Tak Penuhi Syarat Mutu Itu Kriminal, Ada Sanksi” , Detik.Com, https://news.detik.com/berita/d-3102164/bpom-jual-pangan-tak-penuhi-syarat-mutu-itukriminal-ada-sanksi, di akses pada tanggal 18 Januari 2020. 17 Yovita Diane Titiesari, (2018), “ Zat Kimia Berbahaya yang Sering Digunakan dalam Makanan”, GueSehat. Com. 17 Yovita Diane Titiesari, (2018), “ Zat Kimia Berbahaya yang Sering Digunakan dalam Makanan”, GueSehat. Com. 18 Pasal 23 Undang-Undang Nomor 28 Tahun 2004 Tentang Keamanan, Mutu dan Gizi Makanan. 19 Kartika Sari Tarigan, “BPOM : Jual Pangan Tak Penuhi Syarat Mutu Itu Kriminal, Ada Sanksi” , Detik.Com, https://news.detik.com/berita/d-3102164/bpom-jual-pangan-tak-penuhi-syarat-mutu-itukriminal-ada-sanksi, di akses pada tanggal 18 Januari 2020. 17 Yovita Diane Titiesari, (2018), “ Zat Kimia Berbahaya yang Sering Digunakan dalam Makanan”, GueSehat. Com. 18 Pasal 23 Undang-Undang Nomor 28 Tahun 2004 Tentang Keamanan, Mutu dan Gizi Makanan. 19 K rtik S ri T rig n “BPOM Ju l P n n T k P nuhi Sy r t Mutu Itu Krimin l Ad S nk i” D tik C m Pasal 23 Undang-Undang Nomor 28 Tahun 2004 Tentang Keamanan, Mutu dan Gizi Makanan. Yovita Diane Titiesari, (2018), “ Zat Kimia Berbahaya yang Sering Digunakan dalam Makanan”, GueSeha Bentuk Perlindungan Hukum Terhadap Tindak Pidana Peredaran Pangan Ber bahaya di Indonesia Standar adalah spesifikasi 440 Jurnal IUS Kajian Hukum dan Keadilan P-ISSN: 2303-3827, E-ISSN: 2477-815X atau persyaratan teknis yang dibakukan, termasuk tata cara dan metode yang disusun berdasarkan konsensus semua pihak yang terkait dengan memperhatikan syarat-syarat keselamatan, keamanan, kesehatan, lingkungan hidup, perkembangan ilmu pengetahuan dan teknologi, serta pengalaman perkembangan masa kini dan masa yang akan datang untuk memperoleh manfaat yang sebesar-besarnya.12 Standar keamanan pangan di Indonesia menurut ahli gizi atau Departemen Kesehatan, setiap jiwa memerlukan sedikitnya 2200 k kalkapital tahun dimana energi tersebut berasal dari berbagai sumber bahan pangan, misalnya karbohidrat, protein, vitamin dan mineral.13 Makanan yang memenuhi standar sudah pasti harus makanan yang sehat. Makanan yang sehat adalah makanan yang mempunyai zat gizi yang cukup dan seimbang, serta tidak mengandung (tercemar) unsur yang dapat membahayakan atau merusak kesehatan. Sangat penting bagi orang tua dalam mengarahkan anak-anak berkaitan dengan memilih makanan jajanan yang sehat dan halal. Mengenalkan dan menanamkan konsep sehat dan halal sejak dini, akan memberikan perkembangan psikologis yang baik terutama pada pembentukan akhlak yang mulia pada diri anak.14 Adapun jenis-jenis kriteria pangan berbahaya yang ada di Indonesia yaitu seperti makanan yang sudah kadaluwarsa tetapi tetap dipergadangkan atau produk makanan yang mengandung zat kimia berbahaya. Produk pangan yang sudah kadaluwarsa artinya produk yang sudah melewati batas tanggal yang telah ditentukan. Tanggal kadaluwarsa adalah tanggal yang dicantumkan pada kemasan makanan yang bertujuan memberi tahu kepada konsumen sebaiknya sebelum kapan makanan tersebut layak untuk dikonsumsi ataupun menunjukkan tanggal terakhir makanan tersebut berada pada posisi kualitas terbaiknya dalam hal-hal yang menyangkut seperti rasa, tekstur, penampilan, bau, dan kandungan gizi. Penurunan kualitas setelah tanggal tersebut terjadi secara gradual. Tanggal berlaku seperti ini digunakan untuk produk-produk makanan yang belum dibuka.15 Pangan bisa termasuk kategori bahaya pasti disebabkan oleh berbagai jenis kerusakan seperti berikut ini :16 Kerusakan biologis merupakan kerusakan bahan pangan yang dikarenakan oleh aktivitas mikroba. Bakteri, kapang, khamir merupakan salah satu contoh mikroba yang dapat merusak bahan-bahan pangan. Mikroba-mikroba tersebut memiliki daya rusak yang sangat tinggi karena bisa menyebabkan degradasi komponen bahan pangan sehingga bersifat toksin dan tidak bagus untuk kesehatan manusia. 2) Kerusakan Mekanis : Kerusakan mekanis merupakan jenis kerusakan bahan pangan yang disebabkan oleh benturan, terjatuh, tekanan dan lain-lain yang menyebabkan kondisi fisik pangan menjadi tidak seperti selayaknya. 3) Kerusakan Fisik : Kerusakan fisik merupakan jenis kerusakan bahan pangan yang disebabkan oleh adanya perlakukan secara fisik yang dialami oleh bahan pangan itu sendiri. 441 Jurnal IUS Kajian Hukum dan Keadilan 441 Jurnal IUS Kajian Hukum dan Keadilan \| Vol. 9 \| Issue 2 \| August 2021 \| hlm, 442~457 Pada tanggal 21 Desember 2015 pihak BPOM menutup seluruh pengawasan pangan Indonesia di sarana distribusi dan retail seperti pasar tradisional, supermarket, dan pembuat parsel dengan rincian sebagai berikut :19 442 Jurnal IUS Kajian Hukum dan Keadilan P-ISSN: 2303-3827, E-ISSN: 2477-815X Tabel 1. Jumlah Pangan Berbahaya di Indonesia No Persen (%) Keterangan 1 28% Makanan tanpa izin edar 2 63% Makanan kedaluwarsa 3 9% Kemasan pangan rusak Tabel 1. Jumlah Pangan Berbahaya di Indonesia Sumber : https://news.detik.com/berita/d-3102164/bpom-jual-pangan-tak-penuhi-syarat-mutu-itu-kriminal-ada-sanksi Jenis pangan kedaluwarsa yang paling banyak ditemukan yaitu mie instan, susu kental manis, bumbu, teh, minuman serbuk dan makanan ringan. Kota yang paling banyak pangan kadaluwarsannya yaitu : banyak pangan kadaluwarsannya yaitu : Tabel. 2 Jumlah Pangan Kadaluwarsa. No Kota Jumlah Keterangan 1 Kupang 27.052 Kemasan 2 Makassar 20.354 Kemasan 3 Jayapura 16.376 Kemasan 4 Manokwari 5.397 Kemasan 5 Sofifi 1.134 Kemasan Sumber : https://news.detik.com/berita/d-3102164/bpom-jual-pangan-tak-penuhi-syarat-mutu-itu-kriminal-ada-sanksi Tabel. 3 Jumlah Pangan Tanpa Izin Edar No Kota Jumlah Keterangan 1 Medan 11.173 Kemasan 2 Pekanbaru 8.326 Kemasan 3 Batam 3.648 Kemasan 4 Bandung 2.735 Kemasan 5 Manokwari 2.157 Kemasan Sumber : https://news.detik.com/berita/d-3102164/bpom-jual-pangan-tak-penuhi-syarat-mutu-itu-kriminal-ada-sanksi Tabel. 2 Jumlah Pangan Kadaluwarsa. No Kota Jumlah Keterangan 1 Kupang 27.052 Kemasan 2 Makassar 20.354 Kemasan 3 Jayapura 16.376 Kemasan 4 Manokwari 5.397 Kemasan 5 Sofifi 1.134 Kemasan Sumber : https://news.detik.com/berita/d-3102164/bpom-jual-pangan-tak-penuhi-syarat-mutu-itu-kriminal-ada-sanksi Tabel. 2 Jumlah Pangan Kadaluwarsa. No Kota Jumlah Keterangan 1 Kupang 27.052 Kemasan 2 Makassar 20.354 Kemasan 3 Jayapura 16.376 Kemasan 4 Manokwari 5.397 Kemasan 5 Sofifi 1.134 Kemasan Sumber : https://news.detik.com/berita/d-3102164/bpom-jual-pangan-tak-penuhi-syarat-mutu-itu-kriminal-ada-sanksi No Tabel. 3 Jumlah Pangan Tanpa Izin Edar No Kota Jumlah Keterangan 1 Medan 11.173 Kemasan 2 Pekanbaru 8.326 Kemasan 3 Batam 3.648 Kemasan 4 Bandung 2.735 Kemasan 5 Manokwari 2.157 Kemasan Sumber : https://news.detik.com/berita/d-3102164/bpom-jual-pangan-tak-penuhi-syarat-mutu-itu-kriminal-ada-sanksi No Jurnal IUS Kajian Hukum dan Keadilan \| Vol. 9 \| Issue 2 \| August 2021 \| hlm, 444~457 urnal IUS Kajian Hukum dan Keadilan \| Vol. 9 \| Issue 2 \| August 2021 \| hlm, 444~457 Tabel. 4 Jumlah Pangan Kemasan Rusak No Kota Jumlah Keterangan 1 Makassar 5.257 Kemasan 2 Jayapura 1.397 Kemasan 3 Mataram 1.274 Kemasan 4 Manokwari 650 Kemasan 5 Pekanbaru 310 Kemasan Sumber : https://news.detik.com/berita/d-3102164/bpom-jual-pangan-tak-penuhi-syarat-mutu-itu-kriminal-ada-sanksi Tabel. 4 Jumlah Pangan Kemasan Rusak Sumber : https://news.detik.com/berita/d-3102164/bpom-jual-pangan-tak-penuhi-syarat-mutu-itu-kriminal-ada-sanksi Kasus ini tidak dapat dianggap tindak pidana yang ringan dan akan dibawa ke ranah hukum dan pastinya juga harus mendapatkan sanksi administratif. Bahkan beberapa kasus pengedaran pangan tak memenuhi syarat dan ketentuan sudah ada yang dijadikan tersangka. 20 Ade Setiawan, (2015), “Strategi Tiongkok Memasarkan Produk-Produk di Indonesia”, Jurnal Ilmu Hubungan Internasional, Vol. 2 No. 2 Edisi Oktober. 21 Elly, (2014),“BPKN Indonesia Belajar Dari Tiongkok”, Indonesia Media. Jurnal IUS Kajian Hukum dan Keadilan \| Vol. 9 \| Issue 2 \| August 2021 \| hlm, 442~457 Di dalam Pasal 62 UU Nomor 8 Tahun 1999 Tentang Perlindungan Konsumen, ancaman hukuman apabila pangan itu diedarkan ilegal adalah 2 tahun penjara atau denda Rp. 4 Milliar. Sedangkan produsen yang mengedarkan bahan pangan kimia berbahaya bisa dikenai ancaman hukuman 5 tahun penjara atau denda Rp. 10 Milliar. Bentuk Perlindungan Hukum Terhadap Tindak Pidana Peredaran Pangan Ber bahaya di Tiongkok 20Tiongkok tumbuh menjadi negara yang menunjukan peningkatan ekonomi yang di atas rata-rata, mampu bertahan dari goncangan krisis ekonomi dunia pada akhir abad ke 20.Walaupun sebagai negara yang sangat maju dan terdapat banyak penduduk, hal tersebut tidak menjamin Tiongkok merupakan negara yang bebas akan pangan berbahaya. Badan perlindungan konsumen nasional di Tiongkok tidak beda jauh dengan Negara Indonesia. Bahkan BPKN Ardian Syah Parman mengatakan konsumen domestik di Tiongkok sama besarnya dengan Indonesia.21 Sesuai dengan pasal 29 Undang-Undang Kemanan Pangan di Tiongkok, merumuskan standar lokal untuk kemanan pangan dan harus mengumpulkan pendapat secara terbuka di depan umum mengenai rencana implementasi tahunan mereka. Departemen kesehatan administrasi pemerintah rakyat provinsi, daerah otonom, dan kota langsung di bawah pemerintah pusat harus melaporkan standar lokal ke departemen administrasi kesehatan di bawah dewan negara untuk catatan dalam waktu 30 hari kerja sejak tanggal penerbitan standar lokal untuk kemanan pangan. Jika standar lokal untuk keamanan pangan dicabut sesuai dengan hukum, departemen administrasi kesehatan pemerintah provinsi, daerah otonomi, dan kota madya rakyat langsung di bawah pemerintah pusat akan segera mengumumkan pencabutan di situs web mereka. Dalam standar keamanan pangan Tiongkok juga dikatakan makanan khusus seperti makanan kesehatan, makanan formula medis khusus, makanan formula bayi dan sebagainya. Bukan makanan khusus lokal dan standar keamanan pangan lokal mungkin tidak ditetapkan untuk mereka. Setelah standar keamanan pangan diumumkan, produsen dan operator pangan dapat 444 Jurnal IUS Kajian Hukum dan Keadilan P-ISSN: 2303-3827, E-ISSN: 2477-815X menerapkan dan mempublikasikan penerapan standar kemanan pangan sebelum tanggal yang ditentukan dalam standar kemanan pangan. Hal ini perusahaan produksi makanan tidak boleh merumuskan standar perusahaan lebih rendah dari yang disyaratkan oleh standar keamanan pangan nasional atau lokal. Jika perusahaan produksi makanan merumuskan standar perusahaan yang indikator kemanan pangannya lebih ketat daripada standar kemanan pangan nasional atau lokal, maka perusahaan tersebut harus melapor kepada departemen administrasi kesehatan pemerintah provinsi, wilayah otonom, atau kota madya secara langsung dibawah catatan pemerintah pusat. Juga ketika perusahaan produksi makanan merumuskan standar perusahaan, itu harus terbuka untuk inspeksi publik. Beberapa kriteria pangan berbahaya di Tiongkok mulai dari makanan yang berasal dari bahan olahan yang sudah membusuk dan juga makanan yang terkontaminasi oleh bakteri yang membahayakan bagi manusia. Hal tersebut bisa terjadi karena pelaku usaha di Tiongkok hanya mengambil keuntungan sepihak tanpa memikirkan dampak ke masyarakat yang mengonsumsinya.22 g p p p y y g g y Di Negara Tiongkok, keikutsertaan masyarakat secara langsung dalam hal perlindungan konsumen masih kurang. 22 Dewan Negara, “Peraturan Pelaksanaa Undang-Undang Keamanan Pangan Republik Rakyat Tiongkok” http:// www.gov.cn/zhengce/content/2019-10/31/content_5447142.html , diakses tanggal 20 Januari 2020. 23 Green Pack,(2017), “Apa itu FDA dan FSC?”, http://www.greenpack.co.id/id/apa-itu-fda-dan-fsc/#. XhLL8hsxXIV, di akses pada tanggal 20 Januari 2020. 24 Ibid Bentuk Perlindungan Hukum Terhadap Tindak Pidana Peredaran Pangan Ber bahaya di Tiongkok Masyarakat terlibat langsung dalam mengawasi penerapan peraturan perundang-undangan, sehingga pada saat konsumen menemukan produk yang tidak sesuai dengan standart nasionalnya, dapat secara langsung melaporkannya ke pengurus asosiasinya yaitu China Food And Drug Administration. Mengatur dan menjamin keamanan produk pangan, kosmetik dan obat-obatan yang dijual kepada konsumen serta menjanjikan produk sesuai dengan standar yang akan diberikan ke pemakai merupakan wewenang dari pihak CFDA. Selain itu, CFDA juga berkewajiban untuk menata bagaimana supaya kemasan makanan ini layak untuk kenyamanan konsumen.23 Produsen yang memproduksi pangan yang tidak memenuhi standart nasionalnya, misalkan pemilik toko bisa langsung dikenakan penalty. Penalty sebuah sanksi yang dapat berupa denda kepada pengusaha. Seperti yang diatur Undang-Undang Keamanan Pangan Tiongkok di Pasal 75 dalam bab 9 dikatakan bahwa, “Perusahaan produksi dan operasi makanan dan unit lainnya memiliki pelanggaran berdasarkan undang-undang keamanan pangan. Selain hukuman sesuai dengan ketentuan Undang-Undang Keamanan Pangan, dalam situasi berikut, hukum unit, penanggung jawab utama, penanggung jawab langsung bahkan penanggung jawab yang bertanggung jawab dan orang lain yang bertanggung jawab langsung akan didenda lebih dari 1 kali dan 10 kali pendapatan yang pernah mereka terima dari unit tahun sebelumnya akan dikurangi”.24 Dalam Pasal 74 Undang-Undang Keamanan Pangan Repulik Rakyat Tiongkok juga mencatumkan bahwa jika makanan yang diproduksi dan dioperasikan oleh produsen atau operator pangan memenuhi keamanan pangan tetapi tidak memenuhi indikator keamanan pangan yang ditetapkan oleh standar perusahaan yang ditandai untuk makanan, maka departemen pengawasan dan administrasi keamanan pangan pemerintah atas tingkat kabupaten harus memberikan peringatan dan berpesan pada pihak operator untuk segera memberhentikan makanan dan memerintahkan perusahaan produksi makanan untuk memperbaikinya. Jika hal ini ia menolak untuk berhenti mengoperasikan atau memperbaiki makanan, maka pihak pemerintah akan menyita makanan yang tidak memenuhi indikator keamanan makanan yang ditentukan Jurnal IUS Kajian Hukum dan Keadilan 445 urnal IUS Kajian Hukum dan Keadilan \| Vol. 9 \| Issue 2 \| August 2021 \| hlm, 446~457 oleh standar perusahaan, tidak hanya itu, pihak operator akan dikenakan denda sebesar 10.000 yuan (Setara dengan Rp. 19.668.900,-) atau bahkan bisa naik menjadi 5 (lima) kali lipatan. Hal tersebut sudah diatur dalam Undang-Undang Kemanan Pangan Tiongkok. Penyelesaian secara tindak pidana jika terjadi kematian pada konsumen, akan dikenakan sanksi minimal penjara atau hukuman mati sesuai dengan keputusan hakim pada pengadilan Rakyat China. 25 Viva, (2008),“Kronologi Skandal Susu Beracun China” https://m.viva.co.id/amp/berita/dunia/679-kronologi-skan dal-susu-beracun-china, di akses pada tanggal 20 Januari 2020. Bentuk Perlindungan Hukum Terhadap Tindak Pidana Peredaran Pangan Ber bahaya di Tiongkok p p g y Sebuah kasus pernah terjadi di Tiongkok pada tahun 2008 yang cukup menghebohkan dan menarik perhatian masyarakat Tiongkok maupun luar negri yaitu skandal susu melamin yang telah menewaskan 4 orang korban dan hampir 53.000 anak jatuh sakit setelah mengonsumsi susu tersebut. Dari laporan Associated Press terdapat hasil sebagai berikut :25 Tabel 5. Hasil Laporan Kasus Susu Melamin No. Waktu Laporan 1 Desember 2007 Sanlu Group Company mendapatkeluhandari para orang tua karena susu formula merk Sanlu yang mereka beli menyebabkan bayi mereka jatuh sakit. 2 Juni 2008 Sanlu mengetahui bahwa produk susu bubuk buatannya terkontaminasi zat kimia berbahaya, melamin. 3 30 Juni 2008 Badan Pengawas Kualitas Makanan dan Karantina menerima laporan mengenai lima bayi yang menderita batu ginjal dan dirawat di rumah sakit anak di Provinsi Hunan. Kelima bayi tersebut mengonsumsi susu bubuk merk Sanlu. Data ini diambil dari situs Badan Pengawas Kualitas Makanan dan Karantina yang sejak saat itu tel ah dihapus. 4 24 Juli 2008 Seorang dokter anak member tahu Badan Pengawas Kualitas Makanan dan Karantina bahwa ada Sembilan kasus penyakit batu ginjal pada anak-anak. Kesembilan anak tersebut juga mengonsumsi susu formula merk Sanlu. Data ini diambil dari situs yang juga telah diha pus. 5 2 Agustus 2008 Sanlu memberitahu pemerintah kota Shi jia zhuang, ibu kota Provinsi Hebei yang merupakan basis perusahaan, bahwa susu produksi mereka tercemar bahan kimia. Dalam rapat pengurus perusahaan, Fonterra-perusa haan kelompok peternak di Selandia Baru dan pemeg ang 43% saham di Sanlu, diberitahu mengenai jatuhnya korban dan meminta penarikan segera terhadap produk Sanlu. 6 6 Agustus 2008 Sanlu menarik produk susu bubuk bayidari distributor, tetapi tidak dari publik. 7 8 Agustus 2008 Olimpiade Beijing dibuka dan berlangsung sampai 24 Agustus 2008. 25 Viva, (2008),“Kronologi Skandal Susu Beracun China” https://m.viva.co.id/amp/berita/dunia/679-kronologi-skan dal-susu-beracun-china, di akses pada tanggal 20 Januari 2020. Tabel 5. Hasil Laporan Kasus Susu Melamin 446 Jurnal IUS Kajian Hukum dan Keadilan P-ISSN: 2303-3827, E-ISSN: 2477-815X 8 5 September 2008 Fonterra memberitahukan Perdana menteri Selandia Baru, Helen Clark mengenai kasus ini. Tiga hari kemu dian, Clark meminta pemerintah Selandia Baru untuk meneruskan informasi ini kepada pemerintah China di Beijing. 9 9 September 2008 Pemerintah kota Shi jia zhuang menginformasikan ke pada pemerintah Provinsi Hebei. Sehari kemudian, pemerintah Provinsi Hebei meneruskannya ke pemer intah pusat. 10 11 Septemter 2008 Sanlu menarik 700 ton susu formula bayi dari pasaran. Bentuk Perlindungan Hukum Terhadap Tindak Pidana Peredaran Pangan Ber bahaya di Tiongkok Pemerintah Hongkong juga melaporkan korban pertama dari luar daratan China, yaitu seorang anak perempuan berusia tiga tahun yang menderita batu ginjal setelah mengonsumsi susu merk Yili. 18 22 September 2008 Ketua Badan Pengawas Kualitas Makanan dan Karan tina China mengundurkan diri. Sumber : https://m.viva.co.id/amp/berita/dunia/679-kronologi-skandal-susu-beracun-china Sanksi pidana pelaku yaitu dijatuhi hukuman mati karena perbuatannya yang menyebabkan ‘kerugian ekstrim bagi masyarakat’. Demikian keputusan Mahkamah Agung, seperti dikutip sebuah harian China. Salah satu dari kedua pelaku tersebut terbukti bersalah memproduksi lebih dari 770 ton bubuk protein yang dicampur dengan zat Melamin antara Juli 2007 dan Agustus 2008. Sekitar 600 ton bubuk ini berhasil dijual. Keduanya dijatuhi hukuman mati karena ia menjual susu bubuk yang terkontaminasi itu, antara lain kepada perusahaan Sanlu yang dulu menjadi sorotan dalam skandal ini dan sekarang sudah bangkrut. Sampai sekarang ada 21 terdakwa dalam kasus ini yang dinyatakan bersalah. Pemimpin perusahaan Sanlu dihukum penjara seumur hidup. Sanksi pidana untuk para pelaku yang menyebabkan tewasnya konsumen tentu akan diadili dengan hasil pengadilan yang dilakukan aparat hukum dalam sidang Komite Tetap Kongres Rakyat Nasional (NPC) dengan berdasar kepada KUHP Rakyat Tiongkok. Seorang terdakwa lain dijatuhi hukuman mati dengan masa percobaan. Ini merupakan hukuman yang biasanya diubah menjadi hukuman penjara seumur hidup. 15 terdakwa lainnya dijatuhkan hukuman penjara antara satu sampai 15 tahun. Seluruhnya ada 22 produsen susu Cina dinyatakan bersalah karena menjual susu yang terkontaminasi Melamin. Mereka dituntut membayar ganti rugi senilai 115 juta Euro. Banyak orang tua dari anak-anak yang sakit mengkritik, ganti rugi ini terlalu sedikit. Bentuk Perlindungan Hukum Terhadap Tindak Pidana Peredaran Pangan Ber bahaya di Tiongkok Pemerintah pusat bersumpah akan memberikan huku man berat bagi mereka yang bertanggung jawab atas masalah ini. Sementara itu, satu orang bayi dilaporkan telah meninggal akibat gagal ginjal dan belasan anak dirawat di rumah sakit. 11 13 September 2008 Wakil Menteri Kesehatan China, Gao Qiang, menga takan 432 bayi yang mengonsumsi susu formula merk Sanlu menderita batu ginjal. Gao menyalahkan Sanlu karena menunda peringatan kepada publik. Pemerintah kemudian melakukan pemeriksaan terhadap semua pro dusen susu di China. 12 15 September 2008 Jumlah korban meningkatkan menjadi lebih dari 1.200 anak, dan dua orang meninggal. Badan Pengawas Kuali tas Makanan dan Karantina China mengatakan bahwa tindakan pencampuran susu dengan melamin terjadi dit ingkat pemasok susu mentah. Pemasok kemudian men jual susu berbahaya tersebut kepada perusahaan susu. Wakil Presiden Direktur Sanlu meminta maaf tetapi tidak menjelaskan mengapa mereka tidak segera mem beritahu public begitu mereka mengetahui informasi ini. 13 16 September 2008 Inspeksi nasional terhadap 109 perusahaan susu for mula bayi menghasilkan 22 perusahaan yang produknya tercemar melamin. Manajer Sanlu, TianWenhua dipe cat. 14 17 September 2008 Dua perusahaan susu besar di China, Mengniu Dairy Company danYili Industrial Group Company menarik produknya. Sementara, menteri kesehatan China me laporkan bahwa tiga bayi telah meninggal dan lebih 6.200 anak jatuh sakit. 15 18 September 2008 Pejabat berwenang menangkap lebih dari 12 orang ter sangka, menambah jumlah tersangka menjadi 18 orang. Polisi menyita hampir seberat 300 kilogram bahan kimia yang diduga berbahaya. mber 2008 Wakil Menteri Kesehatan China, Gao Qiang, menga takan 432 bayi yang mengonsumsi susu formula merk Sanlu menderita batu ginjal. Gao menyalahkan Sanlu karena menunda peringatan kepada publik. Pemerintah kemudian melakukan pemeriksaan terhadap semua pro dusen susu di China. 4 Jurnal IUS Kajian Hukum dan Keadilan 447 urnal IUS Kajian Hukum dan Keadilan \| Vol. 9 \| Issue 2 \| August 2021 \| hlm, 448~457 Jurnal IUS Kajian Hukum dan Keadilan \| Vol. 9 \| Issue 2 \| August 2021 \| hlm, 448~457 16 19 September 2008 Krisis meluas setelah tes yang dilakukan pemerintah menemukan bahwa kandungan melamin juga ditemu kan dalam susucair yang diproduksi oleh tiga perusa haan susu terbesar di China. 17 21 September 2008 Departmen kesehatan melaporkan jumlah korban yang jatuh melonjak menjadi hampir 53.000 anak. Sebanyak 12.892 anak dirawat di rumah sakit dan 104 anak dalam kondisi gawat. 26 Andi Hamzah.Bunga Rampai Hukum Pidana dan Acara Pidana. Ghalia Indonesia Jakarta.2001. hlm. 30. g g 28 Humas BSN, (2016), “Penerapan SNI untuk Mutu dan Keamanan Pangan”, Badan Standardisasi Nasiona 29 BKP “L Ki j P t P k K i d K P ” htt 27 S Alam. 2010. Pengantar Kriminologi. Makassar. Pustaka Refleksi. hlm. 79-80. 28 Humas BSN, (2016), “Penerapan SNI untuk Mutu dan Keamanan Pangan”, Badan Standardisasi Nasional 29 BKP, “Laporan Kinerja Pusat Penganekaragaman Konsumsi dan Keamanan Pangan”, http: //bkp.pertanian.go.id/storage/app/media/Konsumsi/LAKIN%20KONSUMSI%202018.pdf, di akses pada tanggal 25 Januari 2020. 30 Janus Sidabalok, 2014 Hukum Perlindungan Konsumen di Indonesia, PT. Citra Aditya Bakti, Bandung hlm. 19. 29 BKP, “Laporan Kinerja Pusat Penganekaragaman Konsumsi dan Keamanan Pangan”, http: //bkp.pertanian.go.id/storage/app/media/Konsumsi/LAKIN%20KONSUMSI%202018.pdf, di akses pada tanggal 25 Januari 2020. 30 Janus Sidabalok, 2014 Hukum Perlindungan Konsumen di Indonesia, PT. Citra Aditya Bakti, Bandung hlm. 19. 27 S Alam. 2010. Pengantar Kriminologi. Makassar. Pustaka Refleksi. hlm. 79-80. 3. Upaya Penanggulangan Tindak Pidana Peredaran Pangan Berbahaya di Indone sia dan di Tiongkok Hampir di setiap negara pasti pernah mengalami kasus pangan berbahaya karena pangan merupakan hal yang pertama paling penting bagi semua manusia di muka bumi ini. Tentunya setiap negara pernah mengalami kasus pangan yang berbeda, begitu juga dengan cara penanggulanganya juga berbeda-beda. Di Indonesia, bagi pelaku usaha yang tidak bertanggung jawab dalam memperdagangkan hasil produksinya dan menyebabkan kefatalan terhadap konsumen, hal itu bisa disebut suatu tindak kejahatan. Adapun unsur-unsur yang bisa dikategorikan dalam bentuk kejahatan yaitu :26 a. Kelakuan dan akibat (perbuatan) 448 Jurnal IUS Kajian Hukum dan Keadilan 448 P-ISSN: 2303-3827, E-ISSN: 2477-815X b. Hal ikhwal atau keadaan yang menyertai perbuatan c. Keadaan tambahan yang memberatkan pidana d. Unsur melawan huku yang objektif e. Unsur melawan hukum yang subyektif Menurut A.S. Alam penanggulangan kejahatan emperik yang terdiri dari tiga bagian pokok yaitu :27 1) Upaya Pre-Empif, yaitu usaha awal yang dilakukan untuk mencegah agar tidak terjadi tindak pidana dengan menerapkan norma dan nilai yang baik . 1) Upaya Pre-Empif, yaitu usaha awal yang dilakukan untuk mencegah agar tidak terjadi tindak pidana dengan menerapkan norma dan nilai yang baik . 2) Upaya Preventif, yaitu upaya tindak lanjut dari upaya pre-emtif yang masih dalam tataran pencegahan sebelum terjadinya kejahatan. Dalam upaya ini lebih menekankan kepada bagaimana agar menghilangkan kesempatan untuk melakukan kriminal. 3) Upaya Represif, yaitu upaya yang dilakukan pada saat sudah terjadi tindak kejahatan dengan cara menjatuhkan hukuman. Untuk menghadapi bermacam kendala dalam menciptakan produk pangan yang berkualitas, pemerintah menerapkan strategi baru dalam penerapan Standar Nasional Indonesia (SNI). Mulai dari aspek regulasi SNI agar lebih ramah bagi UMKM (usaha mikro, kecil, menengah) yang bermodal kecil dan memiliki teknologi pengolahan sederhana.28 Berikut ini beberapa penanggulangan yang telah diterapkan di Indonesia yaitu :29 Untuk menghadapi bermacam kendala dalam menciptakan produk pangan yang berkualitas, pemerintah menerapkan strategi baru dalam penerapan Standar Nasional Indonesia (SNI). Mulai dari aspek regulasi SNI agar lebih ramah bagi UMKM (usaha mikro, kecil, menengah) yang bermodal kecil dan memiliki teknologi pengolahan sederhana.28 Berikut ini beberapa penanggulangan yang telah diterapkan di Indonesia yaitu :29 y 1) Menerapkan sistem jaminan mutu dan keamanan pangan. Jaminan mutu di Indonesia harus memperhatikan standar kebutuhan gizi masyarakat minimal yaitu 2.200kkal/ kapita/tahun. 1) Menerapkan sistem jaminan mutu dan keamanan pangan. Jaminan mutu di Indonesia harus memperhatikan standar kebutuhan gizi masyarakat minimal yaitu 2.200kkal/ kapita/tahun. 2) M k d k i b ik 2) Menerapkan cara memproduksi pangan yang baik. , ( ), p g , BKP, “Laporan Kinerja Pusat Penganekaragaman Konsumsi dan Keamanan Pangan”, http: bkp.pertanian.go.id/storage/app/media/Konsumsi/LAKIN%20KONSUMSI%202018.pdf, di akses pada tangg nuari 2020. Janus Sidabalok, 2014 Hukum Perlindungan Konsumen di Indonesia, PT. Citra Aditya Bakti, Bandung hlm. 1 3. Upaya Penanggulangan Tindak Pidana Peredaran Pangan Berbahaya di Indone sia dan di Tiongkok ) ) Memastikan ketersediaan pangan yang memenuhi standar keamanan bagi manu ) h d l k d d b 4) Berusaha dalam menciptakan sistem perdagangan pangan yang jujur dan bertanggung jawab. 5) Berusaha dalam mewujudnya tingkat kecukupan pangan dan dengan harga yang ekonomis yang sesuai dengan kebutuhan masyarakat pada umumnya. 6) Tim pemeriksa yang akurat dalam keamanan pangan, serta peralatan yang memadai dalam pemeriksaan terhadap pangan. Langkah-langkah yang ditempuh pemerintah untuk mencapai tujuan kemanan pangan yaitu:30 Langkah-langkah yang ditempuh pemerintah untuk mencapai tujuan kemanan pangan yaitu:30 1) Registrasi dan penilaian; ) 2) Pengawasan produksi; 3) Pengawasan distribusi 4) Pembinaan dan pengembangan usaha; 4) Pembinaan dan pengembangan usaha; 5) Peningkatan dan pengembangan prasarana dan tenaga. Ada tiga prinsip yang menjadi dasar perlindungan hukum bagi konsumen di Indonesia yaitu prinsip perlindungan kesehatan/harta konsumen, prinsip perlindungan atas barang dan harga serta prinsip penyelesaian sengketa secara patut. Di dalam UUPK juga secara tegas memuat prinsip ganti kerugian subjek terbatas dan prinsip tanggung jawab Jurnal IUS Kajian Hukum dan Keadilan 449 urnal IUS Kajian Hukum dan Keadilan \| Vol. 9 \| Issue 2 \| August 2021 \| hlm, 450~457 gugat berdasarkan kesalahan dengan pembuktian terbalik.31 Dalam upaya memberantas kejatahan pangan, pemerintah juga akan memberikan sanksi tegas kepada produsen yang masih melakukan curang dalam menjalankan usahanya. Undang-undang Nomor 8 Tahun 1999, sanksi yang dikenakan dapat dibagi dua yaitu sanksi administratif (denda) dan sanksi pidana. Pertanggung jawaban pidana hanya dapat terjadi setelah seseorang melakukan tindak pidana. Pertanggung jawaban pidana dilakukan atas dasar asas hukum yang tidak tertulis “tiada pidana tanpa kesalahan”. Setiap usaha untuk memidanakan orang, baik orang perseorangan maupun korporasi, karena tindak pidana yang dilakukan orang lain tanpa terlebih dahulu ditetapkan sebagai tindak pidana dalam undang-undang adalah pelanggaran atas asas legatitas.32 Selain menerapkan sanksi pidana, pemerintah Indonesia juga mengadakan regulasi- regulasi dalam bidang pangan. Dasar hukum yang berlaku saat ini menjadi dasar dalam kajian pemerintah daerah. Hal ini dimaksudkan agar rekomendasi dari kajian ini dapat diaplikasikan oleh semua lembaga terkait, termasuk juga pemda.33 Upaya penanggulangan penting dilakukan upaya Penal (tindakan) dan non penal (pencegahan). Pihak BPOM dan kepolisian lebih meningkatkan upaya pencegahan dengan melakukan razia. Memberikan pemahaman dan pengetahuan kepada masyarakat dengan memberi penyuluhan34 tentang pangan berbahaya. Tidak beda jauh dengan negara Indonesia, Tiongkok juga terus meningkatkan sistem perlindungan konsumen supaya tidak terjadi lagi kasus pangan berbahaya yang sangat merugikan konsumen. , , p p p g g , j da, Jakarta, hlm 215. 32 Chairul Huda, “Dari Tiada Pidana Tanpa Kesalahan Menuju Kepada Tiada Pertanggungjawaban Pidana Tanpa Kesalahan” , Penerbit Kencana, Jakarta, 2006, hlm 46. 33 Hariyanto Kurniawan, (2008), “Laporan Lengkap Penelitian Susu Formula IPB”, https://news.okezone.com/ read/2008/02/28/1/87568/inilah-laporan-lengkap-penelitian-susu-formulaipb, di akses pada tanggal 25 Januari 2020. 34 Kartina Pakpahan, Leviyanti, (2020), Pertanggungjawaban Pidana Korporasi Mengedarkan Makanan Olahan Tanpa Izin Edar, http://www.jurnal.utu.ac.id/jcivile/article/view/1927/1403, hlm 10. 450 Jurnal IUS Kajian Hukum dan Keadilan 31 Ahmadi Miru, 2011, Prinsip-prinsip Perlindungan Hukum Bagi Konsumen di Indonesia, Raja Grafindo Persa da, Jakarta, hlm 215. 32 Chairul Huda, “Dari Tiada Pidana Tanpa Kesalahan Menuju Kepada Tiada Pertanggungjawaban Pidana Tanpa Kesalahan” , Penerbit Kencana, Jakarta, 2006, hlm 46. 33 Hariyanto Kurniawan, (2008), “Laporan Lengkap Penelitian Susu Formula IPB”, https://news.okezone.com/ read/2008/02/28/1/87568/inilah-laporan-lengkap-penelitian-susu-formulaipb, di akses pada tanggal 25 Januari 2020. 34 Kartina Pakpahan, Leviyanti, (2020), Pertanggungjawaban Pidana Korporasi Mengedarkan Makanan Olahan Tanpa Izin Edar, http://www.jurnal.utu.ac.id/jcivile/article/view/1927/1403, hlm 10. Ahmadi Miru, 2011, Prinsip-prinsip Perlindungan Hukum Bagi Konsumen di Indonesia, Raja Grafindo Pers karta, hlm 215. h i l d i i d id l h j d i d j b id read/2008/02/28/1/87568/inilah-laporan-lengkap-penelitian-susu-formulaipb, di akses pada tanggal 25 Januari 2020. 34 Kartina Pakpahan, Leviyanti, (2020), Pertanggungjawaban Pidana Korporasi Mengedarkan Makanan Olahan Tanpa Izin Edar, http://www.jurnal.utu.ac.id/jcivile/article/view/1927/1403, hlm 10. 3. Upaya Penanggulangan Tindak Pidana Peredaran Pangan Berbahaya di Indone sia dan di Tiongkok Departement pemerintah Tiongkok berupaya dalam meningkatkan responsif terhadap peraturan pangan yang ada dan mengawasi juga menegakkan kebijakan keamanan pangan. Kantor Penerangan Dewan Tiongkok telah menerbitkan Buku Putih yang isinya mengenai keamanan dalam hal pangan. Buku Putih tersebut membahas tentang cara menggapai keberhasilan di bidang keamanan pangan, jalan keamanan pangan yang berkepribadian Tiongkok, keterbukaan dan kerja sama internasional, dan kebijakan yang terkait. Di dalam Buku Putih juga mengatakan bahwa negara Tiongkok mempertahankan pedoman swasembada bahan pangan domestik, penerapan sistem perlindungan tanah garapan yang sangat ketat, meningkatkan reformasi sisi suplai pertanian dan inovasi sistem dan mekanisme, kemampuan produksi bahan pangan terus bertambah, level modernisasi logistik bahan pangan terus meningkat, struktur pensuplaian bahan pangan terus menjadi optimal, ekonomi industri dibidang pangan berkembang secara stabil, sistem jaminan keamanan bahan pangan yang bertingkat lebih tinggi, lebih terjamin kualitasmya, lebih efektif dan lebih berkelanjutan dibentuk secara bertahap, jaminan keamanan bahan pangan menjadi lebih kuat, jalan keamanan bahan pangan yang berkepribadian Tiongkok menjadi semakin lebar. Buku putih menegaskan, nasib Negara Tiongkok ataupun dunia sangat berhubungan erat dengan jaminan keamanan bahan pangan. Tentunya Tiongkok akan tetap menerapkan pada prinsip terbuka, iklusif, sama derajat, saling menguntungkan, kerja sama dan menang bersama, berupaya membentuk 450 Jurnal IUS Kajian Hukum dan Keadilan P-ISSN: 2303-3827, E-ISSN: 2477-815X konfigurasi baru keterbukaan bahan pangan, bersama dengan berbagai negara untuk berkembang bersama, berupaya untuk memelihara keamanan bahan pangan dunia, memberi kontribusi baru untuk pembangunan komunitas senasib manusia.35 Selain itu, pada Tahun 2003 negara China mendirikan Administrasi Makanan dan Obat-Obatan Negara China (SFDA) dalam upayanya untuk meningkatkan keamanan pangan. SFDA bertanggung jawab bertanggung jawab untuk mengawasi dan mengoordinasikan lembaga kesehatan, makanan, dan obat-obatan. SFDA mencakup sepuluh departemen yang mengatur dan mengawasi berbagai aspek hukum makanan dan obat. Tugas umum dari SFDA adalah :36 1) Untuk menyusun Undang-Undang, peraturan, dan kebijakan kesehatan, untuk mengusulkan program pembangunan kesehatan. 1) Untuk menyusun Undang-Undang, peraturan, dan kebijakan kesehatan, untuk mengusulkan program pembangunan kesehatan. 2) Untuk mengusulkan program kesehatan regional, untuk perencanaan keseluruhan dan untuk mengoordinasikan alokasi sumber daya kesehatan nasional. 2) Untuk mengusulkan program kesehatan regional, untuk perencanaan keseluruhan dan untuk mengoordinasikan alokasi sumber daya kesehatan nasional. 3) Merumuskan program kerja dan kebijakan kesehatan pedesaan, serta perawatan kesehatan ibu dan anak-anak, untuk memandu pelaksanaan program kesehatan primer dan protokol teknis tentang perawatan kesehatan Ibu dan Anak. 4) Untuk menerapkan kebijakan “Pencegahan Pertama” dan untuk melakukan pendidikan kesehatan kepada masyarakat umum. ty review translate by google, diakses pada tanggal 20 Desember 2020. 37 David Wan, “Why were Your Prepackaged Food Products Reject By China?” ?” http://www.cirsreach.com/news/ Why_were_your_prepackaged_food_products_Biscuits_rejected_by_China.html, di akses pada tanggal 25 Januari 2020. 35 China Radio Internasional, (2011) “Tiongkok Luncurkan Buku Putih Keamanan Bahan Pangan” 35 China Radio Internasional, (2011) “Tiongkok Luncurkan Buku Putih Keamanan Bahan Pangan” 36 Hayun Sumirat, “SFDA Review”, https://www.scribd.com/doc/63902866/SFDA-Saudi-Food-and-Drugs-Authori ty-review-translate-by-google, diakses pada tanggal 20 Desember 2020. 37 David Wan, “Why were Your Prepackaged Food Products Reject By China?” ?” http://www.cirsreach.com/news/ Why_were_your_prepackaged_food_products_Biscuits_rejected_by_China.html, di akses pada tanggal 25 Januari 2020. 3. Upaya Penanggulangan Tindak Pidana Peredaran Pangan Berbahaya di Indone sia dan di Tiongkok Untuk mengembangkan program pencegahan dan pengobatan penyakit yang membahayakan kesehatan populasi, untuk mengatur pencegahan komprehensif dan pengobatan penyakit utama, untuk mempubuikasikan daftar karantina penyakit menular dan daftar pengawasan penyakit menular. 5) Untuk memandu reformasi lembaga medis, untuk merumuskan kriteria bagi praktisi medis. Di China juga mengatur aditif makanan harus secara teknis diperlukan dan terbukti aman dan andal berdasarkan penilaian risiko sebelum digunakan. Penggunaan aditif makanan harus memenuhi persyaratan berikut:37 1) Tidak dimaksudkan untuk menutupi tengik makanan; 2) Tidak dimaksudkan untuk menutupi kekurangan kualitas dari makanan itu sendiri atau selama pemrosesan makanan; ) 3) Aditif makanan tidak digunakan untuk pemalsuan; 4) Tidak mengurangi nilai gizi makanan; 5) Kurangi kadar dosis sebanyak mungkin setelah efek yang diantisipasi tercapai 6) Alat bantu pengolahan makanan yang digunakan dalam proses pengolahan makanan harus dihilangkan sebelum produk akhir diproduksi, kecuali jika tingkat residu ditentukan. Dalam upaya meningkatkan sistem keamanan pangan pemerintah Tiongkok juga membuat peraturan tentang bahan yang menjadi kontak langsung dengan makanan. Bahan dan artikel kontak makanan (FCM) di Tiongkok mencakup berbagai bahan dan produk yang bersentuhan dengan makanan atau diharapkan bersentuhan dengan makanan dalam kondisi penggunaan normal dan dapat diperkirakan sebelumnya. Untuk lebih spesifik, mereka termasuk bahan kemasan, wadah, alat dan peralatan serta 4 Jurnal IUS Kajian Hukum dan Keadilan 451 urnal IUS Kajian Hukum dan Keadilan \| Vol. 9 \| Issue 2 \| August 2021 \| hlm, 452~457 tinta cetak, perekat, minyak pelumas, dll. Mungkin menghubungi makanan secara langsung atau tidak langsung yang digunakan selama produksi makanan, pemerosesan, pengemasan, transportasi, penyimpanan, penjualan dan proses penggunaan (tidak termasuk deterjen, desinfektan, dan fasilitas transportasi air publik). Dalam artikel ini, kami akan memberikan pengantar tentang bagaimana mematuhi peraturan kontak makanan di China dengan fokus pada persyaratan umum untuk bahan kontak makanan, aditif kontak makanan, dan jenis khusus dari bahan kontak makanan dan artikel. Bahan kontak makanan (FCM) diatur sebagai produk terkait makanan di China. Undang-undang Keamanan Pangan Tiongkok melarang impor, penggunaan, atau penjualan produk yang terkait dengan makanan (misalnya zat tambahan makanan, bahan pengemas makanan) yang tidak memenuhi Standar Keamanan Pangan nasional yang berlaku. Pada November 2016, The National Health and Family Planning Commission of the People’s Republic of China (NHFPC) menerbitkan 53 standar keamanan pangan nasional yang terkait dengan bahan kontak makanan dan zat tambahan kontak makanan. Standar keamanan pangan tersebut merupakan kerangka peraturan lengkap untuk bahan kontak makanan di China. 38 Cathy Yu, “Food Safety and Regulatory Department”, CIRS, http://www.Cirsreach.com/China_ Chemical_Regulation/Food_Contact_Regulations_Food_Contact_Materials_in_China.html, di akses pada tanggal 25 Januari 2020. 3. Upaya Penanggulangan Tindak Pidana Peredaran Pangan Berbahaya di Indone sia dan di Tiongkok 53 standar nasional tersebut dapat dibagi menjadi 4 kategori (lihat tabel di bawah):38 1 standar untuk persyaratan umum a. GB 4806.1-2016 Standar Keamanan Pangan Nasional: Persyaratan Keselamatan Umum untuk Bahan dan Artikel Kontak Pangan. a. Ini menetapkan persyaratan keselamatan umum yang berlaku untuk mu j ni b h n k nt k m k n n d n rtik l 1 standar untuk persyaratan umum a. GB 4806.1-2016 Standar Keamanan Pangan Nasional: Persyaratan Keselamatan Umum untuk Bahan dan Artikel Kontak Pangan. a. Ini menetapkan persyaratan keselamatan umum yang berlaku untuk semua jenis bahan kontak makanan dan artikel. b. Mulai berlaku: 19 Oktober 2017 1 standar untuk aditif kontak makanan a. GB 9685-2016 Standar Keamanan Pangan Nasional: Standar untuk Penggunaan Aditif dalam Bahan dan Artikel Kontak Pangan; b. Ini menetapkan daftar aditif positif yang dapat digunakan untuk menghasilkan berbagai bahan dan artikel kontak makanan, ruang lingkup penggunaan dan pembatasannya. c. Mengganti GB 9685-2008. 1 standar untuk aditif kontak makanan a. GB 9685-2016 Standar Keamanan Pangan Nasional: Standar untuk Penggunaan Aditif dalam Bahan dan Artikel Kontak Pangan; b. Ini menetapkan daftar aditif positif yang dapat digunakan untuk tif kontak kanan untuk Penggunaan Aditif dalam Bahan dan Artikel Kontak Pangan; b. Ini menetapkan daftar aditif positif yang dapat digunakan untuk menghasilkan berbagai bahan dan artikel kontak makanan, ruang lingkup penggunaan dan pembatasannya. c. Mengganti GB 9685-2008. d. Mulai berlaku: 19 Oktober 2017 P-ISSN: 2303-3827, E-ISSN: 2477-815X 10 standar untuk jenis bahan dan a. GB 4806.3-2016 Standar keamanan pangan nasional untuk produk enamel. 10 standar untuk jenis bahan dan a. GB 4806.3-2016 Standar keamanan pangan nasional untuk produk enamel. u b. GB 4806.4-2016 Standar keamanan pangan nasional untuk produk keramik artikel tertentu b. GB 4806.4-2016 Standar keamanan pangan nasional untuk produk keramik c. GB 4806.5-2016 Standar keamanan pangan nasional untuk produk gelas c. GB 4806.5-2016 Standar keamanan pangan nasional untuk produk gelas d. GB 4806.6-2016 Standar keamanan pangan nasional - resin plastik untuk penggunaan kontak makanan d. GB 4806.6-2016 Standar keamanan pangan nasional - resin plastik untuk penggunaan kontak makanan e. GB 4806.7-2016 Standar keamanan pangan nasional - bahan plastik dan barang-barang untuk penggunaan kontak makanan e. GB 4806.7-2016 Standar keamanan pangan nasional - bahan plastik dan barang-barang untuk penggunaan kontak makanan f. GB 4806.8-2016 Standar keamanan pangan nasional - kertas dan kertas karton untuk penggunaan kontak makanan f. 3. Upaya Penanggulangan Tindak Pidana Peredaran Pangan Berbahaya di Indone sia dan di Tiongkok GB 4806.8-2016 Standar keamanan pangan nasional - kertas dan kertas karton untuk penggunaan kontak makanan g. GB 4806.9-2016 Standar keamanan pangan nasional - bahan logam dan barang untuk penggunaan kontak makanan g. GB 4806.9-2016 Standar keamanan pangan nasional - bahan logam dan barang untuk penggunaan kontak makanan h. GB 4806.10-2016 Standar keamanan pangan nasional - cat dan pelapis untuk penggunaan kontak makanan i. GB 4806.11-2016 Standar keamanan pangan nasional - bahan dan produk karet untuk penggunaan kontak makanan j. GB 14934-2016 Standar keamanan pangan nasional - Peralatan makan disinfeksi a. GB 5009.156-2016 Standar keamanan pangan nasional - prinsip umum metode pretreatment untuk uji migrasi bahan kontak makanan dan artikel b. GB31604.1-2015 Standar keamanan pangan nasional - prinsip umum untuk uji migrasi bahan dan artikel kontak makanan. c. GB 31604.49-2016 Penentuan arsenik, kadmium, kromi um, timbal dan penentuan arsenik, kadmium, kromium, ni kel, timah, antimon dan seng dalam bahan kontak makanan dan barang-barang; dan d. Metode pengujian lain untuk pengujian migrasi bahan ber bahaya spesifik formaldehida, Pb, dan lain-lain. umber : https://www.chemsafetypro.com/Topics/Food_Contact/How_to_Comply_with_Food_Contact_Regulations_in_China.htm Sebagai negara maju yang mencapai seperlima penduduk dunia, menghidupkan diri sendiri justru adalah kontribusi besar bagi dunia. Namun, di samping mengembangkan diri sendiri, Tiongkok terus meningkatkan keterbukaan dan kerja sama internasional, memberi sumbangan positif bagi keamanan bahan pangan dunia, hal itu membuktikan bahwa Tiongkok memikul kewajibannya sebagai negara besar dalam menyelesaikan masalah kelaparan di lingkup seluruh dunia. Perbandingan antara Negara Indonesia dengan Negara Tiongkok sebagai berikut : Jurnal IUS Kajian Hukum dan Keadilan 453 Jurnal IUS Kajian Hukum dan Keadilan \| Vol. 9 \| Issue 2 \| August 2021 \| hlm, 454~45 Bentuk perlindungan hukum diatur dalam : Bentuk perlindungan hukum diatur dalam : Bentuk perlindungan hukum diatur dalam peraturan : - UU Perlindungan Konsumen - UU Pangan - KUHP - Peraturan pemerintah - Peraturan BPOM - Buku Putih - UU Perlindungan Konsumen - CFDA ( China Food and Drug Administration) - CFDA ( China Food and Drug Administration) - SFDA ( State Food and Drug Ad ministration) - SFDA ( State Food and Drug Ad ministration) - Peraturan pemerintah - Peraturan BPOM - NHFPC ( National Health and Family Planning Commission of - NHFPC ( National Health and Family Planning Commission of the people’s Republic of China) Hukuman : the people’s Republic of China) Hukuman : Diatur dalam pasal 62 UU No 8 Tahun 1999 Tentang perlindungan konsumen Diatur dalam UU Keamanan Pangan Re pulik Rakyat Tiongkok Bab 9 Tanggung jawab hukum pasal 74 – 75 yaitu setiap produsen yang melanggar akan dikena kan sanksi penalty berupa denda sebe sar 10.000 yuan dan memberhentikan produksi makanan. Ancaman penjara seumur hidup jika terjadi kematian pada konsumen - Bisa berupa denda (10 Miliyar jika mengedarkan pangan berbahaya) - Bisa berupa hukuman penjara ( 5 tahun penjara jika mengedarkan pangan berbahaya) Lembaga pengawasan : Lembaga pengawasan : 454 Jurnal IUS Kajian Hukum dan Keadilan Lembaga pengawasan : Lembaga pengawasan : CFDA yang bertugas menjadi lembaga dewan asosiasi dalam hal makanan dan obat-obatan di China. Masyarakat China juga bisa melapor di CFDA jika ada ter jadi korban pangan berbahaya. BPOM (badan pengawasan obat dan ma kanan) yang bertugas mengawasi per edaran obat dan makanan di seluruh In donesia. Makanan yang diproduksi harus memiliki sertifikat lulus uji BPOM baru dinyatakan aman untuk di konsumsi masyarakat Indonesia Upaya penanggulangan di Tiongkok yai tu terdapat pada tugas dan wewenang SFDA : Upaya penanggulangan di Tiongkok yai tu terdapat pada tugas dan wewenang SFDA : Upaya penanggulangan dari pemerintah : Upaya penanggulangan dari pemerintah : SIMPULAN Perlindungan Hukum Terhadap Tindak Pidana Peredaran Pangan Berbahaya di Indonesia pada Pasal 8 UU Nomor 8 Tahun 1999 Tentang Perlindungan Konsumen, melarang pelaku usaha memproduksi dan/atau memperdagangkan barang dan/atau jasa tidak memenuhi standar keamaman pangan dan mutu pangan. Peraturan pelaksana yaitu Pasal 23 PP Nomor 28 Tahun 2004 Tentang Keamanan Mutu dan Gizi Makanan. Negara Tiongkok mewajibkan pelaku usaha memperhatikan standar lokal keamanan pangan Pasal 29 Undang-Undang Kemanan Pangan di Tiongkok. Masyarakat ikut serta terlibat langsung mengawasi penerapan peraturan Keamanan Pangan Tiongkok, dapat melaporkan ke dewan pengurus asosiasi China Food And Drug Administration (CFDA). Jika tidak sesuai standar dapat memberikan Penalty denda kepada pengusaha. Upaya Penanggulangan Tindak Pidana Peredaran Pangan Berbahaya di Indonesia dilakukan Upaya Pre-Entif, Preventif dan Represif, menerapkan hukuman pidana penjara dan denda. Di Tiongkok adanya pengawasan oleh SFDA penerbitan 53 standar keamanan pangan nasional, menerapkan hukuman pidana penjara dan denda. P-ISSN: 2303-3827, E-ISSN: 2477-815X P-ISSN: 2303-3827, E-ISSN: 2477-815X P-ISSN: 2303-3827, E-ISSN: 2477-815X Buku Ahmadi Miru, (2011), Prinsip-prinsip Perlindungan Hukum Bagi Konsumen di Indonesia, Raja Grafindo Persada. Andi Hamzah , (2001), Bunga Rampai Hukum Pidana dan Acara Pidana, Ghalia Indonesia, Jakarta. A.S Alam. (2010). Pengantar Kriminologi, Pustaka Refleksi, Makassar. Chairul Huda, (2006), Dari Tiada Pidana Tanpa Kesalahan Menuju Kepada Tiada Pertanggungjawaban Pidana Tanpa Kesalahan, Penerbit Kencana, Jakarta. Christine F. Mamuja, (2016), Pengawasan Mutu dan Keamanan Pangan, Unsrat Press, Manado. H. Ishaq, (2017), Metode Penelitian Hukum, Penerbit Alfabeta dan Bandung. Janus Sidabalok, (2014), Hukum Perlindungan Konsumen di Indonesia, PT. Citra Aditya Bakti, Bandung. Upaya penanggulangan dari pemerintah : Upaya penanggulangan dari pemerintah : - Menerapkan sistem jaminan mutu dan keamanan pangan - Menyusun rangkaian Undang- Undang - Menerapkan cara memproduksi yang lebih baik - Mengusulkan program kesehatan regional - Memastikan ketersediaan pan gan yang memenuhi standar kea manan bagi manusia - Merumuskan program kerja dan kebijakan kesehatan pedesaan, serta perawatan kesehatan ibu dana nak-anak - Menciptakan sistem perdagangan pangan yang jujur - Mewujudkan tingkat kecuku pan pangan dan dengan harga ekonomis yang sesuai kebutuhan masyarakat pada umumnya - Untuk menerapkan kebijakan “Pencegahan Pertama” dan untuk melakukan pendidikan kesehatan kepada masyarakat umum. - Memandu reformasi lembaga me dis, untuk merumuskan kriteria bagi praktisi medis. 454 Jurnal IUS Kajian Hukum dan Keadilan Jurnal dan Karya Ilmiah Ade Setiawan, (2015), Strategi Tiongkok Memasarkan Produk-Produk di Indonesia, Jurnal Ilmu Hubungan Internasional, Vol. 2 No. 2 Edisi Oktober, https://media.neliti.com/ media/publications/32303-ID-strategi-tiongkok-memasarkan-produk-produk- elektronik-di-indonesia-tahun-2010-21.pdf, diakses tanggal 24 Februari 2021. Asri, (2016) “Perlindungan Hukum Bagi Konsumen Terhadap Pangan” , Jurnal IUS, Vol.4,No.2 http://jurnalius.ac.id/ojs/index.php/jurnalIUS/article/view/316/pdf_2, di akses pada tanggal 27 Desember 2019. BBC News (2017), Cina selidiki pusat produksi saus dan bumbu penyedap makanan palsu, Majalah, https://www.bbc.com/indonesia/majalah-38650024, di akses pada tanggal 15 Agustus 2021. Jurnal IUS Kajian Hukum dan Keadilan 455 urnal IUS Kajian Hukum dan Keadilan \| Vol. 9 \| Issue 2 \| August 2021 \| hlm, 456~457 BKP, “Laporan Kinerja Pusat Penganekaragaman Konsumsi dan Keamanan Pangan” http://bkp.pertanian.go.id/storage/app/media/Konsumsi/LAKIN%20 KONSUMSI%202018.pdf, di akses pada tanggal 25 Januari 2020. Cathy Yu, “Food Safety and Regulatory Department”, CIRS http://www.Cirsreach.com/ China_ Chemical_Regulation/Food_Contact_Regulations_Food_Contact_Materials_in_China. html, di akses pada tanggal 25 Januari 2020. China Radio Internasional, (2011) “Tiongkok Luncurkan Buku Putih Keamanan Bahan Pangan”http://indonesian.cri.cn/20191015/309f8831-4c09-dfd4-e1cb- d76f43f588b6.html, di akses pada tanggal 25 Januari 2020. David Wan, “Why were Your Prepackaged Food Products Reject By China?” http://www. cirsreach.com/news/Why_were_your_prepackaged_food_products_Biscuits_ rejected_by_China.html, di akses pada tanggal 25 Januari 2020. Detik Finance, “Skandal Makanan di Tiongkok, dari Daging Busuk Sampai Susu Melamin”https://finance.detik.com/berita-ekonomi-bisnis/d-2646629/ skandal-makanan-di-tiongkok-dari-dari-daging-busuk-sampai-susu-melamin, di akses pada tanggal 20 Desember 2019. Dewi Nurchayati, (2015), “Upaya Meningkatkan Pengetahuan Makanan Sehat”, Jurnal Penelitian PAUDIA, https://core.ac.uk/download/pdf/234023139.pdf , diakses tanggal 03 Maret 2021 Elly, (2014),“BPKN Indonesia Belajar Dari Tiongkok”, Indonesia Media,http://www. indonesiamedia.com/bpkn-indonesia-belajar-dari-tiongkok/, di akses pada tanggal 18 Januari 2020. Elvira Syamsir, (2016) “Apa sih Tanggal Kadaluarsa itu?” , Personal Blog Ilmu Pangan, https://ilmupangan.blogspot.com/2016/09/apa-sih-tanggal-kadaluarsa-itu. html, di akses pada tanggal 5 Januari 2020. Dewan Negara, (2019) “Peraturan Pelaksanaan Undang-Undang Keamanan Pangan Republik Rakyat Tiongkok” http://www.gov.cn/zhengce/cotent/2019-10/31/ content_5447142.html, di akses pada tanggal 20 Januari 2020. Green Pack, “Apa itu FDA dan FSC?” http://www.greenpack.co.id/id/apa-itu-fda-dan- fsc/#.XhLL8hsxXIV, di akses pada tanggal 20 Januari 2020. Hariyanto Kurniawan, (2008), “Laporan Lengkap Penelitian Susu Formula IPB”, Oke News https://news.okezone.com/read/2008/02/28/1/87568/inilah-laporan-leng kap-penelitian-susu-formulaipb, di akses pada tanggal 25 Januari 2020. Hayun Sumirat, “SFDA Review”, https://www.scribd.com/doc/63902866/SFDA-Saudi- Food- and-Drugs-Authority-review-translate-by-google, diakses pada tanggal 20 Desember 2020. Humas BSN, (2016), “Penerapan SNI untuk Mutu dan Keamanan Pangan”, Badan Standardisasi Nasional https://www.bsn.go.id/main/berita/beritadet/7578/ Memperbaiki-Konsultasi-Publik-DalamPenyusunan-Regulasi-Yang-Baik, di akses pada tanggal 24 Januari 2020. Kang Ismed, “Teknologi Makanan”, Pengolahan Pangan.Com https://www. pengolahanpangan.com/2013/12/jenis-jenis-kerusakan-bahan-pangan.html, di akses pada tanggal 6 Januari 2020. 456 Jurnal IUS Kajian Hukum dan Keadilan 456 Jurnal IUS Kajian Hukum dan Keadilan P-ISSN: 2303-3827, E-ISSN: 2477-815X Kartika Sari Tarigan, “BPOM : Jual Pangan Tak Penuhi Syarat Mutu Itu Kriminal, Ada Sanksi” , Detik.Com, https://news.detik.com/berita/d-3102164/bpom-jual- pangan-tak-penuhi-syarat-mutu-itukriminal-ada-sanksi, di akses pada tanggal 18 Januari 2020. Jurnal dan Karya Ilmiah Kartina Pakpahan, Leviyanti, (2020), Pertanggungjawaban Pidana Korporasi Mengedarkan Makanan Olahan Tanpa Izin Edar, http://www.jurnal.utu.ac.id/ jcivile/article/view/1927/1403, diakses 24 Januari 2020. Parliamentary Center, (2011), “Kronologi Kasus Susu Melamin” http:// kebebasaninformasiorg/2011/02/17/kronologi-kasus-%E2%80%9Csusu- formula%E2%80%9D/ , tanggal akses tanggal 20 Desember 2019. Setiono, (2004), Rule Of Law (Supremasi Hukum), Magister Ilmu Hukum Program Pascasarjana Universitas Sebelas Maret, Surakarta. Sovia Hasanah, (2016)“Pidana Bagi Penjual Makanan Berbahaya”, Hukum Online. Com https://www.hukumonline.com/klinik/detail/ulasan/lt5855165331751/ pidana-bagi-penjualmakanan-yangmengandung-bahan-berbahaya/, di akses pada tanggal 20 Desember 2019. Viva, (2008)“Kronologi Skandal Susu Beracun China” https://m.viva.co.id/amp/berita/ dunia/679-kronologi-skandal-susu-beracun-china, di akses pada tanggal 20 Januari 2020. Yhona Paratmanitya , Veriani Aprilia (2016), Kandungan bahan tambahan pangan berbahaya pada makanan jajanan anak sekolah di Kabupaten Bantul, Jurnal gizi dan dietetik Indonesia Vol. 4 No. 1 edisi Januari 2016 https://ejournal.almaata. ac.id/index.php/IJND/ article/view/329, diakses tanggal 15 Agustus 2021. Yovita Diane Titiesari, (2018), “ Zat Kimia Berbahaya yang Sering Digunakan dalam Makanan”, GueSehat.Com, https://www.guesehat.com/4-zat-kimia-berbahaya- yang-sering-digunakan-dalam-makanan, di akses pada tanggal 11 Januari 2020. Zenzen Zaunidhin, (2015) “Standar Kebutuhan Gizi Masyarakat”, Agrotani.Com,https:// www.agrotani.com/standar-kebutuhan-gizi-masyarakat-di-indonesia-2016-2/, diakses pada tanggal 5 Januari 2020. Peraturan Perundang-Undangan Peraturan Perundang-Undangan Peraturan Pemerintah Nomor 28 Tahun 2004 Tentang Keamanan, Mutu dan Gizi Pangan Lembaran Negara Republik Indonesia Tahun 2004 Nomor 107 Tambahan Lembaran Negara Republik Indonesia Nomor 4424. Undang-Undang No. 8 Tahun 1999 Tentang Perlindungan Konsumen Lembaran Negara Republik Indonesia Tahun 1999 Nomor 42 Tambahan Lembaran Negara Republik Undang-Undang No. 8 Tahun 1999 Tentang Perlindungan Konsumen Lembaran Negara Republik Indonesia Tahun 1999 Nomor 42 Tambahan Lembaran Negara Republik Indonesia Nomor 3821. dang-Undang Nomor 28 Tahun 2004 Tentang Keamanan, Mutu dan Gizi Makanan Lembaran Negara Republik Indonesia Tahun Tahun 2004 Nomor 115 Tambahan Lembaran Negara Republik Indonesia Nomor 4430. dang-Undang Nomor 18 Tahun 2012 Tentang Pangan Lembaran Negara Republik Indonesia Tahun 2012 Nomor 227 Tambahan Lembaran Negara Republik Indonesia Nomor 8360. Jurnal IUS Kajian Hukum dan Keadilan 457
https://openalex.org/W2745195379	https://europepmc.org/articles/pmc5542432?pdf=render	English	null	The effects of various diets on glycemic outcomes during pregnancy: A systematic review and network meta-analysis	PloS one	2,017	cc-by	10,474	RESEARCH ARTICLE Methods MEDLINE, EMBASE, Cochrane database, and reference lists of published studies were searched through April 2017. Randomized trials directly comparing two or more diets for 2-weeks were eligible. Bayesian network meta-analysis was performed for fasting glu- cose. Owing to a lack of similar dietary comparisons, a standard pairwise meta-analysis for the other glycemic outcomes was performed. The certainty of the pooled effect estimates was assessed using the GRADE tool. Copyright: © 2017 Ha et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. OPEN ACCESS Citation: Ha V, Bonner AJ, Jadoo JK, Beyene J, Anand SS, de Souza RJ (2017) The effects of various diets on glycemic outcomes during pregnancy: A systematic review and network meta- analysis. PLoS ONE 12(8): e0182095. https://doi. org/10.1371/journal.pone.0182095 Evidence to support dietary modifications to improve glycemia during pregnancy is limited, and the benefits of diet beyond limiting gestational weight gain is unclear. Therefore, a sys- tematic review and network meta-analysis of randomized trials was conducted to compare the effects of various common diets, stratified by the addition of gestational weight gain advice, on fasting glucose and insulin, hemoglobin A1c (HbA1c), and homeostatic model assessment for insulin resistance (HOMA-IR) in pregnant women. Editor: Marly Augusto Cardoso, Universidade de Sao Paulo, BRAZIL Received: January 18, 2017 Accepted: July 12, 2017 Published: August 3, 2017 Received: January 18, 2017 Accepted: July 12, 2017 Published: August 3, 2017 * rdesouz@mcmaster.ca Vanessa Ha1,2, Ashley J. Bonner1,2, Jaynendr K. Jadoo1, Joseph Beyene1,2,3,4, Sonia S. Anand1,2,5,6,7, Russell J. de Souza1,2* 1 Department of Health Research Methods, Evidence, and Impact, Faculty of Health Sciences, McMaster University, Hamilton, Ontario, Canada, 2 Population Genomics Program, Chanchlani Research Centre, McMaster University, Hamilton, Ontario, Canada, 3 The Dalla Lana School of Public Health, Faculty of Medicine, University of Toronto, Toronto, ON, Canada, 4 The Hospital for Sick Children Research Institute, Toronto, ON, Canada, 5 Department of Medicine, Faculty of Health Sciences, McMaster University, Hamilton, Ontario, Canada, 6 Population Health Research Institute, Hamilton Health Sciences, McMaster University, Hamilton, Ontario, Canada, 7 General Hospital Campus, Hamilton, Ontario, Canada a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 * rdesouz@mcmaster.ca Dietary patterns and glycemic outcomes during pregnancy Registration PROSPERO CRD42015026008 Results Data Availability Statement: All relevant data are within the paper and its Supporting Information files. Twenty-one trials (1,865 participants) were included. In general, when given alongside ges- tational weight gain advice, fasting glucose improved in most diets compared to diets that gave gestational weight gain advice only. However, fasting glucose increased in high unsat- urated or monounsaturated fatty acids diets. In the absence of gestational weight gain ad- vice, fasting glucose improved in DASH-style diets compared to standard of care. Although most were non-significant, similar trends were observed for these same diets for the other glycemic outcomes. Dietary comparisons ranged from moderate to very low in quality of evidence. Funding: V.H. received research support from the Canadian Institutes of Health Research (FRN#- 336515), David L Sackett Scholarship, and Ashbaugh Graduate Scholarship. S.S.A. is supported by a Canada Research Chair in Ethnicity and Cardiovascular disease and the Heart and Stroke Foundation of Canada Michael DeGroote Chair in Population Health. The funders had no role 1 / 17 PLOS ONE \| https://doi.org/10.1371/journal.pone.0182095 August 3, 2017 Introduction The need for implementation of effective dietary strategies in gestational diabetes mellitus (GDM) prevention and management has been emphasized by diabetes organizations [1–3]. Most women also prefer to not use medications to manage their diabetes risk during pregnancy [4]. One method of managing GDM risk is the use of dietary strategies. Data from individual randomized trials suggest benefits of dietary strategies in diabetes control [5–7]. The success of diet and lifestyle changes in managing type 2 diabetes mellitus (T2DM), some of its etiology shared with GDM, in high-risk patients further emphasize the importance of dietary strategies in GDM management [8]. Nonetheless, the evidence to support the application of dietary strategies to the treatment of GDM is lacking [1–3]. Further, a clear benefit for dietary strate- gies have not been demonstrated in recent meta-analyses [9, 10]. However, these analyses have usually been limited to single pair-wise dietary comparisons with a small number of partici- pants. Furthermore, single pair-wise comparisons do not lend itself easily to determine if it is the most effective strategy amongst all the possible dietary strategies for GDM control. gy g y g The above concerns are reflected in current dietary guidelines for GDM prevention and management. Recommendations by the Canadian Diabetes Association (CDA) have not been updated in almost a decade and most are based on expert consensus, despite that dietary inter- ventions are recommended as the first-line of therapy [3]. This has been echoed by the Ameri- can Diabetes Association (ADA) and the National Institute of Health and Care Excellence (NICE) in the U.K., both of which claim no evidence-based recommendations can be made given the lack of high-quality research in this area [2, 11]. Although the importance of diet is acknowledged in GDM prevention and management, current dietary recommendations for GDM are sparse, and where it exists, is outdated or based on experts’ opinion [1–3]. Our goal in this study was to conduct a systematic review and network meta-analysis (NMA) of randomized trials to compare and rank the relative efficacy of various diets on gly- cemic outcomes in pregnant women with or without diabetes. Our analysis was stratified based on whether gestational weight gain (GWG) advice was given in addition to the dietary interventions so that the effects of diet can be isolated. Conclusion/Interpretation in study design, data collection and analysis, decision to publish, or preparation of the manuscript. in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Alongside with gestational weight gain advice, most diets, with the exception of a high unsat- urated or a high monounsaturated fatty acid diet, demonstrated a fasting glucose improve- ment compared with gestational weight gain advice only. When gestational weight gain advice was not given, the DASH-style diet appeared optimal on fasting glucose. However, a small number of trials were identified and most dietary comparisons were underpowered to detect differences in glycemic outcomes. Further studies that are high in quality and ade- quately powered are needed to confirm these findings. Competing interests: Dr. de Souza has served as an external resource person to the World Health Organization’s Nutrition Guidelines Advisory Group on trans fats, saturated fats, and polyunsaturated fats. The WHO paid for his travel and accommodation to attend meetings from 2012- 2016 to present and discuss this work. He has also done contract research for the Canadian Institutes of Health Research’s Institute of Nutrition, Metabolism, and Diabetes, Health Canada, and the World Health Organization for which he received remuneration. He has held a grant from the Canadian Foundation for Dietetic Research as a principal investigator, and is a co-investigator on several funded team grants from Canadian Institutes of Health Research. Dr. de Souza received compensation for a lecture on dietary fat given at McMaster Pediatric Nutrition Days in 2016. He also currently serves as an academic editor for PLOS One. This does not alter our adherence to PLOS ONE policies on sharing data and materials. Registration PROSPERO CRD42015026008 Registration PROSPERO CRD42015026008 Study selection Each study identified by the electronic or manual search was screened by title and abstract to assess for inclusion by one reviewer (V.H.). Studies that passed the title/abstract screening were retrieved for full-text review. Eligible studies were randomized trials that examined the effect of one dietary intervention compared to another dietary intervention or standard of care on glycemic outcomes in pregnant women with or without diabetes and who were followed for at least two-weeks. A minimum of two-weeks of follow-up duration was chosen in accor- dance with diabetes guidelines which recommend that dietary therapy should be given for at least two-weeks before the use of insulin therapy [1–3]. Fasting glucose (FG) and insulin (FI), hemoglobin-A1c (HbA1c), and homeostatic model assessment for insulin resistance (HOMA- IR) were glycemic outcomes of interest. No restriction was placed on language. Data extraction and quality assessment Study characteristics and data from eligible studies were independently extracted by two reviewers (V.H. and J.K.J.). Extracted data included article citation, study design, participant characteristics, dietary interventions and macronutrient composition, level of feeding control, institution and country at which the study was conducted, study results, and statistical tests used. To ensure accuracy, extracted data were compared between the two reviewers and any discrepancies were resolved through consensus. The quality of evidence for each dietary comparison was assessed using the Grading of Rec- ommendations, Assessment, Development and Evaluation (GRADE) approach [14]. The over- all quality of evidence for each dietary comparison was rated as high, moderate, low, or very low. Depending on the type of evidence in question, the starting point for GRADE assessment differed. Direct comparisons, where head-to-head comparisons from randomized trials were available, started at high quality of evidence and were downgraded based on the degree of study limitation, imprecision of pooled effect estimates, inconsistency of results, indirectness, and publication bias. First-order indirect comparisons, where two interventions had been indi- vidually compared against one common comparator but not with each other, started at the lower rating of the two dietary comparisons that made up the link and were downgraded based on evidence of intransitivity. Second and higher order indirect comparisons, where 2 common comparators were found between the two interventions being compared, were always rated as very low because of the distance between the two dietary interventions being compared. Data sources and searches MEDLINE, EMBASE, and Cochrane were searched up until April 2017 (S1 Table). A manual search of the references of the included studies was also conducted to identify additional eligi- ble studies. Materials and methods Study protocol The Cochrane Handbook for Systematic Reviews of Interventions (version 5.2) [12] and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) for network 2 / 17 PLOS ONE \| https://doi.org/10.1371/journal.pone.0182095 August 3, 2017 Dietary patterns and glycemic outcomes during pregnancy meta-analyses [13] was followed for analysis and reporting of results, respectively. The PRISMA checklist is provided in the supplemental material (S1 File). The protocol was regis- tered with PROSPERO (CRD42015026008) (S2 File). meta-analyses [13] was followed for analysis and reporting of results, respectively. The PRISMA checklist is provided in the supplemental material (S1 File). The protocol was regis- tered with PROSPERO (CRD42015026008) (S2 File). Dietary patterns and glycemic outcomes during pregnancy Prior to conducting the network meta-analysis, the assumptions of homogeneity and transi- tivity were assessed. Homogeneity, which reflects the degree of similarity between the effect estimates of each trial within the same dietary comparison, was assessed using Higgins criteria for I2 [12]. The I2 was chosen because it quantifies the degree of variation between trials that is due to inter-study heterogeneity and not by chance. Transitivity, which reflects the distribu- tion of effect modifiers between trials, was assessed by examining the distribution of a priori effect modifiers for both direct and indirect dietary comparisons including stage of pregnancy (first, second, or third trimester), diagnosis of GDM (yes or no), pre-pregnancy body weight (as a continuous variable), and ethnicity (Europeans, Asians, Africans, or others). A NMA for FG was performed. Relative effect estimates from the NMA are expressed as median differences (MeD) with 95% credible intervals (CrIs). MeD and their CrIs can be inter- preted in the same manner as traditional mean differences (MD) with 95% confidence inter- vals (CIs). The FG achieved at the end of each dietary intervention for each included trial was extracted and pooled using the Bayesian fixed effects model, with a minimally informative prior distribution for relative treatment effects. A fixed effects model was chosen because it had a lower deviance information criterion (DIC) compared to the random effects model, sug- gesting a better model fit. Non-informative prior distributions were chosen for model parame- ters so that results were driven entirely by the reported data. Analyses were performed using Markov-Chain Monte-Carlo methods, a method that estimates the effect of each dietary com- parison by simulation, using four chains with 200,000 iterations and thinning interval of ten, after a burn-in of 100,000. Convergence of the chains was assessed using the Gelman plot and diagnostic test [15]. Consistency of direct and indirect sources of evidence within the network was assessed using the node-splitting method [16]. Statistical significance was considered when the CrIs did not cross the line of no effect. Surface Under the Cumulative Ranking (SUCRA) values were calculated to assist in deter- mining the probability of a given dietary intervention as being the best overall among the interventions compared, but this does not necessarily reflect that the dietary intervention is good to treat with as other important clinical factors are not considered in the calculation (e.g. patient preferences, cost-effectiveness, etc.). PLOS ONE \| https://doi.org/10.1371/journal.pone.0182095 August 3, 2017 Statistical analysis The network meta-analysis was conducted using R (version 3.2.0, R Project for Statistical Computing) with the gemtc and rjags packages, which interface with Just Another Gibbs Sam- pler (JAGS) software (version 3.4.0). 3 / 17 PLOS ONE \| https://doi.org/10.1371/journal.pone.0182095 August 3, 2017 The closer SUCRA is to 100, the more certain we are that it is the best overall and the closer it is to zero, the more certain we are that it is worst [17]. Ranks, cumulative ranks, and SUCRA values were considered as supplementary measures to the primary effect estimates for each dietary comparison because the former three measures are known to have substantive uncertainty [18]. Standard pair-wise meta-analyses for FI, HbA1c, and HOMA-IR were performed because they lacked a common dietary comparator that connected them to a network plot. Results were expressed as MD with 95% CIs. The glycemic outcome achieved at the end of each dietary intervention for each included trial was extracted and pooled using the fixed effects model as there were <10 studies included per analysis. Significance was considered when p<0.05. Analyses were stratified by whether advice regarding optimal weight gain during pregnancy was given in addition to the dietary intervention (“GWG advice”). Trials were considered to have given participants GWG advice if the investigators established energy requirements so that women would achieve appropriate GWG. Trials were grouped into “trials with GWG advice provided in both dietary arms” if the study was designed to include GWG advice in addition to the dietary interventions. In contrast, trials were grouped into “trials with no GWG advice” if no GWG advice was given at all. Finally, studies were grouped into “trials with GWG advice provided in one of the dietary arms” if only one of the dietary interventions included GWG advice but not the other. Studies, where GWG advice was given in only one of the dietary arm but not in the comparator, were not included in the NMA. Further, studies were not included in the NMA if they did not connect to the network plot due to a lack of a PLOS ONE \| https://doi.org/10.1371/journal.pone.0182095 August 3, 2017 4 / 17 Dietary patterns and glycemic outcomes during pregnancy common comparator. A standard pairwise meta-analysis was performed for these types of studies. common comparator. A standard pairwise meta-analysis was performed for these types of studies. Literature search and study characteristics Of the 5589 studies that were identified, twenty-one studies were included (Fig 1) [5, 6, 19– 37]. Ten trials were designed to include GWG advice in addition to the dietary intervention, five trials included GWG advice in only one of the dietary arms, and seven trials did not report giving any GWG advice in either arm. Participants were predominantly young women (median = 30.6 years [interquartile range (IQR): 29.5 to 30.9 years]) in their second trimester at the start of the study (median = 24.4 weeks [IQR: 20.8 to 28.5 weeks]) with some degree of glucose intolerance (S2 Table). Most par- ticipants were considered overweight based on their pre-pregnancy BMI (median = 26.6 kg/m2 [IQR: 23.5, 27.7]). Smokers were included in one trial only (20% of included participants). Fig 1. Flow of the literature search. https://doi.org/10.1371/journal.pone.0182095.g001 5 / 17 Dietary patterns and glycemic outcomes during pregnancy Overall, the baseline FG (median = 4.9 mmol/L [IQR: 4.7 to 5.0]) and HbA1c (median = 5.7% [IQR: 4.9 to 5.4%]) were within the normal range. The median baseline FI was 99.8 pmol/L (IQR: 63.8 to 135.2 pmol/L) and the median baseline HOMA-IR was 2.2 (IQR: 1.3 to 2.5). Overall, the baseline FG (median = 4.9 mmol/L [IQR: 4.7 to 5.0]) and HbA1c (median = 5.7% [IQR: 4.9 to 5.4%]) were within the normal range. The median baseline FI was 99.8 pmol/L (IQR: 63.8 to 135.2 pmol/L) and the median baseline HOMA-IR was 2.2 (IQR: 1.3 to 2.5). Macronutrient composition was targeted in twenty-three dietary arms. Carbohydrate (CHO) intake was the focus of fifteen dietary arms (a low- glycemic index or load [GI and GL, respec- tively] diet in six arms, a high-fibre diet in three, a low-GI/GL and high-fibre in one, a low-CHO and low GI diet in two, and a low-CHO diet in three). Fat intake was the focus of four dietary arms (low fat in one arm, high monounsaturated fatty acid (MUFA) intake in one, and high unsaturated fat intake in two), a low-CHO and high fat diet in three dietary arms, and a high-fibre and low-fat diet in one dietary arm. Diets that targeted whole patterns of food consumption were the focus of fourteen dietary arms. The Dietary Approach to Stop Hypertension (DASH) diet was used in three dietary arms, healthy eating was used in two dietary arms, calorie restriction only was used in nine dietary arms. Network assumptions The assumptions of homogeneity and transitivity for NMAs were reasonably met. No evidence of inter-study heterogeneity was found between trials of dietary comparisons that did not pro- vide GWG advice (I2 = 0%). Within trials that offered GWG advice, inter-study heterogeneity was low (range: 0 to 45.5%). Further, too few studies reported pre-pregnancy BMI (n = 8 trials) to assess whether the transitivity assumption was violated due to an imbalance on this charac- teristics across trials, but there was no evidence of an imbalanced distribution of effect modifi- ers for GDM diagnosis, ethnicity, and pregnancy stage. Literature search and study characteristics Standard of care, which were dietary arms with no dietary advice given or a standard macronutrient distribution (45–64% of energy from CHO: 10–35% of energy from protein: 20–35% of energy from fat) was followed, was used in seven dietary arms. Six trials were conducted in North America (Canada two, U.S. three, and Mexico one), seven trials were conducted in Europe (Italy and Denmark two each, and Finland, Ireland, and Poland had one each), three trials were conducted in Australia, and six were conducted in Asia (Iran and China had three trials each). The median follow-up duration was 11.0 weeks (IQR: 7.1 to 14.8 weeks). PLOS ONE \| https://doi.org/10.1371/journal.pone.0182095 August 3, 2017 Trials with GWG advice provided in both dietary arms Fasting glucose. GWG advice was given in addition to dietary interventions and had FG reported in nine trials (Fig 2) [5, 21, 22, 25, 28, 31–35]. Where direct comparisons were available, no between diet differences were observed (high unsaturated fat diets vs GWG advice only and high-MUFA diet vs GWG advice) (Fig 3). Using indirect comparisons, in general, FG increased in diets that modified fat quality intake compared with other diets. FG increase was observed in four out of the six dietary compari- sons that prescribed a high unsaturated fat diet and three out of four dietary comparisons that involved a high-MUFA diet. FG was improved when appropriate GWG advice was given alongside dietary advice com- pared with GWG advice only. FG reduction was observed in four of the six dietary compari- sons, two of which were derived from mixed comparisons (LGI/LGL diets vs GWG advice only and low-CHO & high-fat diet vs GWG advice only) and the other two were derived from indirect comparisons (high-fibre & LGI/LGL vs GWG advice only and Healthy Eating vs GWG advice only). The most effective diet to reduce FBG was the low-GI, high-fibre diet (SUCRA = 89.33%), followed by Healthy Eating (SUCRA = 88.17%), and then a low-CHO with a high-fat diet (SUCRA = 65.05%) (S1 Fig). 6 / 17 PLOS ONE \| https://doi.org/10.1371/journal.pone.0182095 August 3, 2017 Dietary patterns and glycemic outcomes during pregnancy Fig 2. Network geometry of trials that provided GWG advice in both arms and reported fasting glucose. Abbreviations: CHO, carbohydrate; LGI, low- glycemic index; LGL, low-glycemic load; GWG, gestational weight gain; MUFA, monounsaturated fatty acids. The colors of each node correspond to a different diet class: orange node represents diets that targeted macronutrient intake, blue nodes represent diets that targeted overall healthy eating, and green nodes represent diets that targeted GWG. The numbers above each line joining two comparators correspond to the number of trials that compare the treatments with the number of included participants expressed in brackets. Thickness of line represent the number of studies included for that dietary comparison. Distances between nodes are not meaningful. Fig 2. Network geometry of trials that provided GWG advice in both arms and reported fasting glucose. Abbreviations: CHO, carbohydrate; LGI, low- glycemic index; LGL, low-glycemic load; GWG, gestational weight gain; MUFA, monounsaturated fatty acids. Trials with GWG advice provided in one of the dietary arms A significant FI reduction was observed when comparing GWG advice to standard of care (MD = -25.00 pmol/L [95% CIs: -46.50, -3.50]) (S7 Fig). No significant FG (S5 Fig) or HbA1c (S6 Fig) effect was observed in any of the dietary comparisons. Trials with GWG advice provided in both dietary arms The colors of each node correspond to a different diet class: orange node represents diets that targeted macronutrient intake, blue nodes represent diets that targeted overall healthy eating, and green nodes represent diets that targeted GWG. The numbers above each line joining two comparators correspond to the number of trials that compare the treatments with the number of included participants expressed in brackets. Thickness of line represent the number of studies included for that dietary comparison. Distances between nodes are not meaningful. Other glycemic outcomes. A high-MUFA diet compared to GWG advice only increased HbA1c (MeD = 0.40% [95% CrIs: 0.12, 0.68]) (S2 Fig). No significant differences in HbA1c, FI, and HOMA-IR were seen between pairs of any other diets (S2–S4 Figs). Insulin therapy. In a post-hoc NMA analysis, based on an indirect comparison, the odds of progressing to insulin therapy to manage hyperglycemia during pregnancy was greater for a LGI diet than to a combined LGI and high-fibre diet (odds ratio [OR] = 5.92 [95% CrI: 1.20, 36.41]). No other diets were associated with the use of insulin therapy (data not shown). PLOS ONE \| https://doi.org/10.1371/journal.pone.0182095 August 3, 2017 Dietary patterns and glycemic outcomes during pregnancy Fig 3. Difference in median effect with 95% credible intervals between diets given in addition to GWG advice on fasting glucose in mmol/l. Abbreviations: CHO, carbohydrate; LGI, low-glycemic index; LGL, low-glycemic load; GWG, gestational weight gain; MUFA, monounsaturated fatty acids. The value in each cell expresses the median difference and its 95% credible intervals between the dietary pattern in the column and the dietary pattern in the row (e.g. the median difference of the high-unsaturated fat diet compared to LGI/LGL diet is 0.33 mmol/L (95% CrIs = 0.08, 0.57 mmol/L). Fig 3. Difference in median effect with 95% credible intervals between diets given in addition to GWG advice on fasting glucose in mmol/l. Abbreviations: CHO, carbohydrate; LGI, low-glycemic index; LGL, low-glycemic load; GWG, gestational weight gain; MUFA, monounsaturated fatty acids. The value in each cell expresses the median difference and its 95% credible intervals between the dietary pattern in the column and the dietary pattern in the row (e.g. the median difference of the high-unsaturated fat diet compared to LGI/LGL diet is 0.33 mmol/L (95% CrIs = 0.08, 0.57 mmol/L). Fig 3. Difference in median effect with 95% credible intervals between diets given in addition to GWG advice on fasting glucose in mmol/l. Abbreviations: CHO, carbohydrate; LGI, low-glycemic index; LGL, low-glycemic load; GWG, gestational weight gain; MUFA, monounsaturated fatty acids. The value in each cell expresses the median difference and its 95% credible intervals between the dietary pattern in the column and the dietary pattern in the row (e.g. the median difference of the high-unsaturated fat diet compared to LGI/LGL diet is 0.33 mmol/L (95% CrIs = 0.08, 0.57 mmol/L). glycemic index; LGL, low glycemic load; GWG, gestational weight gain; MUFA, monounsaturated fatty acids. difference and its 95% credible intervals between the dietary pattern in the column and the dietary pattern in the nsaturated fat diet compared to LGI/LGL diet is 0.33 mmol/L (95% CrIs = 0.08, 0.57 mmol/L). https://doi.org/10.1371/journal.pone.0182095.g003 https://doi.org/10.1371/journal.pone.0182095.g003 with standard of care (MeD = -0.47 mmol/L [95% CrIs: -0.73, -0.21]). Further, a non-signifi- cant FG-effect was observed in a low-GI diet compared to a high-fibre diet in a study that was with standard of care (MeD = -0.47 mmol/L [95% CrIs: -0.73, -0.21]). Further, a non-signifi- cant FG-effect was observed in a low-GI diet compared to a high-fibre diet in a study that was Fig 4. Trials with no GWG advice provided in both dietary arms Fasting glucose. Dietary interventions given with no GWG advice and had FG reported were identified in six trials (Fig 4) [23, 24, 27, 29, 30, 36]. In the absence of GWG advice, an improvement in FG was found in DASH-style diet com- pared to other diets (Fig 5). FG was reduced for the DASH-style diet in an indirect comparison with low-fat diet (MeD = -0.74 mmol/L [95% CrIs: -1.12, -0.36]) and in a direct comparison PLOS ONE \| https://doi.org/10.1371/journal.pone.0182095 August 3, 2017 7 / 17 The most effective diet to reduce FG in the absence of GWG advice was the DASH-style diet (SUCRA = 66.7%), followed by standard of care (SUCRA = 32.5%), and low-fat diet (SUCRA = 0.88%) (S4 Table). The most effective diet to reduce FG in the absence of GWG advice was the DASH-style diet (SUCRA = 66.7%), followed by standard of care (SUCRA = 32.5%), and low-fat diet (SUCRA = 0.88%) (S4 Table). Other glycemic outcomes. There were no significant differences on HbA1c (S8 Fig), FI (S9 Fig), and HOMA-IR (S10 Fig) between diets with the exception of an insulin-reducing effect (MD = -47.60 pmol/L [95% CIs: -77.34, -17.86]; p = 0.002) and a HOMA-IR-reducing effect (MD = -1.90 [95% CIs: -3.08, -0.72]; p = 0.002) in a DASH-style diet compared to stan- dard of care in the absence of GWG advice. Insulin therapy. None of the dietary comparisons showed a significant association to start insulin therapy to manage hyperglycemia during pregnancy in our post-hoc NMA analy- sis (data not shown). GRADE- quality of evidence assessment The quality of the evidence ranged from moderate to very low (S3–S8 Tables). Most compari- sons were downgraded because of serious concerns regarding indirectness and/or imprecision. Network plot of trials that reported fasting glucose and did not provide gestational weight gain advice in both dietary arms. Abbreviations: DASH, Dietary Approach to Stop Hypertension. The colors of each nodep correspond to a different diet class: orange node represents diets that targeted macronutrient composition, blue represents diets that targeted food consumption, and green on weight gain advice. The number above each line correspond to the number of trials that compared the two diets with the number of included participants expressed in brackets. https://doi org/10 1371/journal pone 0182095 g004 Fig 4. Network plot of trials that reported fasting glucose and did not provide gestational weight gain advice in both dietary arms. Abbreviations: DASH, Dietary Approach to Stop Hypertension. The colors of each nodep correspond to a different diet class: orange node represents diets that targeted macronutrient composition, blue represents diets that targeted food consumption, and green on weight gain advice. The number above each line correspond to the number of trials that compared the two diets with the number of included participants expressed in brackets. htt //d i /10 1371/j l 0182095 004 https://doi.org/10.1371/journal.pone.0182095.g004 PLOS ONE \| https://doi.org/10.1371/journal.pone.0182095 August 3, 2017 8 / 17 Dietary patterns and glycemic outcomes during pregnancy Fig 5. Effect of fasting glucose between diets in trials that did not provide gestational weight gain advice in both dietary arms. Abbreviations: DASH, Dietary Approach to Stop Hypertension. Fasting glucose is expressed in mmol/L. The value in each cell expresses the median difference (MeD) in fasting glucose with the 95% credible intervals (CrIs) in brackets between the diet in the column and the diet in the row (e.g. the MeD in fasting glucose between DASH-style diet compared to low-fat diet is -0.74 mmol/L (95% CrIs: -1.12, -0.36). Fig 5. Effect of fasting glucose between diets in trials that did not provide gestational weight gain advice in both dietary arms. Abbreviations: DASH, Dietary Approach to Stop Hypertension. Fasting glucose is expressed in mmol/L. The value in each cell expresses the median difference (MeD) in fasting glucose with the 95% credible intervals (CrIs) in brackets between the diet in the column and the diet in the row (e.g. the MeD in fasting glucose between DASH-style diet compared to low-fat diet is -0.74 mmol/L (95% CrIs: -1.12, -0.36). Fig 5. Effect of fasting glucose between diets in trials that did not provide gestational weight gain advice in both dietary arms. Abbreviations: DASH, Dietary Approach to Stop Hypertension. Fasting glucose is expressed in mmol/L. The value in each cell expresses the median difference (MeD) in fasting glucose with the 95% credible intervals (CrIs) in brackets between the diet in the column and the diet in the row (e.g. the MeD in fasting glucose between DASH-style diet compared to low-fat diet is -0.74 mmol/L (95% CrIs: -1.12, -0.36). Fig 5. Effect of fasting glucose between diets in trials that did not provide gestational weight gain advice in both dietary arms. Abbreviations: DASH, Dietary Approach to Stop Hypertension. Fasting glucose is expressed in mmol/L. The value in each cell expresses the median difference (MeD) in fasting glucose with the 95% credible intervals (CrIs) in brackets between the diet in the column and the diet in the row (e.g. the MeD in fasting glucose between DASH-style diet compared to low-fat diet is -0.74 mmol/L (95% CrIs: -1.12, -0.36). https://doi.org/10.1371/journal.pone.0182095.g005 not analyzed as part of the NMA due to a lack of a common comparator (MD = -0.10 mmol/L [95% CIs: -0.38, 0.18]; p = 0.48) [27]. Discussion Although most diets that were given in addition to GWG advice demonstrated a FG reduction in our NMA, these same findings were not found in trials that had been specifically designed to assess if diets in addition to GWG advice would affect FG. Instead, a null FG-effect had been reported by these trials. This may, however, be due to the small number trials of such trials identified and included in our analysis. High unsaturated fat intake has been found to be cardio-protective but there is uncertainty concerning its relationship with diabetes risk. Although meta-analyses have shown non-signif- icant findings, a trend for increased type 2 diabetes mellitus (T2DM) risk have been noted for polyunsaturated fatty acid intakes (PUFAs), omega-3’s, and foods that are a source of these fatty acids such as fish and other seafood [41, 42]. High fat intake is linked to increased hepatic glucose production by reducing the ability of insulin to suppress endogenous glucose produc- tion [43]. Trials that were included in our analysis showed a positive correlation between unsaturated fat or MUFA intakes with PUFA intakes [31, 34]. Consistent with the above find- ings between PUFAs and T2DM, our analysis found that FG increased in diets that increased unsaturated or MUFA intakes. Insulin therapy is usually initiated after two weeks if women cannot manage their GDM using diet therapy alone [2, 3]. No difference in the use of insulin therapy was found between diets in our analysis except for LGI diets compared to LGI with high-fibre diets. One interpre- tation of this finding is that the examined interventions (diets, GWG advice, and standard of care) were equally effective in preventing the use of insulin. We cannot, however, rule out the more likely possibility that trials achieved suboptimal dietary compliance (as reflected in our GRADE assessment) or that the dietary contrasts were not large enough to detect effects on insulin therapy use. Several limitations were noted in the present study. First, our network meta-analysis included only RCTs which may have limited the number of available dietary comparisons. We had, however, decided not to include non-randomized studies because of concerns that these types of studies are more likely to introduce bias into the effect estimates because of confound- ing arising from the lack of randomization. Discussion We have systematically reviewed and conducted a network meta-analysis of randomized trials to assess the relative effectiveness of various diets on glycemic outcomes in women during pregnancy. Alongside with gestational weight gain advice, most diets, with the exception of a high unsaturated or a high monounsaturated fatty acid diet, demonstrated a fasting glucose improvement compared with gestational weight gain advice only. When gestational weight gain advice was not given, the DASH-style diet appeared optimal on fasting glucose. Similar trends were observed in the other glycemic outcomes. The benefits of diets given in addition to GWG advice or standard of care on FG appeared modest, but we believe that these have important clinical relevance. Reductions in FG of 0.1 mmol/L in the Metformin in Gestational Diabetes Trial [38] and 0.3 mmol/L in a large RCT PLOS ONE \| https://doi.org/10.1371/journal.pone.0182095 August 3, 2017 9 / 17 Dietary patterns and glycemic outcomes during pregnancy were observed when insulin was compared to anti-hyperglycemic medications in pregnant women [39]. Similar magnitudes of FG reductions were observed in our analysis, ranging from -0.27 to -0.77 mmol/L in trials with GWG advice and -0.47 to -0.74 mmol/L in trials with no GWG advice. This is particularly important during pregnancy as most women prefer die- tary approaches to manage FG levels than the use of insulin therapy [4]. Furthermore, our findings build on existing dietary approaches for management of GDM which mostly focus on carbohydrate-counting or limiting caloric intake to manage GDM risk [1–3]. All our dietary comparisons that demonstrated a FG improvement emphasized on the consumption of high- quality (e.g., unrefined, minimally processed foods such as vegetables and fruits, whole grains), healthy foods, and minimizing low-quality foods (e.g., highly processed snack foods, refined grains, fried foods, and high-GI foods). Providing high-quality diets may be more effective to manage FG than GWG advice or standard of care only. Both the quantity and quality of diet have been emphasized as equally important in the management of cardiometabolic risk [40]. Pregnancy is a time of heightened sensitivity and attention to food intake in most women [4], so the ability of healthcare providers to provide accurate and evidence-based advice increases the relevance of our findings. PLOS ONE \| https://doi.org/10.1371/journal.pone.0182095 August 3, 2017 Conclusions Alongside with gestational weight gain advice, most diets, with the exception of a high unsatu- rated or a monounsaturated fatty acid diet, demonstrated a fasting glucose improvement com- pared with gestational weight gain advice only. When gestational weight gain advice was not given, the DASH-style diet appeared optimal on fasting glucose. However, the number of trials is small and most were underpowered to detect differences in FG. To clarify the role of diets in glycemic management during pregnancy, data from larger, high-quality, and well-powered feeding trials of dietary approaches and high-quality prospective cohort studies are required. Nonetheless, diets, with the exception of ones that modify fat intake, may be useful as part of a strategy to improve FG. Discussion A specific barrier to including both randomized and non-randomized studies in a network meta-analysis is that this practice would compro- mise the validity of our network by possibly violating two key assumptions: transitivity (the distribution of patient and study characteristics that are modifiers of treatment effect be suffi- ciently similar across studies) and as such, could possibly affect the consistency of the evidence (agreement of direct and indirect evidence for a given pair of treatments). Second, our PLOS ONE \| https://doi.org/10.1371/journal.pone.0182095 August 3, 2017 10 / 17 Dietary patterns and glycemic outcomes during pregnancy certainty in the pooled effect estimates for each dietary comparison was moderate to very low. For our FG analysis, the quality of evidence was downgraded mostly due to poor (indirect) network connectivity between diets, small sample sizes, or both. For other glycemic outcomes, a lack of similar dietary comparisons precluded us from conducting a useful NMA. Further- more, most dietary comparisons were under-powered to detect a difference in FG, HbA1c, FI, or HOMA-IR as we had found in our post-hoc analysis (data not shown). Third, most of our findings were derived from indirect comparisons rather than direct comparisons. Although we concluded that the assumption of transitivity was reasonably met for indirect comparisons within our study, we were not able to use the less-reported BMI to guide our assessments and as always, the case with indirect comparisons, minor, immeasurable effect-modifying charac- teristics could bias these estimates. Fourth, the generalizability of our results is limited. Most of the included trials were predominantly in young women in their second trimester who were already diagnosed with GDM. Therefore, it is unclear if the studied diets can prevent GDM per se. Certainly, however, based on our results, some diets appeared to be more effective in managing glycemic outcomes than others. Notwithstanding these limitations, many of the die- tary comparisons in our analyses were designed to assess two dietary interventions that may benefit glycemic control; as such comparisons to a usual diet (e.g. typical North American/ European non-therapeutic diet) were few and in this regard, a maintenance in glycemic con- trol after intervention may be noteworthy. PLOS ONE \| https://doi.org/10.1371/journal.pone.0182095 August 3, 2017 Supporting information Pair-wise meta-analysis of diets and fasting insulin in trials where GWG advice was provided in one of the dietary arms. Abbreviations: CHO, carbohydrate; CI, confidence interval; FI, fasting insulin; GWG, gestational weight gain; MD, mean differences; n, sample size. (DOCX) S8 Fig. Pair-wise meta-analysis of diets and HbA1c in trials with no GWG advice provided. Abbreviations: CI, confidence interval; DASH, Dietary Approach to Stop Hypertension; HbA1c, hemoglobin A1c; LGI, low glycemic index; MD, mean differences; n, sample size. Diet # 1 reflects the diet that is first mentioned before “vs” and diet #2 reflects the diet that comes after “vs”. (DOCX) S9 Fig. Pair-wise meta-analysis of diets and fasting insulin in trials with no GWG advice provided. Abbreviations: CI, confidence interval; DASH, Dietary Approach to Stop Hyperten- sion; FI, fasting insulin; LGI, low glycemic index; MD, mean differences; n, sample size. Diet # 1 reflects the diet that is first mentioned before “vs” and diet #2 reflects the diet that comes after “vs”. (DOCX) S10 Fig. Pair-wise meta-analysis of diets and HOMA-IR in trials with no GWG advice pro- vided. Abbreviations: CI, confidence interval; DASH, Dietary Approach to Stop Hypertension; FI, fasting insulin; LGI, low glycemic index; MD, mean differences; n, sample size. Diet # 1 reflects the diet that is first mentioned before “vs” and diet #2 reflects the diet that comes after “vs”. (DOCX) S1 Table. Search strategy used to identify eligible studies. Search was conducted in Novem- ber 2014 and updated in April 2015, February 2016, and April 2017. (DOCX) S3 Fig. Pair-wise meta-analyses of diets and fasting insulin in trials where GWG advice was provided in both dietary arms. Abbreviations: CHO, carbohydrate; CI, confidence inter- val; FI, fasting insulin; GWG, gestational weight gain; MD, mean differences; n, sample size. Diet # 1 reflects the diet that is first mentioned before “vs” and diet #2 reflects the diet that comes after “vs”. (DOCX) S4 Fig. Pair-wise meta-analyses of diets and HOMA-IR in trials where GWG advice was provided in both dietary arms. Abbreviations: CHO, carbohydrate; CI, confidence interval; GWG, gestational weight gain; HOMA-IR, homeostatic model assessment for insulin resis- tance; MD, mean differences; n, sample size. (DOCX) S5 Fig. Pair-wise meta-analysis of diets and fasting glucose in trials where GWG advice was provided in one of the dietary arms. Supporting information S1 Fig. Rank of each diet that were given in addition to GWG advice as being the most effective in reducing fasting glucose. Abbreviations: CHO, carbohydrate; GWG, gestational weight gain; LGI, low glycemic index; LGL, low glycemic load; MUFA, monounsaturated fatty acids; SUCRA, surface under the cumulative ranking. The rankogram displays the probability of each diet achieving a particular rank and the SUCRA value reflects the probability of a given diet as being the most effective in reducing fasting glucose among all the diets being compared. The closer SUCRA is to 100, the more certain we are that it is the best overall and the closer it is to zero, the more certain we are that it is worst. (DOCX) S2 Fig. Pair-wise meta-analyses of diets and HbA1c in trials where GWG advice was pro- vided in both dietary arms. Abbreviations: CHO, carbohydrate; CI, confidence interval; GWG, gestational weight gain; HbA1c, hemoglobin A1c; LGI, low glycemic index; MD, mean differences; MUFA, monounsaturated fatty acids; n, sample size. Diet # 1 reflects the diet that is first mentioned before “vs” and diet #2 reflects the diet that comes after “vs”. (DOCX) PLOS ONE \| https://doi.org/10.1371/journal.pone.0182095 August 3, 2017 11 / 17 Dietary patterns and glycemic outcomes during pregnancy S3 Fig. Pair-wise meta-analyses of diets and fasting insulin in trials where GWG advice was provided in both dietary arms. Abbreviations: CHO, carbohydrate; CI, confidence inter- val; FI, fasting insulin; GWG, gestational weight gain; MD, mean differences; n, sample size. Diet # 1 reflects the diet that is first mentioned before “vs” and diet #2 reflects the diet that comes after “vs”. (DOCX) S4 Fig. Pair-wise meta-analyses of diets and HOMA-IR in trials where GWG advice was provided in both dietary arms. Abbreviations: CHO, carbohydrate; CI, confidence interval; GWG, gestational weight gain; HOMA-IR, homeostatic model assessment for insulin resis- tance; MD, mean differences; n, sample size. (DOCX) S5 Fig. Pair-wise meta-analysis of diets and fasting glucose in trials where GWG advice was provided in one of the dietary arms. Abbreviations: CHO, carbohydrate; CI, confidence interval; FG, fasting glucose; GWG, gestational weight gain; MD, mean differences; n, sample size. (DOCX) S6 Fig. Pair-wise meta-analysis of diets and HbA1c in trials where GWG advice was provided in one of the dietary arms. Abbreviations: CHO, carbohydrate; CI, confidence interval; GWG, gestational weight gain; HbA1c, hemoglobin A1c; MD, mean differences; n, sample size. (DOCX) S7 Fig. Supporting information Abbreviations: "-," not reported; "~," calculated; BMI, body mass index; CHO, carbohydrate; DASH, Dietary Approach to Stop Hypertension; FG, fasting glucose; FI, fasting insulin; GDM, gestational diabetes mellitus; GWG, gestational weight gain; HbA1c, hemoglobin A1c; HOMA-IR, homeostatic model assessment-insulin resistance; IGT, Impaired Glucose Tolerance; LGI, low glycemic index; LGL, low glycemic load; MUFA, monounsaturated fatty acids; Ob, Obese; OW, overweight; T1DM, type 1 Diabe- tes; T2DM, type 2 diabetes; All data expressed as mean ± SD unless otherwise noted. †Dietary definitions: DASH-style intake, diets rich in fruits, vegetables, whole grains, low-fat dairy products but low in saturated fats, cholesterol, refined grains, sweets, and sodium; GWG advice only, advice given to help women achieve optimal GWG; healthy eating, advice fol- lowed general healthy eating guidelines (e.g. Canada’s Food Guide); high-fat, >30% of energy came from fat; high-fibre, >30 g/d of dietary fibre; high unsaturated fat, increased in unsatu- rated fat intake compared to no intervention; low-CHO, <45% energy came from CHO; low- fat, <20% of energy from fat; LGI, low glycemic index; LGL, low glycemic load; standard of care, no dietary advice given or macronutrient intake was 45–56% energy from CHO: 10–35% energy from protein: 20–35% energy from fat. ‡Body weight was reported when BMI is unavailable. δThe number of active smokers during pregnancy. Counts were reported with the percentage of the total participants as smokers reported in brackets. ¶The gestational week at which the participants started the dietary intervention. §Baseline characteristics were based on the number of randomised participants for Grant et al n = 43, Laitenin et al n = 171, Rae et al n = 124, and Valentini et al n = 781. All foods were provided (DOCX) S3 Table. Quality of the evidence in the direct dietary comparisons in the fasting glucose analysis. Abbreviations: CHO, carbohydrate; CrI, credible intervals; DASH, Dietary Approach to Stop Hypertension; FG, fasting glucose; GWG, gestational weight gain; LGI, low glycemic index; LGL, low glycemic load; MeD, median difference; MUFA, monounsaturated fatty acids; n, sample size. aDietary comparison was downgraded because the attrition rate of the included trial was considered to have high risk of bias. bInconsistency could not be assessed because only one trial was included. Supporting information Abbreviations: CHO, carbohydrate; CI, confidence interval; FG, fasting glucose; GWG, gestational weight gain; MD, mean differences; n, sample size. (DOCX) S6 Fig. Pair-wise meta-analysis of diets and HbA1c in trials where GWG advice was provided in one of the dietary arms. Abbreviations: CHO, carbohydrate; CI, confidence interval; GWG, gestational weight gain; HbA1c, hemoglobin A1c; MD, mean differences; n, sample size. (DOCX) S7 Fig. Pair-wise meta-analysis of diets and fasting insulin in trials where GWG advice was provided in one of the dietary arms. Abbreviations: CHO, carbohydrate; CI, confidence interval; FI, fasting insulin; GWG, gestational weight gain; MD, mean differences; n, sample size. (DOCX) S8 Fig. Pair-wise meta-analysis of diets and HbA1c in trials with no GWG advice provided. Abbreviations: CI, confidence interval; DASH, Dietary Approach to Stop Hypertension; HbA1c, hemoglobin A1c; LGI, low glycemic index; MD, mean differences; n, sample size. Diet # 1 reflects the diet that is first mentioned before “vs” and diet #2 reflects the diet that comes after “vs”. (DOCX) S9 Fig. Pair-wise meta-analysis of diets and fasting insulin in trials with no GWG advice provided. Abbreviations: CI, confidence interval; DASH, Dietary Approach to Stop Hyperten- sion; FI, fasting insulin; LGI, low glycemic index; MD, mean differences; n, sample size. Diet # 1 reflects the diet that is first mentioned before “vs” and diet #2 reflects the diet that comes after “vs”. (DOCX) S10 Fig. Pair-wise meta-analysis of diets and HOMA-IR in trials with no GWG advice pro- vided. Abbreviations: CI, confidence interval; DASH, Dietary Approach to Stop Hypertension; FI, fasting insulin; LGI, low glycemic index; MD, mean differences; n, sample size. Diet # 1 reflects the diet that is first mentioned before “vs” and diet #2 reflects the diet that comes after “vs”. (DOCX) S1 Table. Search strategy used to identify eligible studies. Search was conducted in Novem- ber 2014 and updated in April 2015, February 2016, and April 2017. (DOCX) 12 / 17 PLOS ONE \| https://doi.org/10.1371/journal.pone.0182095 August 3, 2017 Dietary patterns and glycemic outcomes during pregnancy S2 Table. Table of study characteristics. PLOS ONE \| https://doi.org/10.1371/journal.pone.0182095 August 3, 2017 Supporting information cAlthough there was evidence of moderate inter-study heterogene- ity (I2 = 44.6%), this was not a cause of concern because the credible intervals of the two included trials substantially overlapped one another and their point estimate laid on the same side of the line of no effect. dNo evidence of inter-study heterogeneity (I2 = 0%). eThe included trial(s) failed to achieve its dietary goals and therefore, the contrast of the dietary interventions may be too small to affect FG. fOptimal information size (OIS) was not met. gThe effect esti- mate crosses the minimally important difference (MID) of ±0.5 mmol/L. hPublication bias could not be assessed because there were <10 included trials. (DOCX) S4 Table. Quality of the evidence in the indirect dietary comparisons in the fasting glucose analysis. Abbreviations: CHO, carbohydrate; CrI, credible intervals; DASH, Dietary Approach to Stop Hypertension; FG, fasting glucose; GWG, gestational weight gain; LGI, low glycemic index; LGL, low glycemic load; MeD, median difference; MUFA, monounsaturated fatty acids. aThere were important differences in GDM status and at the trimester in which the dietary interventions began between the two first order links. bThere were important differences in GDM status, ethnicity, and at the trimester in which the dietary interventions began between the two first order links. cQuality of comparisons was assumed as very low because the link order is 2. dThere were important differences in ethnicity and at the trimester in which the dietary interventions began between the two first order links. eThere were important PLOS ONE \| https://doi.org/10.1371/journal.pone.0182095 August 3, 2017 13 / 17 Dietary patterns and glycemic outcomes during pregnancy differences at the trimester in which the dietary interventions began between the two first order links. fThere were important differences in ethnicity between the two first order links. gThere were important differences in pre-pregnancy BMI and at the trimester in which the dietary interventions began between the two first order links. (DOCX) S5 Table. Quality of evidence in the mixed dietary comparisons in the fasting glucose anal- ysis. Abbreviations: CHO, carbohydrate; CIs, confidence intervals; DASH, Dietary Approach to Stop Hypertension; GWG, gestational weight gain; HbA1c, hemoglobin A1c; LGI, low gly- cemic index; LGL, low glycemic load; MD, mean difference; MUFA, monounsaturated fatty acids; n, sample size. Abbreviations: CHO, carbohydrate; CrI, credible intervals; FG, fasting glucose; GWG, gestational weight gain; LGI, low glycemic index; LGL, low glycemic load; MeD, median difference. (DOCX) S6 Table. Supporting information Quality of the evidence in the direct dietary comparisons in the HbA1c analysis. aInconsistency could not be assessed because only one trial was included. bThere was evidence of high inter-study heterogeneity (I2 = 88%). Further the two included trials showed different effects, one showed protection and the other showed null. cThe included trial(s) failed to achieve its dietary goals and therefore, the contrast of the dietary interventions may be too small to affect HbA1c. dThe effect estimate crosses the minimally important difference (MID) of ±0.3%. eOptimal information size (OIS) was not met. fPublication bias could not be assessed because there were <10 included trials. gNo evidence of inter-study heterogeneity (I2 = 0%). (DOCX) S7 Table. Quality of the evidence in the direct dietary comparisons in the fasting insulin analysis. Abbreviations: CHO, carbohydrate; CIs, confidence intervals; DASH, Dietary Approach to Stop Hypertension; FI, fasting insulin; GWG, gestational weight gain; LGI, low glycemic index; MD, mean difference; MUFA, monounsaturated fatty acids; n, sample size. aDietary comparison was downgraded because the attrition rate of the included trial was con- sidered to have high risk of bias. bInconsistency could not be assessed because only one trial was included. cNo evidence of inter-study heterogeneity (I2 = 0%). dThe included trial(s) failed to achieve its dietary goals and therefore, the contrast of the dietary interventions may be too small to affect FI. eThe effect estimate crosses the minimally important difference (MID) of ±0.5 pmol/L. fOptimal information size (OIS) was not met. gPublication bias could not be assessed because there were <10 included trials. (DOCX) S8 Table. Quality of the evidence in the direct dietary comparisons in the HOMA-IR analysis. Abbreviations: CHO, carbohydrate; CIs, confidence intervals; DASH, Dietary Approach to Stop Hypertension; GWG, gestational weight gain; HOMA-IR, homeostatic model assessment for insulin resistance; LGI, low glycemic index; MD, mean difference; n, sample size. aInconsistency could not be assessed because only one trial was included. bThe effect estimate crosses the mini- mally important difference (MID) of ±1 unit. cOptimal information size (OIS) was not met. dPu- blication bias could not be assessed because there were <10 included trials. eNo evidence of inter- study heterogeneity (I2 = 0%). fThe included trial failed to achieve its dietary goals and therefore, the contrast of the dietary interventions may be too small to affect HOMA-IR. (DOCX) S1 File. PRISMA checklist. (DOCX) S1 File. PRISMA checklist. (DOCX) S1 File. PRISMA checklist. S1 File. PRISMA checklist. (DOCX) Writing – original draft: Vanessa Ha. Writing – review & editing: Vanessa Ha, Ashley J. Bonner, Jaynendr K. Jadoo, Joseph Beyene, Sonia S. Anand, Russell J. de Souza. Supporting information (DOCX) 14 / 17 PLOS ONE \| https://doi.org/10.1371/journal.pone.0182095 August 3, 2017 Dietary patterns and glycemic outcomes during pregnancy Conceptualization: Sonia S. Anand. Data curation: Vanessa Ha, Ashley J. Bonner, Jaynendr K. Jadoo. Investigation: Vanessa Ha, Ashley J. Bonner, Jaynendr K. Jadoo. Methodology: Vanessa Ha, Ashley J. Bonner, Joseph Beyene, Russell J. de Souza. Project administration: Vanessa Ha, Joseph Beyene, Sonia S. Anand, Russell J. de Souza. Resources: Ashley J. Bonner, Joseph Beyene, Sonia S. Anand, Russell J. de Souza. Software: Vanessa Ha, Ashley J. Bonner. Supervision: Joseph Beyene, Sonia S. Anand, Russell J. de Souza. Validation: Vanessa Ha, Ashley J. Bonner, Jaynendr K. Jadoo, Joseph Beyene, Sonia S. Anand, Russell J. de Souza. Visualization: Vanessa Ha. Visualization: Vanessa Ha. Visualization: Vanessa Ha. Writing – original draft: Vanessa Ha. PLOS ONE \| https://doi.org/10.1371/journal.pone.0182095 August 3, 2017 References 1. National Institute for Health and Care Excellence. Diabetes in pregnancy: management from precon- ception to the postnatal period. 2015. Accessed via: nice.org.uk/guidance/ng3. 2. (12) Management of diabetes in pregnancy. Diabetes care. 2015; 38 Suppl:S77–9. Epub 2014/12/30. https://doi.org/10.2337/dc15-S015 PMID: 25537713. 3. Thompson D, Berger H, Feig D, Gagnon R, Kader T, Keely E, et al. Diabetes and pregnancy. Canadian journal of diabetes. 2013; 37 Suppl 1:S168–83. Epub 2014/05/16. https://doi.org/10.1016/j.jcjd.2013. 01.044 PMID: 24070943. 4. Carolan M. Women’s experiences of gestational diabetes self-management: a qualitative study. Mid- wifery. 2013; 29(6):637–45. https://doi.org/10.1016/j.midw.2012.05.013 PMID: 22877761. 5. Hernandez TL, Van Pelt RE, Anderson MA, Reece MS, Reynolds RM, de la Houssaye BA, et al. Women With Gestational Diabetes Mellitus Randomized to a Higher-Complex Carbohydrate/Low-Fat Diet Manifest Lower Adipose Tissue Insulin Resistance, Inflammation, Glucose, and Free Fatty Acids: A Pilot Study. Diabetes care. 2016; 39(1):39–42. Epub 2015/08/01. https://doi.org/10.2337/dc15-0515 PMID: 26223240; PubMed Central PMCID: PMC4686845. 6. Walsh JM, McGowan CA, Mahony R, Foley ME, McAuliffe FM. Low glycaemic index diet in pregnancy to prevent macrosomia (ROLO study): randomised control trial. BMJ. 2012; 345:e5605. Epub 2012/09/ 01. https://doi.org/10.1136/bmj.e5605 PMID: 22936795; PubMed Central PMCID: PMC3431285. 7. Moses RG, Barker M, Winter M, Petocz P, Brand-Miller JC. Can a low-glycemic index diet reduce the need for insulin in gestational diabetes mellitus? A randomized trial. Diabetes care. 2009; 32(6):996– 1000. 10.2337/dc09-0007. PubMed Central PMCID: PMCPMC2681032. https://doi.org/10.2337/dc09- 0007 PMID: 19279301; 8. Hu FB, Manson JE, Stampfer MJ, Colditz G, Liu S, Solomon CG, et al. Diet, lifestyle, and the risk of type 2 diabetes mellitus in women. The New England journal of medicine. 2001; 345(11):790–7. https://doi. org/10.1056/NEJMoa010492 PMID: 11556298. 9. Tieu J, Shepherd E, Middleton P, Crowther CA. Dietary advice interventions in pregnancy for preventing gestational diabetes mellitus. Cochrane Database Syst Rev. 2017; 1:CD006674. https://doi.org/10. 1002/14651858.CD006674.pub3 PMID: 28046205. PLOS ONE \| https://doi.org/10.1371/journal.pone.0182095 August 3, 2017 15 / 17 Dietary patterns and glycemic outcomes during pregnancy 10. Han S, Middleton P, Shepherd E, Van Ryswyk E, Crowther CA. Different types of dietary advice for women with gestational diabetes mellitus. Cochrane Database Syst Rev. 2017; 2:CD009275. https:// doi.org/10.1002/14651858.CD009275.pub3 PMID: 28236296. 11. National Institute for Health and Care Excellence. Maternal and child nutrition. 2014. Accessed on June 7 2017. Available via: https://www.nice.org.uk/guidance/ph11/resources/maternal-and-child-nutrition- pdf-1996171502533. 12. Higgins JPT, Green S (editors). Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 [updated March 2011]. The Cochrane Collaboration, 2011. Available from www.cochrane- handbook.org. 13. References Hutton B, Salanti G, Caldwell DM, Chaimani A, Schmid CH, Cameron C, et al. The PRISMA extension statement for reporting of systematic reviews incorporating network meta-analyses of health care inter- ventions: checklist and explanations. Annals of internal medicine. 2015; 162(11):777–84. Epub 2015/ 06/02. https://doi.org/10.7326/M14-2385 PMID: 26030634. 14. Puhan MA, Schunemann HJ, Murad MH, Li T, Brignardello-Petersen R, Singh JA, et al. A GRADE Working Group approach for rating the quality of treatment effect estimates from network meta-analy- sis. BMJ. 2014; 349:g5630. Epub 2014/09/26. https://doi.org/10.1136/bmj.g5630 PMID: 25252733. 15. Benton D. Sucrose and behavioral problems. Critical reviews in food science and nutrition. 2008; 48 (5):385–401. https://doi.org/10.1080/10408390701407316 PMID: 18464029 16. Dias S, Welton NJ, Caldwell DM, Ades AE. Checking consistency in mixed treatment comparison meta- analysis. Statistics in medicine. 2010; 29(7–8):932–44. Epub 2010/03/10. https://doi.org/10.1002/sim. 3767 PMID: 20213715. 17. Salanti G, Ades AE, Ioannidis JP. Graphical methods and numerical summaries for presenting results from multiple-treatment meta-analysis: an overview and tutorial. J Clin Epidemiol. 2011; 64(2):163–71. https://doi.org/10.1016/j.jclinepi.2010.03.016 PMID: 20688472. 18. Trinquart L, Attiche N, Bafeta A, Porcher R, Ravaud P. Uncertainty in Treatment Rankings: Reanalysis of Network Meta-analyses of Randomized Trials. Annals of internal medicine. 2016; 164(10):666–73. https://doi.org/10.7326/M15-2521 PMID: 27089537. 19. Bonomo M, Corica D, Mion E, Goncalves D, Motta G, Merati R, et al. Evaluating the therapeutic approach in pregnancies complicated by borderline glucose intolerance: a randomized clinical trial. Dia- betic medicine: a journal of the British Diabetic Association. 2005; 22(11):1536–41. Epub 2005/10/26. https://doi.org/10.1111/j.1464-5491.2005.01690.x PMID: 16241919. 20. Rae A, Bond D, Evans S, North F, Roberman B, Walters B. A randomised controlled trial of dietary energy restriction in the management of obese women with gestational diabetes. The Australian & New Zealand journal of obstetrics & gynaecology. 2000; 40(4):416–22. Epub 2001/02/24. PMID: 11194427. 21. Grant SM, Wolever TM, O’Connor DL, Nisenbaum R, Josse RG. Effect of a low glycaemic index diet on blood glucose in women with gestational hyperglycaemia. Diabetes research and clinical practice. 2011; 91(1):15–22. Epub 2010/11/26. https://doi.org/10.1016/j.diabres.2010.09.002 PMID: 21094553. 22. Perichart-Perera O, Balas-Nakash M, Rodriguez-Cano A, Legorreta-Legorreta J, Parra-Covarrubias A, Vadillo-Ortega F. Low Glycemic Index Carbohydrates versus All Types of Carbohydrates for Treating Diabetes in Pregnancy: A Randomized Clinical Trial to Evaluate the Effect of Glycemic Control. Interna- tional journal of endocrinology. 2012; 2012:296017. Epub 2012/12/20. https://doi.org/10.1155/2012/ 296017 PMID: 23251152; PubMed Central PMCID: PMC3517846. 23. Asemi Z, Tabassi Z, Samimi M, Fahiminejad T, Esmaillzadeh A. PLOS ONE \| https://doi.org/10.1371/journal.pone.0182095 August 3, 2017 References https://doi.org/10.6133/apjcn.2015.24.1.19 PMID: 25740743. 32. Afaghi A, Ghanei L, Ziaee A. Effect of low glycemic load diet with and without wheat bran on glucose control in gestational diabetes mellitus: A randomized trial. Indian J Endocrinol Metab. 2013; 17(4):689– 92. https://doi.org/10.4103/2230-8210.113762 PMID: 23961487; PubMed Central PMCID: PMCPMC3743371. 33. Laitinen K, Poussa T, Isolauri E, Nutrition AMI, Intestinal Microbiota G. Probiotics and dietary counsel- ling contribute to glucose regulation during and after pregnancy: a randomised controlled trial. The Brit- ish journal of nutrition. 2009; 101(11):1679–87. https://doi.org/10.1017/S0007114508111461 PMID: 19017418. 34. Lauszus FF, Rasmussen OW, Henriksen JE, Klebe JG, Jensen L, Lauszus KS, et al. Effect of a high monounsaturated fatty acid diet on blood pressure and glucose metabolism in women with gestational diabetes mellitus. European journal of clinical nutrition. 2001; 55(6):436–43. https://doi.org/10.1038/sj. ejcn.1601193 PMID: 11423920. 35. Ma WJ, Huang ZH, Huang BX, Qi BH, Zhang YJ, Xiao BX, et al. Intensive low-glycaemic-load dietary intervention for the management of glycaemia and serum lipids among women with gestational diabe- tes: a randomized control trial. Public Health Nutr. 2015; 18(8):1506–13. https://doi.org/10.1017/ S1368980014001992 PMID: 25222105. 36. Yao J, Cong, Zhu B, and Wang T. Effect of dietary approaches to stop hypertension diet plan on preg- nancy outcome patients with gestational diabetes mellitus. Bangladesh Jounal of Pharmacology 2015; 10(4): 732–8. 37. Reece EA, Hagay Z, Gay LF, O’Connor, DeGennaro N, Homko CJ, Wizaitzer A. A Randomized Clinical Trial of a Fiber-Enriched Diabetic Diet vs the Standard American Diabetes Association-Recommended Diet in the Management of Diabetes Mellitus in Pregnancy. Journal of Maternal-Fetal Investigation 1995; 5: 8–12. 38. Rowan JA, Hague WM, Gao W, Battin MR, Moore MP, Mi GTI. Metformin versus insulin for the treat- ment of gestational diabetes. The New England journal of medicine. 2008; 358(19):2003–15. https:// doi.org/10.1056/NEJMoa0707193 PMID: 18463376. 39. Langer O, Conway DL, Berkus MD, Xenakis EM, Gonzales O. A comparison of glyburide and insulin in women with gestational diabetes mellitus. The New England journal of medicine. 2000; 343(16):1134– 8. https://doi.org/10.1056/NEJM200010193431601 PMID: 11036118. 40. Mozaffarian D, Hao T, Rimm EB, Willett WC, Hu FB. Changes in diet and lifestyle and long-term weight gain in women and men. The New England journal of medicine. 2011; 364(25):2392–404. Epub 2011/ 06/24. https://doi.org/10.1056/NEJMoa1014296 PMID: 21696306; PubMed Central PMCID: PMC3151731. 41. Ley SH, Hamdy O, Mohan V, Hu FB. Prevention and management of type 2 diabetes: dietary compo- nents and nutritional strategies. Lancet. 2014; 383(9933):1999–2007. https://doi.org/10.1016/S0140- 6736(14)60613-9 PMID: 24910231 42. References Favourable effects of the Dietary Approaches to Stop Hypertension diet on glucose tolerance and lipid profiles in gestational diabetes: a randomised clinical trial. The British journal of nutrition. 2013; 109(11):2024–30. Epub 2012/11/15. https://doi.org/10.1017/S0007114512004242 PMID: 23148885. 24. Asemi Z, Samimi M, Tabassi Z, Sabihi SS, Esmaillzadeh A. A randomized controlled clinical trial investi- gating the effect of DASH diet on insulin resistance, inflammation, and oxidative stress in gestational diabetes. Nutrition. 2013; 29(4):619–24. Epub 2013/03/08. https://doi.org/10.1016/j.nut.2012.11.020 PMID: 23466048. 25. Ney D, Hollingsworth DR, Cousins L. Decreased insulin requirement and improved control of diabetes in pregnant women given a high-carbohydrate, high-fiber, low-fat diet. Diabetes care. 1982; 5(5):529– 33. Epub 1982/09/01. PMID: 6329613. 26. Wolff S, Legarth J, Vangsgaard K, Toubro S, Astrup A. A randomized trial of the effects of dietary counseling on gestational weight gain and glucose metabolism in obese pregnant women. Int J Obes (Lond). 2008; 32(3):495–501. Epub 2008/01/30. https://doi.org/10.1038/sj.ijo.0803710 PMID: 18227847. 27. Louie JC, Markovic TP, Perera N, Foote D, Petocz P, Ross GP, et al. A randomized controlled trial investigating the effects of a low-glycemic index diet on pregnancy outcomes in gestational diabetes PLOS ONE \| https://doi.org/10.1371/journal.pone.0182095 August 3, 2017 16 / 17 Dietary patterns and glycemic outcomes during pregnancy mellitus. Diabetes care. 2011; 34(11):2341–6. Epub 2011/09/09. https://doi.org/10.2337/dc11-0985 PMID: 21900148; PubMed Central PMCID: PMC3198285. mellitus. Diabetes care. 2011; 34(11):2341–6. Epub 2011/09/09. https://doi.org/10.2337/dc11-0985 PMID: 21900148; PubMed Central PMCID: PMC3198285. 28. 28. Valentini R, Dalfra MG, Masin M, Barison A, Marialisa M, Pegoraro E, et al. A pilot study on dietary approaches in multiethnicity: two methods compared. International journal of endocrinology. 2012; 2012:985136. Epub 2012/04/17. https://doi.org/10.1155/2012/985136 PMID: 22505892; PubMed Cen- tral PMCID: PMC3312248. 29. Markovic TP, Muirhead R, Overs S, Ross GP, Louie JC, Kizirian N, et al. Randomized Controlled Trial Investigating the Effects of a Low-Glycemic Index Diet on Pregnancy Outcomes in Women at High Risk of Gestational Diabetes Mellitus: The GI Baby 3 Study. Diabetes care. 2016; 39(1):31–8. Epub 2015/ 07/18. https://doi.org/10.2337/dc15-0572 PMID: 26185283. 30. Cypryk K, Kaminska P, Kosinski M, Pertynska-Marczewska M, Lewinski A. A comparison of the effec- tiveness, tolerability and safety of high and low carbohydrate diets in women with gestational diabetes. Endokrynologia Polska. 2007; 58(4):314–9. Epub 2007/12/07. PMID: 18058723. 31. Wang H, Jiang H, Yang L, Zhang M. Impacts of dietary fat changes on pregnant women with gestational diabetes mellitus: a randomized controlled study. Asia Pacific journal of clinical nutrition. 2015; 24 (1):58–64. Epub 2015/03/06. PLOS ONE \| https://doi.org/10.1371/journal.pone.0182095 August 3, 2017 References Wu BT, Dyer RA, King DJ, Richardson KJ, Innis SM. Early second trimester maternal plasma choline and betaine are related to measures of early cognitive development in term infants. PLoS One. 2012; 7 (8):e43448. https://doi.org/10.1371/journal.pone.0043448 PMID: 22916264; PubMed Central PMCID: PMCPMC3423345. 43. Bisschop PH, de Metz J, Ackermans MT, Endert E, Pijl H, Kuipers F, et al. Dietary fat content alters insulin-mediated glucose metabolism in healthy men. The American journal of clinical nutrition. 2001; 73(3):554–9. PMID: 11237931. PLOS ONE \| https://doi.org/10.1371/journal.pone.0182095 August 3, 2017 17 / 17
https://openalex.org/W1968024451	https://www.frontiersin.org/articles/10.3389/fneur.2014.00141/pdf	English	null	Medical Management of Hereditary Optic Neuropathies	Frontiers in neurology	2,014	cc-by	6,810	INTRODUCTION ND1, and 14484/ND6) (17). Other rare mtDNA mutations have been reported in association with LHON (18). These rare muta- tions may need testing for, by complete sequence of mtDNA, if the clinical suspicion of LHON is high and a maternal inheritance is obvious. Hereditary optic neuropathies are a heterogeneous group of dis- eases, which may be either an isolated optic neuropathy or optic neuropathy in the context of a more complex multi-systemic disease. The estimated minimum prevalence of hereditary optic neuropathies is about 1 in 10,000 (1). Hereditary optic neuropathies are a heterogeneous group of dis- eases, which may be either an isolated optic neuropathy or optic neuropathy in the context of a more complex multi-systemic disease. The estimated minimum prevalence of hereditary optic neuropathies is about 1 in 10,000 (1). Dominant optic atrophy, in about 70% of cases, is due to muta- tions in the OPA1 gene. Three other loci have been associated with autosomal DOA and designated as OPA4 (19), OPA5 (20), and OPA8 (21). Furthermore, heterozygous mutations in the OPA3 gene have been associated with DOA and cataract (22), as well as heterozygous mutations in the WSF1 gene have been reported in association with DOA and hearing loss (23). Both OPA3 and WSF1 genes are mostly associated with recessive mutations leading to Costeff (24) and Wolfram syndromes, respectively. Additional loci have been identiﬁed in X-linked (OPA2) (25) and reces- sive (OPA6 and OPA7) hereditary optic neuropathies, the latter recently associated with mutations in TMEM126A (26, 27). Optic neuropathy, as part of a more complex spectrum of clin- ical symptoms, is common in hereditary neurodegenerative disor- ders with evidence of mitochondrial dysfunction [for a review see Ref. (2)] such as OPA1 “plus” syndrome (3), mitochondr- ial encephalomyopathy lactic acidosis and stroke-like syndrome (MELAS) (4), LHON/dystonia/MELAS/Leigh overlapping syn- drome (5, 6), myoclonic epilepsy with ragged red ﬁbers (MERRF) (7), POLG1-related syndrome (8), Friedreich ataxia (9), Charcot– Marie Tooth type 2A (10),Mohr–Tranebjerg syndrome (11),SPG7 syndrome (12), Wolfram syndrome (13), spinocerebellar ataxias (14), and DNMT1-related disorders (15). The next sections will focus on LHON and DOA, the most fre- quent and prototypical hereditary optic neuropathies, for which an extensive body of studies is available. The main non-syndromic hereditary optic neuropathies are Leber’s hereditary optic neuropathy (LHON) and dominant optic atrophy (DOA) and both are primary mitochondrial hereditary optic neuropathies (16). Medical management of hereditary optic neuropathies Chiara La Morgia1,2, Michele Carbonelli 3, Piero Barboni 3,4, Alfredo Arrigo Sadun5 an 1 UOC Clinica Neurologica, IRCCS Istituto delle Scienze Neurologiche di Bologna, Ospedale Bellaria, Bologna, Italy 2 Unità di Neurologia, Dipartimento di Scienze Biomediche e NeuroMotorie (DIBINEM), Università di Bologna, Bologna, Italy 3 Studio Oculistico d’Azeglio, Bologna, Italy 4 Istituto Scientiﬁco San Raffaele, Milan, Italy 5 Doheny Eye Institute, University of California Los Angeles, Los Angeles, CA, USA Chiara La Morgia1,2, Michele Carbonelli 3, Piero Barboni 3,4, Alfredo Arrigo Sadun5 a Hereditary optic neuropathies are diseases affecting the optic nerve. The most com- mon are mitochondrial hereditary optic neuropathies, i.e., the maternally inherited Leber’s hereditary optic neuropathy (LHON) and dominant optic atrophy (DOA). They both share a mitochondrial pathogenesis that leads to the selective loss of retinal ganglion cells and axons, in particular of the papillo-macular bundle. Typically, LHON is characterized by an acute/subacute loss of central vision associated with impairment of color vision and swelling of retinal nerve ﬁbers followed by optic atrophy. DOA, instead, is characterized by a childhood-onset and slowly progressive loss of central vision, worsening over the years, leading to optic atrophy. The diagnostic workup includes neuro-ophthalmologic evaluation and genetic testing of the three most common mitochondrial DNA mutations affecting com- plex I (11778/ND4, 3460/ND1, and 14484/ND6) for LHON and sequencing of the nuclear gene OPA1 for DOA.Therapeutic strategies are still limited including agents that bypass the complex I defect and exert an antioxidant effect (idebenone). Further strategies are aimed at stimulating compensatory mitochondrial biogenesis. Gene therapy is also a promising avenue that still needs to be validated. CLINICAL PRESENTATION OF LHON/DOA AND CLINICAL WORKUP Leber’s hereditary optic neuropathy typically affects young adult males presenting with unilateral or bilateral subacute/acute pain- less loss of central vision. The clinical course typically involves a rapid worsening of visual acuity over a few weeks (nadir reached in about one to 6 months) and subsequent stabilization of visual INTRODUCTION Other rare recessive or X-linked forms of optic neuropathy are also described, in most cases not yet resolved from the genetic point of view. *Correspondence: Correspondence: Chiara La Morgia, UOC Clinica Neurologica, IRCCS Istituto delle Scienze Neurologiche di Bologna, Ospedale Bellaria, Via Altura, 3 40139 Bologna, Italy; Unità di Neurologia, Dipartimento di Scienze Biomediche e NeuroMotorie (DIBINEM), Università di Bologna, Bologna, Italy e-mail: chiaralamorgia@gmail.com; chiara.lamorgia@unibo.it Keywords: optic nerve, LHON, OPA1, DOA, hereditary, optic atrophy, leber, mitochondria Reviewed by: Shlomo Dotan, Hadassah Medical Center, Israel MINI REVIEW ARTICLE published: 31 July 2014 doi: 10.3389/fneur.2014.00141 MINI REVIEW ARTICLE published: 31 July 2014 doi: 10.3389/fneur.2014.00141 MINI REVIEW ARTICLE Edited by: Mohammad Hosein Nowroozzadeh, Shiraz University of Medical Sciences, Iran GENETIC BASIS OF ISOLATED OPTIC NEUROPATHIES GENETIC BASIS OF ISOLATED OPTIC NEUROPATHIES Leber’s hereditary optic neuropathy is a maternally inherited disorder and in 90–95% of cases it is due to one of three mitochondrial DNA (mtDNA) mutations (11778/ND4, 3460/ July 2014 \| Volume 5 \| Article 141 \| 1 www.frontiersin.org www.frontiersin.org La Morgia et al. Hereditary ON loss, entering the chronic phase. Loss of vision can be very severe leading to legal blindness in most cases (17). In the acute phase, there is bilateral sequential vision loss, signiﬁcant impairment of color vision, and usually maintenance of the pupillary light reﬂex. Fundus examination shows, in the acute phase, pseudoedema of the optic disk and retinal nerve ﬁbers initially involving mainly the inferior and superior arcades with teleangectias (Figure 1A) and temporal or diffuse optic disk pallor in later stage of the dis- ease. Visual ﬁelds reveal central or cecocentral scotoma, which progressively worsens over time (28). Optical coherence tomogra- phy (OCT) demonstrates an initial thickening of the temporal and inferior peripapillary ﬁbers, which is followed by the thickening of the superior and,lastly,of the nasal ﬁbers (Figure 1A). The tempo- ral ﬁbers are lost ﬁrst and the nasal ﬁbers are the most spared (29). Fluorescein angiography in the acute phase usually fails to show leakage of ﬂuorescein (hence pseudo disk edema) and this is a useful marker for differentiating LHON from inﬂammatory optic neuropathies with true disk edema that causes ﬂuorescein leakage (28).Another useful clinical sign,which distinguishes LHON from optic neuritis (the most important clinical differential diagnosis for LHON) is the usual absence of a relative afferent pupillary defect. The maintenance of the pupillary light response is partly due to the bilateral symmetry of the disease process and partly to the relative sparing of melanopsin retinal ganglion cells (mRGCs) in LHON, as revealed by post-mortem investigations (30). These cells are about 1% of the total RGCs and they contribute mainly to circadian photo-entrainment, but through the retinal projec- tions to the olivary pretectal nucleus (OPN), they also regulate FIGURE 1 \| (A) Fundus and OCT in the acute-phase of LHON. Fundus exam shows moderate swelling of the retinal nerve ﬁbers mainly involving the superior and inferior arcades around the optic disk associated with loss of ﬁbers in the papillo-macular bundle (temporal pallor), retinal vessels tortuosity, and telangiectasias. Optical coherence tomography (OCT) shows retinal nerve ﬁber layer thinning of the temporal quadrant with thickening of all the other quadrants. GENETIC BASIS OF ISOLATED OPTIC NEUROPATHIES Interestingly, OCT also demonstrates that optic disk size is reduced in DOA compared to controls and this might be explained by the role played by OPA1 in reg- ulating developmental apoptosis and shaping optic nerve head morphology (55). the pupillary light reﬂex. Clinical studies demonstrated the main- tenance of the pupil responses in LHON patients (31, 32). The diagnostic workup in LHON also may include other neurophys- iologic tests such as visual evoked potentials (VEPs) that reveal delayed latency and reduced amplitude of cortical responses and pattern electroretinograms (PERGs), which demonstrate reduced amplitude and delayed latency of responses (33). Some LHON patients may spontaneously recover vision during the clinical course (34). The rate of spontaneous recovery is vari- able for the three LHON mutations being 4–20% for 11778/ND4 and 3460/ND1 mutation, and 50% or more for the 14484/ND6 (34, 35). This recovery may occur after the nadir of the disease at about 6–8 months and up to 5 years after the onset (1). The family history is very helpful since it often reveals the pres- ence of other affected individuals along the maternal lineage, in prevalence males. As for LHON, the clinical examination in DOA demonstrates the maintenance of the pupillary light reﬂex. This ﬁnding is cor- roborated by histological studies showing a relative sparing of mRGCs in DOA as well (30, 56). Leber’s hereditary optic neuropathy is due in the large majority of cases (90–95%) to the three classical LHON mutations, which are homoplasmic (i.e.,all the mtDNA copies harbor the mutation) in most of the families. Despite the genetic risk of disease seems to be equal among all individuals along the maternal lineage, the penetrance is incomplete and the disease affects prevalently males (17). Genetic counseling implies in these patients that the off- spring of a woman carrying the LHON mutation will harbor the mutation, whereas offspring of males are spared. The probability to become affected is roughly 10% in women and 50% in men carrying the 11778/ND4 LHON mutation (17). The reason for the lower penetrance in females is possibly related to the protective role exerted by estrogens on cellular metabolism (36). The risk of recurrence for individuals younger than 50 years of age and mater- nally related to the index case has been calculated by Harding and colleagues (37). GENETIC BASIS OF ISOLATED OPTIC NEUROPATHIES (B) Fundus and OCT of a LHON carrier. Fundus examination reveals mild swelling of the superior and inferior retinal nerve ﬁbers and microangiopathy. OCT demonstrates retinal nerve ﬁber layer thickening more evident in the supero-nasal (OD) and supero-inferior (OS) quadrants. FIGURE 1 \| (A) Fundus and OCT in the acute-phase of LHON. Fundus exam shows moderate swelling of the retinal nerve ﬁbers mainly involving the superior and inferior arcades around the optic disk associated with loss of ﬁbers in the papillo-macular bundle (temporal pallor), retinal vessels tortuosity, and telangiectasias. Optical coherence tomography (OCT) shows retinal nerve ﬁber layer thinning of the temporal quadrant with thickening of all the other quadrants. (B) Fundus and OCT of a LHON carrier. Fundus examination reveals mild swelling of the superior and inferior retinal nerve ﬁbers and microangiopathy. OCT demonstrates retinal nerve ﬁber layer thickening more evident in the supero-nasal (OD) and supero-inferior (OS) quadrants. FIGURE 1 \| (A) Fundus and OCT in the acute-phase of LHON. Fundus exam shows moderate swelling of the retinal nerve ﬁbers mainly involving the superior and inferior arcades around the optic disk associated with loss of ﬁbers in the papillo-macular bundle (temporal pallor), retinal vessels tortuosity, and telangiectasias. Optical coherence tomography (OCT) shows retinal nerve FIGURE 1 \| (A) Fundus and OCT in the acute-phase of LHON. Fundus exam shows moderate swelling of the retinal nerve ﬁbers mainly involving the superior and inferior arcades around the optic disk associated with loss of ﬁbers in the papillo-macular bundle (temporal pallor), retinal vessels tortuosity, and telangiectasias. Optical coherence tomography (OCT) shows retinal nerve July 2014 \| Volume 5 \| Article 141 \| 2 Frontiers in Neurology \| Neuro-Ophthalmology Frontiers in Neurology \| Neuro-Ophthalmology La Morgia et al. Hereditary ON LHON in that the former presents as a slowly progressive loss of vision with a smaller central or cecocentral scotoma (51). The clinical severity is very variable even within the same pedi- gree (52). As in LHON, OCT demonstrates preferential involve- ment of the papillo-macular bundle with a corresponding loss of ﬁbers in the temporal sector of the optic disk and relative sparing of the nasal quadrant (53), which leads to temporal or diffuse optic atrophy. The main differential diagnosis for DOA is normal tension glaucoma, due to the frequent cupping of chronic DOA cases (54). GENETIC BASIS OF ISOLATED OPTIC NEUROPATHIES A speciﬁc mtDNA background, haplogroup J, due to recurrence of non-synonymous variants affecting complex I and III subunit genes, has been associated with higher penetrance of the mutations 11778/ND4 and 14484/ND6 (38). VEPs may be normal or disclose a delayed latency and reduced amplitude of cortical responses (57). Genetic counseling for DOA implies the rules of Mendelian autosomal dominant diseases and so the probability that offspring of an affected individual will inherit the mutation is 50% and this is equal for males and females. Prenatal genetic diagnosis is feasible, as well as pre-implantation diagnosis, thus providing the option to avoid transmission of the mutant OPA1 allele to offspring of an affected parent. As for LHON, DOA also may present a “plus” phenotype, which includes deafness, chronic external ophthalmo- plegia, peripheral neuropathy, cerebellar atrophy, and mitochon- drial myopathy (3, 58). In these cases, muscle biopsy histology and histoenzymology may disclose the hallmarks of mitochondr- ial myopathy with cytochrome c oxidase negative ﬁbers, whereas molecular analysis will reveal the pathological accumulation of multiple mtDNA deletions (58). Moreover, these patients often present increased lactic acid levels after standardized exercise. The OPA1 mutations associated with the “plus” phenotype are in most cases missense mutations affecting the GTP-ase domain of the OPA1 protein (58), as opposed to the majority of mutations associated with non-syndromic DOA, mostly predicted to induce haploinsufﬁciency. Other known environmental risk factors for LHON are smok- ing, alcohol, and the use of some antibiotics (i.e., macrolids, aminoglicosides, ethambutol, isoniazide, and linezolid) (39, 40). Clinical counseling should include avoiding these risk factors. Though usually LHON presents as an isolated optic neuropa- thy, in a few cases, there may be the coexistence of additional signsandsymptoms,suchasmyoclonus,MS-likefeatures,epilepsy, heart conduction defects, peripheral neuropathy, dystonia, and myopathy (41). FOLLOW-UP MANAGEMENT www.frontiersin.org FOLLOW-UP MANAGEMENT Patients with hereditary optic neuropathies should be fol- lowed regularly during the disease course, particularly in the acute/subacute phase. The most important exams for the clini- cal follow-up of these patients are color vision, visual ﬁelds, visual acuity, and OCT. Other ancillary tests that have been shown useful to complete the clinical characterization of LHON patients include brain and muscle magnetic resonance spectroscopy (42, 43) and lactic acid after standardized exercise (44). Visual acuity is usually poor in LHON and some patients may end with bare light perception. DOA patients present a more vari- able clinical severity (52). Visual ﬁelds usually deteriorate from small to very large central scotoma in LHON and may remain small and stable in DOA over long periods of time (59). In LHON, visual ﬁelds may sometimes show reduction in size of the sco- toma or its fenestration (60), both leading to an improvement in visual acuity (Figure 2). However,spontaneous visual recovery has never been described in DOA (61). Only one rare OPA1 mutation has been associated with a reversible LHON-like visual loss (62). Asymptomatic carriers of LHON mutation may also have swollen retinal nerve ﬁbers that can be seen on fundus exami- nation and by OCT (45, 46) (Figure 1B), abnormalities in color vision (47), and in neural conduction along the visual pathways (33, 48). Dominant optic atrophy usually affects both males and females in young childhood, and in the large majority of cases, is due to mutations in the OPA1 gene (49, 50). There is incomplete penetrance and the classical clinical picture of DOA differs from July 2014 \| Volume 5 \| Article 141 \| 3 www.frontiersin.org www.frontiersin.org La Morgia et al. Hereditary ON design predeﬁned primary end-points and dosages (Yu-Wai-Man, personal communication). Idebenone treatment has also been reported to ameliorate color vision in LHON (67). FIGURE 2 \| 30° and 10° visual ﬁeld pattern of recovery in LHON. (A) Example of scotoma fenestration (OS) in a LHON patient carrying the 14484/ND6 mutation (B) Example of scotoma contraction (OS) in a LHON patient carrying the 11778/ND4 mutation and (C) Example of both scotoma fenestration and contraction (OD) in a LHON patient carrying the 11778/ND4 mutation. Very preliminarily, encouraging results with idebenone treat- ment have been also reported in DOA patients (61) for whom spontaneous recovery has not been previously described. FOLLOW-UP MANAGEMENT The rationale for the use of idebenone in DOA is that also in OPA1-related DOA, there is a complex I defect as in LHON (68). Very preliminarily, encouraging results with idebenone treat- ment have been also reported in DOA patients (61) for whom spontaneous recovery has not been previously described. The rationale for the use of idebenone in DOA is that also in OPA1-related DOA, there is a complex I defect as in LHON (68). More recently, other drugs are under investigations, such as the novel quinone molecule, alfa-tocotrienol EPI-743, which has shown promising results in an open-label trial on a limited group of LHON patients (69). These results need to be conﬁrmed in larger controlled trial. More recently, other drugs are under investigations, such as the novel quinone molecule, alfa-tocotrienol EPI-743, which has shown promising results in an open-label trial on a limited group of LHON patients (69). These results need to be conﬁrmed in larger controlled trial. Other therapeutic strategies for LHON may include the use of drugs that might stimulate mitochondrial biogenesis, which has been demonstrated as a possible compensatory mechanism for LHON and other mitochondrial disorders (70). These include bezaﬁbrate and rosiglitazone and novel molecules such as resver- atrol and AICAR (5-aminoimidazole-4-carboxamide ribonucleo- side) (71). For a variety of reasons, gene therapy is a promising avenue for LHON treatment. The eye is easily accessible and provides a privileged space into which inject a viral vector. The most studied approach in LHON has been the AAV-mediated nuclear allotopic expression of mtDNA-encoded genes delivered to mitochondria (72, 73). This approach has been used in mouse models of LHON by the intravitreal delivery of an AAV vector carrying the wild- type ND4 (74, 75). This approach is now starting to be applied in LHON patients. However, it remains somehow controversial in terms of feasibility and efﬁciency (76). Other strategies include the direct delivery within mitochondria of the whole mtDNA or of an AAV vector containing the mtDNA-encoded ND4 subunit gene (77). These approaches are extremely promising but need to be validated. E 2 \| 30° and 10° visual ﬁeld pattern of recovery in LHON. FIGURE 2 \| 30° and 10° visual ﬁeld pattern of recovery in LHON. CLINICAL SIGNIFICANCE Although diseases as LHON and DOA are rare, they constitute a terrible tragedy for the patients who go blind and for their families. Unlike most common causes of blindness, these hereditary optic neuropathies are not associated with old age. Generally, patients with LHON and DOA loose their vision before young adulthood and the disease tend to be bilaterally symmetrical. Hence, typical patients will loose their vision as a teenager or in their 20s and face a lifetime of blindness and disability. Parents and siblings are often part of the process of diagnosis and management. For DOA, being a nuclear-encoded disease, gene therapy strat- egy should be more straightforward and feasible, as demonstrated by the recent encouraging results obtained with the gene ther- apy in Leber’s congenital amaurosis (78, 79). A proof of principle that genetically determined RGCs loss can be rescued by an AAV- based strategy with intravitreal delivery of the viral vector has been provided in an AIF mouse model, the so-called Harlequine mouse (80). There is issue of packaging, the 30 exons of the OPA1 gene into the AAV. Another problem to solve is which isoform of the OPA1 gene should be used as the most efﬁcient to res- cue the phenotype in RGCs. Currently, the speciﬁc role played by the eight OPA1 isoforms has not yet been systematically investi- gated. Another critical point is that the delivery of wild-type OPA1 to RGCs would likely complement haploinsufﬁciency mutations but more investigations are needed on the purported dominant negative mechanism of missense mutations. Nonetheless, gene therapy in DOA is a realistic objective, which will be pursued in the next future, providing some hope for treatment of DOA patients. THERAPEUTIC OPTIONS FOLLOW-UP MANAGEMENT (A) Example of scotoma fenestration (OS) in a LHON patient carrying the 14484/ND6 mutation (B) Example of scotoma contraction (OS) in a LHON patient carrying the 11778/ND4 mutation and (C) Example of both scotoma fenestration and contraction (OD) in a LHON patient carrying the 11778/ND4 mutation. A component of the visual improvement in LHON can be the adaptive phenomenon of eccentric ﬁxation (63). CLINICAL SIGNIFICANCE Frontiers in Neurology \| Neuro-Ophthalmology REFERENCES Exp Eye Res (2014) 120:161–6. doi:10.1016/j.exer.2013.11.005 11. Ha AD, Parratt KL, Rendtorff ND, Lodahl M, Ng K, Rowe DB, et al. The pheno- typic spectrum of dystonia in Mohr-Tranebjaerg syndrome. Mov Disord (2012) 27:1034–40. doi:10.1002/mds.25033 12. Klebe S, Depienne C, Gerber S, Challe G, Anheim M, Charles P, et al. Spastic paraplegia gene 7 in patients with spasticity and/or optic neuropathy. Brain (2012) 135:2980–93. doi:10.1093/brain/aws240 33. Ziccardi L, Sadun F, De Negri AM, Barboni P, Savini G, Borrelli E, et al. Retinal function and neural conduction along the visual pathways in affected and unaf- fected carriers with Leber’s hereditary optic neuropathy. Invest Ophthalmol Vis Sci (2013) 54:6893–901. doi:10.1167/iovs.13-12894 13. Rigoli L, Di Bella C. Wolfram syndrome 1 and Wolfram syndrome 2. Curr Opin Pediatr (2012) 24:512–7. doi:10.1097/MOP.0b013e328354ccdf 34. Stone EM, Newman NJ, Miller NR, Johns DR, Lott MT, Wallace DC. Visual recovery in patients with Leber’s hereditary optic neuropathy and the 11778 mutation. J Clin Neuroophthalmol (1992) 12:10–4. 14. Stricker S, Oberwahrenbrock T, Zimmermann H, Schroeter J, Endres M, Brandt AU,et al. Temporal retinal nerve ﬁber loss in patients with spinocerebellar ataxia 14. Stricker S, Oberwahrenbrock T, Zimmermann H, Schroeter J, Endres M, Brandt AU,et al. Temporal retinal nerve ﬁber loss in patients with spinocerebellar ataxia type 1. PLoS One (2011) 6:e23024. doi:10.1371/journal.pone.0023024 AU,et al. Temporal retinal nerve ﬁber loss in patients with spinocerebellar atax type 1. PLoS One (2011) 6:e23024. doi:10.1371/journal.pone.0023024 type 1. PLoS One (2011) 6:e23024. doi:10.1371/journal.pone.0023024 35. Carelli V, La Morgia C, Valentino ML, Rizzo G, Carbonelli M, De Negri AM, et al. Idebenone treatment in Leber’s hereditary optic neuropathy. Brain (2011) 134:e188. doi:10.1093/brain/awr180 15. Moghadam KK, Pizza F, La Morgia C, Franceschini C, Tonon C, Lodi R, et al. Narcolepsy is a common phenotype in HSAN IE and ADCA-DN. Brain (2014) 37(Pt 6):1643–55. doi:10.1093/brain/awu069 36. Giordano C, Montopoli M, Perli E, Orlandi M, Fantin M, Ross-Cisneros FN, et al. Oestrogens ameliorate mitochondrial dysfunction in Leber’s hereditary optic neuropathy. Brain (2011) 134:220–34. doi:10.1093/brain/awq276 16. Carelli V, La Morgia C, Sadun AA. Mitochondrial dysfunction in optic neu- ropathies: animal models and therapeutic options. Curr Opin Neurol (2013) 26:52–8. doi:10.1097/WCO.0b013e32835c5f0b 37. Harding AE, Sweeney MG, Govan GG, Riordan-Eva P. Pedigree analysis in Leber hereditary optic neuropathy families with a pathogenic mtDNA mutation. Am J Hum Genet (1995) 57:77–86. 17. Carelli V, Ross-Cisneros FN, Sadun AA. Mitochondrial dysfunction as a cause of opticneuropathies.ProgRetinEyeRes (2004)23:53–89.doi:10.1016/j.preteyeres. 2003.10.003 38. REFERENCES A ﬁrst locus for isolated autosomal recessive optic atrophy (ROA1) maps to chromosome 8q. Eur J Hum Genet (2003) 11:966–71. doi:10.1038/sj.ejhg.5201070 5. Bruyn GW, Went LN. A sex-linked heredo-degenerative neurological disorder, associated with Leber’s optic atrophy. I. Clinical studies. J Neurol Sci (1964) 1:59–80. doi:10.1016/0022-510X(64)90054-1 5. Bruyn GW, Went LN. A sex-linked heredo-degenerative neurological disorder, associated with Leber’s optic atrophy. I. Clinical studies. J Neurol Sci (1964) 1:59–80. doi:10.1016/0022-510X(64)90054-1 27. Hanein S, Perrault I, Roche O, Gerber S, Khadom N, Rio M, et al. TMEM126A, encoding a mitochondrial protein, is mutated in autosomal-recessive nonsyn- dromic optic atrophy. Am J Hum Genet (2009) 84:493–8. doi:10.1016/j.ajhg. 2009.03.003 6. Novotny EJ Jr, Singh G,Wallace DC, Dorfman LJ, Louis A, Sogg RL, et al. Leber’s disease and dystonia: a mitochondrial disease. Neurology (1986) 36:1053–60. doi:10.1212/WNL.36.8.1053 28. Newman NJ. Hereditary optic neuropathies: from the mitochondria to the optic nerve. Am J Ophthalmol (2005) 140:517–23. doi:10.1016/j.ajo.2005.03.017 7. Mancuso M, Orsucci D, Angelini C, Bertini E, Carelli V, Comi GP, et al. Phe- notypic heterogeneity of the 8344A>G mtDNA “MERRF” mutation. Neurology (2013) 80:2049–54. doi:10.1212/WNL.0b013e318294b44c 29. Barboni P, Carbonelli M, Savini G, Ramos Cdo V, Carta A, Berezovsky A, et al. Natural history of Leber’s hereditary optic neuropathy: longitudinal analysis of the retinal nerve ﬁber layer by optical coherence tomography. Ophthalmology (2010) 117:623–7. doi:10.1016/j.ophtha.2009.07.026 8. Milone M, Wang J, Liewluck T, Chen LC, Leavitt JA, Wong LJ. Novel POLG splice site mutation and optic atrophy. Arch Neurol (2011) 68:806–11. doi:10. 1001/archneurol.2011.124 30. La Morgia C, Ross-Cisneros FN, Sadun AA, Hannibal J, Munarini A, Man- tovani V, et al. Melanopsin retinal ganglion cells are resistant to neurode- generation in mitochondrial optic neuropathies. Brain (2010) 133:2426–38. doi:10.1093/brain/awq155 9. Fortuna F, Barboni P, Liguori R, Valentino ML, Savini G, Gellera C, et al. Visual system involvement in patients with Friedreich’s ataxia. Brain (2009) 132:116–23. doi:10.1093/brain/awn269 10. Rouzier C, Bannwarth S, Chaussenot A, Chevrollier A, Verschueren A, Bonello- Palot N, et al. The MFN2 gene is responsible for mitochondrial DNA instability and optic atrophy“plus”phenotype. Brain (2012) 135:23–34. doi:10.1093/brain/ awr323 31. Moura AL,Nagy BV,La Morgia C,Barboni P,Oliveira AG,Salomão SR,et al. The pupil light reﬂex in Leber’s hereditary optic neuropathy: evidence for preserva- tion of melanopsin-expressing retinal ganglion cells. Invest Ophthalmol Vis Sci (2013) 54:4471–7. doi:10.1167/iovs.12-11137 32. Kawasaki A, Collomb S, Léon L, Münch M. Pupil responses derived from outer and inner retinal photoreception are normal in patients with hereditary optic neuropathy. REFERENCES 1. Yu-Wai-Man P, Grifﬁths PG, Chinnery PF. Mitochondrial optic neuropathies- disease mechanisms and therapeutic strategies. Prog Retin Eye Res (2011) 30:81–114. doi:10.1016/j.preteyeres.2010.11.002 1. Yu-Wai-Man P, Grifﬁths PG, Chinnery PF. Mitochondrial optic neuropathies- disease mechanisms and therapeutic strategies. Prog Retin Eye Res (2011) 30:81–114. doi:10.1016/j.preteyeres.2010.11.002 23. Rendtorff ND, Lodahl M, Boulahbel H, Johansen IR, Pandya A, Welch KO, et al. Identiﬁcation of p.A684V missense mutation in the WFS1 gene as a frequent cause of autosomal dominant optic atrophy and hearing impairment. Am J Med Genet A (2011) 155A:1298–313. doi:10.1002/ajmg.a.33970 2. Carelli V, La Morgia C, Valentino ML, Barboni P, Ross-Cisneros FN, Sadun AA. Retinal ganglion cell neurodegeneration in mitochondrial inherited disorders. Biochim Biophys Acta (2009) 1787:518–28. doi:10.1016/j.bbabio.2009.02.024 2. Carelli V, La Morgia C, Valentino ML, Barboni P, Ross-Cisneros FN, Sadun AA. Retinal ganglion cell neurodegeneration in mitochondrial inherited disorders. Biochim Biophys Acta (2009) 1787:518–28. doi:10.1016/j.bbabio.2009.02.024 24. Anikster Y, Kleta R, Shaag A, Gahl WA, Elpeleg O. Type III 3-methylglutaconic aciduria (optic atrophy plus syndrome, or Costeff optic atrophy syndrome): identiﬁcation of the OPA3 gene and its founder mutation in Iraqi Jews. Am J Hum Genet (2001) 69:1218–24. doi:10.1086/324651 3. Amati-Bonneau P,Valentino ML, Reynier P, Gallardo ME, Bornstein B, Boissière A, et al. OPA1 mutations induce mitochondrial DNA instability and optic atro- phy “plus” phenotypes. Brain (2008) 131:338–51. doi:10.1093/brain/awm298 3. Amati-Bonneau P,Valentino ML, Reynier P, Gallardo ME, Bornstein B, Boissière A, et al. OPA1 mutations induce mitochondrial DNA instability and optic atro- phy “plus” phenotypes. Brain (2008) 131:338–51. doi:10.1093/brain/awm298 25. Assink JJ, Tijmes NT, ten Brink JB, Oostra RJ, Riemslag FC, de Jong PT, et al. A gene for X-linked optic atrophy is closely linked to the Xp11.4-Xp11.2 region of the X chromosome. Am J Hum Genet (1997) 61:934–9. 4. Pulkes T, Eunson L, Patterson V, Siddiqui A, Wood NW, Nelson IP, et al. The mitochondrial DNA G13513A transition in ND5 is associated with a LHON/MELAS overlap syndrome and may be a frequent cause of MELAS. Ann Neurol (1999) 46:916–9. doi:10.1002/1531-8249(199912)46:6<916::AID- ANA16>3.0.CO;2-R 4. Pulkes T, Eunson L, Patterson V, Siddiqui A, Wood NW, Nelson IP, et al. The mitochondrial DNA G13513A transition in ND5 is associated with a LHON/MELAS overlap syndrome and may be a frequent cause of MELAS. Ann Neurol (1999) 46:916–9. doi:10.1002/1531-8249(199912)46:6<916::AID- ANA16>3.0.CO;2-R 26. Barbet F, Gerber S, Hakiki S, Perrault I, Hanein S, Ducroq D, et al. THERAPEUTIC OPTIONS Therapeutic options for hereditary optic neuropathies are still limited (64) and up to now only one randomized double-blind placebo-controlled study with idebenone in LHON has been concluded (65). Idebenone (a quinone analog of coenzyme Q10) has been reported to be effective in increasing the rate of recovery in LHON especially in patients with discordant visual acuity and if the patients are treated early in the disease course and for a long time (35, 65). These ﬁndings may suggest that there is a speciﬁc win- dow of opportunity for increasing the rate of visual recovery with idebenone treatment. Coenzyme Q10 and idebenone are able to bypass the complex I defect caused by the LHON mutation and permitthetransferof electronsfromNADHdirectlytocomplexIII (66). A new double-blind placebo-controlled trial with idebenone in LHON is about to start with a longer follow-up and different ACKNOWLEDGMENTS h h d Research on hereditary optic neuropathies was supported by Telethon Grants GGP06233 (to Valerio Carelli), GGP11182 (to July 2014 \| Volume 5 \| Article 141 \| 4 Frontiers in Neurology \| Neuro-Ophthalmology La Morgia et al. Hereditary ON Massimo Zeviani and Valerio Carelli), and GPP10005 (to Valerio Carelli), Mitocon Onlus (to Valerio Carelli); Research to Prevent Blindness (RPB), the International Foundation for Optic Nerve Diseases(IFOND),StrugglingWithinLeber’s,ThePoincenotFam- ily, the Eierman Foundation, and a National Eye Institute grant EY03040 (to Alfredo Arrigo Sadun). of a second locus on chromosome 18q12.2-12.3. Arch Ophthalmol (1999) 117:805–10. of a second locus on chromosome 18q12.2-12.3. Arch Ophthalmol (1999) 117:805–10. 20. Barbet F, Hakiki S, Orssaud C, Gerber S, Perrault I, Hanein S, et al. A third locus for dominant optic atrophy on chromosome 22q. J Med Genet (2005) 42:e1. doi:10.1136/jmg.2004.025502 21. Carelli V, Schimpf S, Fuhrmann N, Valentino ML, Zanna C, Iommarini L, et al. A clinically complex form of dominant optic atrophy (OPA8) maps on chromosome 16. Hum Mol Genet (2011) 20:1893–905. doi:10.1093/hmg/ddr071 22. Reynier P, Amati-Bonneau P, Verny C, Olichon A, Simard G, Guichet A, et al. OPA3 gene mutations responsible for autosomal dominant optic atrophy and cataract. J Med Genet (2004) 41:e110. doi:10.1136/jmg.2003.016576 REFERENCES Effect of EPI-743 on the clinical course of the mitochondrial disease Leber hereditary optic neuropathy. Arch Neurol (2012) 69:331–8. doi:10.1001/ archneurol.2011.2972 50. Alexander C, Votruba M, Pesch UE, Thiselton DL, Mayer S, Moore A, et al. OPA1, encoding a dynamin-related GTPase, is mutated in autosomal domi- nant optic atrophy linked to chromosome 3q28. Nat Genet (2000) 26:211–5. doi:10.1038/79944 70. Giordano C, Iommarini L, Giordano L, Maresca A, Pisano A, Valentino ML, et al. Efﬁcient mitochondrial biogenesis drives incomplete penetrance in Leber’s hereditary optic neuropathy. Brain (2014) 137:335–53. doi:10.1093/ brain/awt343 51. Votruba M, Moore AT, Bhattacharya SS. Clinical features, molecular genet- ics, and pathophysiology of dominant optic atrophy. J Med Genet (1998) 35:793–800. doi:10.1136/jmg.35.10.793 71. Wenz T,Williams SL, Bacman SR, Moraes CT. Emerging therapeutic approaches to mitochondrial diseases. Dev Disabil Res Rev (2010) 16:219–29. doi:10.1002/ ddrr.109 52. Cohn AC, Toomes C, Potter C, Towns KV, Hewitt AW, Inglehearn CF, et al. Auto- somal dominant optic atrophy: penetrance and expressivity in patients with OPA1 mutations. Am J Ophthalmol (2007) 143:656–62. doi:10.1016/j.ajo.2006. 12.038 72. Guy J, Qi X, Pallotti F, Schon EA, Manfredi G, Carelli V, et al. Rescue of a mito- chondrial deﬁciency causing Leber hereditary optic neuropathy. Ann Neurol (2002) 52:534–42. doi:10.1002/ana.10354 53. Barboni P, Savini G, Parisi V, Carbonelli M, La Morgia C, Maresca A, et al. Retinal nerve ﬁber layer thickness in dominant optic atrophy measurements by optical coherence tomography and correlation with age. Ophthalmology (2011) 118:2076–80. doi:10.1016/j.ophtha.2011.02.027 73. Bonnet C,Augustin S,Ellouze S,Bénit P,BouaitaA,Rustin P,et al. The optimized allotopic expression of ND1 or ND4 genes restores respiratory chain complex I activity in ﬁbroblasts harboring mutations in these genes. Biochim Biophys Acta (2008) 1783:1707–17. doi:10.1016/j.bbamcr.2008.04.018 54. Votruba M, Thiselton D, Bhattacharya SS. Optic disc morphology of patients with OPA1 autosomal dominant optic atrophy. Br J Ophthalmol (2003) 87:48–53. doi:10.1136/bjo.87.1.48 74. Ellouze S, Augustin S, Bouaita A, Bonnet C, Simonutti M, Forster V, et al. Opti- mized allotopic expression of the human mitochondrial ND4 prevents blindness in a rat model of mitochondrial dysfunction.Am J Hum Genet (2008) 83:373–87. doi:10.1016/j.ajhg.2008.08.013 55. Barboni P, Carbonelli M, Savini G, Foscarini B, Parisi V, Valentino ML, et al. OPA1 mutations associated with dominant optic atrophy inﬂuence optic nerve head size. Ophthalmology (2010) 117:1547–53. doi:10.1016/j.ophtha. 2009.12.042 75. Guy J, Qi X, Koilkonda RD, Arguello T, Chou TH, Ruggeri M, et al. REFERENCES doi:10.1136/jnnp.72.6.805 44. Montagna P, Plazzi G, Cortelli P, Carelli V, Lugaresi E, Barboni P, et al. Abnormal lactate after effort in healthy carriers of Leber’s hereditary optic neuropathy. J Neurol Neurosurg Psychiatry (1995) 58:640–1. doi:10.1136/jnnp.58.5.640 64. Pfeffer G, Horvath R, Klopstock T, Mootha VK, Suomalainen A, Koene S, et al. New treatments for mitochondrial disease-no time to drop our standards. Nat Rev Neurol (2013) 9:474–81. doi:10.1038/nrneurol.2013.129 65. Klopstock T, Yu-Wai-Man P, Dimitriadis K, Rouleau J, Heck S, Bailie M, et al. A randomized placebo-controlled trial of idebenone in Leber’s hereditary optic neuropathy. Brain (2011) 134:2677–86. doi:10.1093/brain/awr170 45. Sadun AA, Salomao SR, Berezovsky A, Sadun F, Denegri AM, Quiros PA, et al. Subclinical carriers and conversions in Leber hereditary optic neuropathy: a prospective psychophysical study. Trans Am Ophthalmol Soc (2006) 104:51–61. 46. Barboni P, Savini G, Feuer WJ, Budenz DL, Carbonelli M, Chicani F, et al. Reti- nal nerve ﬁber layer thickness variability in Leber hereditary optic neuropathy carriers. Eur J Ophthalmol (2012) 22:985–91. doi:10.5301/ejo.5000154 66. Geromel V, Darin N, Chrétien D, Bénit P, DeLonlay P, Rötig A, et al. Coenzyme Q(10) and idebenone in the therapy of respiratory chain diseases: rationale and comparative beneﬁts. Mol Genet Metab (2002) 77:21–30. doi:10.1016/S1096- 7192(02)00145-2 47. Ventura DF, Gualtieri M, Oliveira AG, Costa MF, Quiros P, Sadun F, et al. Male prevalence of acquired color vision defects in asymptomatic carriers of Leber’s hereditary optic neuropathy. Invest Ophthalmol Vis Sci (2007) 48:2362–70. doi:10.1167/iovs.06-0331 67. Rudolph G, Dimitriadis K, Büchner B, Heck S, Al-Tamami J, Seidensticker F, et al. Effects of idebenone on color vision in patients with leber heredi- tary optic neuropathy. J Neuroophthalmol (2013) 33:30–6. doi:10.1097/WNO. 0b013e318272c643 48. Sacai PY, Salomão SR, Carelli V, Pereira JM, Belfort R Jr, Sadun AA, et al. Visual evoked potentials ﬁndings in non-affected subjects from a large Brazilian pedi- gree of 11778 Leber’s hereditary optic neuropathy. Doc Ophthalmol (2010) 121:147–54. doi:10.1007/s10633-010-9241-2 68. Zanna C, Ghelli A, Porcelli AM, Karbowski M, Youle RJ, Schimpf S, et al. OPA1 mutations associated with dominant optic atrophy impair oxidative phosphory- lation and mitochondrial fusion. Brain (2008) 131:352–67. doi:10.1093/brain/ awm335 49. Delettre C, Lenaers G, Griffoin JM, Gigarel N, Lorenzo C, Belenguer P, et al. Nuclear gene OPA1, encoding a mitochondrial dynamin-related pro- tein, is mutated in dominant optic atrophy. Nat Genet (2000) 26:207–10. doi:10.1038/79936 69. Sadun AA, Chicani CF, Ross-Cisneros FN, Barboni P, Thoolen M, Shrader WD, et al. REFERENCES CarelliV,Achilli A,Valentino ML,Rengo C,Semino O,Pala M,et al. Haplogroup effects and recombination of mitochondrial DNA: novel clues from the analy- sis of Leber hereditary optic neuropathy pedigrees. Am J Hum Genet (2006) 78:564–74. doi:10.1086/501236 18. Achilli A, Iommarini L, Olivieri A, Pala M, Hooshiar Kashani B, Reynier P, et al. Rare primary mitochondrial DNA mutations and probable synergistic variants in Leber’s hereditary optic neuropathy. PLoS One (2012) 7:e42242. doi:10.1371/journal.pone.0042242 39. Sadun AA, La Morgia C, Carelli V. Mitochondrial optic neuropathies: our trav- els from bench to bedside and back again. Clin Experiment Ophthalmol (2013) 41:702–12. doi:10.1111/ceo.12086 19. Kerrison JB, Arnould VJ, Ferraz Sallum JM, VageﬁMR, Barmada MM, Li Y, et al. Genetic heterogeneity of dominant optic atrophy, Kjer type: identiﬁcation July 2014 \| Volume 5 \| Article 141 \| 5 www.frontiersin.org La Morgia et al. Hereditary ON genetic studies. Acta Ophthalmol Scand (2005) 83:337–46. doi:10.1111/j.1600- 0420.2005.00448.x 40. Kirkman MA, Yu-Wai-Man P, Korsten A, Leonhardt M, Dimitriadis K, De Coo IF, et al. Gene-environment interactions in Leber hereditary optic neuropathy. Brain (2009) 132:2317–26. doi:10.1093/brain/awp158 genetic studies. Acta Ophthalmol Scand (2005) 83:337–46. doi:10.1111/j.1600- 0420.2005.00448.x 60. Nakamura M, Yamamoto M. Variable pattern of visual recovery of Leber’s hereditary optic neuropathy. Br J Ophthalmol (2000) 84:534–5. doi:10.1136/ bjo.84.5.534 41. Nikoskelainen EK, Marttila RJ, Huoponen K, Juvonen V, Lamminen T, Sonni- nen P, et al. Leber’s “plus”: neurological abnormalities in patients with Leber’s hereditary optic neuropathy. J Neurol Neurosurg Psychiatry (1995) 59:160–4. doi:10.1136/jnnp.59.2.160 61. Barboni P,Valentino ML, La Morgia C,Carbonelli M, Savini G, De Negri A, et al. Idebenone treatment in patients with OPA1-mutant dominant optic atrophy. Brain (2013) 136:e231. doi:10.1093/brain/aws280 42. La Morgia C, Achilli A, Iommarini L, Barboni P, Pala M, Olivieri A, et al. Rare mtDNA variants in Leber hereditary optic neuropathy families with recurrence of myoclonus. Neurology (2008) 70:762–70. doi:10.1212/01.wnl.0000295505. 74234.d0 62. Cornille K, Milea D, Amati-Bonneau P, Procaccio V, Zazoun L, Guillet V, et al. Reversible optic neuropathy with OPA1 exon 5b mutation. Ann Neurol (2008) 63:667–71. doi:10.1002/ana.21376 63. Altpeter EK,Blanke BR,Leo-Kottler B,Nguyen XN,Trauzettel-Klosinski S. Eval- uation of ﬁxation pattern and reading ability in patients with Leber heredi- tary optic neuropathy. J Neuroophthalmol (2013) 33:344–8. doi:10.1097/WNO. 0b013e31829d1f5b 43. Lodi R, Carelli V, Cortelli P, Iotti S, Valentino ML, Barboni P, et al. Phospho- rus MR spectroscopy shows a tissue speciﬁc in vivo distribution of biochemical expression of the G3460A mutation in Leber’s hereditary optic neuropathy. J Neurol Neurosurg Psychiatry (2002) 72:805–7. 79. Testa F, Maguire AM, Rossi S, Pierce EA, Melillo P, Marshall K, et al. Three- year follow-up after unilateral subretinal delivery of adeno-associated virus in patients with Leber congenital amaurosis type 2. Ophthalmology (2013) 120:1283–91. doi:10.1016/j.ophtha.2012.11.048 Citation: La Morgia C, Carbonelli M, Barboni P, Sadun AA and Carelli V (2014) Medical management of hereditary optic neuropathies. Front. Neurol. 5:141. doi: 10.3389/fneur.2014.00141 80. Bouaita A,Augustin S, Lechauve C, Cwerman-Thibault H, Bénit P, Simonutti M, et al. Downregulation of apoptosis-inducing factor in Harlequin mice induces progressive and severe optic atrophy which is durably prevented by AAV2-AIF1 gene therapy. Brain (2012) 135:35–52. doi:10.1093/brain/awr290 This article was submitted to Neuro-Ophthalmology, a section of the journal Frontiers in Neurology. Copyright © 2014 La Morgia, Carbonelli, Barboni, Sadun and Carelli. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permit- ted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Received: 23 May 2014; accepted: 16 July 2014; published online: 31 July 2014. July 2014 \| Volume 5 \| Article 141 \| 7 Received: 23 May 2014; accepted: 16 July 2014; published online: 31 July 2014. Citation: La Morgia C, Carbonelli M, Barboni P, Sadun AA and Carelli V (2014) Medical management of hereditary optic neuropathies. Front. Neurol. 5:141. doi: 10.3389/fneur.2014.00141 This article was submitted to Neuro-Ophthalmology, a section of the journal Frontiers in Neurology. Copyright © 2014 La Morgia, Carbonelli, Barboni, Sadun and Carelli. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permit- ted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. REFERENCES Efﬁciency and safety of AAV-mediated gene delivery of the human ND4 complex I sub- unit in the mouse visual system. Invest Ophthalmol Vis Sci (2009) 50:4205–14. doi:10.1167/iovs.08-3214 56. Perganta G, Barnard AR, Katti C, Vachtsevanos A, Douglas RH, MacLaren RE, et al. Non-image-forming light driven functions are preserved in a mouse model of autosomal dominant optic atrophy. PLoS One (2013) 8:e56350. doi:10.1371/journal.pone.0056350 76. Perales-Clemente E, Fernández-Silva P, Acín-Pérez R, Pérez-Martos A, Enríquez JA. Allotopic expression of mitochondrial-encoded genes in mammals: achieved goal, undemonstrated mechanism or impossible task? Nucleic Acids Res (2011) 39:225–34. doi:10.1093/nar/gkq769 j 57. Holder GE,Votruba M, Carter AC, Bhattacharya SS, Fitzke FW, Moore AT. Elec- trophysiological ﬁndings in dominant optic atrophy (DOA) linking to the OPA1 locus on chromosome 3q 28-qter. Doc Ophthalmol (1999) 95:217–28. 77. Yu H, Koilkonda RD, Chou TH, Porciatti V, Ozdemir SS, Chiodo V, et al. Gene delivery to mitochondria by targeting modiﬁed adenoassoci- ated virus suppresses Leber’s hereditary optic neuropathy in a mouse model. Proc Natl Acad Sci U S A (2012) 109:E1238–47. doi:10.1073/pnas.1119577109 58. Yu-Wai-Man P, Grifﬁths PG, Gorman GS, Lourenco CM, Wright AF, Auer- Grumbach M, et al. Multi-system neurological disease is common in patients with OPA1 mutations. Brain (2010) 133:771–86. doi:10.1093/brain/awq007 78. Maguire AM, Simonelli F, Pierce EA, Pugh EN Jr, Mingozzi F, Bennicelli J, et al. Safety and efﬁcacy of gene transfer for Leber’s congenital amaurosis. N Engl J Med (2008) 358:2240–8. doi:10.1056/NEJMoa0802315 59. Puomila A, Huoponen K, Mäntyjärvi M, Hämäläinen P, Paananen R, Sankila EM, et al. Dominant optic atrophy: correlation between clinical and molecular Frontiers in Neurology \| Neuro-Ophthalmology Frontiers in Neurology \| Neuro-Ophthalmology July 2014 \| Volume 5 \| Article 141 \| 6 La Morgia et al. Hereditary ON Conﬂict of Interest Statement: Dr. Alfredo Arrigo Sadun is the PI on unrestricted researchgrantsfromEdisonPharmaceuticals.Dr.ValerioCarellireceivedreimburse- ments of travel expenses for consulting activities from Edison Pharmaceuticals. All other co-authors have no conﬂict of interest. July 2014 \| Volume 5 \| Article 141 \| 7 www.frontiersin.org
https://openalex.org/W3130094177	https://pure.rug.nl/ws/files/166102027/_18760104_Journal_for_European_Environmental_Planning_Law_Is_Aarhus_Still_a_Progressive_Force_.pdf	English	null	Is Aarhus Still a Progressive Force?	Journal for European environmental & planning law	2,021	cc-by	2,428	University of Groningen University of Groningen Published in: Journal For European Environmental & Planning Law IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record Link to publication in University of Groningen/UMCG research database Citation for published version (APA): Squintani, L. (2021). Is Aarhus Still a Progressive Force? Journal For European Environmental & Planning Law, 18(1-2), 4-7. https://doi.org/10.1163/18760104-18010002 Citation for published version (APA): Squintani, L. (2021). Is Aarhus Still a Progressive Force? Journal For European Environmental & Planning Law, 18(1-2), 4-7. https://doi.org/10.1163/18760104-18010002 Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons). Citation for published version (APA): Squintani, L. (2021). Is Aarhus Still a Progressive Force? Journal For European Environmental & Planning Law, 18(1-2), 4-7. https://doi.org/10.1163/18760104-18010002 1 United Nations, Convention on Access to Information, Public Participation in Decision- Making and Access to Justice in Environmental Matters, Aarhus, Denmark 25 June 1998, UN Treaty Series 2161, 447. 2 Judgment of 7 November 2019, Alain Flausch and others v. Ypourgos Perivallontos kai Energeias and Others, C-280/18, ecli:EU:C:2019:928. 3 L. Squintani, Case Law of the Court of Justice of the European Union and the General Court, JEEPL 2020 (17/2) 229-24. This is an open access article distributed under the terms of the CC BY 4.0 license. Downloaded from Brill.com03/22/2021 01:28:56PM via free access Copyright The publication may also be distributed here under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license. More information can be found on the University of Groningen website: https://www.rug.nl/library/open-access/self-archiving-pure/taverne- amendment. Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to and investigate your claim. Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum. Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum. Download date: 24-10-2024 © Lorenzo Squintani, 2021 \| doi:10.1163/18760104-18010002 This is an open access article distributed under the terms of the CC BY 4.0 license. Downloaded from Brill.com03/22/2021 01:28:56PM via free access © Lorenzo Squintani, 2021 \| doi:10.1163/18760104-18010002 Is Aarhus Still a Progressive Force? Lorenzo Squintani University of Groningen, Groningen, The Netherlands l.squintani@rug.nl The year 2020 will be remembered for the pandemic we are all experiencing first hand. Its effects are visible in many aspects of our lives, including aca- demic productivity. This Journal registered a significant increase of submis- sions over the course of 2020, especially in concomitance with the first wave of the virus. This is why we decided to make the first issue of the 2021 Volume a double issue. Among the various publications included in this double issue, there are three that focus on the Aarhus Convention.1 Anna Berti Suman’s manuscript focuses on the relationship between access to information, citizen sensing and public participation. Vasiliki (Vicky) Karageorgou’s manuscript, instead, focuses on the interrelation between public participation and access to justice in the eia context by building upon a recent Flausch et al Case2 of the Court of Justice, about which we reported in the previous issue of this Journal.3 Magdalena Maria Michalak and Przemyslaw Kledzik’s manuscript, finally, focuses on the editorial 5 parties right in proceedings for issuing and environmental impact assessment decision, by reviewing the Polish experience. When looked at them conjunctively, these papers show the great effects that the Aarhus Convention is having on the environmental governance struc- tures of the signatory states. During the first four lustra of its regulatory work, the Aarhus Convention stood undeniably as a driving force for environmental democracy in Europe, as Wates wrote in 2005 in this Journal.4 Despite many progresses in the regulatory frameworks of the signatory states, more are still to be expected thanks to the continuous work of refinement brought forward by the Aarhus Convention Compliance Committee, internationally, and by the Court of Justice of the European Union, as well by the national courts, in the European Union and national legal orders. 5 A. Aragão, When feelings become scientific facts: valuing cultural ecosystem services and taking them into account in public decision-making, in L. Squintani, J. Darpö, L. Lavrysen & P.T. Stoll, Managing Facts and Feelings in Environmental Governance, 2019, pp. 53–80. 6 G. Perlaviciute and L. Squintani, Public Participation in Climate Policy Making: Toward Reconciling Public Preferences and Legal Frameworks, One Earth, 2020 2(4), 341–348. https://doi.org/(…)j.oneear.2020.03.009 and L. Squintani and G. Perlaviciute, Access to Public Participation: Unveiling the Mismatch between what Law Prescribes and what the Public Wants, in M. Peeters and M. Eliantonoi (eds.), Research Handbook on EU Environmental Law, In Edward Elgar Publishing 2020, 133–147. 7 L. Squintani and H. Schoukens, Towards equal opportunities in public participation in environmental matters in the European Union, in L. Squintani, J. Darpö, L. Lavrysen, and P.T. Stoll (eds.), Managing Facts and Feelings in Environmental Governance, Edward Elgar Publishing 2019, 22–52. 4 J. Wates, The Aarhus Convention: a driving force for environmental democracy, JEEPL 2005 2(1) 3–11. Journal for European Environmental & Planning Law 18 (2021) 4-7 Downloaded from Brill.com03/22/2021 01:28:56PM via free access Is Aarhus Still a Progressive Force? Still, Suman’s plea for a progressive interpretation of the Aarhus Convention public participation pillar, made me think about another manuscript I had the pleasure to read in 2019, in which Aragão pleaded for a human right based interpretation of the participatory pillar of the Convention in order to foster an ecosystem services based approach to nature conservation.5 Both manu- scripts show that a main stream interpretation of the Convention, or proba- bly I should say a conservative reading of the Convention, does not suffice to encompass a reality in which the people – in Aarhus’ jargon, the public – want to play a more active role in shaping the environment in which we live. In my own research, I am also encountering more and more evidence about a mismatch between the Aarhus regulatory framework on public participa- tion and people’s preferences thereabout.6 Besides, together with Schoukens, I advocate for a progressive reading of the Aarhus participatory pillar in order to foster the participatory rights of weak societal groups, such a low educated.7 4 J. Wates, The Aarhus Convention: a driving force for environmental democracy, JEEPL 2005 2(1) 3–11. Journal for European Environmental & Planning Law 18 (2021) 4-7 Downloaded from Brill.com03/22/2021 01:28:56PM via free access 6 SQUINTANI The question thus raises about where the Aarhus Convention stops being a progressive force, and starts being a conservative one. Of course, many manuscripts in this journal showed the important work that, thanks to the Aarhus Convention, is taking place about all the pillars of the Convention, but in particular the access to justice pillar.8 Karageorgou’s manuscript confirms that there is still important work to perform under the public participation pillar. Yet, what if compliance with the access to information and public par- ticipation pillars results being insufficient to ensure a level of environmental democracy that is in line with the expectations of the public, or at least a part thereof? In other words, does full compliance with the Aarhus regulatory framework ensures behavioral and problem-solving effectiveness of this reg- ulatory framework?9 From an academic perspective, it would be relevant to see a stream of pub- lications investigating the boundaries of the Aarhus regulatory framework and to start a dialogue about the need, room, and forms of a possible moderniza- tion of the Aarhus Convention. This Journal would very much welcome man- uscripts on this topic. 8 E.g. Wates, at note 4; V.Koester, Review of Compliance under the Aarhus Convention: A Rather Unique Compliance Mechanism, JEEPL 2005 2(1), p. 31–44; J. Jendrośka, Citizen’s Rights in European Environmental Law: Stock-Taking of Key Challenges and Current Developments in Relation to Public Access to Environmental Information, Participation and Access to Justice, JEEPL 2012 9, p. 71–90; J. Darpö, Article 9.2 of the Aarhus Convention and EU Law, JEEPL 2014 11, p. 367–391; Mariolina Eliantonio and Franziska Grashof, Wir müssen reden! – We Need to Have a Serious Talk!, JEEPL 2016 13(3–4) 325–349; L. Krämer, Access to Environmental Justice: The Double Standards of the ecj, JEEPL 2017 14, p. 159– 185; U.Gieseke, The Aarhus Convention in Practice: Challenges and Perspectives for German Environmental Authorities, JEEPL 2019 20, p. 372–385. 8 E.g. Wates, at note 4; V.Koester, Review of Compliance under the Aarhus Convention: A Rather Unique Compliance Mechanism, JEEPL 2005 2(1), p. 31–44; J. Jendrośka, Citizen’s Rights in European Environmental Law: Stock-Taking of Key Challenges and Current Developments in Relation to Public Access to Environmental Information, Participation and Access to Justice, JEEPL 2012 9, p. 71–90; J. Darpö, Article 9.2 of the Aarhus Convention and EU Law, JEEPL 2014 11, p. 367–391; Mariolina Eliantonio and Franziska Grashof, Wir müssen reden! – We Need to Have a Serious Talk!, JEEPL 2016 13(3–4) 325–349; L. Krämer, Access to Environmental Justice: The Double Standards of the ecj, JEEPL 2017 14, p. 159– 185; U.Gieseke, The Aarhus Convention in Practice: Challenges and Perspectives for German Environmental Authorities, JEEPL 2019 20, p. 372–385. 9 For the difference between legal, behavioral and problem-solving effectiveness, see S. Maljean-Dubois, The Effectiveness of Environmental Law: A Key Topic, in S. Maljean- Dubois (ed.), The Effectiveness of Environmental Law, Intersentia 2017, 1–13, available at https://www.eelf.info/home-7.html. 10 A. Savaresi and L. Perugini, The Land Sector in the 2030 EU Climate Change Policy Framework: a Look At The Future, JEEPL 2019 16(2) 148–164. 11 Judgment of 14 March 2019, Commission v Czech Republic, C-399/17, EU:C:2019:200. For the difference between legal, behavioral and problem-solving effectiveness, see S. Maljean-Dubois, The Effectiveness of Environmental Law: A Key Topic, in S. Maljean- Dubois (ed.), The Effectiveness of Environmental Law, Intersentia 2017, 1–13, available at https://www.eelf.info/home-7.html. 11 Judgment of 14 March 2019, Commission v Czech Republic, C-399/17, EU:C:2019:200. Journal for European Environmental & Planning Law 18 (2021) 4-7 Downloaded from Brill.com03/22/2021 01:28:56PM via free access 10 A. Savaresi and L. Perugini, The Land Sector in the 2030 EU Climate Change Policy Framework: a Look At The Future, JEEPL 2019 16(2) 148–164. 12 With attention to the constraints posed by the attribution principle in this regard, a topic addressed in B. Vanheusden and L. Squintani, Reconciling Conflicting Values: A Call For Research on Instruments to Achieve Quasi-Sustainability, in B. Vanheusden & L. Squintani (eds.), EU Environmental and Planning Law Aspects of Large-Scale Projects, Intersentia 2016, 385 ff. 13 L. Squintani, Tort-Law based Environmental Litigation: a Victory or a Warning?, JEEPL 2018 15(3–4) 277–280; L. Squintani, Addressing the (Lack of) Effectiveness of Environmental Law and the Gap between Law in the Books and Law in Action, JEEPL 2020 17(2) 133–135, and L. Squintani, Environmental Protection and Human Rights: Recent Developments and Recurring Questions, JEEPL 2020 17(3) 265–266. Is Aarhus Still a Progressive Force? The other topics covered in this double issue array from forest protection, with Kathrin Böhling and Maria Fernanda Marques Todeschini’s manuscript complementing the work of Savaresi and Perugini published in the 2019 volume of this Journal,10 to waste management, with David Hadrousek’s manuscript focusing on the manner in which the Court of Justice dealt with the definition of waste in the TPS-NOLO (Geobal) Case,11 an ever green in academic litera- ture as well known. They range from forensic evidences in Poland, in Elżbieta A. Savaresi and L. Perugini, The Land Sector in the 2030 EU Climate Change Policy Framework: a Look At The Future, JEEPL 2019 16(2) 148–164. Judgment of 14 March 2019, Commission v Czech Republic, C-399/17, EU:C:2019:200. Journal for European Environmental & Planning Law 18 (2021) 4-7 Downloaded from Brill.com03/22/2021 01:28:56PM i f Journal for European Environmental & Planning Law 18 (2021) 4-7 Downloaded from Brill.com03/22/2021 01:28:56PM via free access 7 editorial Małgorzata Zebek and Denis Solodov’s manuscript, to the limited integration of environmental concerns into bilateral agreement of the European Union with third states, in Ludwig Krämer’s article.12 Finally, they space from an assessment of the limitations of tort law as a risk management tool in Herman Kasper Gilissen, Elbert de Jong, Marleen van Rijswick and Annemarie van Wezel’s manuscript, a topic hinted at in the Journal more times,13 to a reflec- tion about the latest developments in the field of EU energy policy law in Ruven Fleming and Romain Mauger’s manuscript, thus complementing the EU Green Deal assessment presented in the previous volume of this Journal.14 Małgorzata Zebek and Denis Solodov’s manuscript, to the limited integration of environmental concerns into bilateral agreement of the European Union with third states, in Ludwig Krämer’s article.12 Finally, they space from an assessment of the limitations of tort law as a risk management tool in Herman Kasper Gilissen, Elbert de Jong, Marleen van Rijswick and Annemarie van Wezel’s manuscript, a topic hinted at in the Journal more times,13 to a reflec- tion about the latest developments in the field of EU energy policy law in Ruven Fleming and Romain Mauger’s manuscript, thus complementing the EU Green Deal assessment presented in the previous volume of this Journal.14 We could not think about a better way to start 2021. We wish the readers a pleasant reading! 13 L. 14 L. Krämer, Planning for Climate and the Environment: the EU Green Deal, JEEPL 2020 17(3) 267–306. Journal for European Environmental & Planning Law 18 (2021) 4-7 Downloaded from Brill.com03/22/2021 01:28:56PM via free access Is Aarhus Still a Progressive Force? Squintani, Tort-Law based Environmental Litigation: a Victory or a Warning?, JEEPL 2018 15(3–4) 277–280; L. Squintani, Addressing the (Lack of) Effectiveness of Environmental Law and the Gap between Law in the Books and Law in Action, JEEPL 2020 17(2) 133–135, and L. Squintani, Environmental Protection and Human Rights: Recent Developments and Recurring Questions, JEEPL 2020 17(3) 265–266. Journal for European Environmental & Planning Law 18 (2021) 4-7 Downloaded from Brill.com03/22/2021 01:28:56PM via free access Journal for European Environmental & Planning Law 18 (2021) 4-7 Downloaded from Brill.com03/22/2021 01:28:56PM via free access
https://openalex.org/W2569285461	https://europepmc.org/articles/pmc5215803?pdf=render	English	null	Selection of Reference Genes for qRT-PCR Analysis of Gene Expression in Stipa grandis during Environmental Stresses	PloS one	2,017	cc-by	10,623	RESEARCH ARTICLE OPEN ACCESS Citation: Wan D, Wan Y, Yang Q, Zou B, Ren W, Ding Y, et al. (2017) Selection of Reference Genes for qRT-PCR Analysis of Gene Expression in Stipa grandis during Environmental Stresses. PLoS ONE 12(1): e0169465. doi:10.1371/journal. pone.0169465 Editor: Hemachandra Reddy, Texas Technical University Health Sciences Center, UNITED STATES Editor: Hemachandra Reddy, Texas Technical University Health Sciences Center, UNITED STATES Received: July 13, 2016 Accepted: December 16, 2016 Published: January 5, 2017 Copyright: © 2017 Wan et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Received: July 13, 2016 Accepted: December 16, 2016 Published: January 5, 2017 Copyright: © 2017 Wan et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: All relevant data are within the paper and its Supporting Information files. Funding: This work was supported by the National Basic Research Program of China (973 program, 2014CB138804), the Natural Science Foundation of Inner Mongolia Autonomous Region of China (2014BS0330), and the Central Non-profit Research Institutes Fundamental Research Funds of China (1610332015004). The funders had no Selection of Reference Genes for qRT-PCR Analysis of Gene Expression in Stipa grandis during Environmental Stresses Dongli Wan1☯, Yongqing Wan2☯, Qi Yang2, Bo Zou2, Weibo Ren1, Yong Ding1, Zhen Wang1, Ruigang Wang2, Kai Wang1, Xiangyang Hou1* Dongli Wan1☯, Yongqing Wan2☯, Qi Yang2, Bo Zou2, Weibo Ren1, Yong Ding1, Zhen Wang1, Ruigang Wang2, Kai Wang1, Xiangyang Hou1* 1 Key Laboratory of Grassland Resources and Utilization of Ministry of Agriculture, Institute of Grassland Research, Chinese Academy of Agricultural Sciences, Hohhot, China, 2 College of Life Sciences, Inner Mongolia Agricultural University, Hohhot, China a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 ☯These authors contributed equally to this work. * Houxy16@126.com ☯These authors contributed equally to this work. * Houxy16@126.com Abstract Stipa grandis P. Smirn. is a dominant plant species in the typical steppe of the Xilingole Plateau of Inner Mongolia. Selection of suitable reference genes for the quantitative real- time reverse transcription polymerase chain reaction (qRT-PCR) is important for gene expression analysis and research into the molecular mechanisms underlying the stress responses of S. grandis. In the present study, 15 candidate reference genes (EF1 beta, ACT, GAPDH, SamDC, CUL4, CAP, SNF2, SKIP1, SKIP5, SKIP11, UBC2, UBC15, UBC17, UCH, and HERC2) were evaluated for their stability as potential reference genes for qRT-PCR under different stresses. Four algorithms were used: GeNorm, NormFinder, BestKeeper, and RefFinder. The results showed that the most stable reference genes were different under different stress conditions: EF1beta and UBC15 during drought and salt stresses; ACT and GAPDH under heat stress; SKIP5 and UBC17 under cold stress; UBC15 and HERC2 under high pH stress; UBC2 and UBC15 under wounding stress; EF1beta and UBC17 under jasmonic acid treatment; UBC15 and CUL4 under abscisic acid treatment; and HERC2 and UBC17 under salicylic acid treatment. EF1beta and HERC2 were the most suitable genes for the global analysis of all samples. Furthermore, six target genes, SgPOD, SgPAL, SgLEA, SgLOX, SgHSP90 and SgPR1, were selected to validate the most and least stable reference genes under different treatments. Our results provide guidelines for reference gene selection for more accurate qRT-PCR quanti- fication and will promote studies of gene expression in S. grandis subjected to environ- mental stress. Reference Genes Selection in Stipa grandis green barrier and plays an important role in sandstorm prevention [3]. To date, studies of S. grandis have mostly focused on ecological aspects, such as the relationship between the photo- synthetic capacity of S. grandis and the soil moisture gradient [4], the effects of grazing inten- sity and duration on the biomass of S. grandis [5], and the effects of global warming and elevated CO2 on plant growth [6]. The physiological responses of S. grandis to high tempera- ture and drought have also been reported [2]. Several genetic diversity studies have been car- ried out on S. grandis related to grazing and location [1, 3, 7]. Transcriptome analysis was carried out to reveal the gene expression profiles of overgrazed and non-grazed S. grandis in a typical grassland; the results showed that many stress-related genes had differential expression under long-term grazing conditions [8]. However, the mechanism underlying S. grandis’s responses to grazing and related stresses remain unclear. To further understand the stress response mechanism of S. grandis, details of the expression patterns of more stress-related genes are necessary. role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. Quantitative real-time reverse transcription PCR (qRT-PCR) is an efficient tool for expression analysis because of its high-throughput, accuracy, sensitivity, and ease of use compared with traditional methods, such as northern blotting and semi-quantitative reverse transcription-PCR [9–11]. The inevitable errors or experimental deviations caused by sample preparation and data analysis can affect the reliability of the results; therefore, it is essential to normalize the data using appropriate internal reference genes [12]. The stability of the ref- erence gene’s expression is crucial for the accurate normalization of target gene expression following qRT-PCR quantification [13]. Extensive studies on reference gene stability and selection have been reported in various plants for different tissues and development stages, and under diverse stresses [14–22]. Traditionally, reference genes such as 18S rRNA, GAPDH (glyceraldehyde-3-phosphate dehydrogenase), ACT (β or γ actin), TUB (α or β tubu- lin), EF1α (elongation factor 1α), and UBQ (poly-ubiquitin) have been used most frequently as internal control genes [13, 17]. Recently, taking advantage of a variety of databases com- prising data from microarrays, expressed sequence tags, and transcriptomes, other stably expressed genes have been used as novel reference genes for qRT-PCR, such as CUL (Cullin), UBCP (ubiquitin carrier protein) [17], CAP (adenylyl cyclase-associated protein), SKIP1 (ASK-interacting protein 1) [22], F-BOX (F-box/kelch-repeat protein) [23], PP2A (protein phosphatase 2A) [24], SAMDC (s-adenosyl methionine decarboxylase) [12, 25], and SNF (sucrose non fermenting-1 protein kinase) [26]. However, the transcript levels of these refer- ence genes vary among different conditions to some extent [20, 27]. Thus, the selection of ideal reference genes to normalize target gene expression is important to improve the reli- ability of qRT-PCR results. Quantitative real-time reverse transcription PCR (qRT-PCR) is an efficient tool for expression analysis because of its high-throughput, accuracy, sensitivity, and ease of use compared with traditional methods, such as northern blotting and semi-quantitative reverse transcription-PCR [9–11]. The inevitable errors or experimental deviations caused by sample preparation and data analysis can affect the reliability of the results; therefore, it is essential to normalize the data using appropriate internal reference genes [12]. The stability of the ref- erence gene’s expression is crucial for the accurate normalization of target gene expression following qRT-PCR quantification [13]. Extensive studies on reference gene stability and selection have been reported in various plants for different tissues and development stages, and under diverse stresses [14–22]. Traditionally, reference genes such as 18S rRNA, In this study, 15 candidate reference genes that could be used potentially as internal control genes to normalize gene expression were selected. The expression stabilities of these reference genes under various abiotic stresses and three hormone treatments were assessed using four algorithms (geNorm, NormFinder, Bestkeeper, and RefFinder). According to our analysis, we were able to recommend the most suitable reference genes to quantify gene expression in S. grandis under various experimental conditions. Introduction Stipa grandis P. Smirn. is a C3 perennial bunchgrass [1] that is distributed widely in eastern Eurasian steppes and the middle Eurasian steppe zone [2]. This species is dominant in the typi- cal steppe of the Xilingole Plateau of Inner Mongolia [1]. The S. grandis steppe acts as a natural PLOS ONE \| DOI:10.1371/journal.pone.0169465 January 5, 2017 1 / 18 RNA isolation and cDNA synthesis Total RNA was extracted from the samples using the Trizol regent. The purity and integrity of the RNA samples were estimated by calculating their A260/280 absorbance ratios and by aga- rose gel electrophoresis analysis. Total RNA (0.5 μg), pretreated with RNase-free DNase I (Takara, Dalian, China), was reverse transcribed into cDNA using an M-MLV reverse tran- scriptase kit (Takara) and oligo (dT) 18 primers (Takara) following the manufacturer’s instruc- tions. The generated cDNAs were diluted 20-fold for qRT-PCR analysis. Reference gene selection and primer design The sequences of the candidate reference genes used in this study were selected based on our previous transcriptome analysis [8], and included EF1 beta, ACT, GAPDH, SamDC, CUL4, CAP, SNF2, SKIP1, SKIP5, SKIP11, UBC2, UBC15, UBC17, UCH, and HERC2. According to these sequences (S1 Text), specific primers for qRT-PCR were designed using the Premier5 software, with a melting temperature (T) of 59–61˚C, a length of 18–25 bp, an amplicon prod- uct size of 75–160 bp, and a GC content of 40–60%; the primers are listed in Table 1. qRT-PCR analysis The qRT-PCR was performed on a LightCycler 480 Real Time PCR system (Roche, Basel, Swit- zerland) with a final volume of 20 μL per reaction, as previously described [30]. Each 20-μL reaction mixture contained 5 μL of diluted cDNA, 10 μL of SYBR Premix Ex Taq II (Takara), 0.8 μL of each primer, and 3.4 μL of sterile distilled water. The thermal cycling program was 95˚C for 30 s, followed by 40 cycles of 95˚C for 5 s, 60˚C for 30 s, and 72˚C for 15 s. Melt- curve analyses were performed using a program with constant heating from 70˚C to 95˚C, fol- lowed by 95˚C for 15 s. Each reaction was carried out in three technical replicates. Hormone and abiotic stress treatments Four-week-old plants were used in the following experiments, and the different treatment con- ditions were mainly achieved according to methods described previously [12, 22, 28, 29]. Before treatment with hormones, salinity, high pH, and drought, S. grandis plants were removed from the soil gently, washed clean with tap water, and then immediately subjected to the different treatments. For plant hormone, high pH and salinity treatments, the plants were dipped separately in solutions of 100 μM salicylic acid (SA), 100 μM jasmonic acid (JA), 100 μM abscisic acid (ABA), high pH (pH 10), and 300 mM NaCl. For the drought treatment, the plants were placed on filter paper and incubated in a growth chamber. For the wounding treatment, plants growing in pots were wounded using tweezers and placed in a growth cham- ber. For cold and heat treatments, plants growing in pots were maintained at 4˚C or 42˚C, respectively. Untreated plants were used as the control. Shoot samples (from four to five plants) were taken at 1, 3, 6, 12, 24, and 48 h time points for all treatments, and an additional 0.5-h time point for drought and wounding treatments, and were frozen immediately in liquid nitrogen and stored at −80˚C until RNA extraction. Three independent biological experiments were performed. Plant materials and growth conditions S. grandis seeds were collected from the Baiyinxile Livestock Farm of Xilinhot, Inner Mongolia (116˚40´E, 43˚33´N). The seeds were sown in pots filled with a soil mixture (rich soil: vermicu- lite, 1:1, v/v), and incubated in a growth chamber at 23˚C with a 16-h light/8-h dark cycle. No specific permits were required for the described field studies. g No specific permits were required for the described field studies. 2 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0169465 January 5, 2017 Reference Genes Selection in Stipa grandis BestKeeper uses the standard deviation (SD) and coefficient of variance (CV) of the Ct values to calculate the stability of a candidate reference gene; the highest SD and CV val- ues equate to the least stable reference genes [32]. Gene symbol Gene name Primer sequences (5 3 ) Amplicon length (bp) Product melting temperature (˚C) Ampliﬁcation efﬁciency (%) EF1beta Elongation factor 1-beta CCATCAACGAATACGTCCAGAG CCCCAGTTCCAAGAATCCACTA 97 83.6 104.1 CUL4 Cullin-4 GTGGAAGGCGTCTGATGTGG GCTTAGCTTTTGTGCGTCGTT 140 82.1 107.3 UCH Ubiquitin carboxyl-terminal hydrolase 6 GAAACTATCTGGTGGAGGCGACT GGTATGCTTGACCCTGATAAATGTAG 129 82.6 106 UBC2 Ubiquitin-conjugating enzyme E2 2 AACATCACGGTCTGGAACGC GTTGGTGGCTTGTTGGGATAA 113 86.4 104.2 UBC17 Ubiquitin-conjugating enzyme E2-17 TATCCGTTCAAGCCTCCAAAG GGTCGGTAAGTAGTGAGCAGATT 160 82.2 101.9 HERC2 HECT domain E3 ubiquitin ligases ATGGGCTTCGGGAGTTTAGG GCTTATTTGGGAGACATAGTGACC 105 85.2 107.3 SNF2 Sucrose non fermenting-1 protein kinase GCGAGCGTTGCTACAGGAGA CCTCAGTTCGAGCAGCCAAA 101 84.3 105.7 CAP Cyclase-associated protein AGCTCCAGAAACCTGCATCAA GGCCTCCTTCCTTCAGTCAAA 100 81.9 106.2 SKIP1 ASK-interacting protein GGTTCTGGACCTGTTTGGCT TCTCCAGGTTCTTCAGATTAGCA 80 81.1 99.3 SKIP5 ASK-interacting protein TGTGAGCAAAACCCTCTGGAT TTGAGCGATGAGAATGAGTCATACT 80 80.3 105.2 SKIP11 ASK-interacting protein TTGGGGTATTGCATTCCGAG CCAGAAGGTTTGCTTCCGAT 147 86.6 102.7 SamDC S-adenosylmethionine decarboxylase proenzyme TGTCGTGGTTGTTGTTTTGGTT TGCTCACTGTGGTAGACGCAG 108 82.3 103.6 UBC15 Ubiquitin-conjugating enzyme 15 GACCGCTATGTTAGGAACTGCC AGAAGAGCCAAACCAATGCC 100 83.5 104.9 GAPDH Glyceraldehyde-3-phosphate dehydrogenase GGTATGTCCTTCCGTGTTCC CGCCTTGATAGCCTTCTTGA 105 81.5 101.4 ACT Actin-7 GCATGAAGGTGAAGGTGGTT TCATCGTACTCTGCCTTGGA 121 84.2 103.4 doi:10.1371/journal.pone.0169465.t001 PLOS ONE \| DOI:10.1371/journal.pone.0169465 January 5, 2017 4 / 18 geNorm determines the most stable reference genes from a set of tested genes in a given sample set under a certain experimental condition and calculates the gene expression normali- zation factor via geometric averaging of all the reference genes [11]. geNorm evaluates gene expression stability by the M value, and ranks the tested genes according to their expression stability by stepwise exclusion of the gene with the highest M value. The lower the M value, the more stable the gene’s expression. The optimal number of reference genes was determined by pairwise variation (V) with a threshold of 0.15 [11]. If the value of Vn/n+1 was below 0.15, additional genes were not required for normalization. NormFinder is an algorithm based on a mathematical model of gene expression and uses a solid statistical framework to estimate the expression variation of candidate normalization genes, as well as the variation in sample sub- groups, to identify the optimal normalization gene among a set of candidates [31]. A stability value, which is a direct measure of the estimated expression variation for each gene, is pro- vided by NormFinder; thus, the systematic errors introduced during the process of normaliza- tion can be evaluated. Stability analysis of reference genes First, three analytical tools, geNorm (version 3.5) [11], NormFinder (version 0.953) [31], and BestKeeper (version 1) [32], were used to calculate the expression stabilities of the 15 reference genes in all samples from the different treatments. 3 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0169465 January 5, 2017 PLOS ONE \| DOI:10.1371/journal.pone.0169465 January 5, 2017 Reference Genes Selection in Stipa grandis Table 1. Candidate reference genes and characteristics of qRT-PCR primers in S. grandis. Gene symbol Gene name Primer sequences (50-30) Amplicon length (bp) Product melting temperature (˚C) Ampliﬁcation efﬁciency (%) EF1beta Elongation factor 1-beta CCATCAACGAATACGTCCAGAG CCCCAGTTCCAAGAATCCACTA 97 83.6 104.1 CUL4 Cullin-4 GTGGAAGGCGTCTGATGTGG GCTTAGCTTTTGTGCGTCGTT 140 82.1 107.3 UCH Ubiquitin carboxyl-terminal hydrolase 6 GAAACTATCTGGTGGAGGCGACT GGTATGCTTGACCCTGATAAATGTAG 129 82.6 106 UBC2 Ubiquitin-conjugating enzyme E2 2 AACATCACGGTCTGGAACGC GTTGGTGGCTTGTTGGGATAA 113 86.4 104.2 UBC17 Ubiquitin-conjugating enzyme E2-17 TATCCGTTCAAGCCTCCAAAG GGTCGGTAAGTAGTGAGCAGATT 160 82.2 101.9 HERC2 HECT domain E3 ubiquitin ligases ATGGGCTTCGGGAGTTTAGG GCTTATTTGGGAGACATAGTGACC 105 85.2 107.3 SNF2 Sucrose non fermenting-1 protein kinase GCGAGCGTTGCTACAGGAGA CCTCAGTTCGAGCAGCCAAA 101 84.3 105.7 CAP Cyclase-associated protein AGCTCCAGAAACCTGCATCAA GGCCTCCTTCCTTCAGTCAAA 100 81.9 106.2 SKIP1 ASK-interacting protein GGTTCTGGACCTGTTTGGCT TCTCCAGGTTCTTCAGATTAGCA 80 81.1 99.3 SKIP5 ASK-interacting protein TGTGAGCAAAACCCTCTGGAT TTGAGCGATGAGAATGAGTCATACT 80 80.3 105.2 SKIP11 ASK-interacting protein TTGGGGTATTGCATTCCGAG CCAGAAGGTTTGCTTCCGAT 147 86.6 102.7 SamDC S-adenosylmethionine decarboxylase proenzyme TGTCGTGGTTGTTGTTTTGGTT TGCTCACTGTGGTAGACGCAG 108 82.3 103.6 UBC15 Ubiquitin-conjugating enzyme 15 GACCGCTATGTTAGGAACTGCC AGAAGAGCCAAACCAATGCC 100 83.5 104.9 GAPDH Glyceraldehyde-3-phosphate dehydrogenase GGTATGTCCTTCCGTGTTCC CGCCTTGATAGCCTTCTTGA 105 81.5 101.4 ACT Actin-7 GCATGAAGGTGAAGGTGGTT TCATCGTACTCTGCCTTGGA 121 84.2 103.4 doi:10.1371/journal.pone.0169465.t001 geNorm determines the most stable reference genes from a set of tested genes in a given sample set under a certain experimental condition and calculates the gene expression normali- zation factor via geometric averaging of all the reference genes [11]. geNorm evaluates gene expression stability by the M value, and ranks the tested genes according to their expression stability by stepwise exclusion of the gene with the highest M value. The lower the M value, the more stable the gene’s expression. The optimal number of reference genes was determined by pairwise variation (V) with a threshold of 0.15 [11]. If the value of Vn/n+1 was below 0.15, additional genes were not required for normalization. NormFinder is an algorithm based on a mathematical model of gene expression and uses a solid statistical framework to estimate the expression variation of candidate normalization genes, as well as the variation in sample sub- groups, to identify the optimal normalization gene among a set of candidates [31]. A stability value, which is a direct measure of the estimated expression variation for each gene, is pro- vided by NormFinder; thus, the systematic errors introduced during the process of normaliza- tion can be evaluated. In geNorm and NormFinder, the quantities obtained from Ct (cycle threshold) values via the delta-Ct method were used for data input, whereas Bestkeeper used raw Ct values. In geNorm and NormFinder, the quantities obtained from Ct (cycle threshold) values via the delta-Ct method were used for data input, whereas Bestkeeper used raw Ct values. BestKeeper uses the standard deviation (SD) and coefficient of variance (CV) of the Ct values to calculate the stability of a candidate reference gene; the highest SD and CV val- ues equate to the least stable reference genes [32]. 4 / 18 Reference Genes Selection in Stipa grandis Finally, RefFinder (http://fulxie.0fees.us/?type=reference), which is a comprehensive web- based tool that integrates geNorm, NormFinder, BestKeeper, and the comparative delta Ct method, was applied to determine the most suitable reference genes for the overall final rank- ing [33]. Performance of primers and expression profiles of candidate reference genes Fifteen fragments of candidate reference genes (Table 1), obtained from previous S. grandis transcriptome data [8], that exhibited stable expression after grazing (within a 2-fold change in expression level) were selected. Agarose gel electrophoresis and melting curve analysis showed that a specific fragment of the expected size and a single peak were observed (S1 Fig), respec- tively, after reverse transcriptase-PCR amplification using cDNA templates with each primer pair. The PCR amplification efficiencies of each primer pair across all samples ranged from 99.3% to 107.3% (Table 1). The expression levels of the candidate reference genes were determined by qRT-PCR across all samples (including normal and different stress conditions) (Fig 1), and a wide range of Ct values, from 15.25 to 30.03, were found, with the values for most of the genes lying between 20 and 26. SamDC exhibited the highest expression abundance, with a mean Ct of 16.91, while SKIP5 had the lowest expression level, with a mean Ct of 26.08 across all samples. SamDC (3.40 cycles) and UCH (4.33 cycles) showed the smallest expression variations across all the samples. In contrast, SKIP5 (6.34 cycles) and SKIP11 (6.21 cycles) exhibited the largest relative expression variations among all the samples. Validation of reference gene expression Peroxidase (POD/POX, EC:1.11.1.7), phenylalanine ammonia-lyase (PAL), late embryogenesis abundant protein (LEA), lipoxygenase (LOX), heat shock protein 90 (HSP90), and pathogene- sis-related protein 1 (PR1) encoding genes were obtained from the S. grandis transcriptome library [8] and used to confirm the utility of the validated candidate reference genes for qRT-PCR. The transcript levels of SgPOD and SgPAL under all treatments, SgLEA under drought, cold, salt, high pH and ABA treatments, SgLOX under wounding and JA treatments, SgHSP90 under heat treatment, and SgPR1 under SA treatment were quantified using the most and least stable reference genes, respectively. The primers were as follows: SgPOD: TCGACTTCGCTGTTCCCACT/CCTTTTGCGGTGAAGTTTCC g SgPR1: GTCGTCTGCAACAACAACGG/GGGACTACTGCGAAGCAATCA g SgPAL: AGCAGCACAACCAGGACGTA/GTCACGCAGTTCTTGACGGA g SgLEA: CTTCCACTTCCATGATTTACGC/CATCCATACAGACCCCAGACAC g SgLOX: CTGGAGGGCATCGAGCAC/GGGTAGAGCAGCATGTAGGG g SgHSP90: CGGATCTTGTGAACAACCTCG/GTAGAAACCAACACCAAACTGACC The 2−ΔΔCT method [34] was used to calculate the relative expression levels. Each biological sample was analyzed using three technical replicates. geNorm analysis of reference genes After SA and wounding treatments, the M values of the combinations EF1beta and HERC2, and UBC2 and UBC15 were the lowest, The optimal number of reference genes for normalization was calculated by pairwise varia- tion (V). Under various conditions, the Vn/n+1 values were lower than 0.15 (Fig 3), which suggested that adding a further reference gene would not improve the normalization quality and a combination of the two most stable reference genes was sufficient for normalization. NormFinder analysis of reference gene expression stability Based on the results calculated by NormFinder (Table 2), the three highest ranked genes under ABA treatment were CUL4, UBC15, and UBC17; under cold treatment were SKIP5, HERC2, and UBC17; under heat treatment were ACT, SKIP1, and GAPDH; under JA treatment were UBC17, EF1beta, and UBC15; under salt treatment were UBC15, EF1beta, and HERC2; under high pH treatment were UBC15, EF1beta, and SKIP5; under SA treatment were HERC2, UBC17, and EF1beta; under wounding treatment were UBC2, UBC15, and UBC17; and under drought treatment were EF1beta, UBC15, and GAPDH. According to all the samples in the study, UBC15, HERC2, and EF1beta were the most stable genes. In addition, the least stable genes ranked by NormFinder and geNorm were consistent: for cold and heat stresses they were SKIP1 and SKIP5, and for the other treatments it was SKIP11. geNorm analysis of reference genes The geNorm results are shown in Fig 2. For the drought treatment, UBC15 and UBC17 were the most stably expressed genes with the lowest M value of 0.0844. For ABA, salt, and JA treat- ments, the most stably expressed genes were EF1beta and UBC15, with M values of 0.096, 5 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0169465 January 5, 2017 Reference Genes Selection in Stipa grandis Fig 1. Distribution of Ct values of candidate reference genes in S. grandis plants obtained by qRT-PCR. Ct values for each reference gene are tested in all samples. Boxes represent the first and third quartiles of the data. Red lines across the boxes indicate the median Ct values. Whiskers show the maximum and minimum values. doi:10 1371/journal pone 0169465 g001 Fig 1. Distribution of Ct values of candidate reference genes in S. grandis plants obtained by qRT-PCR. Ct values for each reference gene are tested in all samples. Boxes represent the first and third quartiles of the data. Red lines across the boxes indicate the median Ct values. Whiskers show the maximum and minimum values. Fig 1. Distribution of Ct values of candidate reference genes in S. grandis plants obtained by qRT-PCR. Ct values for each reference gene are tested in all samples. Boxes represent the first and third quartiles of the data. Red lines across the boxes indicate the median Ct values. Whiskers show the maximum and minimum values. doi:10.1371/journal.pone.0169465.g001 doi:10.1371/journal.pone.0169465.g001 0.106, and 0.117, respectively. For heat treatment, HERC2 and ACT had the lowest M value of 0.137. Under cold treatment, UBC17 and SKIP5 had the lowest M value, and for high pH treat- ment, UBC17 and CUL4 were the most stable genes. After SA and wounding treatments, the M values of the combinations EF1beta and HERC2, and UBC2 and UBC15 were the lowest, respectively. Under heat and cold treatments, the least stable genes were SKIP5 and SKIP1, respectively. In the other seven treatments, SKIP11 was the most unstable gene. Across all sam- ples, UBC17 and CUL4 ranked as the most stable genes, while SKIP11 was the least stable gene. 0.106, and 0.117, respectively. For heat treatment, HERC2 and ACT had the lowest M value of 0.137. Under cold treatment, UBC17 and SKIP5 had the lowest M value, and for high pH treat- ment, UBC17 and CUL4 were the most stable genes. PLOS ONE \| DOI:10.1371/journal.pone.0169465 January 5, 2017 Bestkeeper analysis of reference gene expression stability The pairwise variation (Vn/n +1) between the normalization factors NFn and NFn+1 was applied to determine the optimal number of reference genes recommended to normalize qRT-PCR data in S. grandis under various stresses. If the Vn/n+1 values were below 0.15, an additional (n + 1) reference was not required. doi:10.1371/journal.pone.0169465.g003 Fig 3. Pairwise variation (V) analysis of the reference genes in S. grandis using geNorm. The pairwise variation (Vn/n +1) between the normalization factors NFn and NFn+1 was applied to determine the optimal number of reference genes recommended to normalize qRT-PCR data in S. grandis under various stresses. If the Vn/n+1 values were below 0.15, an additional (n + 1) reference was not required. doi:10 1371/journal pone 0169465 g003 Fig 3. Pairwise variation (V) analysis of the reference genes in S. grandis using geNorm. The pairwise variation (Vn/n +1) between the normalization factors NFn and NFn+1 was applied to determine the optimal number of reference genes recommended to normalize qRT-PCR data in S. grandis under various stresses. If the Vn/n+1 values were below 0.15, an additional (n + 1) reference was not required. RefFinder ranking of the most stable genes RefFinder, which combines the geNorm, NormFinder, and BestKeeper outputs, was applied to generate a comprehensive final ranking of the candidate references. As shown in Table 4, EF1beta (drought and JA), ACT (heat), SKIP5 (cold), UBC15 (salt, high pH, and ABA), UBC2 (wounding), and HERC2 (SA) were ranked as the top stable genes under the different treat- ments. Across all samples, EF1beta, HERC2, and UCH were suggested as the three best refer- ence genes for normalizing the expression of target genes. Bestkeeper analysis of reference gene expression stability As shown in Table 3, Bestkeeper ranked UBC17 as the most stably expressed gene following cold, wounding, and ABA treatments, while UCH had the lowest SD value under drought PLOS ONE \| DOI:10.1371/journal.pone.0169465 January 5, 2017 6 / 18 Reference Genes Selection in Stipa grandis Fig 2. Average expression stability (M) evaluated using geNorm. M values and rankings are indicated for S. grandis under different stresses. (A) Drought treatment. (B) Heat treatment. (C) Cold treatment. (D) Salt treatment. (E) High pH treatment. (F) Wounding treatment. (G) JA treatment. (H) ABA treatment. (I) SA treatment. (J) Total. doi:10.1371/journal.pone.0169465.g002 Fig 2. Average expression stability (M) evaluated using geNorm. M values and rankings are indicated for S. grandis under different stresses. (A) Drought treatment. (B) Heat treatment. (C) Cold treatment. (D) Salt treatment. (E) High pH treatment. (F) Wounding treatment. (G) JA treatment. (H) ABA treatment. (I) SA treatment. (J) Total. Fig 2. Average expression stability (M) evaluated using geNorm. M values and rankings are indicated for S. grandis under different stresses. (A) Drought treatment. (B) Heat treatment. (C) Cold treatment. (D) Salt treatment. (E) High pH treatment. (F) Wounding treatment. (G) JA treatment. (H) ABA treatment. (I) SA treatment. (J) Total. doi:10.1371/journal.pone.0169465.g002 doi:10.1371/journal.pone.0169465.g002 and salt stresses. GAPDH, SamDC, UBC2, and SKIP1 were ranked as the most stably expressed genes under heat, high pH, JA, and SA treatments, respectively. When all the sam- ples were analyzed together, UCH was rated as the most stable gene with the lowest SD value. and salt stresses. GAPDH, SamDC, UBC2, and SKIP1 were ranked as the most stably expressed genes under heat, high pH, JA, and SA treatments, respectively. When all the sam- ples were analyzed together, UCH was rated as the most stable gene with the lowest SD value. 7 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0169465 January 5, 2017 Reference Genes Selection in Stipa grandis Fig 3. Pairwise variation (V) analysis of the reference genes in S. grandis using geNorm. The pairwise variation (Vn/n +1) between the normalization factors NFn and NFn+1 was applied to determine the optimal number of reference genes recommended to normalize qRT-PCR data in S. grandis under various stresses. If the Vn/n+1 values were below 0.15, an additional (n + 1) reference was not required. doi:10 1371/journal pone 0169465 g003 Fig 3. Pairwise variation (V) analysis of the reference genes in S. grandis using geNorm. Validation of candidate reference genes To validate the utility of the candidate reference genes selected using RefFinder, the expression patterns of six target genes identified from RNA-Seq data, including SgPOD, SgPAL, SgLEA, SgLOX, SgHSP90, and SgPR1, in response to different treatments were examined. When nor- malization was performed using the two most stable genes under each treatment, EF1beta and UBC15 (drought), SKIP5 and UBC17 (cold), UBC15 and EF1beta (salt), ACT and GAPDH (heat), UBC15 and HERC2 (high pH), UBC2 and UBC15 (wounding), EF1beta and UBC17 (JA), UBC15 and CUL4 (ABA), and HERC2 and UBC17 (SA), the expression of SgPOD was 8 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0169465 January 5, 2017 Reference Genes Selection in Stipa grandis Table 2. Stabilities of candidate reference genes ranked by NormFinder. Rank Cold Drought Heat Salt High pH Wounding ABA JA SA Total 1 SKIP5 EF1beta ACT UBC15 UBC15 UBC2 CUL4 UBC17 HERC2 UBC15 Stability value 0.076 0.009 0.097 0.033 0.066 0.036 0.056 0.075 0.032 0.154 2 HERC2 UBC15 SKIP1 EF1beta EF1beta UBC15 UBC15 EF1beta UBC17 HERC2 Stability value 0.111 0.112 0.118 0.037 0.158 0.036 0.094 0.098 0.051 0.172 3 UBC17 GAPDH GAPDH HERC2 SKIP5 UBC17 UBC17 UBC15 EF1beta EF1beta Stability value 0.117 0.128 0.133 0.125 0.234 0.040 0.107 0.126 0.052 0.187 4 CUL4 ACT HERC2 UCH HERC2 EF1beta ACT CUL4 UBC2 ACT Stability value 0.137 0.130 0.145 0.178 0.249 0.078 0.122 0.138 0.118 0.196 5 UBC15 HERC2 UCH SKIP1 UCH CUL4 SNF2 UBC2 SKIP1 SKIP1 Stability value 0.140 0.131 0.174 0.182 0.254 0.079 0.123 0.148 0.136 0.237 6 UCH SamDC CUL4 CAP UBC2 UCH HERC2 SNF2 CUL4 CAP Stability value 0.173 0.144 0.185 0.211 0.263 0.097 0.124 0.149 0.150 0.240 7 ACT CAP SNF2 GAPDH CUL4 SKIP1 EF1beta ACT UBC15 UBC2 Stability value 0.188 0.154 0.246 0.325 0.263 0.107 0.130 0.170 0.157 0.261 8 GAPDH UBC2 CAP SNF2 ACT ACT UCH SKIP1 UCH GAPDH Stability value 0.189 0.159 0.259 0.336 0.263 0.110 0.138 0.186 0.169 0.268 9 SKIP11 UBC17 UBC15 ACT UBC17 HERC2 UBC2 UCH SNF2 UCH Stability value 0.194 0.173 0.264 0.348 0.298 0.136 0.198 0.194 0.190 0.289 10 SamDC SKIP1 EF1beta SKIP5 CAP SNF2 CAP HERC2 ACT UBC17 Stability value 0.212 0.234 0.308 0.387 0.325 0.139 0.200 0.203 0.237 0.296 11 SNF2 UCH UBC2 UBC2 GAPDH CAP SKIP1 GAPDH SamDC SNF2 Stability value 0.220 0.270 0.317 0.445 0.349 0.153 0.242 0.224 0.254 0.299 12 EF1beta CUL4 UBC17 CUL4 SKIP1 SamDC GAPDH SKIP5 CAP CUL4 Stability value 0.233 0.299 0.333 0.499 0.350 0.266 0.257 0.259 0.260 0.309 13 CAP SKIP5 SamDC UBC17 SNF2 SKIP5 SKIP5 SamDC SKIP5 SamDC Stability value 0.235 0.328 0.346 0.530 0.397 0.334 0.291 0.346 0.284 0.354 14 UBC2 SNF2 SKIP11 SamDC SamDC GAPDH SamDC CAP GAPDH SKIP5 Stability value 0.256 0.381 0.556 0.542 0.471 0.351 0.330 0.351 0.301 0.371 15 SKIP1 SKIP11 SKIP5 SKIP11 SKIP11 SKIP11 SKIP11 SKIP11 SKIP11 SKIP11 Stability value 0.311 0.761 0.728 0.882 0.807 0.850 0.367 0.411 0.567 0.681 doi:10.1371/journal.pone.0169465.t002 Table 2. Stabilities of candidate reference genes ranked by NormFinder. affected by different treatments (Fig 4), and peak points were observed at 3 h under drought, at 6 h under cold, at 3 h under salt, at 1 h under high pH, at 48 h under JA, and at 3 h under SA, whereas minimums were found at 24 h under heat and at 48 h under ABA and wounding treatments. It is worth noting that the transcripts of SgLEA were intensively induced by drought, cold, salt, high pH, and ABA treatments (Fig 5A–5E). Wounding and JA treatments caused transcript accumulation of SgLOX (Fig 5G and 5H), and heat stress increased the expression level of SgHSP90 (Fig 5F). An increased transcript level of SgPR1 was observed at 3 h after SA treatment compared with the control, and a peak point was observed at 12 h follow- ing a decrease in expression at 6 h (Fig 5I). Additionally, upregulated expression of SgPAL was observed after JA, high pH, and SA treatments, whereas SgPAL expression was downregulated by the other treatments (S2 Fig). Similar expression patterns were obtained using the most sta- ble single genes and combinations of two genes for normalization under the above treatments, while different expression patterns were observed when using the least stable reference genes compared with the most stable genes for target gene normalization. affected by different treatments (Fig 4), and peak points were observed at 3 h under drought, at 6 h under cold, at 3 h under salt, at 1 h under high pH, at 48 h under JA, and at 3 h under SA, whereas minimums were found at 24 h under heat and at 48 h under ABA and wounding treatments. It is worth noting that the transcripts of SgLEA were intensively induced by drought, cold, salt, high pH, and ABA treatments (Fig 5A–5E). Wounding and JA treatments caused transcript accumulation of SgLOX (Fig 5G and 5H), and heat stress increased the expression level of SgHSP90 (Fig 5F). An increased transcript level of SgPR1 was observed at 3 h after SA treatment compared with the control, and a peak point was observed at 12 h follow- ing a decrease in expression at 6 h (Fig 5I). Additionally, upregulated expression of SgPAL was observed after JA, high pH, and SA treatments, whereas SgPAL expression was downregulated by the other treatments (S2 Fig). PLOS ONE \| DOI:10.1371/journal.pone.0169465 January 5, 2017 Similar expression patterns were obtained using the most sta- ble single genes and combinations of two genes for normalization under the above treatments, while different expression patterns were observed when using the least stable reference genes compared with the most stable genes for target gene normalization. 9 / 18 Reference Genes Selection in Stipa grandis Table 3. Stabilities of candidate reference genes ranked by Bestkeeper. Rank Cold Drought Heat Salt High pH Wounding ABA JA SA total 1 UBC17 UCH GAPDH UCH SamDC UBC17 UBC17 UBC2 SKIP1 UCH SD 0.11 0.32 0.46 0.35 0.65 0.16 0.08 0.19 0.11 0.41 2 SKIP5 GAPDH EF 1beta HERC2 UCH CUL4 CUL4 EF 1beta UBC17 SamDC SD 0.13 0.47 0.48 0.37 0.76 0.22 0.11 0.21 0.12 0.52 3 SamDC SamDC UBC17 UBC15 UBC15 ACT HERC2 GAPDH HERC2 SNF2 SD 0.14 0.50 0.52 0.38 0.97 0.24 0.13 0.22 0.14 0.54 4 ACT UBC2 UCH EF 1beta HERC2 UBC2 UBC15 ACT EF 1beta ACT SD 0.16 0.55 0.54 0.41 1.00 0.25 0.13 0.23 0.15 0.54 5 UBC15 SKIP5 ACT SamDC SKIP5 UBC15 ACT UCH UBC15 EF 1beta SD 0.18 0.55 0.57 0.42 1.01 0.25 0.14 0.24 0.15 0.55 6 HERC2 ACT HERC2 SKIP5 UBC2 EF 1beta SNF2 UBC15 UBC2 SKIP5 SD 0.20 0.56 0.57 0.43 1.01 0.28 0.16 0.24 0.19 0.59 7 GAPDH CAP UBC2 CAP GAPDH SKIP1 CAP UBC17 UCH GAPDH SD 0.22 0.59 0.58 0.48 1.02 0.29 0.16 0.25 0.20 0.60 8 SKIP11 EF 1beta CUL4 SKIP1 SKIP1 SKIP5 UCH CUL4 GAPDH HERC2 SD 0.24 0.61 0.58 0.48 1.02 0.29 0.16 0.29 0.22 0.61 9 CUL4 SNF2 SKIP11 ACT CUL4 UCH EF 1beta SKIP11 ACT SKIP1 SD 0.25 0.63 0.59 0.52 1.04 0.30 0.20 0.33 0.24 0.67 10 UCH SKIP1 UBC15 GAPDH SNF2 SNF2 UBC2 SNF2 SNF2 UBC15 SD 0.26 0.65 0.60 0.53 1.05 0.33 0.22 0.34 0.25 0.67 11 SKIP1 UBC15 SNF2 UBC2 EF 1beta HERC2 SKIP1 SKIP5 SamDC UBC2 SD 0.29 0.69 0.61 0.56 1.07 0.34 0.25 0.35 0.30 0.67 12 SNF2 HERC2 SKIP1 SNF2 UBC17 CAP GAPDH SKIP1 CUL4 CAP SD 0.35 0.72 0.61 0.62 1.08 0.39 0.29 0.36 0.31 0.72 13 UBC2 UBC17 CAP UBC17 CAP GAPDH SKIP5 HERC2 SKIP5 UBC17 SD 0.36 0.74 0.65 0.66 1.08 0.41 0.32 0.41 0.32 0.73 14 CAP CUL4 SamDC CUL4 ACT SamDC SKIP11 SamDC CAP CUL4 SD 0.37 0.89 0.72 0.72 1.09 0.41 0.34 0.42 0.36 0.75 15 EF 1beta SKIP11 SKIP5 SKIP11 SKIP11 SKIP11 SamDC CAP SKIP11 SKIP11 SD 0.40 1.02 0.91 0.80 1.19 0.94 0.36 0.56 0.63 1.04 SD represents standard deviation doi 10 1371/journal pone 0169465 t003 Table 3. Stabilities of candidate reference genes ranked by Bestkeeper. PLOS ONE \| DOI:10.1371/journal.pone.0169465 January 5, 2017 Reference Genes Selection in Stipa grandis Fig 4. Validation of the most and least stable reference genes for the normalization of SgPOD expression. The most and least stable reference genes were selected for normalization following different stress treatments for 0, 1, 3, 6, 12, 24, and 48 h, with an additional 0.5-h time point for the drought and wounding treatments. (A) Drought. (B) Cold. (C) Salt. (D) High pH. (E) ABA. (F) Heat. (G) Wounding. (H) JA. (I) SA. The error bars represent the mean of three technical replicates ± SD. the reference gene chosen for normalization of gene expression data [22]; therefore, it is crucial to select stably expressed genes as internal references [13]. Base on previous transcriptome data [8], we selected 15 genes that exhibited relatively low levels of transcript variation after grazing stress and evaluated their expression stability under different stresses in S. grandis. Four statistical algorithms, geNorm, NormFinder, BestKeeper, and RefFinder, were used to evaluate the stability of these genes. All the genes displayed differ- ent expression patterns following the various treatments. Moreover, high stability levels were observed using geNorm analysis; all the M values were below the accepted threshold of 1.5 (Fig 2), which indicated that all the genes could be used as internal control genes for normalization. However, according to different algorithms and analytical procedures, several differences were observed in the ranking of the most stable genes, which was similar to the results for other Fig 4. Validation of the most and least stable reference genes for the normalization of SgPOD expression. The most and least stable reference genes were selected for normalization following different stress treatments for 0, 1, 3, 6, 12, 24, and 48 h, with an additional 0.5-h time point for the drought and wounding treatments. (A) Drought. (B) Cold. (C) Salt. (D) High pH. (E) ABA. (F) Heat. (G) Wounding. (H) JA. (I) SA. The error bars represent the mean of three technical replicates ± SD. doi:10.1371/journal.pone.0169465.g004 PLOS ONE \| DOI:10.1371/journal.pone.0169465 January 5, 2017 11 / 18 Fig 4. Validation of the most and least stable reference genes for the normalization of SgPOD expression. The most and least stable reference genes were selected for normalization following different stress treatments for 0, 1, 3, 6, 12, 24, and 48 h, with an additional 0.5-h time point for the drought and wounding treatments. (A) Drought. (B) Cold. (C) Salt. (D) High pH. Discussion qRT-PCR combines reverse transcription (RT) with the quantitative real-time polymerase chain reaction (qPCR) [35]. The accuracy of qRT-PCR is influenced strongly by the stability of qRT-PCR combines reverse transcription (RT) with the quantitative real-time polymerase chain reaction (qPCR) [35]. The accuracy of qRT-PCR is influenced strongly by the stability of Table 4. Stabilities of candidate reference genes ranked by RefFinder. Drought Heat Cold Salt High pH Wounding JA ABA SA Total Best EF1beta ACT SKIP5 UBC15 UBC15 UBC2 EF1beta UBC15 HERC2 EF1beta UBC15 GAPDH UBC17 EF1beta HERC2 UBC15 UBC17 CUL4 UBC17 HERC2 GAPDH HERC2 HERC2 HERC2 EF1beta UBC17 UBC15 UBC17 EF1beta UCH HERC2 SKIP1 UBC15 UCH CUL4 UCH UBC2 EF1beta SKIP1 ACT SamDC UCH ACT SKIP1 UBC17 CUL4 CUL4 SNF2 UBC2 UBC15 Worst SKIP11 SKIP5 SKIP1 SKIP11 SKIP11 SKIP11 CAP SKIP11 SKIP11 SKIP11 doi:10.1371/journal.pone.0169465.t004 Table 4. Stabilities of candidate reference genes ranked by RefFinder. PLOS ONE \| DOI:10.1371/journal.pone.0169465 January 5, 2017 10 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0169465 January 5, 2017 (E) ABA. (F) Heat. (G) Wounding. (H) JA. (I) SA. The error bars represent the mean of three technical replicates ± SD. doi:10.1371/journal.pone.0169465.g004 the reference gene chosen for normalization of gene expression data [22]; therefore, it is crucial to select stably expressed genes as internal references [13]. Base on previous transcriptome data [8], we selected 15 genes that exhibited relatively low levels of transcript variation after grazing stress and evaluated their expression stability under different stresses in S. grandis. Four statistical algorithms, geNorm, NormFinder, BestKeeper, and RefFinder, were used to evaluate the stability of these genes. All the genes displayed differ- ent expression patterns following the various treatments. Moreover, high stability levels were observed using geNorm analysis; all the M values were below the accepted threshold of 1.5 (Fig 2), which indicated that all the genes could be used as internal control genes for normalization. However, according to different algorithms and analytical procedures, several differences were observed in the ranking of the most stable genes, which was similar to the results for other PLOS ONE \| DOI:10.1371/journal.pone.0169465 January 5, 2017 11 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0169465 January 5, 2017 Reference Genes Selection in Stipa grandis Fig 5. Validation of the most and least stable reference genes for the normalization of SgLEA, SgLOX, SgHSP90, and SgPR1 expression. The most and least stable reference genes were selected for normalization following different stress treatments for 0, 1, 3, 6, 12, 24, and 48 h, with an additional 0.5-h time point for the drought and wounding treatments. (A–E) SgLEA expression under drought, cold, salt, high pH, and ABA treatments. (F) SgHSP90 expression under heat treatment. (G and H) SgLOX expression under wounding and JA treatments. (I) SgPR1 expression under SA treatment. The error bars represent the mean of three technical replicates ± SD. doi:10.1371/journal.pone.0169465.g005 Fig 5. Validation of the most and least stable reference genes for the normalization of SgLEA, SgLOX, SgHSP90, and SgPR1 expression. The most and least stable reference genes were selected for normalization following different stress treatments for 0, 1, 3, 6, 12, 24, and 48 h, with an additional 0.5-h time point for the drought and wounding treatments. (A–E) SgLEA expression under drought, cold, salt, high pH, and ABA treatments. (F) SgHSP90 expression under heat treatment. (G and H) SgLOX expression under wounding and JA treatments. (I) SgPR1 expression under SA treatment. The error bars represent the mean of three technical replicates ± SD. doi:10.1371/journal.pone.0169465.g005 species [19–21]. For wounding treatment in S. grandis, UBC2 and UBC15 were selected by both geNorm and NormFinder as the two most stable reference genes (Fig 2 and Table 2), but UBC17 and CUL4 were ranked top by Bestkeeper (Table 3); UBC2, UBC15, and UBC17 were determined as the most stable by RefFinder’s comprehensive analysis (Table 4). Across all treatments under study, UBC17, CUL4, and HERC2 for geNorm (Fig 2), UBC15, HERC2, and EF1beta for NormFinder (Table 2), and UCH, SamDC, and SNF2 for Bestkeeper were consid- ered the most suitable reference genes (Table 3). Overall, RefFinder recommended using EF1beta, HERC2, and UCH (Table 4). Traditional reference genes, such as ACT, GAPDH, and EF1α, which are involved in pri- mary metabolism or other cellular processes [19], as well as new candidate reference genes 12 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0169465 January 5, 2017 Reference Genes Selection in Stipa grandis derived from public databases [12, 17, 22, 25, 26], have been validated to allow precise inter- pretation of the qRT-PCR results for specific experimental conditions. However, fluctuating expression of reference genes has been observed under certain experimental conditions. For example, EF1α was identified as the best reference gene under salinity and drought stresses in Glycine max [36]. Nevertheless, EF1α exhibited poor performance in leaves and roots during abiotic stress treatments [15]. In the present study, we also found that the most suitable refer- ence genes varied under different environmental stimuli. The comprehensive combined analy- sis by RefFinder showed that the top three ranked genes included the traditional genes EF1beta, ACT, UBC15, UBC2, GAPDH, and UBC17, as well as the novel genes UCH, SKIP5, HERC2, and CUL4 (Table 4). Previously, EF1beta was identified as a stable reference gene in the developing seeds of the Tung Tree [37]. EF1beta has also been determined to be a good reference in soybean under various stresses [36]. Similarly, we found that EF1beta was in the three top ranked reference genes for drought, salt, high pH, JA, and SA treatments, as well as across all samples (Table 4). ACT, a traditional reference gene, is used commonly as an internal control to normalize target gene expression results from qRT-PCR in various species under different conditions. For instance, ACT was determined to be stably expressed during different floral developmental stages and different temperature treatments in Sweet Osmanthus (Osmanthus fragrans Lour.) [14]. In our study, ACT was identified as the best reference gene by the geNorm and RefFinder programs under heat stress (Fig 2 and Table 4). Ubiquitination is related to post-translational modification of proteins, and three enzymes are involved in this cellular process: E1 (ubiquitin-activating enzymes), E2 (ubiquitin-conju- gating enzymes), and E3 (ubiquitin ligases) [38]. Accumulated data indicate that ubiquitin- conjugating enzymes (UBC) exhibit stable expression in many experimental conditions; thus, they have been used widely as reference genes in different species [12, 14, 39]. Three UBC genes (UBC15, UBC2, and UBC17) were analyzed in this study. Notably, these UBC genes per- formed well under different treatments (excluding heat stress), especially when analyzed by geNorm and RefFinder (Fig 2 and Table 4). UBC15 and UBC17 exhibited better expression sta- bility than UBC2. HERC2, a HECT domain E3 (ubiquitin ligases) encoding gene [40], also showed stable transcript levels under most experimental conditions. Additionally, the de-ubi- quitinating protease UCH (ubiquitin carboxyl-terminal hydrolase) was the top ranked gene using Bestkeeper analysis under drought and salt treatments, as well as in the overall assess- ment (Table 3). PLOS ONE \| DOI:10.1371/journal.pone.0169465 January 5, 2017 Taken together, at least one of the ubiquitin-related genes was ranked in the top three stable genes (Fig 2, Tables 3 and 4). These results indicated the UBC genes could be used more often as an internal control for normalizing gene expression under various stress treatments. SKIP, a spliceosome component and transcriptional activator, is conserved in eukaryotes and has diverse functions in plant development associated with responses to envi- ronmental changes [41]. Among the three SKIP encoding genes examined in this study, SKIP5 was ranked as the most stable gene under cold stress using geNorm (Fig 2), NormFinder (Table 2), and RefFinder (Table 4), and SKIP1 was recommended as the best gene under JA treatment by Bestkeeper analysis (Table 3). Conversely, SKIP11 was ranked the least stable ref- erence gene by the different programs (Fig 2 and Tables 2–4), except for heat and cold stresses by geNorm and NormFinder, heat, cold, ABA, and JA treatments by Bestkeeper, and heat, cold, and JA treatments by RefFinder. These results indicate that SKIP11 is the least suitable gene for use as an internal control in S. grandis under various environmental stimuli. The effectiveness of the most stable reference genes selected using RefFinder was evaluated by normalization of six target genes in samples under different stress conditions. Peroxidases [POD/POX, EC:1.11.1.7] are heme-containing glycoproteins that oxidize a wide range of hydrogen donors, accompanied by the accumulation of H2O2 [42]. Many peroxidase PLOS ONE \| DOI:10.1371/journal.pone.0169465 January 5, 2017 13 / 18 Reference Genes Selection in Stipa grandis isoenzymes are found in higher plants and they play roles in plant defense, growth, and devel- opment [42]. Guaiacol peroxidase (GPX, EC 1.11.1.7) [43], which is generally considered a stress-related enzyme, is involved in plant defense against various stresses, and the activity of GPX is thought to be induced by diverse environmental stimuli [44], such as drought, salinity, and chilling [43, 45, 46]. In the present work, the expression of a peroxidase encoding gene was detected under the different treatments (Fig 4), and the changes in its transcript levels sug- gested that peroxidase plays an important role in S. grandis defense against stressful conditions. The late embryogenesis abundant protein encoding gene LEA has been widely reported to par- ticipate in responses to drought, cold, and salt stresses, and ABA treatment, and plays an important role in plant tolerance to abiotic stress [47–51]. Conclusions This study comprehensively evaluated the utility of 15 potential reference genes for quantifying transcript levels in S. grandis exposed to diverse experimental conditions. The stable reference genes identified in this work will help to improve the accuracy of gene expression analysis and will aid the investigation of stress-response genes and the molecular mechanism of stress toler- ance in this species. Supporting Information S1 Fig. Primer specificity and amplicon size for the candidate reference genes. (TIF) S1 Fig. Primer specificity and amplicon size for the candidate reference genes. (TIF) S2 Fig. Validation of the most and least stable reference genes for the normalization of SgPAL expression under different treatments. (TIF) S1 Text. Sequences of candidate reference genes. (TXT) S1 Text. Sequences of candidate reference genes. (TXT) The significantly increased expres- sion of SgLEA under drought, cold, salt, high pH and ABA treatments (Fig 5A–5E) implies that SgLEA participates in S. grandis resistance responses against stresses. The expression level of SgLOX was elevated by wounding and JA treatments (Fig 5G and 5H), and SgHSP90 expres- sion was enhanced by heat treatment (Fig 5F); the induction of these genes was in accordance with previously reported expression patterns of LOX2 [52–54] and HSP90 [55–57]. Addition- ally, a pathogenesis-related protein 1 encoding gene, SgPR1, which is widely accepted as an SA marker gene [58] and is induced by SA accumulation [59], was selected to assess the utility of the identified reference genes under SA treatment. As a critical defense signaling molecule, SA accumulation is associated with enhanced resistance to pathogen infection [60]. The changes in the transcript level of SgPR1 after SA treatment indicated that SgPR1 could be used as a molecular marker for disease resistance responses in S. grandis (Fig 5I). However, when the target gene normalization was performed with the least stable genes, the expression patterns varied in comparison with most stable reference genes. Taken together, these results suggest that appropriate selection of reference genes is necessary for the normalization of target gene expression to ensure the accuracy of qRT-PCR data. References 1. Wu JB, Gao YB, Bao XY, Gao H, Jia MQ, Li J, et al. Genetic variation among Stipa grandis P. Smirn populations with different durations of fencing in the Inner Mongolian Steppe. Rangeland J. 2010; 32(4): 427–434. 2. Song XL, Wang YH and Lv XM. Responses of plant biomass, photosynthesis and lipid peroxidation to warming and precipitation change in two dominant species (Stipa grandis and Leymus chinensis) from North China Grasslands. Ecology and Evolution. 2016; 6(6): 1871–1882. doi: 10.1002/ece3.1982 PMID: 26933491 3. Zhao NX, Gao YB, Wang JL and Ren AZ. Population structure and genetic diversity of Stipa grandis P. Smirn, a dominant species in the typical steppe of northern China. Biochemical Systematics and Ecol- ogy. 2008; 36(1): 1–10. 4. Xu Z and Zhou G. Responses of photosynthetic capacity to soil moisture gradient in perennial rhizome grass and perennial bunchgrass. BMC Plant Biol. 2011; 11: 21. doi: 10.1186/1471-2229-11-21 PMID: 21266062 5. Ren H, Schonbach P, Wan H, Gierus M and Taube F. Effects of grazing intensity and environmental factors on species composition and diversity in typical steppe of Inner Mongolia, China. PLoS One. 2012; 7(12): e52180. doi: 10.1371/journal.pone.0052180 PMID: 23284925 6. Xu Z, Shimizu H, Ito S, Yagasaki Y, Zou C, Zhou G, et al. Effects of elevated CO(2), warming and pre- cipitation change on plant growth, photosynthesis and peroxidation in dominant species from North China grassland. Planta. 2014; 239(2): 421–435. doi: 10.1007/s00425-013-1987-9 PMID: 24463932 7. Wu JB, Gao YB, Bao XY, Gao H, Jia MQ, Li J, et al. Genetic diversity of Stipa grandis P Smirn popula- tions across the species’ range in the Inner Mongolia Plateau of China. Biochemical Systematics and Ecology. 2010; 38(4): 471–477. 8. Wan D, Wan Y, Hou X, Ren W, Ding Y and Sa R. De novo assembly and transcriptomic profiling of the grazing response in Stipa grandis. PLoS One. 2015; 10(4): e0122641. doi: 10.1371/journal.pone. 0122641 PMID: 25875617 9. Udvardi MK, Czechowski T and Scheible WR. Eleven golden rules of quantitative RT-PCR. Plant Cell. 2008; 20(7): 1736–1737. doi: 10.1105/tpc.108.061143 PMID: 18664613 10. VanGuilder HD, Vrana KE and Freeman WM. Twenty-five years of quantitative PCR for gene expres- sion analysis. Biotechniques. 2008; 44(5): 619–626. doi: 10.2144/000112776 PMID: 18474036 11. Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N, De Paepe A, et al. Accurate normaliza- tion of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol. Author Contributions Author Contributions Conceptualization: DLW YQW XYH. Formal analysis: DLW YQW QY. Funding acquisition: XYH DLW. Investigation: YQW DLW BZ KW. Project administration: XYH DLW. Resources: DLW RGW WBR YD. Validation: DLW YQW. Visualization: DLW YQW ZW. Writing – original draft: DLW. Writing – review & editing: DLW XYH Y Writing – review & editing: DLW XYH YQW. Acknowledgments We thank Dr. Guojing Li (Inner Mongolia Agricultural University) for her valuable sugges- tions on the design of the experiment and the Plant Molecular Biology Laboratory of Inner Mongolia Agricultural University for help with the qRT-PCR analysis. PLOS ONE \| DOI:10.1371/journal.pone.0169465 January 5, 2017 14 / 18 Reference Genes Selection in Stipa grandis Author Contributions Conceptualization: DLW YQW XYH. Formal analysis: DLW YQW QY. Funding acquisition: XYH DLW. Investigation: YQW DLW BZ KW. Project administration: XYH DLW. Resources: DLW RGW WBR YD. Validation: DLW YQW. Visualization: DLW YQW ZW. Writing – original draft: DLW. Writing – review & editing: DLW XYH PLOS ONE \| DOI:10.1371/journal.pone.0169465 January 5, 2017 References 2002; 3(7): RESEARCH0034. 12. Hong SY, Seo PJ, Yang MS, Xiang F and Park CM. Exploring valid reference genes for gene expres- sion studies in Brachypodium distachyon by real-time PCR. BMC Plant Biol. 2008; 8: 112. doi: 10.1186/ 1471-2229-8-112 PMID: 18992143 15 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0169465 January 5, 2017 Reference Genes Selection in Stipa grandis 13. Czechowski T, Stitt M, Altmann T, Udvardi MK and Scheible WR. Genome-wide identification and test- ing of superior reference genes for transcript normalization in Arabidopsis. Plant Physiol. 2005; 139(1): 5–17. doi: 10.1104/pp.105.063743 PMID: 16166256 14. Zhang C, Fu J, Wang Y, Bao Z and Zhao H. Identification of Suitable Reference Genes for Gene Expression Normalization in the Quantitative Real-Time PCR Analysis of Sweet Osmanthus (Osmanthus fragrans Lour.). PLoS One. 2015; 10(8): e0136355. doi: 10.1371/journal.pone.0136355 PMID: 26302211 15. Yang Z, Chen Y, Hu B, Tan Z and Huang B. Identification and validation of reference genes for quantifi- cation of target gene expression with quantitative real-time PCR for tall fescue under four abiotic stresses. PLoS One. 2015; 10(3): e0119569. doi: 10.1371/journal.pone.0119569 PMID: 25786207 16. Muller OA, Grau J, Thieme S, Prochaska H, Adlung N, Sorgatz A, et al. Genome-Wide Identification and Validation of Reference Genes in Infected Tomato Leaves for Quantitative RT-PCR Analyses. PLoS One. 2015; 10(8): e0136499. doi: 10.1371/journal.pone.0136499 PMID: 26313760 17. Ling H, Wu QB, Guo JL, Xu LP and Que YX. Comprehensive Selection of Reference Genes for Gene Expression Normalization in Sugarcane by Real Time Quantitative RT-PCR. PLoS One. 2014; 9(5). 18. Kong Q, Yuan J, Niu P, Xie J, Jiang W, Huang Y, et al. Screening suitable reference genes for normali- zation in reverse transcription quantitative real-time PCR analysis in melon. PLoS One. 2014; 9(1): e87197. doi: 10.1371/journal.pone.0087197 PMID: 24475250 19. Chen Y, Hu B, Tan Z, Liu J, Yang Z, Li Z, et al. Selection of reference genes for quantitative real-time PCR normalization in creeping bentgrass involved in four abiotic stresses. Plant Cell Rep. 2015; 34(10): 1825–1834. doi: 10.1007/s00299-015-1830-9 PMID: 26179072 20. Lacerda AL, Fonseca LN, Blawid R, Boiteux LS, Ribeiro SG and Brasileiro AC. Reference Gene Selec- tion for qPCR Analysis in Tomato-Bipartite Begomovirus Interaction and Validation in Additional Tomato-Virus Pathosystems. PLoS One. 2015; 10(8): e0136820. doi: 10.1371/journal.pone.0136820 PMID: 26317870 21. Bansal R, Mittapelly P, Cassone BJ, Mamidala P, Redinbaugh MG and Michel A. PLOS ONE \| DOI:10.1371/journal.pone.0169465 January 5, 2017 References Pfaffl MW, Tichopad A, Prgomet C and Neuvians TP. Determination of stable housekeeping genes, dif- ferentially regulated target genes and sample integrity: BestKeeper—Excel-based tool using pair-wise correlations. Biotechnol Lett. 2004; 26(6): 509–515. PMID: 15127793 33. Xie F, Xiao P, Chen D, Xu L and Zhang B. miRDeepFinder: a miRNA analysis tool for deep sequencing of plant small RNAs. Plant Mol Biol. 2012. 34. Livak KJ and Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2−ΔΔCT method. Methods. 2001; 25(4): 402–408. doi: 10.1006/meth.2001.1262 PMID: 11846609 35. Bennett J, Hondred D and Register JC III. Keeping qRT-PCR rigorous and biologically relevant. Plant Cell Rep. 2015; 34(1): 1–3. doi: 10.1007/s00299-014-1692-6 PMID: 25304620 36. Ma S, Niu H, Liu C, Zhang J, Hou C and Wang D. Expression stabilities of candidate reference genes for RT-qPCR under different stress conditions in soybean. PLoS One. 2013; 8(10): e75271. doi: 10. 1371/journal.pone.0075271 PMID: 24124481 37. Han XJ, Lu MZ, Chen YC, Zhan ZY, Cui QQ and Wang YD. Selection of Reliable Reference Genes for Gene Expression Studies Using Real-Time PCR in Tung Tree during Seed Development. PLoS One. 2012; 7(8). 38. Morreale FE and Walden H. Types of Ubiquitin Ligases. Cell. 2016; 165(1): 248–248 e241. doi: 10. 1016/j.cell.2016.03.003 PMID: 27015313 39. Chang E, Shi S, Liu J, Cheng T, Xue L, Yang X, et al. Selection of reference genes for quantitative gene expression studies in Platycladus orientalis (Cupressaceae) Using real-time PCR. PLoS One. 2012; 7 (3): e33278. doi: 10.1371/journal.pone.0033278 PMID: 22479379 40. Mohiuddin, Kobayashi S, Keka IS, Guilbaud G, Sale J, Narita T, et al. The role of HERC2 and RNF8 ubi- quitin E3 ligases in the promotion of translesion DNA synthesis in the chicken DT40 cell line. DNA Repair (Amst). 2016; 40: 67–76. 41. Li Y, Xia C, Feng J, Yang D, Wu F, Cao Y, et al. The SNW Domain of SKIP is Required for its Integration into the Spliceosome and its Interaction with the Paf1 Complex in Arabidopsis. Mol Plant. 2016. 42. Yoshida K, Kaothien P, Matsui T, Kawaoka A and Shinmyo A. Molecular biology and application of plant peroxidase genes. Appl Microbiol Biotechnol. 2003; 60(6): 665–670. doi: 10.1007/s00253-002- 1157-7 PMID: 12664144 43. Prasad TK, Anderson MD, Martin BA and Stewart CR. Evidence for Chilling-Induced Oxidative Stress in Maize Seedlings and a Regulatory Role for Hydrogen Peroxide. Plant Cell. 1994; 6(1): 65–74. doi: 10.1105/tpc.6.1.65 PMID: 12244221 44. References Recommended Refer- ence Genes for Quantitative PCR Analysis in Soybean Have Variable Stabilities during Diverse Biotic Stresses. PLoS One. 2015; 10(8): e0134890. doi: 10.1371/journal.pone.0134890 PMID: 26244340 22. Yang Q, Yin J, Li G, Qi L, Yang F and Wang R. Reference gene selection for qRT-PCR in Caragana kor- shinskii Kom. under different stress conditions. Mol Biol Rep. 2014; 41(4): 2325–2334. doi: 10.1007/ s11033-014-3086-9 PMID: 24452712 23. Lin LY, Han XJ, Chen YC, Wu QK and Wang YD. Identification of appropriate reference genes for nor- malizing transcript expression by quantitative real-time PCR in Litsea cubeba. Mol Genet Genomics. 2013; 288(12): 727–737. doi: 10.1007/s00438-013-0785-1 PMID: 24162560 24. Lovdal T and Lillo C. Reference gene selection for quantitative real-time PCR normalization in tomato subjected to nitrogen, cold, and light stress. Anal Biochem. 2009; 387(2): 238–242. doi: 10.1016/j.ab. 2009.01.024 PMID: 19454243 25. Gimeno J, Eattock N, Van Deynze A and Blumwald E. Selection and Validation of Reference Genes for Gene Expression Analysis in Switchgrass (Panicum virgatum) Using Quantitative Real-Time RT-PCR. PLoS One. 2014; 9(3). 26. Chandna R, Augustine R and Bisht NC. Evaluation of Candidate Reference Genes for Gene Expression Normalization in Brassica juncea Using Real Time Quantitative RT-PCR. PLoS One. 2012; 7(5). 27. Gutierrez L, Mauriat M, Pelloux J, Bellini C and van Wuytswinkel O. Towards a systematic validation of references in real-time RT-PCR. Plant Cell. 2008; 20(7): 1734–1735. doi: 10.1105/tpc.108.059774 PMID: 18664615 28. Zhang Y and Turner JG. Wound-induced endogenous jasmonates stunt plant growth by inhibiting mito- sis. PLoS One. 2008; 3(11): e3699. doi: 10.1371/journal.pone.0003699 PMID: 19002244 29. Han X, Feng Z, Xing D, Yang Q, Wang R, Qi L, et al. Two NAC transcription factors from Caragana intermedia altered salt tolerance of the transgenic Arabidopsis. BMC Plant Biol. 2015; 15: 208. doi: 10. 1186/s12870-015-0591-5 PMID: 26297025 30. Wan DL, Li RL, Zou B, Zhang X, Cong JY, Wang RG, et al. Calmodulin-binding protein CBP60g is a positive regulator of both disease resistance and drought tolerance in Arabidopsis. Plant Cell Rep. 2012; 31(7): 1269–1281. doi: 10.1007/s00299-012-1247-7 PMID: 22466450 31. Andersen CL, Jensen JL and Orntoft TF. Normalization of real-time quantitative reverse transcription- PCR data: a model-based variance estimation approach to identify genes suited for normalization, applied to bladder and colon cancer data sets. Cancer Res. 2004; 64(15): 5245–5250. doi: 10.1158/ 0008-5472.CAN-04-0496 PMID: 15289330 16 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0169465 January 5, 2017 Reference Genes Selection in Stipa grandis 32. PLOS ONE \| DOI:10.1371/journal.pone.0169465 January 5, 2017 References Sharma P, Jha AB, Dubey RS and Pessarakli M. Reactive oxygen species, oxidative damage, and anti- oxidative defense mechanism in plants under stressful conditions. Journal of Botany. 2012;2012. 45. Sharma P and Dubey RS. Drought induces oxidative stress and enhances the activities of antioxidant enzymes in growing rice seedlings. Plant Growth Regul. 2005; 46(3): 209–221. 46. Mishra P, Bhoomika K and Dubey RS. Differential responses of antioxidative defense system to pro- longed salinity stress in salt-tolerant and salt-sensitive Indica rice (Oryza sativa L.) seedlings. Proto- plasma. 2013; 250(1): 3–19. doi: 10.1007/s00709-011-0365-3 PMID: 22194018 47. Sharma A, Kumar D, Kumar S, Rampuria S, Reddy AR and Kirti PB. Ectopic Expression of an Atypical Hydrophobic Group 5 LEA Protein from Wild Peanut, Arachis diogoi Confers Abiotic Stress Tolerance in Tobacco. PLoS One. 2016; 11(3): e0150609. doi: 10.1371/journal.pone.0150609 PMID: 26938884 48. Liu Y, Liang J, Sun L, Yang X and Li D. Group 3 LEA Protein, ZmLEA3, Is Involved in Protection from Low Temperature Stress. Front Plant Sci. 2016; 7: 1011. doi: 10.3389/fpls.2016.01011 PMID: 27471509 49. Liu Y, Wang L, Xing X, Sun L, Pan J, Kong X, et al. ZmLEA3, a multifunctional group 3 LEA protein from maize (Zea mays L.), is involved in biotic and abiotic stresses. Plant Cell Physiol. 2013; 54(6): 944–959. doi: 10.1093/pcp/pct047 PMID: 23543751 50. Babu RC, Zhang JX, Blum A, Ho THD, Wu R and Nguyen HT. HVA1, a LEA gene from barley confers dehydration tolerance in transgenic rice (Oryza sativa L.) via cell membrane protection. Plant Science. 2004; 166(4): 855–862. 51. Xu DP, Duan XL, Wang BY, Hong BM, Ho THD and Wu R. Expression of a late embryogenesis abun- dant protein gene, HVA1, from barley confers tolerance to water deficit and salt stress in transgenic rice. Plant Physiol. 1996; 110(1): 249–257. PMID: 12226181 52. Wang CX, Zien CA, Afitlhile M, Welti R, Hildebrand DF and Wang XM. Involvement of phospholipase D in wound-induced accumulation of jasmonic acid in Arabidopsis. Plant Cell. 2000; 12(11): 2237–2246. PMID: 11090221 53. Reymond P, Weber H, Damond M and Farmer EE. Differential gene expression in response to mechan- ical wounding and insect feeding in Arabidopsis. Plant Cell. 2000; 12(5): 707–720. PMID: 10810145 17 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0169465 January 5, 2017 Reference Genes Selection in Stipa grandis 54. Leon-Reyes A, Van der Does D, De Lange ES, Delker C, Wasternack C, Van Wees SC, et al. PLOS ONE \| DOI:10.1371/journal.pone.0169465 January 5, 2017 60. Vlot AC, Dempsey DA and Klessig DF. Salicylic Acid, a Multifaceted Hormone to Combat Disease. Annu Rev Phytopathol. 2009; 47: 177–206. doi: 10.1146/annurev.phyto.050908.135202 PMID: 19400653 References Salicy- late-mediated suppression of jasmonate-responsive gene expression in Arabidopsis is targeted down- stream of the jasmonate biosynthesis pathway. Planta. 2010; 232(6): 1423–1432. doi: 10.1007/s00425- 010-1265-z PMID: 20839007 54. Leon-Reyes A, Van der Does D, De Lange ES, Delker C, Wasternack C, Van Wees SC, et al. Salicy- late-mediated suppression of jasmonate-responsive gene expression in Arabidopsis is targeted down- stream of the jasmonate biosynthesis pathway. Planta. 2010; 232(6): 1423–1432. doi: 10.1007/s00425- 010-1265-z PMID: 20839007 55. Milioni D and Hatzopoulos P. Genomic organization of hsp90 gene family in Arabidopsis. Plant Mol Biol. 1997; 35(6): 955–961. PMID: 9426614 56. Xu J, Xue C, Xue D, Zhao J, Gai J, Guo N, et al. Overexpression of GmHsp90s, a heat shock protein 90 (Hsp90) gene family cloning from soybean, decrease damage of abiotic stresses in Arabidopsis thali- ana. PLoS One. 2013; 8(7): e69810. doi: 10.1371/journal.pone.0069810 PMID: 23936107 57. Yabe N, Takahashi T and Komeda Y. Analysis of tissue-specific expression of Arabidopsis thaliana HSP90-family gene HSP81. Plant Cell Physiol. 1994; 35(8): 1207–1219. PMID: 7697294 58. Fan J, Hill L, Crooks C, Doerner P and Lamb C. Abscisic Acid Has a Key Role in Modulating Diverse Plant-Pathogen Interactions. Plant Physiology. 2009; 150(4): 1750–1761. doi: 10.1104/pp.109.137943 PMID: 19571312 59. Loake G and Grant M. Salicylic acid in plant defence-the players and protagonists. Curr Opin Plant Biol 2007; 10(5): 466–472. doi: 10.1016/j.pbi.2007.08.008 PMID: 17904410 60. Vlot AC, Dempsey DA and Klessig DF. Salicylic Acid, a Multifaceted Hormone to Combat Disease. Annu Rev Phytopathol. 2009; 47: 177–206. doi: 10.1146/annurev.phyto.050908.135202 PMID: 19400653 18 / 18
W2897412005.txt	https://www.nature.com/articles/s41598-018-33892-z.pdf	en		Nanogel tectonic porous 3D scaffold for direct reprogramming fibroblasts into osteoblasts and bone regeneration	Scientific reports	2,018	cc-by	6,108	www.nature.com/scientificreports OPEN Received: 5 April 2018 Accepted: 8 October 2018 Published: xx xx xxxx Nanogel tectonic porous 3D scaffold for direct reprogramming fibroblasts into osteoblasts and bone regeneration Yoshiki Sato1,2, Kenta Yamamoto1,2, Satoshi Horiguchi1,2, Yoshiro Tahara3, Kei Nakai1,2, Shin-ichiro Kotani1, Fumishige Oseko2, Giuseppe Pezzotti1,4, Toshiro Yamamoto2, Tsunao Kishida1, Narisato Kanamura2, Kazunari Akiyoshi3 & Osam Mazda 1 Transplantation of engineered three-dimensional (3D) bone tissue may provide therapeutic benefits to patients with various bone diseases. To achieve this goal, appropriate 3D scaffolds and cells are required. In the present study, we devised a novel nanogel tectonic material for artificial 3D scaffold, namely the nanogel-cross-linked porous (NanoCliP)-freeze-dried (FD) gel, and estimated its potential as a 3D scaffold for bone tissue engineering. As the osteoblasts, directly converted osteoblasts (dOBs) were used, because a large number of highly functional osteoblasts could be induced from fibroblasts that can be collected from patients with a minimally invasive procedure. The NanoCliP-FD gel was highly porous, and fibronectin coating of the gel allowed efficient adhesion of the dOBs, so that the cells occupied the almost entire surface of the walls of the pores after culturing for 7 days. The dOBs massively produced calcified bone matrix, and the culture could be continued for at least 28 days. The NanoCliP-FD gel with dOBs remarkably promoted bone regeneration in vivo after having been grafted to bone defect lesions that were artificially created in mice. The present findings suggest that the combination of the NanoCliP-FD gel and dOBs may provide a feasible therapeutic modality for bone diseases. A large number of elderlies suffer from bone diseases that are associated with bone resorption and low ability of bone remodeling. Non-union of the fractured bone after osteoporotic fracture at a spine or a femoral neck may cause locomotive disability or continuing bedridden state. Surgical removal of a bone tumor may seriously reduce the quality of life (QOL) and activity of daily living (ADL) of the patients. Alveolar bone resorption due to periodontal disease may cause loss of teeth, potentially resulting in systemic infection, eating disturbance, malnutrition, and dementia1,2. Autologous bone graft has been performed to treat a patient with severe bone loss, but the preparation of a sufficient amount of bone for transplantation may cause adverse events such as pain in the patients at the sacrificed donor site3. In this context, an efficient regenerative therapy to treat bone loss and bone resorption disorders is strongly needed. To realize an effective regenerative therapy for bone diseases, it is important to build a three-dimensional bone tissue in culture by combining two elements: (i) a 3D scaffold that could be built in a tailor-made manner to have the specific size and shape of the bone loss lesion in each patient, and (ii) a considerable number of autologous osteoblasts with a high bone forming ability. The 3D scaffold should also have appropriate chemical and physical features so that the osteoblasts efficiently settle onto the scaffold and produce bone tissue. Recently, functional hydrogels have been developed and the hydrogel scaffolds induced cellular adhesion and/or function4. The nanogel is nanometer-sized hydrogel which consists of a naturally occurring polysaccharide, pullulan, modified with a cholesterol moiety5. The nanogel is biodegradable, and safely introduced into human. The nanogel has been reported to provide a feasible drug delivery system that enables efficient transfer of short interfering 1 Department of Immunology, Kyoto Prefectural University of Medicine, Kyoto, Japan. 2Department of Dental Medicine, Kyoto Prefectural University of Medicine, Kyoto, Japan. 3Department of Polymer Chemistry, Graduate school of engineering, Kyoto University, Kyoto, Japan. 4Ceramic Physics Laboratory, Kyoto Institute of Technology, Kyoto, Japan. Correspondence and requests for materials should be addressed to K.A. (email: akiyoshi@bio.polym. kyoto-u.ac.jp) or O.M. (email: mazda@koto.kpu-m.ac.jp) SCIenTIfIC REPOrTS \| (2018) 8:15824 \| DOI:10.1038/s41598-018-33892-z 1 www.nature.com/scientificreports/ RNA into tumor cells6, slow-release of a cytokine in vivo7, etc. Through the nanogel tectonics, we devised chemical cross-linking of the nanogel and succeeded in generating the nanogel-cross-linked (NanoClik) gel with various size, shapes and chemical characteristics8,9. Furthermore, freeze-thawing of the NanoClik gel resulted in nanogel-cross-linked porous (NanoCliP) gel with highly porous structure8. Thus, further modification of the NanoCliP gel may provide a quite suitable scaffold for osteoblasts. Direct reprogramming technologies enable phenotypic conversion of a somatic cell type into another without passing through an intermediate pluripotent state10–13. Recently we succeeded in reprogramming human fibroblasts into osteoblasts (directly converted osteoblasts; dOBs) by transducing four genes each encoding Runx2 (R), Osterix (X), Oct3/4 (O), and L-Myc (L), or three genes (X, O and L)14–16. As fibroblasts can be easily obtained from a small piece of biopsy sample of a patient without any invasive procedure17 and expanded into an enough large number18,19, our procedure may be quite adequate for preparation of a large number of autologous osteoblasts with a high bone forming ability. Therefore, we tried to combine the above two technologies, the nanogel tectonic material and dOBs, to generate a feasible tissue engineering procedure for a novel regeneration therapy of bone diseases. In the present study, we newly developed NanoCliP-FD gel from the NanoCliP gel that we previously reported8, and examined its potential as a scaffold for bone tissue engineering using the dOBs induced by our previously reported procedure14–16. Results dOBs adhered to and proliferated onto the fibronectin-coated NanoCliP-FD gel. We modified the NanoCliP gel that we previously reported8 by first freezing and drying it, and subsequently by hydrating it. By doing so, we obtained a novel nanogel tectonic material, namely, the NanoCliP-FD gel (Fig. 1a). CLSM imaging of rhodamine-labeled NanoCliP and NanoCliP-FD gel revealed that the former had interconnected pores of several hundred micrometers in diameter, while the pores in the latter irregularly interlinked into larger pores (Fig. 1b and Supporting Information Table S1). The rhodamine-labeled NanoCliP-FD gel was coated with either RGDC peptide or fibronectin, and the XOL-transduced cells were seeded into the gel (Fig. 2a). After culturing for 1~7 days, the distribution of the cells was observed (Fig. 2b). On day after cell seeding, viable cells positively stained with calcein-AM were widely dispersed all over the fibronectin-coated NanoCliP-FD gel. The density of the viable cells increased with elapsing culture period; and, 7 to 14 days after the initiation of the culture, the cells almost covered the entire surface of the walls of the pores in the fibronectin-coated NanoCliP-FD gel, so that the red signal of rhodamine was masked. Consistently, visualization of F-actin using fluorescent phalloidin also demonstrated that a large number of cells broadly distributed in and adhered to the walls of the pores in the fibronectin-coated NanoCliP-FD gel at day 1, and almost the entire surface was occupied by cells at day 7 (Fig. 2c). In contrast, calcein-AM-positive viable cells did not adhere to the surface in the RGDC-conjugated NanoCliP-FD gel at day 1. The cells were rarely observed in the NanoCliP-FD gel, and the red signal of rhodamine remained clearly visible, even at days 7 and 14 (Fig. 2b). Fluorescent phalloidin staining showed sphere-like aggregation of the cells, suggesting that the cells adhered to each other rather than to the RGDC-conjugated scaffold (Fig. 2c). Cell growth was examined by a tetrazolium-based assay. As shown in Fig. 2d, the XOL-transduced cells significantly proliferated in the fibronectin-coated NanoCliP-FD gel but not in the RGDC-conjugated NanoCliP-FD gel. These results suggest that the fibronectin coating enabled the XOL-transduced cells to adhere to and proliferate in the NanoCliP-FD gel. Bone matrix production by dOBs in the fibronectin-coated NanoCliP-FD gel in culture. We examined whether the XOL-transduced cells gave rise to dOBs during the culture in the fibronectin-coated NanoCliP-FD gel. Quantitative RT-PCR analysis was performed to evaluate expression levels of mRNA for osteopontin and osteocalcin that are osteoblast-specific proteins composing the bone matrix. As shown in Fig. 3a, mRNA for these genes was highly expressed in the XOL-transduced cells, but not by HDFs, in the fibronectin-coated NanoCliP-FD gel 14 to 21 days after the initiation of the culture. These results strongly suggest that the XOL-transduced cells were successfully converted into dOBs in this 3D scaffold to form the fibronectin-coated NanoCliP-FD gel with dOBs (Fig. 2a). Next we analyzed whether the dOBs produced calcified bone matrix in the fibronectin-coated NanoCliP-FD gel. As shown in Fig. 3b, Alizarin red S staining clearly showed massive formation of mineralized bone matrix in the fibronectin-coated NanoCliP-FD gel with dOBs, whereas non-transduced HDFs cultured under the same conditions only faintly did so. The results were confirmed by quantitative measurement of the staining intensity for the samples (Fig. 3c). The OsteoImage assay also showed a huge amount of calcium phosphate deposits in all over the fibronectin-coated NanoCliP-FD gel with dOBs (Fig. 3d), whereas the calcium phosphate was rarely deposited in the fibronectin-coated NanoCliP-FD gel with HDFs. These results strongly suggest that calcified bone matrix was massively produced in the fibronectin-coated NanoCliP-FD gel with dOBs. Bone regeneration in vivo was promoted by the fibronectin-coated NanoCliP-FD gel with dOBs. To estimate whether the fibronectin-coated NanoCliP-FD gel with dOBs can contribute to bone regen- eration in vivo, we transplanted it into immune-deficient mice in which an artificial bone defect lesion was created at the femur site. Three weeks later, micro CT imaging revealed that the bone defect mostly remained in a control group of animals that were not transplanted as well as in another group that received transplantation of the fibronectin-coated NanoCliP-FD gel with HDFs (Fig. 4a). In contrast, transplantation of the fibronectin-coated NanoCliP-FD gel with dOBs resulted in the formation of a large callus that covered a large area of the femur SCIenTIfIC REPOrTS \| (2018) 8:15824 \| DOI:10.1038/s41598-018-33892-z 2 www.nature.com/scientificreports/ 䠄a䠅 CHPOA (+Rh labeling) CHPOA nanogel Self -assembly (+RGDC) HS H R= Pullulan NanoClik gel Crosslinking n SH HS SH PEGSH O O Cholesterol group Freezethaw H O O Acryloyl group NanoCliP gel + Freezedry NanoCliP FD -matrix NanoCliP FD -gel Hydration (+Fibronectin) 䠄b䠅 Whole NanoCliP gel Surface NanoCliP F Cross-section matrix NanoCliP gel 1 mm NanoCliP FD-matrix 1 mm 100 μm NanoCliP F gel NanoCliP F gel NanoCliP FD-gel 1 mm 100 μm Figure 1. The NanoCliP-FD gel used in the study. (a) Preparation of the NanoCliP-FD gel. Rhodamine-labeled or non-labeled CHPOA was used to generate rhodamine-labeled and non-labeled NanoCliP-FD gel. (b) Rhodamine-labeled NanoCliP gel, NanoCliP-FD matrix, and NanoCliP-FD gel were prepared as described in (a). The 3D-composited images of the whole and surface of the samples as well as cross-sectional images of the samples were obtained by CLSM. leaving only small defect lesions. The different outcomes were statistically confirmed by calculation of percent of callus formation (Fig. 4b). Histopathological examinations also confirmed that a large callus was formed to cover most of the area at the bone defect legion in mice transplanted with the NanoCliP-FD gel with dOBs (Fig. 4c). In contrast, a large defect remained at the femur transplanted with the NanoCliP-FD gel with HDFs. These results strongly suggest that the NanoCliP-FD gel with dOBs promoted bone healing in vivo by producing regenerative bone tissue, while NanoCliP-FD gel with HDFs failed to do so. NanoCliP-FD gel did not cause any remarkable adverse event in mice. Fibronectin-coated NanoCliP-FD gel was transplanted into the subcutaneous tissue of mice, and 10 days later sera of the mice were subjected to laboratory tests. As shown in Supplementary Table S3, a significant difference was not detected between the mice and control mice with regard to all the items tested, strongly suggesting that the NanoCliP-FD gel was not toxic. SCIenTIfIC REPOrTS \| (2018) 8:15824 \| DOI:10.1038/s41598-018-33892-z 3 www.nature.com/scientificreports/ 䠄a䠅 20 mm 䠄b䠅 RGDC Fibronec x 2.5 Day 1 Day 7 Day 14 400 μm RGDC x 20 Fibronec Day 1 Day 7 Day 14 20 mm 䠄c䠅 RGDC Fibronec x 2.5 Day 1 Day 7 䠄d䠅 200 μm OD490 RGDC x 20 Fibronec Day 1 Day 7 0.4 0.3 0.2 0.1 0 1 7 14 Days of culture RGDC Fibronec Figure 2. Osteoblasts efficiently adhered to and proliferated on the fibronectin-coated NanoCliP-FD gel. (a) Preparation of NanoCliP-FD gel with dOBs. (b,c) XOL-transduced cells were seeded into rhodamine-labeled NanoCliP-FD gel that had been coated with either RGDC or fibronectin. After culturing for the indicated days, the samples were stained with calcein-AM (b) or phalloidin/Hoechst 33342 (c), and CLSM images at magnifications of objective lenses of x2.5 (upper) and x20 (lower) are shown. (d) XOL-transduced cells were seeded into non-coated (−), RGDC-conjugated, and fibronectin-coated NanoCliP-FD gel. After culturing for the indicated days, cell viability was evaluated by tetrazolium-based assay. Values are means ± SD. N = 3. p < 0.01 vs. day 1. Discussion A variety of biomaterials have been developed so far to build 3D scaffolds for bone tissue engineering20,21. These include (i) bioceramics such as hydroxyl apatite22, (ii) natural macromolecules and their derivatives such as collagen, atelocollagen, chitosan and alginate23–25, (iii) synthetic polymers such as polyethylene glycol26, (iv) hydrogels27,28, and (v) other materials such as carbon nanotubes29,30. A variety of combinations of (i~v) have also been widely used20,21,31–33. They can be shaped into gel-like, sponge-like, mesh-like, or fibrous structures. In the present study, we constructed a novel nanogel tectonic material, the NanoCliP-FD gel, that is a kind of hydrogel fabricated by freeze-thawing and subsequent freeze-drying of chemically cross-linked nanogel consisting of natural macromolecules. The biocompatibility of NanoCliP gel in the use of implantation was evaluated in the previous study8. The H&E and immunohistochemical staining for monocytes and mature macrophages suggested that there was negligible histological evidence of a foreign body response after the implantation of NanoCliP gel. NanoCliP-FD matrix and gel consist of the same concentration of nanogels and cross-linkers and no xenogenic macromolecule is included in it, suggesting that the NanoCliP-FD gel is a biocompatible material. The NanoCliP-FD gel may be biocompatible and non-toxic. So far as we tested, a significant adverse event was not observed in mice SCIenTIfIC REPOrTS \| (2018) 8:15824 \| DOI:10.1038/s41598-018-33892-z 4 www.nature.com/scientificreports/ 䠄a䠅 Osteopon Rela ve mRNA levels 30 20 Osteocalcin 150 100 10 50 0 0 HDFs XLO 14 days HDFs XLO HDFs 28 days XLO HDFs 14 days XLO 28 days 䠄b䠅 Alizarin red S − HDFs XLO XLO 14 days 䠄c䠅 21 days 2.5 Rela ve staining intensity on day 21 2 1.5 1 0.5 0 䠄−䠅 HDFs XLO Low magnifica 䠄d䠅 High magnifica HDFs XLO Figure 3. XOL-transduced cells were successfully converted into dOBs that produced calcified bone matrix in fibronectin-coated NanoCliP-FD gel. (a) RNA was extracted from the HDFs or XOL-transduced cells cultured in fibronectin-coated NanoCliP-FD gel for 14 and 28 days. mRNA levels for the indicated genes were evaluated by real time-RT-PCR. Values are means ± SD. N = 3. p < 0.01 vs. HDFs. (b,c) Fibronectin-coated NanoCliP-FD gel with HDFs and XOL-transduced cells were cultured for the indicated days and stained with Alizarin red S. Some aliquots of the fibronectin-coated NanoCliP-FD gel were cultured without cell seeding (−). Macroscopic images (b) and relative staining intensities (c) are shown. Values are means ± SD. N = 3. p < 0.01 vs. cell-free control. (d) Fibronectin-coated NanoCliP-FD gel with HDFs or XOL-transduced cells was cultured as above, and osteoimage assay was performed 21 days later. Confocal LSM images at low (upper) and high (lower) magnifications are shown. transplanted with fibronectin-coated NanoCliP-FD gel (Supporting Information Table S3). A large amount of NanoCliP-FD gel can be easily and quickly prepared at a low cost. In cell adhesion procedure, the solid-state NanoCliP-FD matrix was rehydrated by adding aqueous culture medium containing cells. The pore structures were still observed in NanoCliP-FD gel, which contains numerous SCIenTIfIC REPOrTS \| (2018) 8:15824 \| DOI:10.1038/s41598-018-33892-z 5 www.nature.com/scientificreports/ 3D Longitudinal Transverse 䠄a䠅 − HDFs dOBs Gra % of callus forma 䠄b䠅 80 ** 60 40 20 0 䠄−䠅 HDFs dOBs 䠄c䠅 H-E Alizarin red S * * 200 μm * * + * + * * * (−) HDFs * + + * * * * + + + dOBs Figure 4. Bone healing was facilitated by transplantation of the fibronectin-coated NanoCliP-FD gel with dOBs. Fibronectin-coated NanoCliP-FD gel with HDFs or dOBs was prepared as in Fig. 3, and transplanted into an artificial segmental bone defect lesion that was created at femoral diaphysis in NOG/SCID mice. Control mice were not transplanted (−). Mice were sacrificed 21 days after the surgery. (a,b) µCT images of the femur were acquired. Serial 10-µm slices (top and middle) and 3D reconstructed (bottom) images (a) and %Callus formation (b) are shown. (c) Serial sections of the tissues were stained with H-E (upper) and Alizarin red S (lower). In (a) triangles and arrows represent bone defect lesions and regenerated bone tissue, respectively. In (b), values are means ± SD. N = 3 mice. *p < 0.01 vs. non-transplantation control. In (c), and +represent regenerated bone tissue and NanoCliP-FD gel, respectively, and arrowheads represent bone defect lesions. large pores that were formed by fusion of pre-existing pores in the NanoCliP gel through freezing-and-drying (Fig. 1b). The enlarged pores allowed cells to easily penetrate into the gel and the cells forcibly contacted with the fibronectin-coated surface of the scaffold. From Fig. 2b and c, the cells effectively adhered onto the fibronectin-coated NanoCliP-FD gel, strongly suggesting that the fibronectin was effectively coated onto the surface of the walls of the expanded pores. The XOL-transduced cells proliferated and converted into dOBs in the NanoCliP-FD gel during the culture (Figs 2d and 3), which may be also due to the efficient adhesion of the cells onto the pores’ walls in the NanoCliP-FD gel. Therefore, this nanogel tectonic material exhibits excellent characteristics appropriate for a 3D scaffold for bone tissue engineering, and is efficiently combined with dOBs, thanks to the large pores formed by the freezing-and-drying treatment. SCIenTIfIC REPOrTS \| (2018) 8:15824 \| DOI:10.1038/s41598-018-33892-z 6 www.nature.com/scientificreports/ The NanoCliP-FD matrix is a dehydrated intermediate to produce NanoCliP-FD gel from the NanoCliP gel (Fig. 1a), and can be conveniently stored under a dry condition at room temperature for a long period. One may utilize the NanoCliP-FD matrix for storage and transport purposes, and the NanoCliP-FD gel can be quickly and easily prepared from the NanoCliP-FD matrix when necessary. This is a great advantage related to the utilization of the NanoCliP-FD gel for transplantation therapy, because NanoCliP-FD matrix can be stored and transported safely and at a low cost. Compared with the fibronectin coating, the RGDC-conjugated NanoCliP-FD gel failed to provide the same efficient adherence of the dOBs onto the gel, so that the cells adhered to each other rather than to the pores’ walls and sphere-like aggregation of the cells appeared (Fig. 2c). The cells less efficiently adhered onto RGDC-conjugated NanoCliP-FD than onto fibronectin-conjugated one, probably because the RGCD was embeded/masked within the gel whereas the fibronectin was on the exterior. Fibronectin may not only allow adhesion of osteoblasts but also stimulate them. A recent report shows that the fibronectin isoforms containing the extradomain A may interact with α4β1 integrin to augment osteoblast differentiation. Other isoforms containing the extradomain B may interact with β3 integrin and promote osteoblast differentiation and mineralization34. The NanoCliP-FD gel can be built to have any desired size and shape, because the cross-linking of CHPOA nanogel into NanoClik gel (Fig. 1a) can be processed in any template with the desired size and shape. The template can be fabricated in a computer-directed tailor-made fashion specific for the bone loss lesion in each patient. In previous studies in which NanoClik gel was used as a scaffold for tissue engineering, the NanoClik gel was also used for slow release of some soluble factors, due to the molecular chaperon activity of the nanogel35–37. The soluble factors included BMP35,36, FGF1836, EP4A35, and W9 peptide37, which may have promoted bone regeneration in vivo. If such a slow release of soluble factors could also be achieved by means of the NanoCliP-FD gel, usefulness and efficacy of the gel in bone tissue regeneration may be further expanded due to the activities of the soluble factors. But this point need to be further examined in future studies. Mesenchymal stem cells (MSCs) are widely used for bone tissue regeneration both in vivo and ex vivo, and remarkable therapeutic benefits of MSC transplantation have been confirmed in various clinical settings such as reconstruction of a large bone defect due to surgical resection of a bone tumor38. However, MSCs are harvested from bone marrow or adipose tissue of patients in a highly invasive manner. The number of MSCs harvested from a patient might not always be satisfactorily high, and proliferation and differentiation capacities of MSCs might be limited, especially in elderly cases39,40. In this respect, an additional option is required and the dOBs may be eligible as alternative cells to be used for transplantation therapy in bone diseases. A potential disadvantage of dOBs is that the dOBs are induced from fibroblasts by transducing Osterix, Oct3/4, and L-myc genes via retroviral vectors. The retroviral sequences integrated into the chromosomes may potentially cause aberrant expression of endogenous oncogenes to transform the dOBs into tumor cells. To overcome this problem, we developed a novel procedure to convert fibroblasts into osteoblasts by means of an addition of a chemical compound instead of gene transfer41. Such transgene-free dOBs may be more advantageous than the present ones, because the former may have further lower risk of tumor formation than the latter. Taken together, the combination of the novel dOBs and the NanoCliP-FD gel may be a quite feasible approach in clinical applications for various bone diseases. Methods Preparation of NanoCliP-FD matrix. The NanoCliP gel was prepared as previously described (Fig. 1a)8. Briefly, the cholesterol-bearing pullulan (CHP), in which pullulan (average molecular weight = 1 × 105 g/mol) was substituted with 1.2 cholesterol moieties per 100 anhydrous glucoside units, was purchased from NOF corporation (Tokyo, Japan). Acryloyl group-modified cholesterol-bearing pullulan (CHPOA) was synthesized using 2-Acryloyloxyethyl isocyanate (Showa Denko, Tokyo, Japan). CHPOA self-assembled into CHPOA nanogel, which was subsequently cross-linked by Michael addition of pentaerythritol tetra (mercaptoethyl) polyoxyethylene (PEGSH) (average molecular weight = 1 × 104 g/mol) (NOF corporation) to form NanoClik gel in Microhematocrit Capillary Tubes with an inner diameter of 1.1 mm (Fisher Scientific) as a template. NanoClik gel was then converted into highly porous NanoCliP gel by freezing-induced phase separation. To prepare NanoCliP freeze-dried-matrix (NanoCliP-FD matrix), NanoCliP gel was quickly frozen in liquid nitrogen, followed by drying in a vacuum. Rhodamine-labeled NanoCliP gel and NanoCliP-FD matrix were prepared as above using acryloyl group-modified rhodamine-labeled CHP (CHPOA-Rh) instead of CHPOA, and rhodamine-labeled NanoCliP-FD matrix was soaked in PBS to form rhodamine-labeled NanoCliP-FD gel. Preparation of RGDC-conjugated and Fibronectin-coated NanoCliP-FD gel. RGDC conjugation was performed as follows. Synthetic RGDC (Arg-Gly-Asp-Cys) peptide (SCRUM Inc., Tokyo, Japan) was added to CHPOA nanogel simultaneously with the addition of PEGSH (Fig. 1a), in such a manner that final concentrations of CHPOA, PEGSH and RGDC peptide were 20 mg/mL, 35 mg/mL and 2 mg/mL, respectively. Subsequently, RGDC-conjugated NanoCliP gel and RGDC-conjugated NanoCliP-FD gel were prepared as above. For fibronectin-coating, NanoCliP-FD matrix was soaked in 50 μg/mL fibronectin solution (Wako laboratory chemicals, Osaka, Japan) for 6 hours, followed by rinsing twice in ethanol and drying (Fig. 1a). The resultant fibronectin-coated NanoCliP-FD matrix was hydrated to form fibronectin-coated NanoCliP-FD gel. Cells. Human dermal fibroblasts (HDFs) were purchased from DS Pharma Biomedical Co., Ltd. (Osaka, Japan) and cultured in Dulbecco’s minimum essential medium (DMEM) supplemented with 100 mM non-essential amino acids, 100 U/ml penicillin 100 μg/ml streptomycin, and 10% fetal bovine serum (FBS) in 5% CO2/95% humidified air at 37 °C (Standard culture conditions). Osterix (X), Oct3/4 (O) and L-myc (L) genes were transduced into the HDFs as described previously with slight modification16. Briefly, HDFs were seeded onto 100 mm culture dish at a density of 5 × 103 cell/dish. Twenty-four hours later, cells were infected with a mixture of X, O SCIenTIfIC REPOrTS \| (2018) 8:15824 \| DOI:10.1038/s41598-018-33892-z 7 www.nature.com/scientificreports/ and L retrovirus vectors freshly prepared using the PLAT-GP packaging cells (Toyobo, Osaka, Japan) as previously described11,12. After culturing for another 24 hours, the XOL-transduced cells were detached from culture dishes by trypsinization and resuspended in DMEM medium supplemented with 50 μg/mL ascorbic acid, 10 mM β-glycerol phosphate, 100 nM dexamethasone and 10% FBS (osteogenic medium). 3D culture. HDFs or XOL-transduced cells resuspended in the osteogenic medium were prepared in such a manner that the approximate cell density was 5 × 105/50 μL. RGD-conjugated, Fibronectin-coated, and non-coated NanoCliP-FD matrices were soaked in the cell suspension, and air bubble was pushed out of the matrices with tweezers to allow infiltration of the cell suspension into the matrices (Fig. 1a). The matrices were incubated in the cell suspension for 2 hours. The resultant samples were transferred into 1.5 mL of fresh osteogenic medium in 12-well culture plates and cultured in 5% CO2/95% humidified air at 37 °C (the standard culture conditions). Calcein-AM staining. The samples were rinsed twice with PBS, and immersed in 0.1 mg/mL of calcein-AM (3′,6′-Di(O-acetyl)-2′,7′-bis[N,N-bis(carboxymethyl)aminomethyl]fluoresceinTetraacetoxymethyl Ester) (Dojindo, Kumamoto, Japan)/PBS solution for 20 min. The samples were rinsed twice with PBS, followed by fixation with 4% formaldehyde for 20 min. After three times washing with distilled water, the samples were observed under Confocal laser scanning microscope (CLSM) (LSM780; Carl Zeiss, Oberkochen, Germany) Phalloidin/Hoechst 33342 staining. The samples were rinsed with PBS, and fixed with 4% formaldehyde for 20 min. After washing with distilled water twice, the samples were immersed in 0.1% Triton X-100 for 5 min. The samples were further washed twice with distilled water, followed by staining with 0.1 mg/mL of Alexa Fluor 488 phalloidin (Molecular Probes, Eugene, OR)/PBS solution for 20 min. After another three-times washing, the samples were observed under CLSM. ® Cell viability analysis. Cell viability was evaluated by a tetrazolium-based assay as previously described42. NanoCliP-FD gel with dOBs was soaked in 500 µL of osteogenic medium in 12-well plates and cultured under the standard culture conditions. One, 7 or 14 days later, the culture supernatant was replaced by 1,000 µl of fresh complete medium containing 2-(2-methoxy-4-nitrophenyl)-3-(4-nitrophenyl)-5-(2, 4-disulfophenyl)-2H-tetrazolium monosodium salt (WST-8; Nacalai Tesque, Kyoto, Japan) was added to the wells. After incubation for an additional 1 hour, the culture supernatant was harvested and OD at 490 nm was measured. Assessment of mineralization. For Alizarin red S staining, samples were fixed in 95% ethanol and stained with Alizarin red S solution (Sigma Aldrich) as previously described16. The OsteoImage mineralization assay was performed using the OsteoImage mineralization assay kit (Lonza) according to the manufacturer’s instruction. Briefly, the samples were rinsed with PBS, and fixed with 95% ethanol for 20 min. After washing, the samples were immersed in OsteoImage Staining Reagent for 30 min, followed by washing and observation under CLSM. ™ Real time RT-PCR. Total RNA was extracted from cells using ISOGEN II (Nippon Gene Co. Ltd.) and reverse-transcribed using ReverTra Ace qPCR RT Master Mix (Toyobo). The resultant cDNA was mixed with Real-Time PCR Master Mix (Applied Biosystems, Waltham, MA) and matching probes/primers specific for human β-actin, osteocalcin and osteopontin genes (Supporting Information Table S2)). Real time PCR was carried out on a 7300 Real Time PCR System (Applied Biosystems). Values (means ± SD) were normalized with respect to the β-actin mRNA level in each sample, and relative values were calculated. Surgery and transplantation. All experimental procedures and protocols for animals conformed to the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by either the Committee for Animal Research of Kyoto Prefectural University of Medicine or the Animal Care. dOBs were induced from HDFs, seeded into fibronectin-coated NanoCliP-FD gel (1 × 1 × 6 mm), and cultured for 7 days as above. After culturing for 7 days, some sample was tested for bone matrix formation by Alizarin red S staining, and the other samples were used for transplantation. Six-week-old male NOG/SCID mice were anesthetized with isoflurane and a segmental bone defect with a size of 1 × 1 × 6 mm was created at the diaphysis of the left femur using a dental drill under pouring water. Radiological and histological analysis. Twenty-one days after the transplantation, mice were euthanized with a lethal dose of isoflurane. Thighs were dissected, fixed with 10% neutral-buffered formalin, and subjected to μCT imaging (Toshiba). Specimens were embedded in SCEM compound (Leica Microsystems), quick-frozen, and cryo-sectioned into 6-μm slices. After staining with H & E and Alizarin Red S, the sections were analyzed with BZ-X710 fluorescence microscope (Keyence). Statistical analysis. Data are expressed as means ± standard deviation (S.D.). Statistical significance was analyzed using Student’s t-test. P < 0.05 was considered significant. References 1. Teixeira, F. B. et al. Periodontitis and Alzheimer’s Disease: A Possible Comorbidity between Oral Chronic Inflammatory Condition and Neuroinflammation. Front Aging Neurosci 9, 327, https://doi.org/10.3389/fnagi.2017.00327 (2017). 2. Nilsson, H., Berglund, J. S. & Renvert, S. Periodontitis, tooth loss and cognitive functions among older adults. Clin Oral Investig https://doi.org/10.1007/s00784-017-2307-8 (2017). 3. Calori, G. M. et al. Non-unions. Clin Cases Miner Bone Metab 14, 186–188, https://doi.org/10.11138/ccmbm/2017.14.1.186 (2017). 4. Pacelli, S. et al. Strategies to develop endogenous stem cell-recruiting bioactive materials for tissue repair and regeneration. Adv Drug Deliv Rev 120, 50–70, https://doi.org/10.1016/j.addr.2017.07.011 (2017). SCIenTIfIC REPOrTS \| (2018) 8:15824 \| DOI:10.1038/s41598-018-33892-z 8 www.nature.com/scientificreports/ 5. Sasaki, Y. & Akiyoshi, K. Nanogel engineering for new nanobiomaterials: from chaperoning engineering to biomedical applications. Chem Rec 10, 366–376, https://doi.org/10.1002/tcr.201000008 (2010). 6. Fujii, H. et al. Cycloamylose-nanogel drug delivery system-mediated intratumor silencing of the vascular endothelial growth factor regulates neovascularization in tumor microenvironment. Cancer Sci. 105, 1616–1625, https://doi.org/10.1111/cas.12547 (2014). 7. Shimizu, T. et al. Nanogel DDS enables sustained release of IL-12 for tumor immunotherapy. Biochem. Biophys. Res. Commun. 367, 330–335, doiS0006-291X(07)02763-5 [pii], https://doi.org/10.1016/j.bbrc.2007.12.112 (2008). 8. Hashimoto, Y., Mukai, S. A., Sawada, S., Sasaki, Y. & Akiyoshi, K. Nanogel tectonic porous gel loading biologics, nanocarriers, and cells for advanced scaffold. Biomaterials 37, 107–115, https://doi.org/10.1016/j.biomaterials.2014.10.045 (2015). 9. Tahara, Y., Mukai, S. A., Sawada, S., Sasaki, Y. & Akiyoshi, K. Nanocarrier-Integrated Microspheres: Nanogel Tectonic Engineering for Advanced Drug-Delivery Systems. Adv Mater 27, 5080–5088, https://doi.org/10.1002/adma.201501557 (2015). 10. Ieda, M. et al. Direct reprogramming of fibroblasts into functional cardiomyocytes by defined factors. Cell 142, 375–386, https://doi. org/10.1016/j.cell.2010.07.002 (2010). 11. Kishida, T. et al. Reprogrammed Functional Brown Adipocytes Ameliorate Insulin Resistance and Dyslipidemia in Diet-Induced Obesity and Type 2 Diabetes. Stem Cell Rep. 5, 569–581, https://doi.org/10.1016/j.stemcr.2015.08.007 S2213-6711(15)00241-6 [pii] (2015). 12. Sowa, Y. et al. Direct Conversion of Human Fibroblasts into Schwann Cells that Facilitate Regeneration of Injured Peripheral Nerve In Vivo. Stem Cells Transl Med 6, 1207–1216, https://doi.org/10.1002/sctm.16-0122 (2017). 13. Wakao, J. et al. Efficient direct conversion of human fibroblasts into myogenic lineage induced by co-transduction with MYCL and MYOD1. Biochem Biophys Res Commun https://doi.org/10.1016/j.bbrc.2017.05.059 (2017). 14. Yamamoto, K. et al. Generation of Directly Converted Human Osteoblasts That Are Free of Exogenous Gene and Xenogenic Protein. J Cell Biochem 117, 2538–2545, https://doi.org/10.1002/jcb.25546 (2016). 15. Mizoshiri, N. et al. Transduction of Oct6 or Oct9 gene concomitant with Myc family gene induced osteoblast-like phenotypic conversion in normal human fibroblasts. Biochem Biophys Res Commun 467, 1110–1116, https://doi.org/10.1016/j.bbrc.2015.10.098 (2015). 16. Yamamoto, K. et al. Direct conversion of human fibroblasts into functional osteoblasts by defined factors. Proc. Natl. Acad. Sci. USA 112, 6152–6157, https://doi.org/10.1073/pnas.1420713112 (2015). 17. Huschtscha, L. I. et al. Enhanced isolation of fibroblasts from human skin explants. Biotechniques 53, 239–244, https://doi. org/10.2144/0000113939 (2012). 18. Cristofalo, V. J., Allen, R. G., Pignolo, R. J., Martin, B. G. & Beck, J. C. Relationship between donor age and the replicative lifespan of human cells in culture: a reevaluation. Proc Natl Acad Sci USA 95, 10614–10619 (1998). 19. Shay, J. W. & Wright, W. E. Hayflick, his limit, and cellular ageing. Nat Rev Mol Cell Biol 1, 72–76, https://doi.org/10.1038/35036093 (2000). 20. Scott, T. G. et al. Advances in bionanomaterials for bone tissue engineering. J Nanosci Nanotechnol 13, 1–22 (2013). 21. Black, C. R. et al. Bone Tissue Engineering. Curr Mol Biol Rep 1, 132–140, https://doi.org/10.1007/s40610-015-0022-2 (2015). 22. Pezzotti, G. et al. Bioactive silicon nitride: A new therapeutic material for osteoarthropathy. Sci Rep 7, 44848, https://doi.org/10.1038/ srep44848 (2017). 23. Shekhter, A. B. et al. Medical applications of collagen and collagen-based materials. Curr Med Chem. https://doi.org/10.2174/09298 67325666171205170339 (2017). 24. Oryan, A. & Sahvieh, S. Effectiveness of chitosan scaffold in skin, bone and cartilage healing. Int J Biol Macromol 104, 1003–1011, https://doi.org/10.1016/j.ijbiomac.2017.06.124 (2017). 25. Noori, A., Ashrafi, S. J., Vaez-Ghaemi, R., Hatamian-Zaremi, A. & Webster, T. J. A review of fibrin and fibrin composites for bone tissue engineering. Int J Nanomedicine 12, 4937–4961, https://doi.org/10.2147/IJN.S124671 (2017). 26. Engebretson, B. & Sikavitsas, V. I. Long-term in vivo effect of PEG bone tissue engineering scaffolds. J Long Term Eff Med Implants 22, 211–218 (2012). 27. Short, A. R. et al. Hydrogels That Allow and Facilitate Bone Repair, Remodeling, and Regeneration. J Mater Chem B 3, 7818–7830, https://doi.org/10.1039/C5TB01043H (2015). 28. Tozzi, G., De Mori, A., Oliveira, A. & Roldo, M. Composite Hydrogels for Bone Regeneration. Materials (Basel) 9, https://doi. org/10.3390/ma9040267 (2016). 29. Yi, H., Ur Rehman, F., Zhao, C., Liu, B. & He, N. Recent advances in nano scaffolds for bone repair. Bone Res 4, 16050, https://doi. org/10.1038/boneres.2016.50 (2016). 30. Das, K. et al. Functionalized carbon nanotubes as suitable scaffold materials for proliferation and differentiation of canine mesenchymal stem cells. Int J Nanomedicine 12, 3235–3252, https://doi.org/10.2147/IJN.S122945 (2017). 31. Zhou, D. et al. Comparative study of porous hydroxyapatite/chitosan and whitlockite/chitosan scaffolds for bone regeneration in calvarial defects. Int J Nanomedicine 12, 2673–2687, https://doi.org/10.2147/IJN.S131251 (2017). 32. Hisanaga, Y. et al. Effect of the combined use of enamel matrix derivative and atelocollagen sponge scaffold on osteoblastic differentiation of mouse induced pluripotent stem cells in vitro. J Periodontal Res 53, 240–249, https://doi.org/10.1111/jre.12511 (2018). 33. Shahbazarab, Z., Teimouri, A., Chermahini, A. N. & Azadi, M. Fabrication and characterization of nanobiocomposite scaffold of zein/chitosan/nanohydroxyapatite prepared by freeze-drying method for bone tissue engineering. Int J Biol Macromol 108, 1017–1027, https://doi.org/10.1016/j.ijbiomac.2017.11.017 (2018). 34. Sens, C. et al. Fibronectins containing extradomain A or B enhance osteoblast differentiation via distinct integrins. J Biol Chem 292, 7745–7760, https://doi.org/10.1074/jbc.M116.739987 (2017). 35. Kamolratanakul, P. et al. Nanogel-based scaffold delivery of prostaglandin E(2) receptor-specific agonist in combination with a low dose of growth factor heals critical-size bone defects in mice. Arthritis Rheum 63, 1021–1033, https://doi.org/10.1002/art.30151 (2011). 36. Fujioka-Kobayashi, M. et al. Cholesteryl group- and acryloyl group-bearing pullulan nanogel to deliver BMP2 and FGF18 for bone tissue engineering. Biomaterials 33, 7613–7620, https://doi.org/10.1016/j.biomaterials.2012.06.075 (2012). 37. Sato, T. et al. Nanogel-crosslinked nanoparticles increase the inhibitory effects of W9 synthetic peptide on bone loss in a murine bone resorption model. Int J Nanomedicine 10, 3459–3473, https://doi.org/10.2147/IJN.S61566 (2015). 38. Liu, Y., Wu, J., Zhu, Y. & Han, J. Therapeutic application of mesenchymal stem cells in bone and joint diseases. Clin Exp Med 14, 13–24, https://doi.org/10.1007/s10238-012-0218-1 (2014). 39. Zhou, S. et al. Age-related intrinsic changes in human bone-marrow-derived mesenchymal stem cells and their differentiation to osteoblasts. Aging Cell 7, 335–343, https://doi.org/10.1111/j.1474-9726.2008.00377.x (2008). 40. Kim, M. et al. Age-related alterations in mesenchymal stem cells related to shift in differentiation from osteogenic to adipogenic potential: implication to age-associated bone diseases and defects. Mech Ageing Dev 133, 215–225, https://doi.org/10.1016/j. mad.2012.03.014 (2012). 41. Yamamoto, K. et al. Direct phenotypic conversion of human fibroblasts into functional osteoblasts triggered by a blockade of the transforming growth factor-beta signal. Sci Rep 8, 8463, https://doi.org/10.1038/s41598-018-26745-2 (2018). 42. Nakanishi, H. et al. Nonviral genetic transfer of Fas ligand induced significant growth suppression and apoptotic tumor cell death in prostate cancer in vivo. Gene Ther 10, 434–442, https://doi.org/10.1038/sj.gt.3301912 (2003). SCIenTIfIC REPOrTS \| (2018) 8:15824 \| DOI:10.1038/s41598-018-33892-z 9 www.nature.com/scientificreports/ Acknowledgements This work was supported by grants from the Japanese Ministry of Education, Culture, Sports, Science and Technology. Author Contributions Y. S., K. Y., Y. T., T. K., K. A., and O. M. designed the study and wrote the paper. Y. S., K. Y., S. H., Y. T., K. N., and S.-I. K. performed the experiments. F. O., G. P., T. Y., and N. K. supervised and advised on the study. Additional Information Supplementary information accompanies this paper at https://doi.org/10.1038/s41598-018-33892-z. Competing Interests: The authors declare no competing interests. Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. © The Author(s) 2018 SCIenTIfIC REPOrTS \| (2018) 8:15824 \| DOI:10.1038/s41598-018-33892-z 10
https://openalex.org/W2109970853	https://amt.copernicus.org/articles/5/1271/2012/amt-5-1271-2012.pdf	Latin	null	Trend analysis of aerosol optical thickness and Ångström exponent derived from the global AERONET spectral observations	Atmospheric measurement techniques	2,012	cc-by	18,853	Correspondence to: J. Yoon (yoon@iup.physik.uni-bremen.de) Received: 19 July 2011 – Published in Atmos. Meas. Tech. Discuss.: 18 August 2011 Revised: 16 May 2012 – Accepted: 16 May 2012 – Published: 6 June 2012 Abstract. Regular aerosol observations based on well- calibrated instruments have led to a better understanding of the aerosol radiative budget on Earth. In recent years, these instruments have played an important role in the determina- tion of the increase of anthropogenic aerosols by means of long-term studies. Only few investigations regarding long- term trends of aerosol optical characteristics (e.g. aerosol op- tical thickness (AOT) and ˚Angstr¨om exponent ( ˚AE)) have been derived from ground-based observations. This paper aims to derive and discuss linear trends of AOT (440, 675, 870, and 1020 nm) and ˚AE (440–870 nm) using AErosol RObotic NETwork (AERONET) level 2.0 spectral observa- tions. Additionally, temporal trends of coarse- and ﬁne-mode dominant AOTs (CdAOT and FdAOT) have been estimated by applying an aerosol classiﬁcation based on accurate ˚AE and ˚Angstr¨om exponent difference ( ˚AED). In order to take into account the fact that cloud disturbance is having a sig- niﬁcant inﬂuence on the trend analysis of aerosols, we intro- duce a weighted least squares regression depending on two weights: (1) monthly standard deviation (σt) and (2) number of observations per month (nt). 1 Introduction Aerosols directly inﬂuence air quality and solar light ex- tinction as well as indirectly inﬂuence the cloud micro- physics and cloud radiative forcing (Twomey, 1977; Pincus and Baker, 1994; Albrecht, 1989; Ackerman et al., 2000; Haywood and Boucher, 2000). Considerable increase of an- thropogenic aerosol from human activities as well as emis- sion pattern change in natural aerosol due to climate change for past decades has complicated assessing the aerosol direct and indirect effects on the Earth climate system. These situ- ations lead to the fact that the impact of aerosols on climate change still remains at “med-low” or “low” level of scientiﬁc understanding (IPCC, 2007). Recently, several studies based on long-term records from the well-calibrated space instruments (e.g. Sea-viewing Wide Field-of-view Sensor – SeaWiFS, Multi-angle Imaging Spec- troRadiometer – MISR, and Moderate Resolution Imaging Spectroradiometer – MODIS, and Along Track Scanning Radiometer – ATSR) have contributed signiﬁcantly to the understanding of global aerosol trends (Yoon et al., 2011; Zhang and Reid, 2010; Yu et al., 2009; Karnieli et al., 2009; Thomas et al., 2010). Related to these trends, a hypothesis of global brightening or dimming has been discussed as well (Wild et al., 2005, 2007; Ohmura, 2006; Stanhill, 2007; Nor- ris and Wild, 2007; Streets et al., 2009). However, aerosol re- trievals based on satellite observations often have serious un- certainties caused by instrument calibration and assumptions within the algorithms (Li et al., 2009; Higurashi and Naka- jima, 1999; Ignatov and Stowe, 2002; Mishchenko et al., 1999a; Jeong et al., 2005; Kahn et al., 2005, 2007; Li et al., 2009; Kokhanovsky and de Leeuw, 2009; von Hoyningen- Huene et al., 2011). For polar-orbiting satellite observations, it is difﬁcult to avoid the bias in aerosol sampling caused by Temporal increase of FdAOTs (440 nm) prevails over newly industrializing countries in East Asia (weighted trends; +6.23 % yr−1 at Beijing) and active agricultural burn- ing regions in South Africa (+1.89 % yr−1 at Mongu). On the other hand, insigniﬁcant or negative trends for FdAOTs are detected over Western Europe (+0.25 % yr−1 at Avignon and −2.29 % yr−1 at Ispra) and North America (−0.52 % yr−1 for GSFC and −0.01 % yr−1 at MD Science Center). Over desert regions, both increase and decrease of CdAOTs (+3.37 % yr−1 at Solar Village and −1.18 % yr−1 at Oua- gadougou) are observed depending on meteorological conditions. J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent frequent cloud disturbance (Remer et al., 1997; Dubovik et al., 2002a; Jeong and Li, 2005; Jeong et al., 2005; Yoon et al., 2011) and coarse temporal resolution of the observation. Therefore, it is necessary to investigate aerosol trends based on ground-based observations. frequent cloud disturbance (Remer et al., 1997; Dubovik et al., 2002a; Jeong and Li, 2005; Jeong et al., 2005; Yoon et al., 2011) and coarse temporal resolution of the observation. Therefore, it is necessary to investigate aerosol trends based on ground-based observations. Over the last two decades, there have been several stud- ies of ˚AE curvature related to the aerosol size distribution (Kaufman, 1993; Eck et al., 1999; O’Neill et al., 2001a,b, 2003, 2005; Schuster et al., 2006; Gobbi et al., 2007; Reid et al., 1999; Kaskaoutis et al., 2007, 2011b). The relationships described in these papers provide a more suitable framework to classify aerosol types using ˚AE and ˚Angstr¨om exponent difference ( ˚AED) since it can minimize the data loss in the classiﬁcation process. Therefore, if considering the signiﬁ- cance of data number in the trend analysis, the present pa- per attempts to analyze the temporal trends of coarse- and ﬁne-mode dominant AOT (CdAOT and FdAOT) separately by applying such aerosol classiﬁcation. The AErosol RObotic NETwork (AERONET) program (http://aeronet.gsfc.nasa.gov/) aims to provide a global dis- tribution of aerosol optical properties and to validate satellite retrievals. Despite aerosols below clouds being underrepre- sented in the AERONET observation database (Remer et al., 1997; Dubovik et al., 2002a), this network of ground obser- vations provides suitable data for trend analysis of aerosol optical thickness (AOT) at main wavelengths (440, 675, 870, and 1020 nm) based on continuous long-term observations with high temporal resolution as well as high accuracy (ac- curacy; ±0.01) (Holben et al., 1998, 2001; Eck et al., 1999; Smirnov et al., 2000). Recently, Karnieli et al. (2009) and de Meij et al. (2010) have discussed AOT trends using long- term AERONET data and have compared them with satel- lite observations (e.g. MISR and MODIS) and model sim- ulations (e.g. Co-operative Programme for Monitoring and Evaluation of the Long-range Transmission of Air Pollutants – EMEP, the Region Emission Inventory – REAS – and the Intergovernmental Panel on Climate Change – RCP 3PD sce- nario). However, they have published no further information about the uncertainty of cloud disturbances and the inﬂu- ences of aerosol classiﬁcation in the AOT trend analysis. J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent The cloud disturbances lead to serious uncertainties in the trend analysis by decreasing the number of aerosol obser- vations (nt) per temporal interval during persistent cloudi- ness, thus leading to problems with the statistical representa- tiveness (Yoon et al., 2011). Therefore, this paper introduces weighted trends using monthly standard deviation and num- ber of observations to reduce cloud uncertainties in the trend analysis. The present study aims to investigate and analyze the long- term trends of AERONET level 2.0 AOT, CdAOT, FdAOT (440, 675, 870, and 1020 nm), and ˚AE (440–870 nm) at sev- eral stations. For this purpose, the second section describes in detail the methodology used for the selection of suitable AERONET stations, the weighted least squares regression to consider the cloud uncertainty, and the classiﬁcation of coarse- and ﬁne-mode dominant aerosols. In the third sec- tion, the aerosol trends at the speciﬁc AERONET stations are discussed regionally. The conclusions are summarized in the ﬁnal section. Published by Copernicus Publications on behalf of the European Geosciences Union. ublished by Copernicus Publications on behalf of the European Geosciences Union. Published by Copernicus Publications on behalf of the European Geosciences Union. 1272 2 Methodology For a reliable analysis of the aerosol trends based on the ground observation, new approaches are introduced: (1) the selection criteria for the AERONET stations having suf- ﬁcient and nearly-complete multi-year data sets, (2) the weighted least squares regression to consider cloud uncer- tainty, and (3) the classiﬁcation of coarse- and ﬁne-mode dominant aerosols. y Additionally, without applying a classiﬁcation of aerosol types, the trend studies are only of limited use in the un- derstanding why the aerosol loading changes in time. With this respect, spectral AOT observations are utilized to derive the ˚Angstr¨om exponent ( ˚AE) ( ˚Angstr¨om, 1929) indicating the mean size of particles. In general, submicron or supermi- cron aerosols have higher or lower ˚AEs accordingly. How- ever, ˚AE is not an ideal indicator to show the exact aver- age size of particles as it also is dependent on aerosol ab- sorption and size distribution (Kaskaoutis and Kambezidis, 2008). Another candidate for aerosol classiﬁcation is utiliz- ing the aerosol optical properties (e.g. volume size distribu- tion and single scattering albedo (SSA)) from AERONET inversion process. However, these could hardly be used to classify aerosol types because of additional retrieval ﬁlters resulting in a large loss of data. For example, the volume size distribution is only valid for solar zenith angle >50◦and SSA needs an additional criterion: AOT (440 nm) > 0.4 (Dubovik et al., 2000). 2.1 Selection criteria for suitable AERONET stations The AERONET program has provided high quality aerosol products for the past decades over roughly 850 global sta- tions. However, not all stations distribute a sufﬁciently large temporal record suitable for a trend analysis. Firstly, we dis- tinguished suitable AERONET stations having a sufﬁciently large record per month. The number of observations (nt) per month (t) basically depends on the seasonal daytime length, the station’s location, the operational instrument sta- tus, the cloud disturbance, and the veriﬁcation process of data quality. To obtain statistically meaningful monthly av- erage values, a large nt is highly required as the sample average based on a larger sample number is closer to the real average. Therefore, we have deﬁned the minimum nt of 300 per month (10 observations per day) to consider a reliable monthly average value. Another important issue in the trend analysis is that the annual data should be complete yearly sets in order to avoid a bias in particular seasons. In Atmos. Meas. Tech., 5, 1271–1299, 2012 www.atmos-meas-tech.net/5/1271/2012/ www.atmos-meas-tech.net/5/1271/2012/ J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1 1 2 Figure 1. The monthly observation numbers at the AEROET stations: (a) Avignon, (b) 3 Banizoumbou, (c) Beijing, (d) Dakar, (e) GSFC, (f) Ispra, (g) Mauna_Loa, (h) 4 Fig. 1. The monthly observation numbers of level 2.0 AOT (440 nm) at the AEROET stations: (a) Avignon, (b) Banizoumbou, (c) Be (d) Dakar, (e) GSFC, (f)Ispra, (g) Mauna Loa, (h) MD Science Center, (i) Mongu, (j) Ouagadougou, (k) SEDE BOKER, (l) Sevi (m) Shirahama, (n) Skukuza, and (o) Solar Village since 1993. The research period for each station is shown by the blue years at ve axis. 1273 el 2.0 AOT (440 nm) at the AEROET stations: (a) Avignon, (b) Banizoumbou, (c) Beijing, 2 Figure 1. The monthly observation numbers at the AEROET stations: (a) Avignon, (b) 3 Fig. 1. The monthly observation numbers of level 2.0 AOT (440 nm) at the AEROET stations: (a) Avignon, (b) Banizoumbou, (c) Beijing, (d) Dakar, (e) GSFC, (f)Ispra, (g) Mauna Loa, (h) MD Science Center, (i) Mongu, (j) Ouagadougou, (k) SEDE BOKER, (l) Sevilleta, (m) Shirahama, (n) Skukuza, and (o) Solar Village since 1993. The research period for each station is shown by the blue years at vertical 2 Figure 1. 2.1 Selection criteria for suitable AERONET stations The monthly observation numbers at the AEROET stations: (a) Avignon, (b) 3 Banizoumbou, (c) Beijing, (d) Dakar, (e) GSFC, (f) Ispra, (g) Mauna Loa, (h) 4 Fig. 1. The monthly observation numbers of level 2.0 AOT (440 nm) at the AEROET stations: (a) Avignon, (b) Banizoumbou, (c) Beijing, (d) Dakar, (e) GSFC, (f)Ispra, (g) Mauna Loa, (h) MD Science Center, (i) Mongu, (j) Ouagadougou, (k) SEDE BOKER, (l) Sevilleta, (m) Shirahama, (n) Skukuza, and (o) Solar Village since 1993. The research period for each station is shown by the blue years at vertical axis. gou, ( ) S _ O , ( ) Sev eta, ( ) since 1993. The research period for each Although a ﬁve-year time series may be insufﬁciently short for a statistically signiﬁcant trend analysis, it is a ﬁrst, prag- matic time span to investigate aerosol temporal change from AERONET observations. Figure 1 shows the nt since 1993 for suitable AERONET stations listed in Table 1. Because each station has a different observation history as well as dif- ferently qualiﬁed data sets (with respect to the above listed conditions), it is difﬁcult to perform the investigation of aerosol trends during the same time period. The research pe- riods when the data sets satisfy the selection criteria for each station are indicated by blue ﬁelds in Fig. 1. Detailed infor- mation about the geolocation and the research periods for the selected AERONET stations are listed in Table 1. _ _ , ( ) g , (j) g Shirahama, (n) Skukuza, and (o) Solar_Villa 6 station is shown by the blue years at vertical ax 7 other words, the absence of continuous monthly averages in the yearly data sets can cause a signiﬁcant uncertainty in the trend analysis. g , ( ) _ , ( ) , ( ) e since 1993. The research period for each Although a ﬁve-year time series may be insufﬁciently short for a statistically signiﬁcant trend analysis, it is a ﬁrst, prag- matic time span to investigate aerosol temporal change from AERONET observations. Figure 1 shows the nt since 1993 for suitable AERONET stations listed in Table 1. Because each station has a different observation history as well as dif- ferently qualiﬁed data sets (with respect to the above listed conditions), it is difﬁcult to perform the investigation of aerosol trends during the same time period. 2.1 Selection criteria for suitable AERONET stations The research pe- riods when the data sets satisfy the selection criteria for each station are indicated by blue ﬁelds in Fig. 1. Detailed infor- mation about the geolocation and the research periods for the selected AERONET stations are listed in Table 1. station is shown by the blue years at vertical ax 7 8 Basically, we have set up the following set of criteria to choose suitable AERONET stations: 9 1. The qualiﬁed monthly average is calculated with a nt larger than 300 per month (10 observations per day). 10 11 2. The complete yearly data set is composed out of more than seven qualiﬁed monthly averages. 3. A suitable AERONET station needs to have more than ﬁve complete years of observation history. 34 Atmos. Meas. Tech., 5, 1271–1299, 2012 www.atmos-meas-tech.net/5/1271/2012/ www.atmos-meas-tech.net/5/1271/2012/ J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1274 1274 J. Yoon et al.: Trend analysis of aerosol optical thickness and Angstrom expon 1 2 Figure 1. (Continued) 3 Fig. 1. Continued. 2 Fig. 1. Continued. 2 Fig. 1. Continued. 35 www.atmos-meas-tech.net/5/1271/2012/ 4 2.2 Weighted least squares regression sampling, so that a weighting factor is used to derive the re- spective trends. Figure 2 depicts the removal ratio of cloud (red line) and quality-unassured (yellow line) observations of AERONET data. If nt is large enough to ignore the other effects (daytime length, station location, and operational in- strument status), then main factors affecting nt are the veriﬁ- cation process of data quality and (mainly) the cloud distur- bance. In Fig. 2, the number of monthly level 2.0 data (nt) correlate negatively with the cloud removal ratio for most of the stations. 5 6 A simple linear model, which is used by minimizing chi- square error statistics, has been adopted in this study. Let Yt be the monthly AERONET level 2.0 AOT. The linear trend model is given by the following equation: 7 Yt = A + B Xt + εt, t = 1, ..., T, (1) (1) 8 9 where A is a constant term, B is the magnitude of the trend per year (Xt = t/12), εt is the noise, t is the month index, and T is the total number of months. However, as mentioned before, in order to analyze reliable trends, this simple model needs to take into account cloud disturbance. We introduce a weighting factor for the trend analysis: χ2(A, B) = T X t=1 (wtt × (Yt −A −B Xt))2 , (2) (2) 10 Each monthly AOT average has been calculated with dif- ferent nt, which is directly related to the number of cloud occurrence. The trend analysis based on monthly averages during cloudy season may strongly be biased through poor where, wtt (= √nt/σt) is the monthly weighting factor de- ﬁned as ratio of number of observations (nt) and monthly Atmos. Meas. Tech., 5, 1271–1299, 2012 35 www.atmos-meas-tech.net/5/1271/2012/ J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1275 J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1 Fig. 1. Continued. 1 Fig. 1. Continued. 1 Fig. 1. Continued. 1 Fig. 1. Continued. gu e . (Continued) le 1. Geolocations and research periods of the suitable AERONET stations for aerosol trend analysis in alpha . (Continued) ations and research periods of the suitable AERONET stations for aerosol trend analysis in alphabetical order. g ( ) 3 4 5 6 7 8 9 10 11 Table 1. 4 2.2 Weighted least squares regression Geolocations and research periods of the suitable AERONET stations for aerosol trend analysis in alphabetical order. Selected AERONET Regions Countries Geolocations Research stations (lat. [◦]/lon. [◦]/alt. [m]) periods (a) Avignon Western Europe France 43.93/4.88/32 2001 ∼2005 (b) Banizoumbou West Africa Niger 13.54/2.66/250 2002 ∼2008 (c) Beijing East Asia China 39.98/116.38/92 2003 ∼2007 (d) Dakar West Africa Senegal 14.39/−16.96/0 2004 ∼2008 (e) GSFC North America USA 38.99/−76.84/87 1995 ∼2008 (f) Ispra Western Europe Italy 45.80/8.63/235 2001 ∼2007 (g) Mauna Loa Free troposphere (Paciﬁc) USA 19.54/−155.58/3397 1998 ∼2009 (h) MD Science Center North America USA 39.28/−76.62/15 2000 ∼2006 (i) Mongu South Africa Zambia −15.25/23.15/1107 2000 ∼2004 (j) Ouagadougou West Africa Burkina Faso 12.20/−1.40/290 2000 ∼2004 (k) SEDE BOKER Middle East Israel 30.86/34.78/480 2003 ∼2008 (l) Sevilleta North America USA 34.35/−106.89/1477 1998 ∼2002 (m) Shirahama East Asia Japan 33.69/135.36/10 2003 ∼2009 (n) Skukuza South Africa South Africa −24.99/31.59/150 2000 ∼2007 (o) Solar Village Middle East Saudi Arabia 24.91/46.40/764 2001 ∼2007 12 13 14 standard deviation (σt). Monthly standard deviation is by it- self a suitable weight as it statistically shows the represen- tativeness (variability or diversity caused by cloud contam- ination or severe aerosol events) of the average. In the fol- lowing we will estimate the cloud uncertainty through the comparison between the weighted and unweighted trends. al., 2000, 2002a,b, 2006; Sinyuk et al., 2007) generates var- ious aerosol characteristics such as volume size distribution and SSA. However, the data are only provided for the fol- lowing conditions: solar zenith angle (θ) > 50◦for volume size distribution, and AOT (440 nm) > 0.4 and θ > 50◦for SSA (Dubovik et al., 2000). Figure 3 shows the normal- ized frequency of AOT (440 nm) and solar zenith angle at the selected AERONET stations. Generally, the normalized frequency (histogram) distributions of AOT (440 nm) are skewed and have long tails towards larger values of AOT. The percentage of the AERONET level 2.0 inversion data (volume size distribution and SSA) to the total observations is indicated as a pie chart on the lower-left hand side in Fig. 3. In most cases, it is difﬁcult to use the volume size distribution and SSA for aerosol classiﬁcation because of a 15 16 2.3 Classiﬁcation of coarse- and ﬁne-mode dominant aerosols 17 18 In order to quantify the change in anthropogenic (gener- ally, ﬁne-mode dominant) and natural (coarse-mode domi- nant) aerosols, an aerosol classiﬁcation needs to be applied as well. The AERONET inversion process (so-called version 2 Dubovik retrievals) (Dubovik and King, 2000; Dubovik et 36 Atmos. Meas. Tech., 5, 1271–1299, 2012 www.atmos-meas-tech.net/5/1271/2012/ J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent Figure 2. The removal ratio of cloud (red line) and quality-unassured (yellow line) cases to AERONET level 2 0 data (blue bar) within each of research period at the several AERONET . The removal ratio of cloud (red line) and quality-unassured (yellow line) cases to AERONET level 2.0 data (blue bar) within each of ch period at the several AERONET stations. Green bars mean that the observation numbers per month are over 1000 times. J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1276 3 Figure 2. The removal ratio of cloud (red line) and quality-unassured (yellow line) cases to 4 AERONET l l 2 0 d t (bl b ) ithi h f h i d t th l AERONET 5 Fig. 2. The removal ratio of cloud (red line) and quality-unassured (yellow line) cases to AERONET level 2.0 data (blue bar) within each of research period at the several AERONET stations. Green bars mean that the observation numbers per month are over 1000 times. Figure 2. The removal ratio of cloud (red line) and quality-unassured (yellow line) cases to 4 Fig. 2. The removal ratio of cloud (red line) and quality-unassured (yellow line) cases to AERONET level 2.0 data (blue bar) within each of research period at the several AERONET stations. Green bars mean that the observation numbers per month are over 1000 times. umbers per month are over 1000 times. the number of retrievals/observations for taking account of cloud disturbance. Therefore, if considering the signiﬁcance of data number in the weighted trend method and thereby try- ing to minimize the data loss, we propose a classiﬁcation of coarse- and ﬁne-mode dominant aerosols using ˚AE and ˚AED retrievals from AERONET level 2.0 direct sun data. ˚AE and stations. Green bars mean that the observation 6 7 8 9 10 low proportion to total observations meeting the conditions mentioned above. The AERONET also provides level 2.0 ﬁne/coarse mode AOTs (500 nm) determined by the spec- tral deconvolution algorithm (SDA) of O’Neill et al. (2003). These values are based on quadratic ﬁt of the spectral AOTs at 5 channels from 380 to 870 nm, and also agree well with the version 2 Dubovik retrievals of ﬁne/coarse mode AOTs (Eck et al., 2010). However, the additional data loss caused by the level 2.0 SDA criteria (http://aeronet.gsfc.nasa.gov/) makes it difﬁcult to analyze the weighted trend based on Atmos. Meas. Tech., 5, 1271–1299, 2012 37 www.atmos-meas-tech.net/5/1271/2012/ on et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1277 Figure 2. (Continued) . Continued. are deﬁned as: = −ln δλ1 δλ3 ln λ1 λ3 , (3) Several investigations have been previously devoted to th curvature of the spectral dependence of AOT in order t derive more accurate aerosol size information. J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent For exam ple, Kaufman (1993) found that the spectral curvature show J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1277 3 Fi 2 (C ti d) 4 Fig. 2. Continued. Figure 2 4 Fig. 2. Continued. Several investigations have been previously devoted to the curvature of the spectral dependence of AOT in order to derive more accurate aerosol size information. For exam- ple, Kaufman (1993) found that the spectral curvature shows a transition from mixed accumulation and coarse particle modes to a dominant accumulation mode. Eck et al. (1999) investigated the wavelength dependence of the optical depth of biomass burning, urban, and desert dust aerosols. O’Neill et al. (2001a,b, 2003, 2005) and Schuster et al. (2006) pre- sented a detailed analysis and compared simulations and ob- servations in order to investigate the relationship between 5 ˚AED are deﬁned as: 6 7 8 ˚AE = −ln δλ1 δλ3 ln λ1 λ3 , (3) ˚AED = −ln δλ1 δλ2 ln λ1 λ2 + ln δλ2 δλ3 ln λ2 λ3 , (4) (3) (4) 8 9 where, δλ is AOT at wavelengths (λ 1 = 440 nm, λ 2 = 675 nm, and λ 3 = 870 nm). www.atmos-meas-tech.net/5/1271/2012/ www.atmos-meas-tech.net/5/1271/2012/ www.atmos-meas-tech.net/5/1271/2012/ 38 Atmos. Meas. Tech., 5, 1271–1299, 2012 1278 J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1 2 Figure 2. (Continued) 3 Fig. 2. Continued. J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1278 2 Figure 2 (Continued) 3 Fig. 2. Continued. Fi 2 3 Fig. 2. Continued. Generally, coarse-mode (ﬁne-mode) dominant aerosols have been classiﬁed by lower (higher) values than ˚AE of at least 1.0 (Kaufman, 1993) or 1.4 (Tanr´e et al., 2001; Pereira et al., 2011; Shinozuka et al., 2011). However, one constant value of ˚AE is not a good threshold to classify the aerosol types (coarse and ﬁne dominant aerosols). Therefore, the new classiﬁcation criteria determined by 50 % ﬁne volume frac- tion effectively discriminates coarse- and ﬁne-mode domi- nant aerosols by higher accuracy (95.73 %) than using other constant ˚AEs (75.30 % for ˚AE of 1.0 and 80.82 % for ˚AE of 1.4) based on the Mie simulation in Fig. 4. Figure 5 shows additional Mie simulations (Mishchenko et al., 1999b, 2002) based on aerosol characteristics of typical aerosols (urban- industrial and mixed, biomass burning, desert dust, oceanic from Dubovik et al., 2002a) to examine the red classiﬁcation line (i.e. a variable ˚AE and ˚AED determined by 50 % ﬁne volume fraction). Coarse-mode dominant aerosols (desert dust and maritime aerosols) have smaller ˚AE and positive ˚AED according to the increase of aerosol loading. As already mentioned, the mean particle size of ﬁne-mode dominant aerosols could increase due to the increase of aerosol loading despite larger ﬁne volume fractions. In addition, the range of ˚AE (440–870 nm) for the typical aerosols from Dubovik et al. (2002a) (horizontal bar chart, bottom of Fig. 5) can ex- plain why only one constant value of ˚AE is not enough to 4 5 6 7 8 9 10 11 12 13 14 aerosol size distribution and spectral dependence of the AOT. Gobbi et al. (2007) have set up a useful straightforward graphical framework applicable to classify aerosol ﬁne mode fraction of the total AOT at 675 nm using ˚AED as a mea- sure of the curvature. They have applied the graphical frame- work to AERONET data and were able characterize different aerosol types such as pollution, mineral dust, and biomass burning. However, none of the above mentioned publications involved their methods in trend analyses. J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1279 J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1279 1 2 Figure 3. Normalized frequency of AOT at 440 nm (δ440) and solar zenith angle (θ) to total 3 observation number (N) at the several AERONET stations. The bin sizes for δ440 and θ are 4 0.01 and 1.0°, respectively. The circle diagram on the lower-left hand means the percentage of 5 AERONET l l 2 0 i i d ( l i di ib i d Si l S i 6 Fig. 3. Normalized frequency of AOT at 440 nm (δ440) and solar zenith angle (θ) to total observation number (N) at the several AERONET stations. The bin sizes for δ440 and θ are 0.01 and 1.0◦, respectively. The circle diagram on the lower-left hand means the percentage o AERONET level 2.0 inversion data (e.g. volume size distribution and single scattering albedo – SSA) to total observations. The volume size distribution is provided under θ > 50◦, and SSA is only valid for the criteria; δ440 > 0.4 and θ > 50◦. 1 2 Figure 3. Normalized frequency of AOT at 440 nm (δ440) and solar zenith angle (θ) to total 3 observation number (N) at the several AERONET stations. The bin sizes for δ440 and θ are 4 0 01 d 1 0° ti l Th i l di th l l ft h d th t f 5 Fig. 3. Normalized frequency of AOT at 440 nm (δ440) and solar zenith angle (θ) to total observation number (N) at the several AERONET stations. The bin sizes for δ440 and θ are 0.01 and 1.0◦, respectively. The circle diagram on the lower-left hand means the percentage of AERONET level 2.0 inversion data (e.g. volume size distribution and single scattering albedo – SSA) to total observations. The volume size distribution is provided under θ > 50◦, and SSA is only valid for the criteria; δ440 > 0.4 and θ > 50◦. lume size distribution and Single Scattering me size distribution is provided under θ >50°, nd θ >50°. border (black thick line) and their variations according to increase of aerosol loading are similar to the simulations. As can be seen in Fig. www.atmos-meas-tech.net/5/1271/2012/ In this study, we build up the classiﬁcation criteria determined by a similar approach as Gobbi et al. (2007). Even though this classiﬁ- cation technique might be problematic as compared to the Dubovik or O’Neill retrievals mentioned before since it re- lies on only 2 channel computations of ˚AE, this is the best way to consider the cloud effects by reducing the data loss in the trend analysis. With Fig. 4, using the Mie theory, we tested the relationship between ˚AE and ˚AED simulated with many bimodal volume size distributions consisting out of mode radii, widths, ﬁne volume fractions, and refractive indices (approximately 25 000 combinations) shown in Ta- ble 2. Usually, negative ˚AED shows a high proportion of ﬁne mode aerosol for the same ˚AE. In this study, we set up the classiﬁcation using both ˚AE and ˚AED by 50 % ﬁne volume fraction to total aerosol. 4 5 6 7 8 9 10 11 12 13 14 aerosol size distribution and spectral dependence of the AOT. Gobbi et al. (2007) have set up a useful straightforward graphical framework applicable to classify aerosol ﬁne mode fraction of the total AOT at 675 nm using ˚AED as a mea- sure of the curvature. They have applied the graphical frame- work to AERONET data and were able characterize different aerosol types such as pollution, mineral dust, and biomass burning. However, none of the above mentioned publications involved their methods in trend analyses. In this study, we build up the classiﬁcation criteria determined by a similar approach as Gobbi et al. (2007). Even though this classiﬁ- cation technique might be problematic as compared to the Dubovik or O’Neill retrievals mentioned before since it re- lies on only 2 channel computations of ˚AE, this is the best way to consider the cloud effects by reducing the data loss in the trend analysis. With Fig. 4, using the Mie theory, we tested the relationship between ˚AE and ˚AED simulated with many bimodal volume size distributions consisting out of mode radii, widths, ﬁne volume fractions, and refractive indices (approximately 25 000 combinations) shown in Ta- ble 2. Usually, negative ˚AED shows a high proportion of ﬁne mode aerosol for the same ˚AE. In this study, we set up the classiﬁcation using both ˚AE and ˚AED by 50 % ﬁne volume fraction to total aerosol. Atmos. Meas. Tech., 5, 1271–1299, 2012 39 www.atmos-meas-tech.net/5/1271/2012/ 39 www.atmos-meas-tech.net/5/1271/2012/ J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 6 for practically all AERONET stations, a difference in percentage of coarse- and ﬁne-mode dominant aerosols is observed due to different regional aerosol sources and atmospheric conditions. All stations over West Africa (Banizoumbou – C 99 % ≫F 1 %, Dakar – C 99 % ≫F 1 %, and Ouagadougou – C 97 % ≫F 3 %) and Middle East (SEDE BOKER – C 71 % > F 29 % and Solar Village – C 97 % ≫F 3 %) are inﬂuenced by coarse-mode dominant aerosols because the regions are close to deserts. In con- trast, industrial pollutant and biomass burning aerosols are dominant over Western Europe (Avignon – C 23 % < F 77 % and Ispra – C 13 % < F 87 %), South Africa (Mongu – AERONET level 2.0 inversion data (e.g. 6 Albedo (SSA)) to total observations. The vo 7 classify coarse- or ﬁne-mode dominant aerosols. These ten- dencies of ˚AE and ˚AED are more apparent when looking at application of AERONET data. and SSA is only valid for the criteria; δ440>0.4 8 Figure 6 shows a scatterplot of ˚AE and ˚AED derived from AERONET datasets at the ﬁfteen stations including the red classiﬁcation line. After applying the classiﬁcation, the percentages of coarse-mode (C) and ﬁne-mode (F) dominant aerosols are shown as a pie chart at the upper-left hand side of Fig. 6. In order to avoid ˚AE errors larger than 30 %, we only take into account those observations having AOT (440 nm) larger than 0.15 (Gobbi et al., 2007). Most ˚AE and ˚AED from AERONET observations in Fig. 6 are generally in good agreement with Mie simulations in Fig. 5. In other words, the majority of them are positioned within the simulation www.atmos-meas-tech.net/5/1271/2012/ Atmos. Meas. Tech., 5, 1271–1299, 2012 1280 J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1 2 Figure 3. (Continued) 3 4 5 6 Fig. 3. Continued. C 6 % ≪F 94 % and Skukuza – C 15 % < F 85 %), and North America (GSFC – C 9 % ≪F 91 % and MD Science Center – C 10 % < F 90 %). J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1280 1 2 Fi 3 (C i d) 3 Fig. 3. Continued Figure 3 3 Fig. 3. Continued. lar Village) might be insigniﬁcant because of incomplete yearly data sets. lar Village) might be insigniﬁcant because of incomplete yearly data sets. 4 5 6 7 8 C 6 % ≪F 94 % and Skukuza – C 15 % < F 85 %), and North America (GSFC – C 9 % ≪F 91 % and MD Science Center – C 10 % < F 90 %). Especially typical anthropogenic aerosols caused by urbanization and industrialization as well as natural aerosols brought in by strong westerly winds are observed over East Asia (Beijing – C 62 % > F 38 % and Shirahama – C 41 % < F 59 %). The classiﬁcation is not applicable to data observed at Mauna Loa and Sevilleta because most AOTs (440 nm) over these stations were less than 0.15. J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent Especially typical anthropogenic aerosols caused by urbanization and industrialization as well as natural aerosols brought in by strong westerly winds are observed over East Asia (Beijing – C 62 % > F 38 % and lar Village) might be insigniﬁcant because of incomplete yearly data sets. 3 Trend analysis J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 3 Trend analysis For this part of study, we have selected ﬁfteen AERONET stations providing datasets meeting the requirements spec- iﬁed in Sect. 2.1. In the following sections, the trends for the stations located in several regions (Western Europe, West Africa, South Africa, Middle East, East Asia, North Amer- ica, and Free troposphere/Paciﬁc) are discussed. The trends of ˚AE (440–870 nm) and AOT (440 nm) at the AERONET stations are shown in Fig. 7. For clariﬁcation, the error bar is scaled by a factor 10 of the standard error (σt/√nt), which is inversely used for the weighted trend analysis. Compari- son between unweighted (blue line and text on the left upper 9 10 11 By applying the aerosol classiﬁcation, it is possible to analyze more reliable trends separately for coarse- and ﬁne-mode dominant aerosols. However, the classiﬁcation is non-applicable for an extremely small aerosol load over Mauna Loa and Sevilleta stations and the trend analysis of FdAOT for stations dominated by coarse-mode aerosols (such as Banizoumbou, Dakar, Ouagadougou, and So- Atmos. Meas. Tech., 5, 1271–1299, 2012 www.atmos-meas-tech.net/5/1271/2012/ www.atmos-meas-tech.net/5/1271/2012/ J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1281 1 Figure 3. (Continued) 2 3 4 5 6 7 8 9 10 11 Fig. 3. Continued. 1 Figure 4. Simulations of the fine volume fraction as a function of Ångström Exponent (440- 2 870 nm) and Ångström Exponent Difference (ÅE(440-675 nm)-ÅE(675-870 nm)) using Mie 3 theory with all combinations of various aerosol optical properties shown in Table 2. 4 5 6 Fig. 4. Simulations of the ﬁne volume fraction as a function of ˚Angstr¨om exponent (440–870 nm) and ˚Angstr¨om exponent dif- ference ( ˚AE (440–675 nm) −˚AE (675–870 nm)) using Mie theory with all combinations of various aerosol optical properties shown in Table 2. 1 Figure 5. Mie simulations (solid, dotted, dashed, dash-dot-dot lines) and range of Ångström 2 Exponent (440-870 nm) for the typical aerosols (urban-industrial and mixed, biomass burning, 3 desert dust, oceanic) summarized in Dubovik et al. (2002). The red spot and red line represent 4 the simulations for AOT mean of the typical aerosols and the classification line for two 5 aerosol types (fine- and coarse-mode dominant aerosols) respectively 6 Fig. 5. Mie simulations (solid, dotted, dashed, dash-dot-dot lines) and range of ˚Angstr¨om exponent (440–870 nm) for the typical aerosols (urban-industrial and mixed, biomass burning, desert dust, oceanic) summarized in Dubovik et al. (2002a). J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1281 1 Figure 3 2 Fig. 3. Continued. Figure 3 2 Fig. 3. Continued. 3 4 5 6 7 8 9 10 11 Figure 4. Simulations of the fine volume fraction as a function of Ångström Exponent (440- 870 nm) and Ångström Exponent Difference (ÅE(440-675 nm)-ÅE(675-870 nm)) using Mie theory with all combinations of various aerosol optical properties shown in Table 2. Fig. 4. Simulations of the ﬁne volume fraction as a function of ˚Angstr¨om exponent (440–870 nm) and ˚Angstr¨om exponent dif- ference ( ˚AE (440–675 nm) −˚AE (675–870 nm)) using Mie theory with all combinations of various aerosol optical properties shown in Table 2. Figure 5. Mie simulations (solid, dotted, dashed, dash-dot-dot lines) and range of Ångström Exponent (440-870 nm) for the typical aerosols (urban-industrial and mixed, biomass burning, desert dust, oceanic) summarized in Dubovik et al. (2002). The red spot and red line represent the simulations for AOT mean of the typical aerosols and the classification line for two aerosol types (fine- and coarse-mode dominant aerosols), respectively. Fig. 5. Mie simulations (solid, dotted, dashed, dash-dot-dot lines) and range of ˚Angstr¨om exponent (440–870 nm) for the typical aerosols (urban-industrial and mixed, biomass burning, desert dust, oceanic) summarized in Dubovik et al. (2002a). The red spot and red line represent the simulations for AOT mean of the typical aerosols and the classiﬁcation line for two aerosol types (ﬁne- and coarse-mode dominant aerosols), respectively. 13 part) and weighted trends (red line and text on the right up- per part) allows to estimate the uncertainty caused by cloud disturbance in the trend analysis. 14 15 16 17 18 43 In this part, the main discussions of the aerosol trends are on the basis of the weighted trends of AOT (440 nm) and ˚AE (440–870 nm). As previously mentioned, a classiﬁcation of coarse- and ﬁne-mode dominant aerosols, (which is based on the Mie theory) is also introduced in the trend analysis in Fig. 9. Finally, the unweighted and weighted trends of ˚AE (440–870 nm), AOT, CdAOT, and FdAOT (440 nm) in percent for most of AERONET stations are indicated on the global map in Figs. 8 and 10, and all speciﬁc values of both trend analyses are summarized in Tables 3 and 4. 3.1 Western Europe 44 The averages of ˚AE shown in Fig. 7 for Avignon (< ˚AE> = 1.43) and Ispra (< ˚AE> = 1.51) over Western Eu- rope were inﬂuenced by industrial, urban, and trafﬁc pol- lutants, such as ammonium salts of sulphate and nitrate (Gonz´alez et al., 2000; Benkovitz et al., 1996; Kambezidis and Kaskaoutis, 2008; Mazzola et al., 2010). The seasonal variation of ˚AE is small as the major source during the year is industrial pollutant. On the other hand, the AOTs over West- ern Europe exhibit a signiﬁcant seasonal variation, which is increasing from spring to summer and decreasing from www.atmos-meas-tech.net/5/1271/2012/ 42 Atmos. Meas. Tech., 5, 1271–1299, 2012 3 Trend analysis The red spot and www.atmos-meas-tech.net/5/1271/2012/ J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1282 Figure 6. Applications of the classification method to the AERONET datasets separated into two aerosol types as fine- and coarse-mode dominant aerosols at the several stations. The i l di h l f h d h f (bl k) d fi d 6. Applications of the classiﬁcation method to the AERONET level 2.0 datasets separated into two aerosol types as ﬁne- and coarse-mo nant aerosols at the several stations. The circle diagram on the upper-left hand means the percentage of coarse (black) and ﬁne-mo dominant aerosols to total observations (N). To avoid large errors in ˚Angstr¨om exponent and its difference from low AOTs, on ONET level 2.0 data with AOT (440 nm) > 0.15 were used. Figure 6. Applications of the classification method to the AERONET datasets separated into 4 two aerosol types as fine- and coarse-mode dominant aerosols at the several stations. The 5 Fig. 6. Applications of the classiﬁcation method to the AERONET level 2.0 datasets separated into two aerosol types as ﬁne- and coarse-mode dominant aerosols at the several stations. The circle diagram on the upper-left hand means the percentage of coarse (black) and ﬁne-mode (red) dominant aerosols to total observations (N). To avoid large errors in ˚Angstr¨om exponent and its difference from low AOTs, only AERONET level 2.0 data with AOT (440 nm) > 0.15 were used. ). To avoid large errors in Ångström Exponent NET level 2.0 data with AOT (440 nm) > 0.15 rope in summer (Gerasopoulos et al., 2003; Bergamo et al., 2008; Venzac et al., 2009). The weighted AOT trends over both stations are insigniﬁcant or decreasing (+0.98 % yr−1 at Avignon and −2.30 % yr−1 at Ispra), most likely due to strict environmental regulations for mitigating climate change and improving air quality (Smith et al., 2001; Streets et al., 2006; Zhao et al., 2008). These tendencies are conﬁrmed once again in negligable or decreasing FdAOT trends at the (red) dominant aerosols to total observations 7 and its difference from low AOTs, only AER 8 were used. 9 10 autumn to winter (Fig. 7). Basically, the AOT depends on the aerosol extinction coefﬁcient (inﬂuenced, e.g. by aerosol types, emission intensity, and relative humidity) and bound- ary layer height. www.atmos-meas-tech.net/5/1271/2012/ Especially industrial pollutants composed of sulphur are enhanced during summer due to stronger solar radiation (Marmer et al., 2007; Karnieli et al., 2009). Addi- tionally, less removal process (e.g. rain/monsoon) as well as a higher boundary layer height causes higher AOTs over Eu- Atmos. Meas. Tech., 5, 1271–1299, 2012 45 www.atmos-meas-tech.net/5/1271/2012/ J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1283 Figure 6. (Continued) . Continued. ons (+0.25 % yr−1 at Avignon and −2.29 % yr−1 at Ispra own in Fig. 9). West Africa ral dust mainly from the Saharan and Sahel regions e most abundant aerosol type year-round inﬂuencing African monsoon in summer from Central Africa (Hao Liu, 1994) and emitted from agricultural activity during dry season (December–February) in West Africa (Johnso al., 2008a). The complex vertical distribution between m eral dust and biomass burning is a difﬁcult condition to sess the shortwave radiative effects in a difﬁcult way (Jo son et al., 2008b). Additionally, low AOT might occu 3 Fi 6 4 Fig. 6. Continued. 5 stations (+0.25 % yr−1 at Avignon and −2.29 % yr−1 at Ispra as shown in Fig. 9). 5 stations (+0.25 % yr−1 at Avignon and −2.29 % yr−1 at Ispra as shown in Fig. 9). African monsoon in summer from Central Africa (Hao and Liu, 1994) and emitted from agricultural activity during the dry season (December–February) in West Africa (Johnson et al., 2008a). The complex vertical distribution between min- eral dust and biomass burning is a difﬁcult condition to as- sess the shortwave radiative effects in a difﬁcult way (John- son et al., 2008b). Additionally, low AOT might occur as a result of efﬁcient wet removal of aerosol particles due to heavy precipitation (Reeves et al., 2010; Huang et al., 2009). In Fig. 7, series of these phenomena over West Africa rep- resent the seasonal pattern in ˚AE and AOT. Insigniﬁcant or 6 3.2 West Africa 7 8 9 Mineral dust mainly from the Saharan and Sahel regions is the most abundant aerosol type year-round, inﬂuencing Banizoumbou, Dakar, and Ouagadougou stations over West Africa (Prospero and Lamb, 2003; Washington and Todd, 2005; Moulin and Chiapello, 2004; Reeves et al., 2010). Be- sides, biomass burning is frequently advected by the West 46 Atmos. Meas. Tech., 5, 1271–1299, 2012 www.atmos-meas-tech.net/5/1271/2012/ 1284 J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1 2 Fi 6 (C i d) 3 Fig. 6. Continued. J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1284 Fi 6 3 Fig. 6. Continued. Fi 6 3 Fig. 6. Continued. 4 5 6 7 8 9 10 decreasing trends of dust aerosol due to decreasing dust ac- tivity (Evan et al., 2007) are observed over most stations in West Africa (+0.22 % yr−1 at Banizoumbou, −1.56 % yr−1 at Dakar, and −1.95 % yr−1 at Ouagadougou), which is con- sistent with the results from the Global Aerosol Climatol- ogy Project (GACP) data (Mishchenko and Geogdzhayev, 2007; Mishchenko et al., 2007), in situ measurement (Chi- apello et al., 2005), AVHRR (Zhao et al., 2008), and TOMS observations (Chiapello and Moulin, 2002). The weighted trends of ˚AE and AOT at Dakar and Ouagadougou are dif- ferent compared to unweighted trends due to frequent cloud disturbance. CdAOT trends over West Africa (+0.28 % yr−1 at Banizoumbou, −1.56 % yr−1 at Dakar, and −1.88 % yr−1 at Ouagadougou in Fig. 9) are generally similar with AOT trends. tible to local burning) are also inﬂuenced by aerosol mix- tures with fossil fuel burning, industrial pollutant, and Ae- olian coarse mode types (Eck et al., 2003). The cloud un- certainty in the trend analysis is insigniﬁcant because the biomass burning generally happens before the beginning of the rain seasons. A noticeable increase of AOT at Mongu (+2.26 % yr−1) is most likely affected by biomass burning (Mishchenko and Geogdzhayev, 2007; Zhao et al., 2008). Accordingly, positive trends of FdAOT over South Africa are observed (+1.89 % yr−1 at Mongu and +0.55 % yr−1 at Skukuza) in Fig. 9. 3.4 Middle East The stations SEDE BOKER and Solar Village are located within the Middle East and provide a long record of mea- surements because of stable and clear-sky weather conditions (Basart et al., 2009). In this region, aerosol size and com- position are dominated by ﬁne-mode pollution emitted from the regional petroleum industry (Zhao et al., 2008; Basart et al., 2009) and mineral dust transported from the Anatolian plateau, Sahara, Negev, and Arabian deserts (Kubilay et al., 2003; Derimian et al., 2006; Sabbah et al., 2006; Smirnov et al., 2002; Tafuro et al., 2006). The latter explains the clear periodical pattern of ˚AE and AOT seen in Fig. 7. The AOT J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent over SEDE BOKER tend to decrease (−2.16 % yr−1) due to a decrease in coarse particles (−2.89 % yr−1 for CdAOT in Fig. 9), while AOT over Solar Village show a strong in- crease (+3.29 % yr−1) in the weighted trend probably related to an increase of mineral dust (+3.37 % yr−1 for CdAOT) and change of atmospheric conditions (e.g. increase of wind speed and relative humidity) (Sabbah and Hasan, 2008). Interestingly, there are clear differences between the un- weighted and weighted trends at both stations due to a large standard error caused by high variability of ˚AE and AOT as well as a small number of observations due to low cloud cover. The weighted FdAOT trends of SEDE BOKER and Solar Village are insigniﬁcant (+0.75 % and −0.10 % yr−1, respectively, as seen in Fig. 9). the weighted trend is +4.59 % yr−1. This trend is consistent with many previous studies (Streets et al., 2000, 2003, 2006; Smith et al., 2001, 2003; Massie et al., 2004; Mishchenko and Geogdzhayev, 2007; Zhao et al., 2008). The weighted trend is larger than the unweighted one (+1.06 % yr−1), as there are smaller weighting factors due to frequent cloud dis- turbance during summer in Beijing. Such increase is also observed in both aerosol types and more pronounced for weighted trends of CdAOT and FdAOT with increases of +7.27 % and +6.23 % yr−1, respectively. Shirahama is in the middle of Japan, far-off large cities, facing the Paciﬁc Ocean (Mukai et al., 2006). Hence, maritime aerosol is predomi- nant, but there are occurrences of mineral dust and/or indus- trial aerosol transported by strong westerly winds from China (Sano et al., 2003; Mukai et al., 2005). The seasonal cycles of ˚AE and AOT in Fig. 7 are similar to those at Beijing due to similar meteorological conditions and aerosol sources. The upward trend of AOT is small (+0.44 % yr−1), while ˚AE in- creases (+2.07 % yr−1) clearly. Due to comparatively small number of observations over Shirahama, only slight differ- ences between weighted and unweighted trends of ˚AE and AOT are observed all year round (see Fig. 2). The magnitudes of the CdAOT and FdAOT trends are, after classiﬁcation, +1.81 % and −0.03 % yr−1, respectively. 3.5 East Asia Many emerging economies are found in East Asia, where, as a consequence, large amounts of anthropogenic aerosols are emitted. Additionally, mineral dust from the deserts in Mon- golia and in Western and Northern China (mainly the Takli- makan and Badain Juran deserts) contributes around 70 % of the total dust emissions in mid-latitude regions. Rapid des- ertiﬁcation caused by climatic variation and human activities additionally increases the aerosol burden due to mineral dust transport (Zhang et al., 2003). ˚AE and AOT at Beijing in Fig. 7 exhibit very clear seasonal cycles, which have been explained by the complex combination of natural and an- thropogenic aerosols, stagnant synoptic meteorological pat- terns, secondary aerosol formation, and hygroscopic growth (Kotchenruther et al., 1999; Dubovik et al., 2002a; Kim et al., 2007). Because of industrialization, urbanization, and desert- iﬁcation over East Asia in the last twenty years, the aerosol loading over Beijing increased rapidly and the magnitude of 11 3.3 South Africa 12 13 14 15 ˚AE as well as AOT over the stations in South Africa (i.e. Mongu and Skukuza) exhibit a strong seasonality due to pronounced wet and dry seasons (see Fig. 7), and the pres- ence of biomass burning aerosol (Tyson, 1986; Swap et al., 1996). Frequent occurrence of burning activity in warm and dry seasons leads to a large biomass burning (Eck et al., 2001). The regions close to Skukuza (where are less suscep- Atmos. Meas. Tech., 5, 1271–1299, 2012 47 www.atmos-meas-tech.net/5/1271/2012/ J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1285 Table 2. Bimodal lognormal volume size distribution ( dV (r) dln r ) parameters and refractive indices (Schuster et al., 2006) used to compute ˚AE (440-870 nm) and ˚AED ( ˚AE (440–675 nm) −˚AE (675–870 nm)) using Mie code in Fig. 4. Parameter∗ Values rﬁne 0.06, 0.09, 0.12, 0.15, 0.18, 0.21, 0.24, 0.27, 0.30 σﬁne 0.38, 0.50 rcoarse 1.9, 2.2, 2.7, 2.8, 3.0, 3.2, 3.4, 3.6, 3.7 σcoarse 0.75, 1.00 n 1.34, 1.37, 1.40, 1.43, 1.47, 1.50, 1.54 k 0.003 Cﬁne/Ctotal 0.01, 0.10, 0.20, 0.30, 0.40, 0.50, 0.60, 0.70, 0.80, 0.90, 0.99 ∗The bimodal lognormal volume size distribution ( dV (r) ) is given by Parameter∗ Values rﬁne 0.06, 0.09, 0.12, 0.15, 0.18, 0.21, 0.24, 0.27, 0.30 σﬁne 0.38, 0.50 rcoarse 1.9, 2.2, 2.7, 2.8, 3.0, 3.2, 3.4, 3.6, 3.7 σcoarse 0.75, 1.00 n 1.34, 1.37, 1.40, 1.43, 1.47, 1.50, 1.54 k 0.003 Cﬁne/Ctotal 0.01, 0.10, 0.20, 0.30, 0.40, 0.50, 0.60, 0.70, 0.80, 0.90, 0.99 ∗The bimodal lognormal volume size distribution ( dV (r) dln r ) is given by dV (r) dln r = Cﬁne √ 2π σﬁne exp " − ln r −ln rﬁne 2σ2 ﬁne # + Ccoarse √ 2π σcoarse exp " −(ln r −ln rcoarse) 2σ2coarse # , where Ctotal,ﬁne,coarse represents the particle volume concentration for total, ﬁne and coarse aerosol modes [µm3 µm−2], rﬁne,coarse is the median or geometric mean radius [µm], and σﬁne,coarse is the variance or width of each mode. n and k represent the real and imaginary parts of the complex refractive index, respectively. ∗The bimodal lognormal volume size distribution ( dV (r) dln r ) is given by dV (r) dln r = Cﬁne √ 2π σﬁne exp " − ln r −ln rﬁne 2σ2 ﬁne # + Ccoarse √ 2π σcoarse exp " −(ln r −ln rcoarse) 2σ2coarse # , where Ctotal,ﬁne,coarse represents the particle volume concentration for total, ﬁne and coarse aerosol modes [µm3 µm−2], rﬁne,coarse is the median or geometric mean radius [µm], and σﬁne,coarse is the variance or width of each mode. n and k represent the real and imaginary parts of the complex refractive index, respectively. where Ctotal,ﬁne,coarse represents the particle volume concentration for total, ﬁne and coarse aerosol modes [µm3 µm−2], rﬁne,coarse is the median or geometric mean radius [µm], and σﬁne,coarse is the variance or width of each mode. J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent n and k represent the real and imaginary parts of the complex refractive index, respectively. the weighted trend is +4.59 % yr−1. This trend is consistent with many previous studies (Streets et al., 2000, 2003, 2006; Smith et al., 2001, 2003; Massie et al., 2004; Mishchenko and Geogdzhayev, 2007; Zhao et al., 2008). The weighted trend is larger than the unweighted one (+1.06 % yr−1), as there are smaller weighting factors due to frequent cloud dis- turbance during summer in Beijing. Such increase is also observed in both aerosol types and more pronounced for weighted trends of CdAOT and FdAOT with increases of +7.27 % and +6.23 % yr−1, respectively. Shirahama is in the middle of Japan, far-off large cities, facing the Paciﬁc Ocean (Mukai et al., 2006). Hence, maritime aerosol is predomi- nant, but there are occurrences of mineral dust and/or indus- trial aerosol transported by strong westerly winds from China (Sano et al., 2003; Mukai et al., 2005). The seasonal cycles of ˚AE and AOT in Fig. 7 are similar to those at Beijing due to similar meteorological conditions and aerosol sources. The upward trend of AOT is small (+0.44 % yr−1), while ˚AE in- creases (+2.07 % yr−1) clearly. Due to comparatively small number of observations over Shirahama, only slight differ- ences between weighted and unweighted trends of ˚AE and AOT are observed all year round (see Fig. 2). The magnitudes of the CdAOT and FdAOT trends are, after classiﬁcation, +1.81 % and −0.03 % yr−1, respectively. over SEDE BOKER tend to decrease (−2.16 % yr−1) due to a decrease in coarse particles (−2.89 % yr−1 for CdAOT in Fig. 9), while AOT over Solar Village show a strong in- crease (+3.29 % yr−1) in the weighted trend probably related to an increase of mineral dust (+3.37 % yr−1 for CdAOT) and change of atmospheric conditions (e.g. increase of wind speed and relative humidity) (Sabbah and Hasan, 2008). Interestingly, there are clear differences between the un- weighted and weighted trends at both stations due to a large standard error caused by high variability of ˚AE and AOT as well as a small number of observations due to low cloud cover. The weighted FdAOT trends of SEDE BOKER and Solar Village are insigniﬁcant (+0.75 % and −0.10 % yr−1, respectively, as seen in Fig. 9). J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1286 Table 3. Total means of ˚Angstr¨om exponent (α) (440–870 nm) and AOTs (δλ) (440, 675, 870, and 1020 nm) and their unweighted/weighted trends at the several AERONET stations. J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent Total Unweighted Weighted Total Unweighted Weighted Total Unweighted Weighted mean trend [yr−1] trend [yr−1] mean trend [yr−1] trend [yr−1] mean trend [yr−1] trend [yr−1] (a) Avignon (b) Banizoumbou (c) Beijing α 1.43 +0.02242 (+1.57 %) +0.01364 (+0.95 %) 0.37 −0.00777 (−2.08 %) −0.01130 (−3.03 %) 1.12 −0.02200 (−1.97 %) −0.02308 (−2.07 %) δ440 0.20 +0.00020 (+0.10 %) +0.00194 (+0.98 %) 0.52 −0.00166 (−0.32 %) +0.00117 (+0.22 %) 0.68 +0.00717 (+1.06 %) +0.03100 (+4.59 %) δ675 0.11 −0.00024 (−0.22 %) +0.00037 (+0.34 %) 0.47 +0.00038 (+0.08 %) +0.00353 (+0.76 %) 0.43 +0.00636 (+1.48 %) +0.02448 (+5.71 %) δ870 0.08 −0.00066 (−0.85 %) −0.00079 (−1.02 %) 0.43 +0.00119 (+0.28 %) +0.00437 (+1.01 %) 0.32 +0.00665 (+2.06 %) +0.02023 (+6.26 %) δ1020 0.06 +0.00021 (+0.33 %) −0.00035 (−0.55 %) 0.41 +0.00056 (+0.14 %) +0.00347 (+0.85 %) 0.27 +0.00673 (+2.46 %) +0.01706 (+6.25 %) (d) Dakar (e) GSFC (f) Ispra α 0.36 −0.01415 (−3.95 %) −0.03264 (−9.12 %) 1.59 +0.00816 (+0.51 %) +0.00442 (+0.28 %) 1.51 −0.00038 (−0.03 %) −0.00178 (−0.12 %) δ440 0.46 −0.01186 (−2.60 %) −0.00710 (−1.56 %) 0.22 −0.00085(−0.38 %) −0.00122(−0.54 %) 0.30 −0.00337 (−1.13 %) −0.00687 (−2.30 %) δ675 0.41 −0.00938 (−2.31 %) −0.00251 (−0.62 %) 0.11 −0.00095 (−0.85 %) −0.00074 (−0.67 %) 0.16 −0.00206 (−1.26 %) −0.00418 (−2.57 %) δ870 0.38 −0.00697 (−1.86 %) +0.00104 (+0.28 %) 0.07 −0.00080 (−1.07 %) −0.00037 (−0.50 %) 0.11 −0.00101 (−0.90 %) −0.00262 (−2.34 %) δ1020 0.36 −0.00649 (−1.83 %) +0.00218 (+0.61 %) 0.06 −0.00062 (−1.05 %) −0.00020 (−0.34 %) 0.09 −0.00094 (−1.06 %) −0.00266 (−2.99 %) (g) Mauna Loa (h) MD Science Center (i) Mongu α 1.21 +0.01180 (+0.97 %) +0.01504 (+1.24 %) 1.68 −0.01228 (−0.73 %) −0.01480 (−0.88 %) 1.53 +0.01601 (+1.05 %) −0.00122 (−0.08 %) δ440 0.02 +0.00026 (+1.36 %) +0.00033 (+1.73 %) 0.24 −0.00177 (−0.74 %) −0.00003 (−0.01 %) 0.28 +0.01580 (+5.70 %) +0.00625 (+2.26 %) δ675 0.01 +0.00003 (+0.39 %) +0.00009 (+1.11 %) 0.12 −0.00089 (−0.75 %) −0.00049 (−0.41 %) 0.14 +0.00592 (+4.30 %) +0.00671 (+4.87 %) δ870 0.01 +0.00004 (+0.46 %) +0.00014 (+1.41 %) 0.08 −0.00020 (−0.26 %) +0.00010 (+0.13 %) 0.09 +0.00370 (+3.99 %) +0.00736 (+7.94 %) δ1020 0.01 +0.00003 (+0.28 %) +0.00008 (+0.91 %) 0.06 −0.00053 (−0.83 %) −0.00083 (−1.30 %) 0.07 +0.00175(+2.49 %) +0.00582(+8.29 %) (j) Ouagadougou (k) SEDE BOKER (l) Sevilleta α 0.42 +0.01331 (+3.14 %) +0.01695 (+4.00 %) 0.88 +0.00611 (+0.69 %) −0.02455 (−2.78 %) 1.19 +0.00516 (+0.43 %) −0.01980 (−1.66 %) δ440 0.51 +0.01747 (+3.42 %) −0.00996 (−1.95 %) 0.20 +0.00111 (+0.56 %) −0.00427 (−2.16 %) 0.08 +0.00271 (+3.28 %) −0.00313 (−3.79 %) δ675 0.47 −0.00367 (−0.78 %) −0.01168 (−2.48 %) 0.14 −0.00007 (−0.05 %) −0.00361 (−2.60 %) 0.05 +0.00096 (+1.98 %) −0.00162 (−3.35 %) δ870 0.41 +0.01427 (+3.48 %) −0.01028 (−2.51 %) 0.12 +0.00055 (+0.45 %) −0.00163 (−1.33 %) 0.04 +0.00115 (+3.01 %) −0.00095 (−2.49 %) δ1020 0.38 +0.01691 (+4.40 %) −0.00720 (−1.87 %) 0.11 −0.00032 (−0.30 %) −0.00112 (−1.05 %) 0.03 +0.00124 (+3.65 %) +0.00057 (+1.68 %) (m) Shirahama (n) Skukuza (o) Solar Village α 1.27 +0.01735 (+1.37 %) +0.02617 (+2.07 %) 1.34 +0.00566 (+0.42 %) −0.01080 (−0.80 %) 0.55 −0.02293 (−4.16 %) −0.00492 (−0.89 %) δ440 0.31 +0.00082 (+0.27 %) +0.00137 (+0.44 %) 0.23 −0.00113 (−0.48 %) +0.00118 (+0.51 %) 0.31 +0.01881 (+6.12 %) +0.01009 (+3.29 %) δ675 0.18 −0.00111 (−0.60 %) −0.00012 (−0.06 %) 0.12 −0.00126 (−1.01 %) −0.00037 (−0.29 %) 0.25 +0.01684 (+6.62 %) +0.00541 (+2.13 %) δ870 0.13 −0.00030 (−0.22 %) −0.00006 (−0.04 %) 0.09 −0.00034 (−0.39 %) +0.00006 (+0.07 %) 0.23 +0.01705 (+7.35 %) +0.00503 (+2.17) δ1020 0.11 −0.00150 (−1.35 %) −0.00165 (−1.49 %) 0.07 −0.00039 (−0.54 %) −0.00027 (−0.37 %) 0.23 +0.01373 (+6.03 %) +0.00140 (+0.62 %) Meas. J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent Tech., 5, 1271–1299, 2012 www.atmos-meas-tech.net/5/1271/2 (j) Ouagadougou (g) Mauna Loa 3.6 North America The stations GSFC and MD Science Center are located on urban and built-up land, while Sevilleta is positioned at shrub land over North America (Liu et al., 2004). Main aerosol type measured at GSFC and MD Science Center is urban- industrial pollution from vehicles and industries. The sea- sonal cycles of ˚AE and AOT in Fig. 7 demonstrate that the variabilities are strongly dependent on the combination www.atmos-meas-tech.net/5/1271/2012/ Atmos. Meas. Tech., 5, 1271–1299, 2012 J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent (k) SEDE BOKER Atmos. Meas. Tech., 5, 1271–1299, 2012 www.atmos-meas-tech.net/5/1271/2012/ J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1287 J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1287 Table 4. Total means of CdAOT and FdAOT (440, 675, 870, and 1020 nm) and their unweighted/weighted trends at the several AERONET stations except (g) Mauna Loa and (l) Sevilleta where the classiﬁcation is non-applicable. Table 4. Total means of CdAOT and FdAOT (440, 675, 870, and 1020 nm) and their unweighted/weighted trends at the several AERONET stations except (g) Mauna Loa and (l) Sevilleta where the classiﬁcation is non-applicable. Table 4. Total means of CdAOT and FdAOT (440, 675, 870, and 1020 nm) and their unweighted/weighted trends at the several AERONET stations except (g) Mauna Loa and (l) Sevilleta where the classiﬁcation is non-applicable. (k) SEDE BOKER p (g) ( ) pp Coarse-mode dominant AOT (CdAOT) Fine-mode dominant AOT (FdAOT) Total Unweighted Weighted Total Unweighted Weighted mean trend [yr−1] trend [yr−1] mean trend [yr−1] trend [yr−1] (a) Avignon δ440 0.14 +0.00374 (+2.59 %) +0.00323 (+2.23 %) 0.18 −0.00032 (−0.18 %) +0.00047 (+0.25 %) δ675 0.09 +0.00211 (+2.48 %) +0.00130 (+1.53 %) 0.09 −0.00062 (−0.66 %) −0.00028 (−0.30 %) δ870 0.07 +0.00125 (+1.87 %) +0.00056 (+0.83 %) 0.06 −0.00098 (−1.51 %) −0.00119 (−1.84 %) δ1020 0.06 +0.00187 (+3.19 %) +0.00102 (+1.75 %) 0.05 −0.00009(−0.18 %) −0.00079(−1.54 %) (b) Banizoumbou∗ δ440 0.52 −0.00160(−0.31 %) +0.00147(+0.28 %) 0.14 −0.00767 (−5.50 %) +0.00287 (+2.05 %) δ675 0.47 +0.00042 (+0.09 %) +0.00382 (+0.81 %) 0.10 −0.00420 (−4.23 %) +0.00256 (+2.57 %) δ870 0.43 +0.00122 (+0.28 %) +0.00467 (+1.08 %) 0.08 −0.00290 (−3.60 %) +0.00090 (+1.12 %) δ1020 0.41 +0.00059 (+0.14 %) +0.00378 (+0.92 %) 0.07 −0.00347 (−4.73 %) −0.00021 (−0.28 %) (c) Beijing δ440 0.47 −0.00391 (−0.84 %) +0.03398 (+7.27 %) 0.72 +0.02673 (+3.74 %) +0.04455 (+6.23 %) δ675 0.32 −0.00231 (−0.73 %) +0.02245 (+7.11 %) 0.43 +0.01856 (+4.30 %) +0.02199 (+5.10 %) δ870 0.26 −0.00081 (−0.32 %) +0.01749 (+6.84 %) 0.30 +0.01409 (+4.62 %) +0.01530 (+5.02 %) δ1020 0.23 −0.00026 (−0.11 %) +0.01485 (+6.52 %) 0.25 +0.01167 (+4.76 %) +0.01135 (+4.63 %) (d) Dakar∗ δ440 0.46 −0.01283 (−2.81 %) −0.00710 (−1.56 %) 0.17 −0.01210 (−7.29 %) −0.00723 (−4.36 %) δ675 0.41 −0.01049 (−2.58 %) −0.00258 (−0.63 %) 0.12 −0.00880(−7.23 %) −0.00511(−4.20 %) δ870 0.38 −0.00815 (−2.16 %) +0.00086 (+0.23 %) 0.10 −0.00528 (−5.27 %) −0.00398 (−3.98 %) δ1020 0.36 −0.00767 (−2.15 %) +0.00198 (+0.56 %) 0.09 −0.00483 (−5.35 %) −0.00378 (−4.19 %) (e) GSFC δ440 0.13 −0.00365 (−2.73 %) −0.00020 (−0.15 %) 0.22 −0.00031 (−0.14 %) −0.00116 (−0.52 %) δ675 0.07 −0.00324 (−4.38 %) −0.00009 (−0.12 %) 0.11 −0.00043 (−0.40 %) −0.00081 (−0.75 %) δ870 0.06 −0.00289 (−4.97 %) +0.00014 (+0.24 %) 0.07 −0.00028 (−0.40 %) −0.00046 (−0.64 %) δ1020 0.05 −0.00226 (−5.27 %) +0.00031 (+0.62 %) 0.05 −0.00010 (−0.18 %) −0.00028 (−0.52 %) (f) Ispra δ440 0.16 +0.00368 (+2.26 %) −0.00193 (−1.18 %) 0.29 −0.00343 (−1.19 %) −0.00659 (−2.29 %) δ675 0.10 +0.00153 (+1.56 %) −0.00266 (−2.69 %) 0.15 −0.00198 (−1.30 %) −0.00423 (−2.78 %) δ870 0.08 +0.00136 (+1.76 %) −0.00155 (−2.01 %) 0.10 −0.00088 (−0.87 %) −0.00228 (−2.25 %) δ1020 0.07 +0.00099 (+1.46 %) −0.00156 (−2.30 %) 0.08 −0.00076 (−0.97 %) −0.00192 (−2.44 %) (h) MD Science Center δ440 0.14 −0.00151 (−1.08 %) +0.00131 (+0.94 %) 0.24 −0.00147 (−0.61 %) −0.00003 (−0.01 %) δ675 0.08 −0.00121 (−1.60 %) +0.00050 (+0.66 %) 0.12 −0.00061 (−0.53 %) +0.00022 (+0.19 %) δ870 0.06 −0.00065 (−1.14 %) +0.00096 (+1.69 %) 0.08 +0.00007 (+0.09 %) +0.00077 (+1.03 %) δ1020 0.05 −0.00107 (−2.06 %) +0.00003 (+0.06 %) 0.06 −0.00027 (−0.45 %) −0.00023 (−0.38 %) (i) Mongu δ440 0.18 −0.01204 (−6.57 %) −0.00055 (−0.30 %) 0.27 +0.01689 (+6.26 %) +0.00509 (+1.89 %) δ675 0.11 −0.00797 (−7.39 %) +0.00036 (+0.33 %) 0.13 +0.00667 (+5.12 %) +0.00552 (+4.23 %) δ870 0.09 −0.00559 (−6.54 %) +0.00141 (+1.65 %) 0.09 +0.00439 (+5.15 %) +0.00603 (+7.08 %) δ1020 0.07 −0.00571 (−7.91 %) +0.00107 (+1.48 %) 0.06 +0.00245 (+3.92 %) +0.00418 (+6.69 %) (j) Ouagadougou∗ δ440 0.51 +0.01705 (+3.32 %) −0.00964 (−1.88 %) 0.22 −0.02445 (−11.30 %) +0.00179 (+0.83 %) δ675 0.48 −0.01380 (−2.85 %) −0.00043 (−0.09 %) 0.19 −0.05651 (−29.18 %) −0.00065 (−0.34 %) δ870 0.41 +0.01405 (+3.41 %) −0.00997 (−2.42 %) 0.14 −0.03348 (−24.50 %) −0.00240 (−1.76 %) δ1020 0.39 +0.01663 (+4.30 %) −0.00726 (−1.88 %) 0.12 −0.02768 (−23.17 %) −0.00109 (−0.91 %) Atmos. (k) SEDE BOKER However, these increasing tendencies at Mauna Loa might be insigniﬁcant because most of AOTs (440 nm) are close to the observation uncertainty (±0.01) (Eck et al., 1999). of natural and anthropogenic aerosols, fuel types, emission characteristic, relative humidity, boundary layer depth, and scavenging by precipitation (Glen et al., 1996; Chen et al., 2001; Dubovik et al., 2002a; Andronache, 2004). The neg- ative trends of AOT (−0.54 % and −0.01 % yr−1 at GSFC and MD Science Center) are consistent with the decrease of industrial emissions in the United States of America (Smith et al., 2001; Streets et al., 2006; Zhao et al., 2008). Sevil- leta station measures a relatively small aerosol loading for the considered time span. The weighted trends are strongly negative (−1.66 % yr−1 for ˚AE and −3.79 % yr−1 for AOT), clearly different from the unweighted ones (+0.43 % yr−1 for ˚AE and +3.28 % yr−1 for AOT). In most cases of monthly averaged AOT at Sevilleta, the values are lower than 0.15, so that the classiﬁcation for the trend analysis was not ap- plied to data of this station. The majority of retrieved aerosol type at GSFC and MD Science Center is ﬁne-mode, and the weighted trends of FdAOT are −0.52 % and −0.01 % yr−1, respectively (see Fig. 9). (k) SEDE BOKER The free troposphere is characterized by being almost cloud-free in the subsiding branch of the Hadley cell as well as a pathway for long-distance transport of aerosols (Garstang and Fitzjarrald, 1999; Schmeissner et al., 2011). In most cases, free tropospheric AOT (440 nm) does not exceed values of 0.05 except when affected by volcano eruption or transported mineral dust and pollution. Therefore, it is difﬁcult to analyze size and type, as the er- ror in ˚AE from low AOTs could be signiﬁcant (Gobbi et al., 2007; Kaskaoutis et al., 2011a). The main factors affect- ing the seasonal pattern of AOT are most likely long-range transported aerosols and seasonal meteorological conditions; the AOT trend for this station is positive (+1.73 % yr−1). In addition, in order to investigate the trends in stratospheric or free tropospheric aerosols, the AOT data from March to May were excluded, and we found that AOTs at 440 nm are increasing by +0.00047 (+3.03 %) for unweighted and +0.00049 (+3.16 %) per year for weighted trend analysis. However, these increasing tendencies at Mauna Loa might be insigniﬁcant because most of AOTs (440 nm) are close to the observation uncertainty (±0.01) (Eck et al., 1999). of Asian mineral dust and pollution in spring (Perry et al., 1999; Eck et al., 2005). The free troposphere is characterized by being almost cloud-free in the subsiding branch of the Hadley cell as well as a pathway for long-distance transport of aerosols (Garstang and Fitzjarrald, 1999; Schmeissner et al., 2011). In most cases, free tropospheric AOT (440 nm) does not exceed values of 0.05 except when affected by volcano eruption or transported mineral dust and pollution. Therefore, it is difﬁcult to analyze size and type, as the er- ror in ˚AE from low AOTs could be signiﬁcant (Gobbi et al., 2007; Kaskaoutis et al., 2011a). The main factors affect- ing the seasonal pattern of AOT are most likely long-range transported aerosols and seasonal meteorological conditions; the AOT trend for this station is positive (+1.73 % yr−1). In addition, in order to investigate the trends in stratospheric or free tropospheric aerosols, the AOT data from March to May were excluded, and we found that AOTs at 440 nm are increasing by +0.00047 (+3.03 %) for unweighted and +0.00049 (+3.16 %) per year for weighted trend analysis. (k) SEDE BOKER Meas. Tech., 5, 1271–1299, 2012 www.atmos-meas-tech.net/5/1271/2012/ www.atmos-meas-tech.net/5/1271/2012/ J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 288 J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exp able 4. Continued. Coarse-mode dominant AOT (CdAOT) Fine-mode dominant AOT (FdAOT) Total Unweighted Weighted Total Unweighted Weighted mean trend [yr−1] trend [yr−1] mean trend [yr−1] trend [yr−1] (k) SEDE BOKER δ440 0.19 +0.00156 (+0.81 %) −0.00560 (−2.89 %) 0.14 +0.00190 (+1.38 %) +0.00104 (+0.75 %) δ675 0.14 +0.00052 (+0.37 %) −0.00528 (−3.77 %) 0.08 +0.00104 (+1.27 %) +0.00016 (+0.20 %) δ870 0.13 +0.00112 (+0.89 %) −0.00383 (−3.04 %) 0.07 +0.00181 (+2.64 %) +0.00114 (+1.67 %) δ1020 0.11 +0.00020 (+0.18 %) −0.00350 (−3.16 %) 0.05 +0.00101 (+1.86 %) +0.00108 (+2.00 %) (m) Shirahama δ440 0.23 +0.00706 (+3.05 %) +0.00420 (+1.81 %) 0.31 +0.00046 (+0.15 %) −0.00009 −0.03 %) δ675 0.15 +0.00294 (+1.99 %) +0.00233 (+1.58 %) 0.17 −0.00163 (−0.95 %) −0.00089 (−0.52 %) δ870 0.12 +0.00253 (+2.13 %) +0.00219 (+1.84 %) 0.12 −0.00073 (−0.61 %) −0.00058 (−0.48 %) δ1020 0.10 +0.00075 (+0.73 %) +0.00059 (+0.58 %) 0.09 −0.00183 (−1.96 %) −0.00193 (−2.06 %) (n) Skukuza δ440 0.13 −0.00111 (−0.83 %) −0.00045 (−0.34 %) 0.23 −0.00083 (−0.36 %) +0.00127 (+0.55 %) δ675 0.08 −0.00183 (−2.16 %) −0.00141 (−1.67 %) 0.12 −0.00104 (−0.86 %) −0.00033 (−0.27 %) δ870 0.07 −0.00128 (−1.95 %) −0.00024 (−0.37 %) 0.08 −0.00014 (−0.17 %) −0.00004 (−0.05 %) δ1020 0.06 −0.00152 (−2.58 %) −0.00049 (−0.84 %) 0.07 −0.00020 (−0.29 %) −0.00036 (−0.53 %) (o) Solar Village∗ δ440 0.31 +0.01848 (+6.00 %) +0.01037 (+3.37 %) 0.13 −0.00487 (−3.72 %) −0.00013 (−0.10 %) δ675 0.26 +0.01649 (+6.46 %) +0.00586 (+2.29 %) 0.10 −0.00431 (−4.30 %) −0.00016 (−0.16 %) δ870 0.23 +0.01657 (+7.11 %) +0.00524 (+2.25 %) 0.08 −0.00208 (−2.46 %) +0.00064 (+0.76 %) δ1020 0.23 +0.01337 (+5.85 %) +0.00195 (+0.85 %)) 0.09 −0.00553 (−6.30 %) −0.00155 (−1.77 %) ∗Trend analysis of FdAOTs represented in italic type might be insigniﬁcant because of incomplete yearly data sets. 1288 Table 4. Continued. Table 4. Continued. ∗Trend analysis of FdAOTs represented in italic type might be insigniﬁcant because of incomplete yearly data sets. of Asian mineral dust and pollution in spring (Perry et al., 1999; Eck et al., 2005). 3.7 Mauna Loa Aerosols measured at Mauna Loa (alt. ∼3397 m) in the Pa- ciﬁc are representative for free tropospheric aerosols. How- ever, the station in the tropical mid-Paciﬁc is also under some inﬂuence of long-range transport (over 6000 to 8000 km) Atmos. Meas. Tech., 5, 1271–1299, 2012 www.atmos-meas-tech.net/5/1271/2012/ J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1289 J. Yoon et al.: Trend analysis of aerosol optical thickness and Angstrom exponent 12 1 2 3 Figure 7. Unweighted (blue one on the left upper part) and weighted (red one on the right 4 upper part) trends of Ångström Exponent (440-870 nm) (α) and AOT (440 nm) (δ440) at the 5 Fig. 7. Unweighted (blue one on the left upper part) and weighted (red one of the richt upper part) trends of ˚Angstr¨om Exponent (440–870 (α) amd AOT (440 nm) (δ440) at the several AERONET stations. The total means of α and δ440 (black one enclosed with parentheses) shown on right vertical axis. The error bar means the 10 times of the standard error, which are used for the weighted trend analysis. 3 i i h d bl h l f d i h d d f h i h d f ˚ Figure 7. Unweighted (blue one on the left upper part) and weighted (red one on the right 4 upper part) trends of Ångström Exponent (440-870 nm) (α) and AOT (440 nm) (δ440) at the 5 Fig. 7. Unweighted (blue one on the left upper part) and weighted (red one of the richt upper part) trends of ˚Angstr¨om Exponent (440–870 nm) (α) amd AOT (440 nm) (δ440) at the several AERONET stations. The total means of α and δ440 (black one enclosed with parentheses) are shown on right vertical axis. The error bar means the 10 times of the standard error, which are used for the weighted trend analysis. several AERONET 6 h ) h 7 4 Summary and conclusion of α and δ440 (black one enclosed with e error bar means the 10 times of the standard ysis. each AERONET station due to the different observation his- tory and condition (see Table 1 and Fig. 1). Weighted trends were derived utilizing monthly standard deviation and num- ber of observations (nt) providing an estimate of trend uncer- tainty (primarily) due to cloud disturbance. For example, if there was a high variability of aerosol loading for a small nt, then signiﬁcant difference between unweighted and weighted trends was observed. parentheses) are shown on right vertical axis. T 7 error, which are used for the weighted trend an 8 9 10 In this study, long-term trends of the level 2.0 ˚Angstr¨om ex- ponent ( ˚AE) (440–870 nm), aerosol optical thickness (AOT), and coarse- and ﬁne-mode dominant AOTs (CdAOT and FdAOT) (440, 675, 870, and 1020 nm) observed at several AERONET stations have been analyzed. Firstly, suitable AERONET stations providing sufﬁciently long-term data se- ries were chosen in order to make a meaningful trend anal- ysis. Unfortunately, the research periods were different for In general, cloud-free aerosol trends in this study are con- sistent with results from other papers (Zhao et al., 2008; www.atmos-meas-tech.net/5/1271/2012/ 48 Atmos. Meas. Tech., 5, 1271–1299, 2012 www.atmos-meas-tech.net/5/1271/2012/ J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om expon 1290 J. Yoon et al.: Trend analysis of aerosol optical thickness and Angstrom expo Figure 7. (Continued) Continued. li et al., 2009; Yoon et al., 2011; de Meij et al., 2010). aerosol loading tends to decrease over West Africa, n Europe and North America, while it is increasing outh Africa and East Asia. Consistent with our ex- ons ﬁne mode dominant aerosol loading primaril rapid desertiﬁcation plays a role are characterized by a siderable increase of coarse-mode dominant aerosol. A applying classiﬁcation, these trends are more apparen particular, the weighted trends of both CdAOT and Fd at Beijing o er East Asia sho considerable increases 49 1 2 3 Figure 7. (Continued) 4 5 6 7 8 9 10 Fig. 7. Continued. Karnieli et al., 2009; Yoon et al., 2011; de Meij et al., 2010). Total aerosol loading tends to decrease over West Africa, Western Europe and North America, while it is increasing over South Africa and East Asia. Consistent with our ex- pectations, ﬁne-mode dominant aerosol loading, primarily created by human activities, is decreasing over those coun- tries having introduced environmental regulations, while it is increasing over regions in emerging economies without such strict regulations. Temporal variation of the loading of coarse-mode dominant aerosol depends strongly on meteo- rological conditions varying with climate change. In partic- ular, those AERONET stations close to the regions where rapid desertiﬁcation plays a role are characterized by a con- siderable increase of coarse-mode dominant aerosol. After applying classiﬁcation, these trends are more apparent. In particular, the weighted trends of both CdAOT and FdAOT at Beijing over East Asia show considerable increases. Due to the high density of population in many strongly pol- luted areas (e.g. 1300 people per 1 km2 in Beijing) and the correlation between aerosol load and mortality (Foster and Kumar, 2011), there is an urgent need for measures to re- duce the aerosol load in large urban agglomerations (aka megacities) worldwide similar to those already introduced in Western Europe. Atmos Meas Tech 5 1271 1299 2012 www atmos meas tech net/5/1271/2012/ 3 Figure 7 4 Fig. 7. Continued. 5 6 7 8 9 10 Karnieli et al., 2009; Yoon et al., 2011; de Meij et al., 2010). Figure 7 3 Fig. 7. Continued. Figure 7 3 Fig. 7. Continued. 4 5 6 7 8 Acknowledgements. This work was supported in part by the CityZen project (megaCITY – Zoom for the Environment: EU Framework Programme 7 of European Commission), the DFG Project Terra, and the University of Bremen. The authors would like to thank NASA AERONET team for the provision of aerosol optical properties data. We are thankful to M. I. Mishchenko for providing the Mie code used in this study. The authors sincerely appreciate your insightful comments and suggestions of the editor (G. de Leeuw), two anonymous referees, and other advisors (D. G. Kaskaoutis and O. Dubovik) for revising the paper. 4 5 6 7 8 Acknowledgements. This work was supported in part by the CityZen project (megaCITY – Zoom for the Environment: EU Framework Programme 7 of European Commission), the DFG Project Terra, and the University of Bremen. The authors would like to thank NASA AERONET team for the provision of aerosol optical properties data. We are thankful to M. I. Mishchenko for providing the Mie code used in this study. The authors sincerely appreciate your insightful comments and suggestions of the editor (G. de Leeuw), two anonymous referees, and other advisors (D. G. Kaskaoutis and O. Dubovik) for revising the paper. Basart, S., P´erez, C., Cuevas, E., Baldasano, J. M., and Gobbi, G. P.: Aerosol characterization in Northern Africa, Northeastern Atlantic, Mediterranean Basin and Middle East from direct-sun AERONET observations, Atmos. Chem. Phys., 9, 8265–8282, doi:10.5194/acp-9-8265-2009, 2009. Benkovitz, C. M., Scholtz, M. T., Pacyna, J., Tarrason, L., Dignon, J., Voldner, E. C., Spiro, P. A., Logan, J. A., and Graedel, T. E.: Global gridded inventories of antropogenic emissions of sulphur and nitrogen, J. Geophys. Res., 101, 29239–29253, doi:10.1029/96JD00126, 1996. Bergamo, A., Tafuro, A. M., Kinne, S., De Tomasi, F., and Perrone, M. R.: Monthly-averaged anthropogenic aerosol direct radiative forcing over the Mediterranean based on AERONET aerosol properties, Atmos. Chem. Phys., 8, 6995–7014, doi:10.5194/acp- 8-6995-2008, 2008. 9 Edited by: G. de Leeuw 10 References Chen, L.-W. A., Doddridge, B. G., Dickerson, R. R., Chow, J. C., Mueller, P. K., Quinn, J., and Butler, W. A.: Seasonal variations in elemental carbon aerosol, carbon monoxide and sulfur diox- ide: Implications for sources, Geophys. Res. Lett., 28, 1711– 1714, doi:10.1029/2000GL012354, 2001. 11 12 Ackerman, A. S., Toon, O. B., Stevens, D. E., Heyms- ﬁeld, A. J., Ramanathan, V., and Welton, E. J.: Reduc- tion of tropical cloudiness by soot, Science, 288, 1042–1047, doi:10.1126/science.288.5468.1042, 2000. 13 Albrecht, B.: Aerosols, cloud microphysics and fractional cloudiness, Science, 245, 1227–1230, doi:10.1126/science.245.4923.1227, 1989. Chiapello, J. and Moulin, C.: TOMS and Meteosat satellite records of the variability of Saharan dust transport over the Atlantic dur- ing the last two decades (1979–1997), Geophys. Res. Lett., 29, 1176, doi:10.1029/2001GL013767, 2002. 14 Andronache, C.: Estimates of sulfate aerosol wet scavenging coefﬁ- cient for locations in the Eastern United States, Atmos. Environ., 38, 795–804, doi:10.1016/j.atmosenv.2003.10.035, 2004. Chiapello, I., Moulin, C., and Prospero, J. M.: Understand- ing the long-term variability of African dust transport across the Atlantic as recorded in both Barbados surface concen- 15 ˚Angstr¨om, A.: On the atmospheric transmission of sun radiation and on dust in the air, Geograf. Ann. Deut., 11, 156–166, 1929. 50 Atmos. Meas. Tech., 5, 1271–1299, 2012 www.atmos-meas-tech.net/5/1271/2012/ Total aerosol loading tends to decrease over West Africa, Western Europe and North America, while it is increasing over South Africa and East Asia. Consistent with our ex- pectations, ﬁne-mode dominant aerosol loading, primarily created by human activities, is decreasing over those coun- tries having introduced environmental regulations, while it is increasing over regions in emerging economies without such strict regulations. Temporal variation of the loading of coarse-mode dominant aerosol depends strongly on meteo- rological conditions varying with climate change. In partic- ular, those AERONET stations close to the regions where rapid desertiﬁcation plays a role are characterized by a con- siderable increase of coarse-mode dominant aerosol. After applying classiﬁcation, these trends are more apparent. In particular, the weighted trends of both CdAOT and FdAOT at Beijing over East Asia show considerable increases. Due to the high density of population in many strongly pol- luted areas (e.g. 1300 people per 1 km2 in Beijing) and the correlation between aerosol load and mortality (Foster and Kumar, 2011), there is an urgent need for measures to re- duce the aerosol load in large urban agglomerations (aka megacities) worldwide similar to those already introduced in Western Europe. Atmos. Meas. Tech., 5, 1271–1299, 2012 49 www.atmos-meas-tech.net/5/1271/2012/ J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 129 1 2 Figure 7. (Continued) 3 4 5 6 7 8 Fig. 7. Continued. Acknowledgements. This work was supported in part by the CityZen project (megaCITY – Zoom for the Environment: EU Framework Programme 7 of European Commission), the DFG Project Terra, and the University of Bremen. The authors would like to thank NASA AERONET team for the provision of aerosol optical properties data. We are thankful to M. I. Mishchenko for providing the Mie code used in this study. The authors sincerely appreciate your insightful comments and suggestions of the editor (G. de Leeuw), two anonymous referees, and other advisors (D. G. Kaskaoutis and O. Dubovik) for revising the paper. Basart, S., P´erez, C., Cuevas, E., Baldasano, J. M., and Gobb G. P.: Aerosol characterization in Northern Africa, Northeaster Atlantic, Mediterranean Basin and Middle East from direct-su AERONET observations, Atmos. Chem. Phys., 9, 8265–828 doi:10.5194/acp-9-8265-2009, 2009. Benkovitz, C. M., Scholtz, M. T., Pacyna, J., Tarrason, L., Digno J., Voldner, E. C., Spiro, P. A., Logan, J. A., and Graede T. E.: Global gridded inventories of antropogenic emission of sulphur and nitrogen, J. Geophys. 9 Edited by: G. de Leeuw www.atmos-meas-tech.net/5/1271/2012/ Res., 101, 29239–2925 doi:10.1029/96JD00126, 1996. Bergamo, A., Tafuro, A. M., Kinne, S., De Tomasi, F., and Perron J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1291 1 2 Fi 7 (C ti d) 3 Fig. 7. Continued. 2 Figure 7 3 Fig. 7. Continued. www.atmos-meas-tech.net/5/1271/2012/ J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1292 1 2 Figure 8. Unweighted and weighted trends of ÅE (440-870 nm) (left square) and AOT (440 3 g. 8. Unweighted and weighted trends of ˚AE (440–870 nm) (left square) and AOT (440 nm) (right diamond) in percent at the major stations cept (a) Avignon over Western Europe, (h) MD Science Center over North America, and (j) Ouagadougou over West Africa. 2 Figure 8. Unweighted and weighted trends of ÅE (440-870 nm) (left square) and AOT (440 3 Fig. 8. Unweighted and weighted trends of ˚AE (440–870 nm) (left square) and AOT (440 nm) (right diamond) in percent at the major stations except (a) Avignon over Western Europe, (h) MD Science Center over North America, and (j) Ouagadougou over West Africa. (h) MD_Science_Center over North America, 5 6 trations and large-scale total ozone mapping spectrometer (TOMS) optical thickness, J. Geophys. Res., 110, D18S10, doi:10.1029/2004JD005132, 2005. nd (j) Ouagadougou over West Africa. Dubovik, O., Smirnov, A., Holben, B. N., King, M. D., Kauf- man, Y. J., Eck, T. F., and Slutsker, I.: Accuracy assessment of aerosol optical properties retrieval from AERONET sun and sky radiance measurements, J. Geophys. Res., 105, 9791–9806, doi:10.1029/2000JD900040, 2000. 6 7 8 de Meij, A., Pozzer, A., and Lelieveld, J.: Global and regional trends in aerosol optical depth based on remote sensing products and pollutant emission estimates between 2000 and 2009, Atmos. Chem. Phys. Discuss., 10, 30731–30776, doi:10.5194/acpd-10- 30731-2010, 2010. Dubovik, O., Holben, B. N., Eck, T. F., Smirnov, A., Kaufman, Y. J., King, M. D., Tanr´e, D., and Slutsker, I.: Variability of absorption and optical properties of key aerosol types observed in world- wide locations, J. Atmos. Sci., 59, 3, 590-608, doi:10.1175/1520- 0469(2002)059<0590:VOAAOP>2.0.CO;2, 2002a. 9 10 Derimian, Y., Karnieli, A., Kaufman, Y. J., Andreae, M. O., An- dreae, T. W., Dubovik, O., Maenhaut, W., Koren, I., and Holben, B. N.: Dust and pollution aerosols over the Negev desert, Israel: Properties, transport, and radiative effect, J. Geophys. Res., 111, D05205, doi:10.1029/2005JD006549, 2006. Dubovik, O., Holben, B. N., Lapyonok, T., Sinyuk, A., Mishchenko, M. I., Yang, P., and Slutsker, I.: Non-spherical aerosol retrieval method employing light scattering by spheroids, Geophys. Res. Lett., 29, 1415–1418, doi:10.1029/2001GL014506, 2002b. Dubovik, O. and King, M. D.: A ﬂexible inversion algorithm for retrieval of aerosol optical properties from Sun and sky radiance measurements, J. Geophys. www.atmos-meas-tech.net/5/1271/2012/ Res., 105, 20673–20696, doi:10.1029/2000JD900282, 2000. 51 Dubovik, O., Sinyuk, A., Lapyonok, T., Holben, B. N., Mishchenko, M., Yang, P., Eck, T. F., Volten, H., Munoz, O., Veihelmann, B., van der Zander, W. J., Leon, J.-F., Sorokin, M., and Slutsker, I.: Atmos. Meas. Tech., 5, 1271–1299, 2012 www.atmos-meas-tech.net/5/1271/2012/ J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1293 on et al.: Trend analysis of aerosol optical thickness and Angstrom exponent 129 Figure 9. Unweighted (blue one on the left upper part) and weighted (red one on the right upper part) trends of Coarse- and Fine-mode dominant AOT (440 nm) (CdAOT and FdAOT) at the several AERONET stations. The total means of CdAOT and FdAOT (black one Unweighted (blue one on the left upper part) and weighted (red one on the right upper part) trends of Coarse- and Fine-mode domina 440 nm) (CdAOT and FdAOT) at the several AERONET stations. The total means of CdAOT and FdAOT (black one enclosed w heses) are shown on right vertical axis. The error bar means the 10 times of the standard error, which are used for the weighted tre is. Trend analysis of FdAOT at (b) Banizoumbou, (d) Dakar, (j) Ouagadougou, and (o) Solar Village might be insigniﬁcant because plete yearly datasets. 3 Figure 9. Unweighted (blue one on the left upper part) and weighted (red one on the right 4 upper part) trends of Coarse- and Fine-mode dominant AOT (440 nm) (CdAOT and FdAOT) 5 at the several AERONET stations. The total means of CdAOT and FdAOT (black one 6 Fig. 9. Unweighted (blue one on the left upper part) and weighted (red one on the right upper part) trends of Coarse- and Fine-mode dominant AOT (440 nm) (CdAOT and FdAOT) at the several AERONET stations. The total means of CdAOT and FdAOT (black one enclosed with parentheses) are shown on right vertical axis. The error bar means the 10 times of the standard error, which are used for the weighted trend analysis. Trend analysis of FdAOT at (b) Banizoumbou, (d) Dakar, (j) Ouagadougou, and (o) Solar Village might be insigniﬁcant because of incomplete yearly datasets. 52 Atmos. Meas. Tech., 5, 1271–1299, 2012 enclosed with parentheses) are shown on right 7 of the standard error, which are used for th 8 FdAOT at (b) Banizoumbou, (d) Dakar, (j) O 9 insignificant because of incomplete yearly data 10 11 Application of spheroid models to account for aerosol particle nonsphericity in remote sensing of desert dust, J. Geophys. Res., 111, D11208, doi:10.1029/2005JD006619, 2006. Eck, T. F., Holben, B. N., Reid, J. S., Dubovik, O., Smirnov, A., O’Neill, N. T., Slutsker, I., and Kinne, S.: Wavelength dependence of the optical depth of biomass burning, urban, and desert dust aerosol, J. Geophys. Res., 104, 31333–31349, doi:10.1029/1999JD900923, 1999. ertical axis. The error bar means the 10 times weighted trend analysis. Trend analysis of uagadougou, and (o) Solar_Village might be ets. Eck, T. F., Holben, B. N., Ward, D. E., Dubovik, O., Reid, J. S., Smirnov, A., Mukelabai, M. M., Hsu, N. C., O’Neill, N. T., and Slutsker, I.: Characterization of the optical proper- ties of biomass burning aerosols in Zambia during the 1997 ZIBBEE Field Campaign, J. Geophys. Res., 106, 3425–3448, doi:10.1029/2000JD900555, 2001. ertical axis. The error bar means the 10 times weighted trend analysis. Trend analysis of uagadougou, and (o) Solar_Village might be ets. Eck, T. F., Holben, B. N., Ward, D. E., Dubovik, O., Reid, J. S., Smirnov, A., Mukelabai, M. M., Hsu, N. C., O’Neill, N. T., and Slutsker, I.: Characterization of the optical proper- ties of biomass burning aerosols in Zambia during the 1997 ZIBBEE Field Campaign, J. Geophys. Res., 106, 3425–3448, doi:10.1029/2000JD900555, 2001. Eck, T. F. Holben, B. N., Ward, D. E., Mukelabai, M. M., Dubovik, O., Smirnov, A., Schafer, J. S., Hsu, N. C., Piketh, enclosed with parentheses) are shown on right 7 of the standard error, which are used for th 8 d (b) i b (d) k (j) Application of spheroid models to account for aerosol particle nonsphericity in remote sensing of desert dust, J. Geophys. Res., 111, D11208, doi:10.1029/2005JD006619, 2006. ertical axis. The error bar means the 10 times weighted trend analysis. Trend analysis of uagadougou, and (o) Solar_Village might be ets. Eck, T. F., Holben, B. N., Ward, D. E., Dubovik, O., Reid, J. S., Smirnov, A., Mukelabai, M. M., Hsu, N. C., O’Neill, N. T., and Slutsker, I.: Characterization of the optical proper- ties of biomass burning aerosols in Zambia during the 1997 ZIBBEE Field Campaign, J. Geophys. Res., 106, 3425–3448, doi:10.1029/2000JD900555, 2001. Eck, T. F. Holben, B. N., Ward, D. E., Mukelabai, M. M., Dubovik, O., Smirnov, A., Schafer, J. S., Hsu, N. C., Piketh, www.atmos-meas-tech.net/5/1271/2012/ www.atmos-meas-tech.net/5/1271/2012/ J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1294 9. Continued. os. Meas. Tech., 5, 1271–1299, 2012 www.atmos-meas-tech.net/5/1271/2012 Fig. 9. Continued. Fig. 9. Continued. Atmos. Meas. Tech., 5, 1271–1299, 2012 Atmos. Meas. Tech., 5, 1271–1299, 2012 www.atmos-meas-tech.net/5/1271/2012/ www.atmos-meas-tech.net/5/1271/2012/ www.atmos-meas-tech.net/5/1271/2012/ J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1295 1 2 Figure 10. Unweighted and weighted trends of Coarse- (left pentagon) and Fine-mode (right 3 circle) dominant AOT (440 nm) in percent at the major stations except (a) Avignon over 4 Western Europe (h) MD Science Center over North America and (j) Ouagadougou over 5 Fig. 10. Unweighted and weighted trends of Coarse- (left pentagon) and Fine-mode (right circle) dominant AOT (440 nm) in percent a the major stations except (a) Avignon over Western Europe, (h) MD Science Center over North America, and (j) Ouagadougou over Wes Africa. Non-applicable cases are shown as a white blank. Trend analysis of FdAOT at (b) Banizoumbou, (d) Dakar, (j) Ouagadougou, an (o) Solar Village might be insigniﬁcant because of incomplete yearly datasets. Figure 10. Unweighted and weighted trends of Coarse- (left pentagon) and Fine-mode (right 3 circle) dominant AOT (440 nm) in percent at the major stations except (a) Avignon over 4 Fig. 10. Unweighted and weighted trends of Coarse- (left pentagon) and Fine-mode (right circle) dominant AOT (440 nm) in percent at the major stations except (a) Avignon over Western Europe, (h) MD Science Center over North America, and (j) Ouagadougou over West Africa. Non-applicable cases are shown as a white blank. Trend analysis of FdAOT at (b) Banizoumbou, (d) Dakar, (j) Ouagadougou, and (o) Solar Village might be insigniﬁcant because of incomplete yearly datasets. s a white blank. Trend analysis of FdAOT at and (o) Solar Village might be insignificant erties of ﬁne/coarse mode aerosol mixtures, J. Geophys. Res., 115, D19205, doi:10.1029/2010JD014002, 2010. West Africa. Non-applicable cases are shown 6 (b) Banizoumbou, (d) Dakar, (j) Ouagadougo 7 because of incomplete yearly data sets. 8 S. J., Queface, A., Le Roux, J., Swap, R. J., and Slutsker, I.: Variability of biomass burning aerosol optical characteristics in southern Africa during the SAFARI 2000 dry season cam- paign and a comparison of single scattering albedo estimates from radiometric measurements, J. Geophys. Res., 108, 8477, doi:10.1029/2002JD002321, 2003. and (o) Solar_Village might be insignificant Evan, A. T., Mahowald, N. M., and Remer, L. s a white blank. Trend analysis of FdAOT at and (o) Solar Village might be insignificant erties of ﬁne/coarse mode aerosol mixtures, J. Geophys. Res., 115, D19205, doi:10.1029/2010JD014002, 2010. J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1296 Glen, W. G., Zelenka, M. P., and Graham, R. C.: Relating mete- orological variables and trends in motor vehicle emissions to monthly urban carbon monoxide concentrations, Atmos. Envi- ron., 30, 4225–4232, doi:10.1016/1352-2310(96)00130-6, 1996. Johnson, B. T., Heese, B., McFarlane, S. A., Chazette, P., Jones, A., and Bellouin, N.: Vertical distribution and radia- tive effects of mineral dust and biomass burning aerosol over West Africa during DABEX, J. Geophys. Res., 113, D00C12, doi:10.1029/2008JD009848, 2008b. Gobbi, G. P., Kaufman, Y. J., Koren, I., and Eck, T. F.: Classiﬁcation of aerosol properties derived from AERONET direct sun data, Atmos. Chem. Phys., 7, 453–458, doi:10.5194/acp-7-453-2007, 2007. Kahn, R., Li, W.-H., Martonchik, J., Bruegge, C., Diner, D., Gaitley, B., Abdou, W., Dubovik, O., Holben, B., Smirnov, S., Jin, Z., and Clark, D.: MISR low-light-level calibration, and implications for aerosol retrieval over dark water, J. Atmos. Sci., 62, 1032–1062, doi:10.1175/JAS3390.1, 2005. Gonz´alez, R. C., Veefkind, J. P., and de Leeuw, G.: Aerosol optical depth over Europe in August 1997 derived from ATSR-2 data, Geophys. Res. Lett., 27, 955–958, doi:10.1029/1999GL010962, 2000. Kahn, R., Garay, M., Nelson, D., Yau, K., Bull, M., and Martonchik, J.: Satellite-derived aerosol optical depth over dark water from MISR and MODIS: Comparisons with AERONET and implica- tions for climatological studies, J. Geophys. Res., 112, D18205, doi:10.1029/2006JD008175, 2007. Hao, W. M. and Liu, M.: Spatial and temporal distribution of tropical biomass burning, Global Biogeochem. Cy., 8, 495–503, doi:10.1029/94GB02086, 1994. Haywood, J. and Boucher, O.: Estimates of the direct and indirect radiative forcing due to tropospheric aerosols: A review, Rev. Geophys., 38, 513–543, doi:10.1029/1999RG000078, 2000. Kambezidisa, H. D. and Kaskaoutisa, D. G.: Aerosol climatology over four AERONET sites: An overview, Atmos. Environ., 42, 1892–1906, doi:10.1016/j.atmosenv.2007.11.013, 2008. Higurashi, A. and Nakajima, T.: Development of a two channel aerosol retrieval algorithm on global scale using NOAA AVHRR, J. Atmos. Sci., 56, 924–941, doi:10.1175/1520- 0469(1999)056<0924:DOATCA>2.0.CO;2, 1999. Karnieli, A., Derimian, Y., Indoitu, R., Panov, N., Levy, R. C., Remer, L. A., Maenhaut, W., and Holben, B. N.: Temporal trend in anthropogenic sulfur aerosol transport from central and eastern Europe to Israel, J. Geophys. Res., 114, D00D19, doi:10.1029/2009JD011870, 2009. Holben, B. N., Eck, T. F., Slutsker, I., Tanr´e, D., Buis, J. P., Set- zer, A., Vermote, E., Reagan, J. A., Kaufman, Y. J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent J., Nakajima, T., Lavenu, F., Jankowiak, I., and Smirnov, A.: AERONET – A federated instrument network and data archive for aerosol char- acteristics, Remote Sens. Environ., 66, 1-16, doi:10.1016/S0034- 4257(98)00031-5, 1998. Kaskaoutis, D. G. and Kambezidis, H. D.: Comparison of the ˚Angstr¨om parameters retrieval in different spectral ranges with the use of different techniques, Meteorol. Atmos. Phys., 99, 233– 246, doi:10.1007/s00703-007-0279-y, 2008. Kaskaoutis, D. G., Kambezidis, H. D., Hatzianastassiou, N., Kos- mopoulos, P. G., and Badarinath, K. V. S.: Aerosol climatology: dependence of the Angstrom exponent on wavelength over four AERONET sites, Atmos. Chem. Phys. Discuss., 7, 7347–7397, doi:10.5194/acpd-7-7347-2007, 2007. Holben, B. N., Tanr´e, D., Smirnov, A., Eck, T. F., Slutsker, I., Abuhassan, N., Newcomb, W. W., Schafer, J. S., Chatenet, B., Lavenu, F., Kaufman, Y. J., Vande Castle, J., Setzer, A., Markham, B., Clark, D., Frouin, R., Halthore, R., Karneli, A., O’Neill, N. T., Pietras, C., Pinker, R. T., Voss, K., and Zibordi, G.: An emerging ground-based aerosol climatology: Aerosol op- tical depth from AERONET, J. Geophys. Res., 106, 12067– 12097, doi:10.1029/2001JD900014, 2001. Kaskaoutis, D. G., Kumar Kharol, S., Sinha, P. R., Singh, R. P., Kambezidis, H. D., Rani Sharma, A., and Badarinath, K. V. S.: Extremely large anthropogenic-aerosol contribution to total aerosol load over the Bay of Bengal during winter season, At- mos. Chem. Phys., 11, 7097–7117, doi:10.5194/acp-11-7097- 2011, 2011, 2011a. Huang, J., Zhang, C., and Prospero, J. M.: African aerosol and large-scale precipitation variability over West Africa, Environ. Res. Lett., 4, 015006, doi:10.1088/1748-9326/4/1/015006, 2009. Kaskaoutis, D. G., Kharol, S. K., Sinha, P. R., Singh, R. P., Badarinath, K. V. S., Mehdi, W., and Sharma, M.: Contrasting aerosol trends over South Asia during the last decade based on MODIS observations, Atmos. Meas. Tech. Discuss., 4, 5275– 5323, doi:10.5194/amtd-4-5275-2011, 2011b. Ignatov, A. and Stowe, L.: Sensitivity and information content of aerosol retrievals from AVHRR: Radiometric factors, Appl. Op- tics, 41, 991–1011, doi:10.1364/AO.41.000991, 2002. IPCC: Changes in Atmospheric Constituents and in Radiative Forc- ing: Issues Related to Aerosols, Cambridge Univ. Press, Cam- bridge, UK, 153–179, 2007. Kaufman, Y. J.: Aerosol optical thickness and atmospheric path ra- diance, J. Geophys. Res., 98, 2677–2692, 1993. Jeong, M.-J. and Li, Z.: Quality, compatibility, and synergy analy- ses of global aerosol products of global aerosol products derived from the advanced very high resolution radiometer and Total Ozone Mapping Spectrometer, J. Geophys. Res., 110, D10S08, doi:10.1029/2004JD004647, 2005. www.atmos-meas-tech.net/5/1271/2012/ A.: Global Aerosols, B. Am. Meteorol. Soc., 88, s1–s135, 2007. Foster, A. and Kumar, N.: Health effects of air quality reg- ulations in Delhi, India, Atmos. Environ., 45, 1675–1683, doi:10.1016/j.atmosenv.2011.01.005, 2011. Eck, T. F., Holben, B. N., Dubovik, O., Smirnov, A., Goloub, P., Chen, H. B., Chatenet, B., Gomes, L., Zhang, X.-Y., Tsay, S.- C., Ji, Q., Giles, D., and Slutsker, I.: Columnar aerosol opti- cal properties at AERONET sites in central eastern Asia and aerosol transport to the tropical mid-Paciﬁc, J. Geophys. Res., 110, D06202, doi:10.1029/2004JD005274, 2005. Garstang, M. and Fitzjarrald, D. R.: Observations of Surface to At- mosphere Interactions in the Tropics, Oxford University Press, New York, USA, 1999. 54 Gerasopoulos, E., Andreae, M. O., Zerefos, C. S., Andreae, T. W., Balis, D., Formenti, P., Merlet, P., Amiridis, V., and Pa- pastefanou, C.: Climatological aspects of aerosol optical prop- erties in Northern Greece, Atmos. Chem. Phys., 3, 2025–2041, doi:10.5194/acp-3-2025-2003, 2003. Eck, T. F., Holben, B. N., Sinyuk, A., Pinker, R. T., Goloub, P., Chen, H., Chatenet, B., Li, Z., Singh, R. P., Tripathi, S. N., Reid, J. S., Giles, D. M., Dubovik, O., O’Neill, N. T., Smirnov, A., Wang, P., and Xia, X.: Climatological aspects of the optical prop- Atmos. Meas. Tech., 5, 1271–1299, 2012 www.atmos-meas-tech.net/5/1271/2012/ J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent V.: Satellite remote sensing reveals regional tropospheric aerosol trends, Opt. Express, 15, 7423–7438, 2007. Pincus, R. and Baker, M. B.: Effect of precipitation on the albedo susceptibility of clouds in the marine boundary layer, Nature, 372, 250–252, doi:10.1038/372250a0, 1994. Mishchenko, M. I., Geogdzhayev, I. V., Cairns, B., Rossow, W. B., and Lacis, A. A.: Aerosol retrievals over the ocean by use of channels 1 and 2 AVHRR data: sensitivity analysis and prelimi- nary results, Appl. Optics, 38, 7325–7341, 1999a. Prospero, J. M. and Lamb, J. P.: African droughts and dust transport to the Caribbean: climate change and implications, Science, 302, 1024–1027, doi:10.1126/science.1089915, 2003. Mishchenko, M. I., Dlugach, J. M., Yanovitskij, E. G., and Za- kharova, N. T.: Bidirectional reﬂectance of ﬂat, optically thick particulate laters: an efﬁcient radiative transfer solution and ap- plications to snow and soil surfaces, J. Quant. Spectrosc. Ra., 63, 409–432, doi:10.1016/S0022-4073(99)00028-X, 1999b. Reeves, C. E., Formenti, P., Aﬁf, C., Ancellet, G., Atti´e, J.-L., Bechara, J., Borbon, A., Cairo, F., Coe, H., Crumeyrolle, S., Fierli, F., Flamant, C., Gomes, L., Hamburger, T., Jambert, C., Law, K. S., Mari, C., Jones, R. L., Matsuki, A., Mead, M. I., Methven, J., Mills, G. P., Minikin, A., Murphy, J. G., Nielsen, J. K., Oram, D. E., Parker, D. J., Richter, A., Schlager, H., Schwarzenboeck, A., and Thouret, V.: Chemical and aerosol characterisation of the troposphere over West Africa during the monsoon period as part of AMMA, Atmos. Chem. Phys., 10, 7575–7601, doi:10.5194/acp-10-7575-2010, 2010. Mishchenko, M. I., Travis, L. D., and Lacis, A. A.: Scattering, Ab- sorption, and Emission of Light by Small Particles, Cambridge University Press, 2002. Mishchenko, M., Geogdzhayev, I. V., Rossow, W. B., Cairns, B., Carlson, B. E., Lacis, A. A., Liu, L., and Travis, L. D.: Long- term satellite record reveals likely recent aerosol trend, Science, 315, 1543, doi:10.1126/science.1136709, 2007. Reid, J. S., Eck, T. F., Christopher, S. A., Hobbs, P. V., and Holben, B. N.: Use of the ˚Angstrom exponent to estimate the variability of optical and physical properties of aging smoke particles in Brazil, J. Geophys. Res., 104, 27473–27489, doi:10.1029/1999JD900833, 1999. Moulin, C. and Chiapello, I.: Evidence of the control of summer atmospheric transport of African dust over the Atlantic by Sahel sources from TOMS satellites (1979–2000), Geophys. Res. Lett., 31, L02107, doi:10.1029/2003GL018931, 2004. Remer, L. A., Gasso, S., Hegg, D. A., Kaufman, Y. J., and Holben, B. J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1297 Li, Z., Zhao, X., Kahn, R., Mishchenko, M., Remer, L., Lee, K.-H., Wang, M., Laszlo, I., Nakajima, T., and Maring, H.: Uncertain- ties in satellite remote sensing of aerosols and impact on monitor- ing its long-term trend: a review and perspective, Ann. Geophys., 27, 2755–2770, doi:10.5194/angeo-27-2755-2009, 2009. O’Neill, N. T., Eck, T. F., Holben, B. N., Smirnov, A., and Dubovick, O.: Bimodal size distribution inﬂuences on the varia- tion of Angstrom derivatives in spectral and optical depth space, J. Geophys. Res., 106, 9787–9806, doi:10.1029/2000JD900245, 2001b. O’Neill, N. T., Eck, T. F., Smirnov, A., Holben, B. N., and Thulasir- aman, S.: Spectral discrimination of coarse and ﬁne mode optical depth, J. Geophys. Res., 108, 4559, doi:10.1029/2002JD002975, 2003. Liu, Y., Sarnat, J. A., Coull, B. A., Koutrakis, P., and Ja- cob, D. J.: Validation of Multiangle Imaging Spectroradiome- ter (MISR) aerosol optical thickness measurements using Aerosol Robotic Network (AERONET) observations over the contiguous United States, J. Geophys. Res., 109, D06205, doi:10.1029/2003JD003981, 2004. O’Neill, N. T., Thulasiraman, S., Eck, T. F., and Reid, J. S.: Robust optical features of ﬁne mode size distributions: Application to the Quebec smoke event of 2002, J. Geophys. Res., 110, D11207, doi:10.1029/2004JD005157, 2005. Marmer, E., Langmann, B., Fagerli, H., and Vestreng, V.: Di- rect shortwave radiative forcing of sulphate aerosol over Eu- rope from 1900 to 2000, J. Geophys. Res., 112, D23S17, doi:10.1029/2006JD008037, 2007. Ohmura, A.: Observed long-term variations of solar irradi- ance at the Earth’s surface, Space Sci. Rev., 125, 111–128, doi:10.1007/s11214-006-9050-9, 2006. Massie, T. S., Torres, O., and Smith, S. J.: Total ozone mapping spectrometer (TOMS) observations of increases in Asian aerosol in winter from 1979 to 2000, J. Geophys. Res., 109, D18211, doi:10.1029/2004JD004620, 2004. Pereira, S. N., Wagner, F., and Silva, A. M.: Seven years of mea- surements of aerosol scattering properties, near the surface, in the southwestern Iberia Peninsula, Atmos. Chem. Phys., 11, 17– 29, doi:10.5194/acp-11-17-2011, 2011. Mazzola, M., Lanconelli, C., Lupi, A., Busetto, M., Vitale, V., and Tomasi, C.: Columnar aerosol optical properties in the Po Val- ley, Italy, from MFRSR data, J. Geophys. Res., 115, D17206, doi:10.1029/2009JD013310, 2010. Perry, K. D., Cahill, T. A., Schnell, R. C., and Harris, J. M.: Long-range transport of anthropogenic aerosols to the National Oceanic and Atmospheric Administration baseline station at Mauna Loa Observatory, Hawaii, J. Geophys. Res., 104, 18521– 18533, doi:10.1029/1998JD100083, 1999. Mishchenko, M. and Geogdzhayev, I. J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent Kim, S.-W., Yoon, S.-C., Kim, J., and Kim, S.-Y.: Seasonal and monthly variations of columnar aerosol optical properties over East Asia determined from multi-year MODIS, LIDAR, and AERONET sun/sky radiometer measurements, Atmos. Environ., 41 1634–1651 doi:10 1016/j atmosenv 2006 10 044 2007 Kim, S.-W., Yoon, S.-C., Kim, J., and Kim, S.-Y.: Seasonal and monthly variations of columnar aerosol optical properties over East Asia determined from multi-year MODIS, LIDAR, and AERONET sun/sky radiometer measurements, Atmos. Environ., 41, 1634–1651, doi:10.1016/j.atmosenv.2006.10.044, 2007. Jeong, M.-J., Li, Z., Chu, D. A., and Tsay, S.-T.: Quality and compatibility analyses of global aerosol products derived from the advanced very high resolution radiometers and the moder- ate imaging spectroradiometer, J. Geophys. Res., 110, D10S09, doi:10.1029/2004JD004648, 2005. Kokhanovsky, A. and de Leeuw, G.: Satellite aerosol remote sensing over land, Springer, 2009. Kotchenruther, R., Hobbs, P. V., and Hegg, D. A.: Humidiﬁ- cation factors for atmospheric aerosols off the mid-Atlantic coast of the United States, J. Geophys. Res., 104, 2239–2251, doi:10.1029/98JD01751, 1999. Johnson, B. T., Osborne, S. R., Haywood, J. M., and Harrison, M. A. J.: Aircraft measurements of biomass burning aerosol over West Africa during DABEX, J. Geophys. Res., 113, D00C06, doi:10.1029/2007JD009451, 2008a. Kubilay, N., Cokacar, T., and Oguz, T.: Optical properties of mineral dust outbreaks over the northeastern Mediterranean, J. Geophys. Res., 108, 4666, doi:10.1029/2003JD003798, 2003. Atmos. Meas. Tech., 5, 1271–1299, 2012 www.atmos-meas-tech.net/5/1271/2012/ J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent Lett., 28, 4555– 4558, doi:10.1029/2001GL013036, 2001. Shinozuka, Y., Redemann, J., Livingston, J. M., Russell, P. B., Clarke, A. D., Howell, S. G., Freitag, S., O’Neill, N. T., Reid, E. A., Johnson, R., Ramachandran, S., McNaughton, C. S., Ka- pustin, V. N., Brekhovskikh, V., Holben, B. N., and McArthur, L. J. B.: Airborne observation of aerosol optical depth during ARC- TAS: vertical proﬁles, inter-comparison and ﬁne-mode fraction, Atmos. Chem. Phys., 11, 3673–3688, doi:10.5194/acp-11-3673- 2011, 2011. Thomas, G. E., Poulsen, C. A., Siddans, R., Sayer, A. M., Car- boni, E., Marsh, S. H., Dean, S. M., Grainger, R. G., and Lawrence, B. N.: Validation of the GRAPE single view aerosol retrieval for ATSR-2 and insights into the long term global AOD trend over the ocean, Atmos. Chem. Phys., 10, 4849–4866, doi:10.5194/acp-10-4849-2010, 2010. Sinyuk, A., Dubovik, O., Holben, B. N., Eck, T. F., Breon, F.- M., Martonchik, J., Kahn, R., Diner, D. J., Vermote, E. F., Roger, J.-C., Lapyonok, T., and Slutsker, I.: Simultaneous re- trieval of aerosol and surface properties from a combination of AERONET and satellite data, Remote Sens. Environ., 107, 90– 108, doi:10.1016/j.rse.2006.07.022, 2007. Twomey, S. A.: The inﬂuence of pollution on the shortwave albedo of clouds, J. Atmos. Sci., 34, 1149–1152, doi:10.1175/1520- 0469(1977)034<1149:TIOPOT>2.0.CO;2, 1977. Tyson, P. D.: Climatic Change and Variability in Southern Africa, Oxford Univ. Press, 220–221,, 1986. Smirnov, A., Holben, B. N., Eck, T. F., Dubovik, O., and Slutsker, I.: Cloud screening and quality control algorithms for the AERONET data base, Remote Sens. Environ., 73, 3, 337-349, doi:10.1016/S0034-4257(00)00109-7, 2000. Venzac, H., Sellegri, K., Villani, P., Picard, D., and Laj, P.: Seasonal variation of aerosol size distributions in the free troposphere and residual layer at the puy de Dˆome station, France, Atmos. Chem. Phys., 9, 1465–1478, doi:10.5194/acp-9-1465-2009, 2009. Smirnov, A., Holben, B. N., Dubovik, O., Neil, N., and Eck, T. F.: Atmospheric aerosol optical properties in the Persian Gulf, Atmos. Sci., 59, 620–634, doi:10.1175/1520- 0469(2002)059<0620:AAOPIT>2.0.CO;2, 2002. von Hoyningen-Huene, W., Yoon, J., Vountas, M., Istomina, L. G., Rohen, G., Dinter, T., Kokhanovsky, A. A., and Burrows, J. P.: Retrieval of spectral aerosol optical thickness over land using ocean color sensors MERIS and SeaWiFS, Atmos. Meas. Tech., 4, 151–171, doi:10.5194/amt-4-151-2011, 2011. Smith, S. J., Pitcher, H., and Wigley, T. M. L.: Global and regional anthropogenic sulfur dioxide emissions, Global Planet. Change, 29, 99–119, doi:10.1016/S0921-8181(00)00057-6, 2001. Washington, R. and Todd, M. J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent N.: Urban/industrial aerosol: Ground-based sun/sky radiome- ter and airborne in situ measurements, J. Geophys. Res., 102, 16849–16859, doi:10.1029/96JD01932, 1997. Mukai, S., Sano, I., and Holben, B.N.: Aerosol properties over Japan by sun/sky photometry, Water Air Soil Pollut., 5, 133–143, doi:10.1007/s11267-005-0731-2, 2005. Mukai, S., Sano, I., Satoh, M., and Holben, B. N.: Aerosol proper- ties and air pollutants over an urban area, Atmos. Res., 82, 643– 651, doi:10.1016/j.atmosres.2006.02.020, 2006. Sabbah, I. and Hasan, F. M.: Remote sensing of aerosols over the Solar Village, Saudi Arabia, Atmos. Res., 90, 170–179, doi:10.1016/j.atmosres.2008.02.004, 2008. Norris, J. R. and Wild, M.: Trends in aerosol radiative effects over Europe inferred from observed cloud cover, solar “dim- ming”, and solar “brightening”, J. Geophys. Res., 112, D08214, doi:10.1029/2006JD007794, 2007. Sabbah, I., Saeed, T., Al Jassar, H. K., and Rao, K. S.: Remote sens- ing of desert dust in Kuwait, J. Sci. Eng., 33, 101–117, 2006. Sano, I., Mukai, S., Okada, Y., Holben, B. N., Ohta, S., and Takamura, T.: Optical properties of aerosols during APEX and ACE-Asia experiments, J. Geophys. Res., 108, 8649, doi:10.1029/2002JD003263, 2003. O’Neill, N. T., Dubovik, O., and Eck, T. F.: A modiﬁed Angstrom coefﬁcient for the characterization of sub-micron aerosols, Appl. Optics, 40, 2368–2374, 2001a. Atmos. Meas. Tech., 5, 1271–1299, 2012 www.atmos-meas-tech.net/5/1271/2012/ J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1298 Swap, R., Garstang, M., Macko, S. A., Tyson, P. D., Maenhaut, W., Artaxo, P., K˚allberg, P., and Talbot, R.: The ong-range transport of southern African aerosols to the tropical South Atlantic, J. Geophys. Res., 101, 23777–23791, 1996. Schmeissner, T., Krejci, R., Str¨om, J., Birmili, W., Wieden- sohler, A., Hochschild, G., Gross, J., Hoffmann, P., and Calderon, S.: Analysis of number size distributions of tropi- cal free tropospheric aerosol particles observed at Pico Espejo (4765 m a.s.l.), Venezuela, Atmos. Chem. Phys., 11, 3319–3332, doi:10.5194/acp-11-3319-2011, 2011. Tafuro, A. M., Barnaba, F., De Tomasi, F., Perrone, M. R., and Gobbi, G. P.: Saharan dust particle properties over the central Mediterranean, Atmos. Res., 81, 67–93, doi:10.1016/j.atmosres.2005.11.008, 2006. Schuster, G. L., Dubovick, O., and Holben, B. N.: Angstrom ex- ponent and bimodal aerosol size distributions, J. Geophys. Res., 111, D07207, doi:10.1029/2005/JD006328, 2006. Tanr´e, D., Br´eon, F. M., Deuz´e, J. L., Herman, M., Goloub, P., Nadal, F., and Marchand, A.: Global observation of anthro- pogenic aerosols from satellite, Geophys. Res. J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent C.: Atmospheric controls on min- eral dust emission from the Bod´el´e depression, chad: the role of the low level jet, Geophys. Res. Lett., 32, L17701, doi:10.1029/2005GL023597, 2005. Smith, S. J., Andres, R., Conception, E., and Lurz, J.: Historical sul- fur dioxide emissions 1850–2000: Methods and results, PNNL Res. Rep. 14537, Pac. Northwest Natl. Lab., Richland, Wash, 2003. Wild, M., Gilgen, H., Roesch, A., Ohmura, A., Long, C. N., Dut- ton, E. G., Forgan, B., Kallis, A., Russak, V., and Tsvetkov, A.: From dimming to brightening: Decadal changes in solar radiation at Earth’s surface, Science, 308, doi:10.1126/science.1103215, 847–850, 2005. Stanhill, G.: A perspective on global warming, dimming, and brightening, EOS Trans. Am. Geophys. Union, 88, 58, doi:10.1029/2007EO050007, 2007. Wild, M., Ohmura, A., and Makowski, K.: Impact of global dim- ming and brightening on global warming, Geophys. Res. Lett., 34, L04702, doi:10.1029/2006GL028031, 2007. Streets, D. G., Tsai, N. Y., Akimoto, H., and Oka, K.: Sulfur dioxide emissions in Asia in the period 1985–1997, Atmos. Environ., 34, 4413–4424, doi:10.1016/S1352-2310(00)00187-4, 2000. Yoon, J., von Hoyningen-Huene, W., Vountas, M., and Burrows, J. P.: Analysis of linear long-term trend of aerosol optical thick- ness derived from SeaWiFS using BAER over Europe and South China, Atmos. Chem. Phys., 11, 12149–12167, doi:10.5194/acp- 11-12149-2011, 2011. Streets, D. G., Bond, T. C., Carmichael, G. R., Fernandes, S. D., Fu, Q., He, D., Klimont, Z., Nelson, S. M., Tsai, N. Y., Wang, M. Q., Woo, J.-H., and Yarber, K. F.: An inventory of gaseous and primary aerosol emissions in Asia in the year 2000, J. Geophys. Res., 108, 8809, doi:10.1029/2002JD003093, 2003. Streets, D. G., Wu, Y., and Chin, M.: Two-decadal aerosol trends as a likely explanation of the global dim- ming/brightening transition, Geophys. Res. Lett., 33, L15806, doi:10.1029/2006GL026471, 2006. Yu, H., Chin, M., Remer, L. A., Kleidman, R. G., Bellouin, N., Bian, H., and Diehl, T.: Variability of marine aerosol ﬁne- mode fraction and estimates of anthropogenic aerosol component over cloud-free oceans from the Moderate Resolution Imaging Spectroradiometer (MODIS), J. Geophys. Res., 114, D10206, doi:10.1029/2008JD010648, 2009. Streets, D. G., Yan, F., Chin, M., Diehl, T., Mahowald, N., Schultz, M., Wild, M., Wu, Y., and Yu, C.: Anthropogenic and natural contributions to regional trends in aerosol op- tical depth, 1980–2006, J. Geophys. Res., 114, D00D18, doi:10.1029/2008JD011624, 2009. Zhang, J. and Reid, J. J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent S.: A decadal regional and global trend analysis of the aerosol optical depth using a data-assimilation grade over-water MODIS and Level 2 MISR aerosol products, Atmos. Chem. Phys., 10, 10949–10963, doi:10.5194/acp-10- Atmos. Meas. Tech., 5, 1271–1299, 2012 www.atmos-meas-tech.net/5/1271/2012/ J. Yoon et al.: Trend analysis of aerosol optical thickness and ˚Angstr¨om exponent 1299 Zhao, T. X.-P., Laszlo, I., Guo, W., Heidinger, A., Cao, C., Je- lenak, A., Tarpley, D., and Sullivan, J.: Study of long-term trend in aerosol optical thickness observed from operational AVHRR satellite instrument, J. Geophys. Res., 113, D07201, doi:10.1029/2007JD009061, 2008. Atmos. Meas. Tech., 5, 1271–1299, 2012 10949-2010, 2010. Zhang, X. Y., Gong, S. L., Zhao, T. L., Arimoto, R., Wang, T. Q., and Zhou, Z. J.: Sources of Asian dust and role of climate change versus desertiﬁcation in Asian dust emission, Geophys. Res. Lett., 30, 2272, doi:10.1029/2003GL018206, 2003. www.atmos-meas-tech.net/5/1271/2012/ Atmos. Meas. Tech., 5, 1271–1299, 2012 Zhang, X. Y., Gong, S. L., Zhao, T. L., Arimoto, R., Wang, T. Q., and Zhou, Z. J.: Sources of Asian dust and role of climate change versus desertiﬁcation in Asian dust emission, Geophys. Res. Lett., 30, 2272, doi:10.1029/2003GL018206, 2003. www.atmos-meas-tech.net/5/1271/2012/ Atmos. Meas. Tech., 5, 1271–1299, 2012 www.atmos-meas-tech.net/5/1271/2012/
https://openalex.org/W4241951625	https://hrcak.srce.hr/file/215936	English	null	Evaluation of Practical Application of Plant Simulation in the Form of Experiments	Naše more	2,015	cc-by	4,174	Evaluation of Practical Application of Plant Simulation in the Form of Experiments Procjena praktične primjene simulacije postrojenja u obliku eksperimenata Lukáš Kamenický Technical University of Kosice Slovakia e-mail: lukas.kamenicky@tuke.sk Michal Hovanec Technical University of Kosice Slovakia e-mail: michal.hovanec@tuke.sk Juraj Sinay Technical University of Kosice Slovakia e-mail: juraj.sinay@tuke.sk Hrozek František Technical University of Kosice Slovakia e-mail: frantisek.hrozek@tuke.sk Hrozek František Technical University of Kosice Slovakia e-mail: frantisek.hrozek@tuke.sk Petr Skřehot Technical University of Kosice Slovakia e-mail: p.skrehot@seznam.cz DOI 10.17818/NM/2015/SI19 UDK 338.36 Review / Pregledni rad Paper accepted / Rukopis primljen: 27. 4. 2015. DOI 10.17818/NM/2015/SI19 UDK 338.36 Review / Pregledni rad Paper accepted / Rukopis primljen: 27. 4. 2015. Summary KEY WORDS product lifecycle managment plant Simulation quality production digital plant digitalization This contribution deals with the Simulation of manufacturing systems. Plant simulation is a modern instrument for optimizing factory production; however, the initial investment is high, especially for human resources, since the creation of a simulation model requires a lot of data from manufacturing, which needs to be analyzed and used as input into the simulation so that the simulation model is as close as possible to the real manufacturing system. Simulation enables changes to first be carried out in the simulation model and the defining of possible problems, through which in this case it is possible to avoid when introducing changes into an operation. The measure of effectiveness of using inputs is the productivity of a manufacturing system, and the aim of organization management is to monitor, analyze and improve a manufacturing system through optimizing for the purpose of increasing its efficiency, ensuring higher quality and a higher volume of production, shortening supply time in the scope of transportation, improving logistics and increasing the profitability of the given organization. KLJUČNE RIJEČI vijek trajanja proizvoda menadžment simulacija postrojenja kvalitetna proizvodnja digitalno postrojenje digitalizacija Mjera učinkovitosti upotrebe ulaznih podataka je produktivnost proizvodnog sustava, a cilj organizacijskog menadžmenta je pratiti, analizirati i poboljšati proizvodni sustav optimizacijom kako bi se povećala učinkovitost, osigurala viša kvaliteta i veća proizvodnja, smanjilo vrijeme nabave tijekom transporta, poboljšala logistika i povećala profitabilnost date organizacije. vijek trajanja proizvoda menadžment simulacija postrojenja kvalitetna proizvodnja digitalno postrojenje digitalizacija Sažetak U ovom radu govori se o simulaciji proizvodnih sustava. Simulacija postrojenja je moderni instrument za optimizaciju tvorničke proizvodnje; međutm, početno ulaganje je veliko, posebice u ljudstvu, budući da pravljenje simulacijskog modela zahtjeva veliku količinu podataka iz proizvodnje, koje treba analizirati i upotrebiti u simulaciji tako da simulacijski model bude što vjeriniji stvarnom proizvodnom sustavu. Simulacija omogućuje da se promjene prije svega provedu u simulacijskom modelu i definiraju mogući problemi, a time je moguće izbjeći iste probleme kada se promjene uvedu u rad. Experiment 1 – shortening the period for the setting of the lacquer equipment; therefore, the work environment in which the process of manufacture took place was selected as the area of optimization in the organization. Optimization related in particular to the shortening of certain times for processing the sheets, transport and logical links, in which time reserves were identified. These were suitable for using and thus increasing the productivity of the entire manufacturing system. The process of setting in the model is represented by the object Buffer, with a processing time of 24 hours. The proposed solution in practice represents placing a drying oven into the manufacturing system between the processes “pressing” and “striking of CC”. The period for setting of the lacquer would in this way be shortened from the original 24 hours to an anticipated 4 hours (2 hours for setting of the lacquer, 2 hours for ventilating the ovens). By creating the experimental model, in the scope of which the period of processing in the object Buffer (see Fig.1) is defined as 4 hours, the influence of these changes will be verified in the total production of the manufacturing system and the impact on the capacity of loading of the object Buffer. The tables below shows the words “KU” and “sluzba” in software as a service and a CC. INTRODUCTIONf into production and verifying the correctness of technological proposals before actually introducing them into a real manufacturing system. Predicting the effects of different changes in the life cycle of a product can be achieved through use of a digital factory as software support, namely by simulation programs such as, for example, plant simulation, which enable predicting the course of the process of planning, transportation, logistics and manufacture in a short time. The use of this tool enables the performing of experiments without direct interference In the scope of conducting the experiments the process of lacquering and pressing of metal sheets and the process of manufacturing crown caps were described [1]. The processing of metal sheets itself runs on automated production 193 “Naše more” 62(3)/2015., pp. 193-199 Experiment 1 – shortening the period for the setting of the lacquer Experiment 2 – shortening the duration of the change of lacquer and colour the change of lacquer/colour), the number of processed entities for the output Service grew by 14.55%, Each change of lacquer and colour lasts 30 minutes; however, it is essential in the scope of carrying out individual orders, because the type of lacquer and the number of colours are specific for each order (see Fig.4). The proposed improvement consists in optimizing the process of changing the lacquer and colours such that the duration of this process is minimized. In the experimental model a proposed shortening of the time needed for a change of the lacquer of object LTG2 and the colour in object LTG1 was introduced from the original 30 minutes to an assumed 25 minutes. The effect of this change was monitored by comparing the statistical output from the experimental model versus the output from the basic model. -- the number of entities processed by lines LTG2 and LTG1 for the purpose of manufacture of CC grew; however, due to the loading of the object Buffer, this was manifested only in the collective in-process production. TOOLS AND METHOD In the following sections of the contribution 5 experiments are carried out in which the independent and common effects of the proposed improvements in the manufacturing process organization using the simulation instrument PLANT SIMULATION will be tested. These are the proposals for improvement: From the above-presented image in comparison with Fig. 2 and Fig. 3 the following effect of the simulated change follows: 1. Shortening of the time necessary for the setting of the lacquer by introducing a drying oven into the process of production between the process of lacquering and the process of manufacture of the CC. -- the loading of object Buffer dropped by 10%, -- the loading of object Buffer dropped by 10%, -- the period of processing of objects SACMI1 and SACMI2 grew by 2.8% for each piece of equipment, which had as a consequence growth in the volume of processed entities for the output of CC by 3.015%. -- the period of processing of objects SACMI1 and SACMI2 grew by 2.8% for each piece of equipment, which had as a consequence growth in the volume of processed entities for the output of CC by 3.015%. 2. Shortening of the time necessary for a change of lacquer and colour on the LTG lines. 3. Introducing of a third series SACMI line for the manufacture of CC, (SACMI - ejectors line, LTG - printing line, CC - Crown Caps . M. Hovanec et al: Evaluation of Practical Application of Plant... Figure 1 Statistical output from the model of experiment 1 Figure 2 Use of the operation time of equipment before the change Figure 1 Statistical output from the model of experiment 1 Figure 2 Use of the operation time of equipment before the change Figure 2 Use of the operation time of equipment before the change Figure 2 Use of the operation time of equipment before the change Figure 2 Use of the operation time of equipment before the change M. Hovanec et al: Evaluation of Practical Application of Plant... Figure 3 Statistical output of the basic simulation model before the change Figure 4 Statistical output of the basic simulation model Experiment 2 – shortening the duration of the change of lacquer and colour Experiment 3 – increasing the production capacity in the manufacture of CC At present the process of CC manufacturing is represented by 2 series of SACMI production lines. The loading capacity of the object Buffer (see Fig. 2) was evaluated as a space for introducing a third series of the SACMI production line. The proposed change was simulated in experiment model 3 by adding object SACMI3. On the basis of the statistical output from this model the effect of the simulated change in the loading of the object Buffer was monitored, as were the amount of processed entities and the loading capacity of objects SACMI1, SACMI2 and SACMI3. From the above-mentioned image, in comparison with Fig.3., the following effect of the simulated change results: -- the average period of processing 1 entity as a service dropped by 5 minutes (represents the time-savings during Figure 5 Statistical output from the model of experiment 3 Figure 6 Loading of the equipment in the model of experiment 3 Figure 6 Loading of the equipment in the model of experiment 3 195 “Naše more” 62(3)/2015., pp. 193-199 -- the share of processed entities in the output Service dropped by 21.74%, From the above-mentioned images, in comparison with Fig. 2 and Fig. 3., the following effect of the simulated changes results: -- the share of processed entities in the output CC grew by 54.27%, -- the volume of the processed entities for the output of CC grew by 6.78%, -- the relative loading of the object Buffer dropped significantly, -- the period of processing SACMI objects dropped to 66.3%, i.e. a reserve for further processing was created, -- the period of processing of the SACMI objects grew by 2.8%. -- the period of processing of the SACMI objects grew by 2.8%. -- the loading capacity of the object Buffer remained unchanged. Experiment 4 – simulation of the combination of Experiment 5 represents the simulation of common introducing of the proposed changes simulated in experiments 2 and 3. The change simulated in the experiment was excluded on the basis of the fact that the achieved benefit (higher volume of processed entities by 3% in the output of CC) is not considered as sufficient with respect to the financial demand of carrying out this change in practice (investment in the drying oven). experiments 1, 2 and 3 In the individual experiments 1, 2 and 3 the effect of the simulated changed was assessed individually in comparison with the basic simulation model of the real manufacturing system. In experiment 4 their common effect on the overall productivity of the system was simulated; thus all proposed changes were implemented in the scope of a single model. On the basis of statistical outputs the effect of the simulated changes on the loading capacity of the SACMI objects, the object Buffer and the amount of processed entities for outputs of CC and Service was monitored. In this experiment the introducing of a third series of the SACMI line and shortening of the time for changing lacquer and colour on the LTG lines and their common effect on productivity in the simulated manufacturing system was simulated. On the basis of statistical outputs the volume of the processed entities in the outputs Service and CC were monitored, as was the loading of the individual objects. From a comparison of the above-mentioned images with Fig. 2 and Fig. 3., the following effect of the simulated changes in experiment 4 results: -- production of objects LTG2 and LTG1 grew by 8.63%, Figure 7 Statistical output from the model of experiment 4 Figure 8 Loading of the equipment in the model of experiment 4 Figure 7 Statistical output from the model of experiment 4 gure 7 Statistical output from the model of experiment 4 Figure 8 Loading of the equipment in the model of experiment 4 Figure 8 Loading of the equipment in the model of experiment 4 M. Hovanec et al: Evaluation of Practical Application of Plant... Experiment 4 – simulation of the combination of Figure 9 Statistical output from the model of experiment 5 Figure 10 Loading of the equipment in the model of experiment 5 Figure 9 Statistical output from the model of experiment 5 Figure 9 Statistical output from the model of experiment 5 Figure 10 Loading of the equipment in the model of experiment 5 0 Loading of the equipment in the model of experiment On the basis of a comparison of the facts presented in Fig. 10 and Fig. 9 and with the statistical outputs from the basic model (Fig. 2 and Fig. 3) the following effect of the simulated changes result: to the object Buffer, 59% of entities head to the output Service). Placement of a third series of SACMI lines creates the space for a change of this ratio in favour of entities further processed for the purpose of CC manufacture. -- the volume of processed entities in the output Service grew by 10.03%, In the second part of experiment 5, the ratio of 100% of the processed entities was divided in the output from object LTG1 in the simulation model as follows: -- the volume of processed entities in the output CC grew by 6.53%, -- 60% head to the object Buffer (entities processed for the purpose of CC manufacture), -- the relative loading capacity of the object Buffer remained unchanged in comparison with the basic model, -- 40% head to the output Service (entities processed as a service). -- a decline in the overall period of processing the SACMI objects to 65.8% of the total operation time. Likewise, the capacity of the object Buffer was changed to 22 entities for the purpose of effective use of the operation time of the SACMI equipment, which (while the previously defined capacity of the object Buffer of 15 entities) expected 34.2% of its total operation time for the input to the process. The reason for the decline in overall processing period for the SACMI objects is the insufficient capacity of the object Buffer (15 entities) and likewise the simultaneous division of the flow of material in the output from object LTG1 (41% of entities head RESULTS, DISCUSSION AND CONCLUSIONSf -- 478 entities in the output Service (a decline by 20.07% in comparisons with the basic model, a decline by 27.36% in comparison with the original model of experiment 5), -- 597 entities in the output CC (growth by 50% in comparison with the basic model, growth by 40.8% in comparisons with the original model of experiment 5).f In the following table the anticipated effect of the proposed improvements on the basic simulation of experiments 1-5 is described. The number of processed entities was monitored for a period of 30 days overall and for the individual outputs Service and CC. The weights of financial demand of the introduced changes in practice (1-low, 2-medium, 3-high) were assigned to the individual experiments, and the generated outputs were recalculated for their possible financial evaluation, because the processed entities in the outputs Service and CC have different selling prices. The selling price presented is inclusive of the price of the initial input – the sheet metal. In experiments 1-3 each of the proposals for improvement was simulated independently, and in experiments 4 and 5 the combination of proposals for improvement was simulated; therefore experiments 1-3 and experiments 4 and 5 will be evaluated separately. The change of the selected evaluations is depicted as a percentage with respect to the basic model (a decline by red, growth by green). Experiments 1 and 3, unlike experiment 2, did not lead to any overall growth of production, but only to a shift of part of the processed sheets in favour of the process of CC manufacture, which is financially more interesting for the organization. For the mentioned reason, experiments 4 and 5, in which a combination of introducing all three of the proposed improvements, were carried out. From the results depicted in Tab. 1 it is evident that experiment 4 achieved better results, which is related to the volume of production of the simulated system; however, after taking the necessary investment costs into account, experiment 5, in which the manufacture system was simulated after introducing of proposals 2 and 3, was selected as the most suitable. The benefit of the experimental Table 1 Comparison of the results of individual experiments Experiment Description Number of processed entities Weight of fin. demand Value of output in EUR Fin. RESULTS, DISCUSSION AND CONCLUSIONSf RESULTS, DISCUSSION AND CONCLUSIONS From experiments 1-3 proposal number 2, the effect of which was simulated in experiment 2, was selected as the most suitable proposal for improvement for introduction into the real manufacturing system. The feasibility of improvement of number two appears in the assumed low costs (in comparison with the costs necessary for carrying out the other proposals) and was the only one that generated growth of production of the simulated manufacturing system; however, only for the output of sheets processed as a service. For the process of CC manufacture this fact means potential for re-evaluating the current capacities of CC manufacture, since CC manufacture brings the organization 65.1% higher financial evaluation than the service itself – lacquering and colour printing of sheet metal. Experiments 1 and 3, unlike experiment 2, did not lead to any overall growth of production, but only to a shift of part of the processed sheets in favour of the process of CC manufacture, which is financially more interesting for the organization. For the mentioned reason, experiments 4 and 5, in which a combination of introducing all three of the proposed improvements, were carried out. From the results depicted in Tab. 1 it is evident that experiment 4 achieved better results, which is related to the volume of production of the simulated system; however, after taking the necessary investment costs into account, experiment 5, in which the manufacture system was simulated after introducing of proposals 2 and 3, was selected as the most suitable. The benefit of the experimental From experiments 1-3 proposal number 2, the effect of which was simulated in experiment 2, was selected as the most suitable proposal for improvement for introduction into the real manufacturing system. The feasibility of improvement of number two appears in the assumed low costs (in comparison with the costs necessary for carrying out the other proposals) and was the only one that generated growth of production of the simulated manufacturing system; however, only for the output of sheets processed as a service. For the process of CC manufacture this fact means potential for re-evaluating the current capacities of CC manufacture, since CC manufacture brings the organization 65.1% higher financial evaluation than the service itself – lacquering and colour printing of sheet metal. Figure 11 Statistical output 2 from the model of experiment 5 Figure 12 Loading of the equipment 2 in the model of experiment 5 Figure 12 Loading of the equipment 2 in the model of experiment 5 “Naše more” 62(3)/2015., pp. 193-199 The most suitable variant is in the table indicated by the pink colour. Furthermore, the change versus all of the previous experiments consisted in a change of capacity of the object Buffer to 22 entities and a change of the ratio of the divided processed entities in the output from object LTG1. This change was shown to be effective, because the total period of processing the SACMI objects grew to 93% from the original 65.8%. This is also significantly manifested in the number of processed entities for the individual outputs: RESULTS, DISCUSSION AND CONCLUSIONSf contribution in EUR CC Service Total CC Service Total 0 basic model 398 (100%) 598 (100%) 996 (100%) - 1321887 1201980 2523867 (100%) - 1 shortening the period of setting of lacquer 410 (+3.02%) 587 (-1.84%) 997 (-) 3 1361742 1179870 2541612 17745 (+0.7%) 2 shortening the period of change of lacquer/colour 398 (-) 685 (+14.55%) 1083 (+8.73%) 1 1321887 1376850 2698736 174869 (+6.93%) 3 increasing prod. capacity of SACMI lines 425 (+6.78%) 570 (-4.68%) 995 (-) 2 1411562 1145700 2557262 33395 (+1.32%) 4 combination of exp. 1,2,3 614 (+54.27%) 468 (-21.74%) 1082 (+8.63%) 1+2+3 2039292 940680 2979972 456105 (+18.07%) 5 combination of exp. 2,3 597 (+50%) 478 (-20.07%) 1075 (+7.93%) 1+2 1982830 960780 2943610 419743 (+16.63%) Table 1 Comparison of the results of individual experiments M. Hovanec et al: Evaluation of Practical Application of Plant... [6] Hovanec M., Sinay J., Pačaiová H. (2014). Application of Proactive Ergonomics Utilizing Digital Plant Methods Based on Augmented Reality as a Tool Improving Prevention for Employees - 2014. In: International Symposium on Occupational Safety and Hygiene: 13. - 14.2.2014: Guimares, Portugalsko P. 182-185 Guimares : SPOSHO, 2014, ISBN : 978-989-98203-2-6 part of the contribution is mainly in verifying the considered improvements in a virtual environment through modern simulation software. The management of the organization thus obtained the opportunity to imagine the possible benefits of the proposed solutions. [7] Piľa, J. - Adamčík, F. - Korba, P. - Antoško, M. (2014). Safety Hazard and Risk in Slovak Aviation Regulations - 2014. In: Our Sea, International Journal of Maritime Science and Technology. Vol. 61, no. 1-2 p. 27-30. - ISSN 1848-6320 This paper was elaborated during realization of the project APVV-0337-11 “Research of new and newly arising risks in the industrial technologies in terms of the integrated safety as an assumption of a sustained development management.” [8] Korba P., Piľa J., Főző, L.,Cibereová, J. (2014). SGEM 2014 : 14th international multidiscilinary scientific geoconference : GeoConference on Informatics, Geoinformatics and Remote Sensing : conference proceedings : volume 1 : 17-26, June, 2014, Albena, Bulgaria. - Sofia : STEF92 Technology Ltd., 2014 P. 399-406. - ISBN 978-619-7105-10-0 This paper was elaborated during realization of the project University Science Park TECHNICOM for Innovation Applications Supported by Knowledge Technology, ITMS: 26220220182, supported by the Research & Development Operational Programme funded by the ERDF.” [9] Hovanec M., Varga M., Sobota B., Pačaiová H. (2012). RESULTS, DISCUSSION AND CONCLUSIONSf Inovatívne trendy a vízie v ergonómii využitím rozšírenej a virtuálnej reality. In: Aktuálne otázky bezpečnosti práce: 25-th International conference, Štrbské Pleso - Vysoké Tatry, 06.-08. 11.2012. p. 1-7, ISBN 978-80-553-1113-5. [10] Lestyánszka, Škurková, K. (2013). Using the shewhart control charts by process control. Production Engineering Archives. - ISSN 2353-5156. - Č. 1, s. 29-31. This contribution originated with support of SOVA DIGITAL a.s. as the exclusive supplier of PLM (Product Lifecycle Management) solutions from Siemens PLM Software for Slovakia. [11] Lisuch, J., Gonos, J. (2014). Use of logistic approach in optimizing the rotary kiln run, Applied Mechanics and Materials. Vol. 483 p. 518-523. ISSN 1660- 9336. [12] Smutná, M., Dulina, Ľ. (2010). Metódy a softvérová podpora v priemyselnej ergonómii, CD. Slovenská ergonomická spoločnosť (SES) 2010, 146 s. ISBN 978-80-970525-6-0. [13] HNÁT, J.: Assembly line balancing problem solved by generic algorithm. In Advanced Industrial Engineering. Wydawnictwo Fundacji Centrum Nowych Technologii, Bielsko-Biała. 2013. p. 7-22. ISBN 978-83-927531-6-2. REFERENCES [1] Bangsow S. (2010). Manufacturing Simulation with Plant Simulation and SimTalk, Second edition, Springer Springer-Verlag Berlin Heidelberg: p. 297, ISBN 978-3-642-05073-2. [14] Mleczko, J. – Mičieta, B. – Dulina, Ľ. (2013). Identification of bottlenecks in the unit make to order production. In: Applied Computer Science Vol. 9, No. 2 (2013), s. 43-56. Lublin, Lublin University of Technology. Institute of Technological Systems of Information. ISSN 1895-3735 (CC) [2] Tuček, J., Majlingová, A. (2004). Vyhodnotenie kvality a presnosti vybraných digitálnych modelov terénu, Aktuální problémy fotogrammetrie a DPZ [elektronický zdroj] : sborník konference. - Praha : FSv ČVUT, 2004. - ISBN 80- 01-03171-3. - 13 s. Technological Systems of Information. ISSN 1895-3735 (CC) [15] Kozuba J. (2013) Impact of human factor on likehood of aircraft accident, Proceding on the International Science Conference TRANSPORT SYSTEMS TELEMATICS – TST-11, Katowice 2011, s. 29-36, ISBN: 978-83-927504-8-2; [3] Hrozek, F., Sobota, B. Szabó, Cs., Korečko, Š., Varga, M., Ivančak, P. (2011). Augmented reality application in parallel computing system, 7th International Workshop on Grid Computing for Complex Problems, Bratislava, Slovakia, 24 –26 October 2011,Ústav Informatiky SAV, 2011, pp.118-125,978-80-970145- 5-1. [16] Martinka J., Hroncová E., Chrebet T., Balog K. (2014) The influence of spruce wood heat treatment on its thermal stability and burning process, European Journal of Wood and Wood Products, Springer Berlin Heidelberg, Volume 72, Issue 4 , pp 477-486, ISSN: 1436-736X; [17] Ščurek, R. (2008): The Improvised Explosive Device threat to air transport, Academic journal Mechanics, Transport, Communications, Tododr Kableshkov Higher School in Sofia, Gabrovo Technical University and the Lyuben Karavelov Higher School of Construction in Sofia, Bulgaria [4] Piľa J., Antoško M., Korba P. (2014). Ergonomy Of An Atco Training Work Place, Croatia, Naše more, ISSN 0469-6255. [5] Korba P., Piľa J. (2013). Aplikácia Cax Systémov Pri Projektovaní Konštrukčných Uzlov Vrtľníka, 1. vyd. - Puławy: Zakład Poligraficzny WISŁA, 191 p. ISBN 978- 83-937543-3-5. [18] Rupová, M. , Skrěhot, P. (2009) Actual questions about safety at work with nanomaterials NANOCON - 1st International Conference, Conference Proceedings 199 “Naše more” 62(3)/2015., pp. 193-199
https://openalex.org/W3111616548	https://biblio.ugent.be/publication/8702584/file/8703291	English	null	Dependence of inclusive jet production on the anti-kT distance parameter in pp collisions at $$ \sqrt{\mathrm{s}} $$ = 13 TeV	The Journal of high energy physics/The journal of high energy physics	2,020	cc-by	22,706	Keywords: Hadron-Hadron scattering (experiments), Jets, QCD ArXiv ePrint: 2005.05159 ArXiv ePrint: 2005.05159 https://doi.org/10.1007/JHEP12(2020)082 Contents 1 Introduction 1 2 The CMS detector 3 3 Jet reconstruction 4 4 Event samples 6 4.1 Collision data 6 4.2 Simulated samples 6 5 Measurement of cross sections and cross section ratios 8 5.1 Unfolding 8 5.2 Nonperturbative corrections for ﬁxed-order calculations 9 6 Experimental uncertainties in the measurement 10 7 Theoretical uncertainties 13 8 Results 14 8.1 Comparison of ratio of cross sections 14 8.2 Variation of the ratio of cross sections with jet size 16 9 Summary 16 The CMS collaboration 23 Contents 1 Introduction 1 2 The CMS detector 3 3 Jet reconstruction 4 4 Event samples 6 4.1 Collision data 6 4.2 Simulated samples 6 5 Measurement of cross sections and cross section ratios 8 5.1 Unfolding 8 5.2 Nonperturbative corrections for ﬁxed-order calculations 9 6 Experimental uncertainties in the measurement 10 7 Theoretical uncertainties 13 8 Results 14 8.1 Comparison of ratio of cross sections 14 8.2 Variation of the ratio of cross sections with jet size 16 9 Summary 16 The CMS collaboration 23 Contents 1 Introduction 1 2 The CMS detector 3 3 Jet reconstruction 4 4 Event samples 6 4.1 Collision data 6 4.2 Simulated samples 6 5 Measurement of cross sections and cross section ratios 8 5.1 Unfolding 8 5.2 Nonperturbative corrections for ﬁxed-order calculations 9 6 Experimental uncertainties in the measurement 10 7 Theoretical uncertainties 13 8 Results 14 8.1 Comparison of ratio of cross sections 14 8.2 Variation of the ratio of cross sections with jet size 16 9 Summary 16 The CMS collaboration 23 3 Jet reconstruction JHEP12(2020)082 The CMS collaboration Published for SISSA by Springer Received: May 11, 2020 Accepted: November 3, 2020 Published: December 11, 2020 Received: May 11, 2020 Accepted: November 3, 2020 Published: December 11, 2020 Dependence of inclusive jet production on the anti-kT distance parameter in pp collisions at √s = 13 TeV JHEP12(2020)082 The CMS collaboration E-mail: cms-publication-committee-chair@cern.ch E-mail: cms-publication-committee-chair@cern.ch Abstract: The dependence of inclusive jet production in proton-proton collisions with a center-of-mass energy of 13 TeV on the distance parameter R of the anti-kT algorithm is studied using data corresponding to integrated luminosities up to 35.9 fb−1 collected by the CMS experiment in 2016. The ratios of the inclusive cross sections as functions of transverse momentum pT and rapidity y, for R in the range 0.1 to 1.2 to those using R = 0.4 are presented in the region 84 < pT < 1588 GeV and \|y\| < 2.0. The results are compared to calculations at leading and next-to-leading order in the strong coupling constant using diﬀerent parton shower models. The variation of the ratio of cross sections with R is well described by calculations including a parton shower model, but not by a leading-order quantum chromodynamics calculation including nonperturbative eﬀects. The agreement between the data and the theoretical predictions for the ratios of cross sections is signiﬁcantly improved when next-to-leading order calculations with nonperturbative eﬀects are used. 1 Introduction Quantum chromodynamics (QCD) is a gauge theory describing the strong interaction be- tween partons (quarks and gluons). Jets are reconstructed using hadron particles produced by the fragmentation of partons in collisions [1]. Thus jets approximate the original par- tons created in short-distance scatterings. The production cross sections for high transverse momentum (pT) partons can be calculated using perturbative QCD (pQCD). Speciﬁcally, predictions for hadron production in proton-proton collisions require models for parton showering [2–4] and nonperturbative (NP) eﬀects such as hadronization [5] and underlying event (UE) [6]. When the ﬁxed-order prediction in pQCD is not adequate, higher-order terms must be included using resummation methods [7–9]. The results of measurements of inclusive jet production cross sections for proton-proton collisions are typically presented using the anti-kT jet algorithm [10] characterized by a dis- tance parameter R, which is a measure of the jet size in rapidity-azimuth plane. Anti-kT – 1 – jets with distance parameter R are referred to as AKn jets, where R = 0.1n. The CMS col- laboration [11] has reported measurements at center-of-mass energy (√s) of 7 TeV [12] and 8 TeV [13] using AK5 and AK7 jets. The CMS results at √s = 13 TeV for AK4 and AK7 jets are reported in ref. [14]. After the application of a correction for NP and electroweak eﬀects, the results for AK7 jets are well described by next-to-leading order (NLO) calcu- lations based on the nlojet++ [15] program used in the fastnlo software package [16]. The prediction from the powheg [17] generator, which also computes matrix elements at NLO and is used with parton showering simulated with pythia8 [18] or herwig++ [19], describes results well for both AK7 and AK4 jets. However, the ATLAS collaboration has measured the production cross sections for both AK4 and AK6 jets and ﬁnds a discrepancy between the measured results and the powheg prediction [20]. The ATLAS collaboration has also compared the measurements of inclusive jet production at √s = 13 TeV with the next-to-next-to-leading order (NNLO) prediction in pQCD [21]. JHEP12(2020)082 The measurement of a jet production cross section as a function of the distance pa- rameter is sensitive to the details of the theoretical modeling of the perturbative and NP processes involved in the evolution of the partons. The measurement of the ratio of cross sections with two jet sizes was ﬁrst performed by the ALICE collaboration with AK2 and AK4 jets [22]. 1 Introduction A similar study was also produced by the CMS collaboration with AK5 and AK7 jets [23]. We explore this topic further in the present paper by extending the mea- surement to various values of jet size. Recently, ALICE collaboration has also measured both the absolute cross sections of inclusive jet production and the ratio of cross sections for R = 0.1–0.6 in 20 < pT < 140 GeV [24]. Dependence of inclusive jet production on the distance parameter is also studied in detail in ref. [25]. Quarks and gluons radiate secondary gluons that can be emitted outside of the catch- ment area of the jet deﬁnition, which is the region in rapidity-azimuth plane contributing to the jet. This lost pT is calculated using a QCD splitting function, with the leading-order (LO) result [26–28] in the small-R approximation (R ≪1) (δpT)q = −CF αSpT π ln 1 R 2 ln 2 −3 8 + O(αS), (1.1) (1.1) for quark-initiated jets and (δpT)g = −αSpT π ln 1 R CA(2 ln 2 −43 96) + TRnf 7 48 + O(αS), (1.2) (1.2) for gluon-initiated jets. Here CF = 4 3 and CA(= 3) are the Casimir factors for quarks and gluons respectively, TR = 1 2 is the SU(3) quantum number, and nf is the number of active quark ﬂavors. Larger values of R capture a larger fraction of the radiation. Properties of jets are also modiﬁed by hadronization, an NP process describing the transition of partons into hadrons. As described in ref. [29], some theoretical models parameterize the eﬀect of hadronization by taking αS(µ) = µlδ(µ −µl), where µl is com- mensurate with the Landau pole, yielding (δpT)had ≃−2CA(µl) πR + O(R), (1.3) (1.3) – 2 – in the small-R limit, where C = CF(CA) for quark (gluon) initiated jets, and A(µl) is related to the scale appearing in the calculations of hadronization. Losses are again minimized at larger values of R. The algorithm deﬁning the jets can also select particles from the underlying event, which in general involves low momentum transfer. These particles typically have low pT. The energy density (ΛUE per unit y) from these sources is approximately uniform over the jet area, and their contribution to the jet pT is approximately given [28, 30] by (δpT)UE ≃1 2ΛUER2, (1.4) (1.4) JHEP12(2020)082 for small R values. Since, as discussed above, the contributions of various perturbative and NP eﬀects de- pend on the jet size, and because radiation and hadronization are diﬀerent for jets initiated by quarks and by gluons, comparisons of jets with diﬀerent cone sizes yield information about these processes, and can be used to improve theoretical calculations. In this paper, we present measurements of the ratio of the cross section for inclusive anti-kT jets with distance parameters of R = 0.1, 0.2, . . . , 1.2 to that of AK4 jets. The re- sults are compared with predictions from diﬀerent Monte Carlo (MC) generators, involving matrix element calculations at diﬀerent orders and utilizing diﬀerent parton shower and hadronization models. Predictions for cross section ratios have also been obtained using a pQCD calculation at NLO that uses the following convention Ratio(R, pT) = dσ(R) dpT −dσ(0.4) dpT dσ(0.4) dpT + 1, (1.5) (1.5) where R is the anti-kT jet distance parameter, and R = 0.4 is taken as the reference jet size. The terms in eq. (1.5) are diﬀerential cross sections for three-jet production and are calculated at ﬁxed-order using nlojet++ with terms up to α4 S [31, 32]. Measurements are restricted to pT < 1588 GeV because of the large experimental uncertainty in the calibration of high energy jets, which was not optimized for the cross section ratios. 2 The CMS detector The central feature of the CMS apparatus is a superconducting solenoid of 6 m internal diameter, providing a magnetic ﬁeld of 3.8 T. Within the solenoid volume are a silicon pixel and strip tracker, a lead tungstate crystal electromagnetic calorimeter (ECAL), and a brass and scintillator hadron calorimeter (HCAL), each composed of a barrel and two endcap sections. Forward calorimeters extend the pseudorapidity (η) coverage provided by the barrel and endcap detectors. Muons are detected in gas-ionization chambers embedded in the steel ﬂux-return yoke outside the solenoid. A more detailed description of the CMS detector, together with a deﬁnition of the coordinate system used and the relevant kinematic variables, can be found in ref. [11]. The silicon tracker measures charged particles within the range \|η\| < 2.5. It consists of 1440 silicon pixel and 15 148 silicon strip detector modules. For nonisolated particles – 3 – of 1 < pT < 10 GeV and \|η\| < 1.4, the track resolutions are typically 1.5% in pT and 25–90 (45–150) µm in the transverse (longitudinal) impact parameter [33]. In the barrel section of the ECAL, an energy resolution of about 1% is achieved for unconverted or late- converting photons that have energies in the range of tens of GeV. The remaining barrel photons have a resolution of about 1.3% up to \|η\| = 1, rising to about 2.5% at \|η\| = 1.4. In the endcaps, the resolution of unconverted or late-converting photons is about 2.5%, while the remaining endcap photons have a resolution between 3 and 4% [34]. In the region \|η\| < 1.74, the HCAL cells have widths of 0.087 in η and 0.087 radians in azimuth (φ). In the η–φ plane, and for \|η\| < 1.48, the HCAL cells map on to 5 × 5 arrays of ECAL crystals to form calorimeter towers projecting radially outwards from close to the nominal interaction point. For \|η\| > 1.74, the coverage of the towers increases progressively to a maximum of 0.174 in ∆η and ∆φ [35]. Within each tower, the energy deposits in ECAL and HCAL cells are summed to deﬁne the calorimeter tower energies, subsequently used to provide the energies and directions of hadronic jets. 2 The CMS detector When combining information from the entire detector, the jet energy resolution amounts typically to 15% at 30 GeV, 10% at 100 GeV, and 5% at 1 TeV at \|η\| < 0.5, while at \|η\| = 2.0 the jet energy resolution increases by 1–2% at low pT [36]. JHEP12(2020)082 Events of interest are selected using a two-tiered trigger system [37]. The ﬁrst level, composed of custom hardware processors, uses information from the calorimeters and muon detectors to select events at a rate of around 100 kHz. The second level, known as the high-level trigger (HLT), consists of a farm of processors running a version of the full event reconstruction software optimized for fast processing that reduces the event rate to around 1 kHz before data storage. 3 Jet reconstruction The CMS particle-ﬂow algorithm [38] reconstructs and identiﬁes each individual particle in an event with an optimized combination of information from the various elements of the CMS detector. The energy of photons is obtained from the ECAL measurement. The energy of electrons is determined from a combination of the electron momentum at the primary interaction vertex as determined by the tracker, the energy of the corresponding ECAL cluster, and the energy sum of all bremsstrahlung photons spatially compatible with originating from the electron track. The energy of muons is obtained from the curvature of the corresponding track as determined using the tracker and the muon system. The energy of charged hadrons is determined from a combination of their momentum measured in the tracker and the matching ECAL and HCAL energy deposits, corrected for zero-suppression eﬀects and for the response function of the calorimeters to hadronic showers. Finally, the energy of neutral hadrons is obtained from the corresponding corrected ECAL and HCAL energy deposits. For each event, hadronic jets are clustered from these reconstructed particles (particle- ﬂow candidates) using the infrared- and collinear-safe anti-kT algorithm [10], as imple- mented in the FastJet package [39]. The jet momentum is determined as the vectorial sum of all particle momenta in the jet, and is found from simulation to be, on average, – 4 – within 5 to 10% of the momentum of the particle-level jets reconstructed using stable par- ticles (lifetime >30 ps) excluding neutrinos, for jet pT > 50 GeV and rapidity \|y\| < 2.5. Additional proton-proton interactions within the same or nearby bunch crossings (pileup) can contribute additional tracks and calorimetric energy depositions, increasing the ap- parent jet momentum. To mitigate this eﬀect, tracks identiﬁed as originating from pileup vertices are discarded and an oﬀset correction [40] is applied to correct for remaining contri- butions [41]. Additional selection criteria are applied to each jet to remove jets potentially dominated by instrumental eﬀects or reconstruction failures [42]. These criteria consist of the following conditions: the energy fraction of the jet carried by neutral hadrons and photons should be less than 90%, the jet should have at least two constituents, and at least one of those should be a charged hadron. This set of criteria is more than 99% eﬃcient for genuine jets. 3 Jet reconstruction JHEP12(2020)082 The missing transverse momentum vector (⃗p miss T ) is deﬁned as the negative vector sum of the pT of all reconstructed particle-ﬂow objects in an event; its magnitude is de- noted using pmiss T . A set of algorithms is used to reject events with anomalous high-pmiss T arising from a variety of reconstruction failures, detector malfunctions, or noncollision backgrounds; these algorithms are designed to identify more than 85–90% of the spurious high-pmiss T events with a misidentiﬁcation rate of less than 0.1% [43]. Jet energy corrections are derived using simulated pythia inclusive jet samples dis- cussed in section 4.2 so that the average measured transverse momentum of jets is the same as the corresponding particle-level jets. This methodology is used to derive nominal jet energy correction factors only for AK4 and AK8 jets. The nominal corrections for AK4 jets are used for AK1 to AK6 jets. For larger jet sizes (R > 0.6) the nominal correction factors derived for AK8 jets are applied. To account for the diﬀerences in the distance parameter, an extra correction factor (CR) is determined in case of each distance parameter for the average pileup condition based purely on simulation and applied to the corresponding jets. A detailed discussion on the derivation of CR is made later in this section. Measurements of the momentum balance in dijet, photon+jet, Z+jet, and multijet events are used to determine any residual diﬀerences between the jet energy scale in data and simulation, and appropriate corrections are made [41]. The in situ techniques are based on the missing transverse momentum projection fraction method, which is fundamentally insensitive to jet size, and on particle-ﬂow reconstruction, whose reliance on tracking and particle-ﬂow hadron calibration further reduces diﬀerences in energy response between jets of diﬀerent radii. Residual corrections are derived using only AK4 jets and applied to jets of all the sizes. The factor CR is derived in the following way. In simulated pythia inclusive jet samples, the detector- and particle-level jets are required to be geometrically matched satisfying ∆R < (0.5×jet size); the ratio of the average detector-level jet pT to the particle- level jet pT is calculated as a function of the particle-level jet pT for all the jet sizes and then used as an extra correction factor CR for both data and simulation. 3 Jet reconstruction The CR factors are also derived using simulated herwig++ inclusive jet samples, so the jet energy response is calibrated to unity for all the jet sizes in herwig++ samples as well. The diﬀerence in CR factors derived using pythia and herwig++ inclusive jet samples is used to estimate – 5 – the systematic uncertainties in CR separately for each of the distance parameters. The CR factors are important to ensure that the jet energy resolution in simulation is properly corrected to match the data without changing the jet energy scale, and that pythia and herwig++ are on equal footing with respect to jet energy scale and resolution when unfolding the data in section 5.1. The value of the CR factor ranges from 0.95 to 1.10 depending on the energy, rapidity, and size of the reconstructed jets; this correction is signiﬁcant only for very small and very large jet sizes. However, CR corrects the jet energy response for diﬀerent jet sizes at the level of simulation only and no dedicated residual correction is derived for data as a function of jet size. Nevertheless, it has been checked that the diﬀerence in average pT between AK8 and AK4 jets pointing in the same direction in data is smaller than the diﬀerence of the same between pythia and herwig++ simulations. JHEP12(2020)082 4.1 Collision data Proton-proton collision data collected by the CMS experiment during 2016, corresponding to an integrated luminosity of 35.9 fb−1, are used for this analysis. The data sample is collected using single-jet triggers, which select events containing at least one AK8 jet, formed from particle-ﬂow candidates, with pT exceeding one of the threshold values listed in table 1. Absolute trigger eﬃciency is measured with a tag-and-probe procedure [44] using the events having a back-to-back dijet topology, where the tag jet is matched to a single-jet trigger, and the eﬃciency is measured using the probe jet. Because of limited bandwidth and storage space, only a fraction of the events satisfying the triggering condition with lower thresholds are recorded. For this reason, in each jet-pT bin, only the trigger that has the highest eﬀective integrated luminosity and is also more than 99% eﬃcient is used. Oﬄine, events are required to contain at least one jet with pT above that value for which the trigger is 99% eﬃcient. These values are also used to deﬁne the pT bins for the measurement. Similarly to other CMS publications [12, 13], we require that pmiss T / P i pT i < 0.3, where the index i runs over all particle-ﬂow candidates in the event and P i pT i denotes the scalar sum of transverse momenta; this rejects calorimeter noise and a part of electroweak backgrounds from the production of W(→lepton) + jets, Z(→lepton) + jets, and top quarks when the top quarks decay to ﬁnal states with leptons. 4.2 Simulated samples The data are compared to predictions from several diﬀerent MC generators, listed below. The data are compared to predictions from several diﬀerent MC generators, listed below. The pythia v8 212 [18] generator computes matrix elements only for 2 →2 Feynman The data are compared to predictions from several diﬀerent MC generators, listed below. The pythia v8.212 [18] generator computes matrix elements only for 2 →2 Feynman diagrams at LO; the missing orders in the perturbation series are approximated using pT- ordered dipole showering. The pythia generator employs the empirical Lund string model to hadronize the partons. The NNPDF2.3 [45] LO parton distribution function (PDF) set – 6 – Trigger pT (GeV) threshold pT (GeV) range for analysis Eﬀective integrated luminosity (fb−1) 40 74–97 0.000050 60 97–133 0.00033 80 133–196 0.00104 140 196–272 0.0105 200 272–330 0.084 260 330–395 0.517 320 395–468 1.54 400 468–548 4.68 450 548–∞ 33.4 Table 1. Trigger pT thresholds and eﬀective integrated luminosity of the HLT triggers based on AK8 jets. These triggers were not active during the initial part of data taking in 2016, thus the maximum integrated luminosity is less than 35.9 fb−1. JHEP12(2020)082 Table 1. Trigger pT thresholds and eﬀective integrated luminosity of the HLT triggers based on AK8 jets. These triggers were not active during the initial part of data taking in 2016, thus the maximum integrated luminosity is less than 35.9 fb−1. Table 1. Trigger pT thresholds and eﬀective integrated luminosity of the HLT triggers based on AK8 jets. These triggers were not active during the initial part of data taking in 2016, thus the maximum integrated luminosity is less than 35.9 fb−1. is used to describe the momentum fractions carried by the partons within the incoming protons; our UE model is the CUETP8M1 tune [46] (CMS Underlying Event Tune for pythia8 based on Monash [47]), which was derived by tuning the model parameters using minimum bias data collected by the CMS collaboration. is used to describe the momentum fractions carried by the partons within the incoming protons; our UE model is the CUETP8M1 tune [46] (CMS Underlying Event Tune for pythia8 based on Monash [47]), which was derived by tuning the model parameters using minimum bias data collected by the CMS collaboration. The herwig++ v2.7.1 [19] generator also calculates only 2 →2 scatterings, but has a diﬀerent fragmentation and hadronization model than pythia. 4.2 Simulated samples It employs angular-ordered showers to radiate the partons and a cluster model to produce the hadrons. The NNPDF3.0 LO PDF set is used, and the UE modeling is described by the CUETHppS1 tune [46]. The MadGraph (MadGraph5_amc@nlo V5 2.2.2) [48] generator provides calcu- lations of matrix elements with up to four outgoing partons in the ﬁnal state at LO. The partons are showered and hadronized with pythia combined with MadGraph, using the MLM merging scheme [49]. The NNPDF3.0 NLO PDF set and the CUETP8M1 UE tune are used here as well. The powheg v2 [17, 50] generator computes the dijet production cross section at NLO in pQCD. Successive parton showering, hadronization, and UE modeling is performed either using pythia with the CUETP8M1 tune (referred to as PH+P8 in the ﬁgures) or herwig++ with the CUETHppS1 and EE5C [51] tunes (referred to as PH+Herwig in the ﬁgures). The NNPDF3.0 NLO PDF set is used for powheg as well, and the value of the hdamp parameter in powheg is 250 GeV. The herwig 7.1.1 [52] generator, used with the NNPDF3.1 NNLO PDF set with αS (mZ) = 0.118, also evaluates the matrix elements for dijet production at NLO, and is matched to the herwig7 parton shower using the FxFx [53] jet merging method. The CH2 tune is used to model UE. This prediction is referred to as herwig7. Fixed-order predictions for dijet production at NLO are computed using nlojet++ within the framework of the fastnlo package. To account for the eﬀects of hadronization, an additional correction factor is used, which will be discussed in section 5.2. This predic- tion is referred to as NLO⊗NP in the ﬁgures. Predictions from nlojet++ are obtained using the CT14NLO PDF set. – 7 – Recently, a prediction for single-inclusive jet production using joint resummation in the threshold energy in the small-R limit has been computed at next-to-leading logarithmic (NLL) accuracy in the framework of Soft Collinear Eﬀective Theory in refs. [9, 54]; the CT14NLO PDF set is also used for this prediction, which is referred to as (NLO+NLL). This prediction is compared with the measurements reported in this paper. 5 Measurement of cross sections and cross section ratios The inclusive jet cross section is calculated as The inclusive jet cross section is calculated as JHEP12(2020)082 d2σ dpTdy = 1 ϵLint Njets ∆pT∆y, (5.1) (5.1) where Njets is the number of jets in a pT and y bin, Lint is the integrated luminosity of the data set, ϵ is the product of trigger and event selection eﬃciencies, and ∆pT and ∆y are the bin widths in pT and y. The widths of pT bins are proportional to the jet energy resolution and increase with jet pT. The ratios of cross sections for the diﬀerent jet sizes with respect to AK4 jets is calculated as the bin-by-bin quotient of the cross sections of AKn (n= 1, 2, . . . , 12) and AK4 jets respectively; in the ratios, all the terms in eq. (5.1) except Njets and ϵ cancel. 5.1 Unfolding To correct for detector ineﬃciencies and resolution, a number of methods available in the RooUnfold package [55] are used to unfold the jet pT spectra. The nominal choice of unfolding technique in this paper is the D’Agostini unfolding [56] with early stopping. Up to 5–8 iterations are used depending on jet size and rapidity region. An alternative method is singular value decomposition (SVD) [57]. A third method is called bin-by-bin [58], which multiplies the particle-level spectra by the ratio between the detector-level spectra in data and simulation. The SVD and bin-by-bin techniques are used to cross-check the result of unfolding with the D’Agostini unfolding. As an additional cross-check, unfolding is also performed using a χ2 minimization without regularization using the TUnfold package [59]. Response matrices between pT spectra of detector-level and generator-level jets are obtained by one-to-one matching of the nearest detector- and particle-level jets, excluding matches with ∆R > (0.5 × jet size), where ∆R denotes the distance between detector- and particle-level jets in the rapidity-azimuth plane. This criterion leads to almost 100% matching eﬃciency between the detector-level and the particle-level jets. Response ma- trices are constructed, for all rapidity and jet sizes, from the CMS detector simulation based on Geant4 [60] using simulated samples from three MC event generators, pythia, herwig++, and MadGraph. For the particle-level results, response matrices based on the pythia simulation are used for the unfolding. The response matrix for AK4 jets in the ﬁrst rapidity region for the pythia sample is shown in ﬁgure 1. The response matrix is diagonal, which shows that unfolding works well. – 8 – (GeV) T Detector-level jet p 200 400 600 800 1000 1200 1400 (GeV) T Particle-level jet p 200 400 600 800 1000 1200 1400 Probability (detector \| particle level) 0 0.1 0.2 0.3 0.4 0.5 T Anti-k R = 0.4 \|y\| < 0.5 Pythia CMS Simulation (GeV) T Jet p 200 400 600 800 1000 1200 1400 (GeV) T Jet p 200 400 600 800 1000 1200 1400 Correlation coefficient 0.4 − 0.2 − 0 0.2 0.4 0.6 0.8 1 T Anti-k R = 0.4 \|y\| < 0.5 Unfolded by Pythia CMS (13 TeV) -1 < 35.9 fb Figure 1. Response matrix constructed from a simulation of a sample generated using pythia, for AK4 jets in the \|y\| < 0.5 bin (left). 5.1 Unfolding A correlation matrix generated after data is unfolded by the D’Agostini unfolding using pythia simulation for AK4 jets (right). (GeV) T Detector-level jet p 200 400 600 800 1000 1200 1400 (GeV) T Particle-level jet p 200 400 600 800 1000 1200 1400 Probability (detector \| particle level) 0 0.1 0.2 0.3 0.4 0.5 T Anti-k R = 0.4 \|y\| < 0.5 Pythia CMS Simulation (GeV) T Jet p 200 400 600 800 1000 1200 1400 (GeV) T Jet p 200 400 600 800 1000 1200 1400 Correlation coefficient 0.4 − 0.2 − 0 0.2 0.4 0.6 0.8 1 T Anti-k R = 0.4 \|y\| < 0.5 Unfolded by Pythia CMS (13 TeV) -1 < 35.9 fb JHEP12(2020)082 Figure 1. Response matrix constructed from a simulation of a sample generated using pythia, for AK4 jets in the \|y\| < 0.5 bin (left). A correlation matrix generated after data is unfolded by the D’Agostini unfolding using pythia simulation for AK4 jets (right). For both the D’Agostini and SVD unfolding techniques, the nearest neighbor pT bi are correlated, and the next-to-nearest bins are anti-correlated (right plot in ﬁgure 1 f For both the D’Agostini and SVD unfolding techniques, the nearest neighbor pT bins are correlated, and the next-to-nearest bins are anti-correlated (right plot in ﬁgure 1 for AK4 jets with the D’Agostini unfolding). Next-to-next-to-nearest bins are again correlated. Several cross-checks are made regarding the unfolding. To investigate possible bias due to the choice of MC generator used to construct the response matrices, event samples are generated using three diﬀerent generators: pythia, herwig++, and MadGraph, fol- lowed by the detector simulation whose output is scaled and smeared independently for each generator to match the energy scale and resolution of jets in data. Detector-level distributions from each of the samples are unfolded using these three response matrices, and the unfolded distributions are compared to the corresponding particle-level distribu- tions. No evidence for signiﬁcant bias is observed. Similarly, the data are unfolded using response matrices from these three simulated samples; the diﬀerences among the unfolded spectra are within systematic uncertainties corresponding to the correction factor CR. The same conclusion holds when comparing the unfolded distributions obtained using diﬀerent unfolding techniques, such as D’Agostini, SVD, bin-by-bin, and χ2 minimization. 5.2 Nonperturbative corrections for ﬁxed-order calculations and powheg+herwig++ (EE5C) samples is deﬁned as the ﬁnal NP correction, and the envelope of the diﬀerences is taken as its uncertainty. Figure 2 depicts the NP corrections for the cross section ratio of the AK2 and AK8 jets with respect to the AK4 jets. Hadronization corrections are larger for smaller jet sizes, and MPI introduces a larger correction for large-R jets. Because both hadronization and MPI are important for low-pT jets, the NP correction is also signiﬁcant in the low-pT portion of phase space; in the high-pT region, the NP correction factor approaches 1. For AK4 jets, the corrections for hadronization and MPI almost cancel, and the resulting NP correction is close to unity throughout the pT range. At around pT = 85 GeV, the correction goes down to 0.8 for AK2 jets, and it goes up to 1.25 for AK8 jets. 5.2 Nonperturbative corrections for ﬁxed-order calculations Fixed-order NLO calculations yield predictions for the partonic ﬁelds, but in experimen- tal measurements, jets are composed of hadrons. To evolve the parton-level prediction to the hadron level, NP corrections are calculated and applied. Although generators such as pythia and herwig come with MC-based phenomenological simulation of these processes, nlojet++ does not. The impact of NP on the nlojet++ prediction is approximated as a multiplicative correction factor as follows. The NP correction is the ratio of an observ- able from a generator, which includes NP eﬀects with hadronization and multiple parton interaction (MPI) processes switched on, to the same observable obtained from the same generator without NP eﬀects, i.e., by switching oﬀhadronization and MPI processes. Simulated powheg+pythia (CUETP8M1 tune) and powheg+herwig++ (EE5C and CUETHppS1 tunes) samples are used to compute NP factors for all the jet sizes in all the pT and rapidity bins. The average NP correction obtained from the powheg+pythia – 9 – (GeV) T Jet p 100 1000 NP correction for cross section ratio 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 (13 TeV) CMS Simulation PH+P8(CUETP8M1) PH+Herwig(EE5C) PH+Herwig(CUETHppS1) Final correction T Anti-k R = 0.2 \|y\| < 0.5 (GeV) T Jet p 100 1000 NP correction for cross section ratio 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 (13 TeV) CMS Simulation PH+P8(CUETP8M1) PH+Herwig(EE5C) PH+Herwig(CUETHppS1) Final correction T Anti-k R = 0.8 \|y\| < 0.5 Figure 2. Nonperturbative correction factor for the cross section ratio of inclusive AK2 (left) and AK8 jets (right) with respect to the AK4 jets in the rapidity bin \|y\| < 0.5. Vertical error bars represent the statistical uncertainty of the NP correction for diﬀerent predictions. (GeV) T Jet p 100 1000 NP correction for cross section ratio 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 (13 TeV) CMS Simulation PH+P8(CUETP8M1) PH+Herwig(EE5C) PH+Herwig(CUETHppS1) Final correction T Anti-k R = 0.2 \|y\| < 0.5 (GeV) T Jet p 100 1000 NP correction for cross section ratio 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 (13 TeV) CMS Simulation PH+P8(CUETP8M1) PH+Herwig(EE5C) PH+Herwig(CUETHppS1) Final correction T Anti-k R = 0.8 \|y\| < 0.5 JHEP12(2020)082 Figure 2. Nonperturbative correction factor for the cross section ratio of inclusive AK2 (left) and AK8 jets (right) with respect to the AK4 jets in the rapidity bin \|y\| < 0.5. Vertical error bars represent the statistical uncertainty of the NP correction for diﬀerent predictions. 6 Experimental uncertainties in the measurement Multiple sources of uncertainty aﬀect the precision of the measurement: statistical, jet en- ergy scale (JES) uncertainties, jet energy resolution (JER) uncertainties, and uncertainties in the pileup condition. We also include systematic uncertainties corresponding to the use of JES corrections derived for one R along with the R-dependent CR factor on jets formed using another R. To estimate the statistical uncertainty in data, the jackknife resampling [61] method is used. In this technique, ten diﬀerent data samples, each containing 90% events of the full data sample, are constructed such that the removed 10% of the events are complementary for each subsample. These subsamples are chosen in such a way that they correspond to very similar phase space regions. The statistical uncertainty is the standard deviation of the ten distributions multiplied by √ 9 = 3. The resulting statistical uncertainty is roughly <1% for jet pT < 1 TeV, and increases at high jet pT. A similar procedure is followed to estimate the statistical uncertainty due to the response matrices used for unfolding. Here also, ten subsets of the simulated sample are considered, each with a nonoverlapping 10% of events removed. The distributions in data are unfolded using each subsample, and the standard deviation of ten unfolded distributions, multiplied by a factor of 3, is the – 10 – statistical uncertainty due to the response matrices; in this case the statistical uncertainty is roughly 0.5–1.0% for the cross section ratio throughout the pT range. The jet energy scale corrections have a number of uncertainties corresponding to the techniques used and the amount of pileup. The JES has an uncertainty of about 1–2% in the central region [62]. The uncertainty is larger in the forward region and at low jet pT. To evaluate the uncertainty in the measurement of the cross section ratio, the JES is varied upwards and downwards by the uncertainties corresponding to diﬀerent sources. The diﬀerence in the unfolded cross section ratios using the nominal and varied JES is the uncertainty. Twenty-seven diﬀerent sources of JES uncertainty are considered individually and added in quadrature. The uncertainty because of JES is very similar for all the jet sizes, except for the pileup component. 6 Experimental uncertainties in the measurement The uncertainties mostly cancel out in the ratio, but there is a small residual, which is about 0.5–1.0% for \|y\| ≤2.0 up to 1 TeV of jet pT and goes up to 2% for very high jet pT. JHEP12(2020)082 To estimate the uncertainty in the ratio of cross sections with respect to that of AK4 jets because of using JES corrections derived for one value of R with jets from other values of R and then applying the CR factor, the standard calibration factors from the AK8 jets are applied to AK1 to AK6 jets, and, for jets of other sizes, the calibration factors for the AK4 jets are used. The CR factors for jets of all sizes are derived for this scenario. The systematic uncertainties in the inclusive jet cross section ratios are evaluated using the diﬀerence between the results obtained by these two procedures. The uncertainty coming from the CR correction is more signiﬁcant for larger jet sizes. The CR calibration factors are derived using both pythia and herwig++ simulations as a function of jet pT in diﬀerent rapidity bins for all jet sizes. The diﬀerence in the resulting CR corrections is an ‘R-dependent’ uncertainty, and it is deﬁned such that it vanishes for AK4 jets, which is used as the reference. The JER and its associated uncertainty are obtained from a dijet balance tech- nique [62]. The JER in data is worse than in simulation. To match the JER in data and simulation, a spreading is added to the jets in simulation. Here also, as in the case for the JES, cross section ratios are obtained using upward and downward variations of the energy resolution factors for simulation while unfolding the data. The diﬀerence with respect to the nominal unfolding is used as an estimate of the uncertainty. The uncertainty due to JER is more important for large-R jets at low pT. The uncertainty also grows in regions of larger rapidities. To match the pileup conditions in data and in MC simulation, pileup proﬁle weighting is performed for the simulated samples. The weighting factors depend on the total inelastic cross section; we vary its nominal value of 67.5 mb [63] up and down by its uncertainty of 2.6% when reconstructing the response matrices, and take the diﬀerence in the unfolded data as the uncertainty. 6 Experimental uncertainties in the measurement This source of systematic uncertainty is larger at low pT for large jet sizes, although its absolute value is small. The uncertainties from diﬀerent experimental sources are added in quadrature, and the total uncertainty is shown in ﬁgure 3 for the cross section ratios of the AK2 and AK8 jets with respect to the AK4 jets. – 11 – (GeV) T Jet p 100 1000 Relative uncertainty 0.04 − 0.02 − 0 0.02 0.04 (13 TeV) -1 < 35.9 fb CMS T Anti-k R = 0.2 \|y\| < 0.5 JES JER Pileup R-dependent R C Total Statistical uncertainty (Data) Statistical uncertainty (RM) (GeV) T Jet p 100 1000 Relative uncertainty 0.04 − 0.02 − 0 0.02 0.04 (13 TeV) -1 < 35.9 fb CMS T Anti-k R = 0.8 \|y\| < 0.5 JES JER Pileup R-dependent R C Total Statistical uncertainty (Data) Statistical uncertainty (RM) Figure 3. Total uncertainty (relative) from experimental sources for the ratio of cross section of inclusive jets of size 0.2 (top) and 0.8 (bottom) with respect to that of AK4 jets in the rapidity bin \|y\| < 0.5. Statistical uncertainties are also overlaid as vertical black (red) bars for data (response matrices, RM, in simulation). (GeV) T Jet p 100 1000 Relative uncertainty 0.04 − 0.02 − 0 0.02 0.04 (13 TeV) -1 < 35.9 fb CMS T Anti-k R = 0.2 \|y\| < 0.5 JES JER Pileup R-dependent R C Total Statistical uncertainty (Data) Statistical uncertainty (RM) (GeV) T Jet p 100 1000 Relative uncertainty 0.04 − 0.02 − 0 0.02 0.04 (13 TeV) -1 < 35.9 fb CMS T Anti-k R = 0.8 \|y\| < 0.5 JES JER Pileup R-dependent R C Total Statistical uncertainty (Data) Statistical uncertainty (RM) JHEP12(2020)082 Figure 3. Total uncertainty (relative) from experimental sources for the ratio of cross section of inclusive jets of size 0.2 (top) and 0.8 (bottom) with respect to that of AK4 jets in the rapidity bin \|y\| < 0.5. Statistical uncertainties are also overlaid as vertical black (red) bars for data (response matrices, RM, in simulation). In the cross section ratio, many of the systematic uncertainties almost cancel, so the ﬁnal uncertainty is small. The statistical component of the uncertainty is also shown in the same ﬁgure. In the cross section ratio, many of the systematic uncertainties almost cancel, so the ﬁnal uncertainty is small. 6 Experimental uncertainties in the measurement The statistical component of the uncertainty is also shown in the same ﬁgure. The experimental systematic uncertainty at low pT and large R is dominated by the pileup uncertainty. The JER uncertainty is also larger there because of additional spreading caused by pileup. At intermediate pT, the uncertainty is dominated by the CR uncertainty; at high pT the JES dominates the experimental uncertainty because the cross sections fall very steeply and event counts are small at high jet pT. The sizes of the statistical – 12 – uncertainties are similar to those of the total systematic uncertainties and are dominated by data at high pT and by the uncertainty in the response matrix because of the number of MC events at intermediate pT. At low pT, the data have similar statistical uncertainties as the simulated sample, since the corresponding triggers are prescaled. Another source of uncertainty, which is relevant only for jets with R > 0.8, is the uncertainty in the trigger eﬃciency correction. The AK8 single-jet triggers are not fully eﬃcient for larger jet sizes near the trigger turn-on points for AK8 jets; an eﬃciency correction is applied for those jet sizes following eq. (5.1). The diﬀerence in the absolute value of the trigger eﬃciency from the curve used to model the variation of trigger eﬃciency as a function of jet pT is the uncertainty. The size of this uncertainty is 0.5–1.0% throughout the pT range. JHEP12(2020)082 7 Theoretical uncertainties Apart from the systematic uncertainties due to experimental sources, theoretical calcula- tions and generators have uncertainties in their predictions for the cross section ratio. For the ﬁxed-order predictions, the contributing factors include the choice of renormalization and factorization scales (scale), the PDF uncertainty (PDF), the uncertainty from αS, and the uncertainty due to the NP corrections (NP correction). In the matrix element computation, the coupling (αS for QCD) is evaluated at an energy scale known as the renormalization scale (µR). Another scale is chosen to compute the PDF, in order to resum initial-state radiation below that scale, called the factorization scale (µF). For the ﬁxed-order calculations, both are set equal to the pT of individual jets. The scale uncertainty is evaluated using the following combinations of factors for (µR, µF): (2, 1), (1, 2), (0.5, 1), (1, 0.5), (2, 2), (0.5, 0.5). The envelope of the variations is the scale uncertainty in the prediction. Scale variations in the numerator and the denominator of the ratio of cross sections are fully correlated since the underlying parton conﬁguration is the same. This is one of the largest sources of theoretical uncertainties. The PDFs are determined using data from several experiments. The PDFs there- fore have uncertainties from the experimental measurements, modeling, and parameteri- zation assumptions. The resulting uncertainty is calculated according to the prescription of CT14 [64] at the 90% conﬁdence level and then scaled to the 68.3% conﬁdence level. The PDF uncertainty is independent of jet size within statistical uncertainties, and thus cancels in the ratios. The cross section measurement for inclusive jets depends on the value of αS. In the nlojet++ prediction, its value (0.118) is varied by ±0.001. The uncertainty is taken as the diﬀerence between the results with varied and nominal values of αS and this diﬀerence is scaled to correspond to ∆αS ≃0.0015, as recommended in ref. [65]. For the jet cross section ratio, the uncertainty due to the αS variation in the numerator and denominator cancels. As mentioned in section 5.2, the envelope of the diﬀerences between the NP correction factors obtained using diﬀerent parton showering algorithms to determine the NP correction is the uncertainty in the NP correction. The uncertainty is signiﬁcant only at low pT. – 13 – (GeV) T Jet p 100 1000 dy w.r.t. 7 Theoretical uncertainties AK4 jets T / dp σ 2 Ratio of d 0 2 4 6 8 (13 TeV) -1 < 35.9 fb CMS R = 1.2 (+2.1) R = 1.1 (+1.75) R = 1.0 (+1.5) R = 0.9 (+1.25) R = 0.8 (+1.0) R = 0.7 (+0.75) R = 0.6 (+0.5) R = 0.5 (+0.25) R = 0.4 R = 0.3 (-0.25) R = 0.2 (-0.5) R = 0.1 (-0.75) Exp. sys. \|y\| < 0.5 Symbol : Data PH+P8(CUETP8M1) Figure 4. Comparison of the ratio of the diﬀerential cross sections of jets of diﬀerent sizes with respect to that of AK4 jets from data and from NLO predictions using powheg+pythia (CUETP8M1 tune) in the region \|y\| < 0.5. Colored symbols indicate data and colored lines repre- sent prediction from simulation. Oﬀsets by the amount written in the parentheses have been added to the corresponding data points to separate the results for diﬀerent jet sizes. (GeV) T Jet p 100 1000 dy w.r.t. AK4 jets T / dp σ 2 Ratio of d 0 2 4 6 8 (13 TeV) -1 < 35.9 fb CMS R = 1.2 (+2.1) R = 1.1 (+1.75) R = 1.0 (+1.5) R = 0.9 (+1.25) R = 0.8 (+1.0) R = 0.7 (+0.75) R = 0.6 (+0.5) R = 0.5 (+0.25) R = 0.4 R = 0.3 (-0.25) R = 0.2 (-0.5) R = 0.1 (-0.75) Exp. sys. \|y\| < 0.5 Symbol : Data PH+P8(CUETP8M1) JHEP12(2020)082 Figure 4. Comparison of the ratio of the diﬀerential cross sections of jets of diﬀerent sizes with respect to that of AK4 jets from data and from NLO predictions using powheg+pythia (CUETP8M1 tune) in the region \|y\| < 0.5. Colored symbols indicate data and colored lines repre- sent prediction from simulation. Oﬀsets by the amount written in the parentheses have been added to the corresponding data points to separate the results for diﬀerent jet sizes. All these uncertainties are added in quadrature, and are collectively referred as the theoretical uncertainty in what follows. The correlation between the experimental and the theoretical uncertainties is not studied. 8.1 Comparison of ratio of cross sections The ratios of cross sections with respect to the AK4 jets are shown in ﬁgure 4 in the central region (\|y\| < 0.5) for all the jet sizes using unfolded data and the prediction from the NLO MC generator powheg with pythia parton showering; they are oﬀset by ﬁxed quantities for clarity. The NLO powheg generator, interfaced with the parton showering model, describes the data well at moderate values of jet size, but there is a deviation at low pT for very large values of jet size. The ratios of the cross sections of inclusive AK2 and AK8 jets with respect to those of AK4 jets are computed at LO and NLO in pQCD, following eq. (1.5), with nlojet++ for the most central region (\|y\| < 0.5). The comparison with data is shown in ﬁgure 5. Both the LO and NLO predictions are systematically below data for AK8 jets and above data for AK2 jets. The NP correction is essential to describe the trend in data below medium jet pT – 14 – (GeV) T Jet p 100 1000 dy w.r.t. AK4 jets T / dp σ 2 Ratio of d 0.6 0.8 1 (13 TeV) -1 < 35.9 fb CMS Data LO NP ⊗ LO NLO NP ⊗ NLO NP ⊗ (NLO+NLL) Exp. sys. Scale+NP unc. T Anti-k R = 0.2 \|y\| < 0.5 Exp. sys. Scale+NP unc. (GeV) T Jet p 100 1000 dy w.r.t. AK4 jets T / dp σ 2 Ratio of d 1 1.5 2 (13 TeV) -1 < 35.9 fb CMS Data LO NP ⊗ LO NLO NP ⊗ NLO NP ⊗ (NLO+NLL) Exp. sys. Scale+NP unc. T Anti-k R = 0.8 \|y\| < 0.5 Exp. sys. Scale+NP unc. gure 5. Comparison of the ratios of diﬀerential cross sections for the AK2 (upper) and AK wer) jets with respect to that of AK4 jets from data and pQCD predictions using nlojet+ the region \|y\| < 0.5. Black symbols indicate data and colored lines represent pQCD predictio atistical uncertainties are shown as vertical bars for the data and the NLO⊗NP prediction. T lowish olive region around data represents the experimental systematic uncertainty whereas t gion shaded in light blue color around NLO⊗NP prediction shows the theoretical uncertainty e prediction. (GeV) T Jet p 100 1000 dy w.r.t. 9 Summary A measurement has been made of the ratio of cross sections of inclusive anti-kT jets of multiple sizes with respect to jets with the distance parameter R = 0.4; this is the ﬁrst such result from the CMS collaboration. Because of cancellation of many experimental and theoretical systematic uncertainties for the ratio, it is more sensitive to perturbative and nonperturbative eﬀects than the absolute cross section measurement; the experimental sys- tematic uncertainty in the cross section ratio is of similar size as the statistical uncertainty, whereas the theoretical uncertainty is dominated by the choice of the renormalization and factorization scales. From the ratio measurement, we observe that the nonperturbative correction is impor- tant in describing the data at low transverse momentum. Thus, the modeling of nonper- turbative eﬀects, such as hadronization and the underlying event has a signiﬁcant impact on the description of the data in diﬀerent regions of phase space. Finally, the variation of the ratio of cross sections with jet size R emphasizes the importance of the inclusion of parton showering algorithms to capture the eﬀects of higher- order terms in the perturbation series by the resummation approach, which are absent in the case of ﬁxed-order computation. This is also demonstrated by the analytic calculations using joint resummation in threshold for single jet production, and jet size. Therefore, this study shows the importance of ﬁnal-state radiation modeled in Monte Carlo simulation to describe the data, and also implies that the diﬀerences between various parton showering and hadronization models are signiﬁcant. 8.1 Comparison of ratio of cross sections AK4 jets T / dp σ 2 Ratio of d 0.6 0.8 1 (13 TeV) -1 < 35.9 fb CMS Data LO NP ⊗ LO NLO NP ⊗ NLO NP ⊗ (NLO+NLL) Exp. sys. Scale+NP unc. T Anti-k R = 0.2 \|y\| < 0.5 Exp. sys. Scale+NP unc. JHEP12(2020)082 ( ) T p (GeV) T Jet p 100 1000 dy w.r.t. AK4 jets T / dp σ 2 Ratio of d 1 1.5 2 (13 TeV) -1 < 35.9 fb CMS Data LO NP ⊗ LO NLO NP ⊗ NLO NP ⊗ (NLO+NLL) Exp. sys. Scale+NP unc. T Anti-k R = 0.8 \|y\| < 0.5 Exp. sys. Scale+NP unc. Figure 5. Comparison of the ratios of diﬀerential cross sections for the AK2 (upper) and AK8 (lower) jets with respect to that of AK4 jets from data and pQCD predictions using nlojet++ in the region \|y\| < 0.5. Black symbols indicate data and colored lines represent pQCD predictions. Statistical uncertainties are shown as vertical bars for the data and the NLO⊗NP prediction. The yellowish olive region around data represents the experimental systematic uncertainty whereas the region shaded in light blue color around NLO⊗NP prediction shows the theoretical uncertainty in the prediction. values. Also, the NLO calculation improves data-theory agreement signiﬁcantly over LO, bringing data and theoretical prediction into agreement within statistical and systematic uncertainties at pT > 1000 GeV for both AK2 and AK8 jets. Resummed calculations bring the theoretical prediction even closer to the data, especially for AK8 jets. The uncertainty – 15 – corresponding to the resummed calculations is within 5% for cross section ratio, and is not shown here to avoid congestion in the ﬁgure. 8.2 Variation of the ratio of cross sections with jet size The cross section is determined as a function of pT for both data and theoretical predic- tions. The numbers are then divided by the cross section for the AK4 jets in the same pT and rapidity window separately for data and each theoretical prediction, and presented in ﬁgure 6, in three ranges of pT for the most central (\|y\| < 0.5) and the most forward (1.5 < \|y\| < 2.0) regions as a function of jet size. Almost all the MC simulations involving resummation via parton shower can describe the trend with jet size seen in data, whereas the LO calculation exhibits diﬀerent behavior. Prediction from NLO calculation, as shown in \|y\| < 0.5, improves signiﬁcantly the description of cross section ratio, as observed in data, for small jet sizes, and lies between the LO prediction and data for large jet sizes. Analytic calculations with joint resummation, available for jet sizes up to 0.8, provide an advancement with respect to ﬁxed-order predictions, and lead to a better agreement with data. Similar behavior is observed in all the rapidity regions reported. JHEP12(2020)082 Acknowledgments \|y\| < 0.5 < 638 GeV T 548 < p CMS (13 TeV) -1 < 35.9 fb Jet size 0.2 0.4 0.6 0.8 1 1.2 Ratio to data 0.8 1 1.2 Figure 6 Comparison of the ratio of cross s JHEP12(2020)082 Jet size 0.2 0.4 0.6 0.8 1 1.2 R /dy of AK4 jets) σ /dy) / (d σ (d 0.5 1 1.5 2 Data Pythia(CUETP8M1) Madgraph+P8(CUETP8M1) Herwig++(CUETHppS1) PH+P8(CUETP8M1) PH+Herwig(EE5C) Herwig7 NP ⊗ LO NP ⊗ (NLO+NLL) Exp. sys. Theo. unc. 1.5 < \|y\| < 2.0 < 468 GeV T 395 < p CMS (13 TeV) -1 < 35.9 fb Jet size 0.2 0.4 0.6 0.8 1 1.2 Ratio to data 0.8 1 1.2 /dy of AK4 jets) σ /dy) / (d σ (d 0.5 1 1.5 2 Data Pythia(CUETP8M1) Madgraph+P8(CUETP8M1) Herwig++(CUETHppS1) PH+P8(CUETP8M1) PH+Herwig(EE5C) Herwig7 NP ⊗ LO NP ⊗ (NLO+NLL) Exp. sys. Theo. unc. 1.5 < \|y\| < 2.0 < 638 GeV T 548 < p CMS (13 TeV) -1 < 35.9 fb Jet size 0.2 0.4 0.6 0.8 1 1.2 Ratio to data 0.8 1 1.2 /dy of AK4 jets) σ /dy) / (d σ (d 0.5 1 1.5 2 Data Pythia(CUETP8M1) Madgraph+P8(CUETP8M1) Herwig++(CUETHppS1) PH+P8(CUETP8M1) PH+Herwig(EE5C) Herwig7 NP ⊗ LO NP ⊗ NLO NP ⊗ (NLO+NLL) Exp. sys. Theo. unc. \|y\| < 0.5 < 468 GeV T 395 < p CMS (13 TeV) -1 < 35.9 fb Jet size 0.2 0.4 0.6 0.8 1 1.2 Ratio to data 0.8 1 1.2 /dy of AK4 jets) σ /dy) / (d σ (d 1 Jet size Jet size /dy of AK4 jets) σ /dy) / (d σ (d 0.5 1 1.5 2 Data Pythia(CUETP8M1) Madgraph+P8(CUETP8M1) Herwig++(CUETHppS1) PH+P8(CUETP8M1) PH+Herwig(EE5C) Herwig7 NP ⊗ LO NP ⊗ (NLO+NLL) Exp. sys. Theo. unc. 1.5 < \|y\| < 2.0 < 638 GeV T 548 < p CMS (13 TeV) -1 < 35.9 fb Jet size 0.2 0.4 0.6 0.8 1 1.2 Ratio to data 0.8 1 1.2 Jet s e /dy of AK4 jets) σ /dy) / (d σ (d 0.5 1 1.5 2 Data Pythia(CUETP8M1) Madgraph+P8(CUETP8M1) Herwig++(CUETHppS1) PH+P8(CUETP8M1) PH+Herwig(EE5C) Herwig7 NP ⊗ LO NP ⊗ NLO NP ⊗ (NLO+NLL) Exp. sys. Theo. unc. \|y\| < 0.5 < 638 GeV T 548 < p CMS (13 TeV) -1 < 35.9 fb Jet size 0.2 0.4 0.6 0.8 1 1.2 Ratio to data 0.8 1 1.2 Jet size Figure 6. Acknowledgments \|y\| < 0.5 < 272 GeV T 196 < p CMS (13 TeV) -1 < 35.9 fb Jet size 0.2 0.4 0.6 0.8 1 1.2 Ratio to data 0.8 1 1.2 /dy of AK4 jets) σ /dy) / (d σ (d 0.5 1 1.5 2 Data Pythia(CUETP8M1) Madgraph+P8(CUETP8M1) Herwig++(CUETHppS1) PH+P8(CUETP8M1) PH+Herwig(EE5C) Herwig7 NP ⊗ LO NP ⊗ NLO NP ⊗ (NLO+NLL) Exp. sys. Theo. unc. \|y\| < 0.5 < 468 GeV T 395 < p CMS (13 TeV) -1 < 35.9 fb J t i 0.2 0.4 0.6 0.8 1 1.2 Ratio to data 0.8 1 1.2 /dy of AK4 jets) σ /dy) / (d σ (d 0.5 1 1.5 2 Data Pythia(CUETP8M1) Madgraph+P8(CUETP8M1) Herwig++(CUETHppS1) PH+P8(CUETP8M1) PH+Herwig(EE5C) Herwig7 NP ⊗ LO NP ⊗ NLO NP ⊗ (NLO+NLL) Exp. sys. Theo. unc. \|y\| < 0.5 < 272 GeV T 196 < p CMS (13 TeV) -1 < 35.9 fb Jet size 0.2 0.4 0.6 0.8 1 1.2 Ratio to data 0.8 1 1.2 /dy of AK4 jets) σ /dy) / (d σ (d 0.5 1 1.5 2 Data Pythia(CUETP8M1) Madgraph+P8(CUETP8M1) Herwig++(CUETHppS1) PH+P8(CUETP8M1) PH+Herwig(EE5C) Herwig7 NP ⊗ LO NP ⊗ (NLO+NLL) Exp. sys. Theo. unc. 1.5 < \|y\| < 2.0 < 272 GeV T 196 < p CMS (13 TeV) -1 < 35.9 fb Jet size 0.2 0.4 0.6 0.8 1 1.2 Ratio to data 0.8 1 1.2 /dy of AK4 jets) σ /dy) / (d σ (d 0.5 1 1.5 2 Data Pythia(CUETP8M1) Madgraph+P8(CUETP8M1) Herwig++(CUETHppS1) PH+P8(CUETP8M1) PH+Herwig(EE5C) Herwig7 NP ⊗ LO NP ⊗ NLO NP ⊗ (NLO+NLL) Exp. sys. Theo. unc. \|y\| < 0.5 < 468 GeV T 395 < p CMS (13 TeV) -1 < 35.9 fb Jet size 0.2 0.4 0.6 0.8 1 1.2 Ratio to data 0.8 1 1.2 /dy of AK4 jets) σ /dy) / (d σ (d 0.5 1 1.5 2 Data Pythia(CUETP8M1) Madgraph+P8(CUETP8M1) Herwig++(CUETHppS1) PH+P8(CUETP8M1) PH+Herwig(EE5C) Herwig7 NP ⊗ LO NP ⊗ (NLO+NLL) Exp. sys. Theo. unc. 1.5 < \|y\| < 2.0 < 468 GeV T 395 < p CMS (13 TeV) -1 < 35.9 fb Jet size 0.2 0.4 0.6 0.8 1 1.2 Ratio to data 0.8 1 1.2 /dy of AK4 jets) σ /dy) / (d σ (d 0.5 1 1.5 2 Data Pythia(CUETP8M1) Madgraph+P8(CUETP8M1) Herwig++(CUETHppS1) PH+P8(CUETP8M1) PH+Herwig(EE5C) Herwig7 NP ⊗ LO NP ⊗ NLO NP ⊗ (NLO+NLL) Exp. sys. Theo. unc. Acknowledgments We congratulate our colleagues in the CERN accelerator departments for the excellent per- formance of the LHC and thank the technical and administrative staﬀs at CERN and at – 16 – /dy of AK4 jets) σ /dy) / (d σ (d 0.5 1 1.5 2 Data Pythia(CUETP8M1) Madgraph+P8(CUETP8M1) Herwig++(CUETHppS1) PH+P8(CUETP8M1) PH+Herwig(EE5C) Herwig7 NP ⊗ LO NP ⊗ (NLO+NLL) Exp. sys. Theo. unc. 1.5 < \|y\| < 2.0 < 272 GeV T 196 < p CMS (13 TeV) -1 < 35.9 fb Jet size 0.2 0.4 0.6 0.8 1 1.2 Ratio to data 0.8 1 1.2 /dy of AK4 jets) σ /dy) / (d σ (d 0.5 1 1.5 2 Data Pythia(CUETP8M1) Madgraph+P8(CUETP8M1) Herwig++(CUETHppS1) PH+P8(CUETP8M1) PH+Herwig(EE5C) Herwig7 NP ⊗ LO NP ⊗ (NLO+NLL) Exp. sys. Theo. unc. 1.5 < \|y\| < 2.0 < 468 GeV T 395 < p CMS (13 TeV) -1 < 35.9 fb Jet size 0.2 0.4 0.6 0.8 1 1.2 Ratio to data 0.8 1 1.2 /dy of AK4 jets) σ /dy) / (d σ (d 0.5 1 1.5 2 Data Pythia(CUETP8M1) Madgraph+P8(CUETP8M1) Herwig++(CUETHppS1) PH+P8(CUETP8M1) PH+Herwig(EE5C) Herwig7 NP ⊗ LO NP ⊗ (NLO+NLL) Exp. sys. Theo. unc. 1.5 < \|y\| < 2.0 < 638 GeV T 548 < p CMS (13 TeV) -1 < 35.9 fb Jet size 0.2 0.4 0.6 0.8 1 1.2 Ratio to data 0.8 1 1.2 /dy of AK4 jets) σ /dy) / (d σ (d 0.5 1 1.5 2 Data Pythia(CUETP8M1) Madgraph+P8(CUETP8M1) Herwig++(CUETHppS1) PH+P8(CUETP8M1) PH+Herwig(EE5C) Herwig7 NP ⊗ LO NP ⊗ (NLO+NLL) Exp. sys. Theo. unc. 1.5 < \|y\| < 2.0 < 272 GeV T 196 < p CMS (13 TeV) -1 < 35.9 fb Jet size 0.2 0.4 0.6 0.8 1 1.2 Ratio to data 0.8 1 1.2 /dy of AK4 jets) σ /dy) / (d σ (d 0.5 1 1.5 2 Data Pythia(CUETP8M1) Madgraph+P8(CUETP8M1) Herwig++(CUETHppS1) PH+P8(CUETP8M1) PH+Herwig(EE5C) Herwig7 NP ⊗ LO NP ⊗ (NLO+NLL) Exp. sys. Theo. unc. 1.5 < \|y\| < 2.0 < 468 GeV T 395 < p CMS (13 TeV) -1 < 35.9 fb 0.2 0.4 0.6 0.8 1 1.2 Ratio to data 0.8 1 1.2 /dy of AK4 jets) σ /dy) / (d σ (d 0.5 1 1.5 2 Data Pythia(CUETP8M1) Madgraph+P8(CUETP8M1) Herwig++(CUETHppS1) PH+P8(CUETP8M1) PH+Herwig(EE5C) Herwig7 NP ⊗ LO NP ⊗ NLO NP ⊗ (NLO+NLL) Exp. sys. Theo. unc. Acknowledgments \|y\| < 0.5 < 638 GeV T 548 < p CMS (13 TeV) -1 < 35.9 fb Jet size 0.2 0.4 0.6 0.8 1 1.2 Ratio to data 0.8 1 1.2 /dy of AK4 jets) σ /dy) / (d σ (d 0.5 1 1.5 2 Data Pythia(CUETP8M1) Madgraph+P8(CUETP8M1) Herwig++(CUETHppS1) PH+P8(CUETP8M1) PH+Herwig(EE5C) Herwig7 NP ⊗ LO NP ⊗ (NLO+NLL) Exp. sys. Theo. unc. 1.5 < \|y\| < 2.0 < 638 GeV T 548 < p CMS (13 TeV) -1 < 35.9 fb Jet size 0.2 0.4 0.6 0.8 1 1.2 Ratio to data 0.8 1 1.2 Figure 6. Comparison of the ratio of cross sections of inclusive jets of various sizes with respect to AK4 jets, as a function of jet size in diﬀerent regions of jet pT in data, and for multiple theoretical predictions in rapidity bins \|y\| < 0.5 (left column) and 1.5 < \|y\| < 2.0 (right column) at particle level. When the dijet production cross section ratio is presented using pure NLO predictions for two jet sizes, the ratio becomes LO at αS; this is quoted as LO⊗NP in the ﬁgure. Points corresponding to a particular prediction are connected via lines to guide the eye. Experimental uncertainties in the ratio of cross sections are shown with bands around the data points, whereas theoretical uncertainties are shown with the bands around the ﬁxed-order predictions. /dy of AK4 jets) σ /dy) / (d σ (d 0.5 1 1.5 2 Data Pythia(CUETP8M1) Madgraph+P8(CUETP8M1) Herwig++(CUETHppS1) PH+P8(CUETP8M1) PH+Herwig(EE5C) Herwig7 NP ⊗ LO NP ⊗ NLO NP ⊗ (NLO+NLL) Exp. sys. Theo. unc. \|y\| < 0.5 < 272 GeV T 196 < p Jet size 0.2 0.4 0.6 0.8 1 1.2 Ratio to data 0.8 1 1.2 /dy of AK4 jets) σ /dy) / (d σ (d 0.5 1 1.5 2 Data Pythia(CUETP8M1) Madgraph+P8(CUETP8M1) Herwig++(CUETHppS1) PH+P8(CUETP8M1) PH+Herwig(EE5C) Herwig7 NP ⊗ LO NP ⊗ NLO NP ⊗ (NLO+NLL) Exp. sys. Theo. unc. \|y\| < 0.5 < 468 GeV T 395 < p CMS (13 TeV) -1 < 35.9 fb Jet size 0.2 0.4 0.6 0.8 1 1.2 Ratio to data 0.8 1 1.2 /dy of AK4 jets) σ /dy) / (d σ (d 0.5 1 1.5 2 Data Pythia(CUETP8M1) Madgraph+P8(CUETP8M1) Herwig++(CUETHppS1) PH+P8(CUETP8M1) PH+Herwig(EE5C) Herwig7 NP ⊗ LO NP ⊗ NLO NP ⊗ (NLO+NLL) Exp. sys. Theo. unc. Acknowledgments Comparison of the ratio of cross sections of inclusive jets of various sizes with respect to AK4 jets, as a function of jet size in diﬀerent regions of jet pT in data, and for multiple theoretical predictions in rapidity bins \|y\| < 0.5 (left column) and 1.5 < \|y\| < 2.0 (right column) at particle level. When the dijet production cross section ratio is presented using pure NLO predictions for two jet sizes, the ratio becomes LO at αS; this is quoted as LO⊗NP in the ﬁgure. Points corresponding to a particular prediction are connected via lines to guide the eye. Experimental uncertainties in the ratio of cross sections are shown with bands around the data points, whereas theoretical uncertainties are shown with the bands around the ﬁxed-order predictions. – 17 – other CMS institutes for their contributions to the success of the CMS eﬀort. In addition, we gratefully acknowledge the computing centers and personnel of the Worldwide LHC Computing Grid for delivering so eﬀectively the computing infrastructure essential to our analyses. Finally, we acknowledge the enduring support for the construction and operation of the LHC and the CMS detector provided by the following funding agencies: BMBWF and FWF (Austria); FNRS and FWO (Belgium); CNPq, CAPES, FAPERJ, FAPERGS, and FAPESP (Brazil); MES (Bulgaria); CERN; CAS, MoST, and NSFC (China); COL- CIENCIAS (Colombia); MSES and CSF (Croatia); RPF (Cyprus); SENESCYT (Ecuador); MoER, ERC IUT, PUT and ERDF (Estonia); Academy of Finland, MEC, and HIP (Fin- land); CEA and CNRS/IN2P3 (France); BMBF, DFG, and HGF (Germany); GSRT (Greece); NKFIA (Hungary); DAE and DST (India); IPM (Iran); SFI (Ireland); INFN (Italy); MSIP and NRF (Republic of Korea); MES (Latvia); LAS (Lithuania); MOE and UM (Malaysia); BUAP, CINVESTAV, CONACYT, LNS, SEP, and UASLP-FAI (Mexico); MOS (Montenegro); MBIE (New Zealand); PAEC (Pakistan); MSHE and NSC (Poland); FCT (Portugal); JINR (Dubna); MON, RosAtom, RAS, RFBR, and NRC KI (Russia); MESTD (Serbia); SEIDI, CPAN, PCTI, and FEDER (Spain); MOSTR (Sri Lanka); Swiss Funding Agencies (Switzerland); MST (Taipei); ThEPCenter, IPST, STAR, and NSTDA (Thailand); TUBITAK and TAEK (Turkey); NASU (Ukraine); STFC (United Kingdom); DOE and NSF (U.S.A.). JHEP12(2020)082 Individuals have received support from the Marie-Curie program and the European Research Council and Horizon 2020 Grant, contract Nos. 675440, 752730, and 765710 (Eu- ropean Union); the Leventis Foundation; the A.P. Acknowledgments Sloan Foundation; the Alexander von Humboldt Foundation; the Belgian Federal Science Policy Oﬃce; the Fonds pour la Forma- tion à la Recherche dans l’Industrie et dans l’Agriculture (FRIA-Belgium); the Agentschap voor Innovatie door Wetenschap en Technologie (IWT-Belgium); the F.R.S.-FNRS and FWO (Belgium) under the “Excellence of Science — EOS” — be.h project n. 30820817; the Beijing Municipal Science & Technology Commission, No. Z191100007219010; the Ministry of Education, Youth and Sports (MEYS) of the Czech Republic; the Deutsche Forschungsgemeinschaft (DFG) under Germany’s Excellence Strategy — EXC 2121 “Quan- tum Universe” — 390833306; the Lendület (“Momentum”) Program and the János Bolyai Research Scholarship of the Hungarian Academy of Sciences, the New National Excellence Program ÚNKP, the NKFIA research grants 123842, 123959, 124845, 124850, 125105, 128713, 128786, and 129058 (Hungary); the Council of Science and Industrial Research, India; the HOMING PLUS program of the Foundation for Polish Science, coﬁnanced from European Union, Regional Development Fund, the Mobility Plus program of the Min- istry of Science and Higher Education, the National Science Center (Poland), contracts Harmonia 2014/14/M/ST2/00428, Opus 2014/13/B/ST2/02543, 2014/15/B/ST2/03998, and 2015/19/B/ST2/02861, Sonata-bis 2012/07/E/ST2/01406; the National Priorities Re- search Program by Qatar National Research Fund; the Ministry of Science and Education, grant no. 14.W03.31.0026 (Russia); the Tomsk Polytechnic University Competitiveness En- hancement Program and “Nauka” Project FSWW-2020-0008 (Russia); the Programa Es- tatal de Fomento de la Investigación Cientíﬁca y Técnica de Excelencia María de Maeztu, grant MDM-2015-0509 and the Programa Severo Ochoa del Principado de Asturias; the – 18 – Thalis and Aristeia programs coﬁnanced by EU-ESF and the Greek NSRF; the Rachada- pisek Sompot Fund for Postdoctoral Fellowship, Chulalongkorn University and the Chu- lalongkorn Academic into Its 2nd Century Project Advancement Project (Thailand); the Kavli Foundation; the Nvidia Corporation; the SuperMicro Corporation; the Welch Foun- dation, contract C-1845; and the Weston Havens Foundation (U.S.A.). Thalis and Aristeia programs coﬁnanced by EU-ESF and the Greek NSRF; the Rachada- pisek Sompot Fund for Postdoctoral Fellowship, Chulalongkorn University and the Chu- lalongkorn Academic into Its 2nd Century Project Advancement Project (Thailand); the Kavli Foundation; the Nvidia Corporation; the SuperMicro Corporation; the Welch Foun- dation, contract C-1845; and the Weston Havens Foundation (U.S.A.). Open Access. This article is distributed under the terms of the Creative Commons Attribution License (CC-BY 4.0), which permits any use, distribution and reproduction in any medium, provided the original author(s) and source are credited. Open Access. Acknowledgments This article is distributed under the terms of the Creative Commons Attribution License (CC-BY 4.0), which permits any use, distribution and reproduction in any medium, provided the original author(s) and source are credited. JHEP12(2020)082 References [1] G.F. Sterman and S. Weinberg, Jets from Quantum Chromodynamics, Phys. Rev. Lett. 39 (1977) 1436 [INSPIRE]. [2] T. Sjöstrand, A Model for Initial State Parton Showers, Phys. Lett. B 157 (1985) 321 [INSPIRE]. [3] T.D. Gottschalk, Backwards evolved initial state parton showers, Nucl. Phys. B 277 (1986) 700 [INSPIRE]. [4] G. Marchesini and B.R. Webber, Simulation of QCD Jets Including Soft Gluon Interference, Nucl. Phys. B 238 (1984) 1 [INSPIRE]. [5] Y.L. Dokshitzer and B.R. Webber, Calculation of power corrections to hadronic event shapes, Phys. Lett. B 352 (1995) 451 [hep-ph/9504219] [INSPIRE]. 6] CDF collaboration, Charged Jet Evolution and the Underlying Event in p¯p Collisions at 1.8 TeV, Phys. Rev. D 65 (2002) 092002 [INSPIRE]. [7] N. Kidonakis, G. Oderda and G.F. Sterman, Threshold resummation for dijet cross-sections, Nucl. Phys. B 525 (1998) 299 [hep-ph/9801268] [INSPIRE]. [8] M.H. Seymour, Jet shapes in hadron collisions: Higher orders, resummation and hadronization, Nucl. Phys. B 513 (1998) 269 [hep-ph/9707338] [INSPIRE]. [9] X. Liu, S.-O. Moch and F. Ringer, Threshold and jet radius joint resummation for single-inclusive jet production, Phys. Rev. Lett. 119 (2017) 212001 [arXiv:1708.04641] [INSPIRE]. [10] M. Cacciari, G.P. Salam and G. Soyez, The anti-kt jet clustering algorithm, JHEP 04 (2008) 063 [arXiv:0802.1189] [INSPIRE]. [11] CMS collaboration, The CMS Experiment at the CERN LHC, 2008 JINST 3 S08004 [INSPIRE]. [12] CMS collaboration, Measurement of the Inclusive Jet Cross Section in pp Collisions at √s = 7 TeV, Phys. Rev. Lett. 107 (2011) 132001 [arXiv:1106.0208] [INSPIRE]. [13] CMS collaboration, Measurement and QCD analysis of double-diﬀerential inclusive jet cross sections in pp collisions at √s = 8 TeV and cross section ratios to 2.76 and 7 TeV, JHEP 03 (2017) 156 [arXiv:1609.05331] [INSPIRE]. [14] CMS collaboration, Measurement of the double-diﬀerential inclusive jet cross section in proton-proton collisions at √s = 13 TeV, Eur. Phys. J. C 76 (2016) 451 [arXiv:1605.04436] [INSPIRE]. – 19 – [15] Z. Nagy, Next-to-leading order calculation of three jet observables in hadron hadron collision, Phys. Rev. D 68 (2003) 094002 [hep-ph/0307268] [INSPIRE]. [16] T. Kluge, K. Rabbertz and M. Wobisch, FastNLO: Fast pQCD calculations for PDF ﬁts, in proceedings of the 14th International Workshop on Deep Inelastic Scattering (DIS 2006), Tsukuba, Japan, 20–24 April 2006, pp. 483–486 [hep-ph/0609285] [INSPIRE]. [17] S. Frixione, P. Nason and C. Oleari, Matching NLO QCD computations with Parton Shower simulations: the POWHEG method, JHEP 11 (2007) 070 [arXiv:0709.2092] [INSPIRE]. [18] T. References Sjöstrand et al., An introduction to PYTHIA 8.2, Comput. Phys. Commun. 191 (2015) 159 [arXiv:1410.3012] [INSPIRE]. JHEP12(2020)082 JHEP12(2020)082 [19] M. Bähr et al., HERWIG++ Physics and Manual, Eur. Phys. J. C 58 (2008) 639 [arXiv:0803.0883] [INSPIRE]. [20] ATLAS collaboration, Measurement of the inclusive jet cross-sections in proton-proton collisions at √s = 8 TeV with the ATLAS detector, JHEP 09 (2017) 020 [arXiv:1706.03192] [INSPIRE]. [21] ATLAS collaboration, Measurement of inclusive jet and dijet cross-sections in proton-proton collisions at √s = 13 TeV with the ATLAS detector, JHEP 05 (2018) 195 [arXiv:1711.02692] [INSPIRE]. [22] ALICE collaboration, Measurement of the inclusive diﬀerential jet cross section in pp collisions at √s = 2.76 TeV, Phys. Lett. B 722 (2013) 262 [arXiv:1301.3475] [INSPIRE]. [23] CMS collaboration, Measurement of the Ratio of Inclusive Jet Cross Sections using the Anti-kT Algorithm with Radius Parameters R = 0.5 and 0.7 in pp Collisions at √s = 7 TeV, Phys. Rev. D 90 (2014) 072006 [arXiv:1406.0324] [INSPIRE]. [24] ALICE collaboration, Measurements of inclusive jet spectra in pp and central Pb-Pb collisions at √sNN = 5.02 TeV, Phys. Rev. C 101 (2020) 034911 [arXiv:1909.09718] [INSPIRE]. [25] J. Bellm et al., Jet Cross Sections at the LHC and the Quest for Higher Precision, Eur. Phys. J. C 80 (2020) 93 [arXiv:1903.12563] [INSPIRE]. [26] F. Aversa, P. Chiappetta, M. Greco and J.P. Guillet, QCD Corrections to Parton-Parton Scattering Processes, Nucl. Phys. B 327 (1989) 105 [INSPIRE]. [27] D. de Florian and W. Vogelsang, Resummed cross-section for jet production at hadron colliders, Phys. Rev. D 76 (2007) 074031 [arXiv:0704.1677] [INSPIRE]. [28] M. Dasgupta, L. Magnea and G.P. Salam, Non-perturbative QCD eﬀects in jets at hadron colliders, JHEP 02 (2008) 055 [arXiv:0712.3014] [INSPIRE]. [29] M. Beneke, Renormalons, Phys. Rept. 317 (1999) 1 [hep-ph/9807443] [INSPIRE]. [30] M. Cacciari, G.P. Salam and G. Soyez, The Catchment Area of Jets, JHEP 04 (2008) 005 [arXiv:0802.1188] [INSPIRE]. [31] Z. Nagy, Three jet cross-sections in hadron hadron collisions at next-to-leading order, Phys. Rev. Lett. 88 (2002) 122003 [hep-ph/0110315] [INSPIRE]. [32] G. Soyez, A Simple description of jet cross-section ratios, Phys. Lett. B 698 (2011) 59 [arXiv:1101.2665] [INSPIRE]. [32] G. Soyez, A Simple description of jet cross-section ratios, Phys. Lett. B 698 (2011) 59 [arXiv:1101.2665] [INSPIRE]. [33] CMS collaboration, Description and performance of track and primary-vertex reconstruction with the CMS tracker, 2014 JINST 9 P10009 [arXiv:1405.6569] [INSPIRE]. References [33] CMS collaboration, Description and performance of track and primary-vertex reconstruction with the CMS tracker, 2014 JINST 9 P10009 [arXiv:1405.6569] [INSPIRE]. – 20 – [34] CMS collaboration, Performance of Photon Reconstruction and Identiﬁcation with the CMS Detector in Proton-Proton Collisions at √s = 8 TeV, 2015 JINST 10 P08010 [arXiv:1502.02702] [INSPIRE]. [35] CMS collaboration, The CMS hadron calorimeter project: Technical Design Report, CERN-LHCC-97-031 (1997) [INSPIRE]. [36] CMS collaboration, Jet energy scale and resolution performances with 13 TeV data, CMS-DP-2016-020 (2016). [37] CMS collaboration, The CMS trigger system, 2017 JINST 12 P01020 [arXiv:1609.02366] [INSPIRE]. JHEP12(2020)082 [38] CMS collaboration, Particle-ﬂow reconstruction and global event description with the CMS detector, 2017 JINST 12 P10003 [arXiv:1706.04965] [INSPIRE]. [39] M. Cacciari, G.P. Salam and G. Soyez, FastJet User Manual, Eur. Phys. J. C 72 (2012) 1896 [arXiv:1111.6097] [INSPIRE]. [40] M. Cacciari and G.P. Salam, Pileup subtraction using jet areas, Phys. Lett. B 659 (2008) 119 [arXiv:0707.1378] [INSPIRE]. [41] CMS collaboration, Jet energy scale and resolution in the CMS experiment in pp collisions at 8 TeV, 2017 JINST 12 P02014 [arXiv:1607.03663] [INSPIRE]. [42] CMS collaboration, Jet algorithms performance in 13 TeV data, CMS-PAS-JME-16-003 (2017) [INSPIRE]. [43] CMS collaboration, Performance of missing transverse momentum reconstruction in proton-proton collisions at √s = 13 TeV using the CMS detector, 2019 JINST 14 P07004 [arXiv:1903.06078] [INSPIRE]. [44] CMS collaboration, Determination of Jet Energy Calibration and Transverse Momentum Resolution in CMS, 2011 JINST 6 P11002 [arXiv:1107.4277] [INSPIRE]. [45] NNPDF collaboration, A Determination of parton distributions with faithful uncertainty estimation, Nucl. Phys. B 809 (2009) 1 [Erratum ibid. 816 (2009) 293] [arXiv:0808.1231] [INSPIRE]. [46] CMS collaboration, Event generator tunes obtained from underlying event and multiparton scattering measurements, Eur. Phys. J. C 76 (2016) 155 [arXiv:1512.00815] [INSPIRE]. [47] P. Skands, S. Carrazza and J. Rojo, Tuning PYTHIA 8.1: the Monash 2013 Tune, Eur. Phys. J. C 74 (2014) 3024 [arXiv:1404.5630] [INSPIRE]. [48] J. Alwall et al., The automated computation of tree-level and next-to-leading order diﬀerential cross sections, and their matching to parton shower simulations, JHEP 07 (2014) 079 [arXiv:1405.0301] [INSPIRE]. [49] J. Alwall et al., Comparative study of various algorithms for the merging of parton showers and matrix elements in hadronic collisions, Eur. Phys. J. C 53 (2008) 473 [arXiv:0706.2569] [INSPIRE]. [50] S. Alioli, K. Hamilton, P. Nason, C. Oleari and E. Re, Jet pair production in POWHEG, JHEP 04 (2011) 081 [arXiv:1012.3380] [INSPIRE]. [51] M.H. Seymour and A. niversité Catholique de Louvain, Louvain-la-Neuve, Belgium O. Bondu, G. Bruno, C. Caputo, P. David, C. Delaere, M. Delcourt, A. Giammanco, V. Lemaitre, J. Prisciandaro, A. Saggio, M. Vidal Marono, P. Vischia, J. Zobec Centro Brasileiro de Pesquisas Fisicas, Rio de Janeiro, Brazil F.L. Alves, G.A. Alves, G. Correia Silva, C. Hensel, A. Moraes, P. Rebello Teles Universidade do Estado do Rio de Janeiro, Rio de Janeiro, Brazil References Siodmok, Constraining MPI models using σeﬀand recent Tevatron and LHC Underlying Event data, JHEP 10 (2013) 113 [arXiv:1307.5015] [INSPIRE]. – 21 – [52] J. Bellm et al., HERWIG 7.0/HERWIG++ 3.0 release note, Eur. Phys. J. C 76 (2016) 196 [arXiv:1512.01178] [INSPIRE]. [53] R. Frederix and S. Frixione, Merging meets matching in MC@NLO, JHEP 12 (2012) 061 [arXiv:1209.6215] [INSPIRE]. [54] X. Liu, S.-O. Moch and F. Ringer, Phenomenology of single-inclusive jet production with jet radius and threshold resummation, Phys. Rev. D 97 (2018) 056026 [arXiv:1801.07284] [INSPIRE]. [55] T. Adye, Unfolding algorithms and tests using RooUnfold, in proceedings of the PHYSTAT 2011, Geneva, Switzerland, 17–20 January 2011, pp. 313–318 [arXiv:1105.1160] [INSPIRE]. JHEP12(2020)082 [56] G. D’Agostini, A Multidimensional unfolding method based on Bayes’ theorem, Nucl. Instrum. Meth. A 362 (1995) 487 [INSPIRE]. JHEP12(2020)082 [57] A. Hocker and V. Kartvelishvili, SVD approach to data unfolding, Nucl. Instrum. Meth. A 372 (1996) 469 [hep-ph/9509307] [INSPIRE]. [58] S. Schmitt, Data Unfolding Methods in High Energy Physics, EPJ Web Conf. 137 (2017) 11008 [arXiv:1611.01927] [INSPIRE]. [59] S. Schmitt, TUnfold: an algorithm for correcting migration eﬀects in high energy physics, 2012 JINST 7 T10003 [arXiv:1205.6201] [INSPIRE]. [60] GEANT4 collaboration, GEANT4 — a simulation toolkit, Nucl. Instrum. Meth. A 506 (2003) 250 [INSPIRE]. [61] B. Efron and R.J. Tibshirani, An introduction to the bootstrap, in Monographs on Statistics and Applied Probability 57, Chapman & Hall/CRC, London U.K. (1993). [62] CMS collaboration, Jet energy scale and resolution performance with 13 TeV data collected by CMS in 2016, CMS-DP-2018-028 (2018). [63] CMS collaboration, Measurement of the inelastic proton-proton cross section at √s = 13 TeV, JHEP 07 (2018) 161 [arXiv:1802.02613] [INSPIRE]. [64] S. Dulat et al., New parton distribution functions from a global analysis of quantum chromodynamics, Phys. Rev. D 93 (2016) 033006 [arXiv:1506.07443] [INSPIRE]. [65] J. Butterworth et al., PDF4LHC recommendations for LHC Run II, J. Phys. G 43 (2016) 023001 [arXiv:1510.03865] [INSPIRE]. – 22 – Yerevan Physics Institute, Yerevan, Armenia A.M. Sirunyan†, A. Tumasyan Institut für Hochenergiephysik, Wien, Austria W. Adam, F. Ambrogi, T. Bergauer, J. Brandstetter, M. Dragicevic, J. Erö, A. Es- calante Del Valle, M. Flechl, R. Frühwirth1, M. Jeitler1, N. Krammer, I. Krätschmer, D. Liko, T. Madlener, I. Mikulec, N. Rad, J. Schieck1, R. Schöfbeck, M. Spanring, D. Spitzbart, W. Waltenberger, C.-E. Wulz1, M. Zarucki JHEP12(2020)082 Institute for Nuclear Problems, Minsk, Belarus V. Drugakov, V. Mossolov, J. Suarez Gonzalez Institute for Nuclear Problems, Minsk, Belarus Universiteit Antwerpen, Antwerpen, Belgium M.R. Darwish, E.A. De Wolf, D. Di Croce, X. Janssen, A. Lelek, M. Pieters, H. Rejeb Sfar, H. Van Haevermaet, P. Van Mechelen, S. Van Putte, N. Van Remortel Université Libre de Bruxelles, Bruxelles, Belgium D. Beghin, B. Bilin, H. Brun, B. Clerbaux, G. De Lentdecker, H. Delannoy, B. Dorney, L. Favart, A. Grebenyuk, A.K. Kalsi, A. Popov, N. Postiau, E. Starling, L. Thomas, C. Vander Velde, P. Vanlaer, D. Vannerom D. Beghin, B. Bilin, H. Brun, B. Clerbaux, G. De Lentdecker, H. Delannoy, B. Dorne L. Favart, A. Grebenyuk, A.K. Kalsi, A. Popov, N. Postiau, E. Starling, L. Thomas, C. Vander Velde, P. Vanlaer, D. Vannerom Ghent University, Ghent, Belgium T. Cornelis, D. Dobur, I. Khvastunov2, M. Niedziela, C. Roskas, D. Trocino, M. Tytgat, W. Verbeke, B. Vermassen, M. Vit Université Catholique de Louvain, Louvain-la-Neuve, Belgium The CMS collaboration The CMS collaboration Yerevan Physics Institute, Yerevan, Armenia A.M. Sirunyan†, A. Tumasyan Vrije Universiteit Brussel, Brussel, Belgium Vrije Universiteit Brussel, Brussel, Belgium F. Blekman, E.S. Bols, S.S. Chhibra, J. D’Hondt, J. De Clercq, D. Lontkovskyi, S. Lowette, F. Blekman, E.S. Bols, S.S. Chhibra, J. D’Hondt, J. De Clercq, D. Lontkovskyi, S. Lowette, I. Marchesini, S. Moortgat, Q. Python, K. Skovpen, S. Tavernier, W. Van Doninck, F. Blekman, E.S. Bols, S.S. Chhibra, J. D Hondt, J. De Clercq, D. Lontkovskyi, S. Lowette, I. Marchesini, S. Moortgat, Q. Python, K. Skovpen, S. Tavernier, W. Van Doninck, P. Van Mulders I. Marchesini, S. Moortgat, Q. Python, K. Skovpen, S. Tavernier, W. Van Doninc P Van Mulders I. Marchesini, S. Moortgat, Q. Python, K. Skovpen, S. Tavernier, W. Van Doninck, P. Van Mulders P. Van Mulders Université Libre de Bruxelles, Bruxelles, Belgium Universidade do Estado do Rio de Janeiro, Rio de Janeiro, Brazil E. Belchior Batista Das Chagas, W. Carvalho, J. Chinellato3, E. Coelho, E.M. Da Costa, G.G. Da Silveira4, D. De Jesus Damiao, C. De Oliveira Martins, S. Fonseca De Souza, L.M. Huertas Guativa, H. Malbouisson, J. Martins5, D. Matos Figueiredo, M. Med- ina Jaime6, M. Melo De Almeida, C. Mora Herrera, L. Mundim, H. Nogima, W.L. Prado Da Silva, L.J. Sanchez Rosas, A. Santoro, A. Sznajder, M. Thiel, E.J. Tonelli Manganote3, F. Torres Da Silva De Araujo, A. Vilela Pereira – 23 – Universidade Estadual Paulistaa, Universidade Federal do ABCb, São Paulo, Brazil C.A. Bernardesa, L. Calligarisa, T.R. Fernandez Perez Tomeia, E.M. Gregoresb, D.S. Lemos, P.G. Mercadanteb, S.F. Novaesa, SandraS. Padulaa Institute for Nuclear Research and Nuclear Energy, Bulgarian Academy of Sciences, Soﬁa, Bulgaria A. Aleksandrov, G. Antchev, R. Hadjiiska, P. Iaydjiev, M. Misheva, M. Rodozov, M. Shopova, G. Sultanov University of Soﬁa, Soﬁa, Bulgaria University of Soﬁa, Soﬁa, Bulgaria M. Bonchev, A. Dimitrov, T. Ivanov, L. Litov, B. Pavlov, P. Petkov JHEP12(2020)082 M. Bonchev, A. Dimitrov, T. Ivanov, L. Litov, B. Pavlov, P. Petkov Beihang University, Beijing, China W. Fang7, X. Gao7, L. Yuan Beihang University, Beijing, China W. Fang7, X. Gao7, L. Yuan Department of Physics, Tsinghua University, Beijing, China M. Ahmad, Z. Hu, Y. Wang Institute of High Energy Physics, Beijing, China Institute of High Energy Physics, Beijing, China G.M. Chen, H.S. Chen, M. Chen, C.H. Jiang, D. Leggat, H. Liao, Z. Liu, A. Spiezia, J. Tao, E. Yazgan, H. Zhang, S. Zhang8, J. Zhao State Key Laboratory of Nuclear Physics and Technology, Peking University, Beijing, China A. Agapitos, Y. Ban, G. Chen, A. Levin, J. Li, L. Li, Q. Li, Y. Mao, S.J. Qian, D. Wang, Q. Wang Zhejiang University, Hangzhou, China M. Xiao Zhejiang University, Hangzhou, China M. Xiao Universidad de Los Andes, Bogota, Colombia C. Avila, A. Cabrera, C. Florez, C.F. González Hernández, M.A. Segura Delgado Universidad de Antioquia, Medellin, Colombia J. Mejia Guisao, J.D. Ruiz Alvarez, C.A. Salazar González, N. Vanegas Arbelaez Universidad de Antioquia, Medellin, Colombia Universidad de Antioquia, Medellin, Colombia J. Mejia Guisao, J.D. Ruiz Alvarez, C.A. Salazar González, N. Vanegas Arbelaez University of Split, Faculty of Electrical Engineering, Mechanical Engineerin and Naval Architecture, Split, Croatia University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, Split, Croatia D. Giljanović, N. Godinovic, D. Lelas, I. Puljak, T. Sculac D. Giljanović, N. Godinovic, D. Lelas, I. Puljak, T. Sculac D. Universidade do Estado do Rio de Janeiro, Rio de Janeiro, Brazil Giljanović, N. Godinovic, D. Lelas, I. Puljak, T. Sculac University of Split, Faculty of Science, Split, Croatia Z. Antunovic, M. Kovac Institute Rudjer Boskovic, Zagreb, Croatia 9 Institute Rudjer Boskovic, Zagreb, Croatia Institute Rudjer Boskovic, Zagreb, Croatia V. Brigljevic, D. Ferencek, K. Kadija, B. Mesic, M. Roguljic, A. Starodumov9, T. Susa University of Cyprus, Nicosia, Cyprus M.W. Ather, A. Attikis, E. Erodotou, A. Ioannou, M. Kolosova, S. Konstantinou, G. Mavromanolakis, J. Mousa, C. Nicolaou, F. Ptochos, P.A. Razis, H. Rykaczewski, D. Tsiakkouri y yp , , yp M.W. Ather, A. Attikis, E. Erodotou, A. Ioannou, M. Kolosova, S. Konstantinou, G l k C l h A k k M.W. Ather, A. Attikis, E. Erodotou, A. Ioannou, M. Kolosova, S. Konstantinou, G. Mavromanolakis, J. Mousa, C. Nicolaou, F. Ptochos, P.A. Razis, H. Rykaczewski, D. Tsiakkouri M.W. Ather, A. Attikis, E. Erodotou, A. Ioannou, M. Kolosova, S. Konstantinou, G. Mavromanolakis, J. Mousa, C. Nicolaou, F. Ptochos, P.A. Razis, H. Rykaczewski, G. Mavromanolakis, J. Mousa, C. Nicolaou, F. Ptochos, P.A. Razis, H. Rykaczewski, D. Tsiakkouri , , , , , y , D. Tsiakkouri – 24 – Charles University, Prague, Czech Republic M. Finger10, M. Finger Jr.10, A. Kveton, J. Tomsa Escuela Politecnica Nacional, Quito, Ecuador E. Ayala Universidad San Francisco de Quito, Quito, Ecuador E. Carrera Jarrin Academy of Scientiﬁc Research and Technology of the Arab Republic of Egypt, Egyptian Network of High Energy Physics, Cairo, Egypt Y. Assran11,12, S. Elgammal12 JHEP12(2020)082 National Institute of Chemical Physics and Biophysics, Tallinn, Estonia S. Bhowmik, A. Carvalho Antunes De Oliveira, R.K. Dewanjee, K. Ehataht, M. Kadastik, M. Raidal, C. Veelken Department of Physics, University of Helsinki, Helsinki, Finland P. Eerola, L. Forthomme, H. Kirschenmann, K. Osterberg, M. Voutilainen Helsinki Institute of Physics, Helsinki, Finland Helsinki Institute of Physics, Helsinki, Finland y , , F. Garcia, J. Havukainen, J.K. Heikkilä, V. Karimäki, M.S. Kim, R. Kinnunen, T. Lampén, K. Lassila-Perini, S. Laurila, S. Lehti, T. Lindén, P. Luukka, T. Mäenpää, H. Siikonen, E. Tuominen, J. Tuominiemi Lappeenranta University of Technology, Lappeenranta, Finland T. Tuuva Deutsches Elektronen-Synchrotron, Hamburg, Germany Deutsches Elektronen-Synchrotron, Hamburg, Germany M. Aldaya Martin, P. Asmuss, I. Babounikau, H. Bakhshiansohi, K. Beernaert, O. Behnke, A. Bermúdez Martínez, D. Bertsche, A.A. Bin Anuar, K. Borras18, V. Botta, A. Campbell, A. Cardini, P. Connor, S. Consuegra Rodríguez, C. Contreras-Campana, V. Danilov, A. De Wit, M.M. Defranchis, C. Diez Pardos, D. Domínguez Damiani, G. Eckerlin, D. Eckstein, T. Eichhorn, A. Elwood, E. Eren, E. Gallo19, A. Geiser, A. Grohsjean, M. Guthoﬀ, M. Haranko, A. Harb, A. Jafari, N.Z. Jomhari, H. Jung, A. Kasem18, M. Kase- mann, H. Kaveh, J. Keaveney, C. Kleinwort, J. Knolle, D. Krücker, W. Lange, T. Lenz, J. Lidrych, K. Lipka, W. Lohmann20, R. Mankel, I.-A. Melzer-Pellmann, A.B. Meyer, M. Meyer, M. Missiroli, G. Mittag, J. Mnich, A. Mussgiller, V. Myronenko, D. Pérez Adán, S.K. Pﬂitsch, D. Pitzl, A. Raspereza, A. Saibel, M. Savitskyi, V. Scheurer, P. Schütze, C. Schwanenberger, R. Shevchenko, A. Singh, H. Tholen, O. Turkot, A. Vagnerini, M. Van De Klundert, R. Walsh, Y. Wen, K. Wichmann, C. Wissing, O. Zenaiev, R. Zlebcik . Aldaya Martin, P. Asmuss, I. Babounikau, H. Bakhshiansohi, K. Beernaert, O. Behnke, Bermúdez Martínez, D. Bertsche, A.A. Bin Anuar, K. Borras18, V. Botta, A. Campbell, Lappeenranta University of Technology, Lappeenranta, Finland T. Tuuva IRFU, CEA, Université Paris-Saclay, Gif-sur-Yvette, France IRFU, CEA, Université Paris-Saclay, Gif-sur-Yvette, France M. Besancon, F. Couderc, M. Dejardin, D. Denegri, B. Fabbro, J.L. Faure, F. Ferri, S. Ganjour, A. Givernaud, P. Gras, G. Hamel de Monchenault, P. Jarry, C. Leloup, E. Locci, J. Malcles, J. Rander, A. Rosowsky, M.Ö. Sahin, A. Savoy-Navarro13, M. Titov . Besancon, F. Couderc, M. Dejardin, D. Denegri, B. Fabbro, J.L. Faure, F. Ferri, Ganjour, A. Givernaud, P. Gras, G. Hamel de Monchenault, P. Jarry, C. Leloup, E. Locci, 13 Malcles, J. Rander, A. Rosowsky, M.Ö. Sahin, A. Savoy-Navarro13, M. Titov Laboratoire Leprince-Ringuet, CNRS/IN2P3, Ecole Polytechnique, Institut Polytechnique de Paris, France S. Ahuja, C. Amendola, F. Beaudette, P. Busson, C. Charlot, B. Diab, G. Falmagne, R. Granier de Cassagnac, I. Kucher, A. Lobanov, C. Martin Perez, M. Nguyen, C. Ochando, P. Paganini, J. Rembser, R. Salerno, J.B. Sauvan, Y. Sirois, A. Zabi, A. Zghiche S. Ahuja, C. Amendola, F. Beaudette, P. Busson, C. Charlot, B. Diab, G. Falmagne, R. Granier de Cassagnac, I. Kucher, A. Lobanov, C. Martin Perez, M. Nguyen, C. Ochando, P P i i J R b R S l J B S Y Si i A Z bi A Z hi h R. Granier de Cassagnac, I. Kucher, A. Lobanov, C. Martin Perez, M. Nguyen, C. Ochando, P. Paganini, J. Rembser, R. Salerno, J.B. Sauvan, Y. Sirois, A. Zabi, A. Zghiche P. Paganini, J. Rembser, R. Salerno, J.B. Sauvan, Y. Sirois, A. Zabi, A. Zghiche Université de Strasbourg, CNRS, IPHC UMR 7178, Strasbourg, France J.-L. Agram14, J. Andrea, D. Bloch, G. Bourgatte, J.-M. Brom, E.C. Chabert, C. Collard, E. Conte14, J.-C. Fontaine14, D. Gelé, U. Goerlach, M. Jansová, A.-C. Le Bihan, N. Tonon, P. Van Hove E. Conte14, J.-C. Fontaine14, D. Gelé, U. Goerlach, M. Jansová, A.-C. Le Bihan, N. Tonon, P. Van Hove E. Conte , J. C. Fontaine , D. Gelé, U. Goerlach, M. Jansová, A. C. Le Bihan, N. Tonon, P. Van Hove P. Van Hove Centre de Calcul de l’Institut National de Physique Nucleaire et de Physique des Particules, CNRS/IN2P3, Villeurbanne, France S. Gadrat – 25 – Université de Lyon, Université Claude Bernard Lyon 1, CNRS-IN2P3, Institut de Physique Nucléaire de Lyon, Villeurbanne, France Université de Lyon, Université Claude Bernard Lyon 1, CNRS-IN2P3, Institut de Physique Nucléaire de Lyon, Villeurbanne, France S. Beauceron, C. Bernet, G. Boudoul, C. Camen, A. Carle, N. Chanon, R. Chierici, D. Contardo, P. Lappeenranta University of Technology, Lappeenranta, Finland T. Tuuva Depasse, H. El Mamouni, J. Fay, S. Gascon, M. Gouzevitch, B. Ille, Sa. Jain, F. Lagarde, I.B. Laktineh, H. Lattaud, A. Lesauvage, M. Lethuillier, L. Mirabito, S. Perries, V. Sordini, L. Torterotot, G. Touquet, M. Vander Donckt, S. Viret S. Beauceron, C. Bernet, G. Boudoul, C. Camen, A. Carle, N. Chanon, R. Chierici, D. Contardo, P. Depasse, H. El Mamouni, J. Fay, S. Gascon, M. Gouzevitch, B. Ille, Sa. Jain, F. Lagarde, I.B. Laktineh, H. Lattaud, A. Lesauvage, M. Lethuillier, L. Mirabito, S. Perries, V. Sordini, L. Torterotot, G. Touquet, M. Vander Donckt, S. Viret Georgian Technical University, Tbilisi, Georgia T. Toriashvili15 Tbilisi State University, Tbilisi, Georgia Z. Tsamalaidze10 JHEP12(2020)082 RWTH Aachen University, I. Physikalisches Institut, Aachen, Germany RWTH Aachen University, I. Physikalisches Institut, Aachen, Germany C. Autermann, L. Feld, M.K. Kiesel, K. Klein, M. Lipinski, D. Meuser, A. Pauls, M. Preuten, M.P. Rauch, J. Schulz, M. Teroerde, B. Wittmer RWTH Aachen University, III. Physikalisches Institut A, Aachen, Germany RWTH Aachen University, III. Physikalisches Institut A, Aachen, Germany M. Erdmann, B. Fischer, S. Ghosh, T. Hebbeker, K. Hoepfner, H. Keller, L. Mastrolorenzo, M. Merschmeyer, A. Meyer, P. Millet, G. Mocellin, S. Mondal, S. Mukherjee, D. Noll, A. Novak, T. Pook, A. Pozdnyakov, T. Quast, M. Radziej, Y. Rath, H. Reithler, J. Roemer, A. Schmidt, S.C. Schuler, A. Sharma, S. Wiedenbeck, S. Zaleski RWTH Aachen University, III. Physikalisches Institut B, Aachen, Germany G. Flügge, W. Haj Ahmad16, O. Hlushchenko, T. Kress, T. Müller, A. Nowack, C. Pistone, O. Pooth, D. Roy, H. Sert, A. Stahl17 University of Hamburg, Hamburg, Germany R. Aggleton, S. Bein, L. Benato, A. Benecke, V. Blobel, T. Dreyer, A. Ebrahimi, F. Feindt, A. Fröhlich, C. Garbers, E. Garutti, D. Gonzalez, P. Gunnellini, J. Haller, A. Hinzmann, A. Karavdina, G. Kasieczka, R. Klanner, R. Kogler, N. Kovalchuk, S. Kurz, V. Kutzner, J. Lange, T. Lange, A. Malara, J. Multhaup, C.E.N. Niemeyer, A. Perieanu, A. Reimers, – 26 – O. Rieger, C. Scharf, P. Schleper, S. Schumann, J. Schwandt, J. Sonneveld, H. Stadie, G. Steinbrück, F.M. Stober, B. Vormwald, I. Zoi Rieger, C. Scharf, P. Schleper, S. Schumann, J. Schwandt, J. Sonneveld, H. Stadie, G. Steinbrück, F.M. Stober, B. Vormwald, I. Zoi Karlsruher Institut fuer Technologie, Karlsruhe, Germany M. Akbiyik, C. Barth, M. Baselga, S. Baur, T. Berger, E. Butz, R. Caspart, T. Chwalek, W. De Boer, A. Dierlamm, K. El Morabit, N. Faltermann, M. Giﬀels, P. Goldenzweig, A. Gottmann, M.A. Harrendorf, F. Hartmann17, U. Husemann, S. Kudella, S. Mitra, M.U. Mozer, D. Müller, Th. Müller, M. Musich, A. Nürnberg, G. Quast, K. Rabbertz, M. Schröder, I. Shvetsov, H.J. Simonis, R. Ulrich, M. Wassmer, M. Weber, C. Wöhrmann, R. Wolf JHEP12(2020)082 Institute of Nuclear and Particle Physics (INPP), NCSR Demokritos, Aghia Paraskevi, Greece G. Anagnostou, P. Asenov, G. Daskalakis, T. Geralis, A. Kyriakis, D. Loukas, G. Paspalaki National and Kapodistrian University of Athens, Athens, Greece M. Diamantopoulou, G. Karathanasis, P. Kontaxakis, A. Manousakis-katsikakis, A. Pana- giotou, I. Papavergou, N. Saoulidou, A. Stakia, K. Theoﬁlatos, K. Vellidis, E. Vourliotis National Technical University of Athens, Athens, Greece G. Bakas, K. Kousouris, I. Papakrivopoulos, G. Tsipolitis University of Ioánnina, Ioánnina, Greece I. Evangelou, C. Foudas, P. Gianneios, P. Katsoulis, P. Kokkas, S. Mallios, K. Manitara, N. Manthos, I. Papadopoulos, J. Strologas, F.A. Triantis, D. Tsitsonis University of Ioánnina, Ioánnina, Greece MTA-ELTE Lendület CMS Particle and Nuclear Physics Group, Eötvös Loránd University, Budapest, Hungary M. Bartók21, R. Chudasama, M. Csanad, P. Major, K. Mandal, A. Mehta, M.I. Nagy, G. Pasztor, O. Surányi, G.I. Veres Wigner Research Centre for Physics, Budapest, Hungary G. Bencze, C. Hajdu, D. Horvath22, F. Sikler, T.Á. Vámi, V. Veszpremi, G. Vesztergombi† Institute of Nuclear Research ATOMKI, Debrecen, Hungary N. Beni, S. Czellar, J. Karancsi21, A. Makovec, J. Molnar, Z. Szillasi Institute of Physics, University of Debrecen, Debrecen, Hungary P. Raics, D. Teyssier, Z.L. Trocsanyi, B. Ujvari Eszterhazy Karoly University, Karoly Robert Campus, Gyongyos, Hungary T. Csorgo, W.J. Metzger, F. Nemes, T. S. Bahinipati24, C. Kar, G. Kole, P. Mal, V.K. Muraleedharan Nair Bindhu, A. Nayak2 D.K. Sahoo24, S.K. Swain University of Hamburg, Hamburg, Germany Novak Indian Institute of Science (IISc), Bangalore, India S. Choudhury, J.R. Komaragiri, P.C. Tiwari National Institute of Science Education and Research, HBNI, Bhubaneswar, India S. Bahinipati24, C. Kar, G. Kole, P. Mal, V.K. Muraleedharan Nair Bindhu, A. Nayak25, D.K. Sahoo24, S.K. Swain – 27 – Panjab University, Chandigarh, India Panjab University, Chandigarh, India S. Bansal, S.B. Beri, V. Bhatnagar, S. Chauhan, R. Chawla, N. Dhingra, R. Gupt S. Bansal, S.B. Beri, V. Bhatnagar, S. Chauhan, R. Chawla, N. Dhingra, R. Gupta, A. Kaur, M. Kaur, S. Kaur, P. Kumari, M. Lohan, M. Meena, K. Sandeep, S. Sharma, J.B. Singh, A.K. Virdi University of Delhi, Delhi, India University of Delhi, Delhi, India A. Bhardwaj, B.C. Choudhary, R.B. Garg, M. Gola, S. Keshri, Ashok Kumar, M. Naimud- din, P. Priyanka, K. Ranjan, Aashaq Shah, R. Sharma A. Bhardwaj, B.C. Choudhary, R.B. Garg, M. Gola, S. Keshri, Ashok Kumar, M. Naimud- din, P. Priyanka, K. Ranjan, Aashaq Shah, R. Sharma Saha Institute of Nuclear Physics, HBNI, Kolkata, India R. Bhardwaj26, M. Bharti26, R. Bhattacharya, S. Bhattacharya, U. Bhawandeep26, D. Bhowmik, S. Dutta, S. Ghosh, M. Maity27, K. Mondal, S. Nandan, A. Purohit, P.K. Rout, G. Saha, S. Sarkar, T. Sarkar27, M. Sharan, B. Singh26, S. Thakur26 JHEP12(2020)082 Indian Institute of Technology Madras, Madras, India Indian Institute of Technology Madras, Madras, India P.K. Behera, P. Kalbhor, A. Muhammad, P.R. Pujahari, A. Sharma, A.K. Sikdar Indian Institute of Technology Madras, Madras, India P.K. Behera, P. Kalbhor, A. Muhammad, P.R. Pujahari, A. Sharma, A.K. Sikdar P.K. Behera, P. Kalbhor, A. Muhammad, P.R. Pujahari, A. Sharma, A.K. Sikdar Bhabha Atomic Research Centre, Mumbai, India D. Dutta, V. Jha, V. Kumar, D.K. Mishra, P.K. Netrakanti, L.M. Pant, P. Shukla Tata Institute of Fundamental Research-A, Mumbai, India T. Aziz, M.A. Bhat, S. Dugad, G.B. Mohanty, N. Sur, RavindraKumar Verma T. Aziz, M.A. Bhat, S. Dugad, G.B. Mohanty, N. Sur, RavindraKumar Verma Tata Institute of Fundamental Research-B, Mumbai, India Tata Institute of Fundamental Research-B, Mumbai, India S. Banerjee, S. Bhattacharya, S. Chatterjee, P. Das, M. Guchait, S. Karmakar, S. Kumar, G. Majumder, K. Mazumdar, N. Sahoo, S. Sawant Indian Institute of Science Education and Research (IISER), Pune, India S. Dube, V. Hegde, B. Kansal, A. Kapoor, K. Kothekar, S. Pandey, A. Rane, A. Rastogi, S. Sharma Institute for Research in Fundamental Sciences (IPM), Tehran, Iran S. Chenarani28, E. Eskandari Tadavani, S.M. Etesami28, M. Khakzad, M. Mohammadi Na- jafabadi, M. Naseri, F. Rezaei Hosseinabadi University College Dublin, Dublin, Ireland M. Felcini, M. Grunewald University College Dublin, Dublin, Ireland M. Felcini, M. Grunewald INFN Sezione di Baria, Università di Barib, Politecnico di Baric, Bari, Italy M. Abbresciaa,b, R. Alya,b,29, C. Calabriaa,b, A. Colaleoa, D. Creanzaa,c, L. Cristellaa,b, N. De Filippisa,c, M. De Palmaa,b, A. Di Florioa,b, W. Elmetenaweea,b, L. Fiorea, A. Gelmia,b, G. Iasellia,c, M. Incea,b, S. Lezkia,b, G. Maggia,c, M. Maggia, G. Minielloa,b, S. Mya,b, S. Nuzzoa,b, A. Pompilia,b, G. Pugliesea,c, R. Radognaa, A. Ranieria, G. Selvaggia,b, L. Silvestrisa, F.M. Simonea,b, R. Vendittia, P. Verwilligena INFN Sezione di Bolognaa, Università di Bolognab, Bologna, Italy G. Abbiendia, C. Battilanaa,b, D. Bonacorsia,b, L. Borgonovia,b, S. Braibant-Giacomellia,b, R. Campaninia,b, P. Capiluppia,b, A. Castroa,b, F.R. Cavalloa, C. Cioccaa, G. Codispotia,b, M. Cuﬃania,b, G.M. Dallavallea, F. Fabbria, A. Fanfania,b, E. Fontanesia,b, P. Giacomellia, p , p pp , , , , p , M. Cuﬃania,b, G.M. Dallavallea, F. Fabbria, A. Fanfania,b, E. Fontanesia,b, P. Giacomellia, – 28 – C. Grandia, L. Guiduccia,b, F. Iemmia,b, S. Lo Meoa,30, S. Marcellinia, G. Masettia, F.L. Navarriaa,b, A. Perrottaa, F. Primaveraa,b, A.M. Rossia,b, T. Rovellia,b, G.P. Sirolia,b, N. Tosia C. Grandia, L. Guiduccia,b, F. Iemmia,b, S. Lo Meoa,30, S. Marcellinia, G. Masettia, F.L. Navarriaa,b, A. Perrottaa, F. Primaveraa,b, A.M. Rossia,b, T. Rovellia,b, G.P. Sirolia,b, N. Tosia INFN Sezione di Cataniaa, Università di Cataniab, Catania, Italy S. Albergoa,b,31, S. Costaa,b, A. Di Mattiaa, R. Potenzaa,b, A. Tricomia,b,31, C. Tuvea,b INFN Sezione di Firenzea, Università di Firenzeb, Firenze, Italy G. Barbaglia, A. Cassesea, R. Ceccarellia,b, V. Ciullia,b, C. Civininia, R. D’Alessandroa,b, E. Focardia,b, G. Latinoa,b, P. Lenzia,b, M. Meschinia, S. Paolettia, G. Sguazzonia, L. Viliania G. Barbagli , A. Cassese , R. Ceccarelli , , V. Ciulli , , C. Civinini , R. D’Alessandro , , E. Focardia,b, G. Latinoa,b, P. Lenzia,b, M. Meschinia, S. Paolettia, G. Sguazzonia, L. Viliania JHEP12(2020)082 INFN Laboratori Nazionali di Frascati, Frascati, Italy L. Benussi, S. Bianco, D. Piccolo INFN Sezione di Genovaa, Università di Genovab, Genova, Italy M. Bozzoa,b, F. Ferroa, R. Mulargiaa,b, E. Robuttia, S. Tosia,b INFN Sezione di Milano-Bicoccaa, Università di Milano-Bicoccab, Milano, Italy A. Benagliaa, A. Beschia,b, F. Brivioa,b, V. Cirioloa,b,17, S. Di Guidaa,b,17, M.E. Dinardoa,b, P. Dinia, S. Gennaia, A. Ghezzia,b, P. Govonia,b, L. Guzzia,b, M. Malbertia, S. Malvezzia, D. Menascea, F. Montia,b, L. Moronia, M. Paganonia,b, D. Pedrinia, S. Ragazzia,b, T. Tabarelli de Fatisa,b, D. Zuoloa,b INFN Sezione di Napolia, Università di Napoli ‘Federico II’b, Napoli, Italy, Università della Basilicatac, Potenza, Italy, Università G. Marconid, Roma, Italy S. Buontempoa, N. Cavalloa,c, A. De Iorioa,b, A. Di Crescenzoa,b, F. Fabozzia,c, F. Fienga b b d 17 17 S. Buontempoa, N. Cavalloa,c, A. De Iorioa,b, A. Di Crescenzoa,b, F. Fabozzia,c, F. Fiengaa, G. Galatia, A.O.M. Iorioa,b, L. Listaa,b, S. Meolaa,d,17, P. Paoluccia,17, B. Rossia, C. Sciaccaa,b, E. Voevodinaa,b G. Galatia, A.O.M. Iorioa,b, L. Listaa,b, S. Meolaa,d,17, P. Paoluccia,17, B. Rossia, C. Sciaccaa,b, E. Voevodinaa,b INFN Sezione di Padovaa, Università di Padovab, Padova, Italy, Università di Trentoc, Trento, Italy P. Azzia, N. Bacchettaa, D. Biselloa,b, A. Bolettia,b, A. Bragagnoloa,b, R. Carlina,b, P. Checchiaa, P. De Castro Manzanoa, T. Dorigoa, U. Dossellia, F. Gasparinia,b, U. Gasparinia,b, A. Gozzelinoa, S.Y. Hoha,b, P. Lujana, M. Margonia,b, A.T. Meneguzzoa,b, J. Pazzinia,b, M. Presillab, P. Ronchesea,b, R. Rossina,b, F. Simonettoa,b, A. Tikoa, M. Tosia,b, M. Zanettia,b, P. Zottoa,b, G. Zumerlea,b p , , , j , g , g , J. Pazzinia,b, M. Presillab, P. Ronchesea,b, R. Rossina,b, F. Simonettoa,b, A. Tikoa, M. Tosia,b, M. Zanettia,b, P. Zottoa,b, G. Zumerlea,b INFN Sezione di Paviaa, Università di Paviab, Pavia, Italy INFN Sezione di Paviaa, Università di Paviab, Pavia, Italy INFN Sezione di Paviaa, Università di Paviab, Pavia, Italy A. Braghieria, D. Fiorinaa,b, P. Montagnaa,b, S.P. Rattia,b, V. Rea, M. Ressegottia,b, C. Riccardia,b, P. Salvinia, I. Vaia, P. Vituloa,b A. Braghieria, D. Fiorinaa,b, P. Montagnaa,b, S.P. Rattia,b, V. Rea, M. Ressegottia C. Riccardia,b, P. Salvinia, I. Vaia, P. Vituloa,b INFN Sezione di Perugiaa, Università di Perugiab, Perugia, Italy M. Biasinia,b, G.M. Bileia, D. Ciangottinia,b, L. Fanòa,b, P. Laricciaa,b, R. Leonardia,b, E. Manonia, G. Mantovania,b, V. Mariania,b, M. Menichellia, A. Rossia,b, A. Santocchiaa,b, D. Spigaa – 29 – INFN Sezione di Pisaa, Università di Pisab, Scuola Normale Superiore di Pisac, Pisa, Italy K. Androsova, P. Azzurria, G. Bagliesia, V. Bertacchia,c, L. Bianchinia, T. Boccalia, K. Androsova, P. Azzurria, G. Bagliesia, V. Bertacchia,c, L. Bianchinia, T. Boccalia, R Castaldia M A Cioccia,b R Dell’Orsoa S Donatoa G Fedia L Gianninia,c K. Androsova, P. Azzurria, G. Bagliesia, V. Bertacchia,c, L. Bianchinia, T. Boccalia, R. Castaldia, M.A. Cioccia,b, R. Dell’Orsoa, S. Donatoa, G. Fedia, L. Gianninia,c, A. Giassia, M.T. Grippoa, F. Ligabuea,c, E. Mancaa,c, G. Mandorlia,c, A. Messineoa,b, F P ll a A Ri ia,b G R l di32 S R Ch dh A S ib a P S l a R. Castaldia, M.A. Cioccia,b, R. Dell’Orsoa, S. Donatoa, G. Fedia, L. Gianninia,c, A. Giassia, M.T. Grippoa, F. Ligabuea,c, E. Mancaa,c, G. Mandorlia,c, A. Messineoa,b, F. Pallaa, A. Rizzia,b, G. Rolandi32, S. Roy Chowdhury, A. Scribanoa, P. Spagnoloa, R. Tenchinia, G. Tonellia,b, N. Turinia, A. Venturia, P.G. Verdinia A. Giassia, M.T. Grippoa, F. Ligabuea,c, E. Mancaa,c, G. Mandorlia,c, A. Messineoa,b, F. Pallaa, A. Rizzia,b, G. Rolandi32, S. Roy Chowdhury, A. Scribanoa, P. Spagnoloa, R. Tenchinia, G. Tonellia,b, N. Turinia, A. Venturia, P.G. Verdinia INFN Sezione di Romaa, Sapienza Università di Romab, Rome, Italy F. Cavallaria, M. Cipriania,b, D. Del Rea,b, E. Di Marcoa,b, M. Diemoza, E. Longoa,b, P. Meridiania, G. Organtinia,b, F. Pandolﬁa, R. Paramattia,b, C. Quarantaa,b, S. Rahatloua,b, C. Rovellia, F. Santanastasioa,b, L. Soﬃa,b INFN Sezione di Romaa, Sapienza Università di Romab, Rome, Italy INFN Sezione di Romaa, Sapienza Università di Romab, Rome, Italy F. Cavallaria, M. Cipriania,b, D. Del Rea,b, E. Di Marcoa,b, M. Diemoza, E. Longoa,b, P M idi ia G O i ia b F P d lﬁa R P ia b C Q a b F. Cavallaria, M. Cipriania,b, D. Del Rea,b, E. Di Marcoa,b, M. Diemoza, E. Longoa F. Cavallaria, M. Cipriania,b, D. Del Rea,b, E. INFN Sezione di Triestea, Università di Triesteb, Trieste, Italy S. Belfortea, V. Candelisea,b, M. Casarsaa, F. Cossuttia, A. Da Rolda,b, G. Della Riccaa,b, F. Vazzolera,b, A. Zanettia INFN Sezione di Paviaa, Università di Paviab, Pavia, Italy Di Marcoa,b, M. Diemoza, E. Longoa,b, P. Meridiania, G. Organtinia,b, F. Pandolﬁa, R. Paramattia,b, C. Quarantaa,b, S. Rahatloua,b, C. Rovellia, F. Santanastasioa,b, L. Soﬃa,b JHEP12(2020)082 , p , , , , g P. Meridiania, G. Organtinia,b, F. Pandolﬁa, R. Paramattia,b, C. Quarantaa S. Rahatloua,b, C. Rovellia, F. Santanastasioa,b, L. Soﬃa,b INFN Sezione di Torinoa, Università di Torinob, Torino, Italy, Università del Piemonte Orientalec, Novara, Italy INFN Sezione di Torinoa, Università di Torinob, Torino, Italy, Università d Piemonte Orientalec, Novara, Italy N. Amapanea,b, R. Arcidiaconoa,c, S. Argiroa,b, M. Arneodoa,c, N. Bartosika, R. Bellana N. Amapanea,b, R. Arcidiaconoa,c, S. Argiroa,b, M. Arneodoa,c, N. Bartosika, R. Bellana,b, A. Bellora, C. Biinoa, A. Cappatia,b, N. Cartigliaa, S. Comettia, M. Costaa,b, p , , g , , , , A. Bellora, C. Biinoa, A. Cappatia,b, N. Cartigliaa, S. Comettia, M. Costaa,b, R. Covarellia,b, N. Demariaa, B. Kiania,b, F. Legger, C. Mariottia, S. Masellia, E. Migliorea,b, V. Monacoa,b, E. Monteila,b, M. Montenoa, M.M. Obertinoa,b, G. Ortonaa,b, L. Pachera,b, N. Pastronea, M. Pelliccionia, G.L. Pinna Angionia,b, A. Romeroa,b, M. Ruspaa,c, R. Salvaticoa,b, V. Solaa, A. Solanoa,b, D. Soldia,b, A. Staianoa INFN Sezione di Triestea, Università di Triesteb, Trieste, Italy Kyungpook National University, Daegu, Korea Kyungpook National University, Daegu, Korea B. Kim, D.H. Kim, G.N. Kim, J. Lee, S.W. Lee, C.S. Moon, Y.D. Oh, S.I. Pak, S. Sekmen, D.C. Son, Y.C. Yang Chonnam National University, Institute for Universe and Elementary Particles, Kwangju, Korea H. Kim, D.H. Moon, G. Oh H. Kim, D.H. Moon, G. Oh Hanyang University, Seoul, Korea B. Francois, T.J. Kim, J. Park Hanyang University, Seoul, Korea B. Francois, T.J. Kim, J. Park B. Francois, T.J. Kim, J. Park Korea University, Seoul, Korea S. Cho, S. Choi, Y. Go, S. Ha, B. Hong, K. Lee, K.S. Lee, J. Lim, J. Park, S.K. Park, Y. Roh, J. Yoo Kyung Hee University, Department of Physics, Seoul, Republic of Korea J. Goh Sejong University, Seoul, Korea H.S. Kim – 30 – Seoul National University, Seoul, Korea J. Almond, J.H. Bhyun, J. Choi, S. Jeon, J. Kim, J.S. Kim, H. Lee, K. Lee, S. Lee, K. Nam, M. Oh, S.B. Oh, B.C. Radburn-Smith, U.K. Yang, H.D. Yoo, I. Yoon, G.B. Yu Seoul National University, Seoul, Korea University of Seoul, Seoul, Korea D. Jeon, H. Kim, J.H. Kim, J.S.H. Lee, I.C. Park, I.J Watson D. Jeon, H. Kim, J.H. Kim, J.S.H. Lee, I.C. Park, I.J Watson Sungkyunkwan University, Suwon, Korea Y. Choi, C. Hwang, Y. Jeong, J. Lee, Y. Lee, I. Yu Sungkyunkwan University, Suwon, Korea Y. Choi, C. Hwang, Y. Jeong, J. Lee, Y. Lee, I. Yu Riga Technical University, Riga, Latvia V. Veckalns33 Riga Technical University, Riga, Latvia V. Veckalns33 JHEP12(2020)082 Vilnius University, Vilnius, Lithuania V. Dudenas, A. Juodagalvis, A. Rinkevicius, G. Tamulaitis, J. Vaitkus National Centre for Particle Physics, Universiti Malaya, Kuala Lumpur, Malaysia Z.A. Ibrahim, F. Mohamad Idris34, W.A.T. Wan Abdullah, M.N. Yusli, Z. Zolkapli Universidad de Sonora (UNISON), Hermosillo, Mexico J.F. Benitez, A. Castaneda Hernandez, J.A. Murillo Quijada, L. Valencia Palomo Centro de Investigacion y de Estudios Avanzados del IPN, Mexico City, Mexico H. Castilla-Valdez, E. De La Cruz-Burelo, I. Heredia-De La Cruz35, R. Lopez-Fernandez, A. Sanchez-Hernandez Universidad Iberoamericana, Mexico City, Mexico S. Carrillo Moreno, C. Oropeza Barrera, M. Ramirez-Garcia, F. Vazquez Valencia Benemerita Universidad Autonoma de Puebla, Puebla, Mexico J. Eysermans, I. Pedraza, H.A. Salazar Ibarguen, C. Uribe Estrada Universidad Autónoma de San Luis Potosí, San Luis Potosí, Mexico A. Morelos Pineda University of Montenegro, Podgorica, Montenegro J. Mijuskovic, N. Raicevic University of Auckland, Auckland, New Zealand D. Krofcheck University of Auckland, Auckland, New Zealand D. Portugal M. Araujo, P. Bargassa, D. Bastos, A. Di Francesco, P. Faccioli, B. Galinhas, M. Gallinar JHEP12(2020)082 M. Araujo, P. Bargassa, D. Bastos, A. Di Francesco, P. Faccioli, B. Galinhas, M. Gallinaro, J. Hollar, N. Leonardo, T. Niknejad, J. Seixas, K. Shchelina, G. Strong, O. Toldaiev, J. Varela J. Hollar, N. Leonardo, T. Niknejad, J. Seixas, K. Shchelina, G. Strong, O. Toldaiev, J. Varela J. Hollar, N. Leonardo, T. Niknejad, J. Seixas, K. Shchelina, G. Strong, O. Toldaiev, J Varela National Centre for Nuclear Research, Swierk, Poland H. Bialkowska, M. Bluj, B. Boimska, M. Górski, M. Kazana, M. Szleper, P. Zalewski Institute of Experimental Physics, Faculty of Physics, University of Warsaw, Warsaw, Poland K. Bunkowski, A. Byszuk36, K. Doroba, A. Kalinowski, M. Konecki, J. Krolikowski, M. Misiura, M. Olszewski, M. Walczak Kyungpook National University, Daegu, Korea Krofcheck University of Canterbury, Christchurch, New Zealand S. Bheesette, P.H. Butler National Centre for Physics, Quaid-I-Azam University, Islamabad, Pakistan A. Ahmad, M. Ahmad, Q. Hassan, H.R. Hoorani, W.A. Khan, M.A. Shah, M. Shoaib, M. Waqas AGH University of Science and Technology Faculty of Computer Science, Electronics and Telecommunications, Krakow, Poland V. Avati, L. Grzanka, M. Malawski – 31 – National Centre for Nuclear Research, Swierk, Poland Joint Institute for Nuclear Research, Dubna, Russia V. Alexakhin, Y. Ershov, I. Golutvin, I. Gorbunov, A. Kamenev, V. Karjavine, A. Lanev, A. Malakhov, V. Matveev37,38, P. Moisenz, V. Palichik, V. Perelygin, M. Savina, S. Shma- tov, S. Shulha, N. Skatchkov, V. Smirnov, N. Voytishin, B.S. Yuldashev39, A. Zarubin Petersburg Nuclear Physics Institute, Gatchina (St. Petersburg), Russia L. Chtchipounov, V. Golovtcov, Y. Ivanov, V. Kim40, E. Kuznetsova41, P. Levchenko, V. Murzin, V. Oreshkin, I. Smirnov, D. Sosnov, V. Sulimov, L. Uvarov, A. Vorobyev Institute for Nuclear Research, Moscow, Russia Yu. Andreev, A. Dermenev, S. Gninenko, N. Golubev, A. Karneyeu, M. Kirsanov, N. Krasnikov, A. Pashenkov, D. Tlisov, A. Toropin Yu. Andreev, A. Dermenev, S. Gninenko, N. Golubev, A. Karneyeu, M. Kirsanov, N. Krasnikov, A. Pashenkov, D. Tlisov, A. Toropin Institute for Theoretical and Experimental Physics named by A.I. Alikhanov of NRC ‘Kurchatov Institute’, Moscow, Russia Institute for Theoretical and Experimental Physics named by A.I. Alikhanov of NRC ‘Kurchatov Institute’, Moscow, Russia V. Epshteyn, V. Gavrilov, N. Lychkovskaya, A. Nikitenko42, V. Popov, I. Pozdnyakov, of NRC ‘Kurchatov Institute’, Moscow, Russia V. Epshteyn, V. Gavrilov, N. Lychkovskaya, A. Nikitenko42, V. Popov, I. Pozdnyakov, G. Safronov, A. Spiridonov, A. Stepennov, M. Toms, E. Vlasov, A. Zhokin V. Epshteyn, V. Gavrilov, N. Lychkovskaya, A. Nikitenko42, V. Popov, I. Pozdnyakov, G. Safronov, A. Spiridonov, A. Stepennov, M. Toms, E. Vlasov, A. Zhokin Moscow Institute of Physics and Technology, Moscow, Russia T. Aushev National Research Nuclear University ‘Moscow Engineering Physics Institute’ (MEPhI), Moscow, Russia M. Chadeeva43, P. Parygin, D. Philippov, E. Popova, V. Rusinov P.N. Lebedev Physical Institute, Moscow, Russia P.N. Lebedev Physical Institute, Moscow, Russia V. Andreev, M. Azarkin, I. Dremin, M. Kirakosyan, A. Terkulov V. Andreev, M. Azarkin, I. Dremin, M. Kirakosyan, A. Terkulov Skobeltsyn Institute of Nuclear Physics, Lomonosov Moscow State University, Moscow, Russia Moscow, Russia A. Belyaev, E. Boos, M. Dubinin44, L. Dudko, A. Ershov, A. Gribushin, V. Klyukhin, O K d l I L khti S Ob t S P t h k V S i A S i i A. Belyaev, E. Boos, M. Dubinin44, L. Dudko, A. Ershov, A. Gribushin, V. Klyukhi Belyaev, E. Boos, M. Dubinin44, L. Dudko, A. Ershov, A. Gribushin, V. Klyukhin, Kodolova I Lokhtin S Obraztsov S Petrushanko V Savrin A Snigirev A. Belyaev, E. Boos, M. Dubinin44, L. Dudko, A. Ershov, A. Gribushin, V. Klyukhin, O. Kodolova, I. Lokhtin, S. Obraztsov, S. Petrushanko, V. Savrin, A. Snigirev Belyaev, E. Boos, M. Dubinin44, L. Dudko, A. Ershov, A. Gribushin, V. Klyukhin, Kodolova, I. Lokhtin, S. Obraztsov, S. Petrushanko, V. Savrin, A. Snigirev Novosibirsk State University (NSU), Novosibirsk, Russia A. Barnyakov45, V. Blinov45, T. Dimova45, L. Kardapoltsev45, Y. Skovpen45 Novosibirsk State University (NSU), Novosibirsk, Russia A. Barnyakov45, V. Blinov45, T. Dimova45, L. Kardapoltsev45, Y. Skovpen45 – 32 – Institute for High Energy Physics of National Research Centre ‘Kurchatov Institute’, Protvino, Russia I. Azhgirey, I. Bayshev, S. Bitioukov, V. Kachanov, D. Konstantinov, P. Mandrik, V. Petrov, R. Ryutin, S. Slabospitskii, A. Sobol, S. Troshin, N. Tyurin, A. Uzunian, A. Volkov National Research Tomsk Polytechnic University, Tomsk, Russia A. Babaev, A. Iuzhakov, V. Okhotnikov Tomsk State University, Tomsk, Russia V. Borchsh, V. Ivanchenko, E. Tcherniaev JHEP12(2020)082 University of Belgrade: Faculty of Physics and VINCA Institute of Nuclear Sciences, Serbia P. Adzic46, P. Cirkovic, M. Dordevic, P. Milenovic, J. Milosevic, M. Stojanovic University of Ruhuna, Department of Physics, Matara, Sri Lanka W.G.D. Dharmaratna, N. Wickramage Centro de Investigaciones Energéticas Medioambientales y Tecnológica (CIEMAT), Madrid, Spain Centro de Investigaciones Energéticas Medioambientales y Tecnológicas (CIEMAT), Madrid, Spain M. Aguilar-Benitez, J. Alcaraz Maestre, A. Álvarez Fernández, I. Bachiller, M. Barrio Luna, J.A. Brochero Cifuentes, C.A. Carrillo Montoya, M. Cepeda, M. Cerrada, N. Colino, B. De La Cruz, A. Delgado Peris, C. Fernandez Bedoya, J.P. Fernández Ramos, J. Flix, M.C. Fouz, O. Gonzalez Lopez, S. Goy Lopez, J.M. Hernandez, M.I. Josa, D. Moran, Á. Navarro Tobar, A. Pérez-Calero Yzquierdo, J. Puerta Pelayo, I. Redondo, L. Romero, S. Sánchez Navas, M.S. Soares, A. Triossi, C. Willmott Universidad Autónoma de Madrid, Madrid, Spain Universidad Autónoma de Madrid, Madrid, Spain C. Albajar, J.F. de Trocóniz, R. Reyes-Almanza Universidad Autónoma de Madrid, Madrid, Spain C. Albajar, J.F. de Trocóniz, R. Reyes-Almanza Universidad de Oviedo, Instituto Universitario de Ciencias y Tecnologías Espaciales de Asturias (ICTEA), Oviedo, Spain B. Alvarez Gonzalez, J. Cuevas, C. Erice, J. Fernandez Menendez, S. Folgueras, I. Gon- zalez Caballero, J.R. González Fernández, E. Palencia Cortezon, V. Rodríguez Bouza, S. Sanchez Cruz Instituto de Física de Cantabria (IFCA), CSIC-Universidad de Cantabri Santander, Spain Instituto de Física de Cantabria (IFCA), CSIC-Universidad de Cantabria, Santander, Spain I.J. Cabrillo, A. Calderon, B. Chazin Quero, J. Duarte Campderros, M. Fernandez, P.J. Fernández Manteca, A. García Alonso, G. Gomez, C. Martinez Rivero, P. Mar- tinez Ruiz del Arbol, F. Matorras, J. Piedra Gomez, C. Prieels, T. Rodrigo, A. Ruiz-Jimeno, L. Russo47, L. Scodellaro, I. Vila, J.M. Vizan Garcia University of Colombo, Colombo, Sri Lanka D.U.J. Sonnadara University of Ruhuna, Department of Physics, Matara, Sri Lanka W.G.D. Dharmaratna, N. Wickramage – 33 – CERN, European Organization for Nuclear Research, Geneva, Switzerland D. Abbaneo, B. Akgun, E. Auﬀray, G. Auzinger, J. Baechler, P. Baillon, A.H. Ball, D. Barney, J. Bendavid, M. Bianco, A. Bocci, P. Bortignon, E. Bossini, C. Botta, E. Bron- dolin, T. Camporesi, A. Caratelli, G. Cerminara, E. Chapon, G. Cucciati, D. d’Enterria, A. Dabrowski, N. Daci, V. Daponte, A. David, O. Davignon, A. De Roeck, M. Deile, M. Dobson, M. Dünser, N. Dupont, A. Elliott-Peisert, N. Emriskova, F. Fallavollita48, D. Fasanella, S. Fiorendi, G. Franzoni, J. Fulcher, W. Funk, S. Giani, D. Gigi, A. Gilbert, K. Gill, F. Glege, L. Gouskos, M. Gruchala, M. Guilbaud, D. Gulhan, J. Hegeman, C. Heidegger, Y. Iiyama, V. Innocente, T. James, P. Janot, O. Karacheban20, J. Kaspar, J. Kieseler, M. Krammer1, N. Kratochwil, C. Lange, P. Lecoq, C. Lourenço, L. Malgeri, M. Mannelli, A. Massironi, F. Meijers, J.A. Merlin, S. Mersi, E. Meschi, F. Moort- gat, M. Mulders, J. Ngadiuba, J. Niedziela, S. Nourbakhsh, S. Orfanelli, L. Orsini, F. Pantaleo17, L. Pape, E. Perez, M. Peruzzi, A. Petrilli, G. Petrucciani, A. Pfeiﬀer, M. Pierini, F.M. Pitters, D. Rabady, A. Racz, M. Rieger, M. Rovere, H. Sakulin, C. Schäfer, C. Schwick, M. Selvaggi, A. Sharma, P. Silva, W. Snoeys, P. Sphicas49, J. Steggemann, S. Summers, V.R. Tavolaro, D. Treille, A. Tsirou, G.P. Van Onsem, A. Vartak, M. Verzetti, W.D. Zeuner JHEP12(2020)082 Paul Scherrer Institut, Villigen, Switzerland L. Caminada50, K. Deiters, W. Erdmann, R. Horisberger, Q. Ingram, H.C. Kaestli, D. Kotlinski, U. Langenegger, T. Rohe, S.A. Wiederkehr ETH Zurich — Institute for Particle Physics and Astrophysics (IPA), Zurich, Switzerland M. Backhaus, P. Berger, N. Chernyavskaya, G. Dissertori, M. Dittmar, M. Donegà, C. Dorfer, T.A. Gómez Espinosa, C. Grab, D. Hits, W. Lustermann, R.A. Manzoni, M.T. Meinhard, F. Micheli, P. Musella, F. Nessi-Tedaldi, F. Pauss, G. Perrin, L. Per- rozzi, S. Pigazzini, M.G. Ratti, M. Reichmann, C. Reissel, T. Reitenspiess, D. Ruini, D.A. Instituto de Física de Cantabria (IFCA), CSIC-Universidad de Cantabri Santander, Spain Sanz Becerra, M. Schönenberger, L. Shchutska, M.L. Vesterbacka Olsson, R. Wallny, D.H. Zhu Universität Zürich, Zurich, Switzerland Universität Zürich, Zurich, Switzerland T.K. Aarrestad, C. Amsler51, D. Brzhechko, M.F. Canelli, A. De Cosa, R. Del Burgo, B. Kilminster, S. Leontsinis, V.M. Mikuni, I. Neutelings, G. Rauco, P. Robmann, K. Schweiger, C. Seitz, Y. Takahashi, S. Wertz, A. Zucchetta National Central University, Chung-Li, Taiwan T.H. Doan, C.M. Kuo, W. Lin, A. Roy, S.S. Yu National Taiwan University (NTU), Taipei, Taiwan P. Chang, Y. Chao, K.F. Chen, P.H. Chen, W.-S. Hou, Y.y. Li, R.-S. Lu, E. Paganis, National Taiwan University (NTU), Taipei, Taiwan National Taiwan University (NTU), Taipei, Taiwan P. Chang, Y. Chao, K.F. Chen, P.H. Chen, W.-S. Hou, Y.y. Li, R.-S. Lu, E. Paganis, A. Psallidas, A. Steen P. Chang, Y. Chao, K.F. Chen, P.H. Chen, W.-S. Hou, Y.y. Li, R.-S. Lu, E. Paganis, A. Psallidas, A. Steen Chulalongkorn University, Faculty of Science, Department of Physics, Bangkok, Thailand B. Asavapibhop, C. Asawatangtrakuldee, N. Srimanobhas, N. Suwonjandee B. Asavapibhop, C. Asawatangtrakuldee, N. Srimanobhas, N. Suwonjandee – 34 – Çukurova University, Physics Department, Science and Art Faculty, Adana, Turkey A. Bat, F. Boran, A. Celik52, S. Cerci53, S. Damarseckin54, Z.S. Demiroglu, F. Dolek, C. Dozen55, I. Dumanoglu, G. Gokbulut, EmineGurpinar Guler56, Y. Guler, I. Hos57, C. Isik, E.E. Kangal58, O. Kara, A. Kayis Topaksu, U. Kiminsu, G. Onengut, K. Ozdemir59, S. Ozturk60, A.E. Simsek, D. Sunar Cerci53, U.G. Tok, S. Turkcapar, I.S. Zorbakir, C. Zorbilmez Middle East Technical University, Physics Department, Ankara, Turkey B. Isildak61, G. Karapinar62, M. Yalvac JHEP12(2020)082 Bogazici University, Istanbul, Turkey I.O. Atakisi, E. Gülmez, M. Kaya63, O. Kaya64, Ö. Özçelik, S. Tekten, E.A. Yetkin Istanbul Technical University, Istanbul, Turkey A. Cakir, K. Cankocak, Y. Komurcu, S. Sen66 Istanbul Technical University, Istanbul, Turkey A. Cakir, K. Cankocak, Y. Komurcu, S. Sen66 Istanbul Technical University, Istanbul, Tu A. Cakir, K. Cankocak, Y. Komurcu, S. Sen66 Istanbul University, Istanbul, Turkey B. Kaynak, S. Ozkorucuklu Istanbul University, Istanbul, Turkey B. Kaynak, S. Ozkorucuklu Institute for Scintillation Materials of National Academy of Science of Ukraine, Kharkov, Ukraine B. Grynyov National Scientiﬁc Center, Kharkov Institute of Physics and Technology, Kharkov, Ukraine L. Levchuk University of Bristol, Bristol, United Kingdom E. Bhal, S. Bologna, J.J. Brooke, D. Burns67, E. Clement, D. Cussans, H. Flacher, J. Goldstein, G.P. Heath, H.F. Heath, L. Kreczko, B. Krikler, S. Paramesvaran, B. Penning, T. Sakuma, S. Seif El Nasr-Storey, V.J. Smith, J. Instituto de Física de Cantabria (IFCA), CSIC-Universidad de Cantabri Santander, Spain Taylor, A. Titterton Rutherford Appleton Laboratory, Didcot, United Kingdom K.W. Bell, A. Belyaev68, C. Brew, R.M. Brown, D.J.A. Cockerill, J.A. Coughlan, K. Harder, S. Harper, J. Linacre, K. Manolopoulos, D.M. Newbold, E. Olaiya, D. Petyt, T. Reis, T. Schuh, C.H. Shepherd-Themistocleous, A. Thea, I.R. Tomalin, T. Williams, W.J. Womersley Imperial College, London, United Kingdom R. Bainbridge, P. Bloch, J. Borg, S. Breeze, O. Buchmuller, A. Bundock, GurpreetS- ingh CHAHAL69, D. Colling, P. Dauncey, G. Davies, M. Della Negra, R. Di Maria, P. Everaerts, G. Hall, G. Iles, M. Komm, C. Laner, L. Lyons, A.-M. Magnan, S. Malik, A. Martelli, V. Milosevic, A. Morton, J. Nash70, V. Palladino, M. Pesaresi, D.M. Raymond, A. Richards, A. Rose, E. Scott, C. Seez, A. Shtipliyski, M. Stoye, T. Strebler, A. Tapper, K. Uchida, T. Virdee17, N. Wardle, D. Winterbottom, J. Wright, A.G. Zecchinelli, S.C. Zenz – 35 – Brunel University, Uxbridge, United Kingdom J.E. Cole, P.R. Hobson, A. Khan, P. Kyberd, C.K. Mackay, I.D. Reid, L. Teodorescu, S. Zahid Baylor University, Waco, U.S.A. Baylor University, Waco, U.S.A. K. Call, B. Caraway, J. Dittmann, K. Hatakeyama, C. Madrid, B. McMaster, N. Pastika, C. Smith Catholic University of America, Washington, DC, U.S.A. Catholic University of America, Washington, DC, U.S.A. R. Bartek, A. Dominguez, R. Uniyal, A.M. Vargas Hernandez JHEP12(2020)082 The University of Alabama, Tuscaloosa, U.S.A. A. Buccilli, S.I. Cooper, C. Henderson, P. Rumerio, C. West A. Buccilli, S.I. Cooper, C. Henderson, P. Rumerio, C. West Boston University, Boston, U.S.A. A. Albert, D. Arcaro, Z. Demiragli, D. Gastler, C. Richardson, J. Rohlf, D. Sperka, I. Suarez, L. Sulak, D. Zou Brown University, Providence, U.S.A. G. Benelli, B. Burkle, X. Coubez18, D. Cutts, Y.t. Duh, M. Hadley, U. Heintz, J.M. Hogan71, K.H.M. Kwok, E. Laird, G. Landsberg, K.T. Lau, J. Lee, Z. Mao, M. Narain, S. Sagir72, R. Syarif, E. Usai, D. Yu, W. Zhang University of California, Davis, Davis, U.S.A. R. Band, C. Brainerd, R. Breedon, M. Calderon De La Barca Sanchez, M. Cherto R. Band, C. Brainerd, R. Breedon, M. Calderon De La Barca Sanchez, M. Chertok, J. Conway, R. Conway, P.T. Cox, R. Erbacher, C. Flores, G. Funk, F. Jensen, W. Ko, J. Conway, R. Conway, P.T. Cox, R. Erbacher, C. Flores, G. Funk, F. Jensen, W. K O. Kukral, R. Lander, M. Mulhearn, D. Pellett, J. Pilot, M. Shi, D. Taylor, K. Tos, M. Tripathi, Z. Wang, F. Zhang M. Tripathi, Z. Wang, F. Zhang University of California, Los Angeles, U.S.A. M. Bachtis, C. Bravo, R. Cousins, A. Dasgupta, A. Florent, J. Hauser, M. Ignatenko, N. Mccoll, W.A. Nash, S. Regnard, D. Saltzberg, C. Schnaible, B. Stone, V. Valuev University of California, Riverside, Riverside, U.S.A. K. Burt, Y. Chen, R. Clare, J.W. Gary, S.M.A. Ghiasi Shirazi, G. Hanson, G. Karapostoli, E. Kennedy, O.R. Long, M. Olmedo Negrete, M.I. Paneva, W. Si, L. Wang, S. Wimpenny, B.R. Yates, Y. Zhang B.R. Yates, Y. Zhang University of California, San Diego, La Jolla, U.S.A. Zientek Fermi National Accelerator Laboratory, Batavia, U.S.A. S. Abdullin, M. Albrow, M. Alyari, G. Apollinari, A. Apresyan, A. Apyan, S. Bane S. Abdullin, M. Albrow, M. Alyari, G. Apollinari, A. Apresyan, A. Apyan, S. Banerjee, L.A.T. Bauerdick, A. Beretvas, D. Berry, J. Berryhill, P.C. Bhat, K. Burkett, J.N. But- ler, A. Canepa, G.B. Cerati, H.W.K. Cheung, F. Chlebana, M. Cremonesi, J. Duarte, V.D. Elvira, J. Freeman, Z. Gecse, E. Gottschalk, L. Gray, D. Green, S. Grünendahl, O. Gutsche, AllisonReinsvold Hall, J. Hanlon, R.M. Harris, S. Hasegawa, R. Heller, J. Hirschauer, B. Jayatilaka, S. Jindariani, M. Johnson, U. Joshi, T. Klijnsma, B. Klima, M.J. Kortelainen, B. Kreis, S. Lammel, J. Lewis, D. Lincoln, R. Lipton, M. Liu, T. Liu, J. Lykken, K. Maeshima, J.M. Marraﬃno, D. Mason, P. McBride, P. Merkel, S. Mrenna, S. Nahn, V. O’Dell, V. Papadimitriou, K. Pedro, C. Pena, G. Rakness, F. Ravera, L. Ristori, B. Schneider, E. Sexton-Kennedy, N. Smith, A. Soha, W.J. Spalding, L. Spiegel, S. Stoynev, J. Strait, N. Strobbe, L. Taylor, S. Tkaczyk, N.V. Tran, L. Uplegger, E.W. Vaandering, C. Vernieri, R. Vidal, M. Wang, H.A. Weber J. Strait, N. Strobbe, L. Taylor, S. Tkaczyk, N.V. Tran, L. Uplegger, E.W. Vaander C. Vernieri, R. Vidal, M. Wang, H.A. Weber C. Vernieri, R. Vidal, M. Wang, H.A. Weber University of Florida, Gainesville, U.S.A. , y, , , C , C , V C p , F. Errico, R.D. Field, S.V. Gleyzer, B.M. Joshi, M. Kim, J. Konigsberg, A. Korytov, K.H. Lo, P. Ma, K. Matchev, N. Menendez, G. Mitselmakher, D. Rosenzweig, K. Shi, J. Wang, S. Wang, X. Zuo K.H. Lo, P. Ma, K. Matchev, N. Menendez, G. Mitselmakher, D. Rosenzweig, K. Shi, J. Wang, S. Wang, X. Zuo Florida International University, Miami, U.S.A. Y.R. Joshi Florida State University, Tallahassee, U.S.A. T. Adams, A. Askew, S. Hagopian, V. Hagopian, K.F. Johnson, R. Khurana, T. Kolberg, G. Martinez, T. Perry, H. Prosper, C. Schiber, R. Yohay, J. Zhang T. Adams, A. Askew, S. Hagopian, V. Hagopian, K.F. Johnson, R. Khur G. Martinez, T. Perry, H. Prosper, C. Schiber, R. Yohay, J. Zhang Florida Institute of Technology, Melbourne, U.S.A. M.M. Baarmand, M. Hohlmann, D. Noonan, M. Rahmani, M. Saunders, F. Yumiceva University of Illinois at Chicago (UIC), Chicago, U.S.A. M.R. Adams, L. Apanasevich, R.R. Betts, R. Cavanaugh, X. Chen, S. Dittmer, O. Evdoki- mov, C.E. Gerber, D.A. Hangal, D.J. Hofman, K. Jung, C. Mills, T. University of California, San Diego, La Jolla, U.S.A. University of California, San Diego, La Jolla, U.S.A. J.G. Branson, P. Chang, S. Cittolin, S. Cooperstein, N. Deelen, M. Derdzinski, R. Gerosa, D. Gilbert, B. Hashemi, D. Klein, V. Krutelyov, J. Letts, M. Masciovecchio, S. May, S. Padhi, M. Pieri, V. Sharma, M. Tadel, F. Würthwein, A. Yagil, G. Zevi Della Porta University of California, Santa Barbara — Department of Physics, Santa Barbara, U.S.A. N. Amin, R. Bhandari, C. Campagnari, M. Citron, V. Dutta, M. Franco Sevilla, J. Incan- dela, B. Marsh, H. Mei, A. Ovcharova, H. Qu, J. Richman, U. Sarica, D. Stuart, S. Wang – 36 – California Institute of Technology, Pasadena, U.S.A. California Institute of Technology, Pasadena, U.S.A. D. Anderson, A. Bornheim, O. Cerri, I. Dutta, J.M. Lawhorn, N. Lu, J. Mao, H.B. Newman, T.Q. Nguyen, J. Pata, M. Spiropulu, J.R. Vlimant, S. Xie, Z. Zhang, R.Y. Zhu D. Anderson, A. Bornheim, O. Cerri, I. Dutta, J.M. Lawhorn, N. Lu, J. Mao, H.B. Newma D. Anderson, A. Bornheim, O. Cerri, I. Dutta, J.M. Lawhorn, N. Lu, J. Mao, H.B. Newman, T.Q. Nguyen, J. Pata, M. Spiropulu, J.R. Vlimant, S. Xie, Z. Zhang, R.Y. Zhu T.Q. Nguyen, J. Pata, M. Spiropulu, J.R. Vlimant, S. Xie, Z. Zhang, R.Y. Zhu Carnegie Mellon University, Pittsburgh, U.S.A. M.B. Andrews, T. Ferguson, T. Mudholkar, M. Paulini, M. Sun, I. Vorobiev, M. Weinberg Carnegie Mellon University, Pittsburgh, U.S.A. Carnegie Mellon University, Pittsburgh, U.S.A. M B A d T F T M dh lk M P li i arnegie Mellon University, Pittsburgh, U.S.A. .B. Andrews, T. Ferguson, T. Mudholkar, M. Paulini, M. Sun, I. Vorobiev, M. Weinberg M.B. Andrews, T. Ferguson, T. Mudholkar, M. Paulini, M. Sun, I. Vorobiev, M. Weinberg University of Colorado Boulder, Boulder, U.S.A. University of Colorado Boulder, Boulder, U.S.A. J.P. Cumalat, W.T. Ford, E. MacDonald, T. Mulholland, R. Patel, A. Perloﬀ, K. Stenson, K.A. Ulmer, S.R. Wagner Cornell University, Ithaca, U.S.A. JHEP12(2020)082 J. Alexander, Y. Cheng, J. Chu, A. Datta, A. Frankenthal, K. Mcdermott, J J. Alexander, Y. Cheng, J. Chu, A. Datta, A. Frankenthal, K. Mcdermott, J.R. Patterson, D Q h A R d S M T Z T J Th P Witti h M Zi t k J. Alexander, Y. Cheng, J. Chu, A. Datta, A. Frankenthal, K. Mcdermott, J.R. Patterson, D. Quach, A. Ryd, S.M. Tan, Z. Tao, J. Thom, P. Wittich, M. Zientek D. Quach, A. Ryd, S.M. Tan, Z. Tao, J. Thom, P. Wittich, M. University of California, San Diego, La Jolla, U.S.A. Roy, M.B. Tonjes, N. Varelas, J. Viinikainen, H. Wang, X. Wang, Z. Wu – 37 – The University of Iowa, Iowa City, U.S.A. M. Alhusseini, B. Bilki56, W. Clarida, K. Dilsiz73, S. Durgut, R.P. Gandrajula, M. Hayt- myradov, V. Khristenko, O.K. Köseyan, J.-P. Merlo, A. Mestvirishvili74, A. Moeller, J. Nachtman, H. Ogul75, Y. Onel, F. Ozok76, A. Penzo, C. Snyder, E. Tiras, J. Wetzel The University of Iowa, Iowa City, U.S.A. Johns Hopkins University, Baltimore, U.S.A. p y, , B. Blumenfeld, A. Cocoros, N. Eminizer, A.V. Gritsan, W.T. Hung, S. Kyriacou, P. Mak- simovic, J. Roskes, M. Swartz The University of Kansas, Lawrence, U.S.A. The University of Kansas, Lawrence, U.S.A. C. Baldenegro Barrera, P. Baringer, A. Bean, S. Boren, J. Bowen, A. Bylinkin, T. Isidori, S. Khalil, J. King, G. Krintiras, A. Kropivnitskaya, C. Lindsey, D. Majumder, W. Mcbrayer, N. Minafra, M. Murray, C. Rogan, C. Royon, S. Sanders, E. Schmitz, J.D. Tapia Takaki, Q. Wang, J. Williams, G. Wilson JHEP12(2020)082 The Ohio State University, Columbus, U.S.A. The Ohio State University, Columbus, U.S.A. J. Alimena, B. Bylsma, L.S. Durkin, B. Francis, C. Hill, W. Ji, A. Lefeld, T.Y. Ling, B.L. Winer Princeton University, Princeton, U.S.A. G. Dezoort, P. Elmer, J. Hardenbrook, N. Haubrich, S. Higginbotham, A. Kalogeropoulos, S. Kwan, D. Lange, M.T. Lucchini, J. Luo, D. Marlow, K. Mei, I. Ojalvo, J. Olsen, C. Palmer, P. Piroué, J. Salfeld-Nebgen, D. Stickland, C. Tully, Z. Wang G. Dezoort, P. Elmer, J. Hardenbrook, N. Haubrich, S. Higginbotham, A. Kalogeropoulo S. Kwan, D. Lange, M.T. Lucchini, J. Luo, D. Marlow, K. Mei, I. Ojalvo, J. Olsen, C. Palmer, P. Piroué, J. Salfeld-Nebgen, D. Stickland, C. Tully, Z. Wang C. Palmer, P. Piroué, J. Salfeld-Nebgen, D. Stickland, C. Tully, Z. W Northwestern University, Evanston, U.S.A. Northwestern University, Evanston, U.S.A. S. Bhattacharya, J. Bueghly, T. Gunter, K.A. Hahn, N. Odell, M.H. Schmitt, K. Sung, M. Trovato, M. Velasco University of Notre Dame, Notre Dame, U.S.A. R. Bucci, N. Dev, R. Goldouzian, M. Hildreth, K. Hurtado Anampa, C. Jessop, D.J. Kar- mgard, K. Lannon, W. Li, N. Loukas, N. Marinelli, I. Mcalister, F. Meng, C. Mueller, Y. Musienko37, M. Planer, R. Ruchti, P. Siddireddy, G. Smith, S. Taroni, M. Wayne, A. Wightman, M. Wolf, A. Woodard JHEP12(2020)082 Kansas State University, Manhattan, U.S.A. Kansas State University, Manhattan, U.S.A. S. Duric, A. Ivanov, K. Kaadze, D. Kim, Y. Maravin, D.R. Mendis, T. Mitchell, A. Modak, A. Mohammadi Lawrence Livermore National Laboratory, Livermore, U.S.A. F. Rebassoo, D. Wright University of Maryland, College Park, U.S.A. University of Maryland, College Park, U.S.A. A. Baden, O. Baron, A. Belloni, S.C. Eno, Y. Feng, N.J. Hadley, S. Jabeen, G.Y. Jeng, R.G. Kellogg, J. Kunkle, A.C. Mignerey, S. Nabili, F. Ricci-Tam, M. Seidel, Y.H. Shin, A. Skuja, S.C. Tonwar, K. Wong Massachusetts Institute of Technology, Cambridge, U.S.A. D. Abercrombie, B. Allen, A. Baty, R. Bi, S. Brandt, W. Busza, I.A. Cali, M. D’Alfonso, G. Gomez Ceballos, M. Goncharov, P. Harris, D. Hsu, M. Hu, M. Klute, D. Kovalskyi, Y.-J. Lee, P.D. Luckey, B. Maier, A.C. Marini, C. Mcginn, C. Mironov, S. Narayanan, X. Niu, C. Paus, D. Rankin, C. Roland, G. Roland, Z. Shi, G.S.F. Stephans, K. Sumorok, K. Tatar, D. Velicanu, J. Wang, T.W. Wang, B. Wyslouch University of Minnesota, Minneapolis, U.S.A. R.M. Chatterjee, A. Evans, S. Guts†, P. Hansen, J. Hiltbrand, Sh. Jain, Y. Kubota, Z. Lesko, J. Mans, M. Revering, R. Rusack, R. Saradhy, N. Schroeder, M.A. Wadud University of Mississippi, Oxford, U.S.A. J.G. Acosta, S. Oliveros University of Nebraska-Lincoln, Lincoln, U.S.A. K. Bloom, S. Chauhan, D.R. Claes, C. Fangmeier, L. Finco, F. Golf, R. Kamalieddin, I. Kravchenko, J.E. Siado, G.R. Snow†, B. Stieger, W. Tabb State University of New York at Buﬀalo, Buﬀalo, U.S.A. G A l C H i t I I h ili A Kh hil C M L D N A P k State University of New York at Buﬀalo, Buﬀalo, U.S.A. State University of New York at Buﬀalo, Buﬀalo, U.S.A. G. Agarwal, C. Harrington, I. Iashvili, A. Kharchilava, C. McLean, D. Nguyen, A. Parker, J. Pekkanen, S. Rappoccio, B. Roozbahani G. Agarwal, C. Harrington, I. Iashvili, A. Kharchilava, C. McLean, D. Nguyen, A. Parke J. Pekkanen, S. Rappoccio, B. Roozbahani – 38 – Northeastern University, Boston, U.S.A. Northeastern University, Boston, U.S.A. G. Alverson, E. Barberis, C. Freer, Y. Haddad, A. Hortiangtham, G. Madigan, B. Marzoc- chi, D.M. Morse, T. Orimoto, L. Skinnari, A. Tishelman-Charny, T. Wamorkar, B. Wang, A. Wisecarver, D. Wood University of Puerto Rico, Mayaguez, U.S.A. S. Malik, S. Norberg University of Puerto Rico, Mayaguez, U.S.A. S. Malik, S. Norberg Purdue University, West Lafayette, U.S.A. A. Barker, V.E. Barnes, S. Das, L. Gutay, M. Jones, A.W. Jung, A. Khatiwada, B. Ma- hakud, D.H. Miller, G. Negro, N. Neumeister, C.C. Peng, S. Piperov, H. Qiu, J.F. Schulte, N. Trevisani, F. Wang, R. Xiao, W. Xie Purdue University Northwest, Hammond, U.S.A. T. Cheng, J. Dolen, N. Parashar Rice University, Houston, U.S.A. Rice University, Houston, U.S.A. U. Behrens, K.M. Ecklund, S. Freed, F.J.M. Geurts, M. Kilpatrick, Arun Kumar, W. Li, B.P. Padley, R. Redjimi, J. Roberts, J. Rorie, W. Shi, A.G. Stahl Leiton, Z. Tu, A. Zhang University of Rochester, Rochester, U.S.A. University of Rochester, Rochester, U.S.A. A. Bodek, P. de Barbaro, R. Demina, J.L. Dulemba, C. Fallon, T. Ferbel, M. Galanti, A. Garcia-Bellido, O. Hindrichs, A. Khukhunaishvili, E. Ranken, R. Taus Rutgers, The State University of New Jersey, Piscataway, U.S.A. B. Chiarito, J.P. Chou, A. Gandrakota, Y. Gershtein, E. Halkiadakis, A. Hart, M. Heindl, E. Hughes, S. Kaplan, I. Laﬂotte, A. Lath, R. Montalvo, K. Nash, M. Osherson, H. Saka, S. Salur, S. Schnetzer, S. Somalwar, R. Stone, S. Thomas B. Chiarito, J.P. Chou, A. Gandrakota, Y. Gershtein, E. Halkiadakis, A. Hart, M. Heind E. Hughes, S. Kaplan, I. Laﬂotte, A. Lath, R. Montalvo, K. Nash, M. Osherson, H. Sak S Salur S Schnetzer S Somalwar R Stone S Thomas E. Hughes, S. Kaplan, I. Laﬂotte, A. Lath, R. Montalvo, K. Nash, M. Osherson, H. Saka, S. Salur, S. Schnetzer, S. Somalwar, R. Stone, S. Thomas – 39 – University of Tennessee, Knoxville, U.S.A. H. Acharya, A.G. Delannoy, S. Spanier University of Tennessee, Knoxville, U.S.A. H. Acharya, A.G. Delannoy, S. Spanier Texas A&M University, College Station, U.S.A. O. Bouhali77, M. Dalchenko, M. De Mattia, A. Delgado, S. Dildick, R. Eusebi, J. Gilmor O. Bouhali77, M. Dalchenko, M. De Mattia, A. Delgado, S. Dildick, R. Eusebi, J. Gilmore, T. Huang, T. Kamon78, S. Luo, S. Malhotra, D. Marley, R. Mueller, D. Overton, L. Perniè, D. Rathjens, A. Safonov Texas Tech University, Lubbock, U.S.A. N. Akchurin, J. Damgov, F. De Guio, S. Kunori, K. Lamichhane, S.W. Lee, T. Mengke, S. Muthumuni, T. Peltola, S. Undleeb, I. Volobouev, Z. Wang, A. Whitbeck JHEP12(2020)082 Vanderbilt University, Nashville, U.S.A. S. Greene, A. Gurrola, R. Janjam, W. Johns, C. Maguire, A. Melo, H. Ni, K. Padeken, F. Romeo, P. Sheldon, S. Tuo, J. Velkovska, M. Verweij F. University of Puerto Rico, Mayaguez, U.S.A. S. Malik, S. Norberg Romeo, P. Sheldon, S. Tuo, J. Velkovska, M. Verweij University of Virginia, Charlottesville, U.S.A. M.W. Arenton, P. Barria, B. Cox, G. Cummings, J. Hakala, R. Hirosky, M. Joyce, A. Ledovskoy, C. Neu, B. Tannenwald, Y. Wang, E. Wolfe, F. Xia Wayne State University, Detroit, U.S.A. R. Harr, P.E. Karchin, N. Poudyal, J. Sturdy, P. Thapa Wayne State University, Detroit, U.S.A. R. Harr, P.E. Karchin, N. Poudyal, J. Sturdy, P. Thapa University of Wisconsin — Madison, Madison, WI, U.S.A. T. Bose, J. Buchanan, C. Caillol, D. Carlsmith, S. Dasu, I. De Bruyn, L. Dodd, F. Fiori, C. Galloni, B. Gomber79, H. He, M. Herndon, A. Hervé, U. Hussain, P. Klabbers, A. Lanaro, A. Loeliger, K. Long, R. Loveless, J. Madhusudanan Sreekala, D. Pinna, T. Ruggles, A. Savin, V. Sharma, W.H. Smith, D. Teague, S. Trembath-reichert, N. Woods University of Wisconsin — Madison, Madison, WI, U.S.A. T. Bose, J. Buchanan, C. Caillol, D. Carlsmith, S. Dasu, I. De Bruyn, L. Dodd, F. Fiori, C. Galloni, B. Gomber79, H. He, M. Herndon, A. Hervé, U. Hussain, P. Klabbers, A. Lanaro, A. Loeliger, K. Long, R. Loveless, J. Madhusudanan Sreekala, D. Pinna, T. Ruggles, C. Galloni, B. Gomber , H. He, M. Herndon, A. Hervé, U. Hussain, P. Klabbers, A. Lanaro, A. Loeliger, K. Long, R. Loveless, J. Madhusudanan Sreekala, D. Pinna, T. Ruggles, A S Sh S h S b h h d †: Deceased 1: Also at Vienna University of Technology, Vienna, Austria 1: Also at Vienna University of Technology, Vienna, Austria 2: Also at IRFU, CEA, Université Paris-Saclay, Gif-sur-Yvette, France 2: Also at IRFU, CEA, Université Paris-Saclay, Gif-sur-Yvette, France 3: Also at Universidade Estadual de Campinas, Campinas, Brazil 3: Also at Universidade Estadual de Campinas, Campinas, Brazil 4: Also at Federal University of Rio Grande do Sul, Porto Alegre, Brazil 4: Also at Federal University of Rio Grande 5: Also at UFMS, Nova Andradina, Brazil 6: Also at Universidade Federal de Pelotas, Pelotas, Brazil 6: Also at Universidade Federal de Pelotas, Pelotas, Brazil 7: Also at Université Libre de Bruxelles, Bruxelles, Belgium 7: Also at Université Libre de Bruxelles, Bruxelles, Belgium 8: Also at University of Chinese Academy of Sciences, Beijing, China 8: Also at University of Chinese Academy of Sciences, Beijing, China 9: Also at Institute for Theoretical and Experimental Physics named by A.I. University of Puerto Rico, Mayaguez, U.S.A. S. Malik, S. Norberg Alikhanov of NRC ‘Kurchatov Institute’, Moscow, Russia 9: Also at Institute for Theoretical and Exp ‘Kurchatov Institute’, Moscow, Russia 10: Also at Joint Institute for Nuclear Research, Dubna, Russia 10: Also at Joint Institute for Nuclear Research, Dubna, Russia 10: Also at Joint Institute for Nuclear Res 11: Also at Suez University, Suez, Egypt 11: Also at Suez University, Suez, Egypt 12: Now at British University in Egypt, Cairo, Egypt 12: Now at British University in Egypt, Cairo, Egypt 13: Also at Purdue University, West Lafayette, U.S.A. 13: Also at Purdue University, West Lafayette, U.S.A. 14: Also at Université de Haute Alsace, Mulhouse, France 14: Also at Université de Haute Alsace, Mulhouse, France 15: Also at Tbilisi State University, Tbilisi, Georgia 15: Also at Tbilisi State University, Tbilisi, Georgia 16: Also at Erzincan Binali Yildirim University, Erzincan, Turkey 16: Also at Erzincan Binali Yildirim University, Erzincan, Turkey 17: Also at CERN, European Organization for Nuclear Research, Geneva, Switzerland – 40 – 18: Also at RWTH Aachen University, III. University of Puerto Rico, Mayaguez, U.S.A. S. Malik, S. Norberg Physikalisches Institut A, Aachen, Germany 19: Also at University of Hamburg, Hamburg, Germany 20: Also at Brandenburg University of Technology, Cottbus, Germany 21: Also at Institute of Physics, University of Debrecen, Debrecen, Hungary, Debrecen, Hunga 22: Also at Institute of Nuclear Research ATOMKI, Debrecen, Hungary 23: Also at MTA-ELTE Lendület CMS Particle and Nuclear Physics Group, Eötvös Loránd University, Budapest, Hungary, Budapest, Hungary Also at IIT Bhubaneswar, Bhubaneswar, India, 24: Also at IIT Bhubaneswar, Bhubaneswar, India, Bhubaneswar, India 25: Also at Institute of Physics, Bhubaneswar, India 26: Also at Shoolini University, Solan, India 27: Also at University of Visva-Bharati, Santiniketan, India 27: Also at University of Visva-Bharati, Santin JHEP12(2020)082 28: Also at Isfahan University of Technology, Isfahan, Iran, Isfahan, Iran b 29: Now at INFN Sezione di Baria, Università di Barib, Politecnico di Baric, Bari, Italy 30: Also at Italian National Agency for New Technologies, Energy and Sustainable Economic Development, Bologna, Italy 31: Also at Centro Siciliano di Fisica Nucleare e di Struttura Della Materia, Catania, Italy 32: Also at Scuola Normale e Sezione dell’INFN, Pisa, Italy Also at Riga Technical University, Riga, Latvia, 33: Also at Riga Technical University, Riga, Latvia, Riga, Latvia 34: Also at Malaysian Nuclear Agency, MOSTI, Kajang, Malaysia 35: Also at Consejo Nacional de Ciencia y Tecnología, Mexico City, Mexico 36: Also at Warsaw University of Technology, Institute of Electronic Systems, Warsaw, Polan 37: Also at Institute for Nuclear Research, Moscow, Russia 38: Now at National Research Nuclear University ‘Moscow Engineering Physics Institute’ (MEPhI), Moscow, Russia 39: Also at Institute of Nuclear Physics of the Uzbekistan Academy of Sciences, Tashkent, Uzbekistan 40: Also at St. Petersburg State Polytechnical University, St. Petersburg, Russia 41: Also at University of Florida, Gainesville, U.S.A. 42: Also at Imperial College, London, United Kingdom 43: Also at P.N. Lebedev Physical Institute, Moscow, Russia 44: Also at California Institute of Technology, Pasadena, U.S.A. University of Puerto Rico, Mayaguez, U.S.A. S. Malik, S. Norberg 45: Also at Budker Institute of Nuclear Physics, Novosibirsk, Russia 46: Also at Faculty of Physics, University of Belgrade, Belgrade, Serbia 47: Also at Università degli Studi di Siena, Siena, Italy 48: Also at INFN Sezione di Paviaa, Università di Paviab, Pavia, Italy, Pavia, Italy 49: Also at National and Kapodistrian University of Athens, Athens, Greece 50: Also at Universität Zürich, Zurich, Switzerland 51: Also at Stefan Meyer Institute for Subatomic Physics, Vienna, Austria, Vienna, Austria 52: Also at Burdur Mehmet Akif Ersoy University, BURDUR, Turkey 53: Also at Adiyaman University, Adiyaman, Turkey 54: Also at Şirnak University, Sirnak, Turkey 55: Also at Department of Physics, Tsinghua University, Beijing, China, Beijing, China 55: Also at Department of Physics, Tsinghua Un 56: Also at Beykent University, Istanbul, Turkey, Istanbul, Turkey 57: Also at Istanbul Aydin University, Application and Research Center for Advanced Studies (A & R C f Ad d S d ) I b l T k 57: Also at Istanbul Aydin University, Application and Research Center for Advanced Studies (App. & Res. Cent. for Advanced Studies), Istanbul, Turkey 57: Also at Istanbul Aydin University, Application and Research (App. & Res. Cent. for Advanced Studies), Istanbul, Turkey 58: Also at Mersin University, Mersin, Turkey 59: Also at Piri Reis University, Istanbul, Turkey 60: Also at Gaziosmanpasa University, Tokat, Turkey – 41 – 61: Also at Ozyegin University, Istanbul, Turkey 62: Also at Izmir Institute of Technology, Izmir, Turkey 63: Also at Marmara University, Istanbul, Turkey 64: Also at Kafkas University, Kars, Turkey 65: Also at Istanbul Bilgi University, Istanbul, Turkey 66: Also at Hacettepe University, Ankara, Turkey 67: Also at Vrije Universiteit Brussel, Brussel, Belgium 68: Also at School of Physics and Astronomy, University of Southampton, Southampton, United Kingdom 69: Also at IPPP Durham University, Durham, United Kingdom JHEP12(2020)082 71: Also at Bethel University, St. Paul, Minneapolis, U.S.A., St. Paul, U.S.A. 71: Also at Bethel University, St. Paul, Minneapolis, U.S.A., St. Paul, U.S.A. University of Puerto Rico, Mayaguez, U.S.A. S. Malik, S. Norberg 72: Also at Karamanoğlu Mehmetbey University, Karaman, Turkey 72: Also at Karamanoğlu Mehmetbey University, Karaman, Turkey 73: Also at Bingol University, Bingol, Turkey 74: Also at Georgian Technical University, Tbilisi, Georgia 74: Also at Georgian Technical University, Tbilisi, Georgia 74: Also at Georgian Technical University, Tb 75: Also at Sinop University, Sinop, Turkey 76: Also at Mimar Sinan University, Istanbul, Istanbul, Turkey 76: Also at Mimar Sinan University, Istanbul 76: Also at Mimar Sinan University, Istanbul, Istanbul, Turkey 77: Also at Texas A&M University at Qatar, Doha, Qatar 77: Also at Texas A&M University at Qatar, Doha, Qatar 78: Also at Kyungpook National University, Daegu, Korea, Daegu, Korea 78: Also at Kyungpook National University, Daegu, Korea, Daegu, Korea 79: Also at University of Hyderabad, Hyderabad, India 79: Also at University of Hyderabad, Hyderabad, India – 42 –
https://openalex.org/W2069581308	https://europepmc.org/articles/pmc4199438?pdf=render	English	null	External Built Residential Environment Characteristics that Affect Mental Health of Adults	Journal of urban health	2,014	cc-by	11,678	Journal of Urban Health: Bulletin of the New York Academy of Medicine, Vol. 91, No. 5 doi:10.1007/s11524-013-9852-5 Journal of Urban Health: Bulletin of the New York Academy of Medicine, Vol. 91, No. 5 doi:10.1007/s11524-013-9852-5 Journal of Urban Health: Bulletin of the New York Academy of Medicine, Vol. 91, No. 5 doi:10.1007/s11524-013-9852-5 * 2014 The Author(s). This article is published with open access at Springerlink.com * 2014 The Author(s). This article is published with open access at Springerlink.com * 2014 The Author(s). This article is published with open access at Springerlink.com External Built Residential Environment Characteristics that Affect Mental Health of Adults Charles Ochodo, D. M. Ndetei, W. N. Moturi, and J. O. Otieno ABSTRACT External built residential environment characteristics include aspects of building design such as types of walls, doors and windows, green spaces, density of houses per unit area, and waste disposal facilities. Neighborhoods that are characterized by poor quality external built environment can contribute to psychosocial stress and increase the likelihood of mental health disorders. This study investigated the relationship between characteristics of external built residential environment and mental health disorders in selected residences of Nakuru Municipality, Kenya. External built residential environment characteristics were investigated for 544 residents living in different residential areas that were categorized by their socioeconomic status. Medically validated interview schedules were used to determine mental health of residents in the respective neighborhoods. The relationship between characteristics of the external built residential environment and mental health of residents was determined by multivariable logistic regression analyses and chi-square tests. The results show that walling materials used on buildings, density of dwelling units, state of street lighting, types of doors, states of roofs, and states of windows are some built external residential environment characteristics that affect mental health of adult males and females. Urban residential areas that are characterized by poor quality external built environment substantially expose the population to daily stressors and inconveniences that increase the likelihood of developing mental health disorders. KEYWORDS External built environment, Urban residential areas, Socioeconomic status, Psychosocial stress, Mental health disorders Ochodo and Moturi are with the Department of Environmental Science, Egerton University, P.O Box 536, Egerton, Njoro, Kenya; Ndetei is with the Department of Psychiatry, University of Nairobi, Nairobi, Kenya; Ndetei is with the Africa Mental Health Foundation (AMHF), Nairobi, Kenya; Otieno is with the Department of Geography, Egerton University, Egerton, Njoro, Kenya. Correspondence: Charles Ochodo, Department of Environmental Science, Egerton University, P.O Box 536, Egerton, Njoro, Kenya. (E-mail: ochodoc@yahoo.com) INTRODUCTION External built residential environment encompasses the built surrounding of dwelling units that provide the setting for human activities and range in scale from immediate to neighborhood and large-scale civic surroundings. On the other hand, internal built residential environment refers to the man-made components of the inside locations of dwelling units. The potential for built environments to promote health outcomes is linked to their effectiveness in facilitating stress coping and restoration. Good designs of the built environment can reduce anxiety, lower blood pressure, and reduce pain. Psychologically unsupportive surroundings are linked to negative effects such as higher occurrence of delirium, depression, and greater need for pain medication.1,2 908 EXTERNAL BUILT RESIDENTIAL ENVIRONMENT CHARACTERISTICS 909 Facility design should promote and maintain user independence, acceptability, convenience, and controllability. There should be a degree of conﬁdentiality and privacy for occupants of a residential building particularly in toilets, bedrooms, and bathrooms.3 Access to external areas that promote a sense of normality through large windows, pleasant outdoor views, balconies, and courtyard areas are conducive to a healthy mental state.3,4 Residents who are continuously exposed to poor external built environment characteristics are likely to experience psychosocial stress that would in turn increase their likelihood of developing mental health disorders (Fig. 1). Mental health problems such as anxiety, depression, attention deﬁcit hyperactivity disorder, substance abuse, and aggressive behavior relate to the built environment particularly to poor urban planning and inadequate housing.5,6 Dilapidated housing: leaking pipes, peeling paint, cracks, and holes in walls or ceilings are stressors that can adversely affect mental health.7,8 Inequities in construction and maintenance of low income housing result in insufﬁcient housing, poor quality housing, overcrowding, higher levels of population density, and health problems.6,9 Housing designs that typically incorporate aspects of structural quality, maintenance, upkeep, amenities (such as private bathrooms, toilets, balconies), and prevention of physical hazards are particularly associated with good mental health. Humke et al. (1995) and Adam (2002) are of the view that insecurity often accompanies poor housing. INTRODUCTION The occupants of such housing are often low income earners who are concerned about housing tenure and constantly experience difﬁculties with repairs.10,11 Features of urban physical planning that enhance a sense of community include those which afford sufﬁcient privacy, ensure residents have easier access to amenities, parks, recreation facilities, offer pedestrian friendly spaces, provide streetscapes so that houses have views of the surrounding neighborhood, encourage open verandas, and low fences and restrict motor trafﬁc.12 The location of residential buildings in areas that limit availability and access to services can have External built environment characteristics (Independent variables) Poor state of roofs Insecure: doors, windows Narrow access pathways Lack of street lights Inadequate: garbage disposal, sewage disposal and water provision facilities Crowded housing, dissatisfaction with: green spaces and shopping facilities Mental Health Disorders (Dependent variable) Psychosocial stress (Intervening variable) FIGURE 1. Conceptual model. Mental Health Disorders Psychosocial stress (Intervening variable) (Dependent variable) External built environment characteristics (Independent variables) FIGURE 1. Conceptual model. External built environment characteristics (Independent variables) FIGURE 1. Conceptual model. 910 OCHODO ET AL. adverse impacts on both physical and mental health. Seclusion has the potential of generating anxiety over accessibility to services.13,14 adverse impacts on both physical and mental health. Seclusion has the potential of generating anxiety over accessibility to services.13,14 The immediate environment surrounding a building can inﬂuence health. Views of nature are particularly beneﬁcial and are associated with less anxiety.15,16 Esthetic appeal including plants and ﬂowers, landscaping, and artwork on walls are important for positive mental health outcomes. Residents appreciate good quality design as it makes them to be more relaxed.17 Relocation from low income neighborhoods to middle income areas is associated with enhanced mental health for both adults and children.13,18 The most frequently mentioned positive garden qualities are visual nature elements especially trees, greenery, ﬂowers, and water. Respondents strongly associate these natural features with restorative inﬂuences on their moods.19 By contrast, a characteristic that usually lowers open spaces' effectiveness in reducing stress is a predominance of hard space such as concrete or starkly built content.20 Green spaces can have a positive impact on mental health through providing places for meetings and social interaction and places for exercise and relaxation. INTRODUCTION These features include poor structural designs that often consist of EXTERNAL BUILT RESIDENTIAL ENVIRONMENT CHARACTERISTICS 911 mud, iron sheets, or cardboard walls; lack of access roads or poorly kept pathways; absence of sewerage facilities and uncontrolled garbage dumping; lack of electricity connections and inadequate water supply. All these attributes impact negatively on mental health. mud, iron sheets, or cardboard walls; lack of access roads or poorly kept pathways; absence of sewerage facilities and uncontrolled garbage dumping; lack of electricity connections and inadequate water supply. All these attributes impact negatively on mental health. STUDY SITE The location of the study was Nakuru Municipality which is one of the divisions within Nakuru County in Kenya. The municipality is located between longitudes 36 and 36°10′E and latitude 0°10 and 0°20′S at 1,859 m above the sea level. Nakuru is one of the fastest growing municipalities in Kenya; it is a modern and cosmopolitan town with several commercial centers. Its construction and transport industries are expanding rapidly.35 As a consequence of this growth, the population of the town has rapidly increased. According to the 2009 population census results, Nakuru town had a multiracial population of 224,743 people; 113,975 were males and 110,768 were females. The population density was 5,773 persons per square kilometer and there were 68,469 households.36 Most of the people live in high- density low income residential areas. Whereas many of the residents in the low income neighborhoods are tenants, a few of those in the middle income residential areas and most of the people in the high income areas have put up their own houses.37 METHOD This study was conducted through a cross-sectional social survey of residents in the low, middle, and high income neighborhoods of Nakuru Municipality in Kenya. Income levels of respective neighborhoods were determined by their type of housing that took into consideration types of walling and rooﬁng materials, design of houses, and amenities as prescribed by the Kenya National Bureau of Statistics.17The total number of households in the four residential areas that were sampled (Milimani, Section 58, Shabab, and Rhonda) was 61,220.36 A total of 544 respondents from the respective residential areas representing the low, middle, and high income neighborhoods were selected through a multistage random sampling process. Among other factors, inclusion criteria depended on respondents' continuous residence in a given neighborhood for more than 1 year. This is because studies have shown that after living in a given neighborhood for more than 1 year, residents' exhibit behavior and health states that demonstrate either successful or unsuccessful adjustment to prevailing environmental character- istics.38,39 Nearly equal numbers of the low, middle, and high income households were interviewed because of stratiﬁcation based on income levels. In all the three types of neighborhoods, different households that had at least one adult male or one adult female aged over 18 years or above were randomly selected. Only one member of each household who satisﬁed the preceding criteria was interviewed. In each category (males and females), a near equal number of respondents was interviewed. A total of 544 respondents from the respective residential areas representing the low, middle, and high income neighborhoods were selected through a multistage random sampling process. Among other factors, inclusion criteria depended on respondents' continuous residence in a given neighborhood for more than 1 year. This is because studies have shown that after living in a given neighborhood for more than 1 year, residents' exhibit behavior and health states that demonstrate either successful or unsuccessful adjustment to prevailing environmental character- istics.38,39 Nearly equal numbers of the low, middle, and high income households were interviewed because of stratiﬁcation based on income levels. In all the three types of neighborhoods, different households that had at least one adult male or one adult female aged over 18 years or above were randomly selected. Only one member of each household who satisﬁed the preceding criteria was interviewed. In each category (males and females), a near equal number of respondents was interviewed. INTRODUCTION They provide pleasant visual experiences and improve air quality and, if occupied by trees, act as barriers for reduction of environmental noise and screen off particulate pollutants apart from absorbing carbon dioxide that is responsible for global warming.21 Encouraging green space development promotes a sense of community, reduces violence, and improves mental health.21 , p The quality of residential units is often related to the neighborhood in which they are situated. Poor quality housing is often located in neighborhoods with multiple indicators of urban decay such as dilapidated roads, vandalized utilities, grafﬁti, and garbage that can be harmful to health.22 Weich et al. (2002) found that housing areas with deck access had higher levels of depression.23 Deck access residences are multi-family units in which individual apartments open directly onto open walkways connected to a central staircase. These walkways are often public anonymous spaces with little evidence of residential social control or ownership. Such environments can impact on both mental and physical health through a reduction in physical activity, increased anxiety, and social disorder among residents.23–26 Wandersman et al. (1990) and Leventhal et al. (2000) argue that neighborhood quality has mental health impacts on children and their families independent of household socioeconomic status.27,28 Overcrowded housing conditions contribute to higher mortality rates, stress, infectious disease risk, and poor childhood development.29 Overcrowding is a major problem of the built environment especially in slums and squalid environments. It arises from poverty, overpopulation, and inefﬁcient accommodation. It is a problem that has social and health effects.30 Chombant (1975) found that when each person had less than 8 to 10 m2 of space, then instances of physical illness and irregular behavior doubled compared to those in less crowded homes.31 Tripling of college students shows evidence of association between crowding and psychological distress. Similarly, residential density studies reveal evidence of psychological distress.32 Low class neighborhoods that were studied were often characterized by high population densities that exhibited tendencies of crowding both indoors and outdoors. Whereas residential density may not be related to psychological distress at initial occupancy; 6 months later, a positive correlation can be discerned.33,34 Low-cost housing in the developing countries like Kenya has peculiar features that may impact more on mental health than those of similar houses in the developed world. MENTAL HEALTH ASSESSMENT Mental health of respondents was assessed by the use of two instruments as follows: the Mini-International Neuropsychiatric Interview (M.I.N.I) plus and the Alcohol, Smoking, and Substance Involvement Screening Test. The M.I.N.I plus is a short structured diagnostic interview developed jointly by psychiatrists and clinicians for Diagnosis and Statistical Manual-IV and International Classiﬁcation of Diseases-10 psychiatric disorders. In contrast to the usual clinical interviews, structured diagnostic interviews allow comparison across clinical centers or different environ- mental settings. In addition, they have the ability to reduce variability in diagnosis.40–42 The Alcohol, Smoking, and Substance Involvement Screening Test is an interview schedule that is designed to detect and manage substance use and related problems in primary and general medical care settings. It has undergone signiﬁcant testing in three sequential phases (I, II, and III) to ascertain its psychometric properties.43 Mental health of respondents was assessed by the use of two instruments as follows: the Mini-International Neuropsychiatric Interview (M.I.N.I) plus and the Alcohol, Smoking, and Substance Involvement Screening Test. The M.I.N.I plus is a short structured diagnostic interview developed jointly by psychiatrists and clinicians for Diagnosis and Statistical Manual-IV and International Classiﬁcation of Diseases-10 psychiatric disorders. In contrast to the usual clinical interviews, structured diagnostic interviews allow comparison across clinical centers or different environ- mental settings. In addition, they have the ability to reduce variability in diagnosis.40–42 The Alcohol, Smoking, and Substance Involvement Screening Test h d l h d d d d b d METHOD The age range of respondents was 18–80 years, 274 were males and 270 were females. Out of the 274 males, 90, 96, and 88 belonged to the low, middle, and high income residential areas, respectively. As for the females, 95 were from the low income residential areas, 87 belonged to the middle income, and 88 were residents of the high income residential areas. Overall, 185 respondents were from the low income areas, 183 came from the middle income areas, and 176 belonged to the The age range of respondents was 18–80 years, 274 were males and 270 were females. Out of the 274 males, 90, 96, and 88 belonged to the low, middle, and high income residential areas, respectively. As for the females, 95 were from the low income residential areas, 87 belonged to the middle income, and 88 were residents of the high income residential areas. Overall, 185 respondents were from the low income areas, 183 came from the middle income areas, and 176 belonged to the 912 OCHODO ET AL. high income residential areas. Respondents were interviewed using structured questionnaires for environmental characteristics and validated lay mental health instruments. high income residential areas. Respondents were interviewed using structured questionnaires for environmental characteristics and validated lay mental health instruments. ETHICAL CONSIDERATIONS Ethical clearance for this study was obtained from the Kenyatta National Hospital Ethics Review Board. The nature and objectives of the study were explained to the respondents who voluntarily gave informed consent to be interviewed. Participants' names were coded to maintain conﬁdentiality. DATA ANALYSIS The external built environment characteristics across the three categories of residential areas, low, middle, and high income, were analyzed by the use of descriptive statistics, while chi-square tests (P≤0.05) were used to determine the prevalence of speciﬁc mental health disorders of respondents. Mental health was assessed on the basis of one having lived in the neighborhood of residence continuously for the immediate preceding 12 months or longer. Multivariable logistic regression models (P≤0.05) were used to assess the relationship between the urban built external environment characteristics and the likelihood of developing speciﬁc mental health disorders. To avoid possible confounding, sociodemographic characteristics (age, sex, marital status, number of children and number of occupants in a household, daily food expenditure per person, level of education, type of employment, and type of residence) were controlled for in the models. Characteristics of External Built Environment h l b l h Characteristics of External Built Environment The external built environment characteristics studied were as follows: types of walls, states of roofs, heights of roofs above the ground, states of windows, types of windows, types of doors, access pathways for walking or vehicles, and street EXTERNAL BUILT RESIDENTIAL ENVIRONMENT CHARACTERISTICS 913 lighting. Others were garbage disposal facilities, sewage disposal facilities, pit latrines, external bathrooms, availability of green spaces, and water facilities (Table 1). lighting. Others were garbage disposal facilities, sewage disposal facilities, pit latrines, external bathrooms, availability of green spaces, and water facilities (Table 1). Height of Roofs h l The low income houses had 98.5 % of roof heights that were less than 10 ft above the ground compared to the middle income ones that had 1.5 % roofs with such heights. The recommended height of roofs for residential houses should be at least 10 ft above the ground.44 Houses whose roof heights are less than 10 ft concentrate heat during the day and reduce indoor air circulation thus affecting ventilation. These attributes can engender chronic discomfort that may affect mental health of occupants. aHouses had private ﬂush toilets unlike the low income houses, 98.9 % of which lacked these toilets Windows and Doors Whereas all housing units in the middle and high income residential areas had the recommended sizes of windows, none of those in the low income neighborhood had these windows (Table 1). In addition, 99.1 % of the windows in the low income houses were insecure owing to their wooden content. This sharply contrasted with the middle and high income ones that had only 0.9 % and no insecure windows, respectively. p y Allied to windows were doors, 73.5 % of which were insecure in the low income houses. Only 20.3 and 6.1 %, respectively, of doors in the middle and high income houses were insecure. These categories of houses mostly had fences and hence did not experience the level of insecurity that characterized the low income ones. TABLE 1 Characteristics of the external built environment External environment characteristics (Residential areas) Low income % Middle income % High income % Housing units: With leaking roofs 74.1 13.7 12.2 With non-stone walls (mud/plastered mud) 100 0.0 0.0 Whose roof height Above the ground is less than 10 ft 98.5 1.5 0.0 Without recommended sizes of windows 100 0.0 0.0 Without secure windows 99.1 0.9 0.0 Without secure doors 73.5 20.3 6.1 Whose access pathways for walking are less than 2-m wide 66.3 31.8 2.0 Whose access pathways for vehicles are less than 6-m wide 78.2 20.9 0.9 With sufﬁcient street lighting 3.0 49.2 47.7 Without garbage disposal facilities 97.7 2.2 0.0 Without sewage disposal facilities 98.6 1.4 0.0 With inadequate pit latrines 33.8 38.6a 27.6a With inadequate bathrooms 99.6 0.4 0.0 Whose occupants walk beyond 500 m to access water 100 0.0 0.0 That are overcrowded 65.7 34.3 0.0 Residents who are dissatisﬁed with available green spaces 64.6 27.9 7.5 Residents who are dissatisﬁed with shopping facilities 47.4 7.4 45.3 aHouses had private ﬂush toilets unlike the low income houses, 98.9 % of which lacked these toilets TABLE 1 Characteristics of the external built environment Street Lighting Th ddl The middle income neighborhood had the best street lighting rate with 49.2 % of the streets sufﬁciently lit at night. Overall, lack of street lighting or inadequate lighting was a notable problem in residential areas. Whereas most areas had not had any street lighting programs; some of the areas that were initially lit were plunged into darkness as a result of lack of maintenance of lighting systems or vandalism. Pathways should be well lit at night. Dark or unlit roads are insecure in most neighborhoods. Such insecurity causes residents to worry when accessing the dark pathways. People are more likely to maximize the use of outdoor spaces if the neighborhood is perceived to be safe.48 Street lighting improvements show crime reduction effects and increase conﬁdence of residents at night.49 Access Pathways In the low income neighborhoods, 66.3 % of walking pathways were less than 2-m wide. As for access pathways for vehicles, 78.2 % of those in the low income residences were less than the recommended 6-m width, while in the high income residences, only 0.9 % of the vehicle access pathways were less than 6-m wide; this demonstrates the importance of motorized transport among high income residents. Access pathways for walking and vehicles should be at least 2 and 6-m wide, respectively.47These pathways provide escape routes for residents in cases of emergency. Smaller pathways leading into dwelling environments are an indicator of overcrowding and can prompt feelings of claustrophobia. Where roads are less than 6-m wide, by-passing oncoming vehicles is not possible for drivers. In this case, they have to reverse, give way, or stop at junctions. This can be a security risk in some neighborhoods and may lead to mental distress. Narrow walking pathways also expose residents to inconveniences from bicycle and motorcycle taxis that are frequent in the low income neighborhoods and are common means of transport in the middle income residences. External environment characteristics 914 OCHODO ET AL. The state, size, and types of windows and doors determine indoor temperatures and security of households. Wooden windows and doors were found to be less secure as they could be easily broken by solid objects. Families living in houses with such facilities were likely to experience anxiety and interrupted sleep. Lack of fencing, mud walls, and absence of bar grills on windows and doors increased vulnerability of occupants to external attacks. Households experienced discomfort particularly at night due to insecurity that they were exposed to owing to their types of doors and windows. Noise and intrusion from passers-by was not uncommon. Aggression among inhabitants was high in order to deter potential intruders. Suspicion of strangers was heightened and unwelcome remarks relayed in discourteous language were other manifestations of persistent discomfort. Small wooden windows also reduced lighting and ventilation, thus contributing to high internal temperatures during daylight. Lack of windows or small windows is associated with higher levels of anxiety, depression, and delirium. According to Verdeberber and Leather et al., windows that allow enough light into the room are linked to favorable mental health outcomes.45,46 Sewage Disposal Many of the low income housing units (98.6 %) had no effective sewage disposal facilities; they mostly relied on poorly constructed pit latrines. However, if they had enough water, adequate sewage disposal facilities linked to ﬂush toilets would have served them better since the few pit latrines they had easily ﬁlled up due to large numbers of tenants. The middle income dwelling units were mainly connected to the municipal sewer line, while the high income residences mostly used septic tanks. Major problems with the municipal sewer line were recurrent blockages that led to sewage overﬂowing on roads and walk ways. This overﬂow sometimes found its way into housing compounds. In addition, the environment had a persistent offensive smell that affected residents. Sewage disposal and treatment plants are recommended for all settlements with a population of 3,000 people or more. In residences where an integrated sewage system is not available, septic tanks should be used.47 Garbage Disposal In the low income residential areas where 97.7 % of households had no garbage disposal facilities, garbage was discarded anywhere outside the houses especially in pathways and drainage channels where wind would blow it to far off places. In the EXTERNAL BUILT RESIDENTIAL ENVIRONMENT CHARACTERISTICS 915 middle income residences, only 2.2 % of the households lacked garbage disposal facilities, those that had these facilities predominantly relied on paid up private garbage collectors. The latter were reported to be largely reliable. However, dogs and cats tore off the refuse paper bags placed outside residential gates thus scattering contents. Together with refuse from shopping premises that resulted from shoppers carelessly disposing wrappings such as polythene paper bags, the torn refuse bags contributed to unsightly garbage strewn in the vicinity of shopping premises and markets in the middle income residential areas. The high income households all had garbage disposal facilities. Dug out pits into which garbage was dumped and periodically burnt were the predominant means of disposing or reducing garbage. The burning of garbage emitted smoke that polluted air, thus inconveniencing neighbors. Complaints arising from this practice were, however, few due to low population in the high income areas. p p g Even though a few high income households used private garbage collectors, they were not the preferred method mainly because of possible security breaches that they could occasion. The municipal authorities hardly collected garbage in all the three residential areas; this happened in spite of the residents regularly paying land rates that entitled them to services such as garbage collection. Garbage or refuse collection facilities in residential areas should be adequate. Garbage should be removed from the dwelling units daily and disposed of in a sanitary manner. Disposal sites should be located on the leeward side and have a 100 m protection belt preferably comprising of trees.47 Garbage dump sites within or near residential units are an eyesore, emit offensive odors, and provide breeding grounds for organisms that cause or transmit diseases. Quite often in the low income neighborhoods where dump sites were located near houses, wind blew light garbage such as paper bags to the door steps. Dogs or cats could also carry certain rotting edible garbage to residences. In a number of instances, slum residents relieved themselves in garbage dump sites particularly when they were located close to pit latrines. Water Facilities The low income families had no water facilities within their housing units; water taps were either located next to their houses or at central points to serve many households. Residents who had to walk beyond 500 m to access water were all within this income group (Table 1). In the informal settlements, watering facilities 916 OCHODO ET AL. should be located within 500 m of dwelling units.47 With few far-placed water facilities, families spend long and stressful hours collecting water. Less water is drawn for household chores and this affects hygiene levels. Extended hours spent standing in the sun and dirty environments strain patience particularly where quantities of water decline and some residents foresee that they could miss it altogether. Usually, water brawls ensue at watering points that serve many households. Mothers who cannot afford maids worry about small children left back in the house on their own as they can suffer harm. Overcrowding d Overcrowding characterized 65.7 % of housing units in the low income neighborhoods. Overcrowding was also noted in 34.3 % of the middle income households; frequently, high-rise ﬂats were crammed into a quarter of an acre of land leaving no room for vehicle parks or wash lines. The high income housing units were singly placed in their compounds and hence did not exhibit the features associated with overcrowding; however, residents complained of loneliness and less social interaction even during periods of bereavement. Ironically, they were visited more by their friends and relatives from the low and middle income neighborhoods. In many instances, neighbors in the high income residential areas including next door ones were not familiar with each other, neither did they express willingness to know each other. Population density and the number of dwelling units in a given area determine overcrowding. For multifamily housing units, the recommended number of houses per hectare is as follows: low-density—50, medium-density—60, high-density—70, and extreme high-density—133.47 Informal settlements invariably exceed 400 dwelling units per hectare. 47 Overcrowding of dwelling units cause distress to residents as amenities such as latrines, bathrooms, and water are severely deﬁcient. There is lack of space for movement or airing clothes. Littering of neighbors' frontages occurs, noise pollution, internal insecurity, and recurrent disagreements, e.g., due to activities of children are common. Quite often, neighborhood brawls resulted. This created fear among minors to the extent that some of them suffered from severe separation anxiety disorder. They felt extremely insecure when their parents or guardians were away or when they had to leave home for other places like school. An association exists between poor mental health and lack of space within the home as well as lack of social space for interaction outside the home.50 Multi- occupation dwellings and ﬂats, particularly high rise ﬂats, are some of the housing risk factors associated with poor mental health.51 Shopping Facilities Shopping facilities were not a major source of distress to most of the residents in the middle income category where only 7.4 % of the households expressed dissatisfac- tion with the available shops. The low level of dissatisfaction with shopping facilities among the middle income residents was attributed to the availability of shops and supermarkets in the vicinity of their dwelling units. In addition, they had the necessary ﬁnancial capability to purchase most of the goods in standard packaging sizes. On the other hand, 47.4 % of residents in the low income neighborhood and 45.3 % of those in the high income residential areas expressed dissatisfaction with shopping facilities but for different reasons. The high income residents cited the far placed supermarkets and shops in town where they obtained most of their supplies as being inconvenient. While the low income residents who had shops and grocery kiosks very close to their door steps could not ﬁnd certain items they required; the usual packaging of commodities like ﬂour or sugar also appeared to be unaffordable to some of them. Shopkeepers therefore had to repackage items in smaller units. Green Spaces In the low income residential areas, 64.6 % of the population was dissatisﬁed with available green spaces compared to 27.9 % of those in the middle income and only 7.5 % of the high income ones. In the middle and high income neighborhoods, concrete compounds replaced green spaces while in the low income residential areas, open grounds that generated choking dust in the dry season or gave rise to muddy pathways dominated the small restricted compounds. Green spaces provide esthetic appeal; grass and ﬂower gardens denote tranquility. According to Ulrich (2002), viewing of plants and ﬂowers ameliorates stress within 5 minutes or less.52 Viewing nature for longer periods helps to calm residents and can foster improvement in mental health. Green vegetation was found to promote EXTERNAL BUILT RESIDENTIAL ENVIRONMENT CHARACTERISTICS 917 positive attention inﬂuence on both females and males even though females manifested stronger inﬂuence.52 positive attention inﬂuence on both females and males even though females manifested stronger inﬂuence.52 Prevalence of Mental Health Disorders among the Adult Residents of the Low, Middle, and High Income Neighborhoods 918 TABLE 2 Prevalence of mental health disorders among adults in the low, middle, and high income residential areas Mental health disorder (Residential area) Chi-square (χ2) P≤ 0.05 Low income % Middle income % High income % Depression 31.4 24 17.6 9.240 0.01 Dysthymia 3.2 2.7 8.5 8.064 0.02 BMD 2.2 1.1 3.4 2.234 0.33 Panic disorder 8.1 1.6 9.7 10.891 0.004 PTSD 8.7 1.6 6.3 8.944 0.01 AA&D 0.5 6.0 2.8 9.163 0.01 GAD 7.0 1.6 14.2 20.431 0.00 BMD bipolar mood disorder, PTSD post-traumatic stress disorder, AA&D alcohol abuse and dependence, GAD generalized anxiety disorder TABLE 2 Prevalence of mental health disorders among adults in the low, middle, and high ncome residential areas BMD bipolar mood disorder, PTSD post-traumatic stress disorder, AA&D alcohol abuse and dependence, GAD generalized anxiety disorder suicidal attempts or actual suicide. Depression affects more women than men. Major depression is manifested by a combination of symptoms that are severe enough to interfere with the ability to work, sleep, eat, and enjoy pleasurable activities. It can occur once, twice, or several times in a lifetime.53 suicidal attempts or actual suicide. Depression affects more women than men. Major depression is manifested by a combination of symptoms that are severe enough to interfere with the ability to work, sleep, eat, and enjoy pleasurable activities. It can occur once, twice, or several times in a lifetime.53 Bipolar Mood Disorder The overall incidence of this disorder in the three neighborhoods was 2.2 % similar to that in the low income class. However, the middle income residents had a lower level of 1.1 % unlike the high income ones who expressed the highest level of occurrence at 3.4 %. There was no signiﬁcant variation in the prevalence of this disorder across the three categories of residential areas neither did any two of these residential areas express a signiﬁcant difference in the level of bipolar mood disorder. Likewise, there was no signiﬁcant variation in the disorder between males and females who manifested a rating of 1.8 and 2.6 %, respectively. Bipolar mood disorder was signiﬁcantly associated with one characteristic of the external environment. Residents whose house doors were made of steel were found to be 0.113 times less susceptible to bipolar mood disorder than those whose doors were wooden (Table 3). The signiﬁcant sociodemographic factors that were associated with bipolar mood disorder were being employed or unemployed and type of residence. Prevalence of Mental Health Disorders among the Adult Residents of the Low, Middle, and High Income Neighborhoods Depression There was a signiﬁcant difference in the levels of depression across the three income groups (Table 2). However, between the low and middle income neighborhoods, the level of depression was not signiﬁcantly different (χ2=2.452, P= 0.117). Likewise, between the middle and high income residential areas, there was no signiﬁcant difference in the prevalence of depression (χ2=2.244, P=0.134). The low income neighborhoods had signiﬁcantly higher levels of depression than the high income residential areas (χ2=9.164, P=0.002). This implies that environmental characteristics that are associated with depression were signiﬁcantly different in the two residential set ups. Of the females interviewed, 29.3 % expressed feelings consistent with depression compared to 19.7 % of the men. On average, 24.5 % prevalence of depression was recorded among the residents of the three neighborhoods. g Type of walling material, access pathways for walking, and availability of shopping facilities were signiﬁcantly associated with depression (Table 3). Respon- dents who lived in houses built of stones were 0.093 times less likely to experience depression in comparison to those who lived in non-stone wall houses. Residents who had access foot paths measuring more than 2-m wide were 0.592 times less likely to develop depression. Those who expressed dissatisfaction with shopping facilities were 2.307 times more likely to suffer from depression compared to those who were satisﬁed with these facilities. The signiﬁcant confounders that were associated with depression were sex, type of residence (tenant or own house), and level of education. According to Ndetei et al., depression is a common mental health disorder that affects people of all races, socioeconomic, and cultural backgrounds.53 It is important not only because of lost productivity and the pain it causes to the affected individuals and their families but also due to its cost to human life through OCHODO ET AL. Prevalence of Mental Health Disorders among the Adult Residents of the Low, Middle, and High Income Neighborhoods Bipolar disorder also known as manic depressive disorder is a psychiatric diagnosis that describes a category of mood disorders deﬁned by the presence of one or more episodes of abnormally elevated energy levels, cognition and mood with or without one or more depressive episodes. The elevated moods are clinically referred to as mania or if milder, hypomania. Persons affected by manic episodes commonly experience depression.54 Panic Disorder Panic disorder was highest among the high and low income residents: 9.7 and 8.1 %, respectively, while males and females were equally affected. Overall, 6.4 % of the respondents experienced the disorder. Like the low and middle income residential areas which indicated a signiﬁcant difference in the levels of panic disorder (χ2=8.275, P=0.004), the middle and high income residential areas also had a signiﬁcant variation in the prevalence of this disorder EXTERNAL BUILT RESIDENTIAL ENVIRONMENT CHARACTERISTICS 919 TABLE 3 Relationship between the external built environment characteristics and mental health disorders: multivariable logistic regression analyses Mental health disorders Characteristics of external built environment Exp (β) 95.0 % CI for Exp (β) P ≤ 0.05 Lower Upper Depression Type of walling material 0.093 0.009 0.967 0.047 Access pathways for walking 0.592 0.354 0.990 0.046 Shopping facilities 2.307 1.049 3.955 0.036 Bipolar mood disorder Type of doors 0.113 0.016 0.772 0.026 Panic disorder State of roof 0.294 0.106 0.8I7 0.019 State of windows 8.344 1.975 35.248 0.004 Functional bathrooms 0.089 0.009 0.918 0.042 Dissatisﬁed with available green spaces 4.262 1.215 14.945 0.024 Post-traumatic stress disorder State of roof 0.353 0.137 0.911 0.031 Distance traveled to access water 4.760 1.791 12.368 0.002 Alcohol abuse and dependence Type of doors 0.081 0.011 0.600 0.014 Access pathways for walking 386.4256 5.578 26767.992 0.006 Lack of street lighting 15.623 1.921 127.059 0.010 Dissatisfaction with available green spaces 0.002 0.000 0.125 0.003 Generalized anxiety disorder State of roof 0.322 0.122 0.850 0.022 Distance traveled to access water 5.327 1.855 15.299 0.002 All models were adjusted for age, sex, marital status, number of children, number of occupants in a house, level of education, employment, and type of residence TABLE 3 Relationship between the external built environment characteristics and menta health disorders: multivariable logistic regression analyses All models were adjusted for age, sex, marital status, number of children, number of occupants in a house, level of education, employment, and type of residence (χ2=10.969, P=0.001). Prevalence of Mental Health Disorders among the Adult Residents of the Low, Middle, and High Income Neighborhoods However, between the low and middle income residential areas, there was a signiﬁcant variation in the levels of PTSD (χ2=9.231, P=0.002), likewise, a signiﬁcant variation was recorded between the middle and high income residential areas (χ2= 5.089, P=0.024). Males and females expressed similar patterns of PTSD: 5.8 and 5.2 %, respectively. Living in houses that had non-leaking roofs reduced susceptibility to PTSD by 0.353 times in relation to living in houses with leaking roofs. Residents who traveled more than 500 m to access household water increased their chances of experiencing PTSD by 4.760 units compared to those who accessed water within a distance of 500 m. PTSD is a severe anxiety disorder that develops after exposure to any event that results into psychological trauma. This event may involve the threat of death to oneself or someone else or to ones or someone else's physical, sexual, or psychological integrity overwhelming the individual's ability to cope.40,56,57 Alcohol and Drug Abuse Across the three categories of income, alcohol abuse and dependence was lowest in the low income residential areas (0.5 %) and highest in the middle income residential areas at 6.0 %. Whereas there was a signiﬁcant variation in the level of alcohol abuse and dependence between the low and middle income residential areas (χ2=8.727, P=0.003), there was no signiﬁcant variation in the prevalence of this disorder between the low and high income residential areas (χ2=2.920, P=0.087). Similarly, no signiﬁcant difference was recorded between the middle and high income residential areas (χ2=2.117, P=0.146). Living in a house with steel doors reduced susceptibility to alcohol abuse by 0.081 times in comparison to living in a house with wooden doors. Houses with steel doors are more secure than those with wooden doors. Residents living in houses with wooden doors experience anxiety and fear of attacks by robbers or thieves at night. In order to assuage this fear, some of them turn to alcohol. Indeed, some residents that took alcohol in high quantities acquired temporary courage and even slept with their doors unlatched, some even spent the night on their doorsteps. Residents living in areas with access footpaths that exceeded 2-m width were 386.426 times more prone to alcohol abuse and dependence than those who lived in areas with smaller access footpaths. Prevalence of Mental Health Disorders among the Adult Residents of the Low, Middle, and High Income Neighborhoods However, there was no signiﬁcant difference in the levels of panic disorder between the low and high income residential areas (χ2=0.269, P=0.604). (χ2=10.969, P=0.001). However, there was no signiﬁcant difference in the levels of panic disorder between the low and high income residential areas (χ2=0.269, P=0.604). Living in a house with very small windows increased the odds of experiencing panic disorder by 8.344 times as opposed to living in one with standard size windows. Residents who had adequate bathrooms reduced their chances of developing panic disorder due to lack of these facilities by 0.089 times. Dissatisfaction with available green spaces increased the chances of developing panic disorder by 4.262 times in contrast to satisfaction with these spaces. Level of education, daily food expenditure per person, and type of residence were signiﬁcant sociodemographic characteristics that were associated with panic disorder. Panic disorder is characterized by recurring severe panic attacks; sometimes, it is accompanied by signiﬁcant behavior change lasting for a month or longer and on- going worry about the implications or concern about experiencing subsequent attacks also called anticipatory attacks. Panic attacks are entirely unpredictable, hence individuals become stressed, anxious, or worried wondering when the next attack will occur.55 The three known types of panic disorder are as follows: unexpected, situationally bounded, and situationally predisposed.37 Post-Traumatic Stress Disorder Across the three income groups, the level of post- traumatic stress disorder (PTSD) was 5.5 %. The low income residents had the 920 OCHODO ET AL. highest incidence recorded at 8.7 %, while its incidence among the high income residents was 6.3 %. The prevalence of this disorder was not signiﬁcantly different between the low and high income residential areas (χ2=0.750, P=0.387). However, between the low and middle income residential areas, there was a signiﬁcant variation in the levels of PTSD (χ2=9.231, P=0.002), likewise, a signiﬁcant variation was recorded between the middle and high income residential areas (χ2= 5.089, P=0.024). Males and females expressed similar patterns of PTSD: 5.8 and 5.2 %, respectively. highest incidence recorded at 8.7 %, while its incidence among the high income residents was 6.3 %. The prevalence of this disorder was not signiﬁcantly different between the low and high income residential areas (χ2=0.750, P=0.387). Prevalence of Mental Health Disorders among the Adult Residents of the Low, Middle, and High Income Neighborhoods Likewise, residents who lived in neighborhoods with sufﬁcient street lighting were 15.623 times more likely to abuse alcohol than those who lived in areas with insufﬁcient street lighting. Residents who were dissatisﬁed with green spaces in their compounds were 0.002 times less likely to abuse alcohol than those who were satisﬁed with these spaces. The signiﬁcant confounders were number of children in a household, number of occupants of a household, age, and daily food expenditure per person. Alcohol dependence is a condition that is characterized by harmful consequences of repeated alcohol use, a pattern of compulsive alcohol use, and physiological dependence on alcohol, i.e., tolerance or symptoms of alcohol withdrawal. This disorder is only diagnosed when these patterns of behavior become persistent, very disabling, or distressing.57,58 Generalized Anxiety Disorder The level of generalized anxiety disorder (GAD) across the income groups was established to be 7.5 %. The level of this disorder was EXTERNAL BUILT RESIDENTIAL ENVIRONMENT CHARACTERISTICS 921 signiﬁcantly different between every two of the three categories of incomes that were studied. The chi-square and P values for the respective pairs of residential areas being low and middle, 6.421, 0.011; low and high, 4.933, 0.026; and middle and high, 19.697, 0.000. Both males and females had a near equal GAD prevalence of 7.7 and 7.4 %, respectively. This ﬁnding is consistent with Cameron's assertion that in the developing nations; the prevalence of GAD is about equal in both sexes. Cameron also established that in the developed countries, women are two to three times more likely to suffer from GAD than men.43 signiﬁcantly different between every two of the three categories of incomes that were studied. The chi-square and P values for the respective pairs of residential areas being low and middle, 6.421, 0.011; low and high, 4.933, 0.026; and middle and high, 19.697, 0.000. Both males and females had a near equal GAD prevalence of 7.7 and 7.4 %, respectively. This ﬁnding is consistent with Cameron's assertion that in the developing nations; the prevalence of GAD is about equal in both sexes. Cameron also established that in the developed countries, women are two to three times more likely to suffer from GAD than men.43 The state of roof and distance traveled to access water were signiﬁcantly associated with GAD. Prevalence of Mental Health Disorders among the Adult Residents of the Low, Middle, and High Income Neighborhoods Residents who lived in houses with non-leaking roofs were 0.322 times less likely to experience GAD than those who lived in houses with leaking roofs. Those who walked more than 500 m to access water for household use were 5.327 times more likely to experience GAD compared to those who accessed water within a distance of less than 500 m. Number of children in a household, age, type of residence, marital status, and level of education were signiﬁcant sociodemographic factors. GAD exceeds the usual anxiety people experience every day. It is a chronic and exaggerated worry that is also accompanied by tension which is not provoked by any particular event. Affected individuals always anticipate disaster; often, they worry excessively about almost anything including health, money, family, or work, though most of the time it is difﬁcult to justify the worry. The simple thought of getting through the day provokes anxiety. Affected people are aware that their anxiety is more intense than the situation warrants.53,59 DISCUSSION This study assessed a wide range of characteristics of the external built residential environment found in poor developing countries against a number of speciﬁc mental health disorders that were identiﬁed among respondents in three distinct types of residential areas. The attributes that were signiﬁcantly associated with mental health disorders were the type of walling material used on buildings, access pathways for walking whose sizes were related to residential housing densities, availability of shopping facilities, types of doors, states of roofs, and states of windows. Others were availability of functional bathrooms, dissatisfaction with available green spaces, lack of street lighting in the neighborhood, and distance traveled to access water. Types of walling material determine the level of draught in a house and security of the occupants. Mud or iron sheet wall houses allow excess wind ﬂow internally; this predisposes residents to ﬂu and pneumonia. Such walls can be easily breached by attackers. The combination of too much cold and fear of attack contribute to troubled incomplete sleep. Repeated inadequate sleep can lead to disorientation that may develop into depression. Narrow access footpaths characterize crowded residences and reduce the possibility of escape in case of misfortunes such as estate ﬁres or ﬂoods. Residents develop feelings of being trapped-in that may lead to claustrophobia. Constant preoccupation with personal safety is a stress factor that creates anxiety and affects mental health. In residential areas, shops mainly supply foodstuffs but may also provide items like medicines and objects of use in households, e.g., brooms, tooth brushes, and toothpaste. In Nakuru Municipality, like in the rest of Kenya, many of the residents 922 OCHODO ET AL. tend to purchase food items almost on a daily basis. For instance, milk for breakfast may be purchased every morning while vegetables for supper may be acquired every evening. When shopping or market facilities are not conveniently placed, distress is likely to set in as residents search for these items. tend to purchase food items almost on a daily basis. For instance, milk for breakfast may be purchased every morning while vegetables for supper may be acquired every evening. When shopping or market facilities are not conveniently placed, distress is likely to set in as residents search for these items. DISCUSSION Prevalence of bipolar mood disorder is dependent on environmental character- istics that vary in different categories of income groups and genetic factors that are randomly distributed in populations.60 Individual psychosocial attributes including harsh environmental conditions can interact with genetic dispositions to inﬂuence the development and course of this disorder.61 Wooden house doors are not secure hence residents living in houses with such doors experience endless fear of attack by thieves especially at night. The absence of secure fencing around the residences adds to the worry for personal safety. This insecurity constitutes a harsh environment that can interact with the residents' genetic factors to inﬂuence the development of bipolar mood disorder. Living in a house with small windows was signiﬁcantly associated with panic disorder. Small windows limit alternative routes for escape in case of ﬁre outbreaks or other misfortunes, they also cause stufﬁness and induced darkness in the house as they do not allow effective ventilation and sufﬁcient daylight indoors. Bathrooms are necessary for personal hygiene particularly at the beginning and end of the day. Where they are inadequate, residents become restless as they await their turn to use them. Sometimes, they are critically inadequate and so residents have to bathe in the open or in their houses. Both of these situations create panic as personal privacy or safety is not guaranteed especially in crowded residential areas. Bathing indoors means that visitors are unwelcome for the time being; besides, other housemates must stay outdoors and if they happen to barge in, embarrassment can occur. This creates anxiety that can lead to panic disorder. Leaking roofs cause dampness that can lead to pneumonia among the affected residents. Reports from the Rift Valley Provincial General Hospital in Kenya indicate that pneumonia is one of the leading killer diseases in the Central Rift Valley region particularly in the cold and wet season.62 Deaths of loved ones traumatize residents leading to post-traumatic stress disorder as the cold weather approaches. Leaking roofs lead to damage of property, interfere with sleep, and affect the health of residents. According to Gingles et al., damp houses are associated with fatigue, headache, chronic anxiety, and depression.59 In the low income residential areas, water taps are few and far between. Frequently, water brawls occur and can lead to fatalities. DISCUSSION Conﬂicts that were at times fatal were witnessed in the low income residential areas in the aftermath of the 2007 General Elections in Kenya. According to residents, some of their kin were attacked and killed as they went out in search of water. The memories of these events were still vivid in their minds at the time of conducting this research and increased their likelihood of developing post-traumatic stress disorder. Quite often, water was not enough and families were forced to use little of it while foregoing some activities that required its use. Anxiety arose from the possibility of missing it altogether even after spending long hours in winding queues. Children left in the houses unattended also contributed to anxiety as they could burn in the charcoal braziers that were left with blazing ﬁres cooking the midday meals, they could also hurt themselves using objects in the house like sharp knives or poisonous chemicals stored within their reach. Areas with standard footpaths and sufﬁcient street lighting are likely to be inhabited by richer folks who possess extra disposable income that they can use to EXTERNAL BUILT RESIDENTIAL ENVIRONMENT CHARACTERISTICS 923 purchase alcohol. Conversely, most residents who were dissatisﬁed with available green spaces were found in the low income residential areas. Even though some of them did accept that they took alcohol, they could not afford to consume it in high quantities on a regular basis like their middle or high class counterparts. Their capacity to abuse alcohol was curtailed by default of lacking the ﬁnancial means to access it. purchase alcohol. Conversely, most residents who were dissatisﬁed with available green spaces were found in the low income residential areas. Even though some of them did accept that they took alcohol, they could not afford to consume it in high quantities on a regular basis like their middle or high class counterparts. Their capacity to abuse alcohol was curtailed by default of lacking the ﬁnancial means to access it. Since it is the low income residential areas that were mainly affected by lack of physical planning and inadequate amenities, residents who lived there had a higher likelihood of suffering from various mental health disorders identiﬁed in this study. DISCUSSION Even though many residents of the low income areas consumed alcohol, the possibility of alcohol abuse and dependence was found to be higher in the middle and high income neighborhoods owing to their enhanced purchasing power. Judicious taxation of this beverage can therefore check its consumption. p Lack of shopping facilities in the high income residential areas was signiﬁcantly associated with occurrence of depression in spite of the residents' ability to buy food items in large quantities and refrigerate. This association can be explained by the eating habits of Kenyans who prefer fresh food items to refrigerated ones. As a result, living in high income residential areas places restrictions on their buying habits. Galea et al. provide an explanation that may account for the association between characteristics of the urban built residential environment and mental health disorders.38 Their psychosocial stress explanation suggests that living in neighborhoods characterized by poor quality built environment is associated with psychosocial stress that in turn may contribute to mental health disorders. 63,64 Urban residential areas characterized by poor quality built environment substantially expose inhabitants to daily stressors and inconveniences that can lead to greater social strain and a higher likelihood of developing mental health disorders. It is also possible that densely populated urban residential areas amplify social deﬁciencies and stressors.65,66 Thus, high populations in areas characterized by deteriorating built environment characteristics can exacerbate both social and poor mental health consequences. CONCLUSION Since dense populations in low income neighborhoods aggravate social strain and consequent mental health disorders, easing of population pressure in these areas by stemming rural–urban migration and improving overall neighborhood economic outlook should be prioritized. Developing rural infrastructure alongside provision of basic amenities like electricity and water can encourage establishment of agro-based industries that would reduce rural–urban migration. ACKNOWLEDGMENT I thank my two able research assistants Mr. Dan Okoo and Mr. Ben Mwangi both of the Kenya National Bureau of Statistics, Nakuru County for their enduring stamina that ensured successful completion of data collection. I am most grateful to the residents of Nakuru Municipality (Rhonda, Shabaab, Kenlands, Section 58, and Milimani residential areas) for their notable cooperation in responding to various questions of the study. Open Access This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited. Limitations of the Study In this study, it is possible that certain multiple risk factors clustered at different levels thus becoming difﬁcult to investigate. Factors occurring at the individual, household, and neighborhood levels could all have an impact on urban residents' mental health. Residents could disproportionately experience individual risk factors associated with mental illness such as unemployment that is associated with poverty. The study sample comprised of low, middle, and high income groups and the unemployed; this reduced bias arising from poverty. Income was deﬁned by attributes of the external built environment. Therefore, the rich people who chose to live in neighborhoods with poor external built environment characteristics were exposed to the prevailing adverse environmental characteristics and could consequently experience poor mental health. Other confounding factors that could inﬂuence mental health were age, sex, period in residence, size of household, and house tenure (tenancy, leasehold, or freehold), all of which were individually analyzed in order to establish their relationship with mental health states of residents. OCHODO ET AL. 924 5. Raffestine C, Lawrence R. An ecological perspective on housing, health, and wellbeing. J Sociol Soc Welf. 1990; 17: 143–160. CONCLUSION This study demonstrates distinct associations between explicit characteristics of the urban external built residential environment and the likelihood of speciﬁc mental health disorders in adult males and females. It proves that low income residential areas have a higher concentration of poor external built environment characteristics and therefore expose residents to increased possibilities of developing mental health disorders. It is therefore necessary to improve qualities of the external built residential environment as part of key interventions to safeguard public mental health. Urban authorities can use available resources to upgrade the built residential environment. Such upgrading does not necessarily have to involve construction of new houses similar to those in the middle or high income residential areas. For a start, modiﬁcation of available living structures to protect residents from the elements, decongesting residential estates by adhering to prescribed housing densities, provision of street lighting, sanitary, and garbage management facilities alongside maintenance of access pathways would sufﬁce. These interventions call for close partnership between urban physical planners and public health professionals. This study demonstrates distinct associations between explicit characteristics of the urban external built residential environment and the likelihood of speciﬁc mental health disorders in adult males and females. It proves that low income residential areas have a higher concentration of poor external built environment characteristics and therefore expose residents to increased possibilities of developing mental health disorders. It is therefore necessary to improve qualities of the external built residential environment as part of key interventions to safeguard public mental health. Urban authorities can use available resources to upgrade the built residential environment. Such upgrading does not necessarily have to involve construction of new houses similar to those in the middle or high income residential areas. For a start, modiﬁcation of available living structures to protect residents from the elements, decongesting residential estates by adhering to prescribed housing densities, provision of street lighting, sanitary, and garbage management facilities alongside maintenance of access pathways would sufﬁce. These interventions call for close partnership between urban physical planners and public health professionals. Since dense populations in low income neighborhoods aggravate social strain and consequent mental health disorders, easing of population pressure in these areas by stemming rural–urban migration and improving overall neighborhood economic outlook should be prioritized. Developing rural infrastructure alongside provision of basic amenities like electricity and water can encourage establishment of agro-based industries that would reduce rural–urban migration. 4. Douglas CH, Douglas MR. Attitudes to and Perceptions of the Built Environment: Considerations for Better Design Technologies. Shefﬁeld, United Kingdom: University of Shefﬁeld; 2003. 3. The Picker Institute (Boston MA) and the Center for Health Design Working Paper. Consumer Perceptions of the Healthcare Environment, an investigation to Determine What Matters. Boston MA: The Picker Institute and the Center for Health Design; 1998. 2. Jackson RJ. The impact of the built environment on health: an emerging ﬁeld. Am J Public Health. 2003; 93(9): 1382–1384. 1. Ulrich RS. Effects of interior design on wellness: theory and recent scientiﬁc research. J Health Care Inter Des. 1991; 3: 1997–109. 2. Jackson RJ. The impact of the built environment on health: an emerging ﬁeld. Am J Public Health. 2003; 93(9): 1382–1384. 3. The Picker Institute (Boston MA) and the Center for Health Design Working Paper. Consumer Perceptions of the Healthcare Environment, an investigation to Determine What Matters. Boston MA: The Picker Institute and the Center for Health Design; 1998. 4. Douglas CH, Douglas MR. Attitudes to and Perceptions of the Built Environment: Considerations for Better Design Technologies. Shefﬁeld, United Kingdom: University of Shefﬁeld; 2003. 5. Raffestine C, Lawrence R. An ecological perspective on housing, health, and wellbeing. J Sociol Soc Welf. 1990; 17: 143–160. 1. Ulrich RS. Effects of interior design on wellness: theory and recent scientiﬁc research. J Health Care Inter Des. 1991; 3: 1997–109. REFERENCES 1. Ulrich RS. Effects of interior design on wellness: theory and recent scientiﬁc research. J Health Care Inter Des. 1991; 3: 1997–109. 2. Jackson RJ. The impact of the built environment on health: an emerging ﬁeld. Am J Public Health. 2003; 93(9): 1382–1384. 3. The Picker Institute (Boston MA) and the Center for Health Design Working Paper. Consumer Perceptions of the Healthcare Environment, an investigation to Determine What Matters. Boston MA: The Picker Institute and the Center for Health Design; 1998. 4. Douglas CH, Douglas MR. Attitudes to and Perceptions of the Built Environment: Considerations for Better Design Technologies. Shefﬁeld, United Kingdom: University of Shefﬁeld; 2003. 5. Raffestine C, Lawrence R. An ecological perspective on housing, health, and wellbeing. J Sociol Soc Welf. 1990; 17: 143–160. EXTERNAL BUILT RESIDENTIAL ENVIRONMENT CHARACTERISTICS 925 6. Fullilove MT, Fullilove RE. What's housing got to do with it? Am J Public Health. 2000; 90: 183–184. 7. Rauh VA, Chew GL, Garﬁnkel RS. Deteriorated housing contributes to high cockroach allergen levels in inner city households. Environ Health Perspect. 2002; 110: 323–327. 8. Lehman I, Thoelke A, Weiss MT. Cell reactivity in neonates from an East and West German City: results of the Lisa study. Allergy. 2002; 57: 129–136. 9. Leaderer BP, Belanger K, Triche E. Dust mite, cockroach, cat and dog allergen concentrations in homes of asthmatic children in northeastern United States: impact of socioeconomic factors and population density. Environ Health Perspect; 2002; 110(4): 419–425. 10. Humke C, Schaefer C. Relocation: a review of the effects of residential mobility on children and adolescents. J Psychol Hum Behav. 1995; 32: 16–24. 11. Adam EK, Chase-Lamsdale PL. Home sweet home: parental separation, residential move, and adjustment problems in low-income girls. Dev Psychol. 2002; 38: 792–805. 12. Butterworths I. The relationship between the built environment and wellbeing: a literature review. Melbourne, Australia: The Victoria Health Promotion Foundation; 2000 13. Dalgard OS, Tambs K. Urban environment and mental health: a longitudinal study. J Psychiatry. 1997; 171: 530–536. 14. Wainwright, NWJ. Surtees, PG. Places, people and their physical and mental functional health. J Epidemiol Community Health. 2004; 58: 333–339. 15. World Health Organization WHO Technical Meeting on the Immediate Housing Environment. Freiburg: European Center for Environment and Health Ofﬁce 2002 15. World Health Organization WHO Technical Meeting on the Immediate Housing Environment. Freiburg: European Center for Environment and Health Ofﬁce 2002 16. Frumkin H. Healthy places: exploring the evidence. REFERENCES Am J Public Health. 2003; 93(9): 1451 1456 16. Frumkin H. Healthy places: exploring the evidence. Am J Public Health. 2003; 93(9) 1451–1456. 17. Sloane PD, Mitchell MC, Presser JS, et al. Environmental correlates of Alzheimer's disease special care units. Am Geriatr Soci. 1998; 46(7): 862–869. 18. Johansson MP, Ladd HF, Ludwig J. The beneﬁts and cost of residential mobility programs for the poor. J Hous Study. 2002; 17: 125–138. p g p J y 19. Cooper MC, Barnes M. Gardens in health care facilities: uses, therapeutic benefits, and design recommendations. Martinez C.A. The Center for Health Design, Inc. 1995 20. Ulrich RS. Effects of gardens on health outcomes: theory and research. In: Cooper C, Barnes M, eds. Healing Gardens, Therapeutic Benefits and Design Recommendations (27–86). New York, USA: John Wiley and Sons; 1999. 21. Teresa L, Claire G, Owen M, et al. Health impacts of the built environment: a review. The Institute of Public Health in Ireland 2006 22. Mueller BA, Rivara FP, Lii SM, et al. Environmental factors and the risk for childhood pedestrian–motor vehicle collusion occurrence. Am J Epidemiol. 1990; 132: 550–560. 23. Weich S, Blanchard M, Prince M. Mental health and the built environment: cross-sectional survey of individual and contextual risk factors for depression. J Psychiatry. 2002; 180: 428–433. p J y y ; 24. Semenza JC. The intersection of urban planning, art and public health: the Sunny Side Piazza. Am J Public Health. 2003; 93(9). J 25. CABE S. Decent Parks? Decent Behavior? The Link Between Equality of Parks and User Behavior. London, United Kingdom; 2005 26. Ellaway A, Macintyre S, Xavier B. Grafﬁti, greenery, and obesity in adults, secondary analysis of a European cross-sectional survey. BMJ. 2005; 331: 611–612. 27. Wandersman A, Nation M. Urban neighborhoods and mental health. J Am Psychol 1998; 53: 647–656. 28. Leventhal T, Brooks-Gunn J. Neighborhoods they live in: the effects of neighborhood residence on child and adolescent outcomes. Psychol Bull. 2000; 126: 309–337. 29. Allen D. Impact of Our Built Environment on Public Health. Charlotte, USA: Nationa Institute of Health Services; 2004 30. Baum A, Paulus PB. Crowding. In: Stokols D, Altman I, eds. Handbook of Environmental Psychology. New York, USA: Wiley; 1987: 533–570. 31. Chombant P, Chombant M. The Danger Signals of Overcrowding. In: Asbell B. Think Magazine (July- August pp. 30–33). New York, USA: Armon Inc; 1975. 926 OCHODO ET AL. 32. Wener RE, Keys C. REFERENCES The effects of changes in jail population densities on crowding, sick call, and social behavior. J Appl Soc Psychol. 1988; 18: 852–866. 33. Lepore SJ, Evans GW, Schneider M. The dynamic role of social support in the link between chronic stress and psychological distress. J Soc Psychol. 1991; 61: 899– 909. 34. Evans GW. Environmental Stress and Health. In: Baum A, Revenson T, Singer J.E eds. Handbook of Health Psychology. Mawhah. 2001; NJ: 571–610. 35. Kenya Advisor. com. The Independent Kenya Travel Guide. http://www.kenya- advisor.com/kenya-cities. html; 2005. 36. Kenya National Bureau of Statistics. Population and Housing Census Vol. 1A. Nairobi: Government Printer; 2009. 37. Ministry of Planning. Nakuru District Development Plan, 2002–2008. Nairobi. Government Printer; 2002. 38. Galea S, Freudenberg N, Vlahor D. Cities and population health. Soc Sci Med J. 2005; 60: 1017–1033. 39. Guite HF, Clarke C, Ackrill G. The Impact of the Physical and Urban Environment on Mental Wellbeing. London, United Kingdom: Queen Mary University, Centre for Psychiatry; 2006. 40. Satcher D. Mental Health: A Report of the Surgeon General. Surgeon General of the United States of America; 1999 41. Sheehan DV, Janavs J. The Mini International Neuropsychiatric Interview for Childre and Adolescents. Tampa, USA: University of South Florida; 2008. 42. Sheehan DV, Janavs J. The Mini International Neuropsychiatric Interview. Tampa, USA University of South Florida; 2010. 43. World Health Organization. Alcohol, Smoking and Substance Involvement Screening Te (ASSIST). Geneva, Switzerland: World Health Organization; 1997. 44. World Health Organization. Community Health Nursing. In: Basavanthappa BT. Environmental Health. New Delhi, India: Jaypee Brothers Medical Publishers; 2008. 45. Verderberber S. Dimensions of person–window transactions in the hospital environment. Environ Behav. 1986; 18(4): 450–466. 46. Leather P, Pyrgas M, Beale D, et al. Windows in the workplace: sunlight, view, and occupational stress. Environ Behav. 1987; 30(6): 739–762. 47. Ministry of Lands. Physical Planning Handbook. Nairobi: Government Printer; 2008. 47. Ministry of Lands. Physical Planning Handbook. Nairobi: Government Printer; 2008. 48. Dorfman SF. Exploring the Built Environment. In: http://www.medspace.com/viewarticle/ 489023 M d P bli H l h d P i 2006 48. Dorfman SF. Exploring the Built Environment. In: http://www.medspace.com/viewarticl 489023. Medscape Public Health and Prevention; 2006. 49. Cave B. Rapid Review of Health Evidence for the Draft London Plan; Based on – Towards the London Plan: Initial Proposals for the Mayor’s Spatial Development Strategy. Greater London Authority and the London Health Observatory; 2001 50. World Health Organization. REFERENCES Fourth Ministerial Conference on Environment and Health. Budapest; 2004. p 51. Task N, Taylor L, Mulvihill C, et al. A Review of Reviews of Interventions for Improving Health. National Institute for Health and Clinical Excellence: Evidence Brieﬁng; 2005. 52. Ulrich RS. Health Beneﬁts of Gardens in Hospitals. Paper for conference. Plants fo People: International Exhibition Florida; 2002. p 53. Ndetei DM. Your A-Z on Mental Health. Nairobi: Acrodile Publishing Limited, Kenya; 2010. 53. Ndetei DM. Your A-Z on Mental Health. Nairobi: Acrodile Publishing Limited, Kenya; 2010. 54. Basco MR. The Bipolar Work Book: Tools for Controlling Your Mood Swings. New York, USA: The Guillford Press; 2006. 54. Basco MR. The Bipolar Work Book: Tools for Controlling Your Mood Swings. Ne York, USA: The Guillford Press; 2006. , ; 55. Barker P. Psychiatric and mental health nursing: the craft of caring. Arnold; 2003: 215 55. Barker P. Psychiatric and mental health nursing: the craft of caring. Arnold; 2003: 215. 56. American Psychiatric Association. Diagnosis and Statistical Manual of Mental Disorders. 4th ed. Washington DC: American Psychiatric Association; 1994. y g g ; 56. American Psychiatric Association. Diagnosis and Statistical Manual of Mental Disorders. 4th ed. Washington DC: American Psychiatric Association; 1994. g y ; 57. Brunet A, Akerib V, Birmes P. Don't throw away the baby with the bathwater, PTSD is not over-diagnosed. Can J Psychiatry. 2007; 52(8): 501–502. 58. Bablor TF. Substance related problems in the context of international classiﬁcation systems. In: Lader M, Edwards G, Drummonds DC, eds. The Nature of Alcohol and Drug Related Problems. New York, USA: Oxford University Press; 1992. EXTERNAL BUILT RESIDENTIAL ENVIRONMENT CHARACTERISTICS 927 59. Gingles EJ, McErlain MS, McPeake JWR, et al. Health and Housing Study. Northern Ireland Housing Executive: Department of Public Health Medicine, Eastern Health, Social Services Board and Research Unit; 1995 60. Lam D, Wright K, Smith N. Dysfunctional assumptions in bipolar disorder. J Affect Disord. 2004; 79(1–3): 193–199. 61. Serretti A, Mandelli L. The genetics of bipolar disorder: genome hot regions, genes, new potential candidates, and future directions. Mol Psychiatry. 2008; 13(8): 742–771. 62. Rift Valley Provincial General Hospital. Annual Reports: 2008–2010. Rift Valley Provincial General Hospital, Kenya; 2010. 63. Scheck C, Kinicki A, Davy JA. Longitudinal study of a multivariate model of the stress process using structured equation modeling. Homan Relat. 1995; 48: 1481–1510. 64. Amick BC, Kawachi I, Coakley EH. ; 66. Eaton WW. The Sociology of Mental Disorders, 2nd ed. New York, USA; 1986 REFERENCES Relationship of job strain and strain to health status in a cohort of women in the United States. Environ Health. 1998; 24: 54–61. 65. Frei A, Schenker T, Finzen A. The weather effect and assisted suicide. Suicide Life Threat Behav. 2003; 33: 192–200. ; 66. Eaton WW. The Sociology of Mental Disorders, 2nd ed. New York, USA; 198
https://openalex.org/W2792092363	https://www.nature.com/articles/s41598-018-22799-4.pdf	English	null	The development of an extra-anatomic tissue-engineered artery with collateral arteries for therapeutic angiogenesis in ischemic hind limb	Scientific reports	2,018	cc-by	8,686	The development of an extra- anatomic tissue-engineered artery with collateral arteries for therapeutic angiogenesis in ischemic hind limb Received: 31 October 2017 Accepted: 1 March 2018 Published: xx xx xxxx Xu Zhou 1, Yinlong Zhang2, Hongfei Wang3, Bin Zhao2, Jinling Wang4, Guoliang Yan5, Shuangyue Xu6, Yuanyuan Zhou2, Hongyi Liu2, Yifei Zheng2, Wei Quan2, Jianyin Zhou7, Yun Liu7, Maochuan Zhen7, Xuan Zhu8 & Yilin Zhao1 To develop tissue-engineered arteries (TEAs) with collateral arteries(CAs) in ischemic hind limb goat models(IHLMs). The IHLMs created by removing femoral arteries were divided into non-treated control group(NG); non-catheter group (NCG) in which TEA was anastomosed to external iliac artery(EIA), and surrounded with collagen sponge containing autologous MSCs and VEGF-gelatin microspheres, the distal end of TEA was ligated; catheter group(CG) which received the same procedure as NCG, also received heparin infusion through catheter in EIA. TEA patency was assessed weekly by Ultrasound. The TEA and CAs were assessed by angiography, gross examination, histology and electron microscopy. In CG, TEAs remained patent for 1 month, but became partly occluded 1 week after catheter withdrawn. In NCG, TEAs were occluded 1 week after implantation. Angiography demonstrated that communication between CAs arising from the TEAs and the native vessels was established in both groups. NCG had fewer CAs than CG (P < 0.01). At 40 days, TEAs in CG demonstrated of endothelium formation, smooth muscle cells infiltration and collagen regeneration. The CG had more capillaries and mature vessels in adventia of TEAs than NCG (P < 0.01). CG group also had more vessels around TEAs than NCG (P < 0.01) or NG (P < 0.001). A large number of chronic peripheral artery disease (PAD) patients are not candidates for the conventional revas- cularization techniques of angioplasty, stenting, and peripheral artery bypass grafting1–4, but advances continue to be made in the prevention and treatment of PAD5–7. Alternative noninvasive methods now available include localized delivery of therapeutic growth factors8,9and autologous stem cell transplantation10. The administration of VEGF and bFGF have demonstrated beneficial results in reversing ischemia and improving regional blood perfusion in extremities in both animal and clinical trials11. However, growth fac- tors such as VEGF or bFGF have poor stability in vivo and consequently, short-lasting biological activity12. Sustained-release systems, with gelatin as the drug carrier13–15, have been developed that ensure gradual, localized delivery of growth factors over the span of a few weeks. 1Oncology and Vascular Interventional Radiology, Zhongshan Hospital, Xiamen University, Xiamen, Fujian Province, P. R. China. 2Medical College, Xiamen University, Xiamen, Fujian Province, P. R. China. www.nature.com/scientificreports www.nature.com/scientificreports www.nature.com/scientificreports Received: 31 October 2017 Accepted: 1 March 2018 Published: xx xx xxxx Methods P ti Preparation and characterization of nanosized decellularized artery scaffolds (NDAs). Fresh goat carotid arteries, with lumen diameters of 4.0 ± 0.2 mm, were harvested from the local slaughterhouse. The procedure for preparation and examination of NDAs, the DNA quantification and mechanical testing were based on our previous published work24,25. The 5-0 polypropylene sutures (Ethilon, USA) were placed in each side of the scaffold at 1 mm from the edge. A constant elongation (5 mm/min) was applied along the longitudinal axis of the scaffold until the sutures were pulled through the vessel edge, and the maximum suture strength was recorded by Instron mechanical tester (Norwood, MA). Fifty-millimeter scaffold samples were attached via tapered adapters directly into the pressurized tube and tensioned and submerged in distilled water. The system was progressively pressurized until vessel failure when air bubbles from the water can be observed. The burst pressure was recorded by a manometer with data acquisition system (Testo, Germany). The porosity of the scaffold was measured using the following formula: Porosity (%) = (bulk volume − true volume)/bulk volume × 100%. The true volume of scaffold were analyzed using True Density Analyzer with helium gas (Beijing Builder Electronic Technology, China). Bulk volume of the scaffold was calculated by measuring volume. Preparation and characterization of gelatin microspheres (GMs). Gelatin (2 g; Sigma-Aldrich, St Louis, MO, USA) was dissolved in15 ml phosphate-buffered saline (PBS; pH 7.4). The preparation and examina- tion of GMs as described by our previous our previous published work26,27. Preparation and characterization of collagen scaffolds (CSs). Type I collagen (100 mg; Sigma-Aldrich) was mixed in 40 ml 0.1% acetic acid solution at 2000 rpm for 30 min to prepare collagen protein solution. The preparation and examination of CSs as described by our previous our previous published work24,28. Rehydration analysis. Dried samples of NDAs, GMs, and CSs were immersed in PBS. After 24 h, the sam- ples were removed from the solution and the surface was blotted dry with a piece of tissue; the samples were then weighed. Rehydration analysis. Dried samples of NDAs, GMs, and CSs were immersed in PBS. After 24 h, the sam- ples were removed from the solution and the surface was blotted dry with a piece of tissue; the samples were then weighed. Cytotoxicity of NDAs, GMs, and CSs. www.nature.com/scientificreports/ www.nature.com/scientificreports/ However, questions remain regarding the efficacy of treatment with such systems, especially in patients who have advanced PAD with many occluded arteries. For example, after revascularization has been induced with the administration of angiogenic factors, what will be the source of blood supply to the new collateral vessels? g g pp y One approach might be to combine the Vineberg procedure, which has been virtually neglected for decades, with therapeutic angiogenesis16. In animal experiments, Arthur Vineberg dissected the internal thoracic artery from the chest wall, ligated the distal end, and pulled it into a tunnel created in the superficial myocardium17. He found that the artery sprouted new collateral vessels that connected with the native myocardial blood vessels18. The Vineberg operation has been used as a last resort in patients with diffuse atheromatous lesions who are unsuited for direct revascularization with methods such as angioplasty, stenting, and bypass surgery19. g p y g yp g y However, the Vineberg procedure is not suitable for PAD patients, as autologous arteries are unavailable for use in this location. Completely synthetic grafts, such as Dacron® and ePTFE vascular grafts, are also ruled out as they cannot develop collateral branches to establish communication with the native vasculature in the ischemic extremity. Tissue-engineered (TE) blood vessels, with decellularized vascular grafts used as a scaffold, could be a possible solution; this acellular tissue retains its natural mechanical properties and facilitates neovascularization and recellularization after transplantation20. t p Therefore, in this study, we applied the Vineberg procedure in an ischemic hind limb animal model, using a TE artery surrounded by collagen complex containing autologous MSCs and gelatin microspheres formulated for sustained VEGF release21–23. We studied whether this extra-anatomic artery could develop collateral artery branches to communicate with local vessels and recover the blood supply to the ischemic hind limb. Methods P ti A 96-well microplate containing 200 μl culture medium per well was taken and seeded with mesenchymal stem cells (MSCs) at a density of 1.0 × 104 cells per well. This was kept at 37 °C in a 5% CO2 atmosphere for 24 h, after which the culture medium was replaced with fresh medium con- taining 200 μg/ml of NDAs, GMs, or CSs pieces; normal medium was used as a control. After 1, 3, 7, and 9 days, the medium was removed and the cells were detached by 0.05% trypsin with ethylene diamine tetraacetic acid (EDTA), and3-(4,5-dimethyl-2-thiazolyl)-2, 5-diphenyl-2-H-tetrazolium bromide (MTT)assay was performed. Cytotoxicity of NDAs, GMs, and CSs. A 96-well microplate containing 200 μl culture medium per well was taken and seeded with mesenchymal stem cells (MSCs) at a density of 1.0 × 104 cells per well. This was kept at 37 °C in a 5% CO2 atmosphere for 24 h, after which the culture medium was replaced with fresh medium con- taining 200 μg/ml of NDAs, GMs, or CSs pieces; normal medium was used as a control. After 1, 3, 7, and 9 days, the medium was removed and the cells were detached by 0.05% trypsin with ethylene diamine tetraacetic acid (EDTA), and3-(4,5-dimethyl-2-thiazolyl)-2, 5-diphenyl-2-H-tetrazolium bromide (MTT)assay was performed. Incorporation of VEGF in GMs (GMs-VEGF). Recombinant human VEGF (rhVEGF; Applied Biosystems, California, USA) solution (20 μg/mL) was dropped onto the GMs. The GMs-VEGF was kept at room temperature for 4 h and then lyophilized for 24 h. Incorporation of VEGF in GMs (GMs-VEGF). Recombinant human VEGF (rhVEGF; Applied Biosystems, California, USA) solution (20 μg/mL) was dropped onto the GMs. The GMs-VEGF was kept at room temperature for 4 h and then lyophilized for 24 h. Isolation of MSCs. MSCs were isolated from goat bone marrow (Xiamen University Animal Center). The procedure for preparation of MSCs was based on our previous published work25. Briefly, bone marrow aspirated from the anterior iliac crest was diluted with heparinized phosphate-buffered saline (PBS). Ficoll density gradi- ent centrifugation was used to separate MSCs from other cells. The residual cells were washed three times with PBS and incubated at 37 °C, 90% humidity, and 5% CO2. Second-passage cells at subconfluence were used for all experiments. All the cells used for experiments were labeled with PKH67(Sigma-Aldrich). Isolation of MSCs. MSCs were isolated from goat bone marrow (Xiamen University Animal Center). The development of an extra- anatomic tissue-engineered artery with collateral arteries for therapeutic angiogenesis in ischemic hind limb 3Vascular surgery, Changhai Hospital, Second Military Medical University, Shanghai, P. R. China. 4Emergency, Zhongshan Hospital, Xiamen University, Xiamen, Fujian Province, P. R. China. 5Basic Medical Department of Medical College, Xiamen University, Xiamen, Fujian Province, P. R. China. 6Organ Transplantation Institute, Xiamen University, Fujian Key Laboratory of Organ and Tissue Regeneration, Xiamen, Fujian Province, P. R. China. 7Hepatology Surgery, Zhongshan Hospital, Xiamen University, Fujian Provincial Key Laboratory of Chronic Liver Disease and Hepatocellular Carcinoma, Xiamen, Fujian Province, P. R. China. 8College of Pharmacy, Xiamen University, Xiamen, Fujian Province, P. R. China. Xu Zhou, Yinlong Zhang and Hongfei Wang contributed equally to this work. Correspondence and requests for materials should be addressed to Y.Zhao. (email: zyllbz@gmail.com) or X.Zhu. (email: zhuxuan@xmu.edu.cn) SCIEnTIFIC REPOrTS \| (2018) 8:4627 \| DOI:10.1038/s41598-018-22799-4 SCIEnTIFIC REPOrTS \| (2018) 8:4627 \| DOI:10.1038/s41598-018-22799-4 1 Methods P ti At the designated time intervals, the tubes were taken out and centrifuged at 12 × 103 rpm for 15 min, and the superna- tant was transferred into a test tube and stored at −20 °C till VEGF analysis was performed. Equal amount of fresh PBS was added to the original tubes till the next measurement. The amount of released VEGF was analyzed using a human VEGF ELISA kit(Abcam, Cambridge, UK) according to the manufacturer’s instructions. Tissue-engineered artery preparation. MSCs suspensions (1 × 106 cells/ml) were stained with PKH67 (Sigma-Aldrich) and seeded into the 3 cm long NDAs. The constructions were cultured for 9 days, and observed under fluorescence microscopy. MSCs proliferation was analyzed by MTT assays at regular intervals. Vineberg procedure. Goat experiment was approved by Laboratory Animal Management Ethics Committee of Xiamen University. Animal care and experimental procedures were carried out strictly in compliance with the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health(NIH Publications No. 8023, revised 1978).h The adult goats (mean weight, 20 ± 2 kg) were anesthetized with intramuscular ketamine (30 mg/kg) and intravenous pentobarbital (30 mg/kg) and ventilated with a mixture of O2, N2, and isoflurane during the oper- ation. The acute ischemic hind limb goat model was created by ligating the external iliac and popliteal arteries, and removing the femoral artery between the ligatures. The goat models were divided into a catheter group (CG; n = 3), a non-catheter group (NCG; n = 3), and a control group (NG; n = 3). The control group did not receive any treatment. In the non-catheter group, the prepared TE artery was anastomosed to the external iliac artery using 6-0 polypropylene sutures (Ethicon, USA). The distal end was ligated and surrounded with CSs containing autologous MSCs and sustained-release GMs-VEGF formulations. In the catheter group, the same procedure was performed but, in addition, a catheter was placed in the external iliac artery for continuous heparin infusion (at the rate of 5 u/kg/h) to maintain anticoagulation within the TE artery. No anticoagulants were used in the other groups. The implanted TE vascular conduits were approximately 30 mm in length and had an average inner diameter of 4 mm. Monitoring, evaluation, and harvest of the graft. To determine graft patency after implantation, the animals were examined by Doppler ultrasound every week. Methods P ti The procedure for preparation of MSCs was based on our previous published work25. Briefly, bone marrow aspirated from the anterior iliac crest was diluted with heparinized phosphate-buffered saline (PBS). Ficoll density gradi- ent centrifugation was used to separate MSCs from other cells. The residual cells were washed three times with PBS and incubated at 37 °C, 90% humidity, and 5% CO2. Second-passage cells at subconfluence were used for all experiments. All the cells used for experiments were labeled with PKH67(Sigma-Aldrich). Characterization of MSCs. The immunophenotype of MSCs was determined by flow cytometry accord- ing to previous described methods29,30. Second-passage MSCs were harvested, non specific binding was blocked and the cells were stained with different antibodies against the human antigens: CD29-FITC, CD90- FITC and CD45-FITC(eBioscience, California, USA). Nonspecific isotype matched antibodies were used as controls. The SCIEnTIFIC REPOrTS \| (2018) 8:4627 \| DOI:10.1038/s41598-018-22799-4 2 www.nature.com/scientificreports/ fluorescence intensity of the cells was evaluated by flow cytometer and the data were analyzed with the CytExpert software (Beckman Coulter, California, USA). Results are presented as percent positively stained cells. Incorporation of MSCs and GMs-VEGF into CSs (CSs complex). MSCs and GMs-VEGF were seeded into collagen sponges by the agitated seeding method. Collagen complex (1 × 1 × 0.3 cm3) was placed into a dish, and 1 × 106 MSCs in 1 ml suspension and 50 mg GMs-VEGF were evenly injected into the CSs complex by blunt-end needle syringes. The collagen complex was placed in a 50 ml tube and agitated at 300 rpm on an orbital shaker for 6 h at 37 °C. The complex was then incubated with 2 ml DMEM supplemented with 10% FBS and 100 u/ ml penicillin–streptomycin in a dish for 2 h at 37 °C. The collagen complex was routinely examined under light and fluorescence microscopes; H&E staining and SEM were used for further evaluation of the collagen complex. The MSCs proliferation within the collagen complex was assessed at designated intervals using MTT assay. In vitro release of VEGF. GMs-VEGF and CSs-GMs-VEGF specimens were placed in 5 ml centrifuge tubes containing 2 ml PBS (pH 7.4). The tubes were maintained at 37 °C with gentle shaking for up to 14 days. Methods P ti Angiography was performed 1 month after surgery, and the catheter (in the catheter group) was removed after angiography. All animals were sacrificed at the end of 40 days for gross examination and histological analysis of the graft. The collateral vessels arising from the TE artery wall were counted on the gross specimen and in the angiography images. Histological and immunohistochemical examination. Before implantation, the NDAs, CSs com- plex, and TE artery were all histologically examined after H&E, Masson’s trichrome, and Van Gieson’s staining. Segments of the implanted vascular graft and surrounding tissue were harvested 40 days after implantation, fixed in 10% buffered formaldehyde solution, dehydrated with a graded ethanol series, embedded in paraffin, cut into 5-μm-thick sections, and stained with H&E. Masson’s trichrome and van Gieson’s stains were used to examine collagen and elastin deposition in the harvested vascular conduits. Immunostaining for von Willebrand Factor (vWF) and α-Smooth Muscle-Actin (α-SMA) was also performed; brown granular deposits indicated the pres- ence of vWF or α-SMA. The vessels per unit area were counted in five randomly chosen fields per slide of each portion by three blinded pathologists. The average number of vessels in one portion was used for assessment of vascular number. Vessels with diameter >50 μm were considered mature. μ We harvested the samples at proximal, middle and distal part of TE arteries and got the samples at 0, 4 and 8 o ‘clock points at each part. The sections of specimens were stained with immunostaining for vWF, the coverage of endothelium in TE arteries was calculated according to the formula (the length of positive vWF staining/total detected length %) in five randomly chosen fields per slide by three blinded pathologists. In vivo differentiation of MSCs. The samples were fixed and cut into 5-µm-thick sections. The slides were examined for fluorescence to trace the implanted MSCs, and then incubated with vWF antibody and FITC-conjugated secondary antibody (Boster, Wuhan, China) to detect endothelial cells. The signals were visual- ized using a confocal laser scanning microscope (Leica, Wetzlar, DE). Electron microscopy. For SEM, specimens were fixed in 1% buffered glutaraldehyde and 0.1% buffered formaldehyde for 1 and 24 h, respectively, dehydrated with a graded ethanol series, and dried. The dried samples were mounted on an aluminum stub and coated with gold using a sputter coater before SEM (TESCAN, Czech Republic). SCIEnTIFIC REPOrTS \| (2018) 8:4627 \| DOI:10.1038/s41598-018-22799-4 3 www.nature.com/scientificreports/ Figure 1. Methods P ti Nanosized decellularized arteries (NDAs). (A) Biological structures, (1) Gross view, (2) H&E (bar: 200 μm), (3) Masson’s trichrome staining (bar: 200 μm), (4) Van Gieson’s staining (bar: 200 μm). (B) Test, (1) DNA content (P < 0.001), (2) Suture strength (P > 0.05) Fresh arteries (FAs), (3) Burst pressure (P > 0.05), (4) Rehydration (P < 0.001). (C) Microstructure, (1) Cross-section of NDAs (SEM, bar: 50 μm), (2) Intima (SEM, bar: 10 μm), (3) Subintima and media (SEM, bar: 10 μm), (4) Adventitia (SEM, bar: 10 μm). (D) Cell seeding, (1) Immunophenotype of MSCs by flow cytometry, (2) Cytotoxicity (P > 0.05), (3) Cell proliferation on NDAs, (4) MSCs (green) seeded NDAs (H&E, bar: 200 μm), (5) MSCs seeded NDAs (bar: 100 μm), (6) Intima of MSCs seeded NDAs (SEM, bar: 50 μm), (7) Adventitia of MSCs seeded NDAs (SEM, bar: 10 μm). Figure 1. Nanosized decellularized arteries (NDAs). (A) Biological structures, (1) Gross view, (2) H&E (bar: 200 μm), (3) Masson’s trichrome staining (bar: 200 μm), (4) Van Gieson’s staining (bar: 200 μm). (B) Test, (1) DNA content (P < 0.001), (2) Suture strength (P > 0.05) Fresh arteries (FAs), (3) Burst pressure (P > 0.05), (4) Rehydration (P < 0.001). (C) Microstructure, (1) Cross-section of NDAs (SEM, bar: 50 μm), (2) Intima (SEM, bar: 10 μm), (3) Subintima and media (SEM, bar: 10 μm), (4) Adventitia (SEM, bar: 10 μm). (D) Cell seeding, (1) Immunophenotype of MSCs by flow cytometry, (2) Cytotoxicity (P > 0.05), (3) Cell proliferation on NDAs, (4) MSCs (green) seeded NDAs (H&E, bar: 200 μm), (5) MSCs seeded NDAs (bar: 100 μm), (6) Intima of MSCs seeded NDAs (SEM, bar: 50 μm), (7) Adventitia of MSCs seeded NDAs (SEM, bar: 10 μm). Statistical analysis. The suture retention test and burst pressure test were repeated at least three times. Data were expressed as means ± SD. The Student’s t test was used for comparison between groups. Statistical signifi- cance was at P < 0.05. Results Nanosized fiber decellularized scaffolds. The NDAs were white, with lumen diameter of 4.0 ± 0.5 mm and wall thickness of about 200 ± 25 μm (Fig. 1A1). All NDAs retained their original shape in PBS. H&E staining showed that all cellular components had been completely removed from the fresh arteries and that the remaining acelluar scaffold was a multilayer structure(Fig. 1A2). Masson’s trichrome staining (Fig. 1A3) and Van Gieson staining (Fig. 1A4) demonstrated the existence of collagen and elastin, respectively, throughout the scaffold. Masson’s trichrome staining also confirmed that smooth muscle and cytoplasm had been totally removed in the NDAs. DNA content was 1.62 ± 0.11 × 10−5 in NDAs vs. 27.78 ± 2.62 × 10−5 in fresh artery (Fig. 1B1). Mechanical testing showed no significant difference in suture retention between fresh artery and NDAs (0.61 ± 0.03N vs. 0.57 ± 0.02N, respectively; P > 0.05; Fig. 1B2). Burst pressure was also comparable between NDAs and fresh artery (297.0 ± 9.3 KPa vs. 310.9 ± 9.9 KPa, respectively; P > 0.05; Fig. 1B3). After rehydration, the NDAs showed a 6-fold increase in weight (P < 0.001; Fig. 1B4). SEM confirmed that no cells were left within the NDAs, and that the architecture mainly consisted of nanoscale- to microscale-sized collagen fibers. The cross-section of the NDAs revealed a porous structure with a uniform pore size. In the cross-sections the average distance between collagen fibers was about 20 ± 5 μm (Fig. 1C1−4).i MSCs exhibited similar fibroblastic morphology in normal culture. Flow cytometry analysis demonstrated hat most MSCs were positive for CD29(99.93%) and CD90(99.78%) and negative for CD45(97.88%) (Fig. 1D1) SCIEnTIFIC REPOrTS \| (2018) 8:4627 \| DOI:10.1038/s41598-018-22799-4 4 www.nature.com/scientificreports/ Figure 2. sustained-releasing system. (A) GMs, (1) Gross appearance, (2) SEM (bar: 5 μm), (3) Size distribution, (4) GMs-VEGF (SEM, bar: 5 μm). (B) CSs, (1) Gross appearance, (2) SEM (bar: 500 μm), (3) SEM (bar: 10 μm), (4) Distribution of pore size. (C) Character, (1) Rehydration of GMs, CSs and GMs-CSs (P < 0.001), (2) VEGF Releasing of from GMs-VEGF and CSs-GMs-VEGF (P < 0.05), (3) Cytotoxicity of GMs and CSs (P > 0.05), (4) Cell proliferation on CSs. Results (D) CSs complex (GMs-MSCs-CSs), (1) Gross appearance, (2) H&E (bar: 50 μm), (3) MSCs (green) proliferated on the NDAs (bar: 50 μm), (4) SEM (bar: 200 μm), (5) GMs in the CSs complex (SEM, bar: 20 μm), (6) MSCs and GMs in CSs complex (SEM, bar: 20 μm), (7) MSCs and ECM secreted by MSCs in CSs complex (SEM, bar: 50 μm), (8) ECM of MSCs (SEM, bar: 100 μm), CSs (Black arrow), GMs (White arrow), MSCs (Red arrow), ECM (Blue arrow). Figure 2. sustained-releasing system. (A) GMs, (1) Gross appearance, (2) SEM (bar: 5 μm), (3) Size distribution, (4) GMs-VEGF (SEM, bar: 5 μm). (B) CSs, (1) Gross appearance, (2) SEM (bar: 500 μm), (3) SEM (bar: 10 μm), (4) Distribution of pore size. (C) Character, (1) Rehydration of GMs, CSs and GMs-CSs (P < 0.001), (2) VEGF Releasing of from GMs-VEGF and CSs-GMs-VEGF (P < 0.05), (3) Cytotoxicity of GMs and CSs (P > 0.05), (4) Cell proliferation on CSs. (D) CSs complex (GMs-MSCs-CSs), (1) Gross appearance, (2) H&E (bar: 50 μm), (3) MSCs (green) proliferated on the NDAs (bar: 50 μm), (4) SEM (bar: 200 μm), (5) GMs in the CSs complex (SEM, bar: 20 μm), (6) MSCs and GMs in CSs complex (SEM, bar: 20 μm), (7) MSCs and ECM secreted by MSCs in CSs complex (SEM, bar: 50 μm), (8) ECM of MSCs (SEM, bar: 100 μm), CSs (Black arrow), GMs (White arrow), MSCs (Red arrow), ECM (Blue arrow). The proliferation rate of MSCs in medium containing NDAs was not significantly different from the rate in con- trol medium (Fig. 1D2). The porosity and variations in the size of the collagen fibers provided a three-dimensional space which was more conducive for MSCs proliferation than a plain dish (P < 0.05; Fig. 1D3). H&E staining (Fig. 1D4), fluorescent detection (Fig. 1D5), and SEM (Fig. 1D6) demonstrated MSCs adhesion and proliferation within the NDAs. Moreover, MSCs were also shown to secrete nanosized fibers of extracellular matrix (ECM) on the NDAs (Fig. 1D7). Gelatin microspheres, collagen scaffold, and the CSs complex. On gross examoniation, GMs appeared as a yellow, well-dispersed powder (Fig. 2A1). On SEM, the GMs were spherical, with an average diam- eter of 10 ± 5 μm (Fig. 2A2); the diameter was <5 μm in 29.0% of GMs, 5–10 μm in 52.7%, and >10 μm in 18.3% (Fig. 2A3). Results After rehydration, the GMs showed a 6-fold increase in weight (P < 0.001). For the incorporation of VEGF in GMs, entrapment efficiency of 89.0 ± 2.7%. The spherical shape of the GMs showed no change after loading VEGF (Fig. 2A4). The collagen scaffold was white, spongy, multiporous, and soft (Fig. 2B1–3). The pore size was mostly (78.0%) between 100 and 500 μm (Fig. 2B4). After rehydration the CSs showed a 25-fold increase SCIEnTIFIC REPOrTS \| (2018) 8:4627 \| DOI:10.1038/s41598-018-22799-4 5 www.nature.com/scientificreports/ Figure 3. Implantation and following examination. (A) Implantation, (1) TE artery in CG, (2) TE artery was surrounded by CSs complex in CG, (3) TE artery in NCG, (4) TE artery was surrounded by CSs complex in NCG, (5) The catheter secured on the skin for heparin infusion. (B) Angiography and anatomy, (1) Angiography of CG, (2) Angiography of NCG, (3) Collateral arteries (CAs) in NG, NCG and CG by angiography, (4) CAs arising from TE artery in CG, (5) CAs arising from TE artery in NCG, (6) Quantification of CAs. (C) Gross examination, (1) TE artery in CG, (2) Intima of TE artery in CG, (3) TE artery in NCG, (4) Thrombosis of TE artery in NCG, TE artery (Black arrow), CA (White arrow), catheter (Red arrow) and thrombosis (Blue arrow). Figure 3. Implantation and following examination. (A) Implantation, (1) TE artery in CG, (2) TE artery was surrounded by CSs complex in CG, (3) TE artery in NCG, (4) TE artery was surrounded by CSs complex in NCG, (5) The catheter secured on the skin for heparin infusion. (B) Angiography and anatomy, (1) Angiography of CG, (2) Angiography of NCG, (3) Collateral arteries (CAs) in NG, NCG and CG by angiography, (4) CAs arising from TE artery in CG, (5) CAs arising from TE artery in NCG, (6) Quantification of CAs. (C) Gross examination, (1) TE artery in CG, (2) Intima of TE artery in CG, (3) TE artery in NCG, (4) Thrombosis of TE artery in NCG, TE artery (Black arrow), CA (White arrow), catheter (Red arrow) and thrombosis (Blue arrow). in weight (P < 0.001; Fig. 2C1).We analyzed the VEGF release profile of GMs-VEGF and CSs-GMs-VEGF over a period of 14 days (Fig. 2C2). The VEGF release was significantly slower in CSs-GMs-VEGF than in GMs-VEGF (P < 0.01). Results Release from GMs-VEGF occurred in a sudden burst, with 27.10% of VEGF being released during the first 3 days. At the end of 14 days, 80.80% had been released. In contrast, VEGF release from CSs-GMs-VEGF presented a sustained profile, with only 17.57% of VEGF being released in the first 3 days; at the end of 14 days, 59.57% of the impregnated VEGF had been released. MSCs proliferation assays were used to evaluate the cyto- toxicity of GMs and CSs. The proliferation rate of MSCs cultured with GMs or CSs was not significantly different from that of MSCs cultured in control medium (Fig. 2C3). MSCs proliferation was significantly faster on the CSs than in a plain dish (P < 0.01; Fig. 2C4), indicating that the three-dimensional architecture of CSs is more suitable for cellular adhesion and proliferation.t p After multipoint injection seeding and agitation, GMs-VEGF and MSCs were evenly distributed within CSs (Fig. 2D1). The density of GMs-VEGF was about 34.0 ± 1.4 mg/cm3. H&E staining (Fig. 2D2) and fluorescence microscopy (Fig. 2D3) showed that GMs and MSCs were uniformly distributed within the CSs. SEM showed that the GMs and MSCs were evenly distributed in the pores of the CSs (Fig. 2D4), Both GMs (Fig. 2D5) and MSCs (Fig. 2D6) were found attached to the fibers of the CSs. SEM also demonstrated that MSCs secreted nanosized extracellular matrix fibers, which connected with the structure of CSs (Fig. 2D7–8). TE artery patency and collateral artery development. Angiography was performed to assess the patency of the implanted TE arteries and collaterals formation. While the TE arteries in the non-catheter group presented thrombosis and occlusion by the seventh day after transplantation, the arteries in the catheter group were still patent at 1 month after surgery. Many collateral arteries were visible around the TE arteries (Fig. 3B1). Collateral arteries were also visible in the non-catheter group although the grafts were completely occluded (Fig. 3B2). Angiography revealed 5 ± 1 collateral blood vessels in the control group, 10 ± 2 in the non-catheter group, and 18 ± 2 in the catheter group (Fig. 3B3); the difference between the groups was statistically significant (P < 0.01). With the assumption that the patent TE arteries in the catheter group had achieved endothelial integ- rity, we removed the catheters. SCIEnTIFIC REPOrTS \| (2018) 8:4627 \| DOI:10.1038/s41598-018-22799-4 www.nature.com/scientificreports/ www.nature.com/scientificreports/ Figure 4. Regeneration of TE artery in CG and NCG. (A) H&E (bar: 200 μm), Masson’s trichrome (bar: 200 μm) and Intima of TE artery (SEM, bar: 50 μm in CG, bar: 100 μm in NCG). (B) Endothelial layer, smooth muscle cell layer and vascularization within the adventitia of TE arteries (IHC, bar: 100 μm). (C) Quantification, (1) Continuity of endothelial layer, (2) Quantification of SMCs of TE artery, (3) Quantification of vessels formation in the adventitia of TE artery. (D) Colocalization of MSCs and ECs (I green indicated MSCs, II red showed ECs, III merge), (1) Intima of TE artery in CG, (2) Intima of TE artery in NCG, Endothelium (White arrow), SMC (Black arrow), blood vessels of adventitia (Red arrow). Figure 4. Regeneration of TE artery in CG and NCG. (A) H&E (bar: 200 μm), Masson’s trichrome (bar: 200 μm) and Intima of TE artery (SEM, bar: 50 μm in CG, bar: 100 μm in NCG). (B) Endothelial layer, smooth muscle cell layer and vascularization within the adventitia of TE arteries (IHC, bar: 100 μm). (C) Quantification, (1) Continuity of endothelial layer, (2) Quantification of SMCs of TE artery, (3) Quantification of vessels formation in the adventitia of TE artery. (D) Colocalization of MSCs and ECs (I green indicated MSCs, II red showed ECs, III merge), (1) Intima of TE artery in CG, (2) Intima of TE artery in NCG, Endothelium (White arrow), SMC (Black arrow), blood vessels of adventitia (Red arrow). closely integrated with the surrounding tissue. Dissecting or separating the TE arteries from the surrounding tissue was difficult and caused obvious bleeding. In the catheter group, many collateral arteries had formed from the wall of the TE artery. The collateral vessels were obvious on gross view, and angiography had earlier demon- strated contrast perfusion from the TE arteries and their collateral vessels into the native distal vessels (Fig. 3B4); unfortunately the TE arteries had become occluded and thrombosed after the heparin infusion catheter was pulled out. The non-catheter group also showed collateral vessels formation around the TE arteries (Fig. 3B5). On gross observation there were significantly more collateral branches arising from the TE arteries in the catheter group (5 ± 1 vs. 3 ± 1 in the non-catheter group; P < 0.05); no new blood vessels were visible in the control group (Fig. 3B6). Results However, 1 week after catheter removal, sonography showed that the TE arteries had become occluded.ht The TE arteries were harvested at 40 days after transplantation. In both the catheter and the non-cath groups, the CSs-GMS-VEGF-MSCs complex was found to have completely degraded. The TE arteries w SCIEnTIFIC REPOrTS \| (2018) 8:4627 \| DOI:10.1038/s41598-018-22799-4 6 www.nature.com/scientificreports/ y ( g ) On examination of the luminal surface of the TE arteries, 91.7% was covered by endothelial cells in the cath- eter group vs. 26.7% in the non-catheter group (Fig. 4C1). There were about 2.275 ± 0.338 × 103 cells/mm2 in the catheter group vs. 1.292 ± 0.191 × 103 cells/mm2 in non-catheter group (Fig. 4C2). Blood vessel formation in the adventitia of TE arteries was more in the catheter group (13 ± 2 vs. 6 ± 1 in the non-catheter group (Fig. 4C3). The results of immunofluorescent staining also supported these findings. The endothelial cell lining on the intimal surface of the TE arteries in the catheter group (Fig. 4D1) was almost continuous, whereas it was discontinuous in the non-catheter group (Fig. 4D2). The data also indicated that some implanted green fluorescent MSCs on the intima of TE arteries of catheter and non-catheter group were co-localized with red fluorescent endothelial cells on the intima. The results proved that MSCs differentiated into endothelial cells. Angiogenesis, vasculogenesis, and arteriogenesis in the tissue around the TE artery. H&E and immunohistochemical staining showed that a large number blood vessels had formed in the tissues sur- rounding the TE arteries in both groups; however, there were few blood vessels in the tissues of the control group (Fig. 5A). The mean number of blood vessels was 15 ± 2 in the catheter group, 14 ± 1 in the non-catheter group, and 4 ± 1 in the control group (Fig. 5B1); the difference between the catheter and non-catheter groups was not statistically significant (P > 0.05). Immunohistochemical analysis for vWF found 8 ± 1 blood vessels in the catheter group vs. 8 ± 1 blood vessels in the non-catheter group and 4 ± 1 blood vessels in the control group (Fig. 5B2). Immunofluorescent analysis revealed more capillaries in the tissues around TE arteries in the catheter and non-catheter groups than in the control group. However, immunohistochemical staining for α-SMA indi- cated a significant difference between the two groups (9 ± 1 blood vessels in the catheter group vs. 6 ± 1 in the non-catheter group; P < 0.01); there were 4 ± 1 vessels in the control group (Fig. 5B3), which suggested that more mature blood vessels formed in the catheter group than in other groups. Discussion In this study, we employed the Vineberg procedure to implant a TE artery and, with sustained delivery of angio- genic growth factor and autologous MSCs, successfully established adequate blood supply to ischemic hind limbs of goats. Angiogenesis, vasculogenesis, and arteriogenesis were stimulated around the TE artery and a dense collateral circulation formed, with wide communication with the native vasculature.l When the distal part of the TE artery is ligated, preventing outflow, the blood within the artery will coagulate. To prevent this, we inserted a catheter into the artery and instituted continuous heparin infusion in one group. These TE arteries remained patent until the catheter was removed at the end of 1 month, whereas thrombo- sis and occlusion occurred within 7 days in the TE vascular grafts that did not receive the heparin infusion. Branches arising from the artery help maintain steady blood flow through the artery and thus prevent thrombo- sis. Following implantation of the TE artery, the sustained release of VEGF from the gelatin microspheres and the autologous MSCs within the CSs stimulated the growth of collateral vessels, which formed connections with the native microcirculation. The importance of a good collateral circulation has long been known. The mechanisms governing the recruitment, growth, and proliferation of collateral vessels differs from those regulating angiogenesis and vas- culogenesis31. In this study, for the development of a collateral circulation, new vessels had to sprout from the extra-anatomic TE artery wall, which was composed of natural ECM and therefore had ideal biocompatibility and mechanical retention characteristics. The decellularized vascular scaffold provided structural support and allowed for MSCs migration, growth, and differentiation and cellular ECM production32. It was expected that parts of scaffolds would degrade and allow collateral vessels development from the degraded locations33. pf g p g It is now accepted that neovascularization in the adult occurs by both angiogenesis and vasculogenesis34. In this study, angiogenesis and vasculogenesis were demonstrated around the TE artery and ischemic muscle. Our own preliminary studies and work by others have established that to achieve a mature vasculature, sustained release of cytokines and endothelial progenitor cells are needed at the ischemic site in a pattern that mimics the natural processes35,36. Previous studies have demonstrated that inflammatory cytokines stimulate the develop- ment of arteroles, venules, and capillaries. In this research, we used a collagen scaffold complex containing gelatin microspheres formulated for sustained VEGF release to promote angiogenesis. www.nature.com/scientificreports/ The TE arteries in the catheter group had smooth luminal surfaces (Fig. 3C1–2), but in the non-catheter group the lumina were completely occluded (Fig. 3C3–4). closely integrated with the surrounding tissue. Dissecting or separating the TE arteries from the surrounding tissue was difficult and caused obvious bleeding. In the catheter group, many collateral arteries had formed from the wall of the TE artery. The collateral vessels were obvious on gross view, and angiography had earlier demon- strated contrast perfusion from the TE arteries and their collateral vessels into the native distal vessels (Fig. 3B4); unfortunately the TE arteries had become occluded and thrombosed after the heparin infusion catheter was pulled out. The non-catheter group also showed collateral vessels formation around the TE arteries (Fig. 3B5). On gross observation there were significantly more collateral branches arising from the TE arteries in the catheter group (5 ± 1 vs. 3 ± 1 in the non-catheter group; P < 0.05); no new blood vessels were visible in the control group (Fig. 3B6). The TE arteries in the catheter group had smooth luminal surfaces (Fig. 3C1–2), but in the non-catheter group the lumina were completely occluded (Fig. 3C3–4). TE artery regeneration and its adventitial angiogenesis. In the catheter group, H&E stained sections of the TE arteries showed the presence of all three layers of the arterial wall (endothelium, media, and adventi- tia). Masson’s trichrome staining confirmed the presence of regular collagen and smooth muscle throughout the vascular graft, and SEM confirmed that the endothelium at 40 days post-implantation was similar to the native endothelium. In the non-catheter group, H&E staining of TE arteries showed that the three-layer structure of the arterial wall was incomplete. Masson’s trichrome staining showed irregular laying down of collagen, and SEM demonstrated an incomplete and rough endothelium (Fig. 4A). On immunohistochemical analysis the luminal surface of the TE arteries in the catheter group was positive for vWF, indicating endothelial regeneration, and the medial layer was positive for α-SMA, indicating smooth muscle regeneration. Relatively complete smooth muscle SCIEnTIFIC REPOrTS \| (2018) 8:4627 \| DOI:10.1038/s41598-018-22799-4 7 www.nature.com/scientificreports/ cell layer has been formed in the TE artery of the catheter group. In the non-catheter group, it shows the poor endothelial regeneration on the luminal surface and the scattered and disorderly arrangement of smooth muscle cells in TE artery (Fig. 4B). www.nature.com/scientificreports/ Some of the implanted MSCs in the col- lagen complex around the TE arteries were found to have differentiated into endothelial cells (Fig. 5C). Discussion (C) Colocalization of MSCs and ECs (I green indicated MSCs, II red showed ECs, III merge), (1) Endothelium of TE artery and angiogenesis around TE artery in CG, (2) Endothelium of TE artery and angiogenesis around TE artery in NCG, blood vessels (White arrow), ECs positive vessel (Black arrow), SMCs positive vessel (Red arrow). Figure 5. Angiogenesis of surrounding tissues in CG, NCG and NG. (A) Angiogenesis (HE, bar: 200 μm in CG/ NCG, bar: 100 μm in NG), ECs (IHC, bar: 100 μm) and SMCs (IHC, bar: 100 μm). (B) Quantification, (1) New vessels formation (HE), (2) ECs positive vessels, (3) SMCs positive vessels. (C) Colocalization of MSCs and ECs (I green indicated MSCs, II red showed ECs, III merge), (1) Endothelium of TE artery and angiogenesis around TE artery in CG, (2) Endothelium of TE artery and angiogenesis around TE artery in NCG, blood vessels (White arrow), ECs positive vessel (Black arrow), SMCs positive vessel (Red arrow). Conclusion In this study, we employed the Vineberg procedure in an ischemic hind limb model and successfully developed an extra-anatomic TE artery with a large number of collateral artery branches. We used MSCs and sustained-release GMs-VEGF formulations to stimulate angiogenesis, vasculogenesis, and arteriogenesis. Although some throm- bosis occurred in the TE arteries 1 month after implantation, we have demonstrated the feasibility of the tech- nique. Full development of the TE artery, with complete endothelialization and many outflow collaterals, may require many more weeks. Discussion This system showed a stable VEGF release profile and performed better than collagen scaffold or gelatin microsphere alone. pi p gf g p In this study, autologous MSCs were used to achieve vasculogenesis. Vasculogenesis is the process of in situ formation of blood vessels from endothelial progenitor cells or angioblasts. It has been demonstrated that bone-marrow-derived endothelial precursor cells circulate in the peripheral blood and can incorporate into areas of neovascularization in the adult. Initially, mesenchymal cells differentiate in situ into primitive endothelial cells that generate a functioning vascular labyrinth36,37. g g y Our results show that the primitive vascular plexus subsequently develops into a complex, interconnecting network of mature blood vessels. Previous investigations have shown that MSCs can promote therapeutic angio- genesis and vasculogenesis, playing a key role in capillary endothelium formation and differentiating into smooth muscle cells or pericytes38,39. We hypothesized that autologous MSCs would not stimulate an immunologic reac- tion and that they would gather around the TE artery and facilitate the formation of arterial branches. We showed that the MSCs could differentiate into endothelial cells, which were involved in vasculogenesis around the TE arteries. SCIEnTIFIC REPOrTS \| (2018) 8:4627 \| DOI:10.1038/s41598-018-22799-4 8 www.nature.com/scientificreports/ Figure 5. Angiogenesis of surrounding tissues in CG, NCG and NG. (A) Angiogenesis (HE, bar: 200 μm in CG/ NCG, bar: 100 μm in NG), ECs (IHC, bar: 100 μm) and SMCs (IHC, bar: 100 μm). (B) Quantification, (1) New vessels formation (HE), (2) ECs positive vessels, (3) SMCs positive vessels. (C) Colocalization of MSCs and ECs (I green indicated MSCs, II red showed ECs, III merge), (1) Endothelium of TE artery and angiogenesis around TE artery in CG, (2) Endothelium of TE artery and angiogenesis around TE artery in NCG, blood vessels (White arrow), ECs positive vessel (Black arrow), SMCs positive vessel (Red arrow). Figure 5. Angiogenesis of surrounding tissues in CG, NCG and NG. (A) Angiogenesis (HE, bar: 200 μm in CG/ NCG, bar: 100 μm in NG), ECs (IHC, bar: 100 μm) and SMCs (IHC, bar: 100 μm). (B) Quantification, (1) New vessels formation (HE), (2) ECs positive vessels, (3) SMCs positive vessels. References 1. Liang, G. Z. & Zhang, F. X. Novel devices and specialized techniques in recanalization of peripheral artery chronic total occlusions (CTOs) - A literature review. International journal of cardiology 165, 423–429, https://doi.org/10.1016/j.ijcard.2012.03.033 (2013). 2. Stoner, M. C. et al. Reporting standards of the Society for Vascular Surgery for endovascular treatment of chronic lower extremity peripheral artery disease. Journal of vascular surgery 64, e1–e21, https://doi.org/10.1016/j.jvs.2016.03.420 (2016). 3. Reinecke, H. et al. Peripheral arterial disease and critical limb ischaemia: still poor outcomes and lack of guideline adherence. European heart journal 36, 932–938A, https://doi.org/10.1093/eurheartj/ehv006 (2015). www.nature.com/scientificreports/ & Ikada, Y. Neovascularization effect of biodegradable gelatin microspheres incorpor basic fibroblast growth factor. Journal of biomaterials science. Polymer edition 10, 79–94 (1999). Th Th b l l l f b l T i 16. Thomas, J. L. The Vineberg legacy: internal mammary artery implantation from inception to obsolescence. Texas Heart Institute journal 26, 107–113 (1999).h j 17. Vineberg, A. M. The Vineberg operation. I. Revascularization of the heart. Jama 195, Suppl: 43–47 c (1966). h 18. Lobo Filho, J. G., Forte, A. J. & Leitao, M. C. Angiographic follow-up of myocardial revascularization using the vineberg procedure correlated with intraoperative imaging. Arquivos brasileiros de cardiologia 90, e8–9 (2008). Forte, A. J. & Leitao, M. C. Angiographic follow-up of myocardial re h 18. Lobo Filho, J. G., Forte, A. J. & Leitao, M. C. Angiographic follow-up of myocardial revasculari correlated with intraoperative imaging. Arquivos brasileiros de cardiologia 90, e8–9 (2008). 19. Johnson, W. D., Chekanov, V. S., Nikolaychik, V. V. & Kipshidze, N. Vineberg procedure combined with therapeutic angiogenesis: old wine in a new bottle. The Annals of thoracic surgery 72, 1443–1444 (2001). h f g y 0. Schmidt, C. E. & Baier, J. M. Acellular vascular tissues: natural biomaterials for tissue repair and tissue engineering. Biomaterials 21 2215–2231 (2000). ( ) 1. Tan, Q. et al. Controlled release of chitosan/heparin nanoparticle-delivered VEGF enhances regeneration of decellularized tissuef 21. Tan, Q. et al. Controlled release of chitosan/heparin nanoparticle-delivered VEGF enhances regeneration of decellularized tissue- engineered scaffolds. Int J Nanomed 6, 929–942, https://doi.org/10.2147/Ijn.S18753 (2011).f f 22. Hirata, E. et al. 3D collagen scaffolds coated with multiwalled carbon nanotubes: initial cell attachment to internal surface. Journal of biomedical materials research. Part B, Applied biomaterials 93, 544–550, https://doi.org/10.1002/jbm.b.31613 (2010). 23. Matsumura, G., Miyagawa-Tomita, S., Shin’oka, T., Ikada, Y. & Kurosawa, H. First evidence that bone marrow cells contribute t construction of tissue-engineered vascular autografts in vivo. Circulation 108, 1729–1734, https://doi.org/10.116 cir.0000092165.32213.61 (2003).f 4. Wang, J. et al. A cellular delivery system fabricated with autologous BMSCs and collagen scaffold enhances angiogenesis and perfusion in ischemic hind limb. Journal of biomedical materials research. Part A 100, 1438–1447, https://doi.org/10.1002 jbm.a.34081 (2012).hf j 25. Zhao, Y. et al. The development of a tissue-engineered artery using decellularized scaffold and autologous ovine mesenchymal cells. Biomaterials 31, 296–307, https://doi.org/10.1016/j.biomaterials.2009.09.049 (2010).ti p g j 26. Zhou, X. et al. References 1. Liang, G. Z. & Zhang, F. X. Novel devices and specialized techniques in recanalization of peripheral artery chronic total occlusion (CTOs) - A literature review. International journal of cardiology 165, 423–429, https://doi.org/10.1016/j.ijcard.2012.03.033 (2013). 2. Stoner, M. C. et al. Reporting standards of the Society for Vascular Surgery for endovascular treatment of chronic lower extremity peripheral artery disease. Journal of vascular surgery 64, e1–e21, https://doi.org/10.1016/j.jvs.2016.03.420 (2016). SCIEnTIFIC REPOrTS \| (2018) 8:4627 \| DOI:10.1038/s41598-018-22799-4 9 www.nature.com/scientificreports/ www.nature.com/scientificreports/ 4. Park, S. J. et al. Role of surgical treatment for peripheral arterial disease in endovascular era. Journal of the Korean Surgical Society 84, 353–359, https://doi.org/10.4174/jkss.2013.84.6.353 (2013).f 4. Park, S. J. et al. Role of surgical treatment for peripheral arterial disease in endovascular era. Journal of the Korean Surgical Society 84, 353–359, https://doi.org/10.4174/jkss.2013.84.6.353 (2013). 5 Panico A Jafferani A Shah F & Dieter R S Advances in peripheral arterial disease endovascular revascularization Cardiology 84, 353–359, https://doi.org/10.4174/jkss.2013.84.6.353 (2013). 5. Panico, A., Jafferani, A., Shah, F. & Dieter, R. S. Advances in peripheral arterial disease endovascular revascularization. Cardiology l h d l p g j 5. Panico, A., Jafferani, A., Shah, F. & Dieter, R. S. Advances in peripheral arterial disease endovascular revascularization. Cardiology clinics 33, 89–98, https://doi.org/10.1016/j.ccl.2014.09.002 (2015).h p g j ( ) 6. Annex, B. H. Therapeutic angiogenesis for critical limb ischaemia. Nature reviews. Cardiology 10, 387–396, https://doi.org/10.1038 nrcardio.2013.70 (2013).h 7. Harjai, K. J., Chowdhury, P. & Grines, C. L. Therapeutic angiogenesis: a fantastic new adventure. Journal of interventional cardiology 15, 223–229 (2002). 8. Deev, R., Plaksa, I., Bozo, I. & Isaev, A. Results of an International Postmarketing Surveillance Study of pl-VEGF165 Safety and Efficacy in 210 Patients with Peripheral Arterial Disease. American journal of cardiovascular drugs: drugs, devices, and other interventions, https://doi.org/10.1007/s40256-016-0210-3 (2017). p g 9. Mikroulis, D., Papanas, N., Maltezos, E. & Bougioukas, G. Angiogenic growth factors in the treatment of peripheral arterial disease. Current vascular pharmacology 5, 195–209 (2007). p gy 0. Lachmann, N. & Nikol, S. Therapeutic angiogenesis for peripheral artery disease: stem cell therapy. VASA. Zeitschrift fu Gefasskrankheiten 36, 241–251, https://doi.org/10.1024/0301-1526.36.4.241 (2007). f p g 1. Staton, C. A. et al. Current methods for assaying angiogenesis in vitro and in vivo. International journal of experimental pathology 85 233–248, https://doi.org/10.1111/j.0959-9673.2004.00396.x (2004). p g j ( ) 12. Zhao, H. Y. et al. Research Advances in Tissue Engineering Materials for Sustained Release of Growth Factors. BioMed research international 2015, 808202, https://doi.org/10.1155/2015/808202 (2015). g 13. Jin, Y. et al. Biodegradable gelatin microspheres enhance the neuroprotective potency of osteopontin via quick and sustained re in the post-ischemic brain. Acta biomaterialia 10, 3126–3135, https://doi.org/10.1016/j.actbio.2014.02.045 (2014). 14. Layman, H. et al. The effect of the controlled release of basic fibroblast growth factor from ionic gelatin-based hydrogels on angiogenesis in a murine critical limb ischemic model. Biomaterials 28, 2646–2654, https://doi.org/10.1016/j.biomaterials.2007. 01.044 (2007).f 15. Tabata, Y., Hijikata, S., Muniruzzaman, M. www.nature.com/scientificreports/ Engineered Vascular Graft Using Nanoscale Decellularized Arteries Modified with Controlled-Release Heparin and Vascular Endothelial Growth Factor. Journal of Biomaterials and Tissue Engineering 6, 870–882, https://doi.org/10.1166/ jbt.2016.1516 (2016). j ( ) 7. Zhao, Y. et al. Preparation of gelatin microspheres encapsulated with bFGF for therapeutic angiogenesis in a canine ischemic hind limb. Journal of biomaterials science. Polymer edition 22, 665–682, https://doi.org/10.1163/092050610x489880 (2011).hi 28. Zhou, J. et al. Therapeutic angiogenesis using basic fibroblast growth factor in combination with a collagen matrix in ch hindlimb ischemia. The Scientific World Journal 2012, 652794, https://doi.org/10.1100/2012/652794 (2012).i h fi p g 29. Rolandsson Enes, S. et al. MSC from fetal and adult lungs possess lung-specific properties compared to bone marrow-derived MSC. Scientific reports 6, 29160, https://doi.org/10.1038/srep29160 (2016).f Scientific reports 6, 29160, https://doi.org/10.1038/srep29160 (2 fi p p g p ( ) 30. Kadekar, D., Kale, V. & Limaye, L. Differential ability of MSCs isolated from placenta and cord as feeders for supporting ex vivo expansion of umbilical cord blood derived CD34(+) cells. Stem cell research & therapy 6, 201, https://doi.org/10.1186/s13287-015- 0194-y (2015).t y ( ) 31. Syedain, Z. et al. Tissue engineering of acellular vascular grafts capable of somatic growth in young lambs. Nature communications 7, 12951, https://doi.org/10.1038/ncomms12951 (2016). p g 32. Belotti, D., Foglieni, C., Resovi, A., Giavazzi, R. & Taraboletti, G. Targeting angiogenesis with compounds from the extracel matrix. The international journal of biochemistry & cell biology 43, 1674–1685, https://doi.org/10.1016/j.biocel.2011.08.012 (20 h j f y gy g j 33. Tao, Z. et al. Coexpression of VEGF and angiopoietin-1 promotes angiogenesis and cardiomyocyte proliferation reduces apoptosis in porcine myocardial infarction (MI) heart. Proceedings of the National Academy of Sciences of the United States of America 108, 2064–2069, https://doi.org/10.1073/pnas.1018925108 (2011). p g p 34. Patan, S. Vasculogenesis and angiogenesis. Cancer treatment and research 117, 3–32 (2004). Vasculogenesis and angiogenesis. Cancer treatment and research 11 34. Patan, S. Vasculogenesis and angiogenesis. Cancer treatment an g g g 35. Smadja, D. M. et al. Increased VEGFR2 expression during human late endothelial progenitor cells expansion enhances in vitro angiogenesis with up-regulation of integrin alpha(6). Journal of cellular and molecular medicine 11, 1149–1161, https://doi. org/10.1111/j.1582-4934.2007.00090.x (2007). 35. Smadja, D. M. et al. Increased VEGFR2 expression during human late endothelial progenitor cells expansion enhances in vitro angiogenesis with up-regulation of integrin alpha(6). Journal of cellular and molecular medicine 11, 1149–1161, https://doi. org/10.1111/j.1582-4934.2007.00090.x (2007). g j ( ) 36. Guerin, C. L. Author Contributions Xu Zhou, Yinlong Zhang, Hongfei Wang, Bin Zhao, Jinling Wang, Guoliang Yan, Shuangyue Xu, Yuanyuan Zhou, Hongyi Liu, Yifei Zheng, Wei Quan, Jianyin Zhou,Yun Liu, Maochuan Zhen performed the experiments, collected and analyzed the data; Xuan Zhu partly supported this research, and Yilin Zhao supported, designed this research and wrote the manuscript. www.nature.com/scientificreports/ et al. Bone-marrow-derived very small embryonic-like stem cells in patients with critical leg ischaemia: eviden vasculogenic potential. Thrombosis and haemostasis 113, 1084–1094, https://doi.org/10.1160/th14-09-0748 (2015). g ph p g 37. Nassiri, S. M. & Rahbarghazi, R. Interactions of mesenchymal stem cells with endothelial cells. Stem cells and development 23, 319–332, https://doi.org/10.1089/scd.2013.0419 (2014). gh g 37. Nassiri, S. M. & Rahbarghazi, R. Interactions of mesenchymal stem cells with endothelial cells. Stem cells and development 23, 319–332, https://doi.org/10.1089/scd.2013.0419 (2014). g 38. Gu, W., Hong, X., Potter, C., Qu, A. & Xu, Q. Mesenchymal stem cells and vascular regeneration. Microcirculation (New York, N.Y. 1994) 24, https://doi.org/10.1111/micc.12324 (2017). 38. Gu, W., Hong, X., Potter, C., Qu, A. & Xu, Q. Mesenchymal stem cells and vascular regeneration. Microcirculation (New York, N.Y. 1994) 24, https://doi.org/10.1111/micc.12324 (2017). 1994) 24, https://doi.org/10.1111/micc.12324 (2017). p g 9. Moraes, F. et al. Endothelial cell-dependent regulation of arteriogenesis. Circulation research 113, 1076–1086, https://doi org/10.1161/circresaha.113.301340 (2013). SCIEnTIFIC REPOrTS \| (2018) 8:4627 \| DOI:10.1038/s41598-018-22799-4 10 www.nature.com/scientificreports/ Acknowledgementsh g This work was supported by the National Natural Science Foundation of China [grant number 81170290 and 81770294]. SCIEnTIFIC REPOrTS \| (2018) 8:4627 \| DOI:10.1038/s41598-018-22799-4 © The Author(s) 2018 Additional Informationh Competing Interests: The authors declare no competing interests. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps an institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not per- mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. © The Author(s) 2018 SCIEnTIFIC REPOrTS \| (2018) 8:4627 \| DOI:10.1038/s41598-018-22799-4 11
https://openalex.org/W2075027588	https://www.frontiersin.org/articles/10.3389/fphys.2014.00397/pdf	English	null	Intermittent hypoxic resistance training: does it provide added benefit?	Frontiers in physiology	2,014	cc-by	3,614	Edited by: Edited by: J.-P. Jin, Wayne State University School of Medicine, USA Reviewed by: Fan Ye, University of Florida, USA J.-P. Jin, Wayne State University School of Medicine, USA Keywords: muscle, hypertrophy, strength, metabolic stress, training adaptation FINDINGS FROM IHRT STUDIES To date, six separate investigations have examined the impact of IHRT on hypertrophic and strength responses, resulting in seven published papers comparing adaptive responses follow- ing IHRT to a normoxic control group (Friedmann et al., 2003; Nishimura et al., 2010; Manimmanakorn et al., 2013a,b; Ho et al., 2014b; Kon et al., 2014; Kurobe et al., 2014). Research using low-load resistance training (20% 1RM) combined with moderate hypoxia (frac- tion of inspired oxygen [FIO2] adjusted to maintain arterial oxygen saturation at 80%) and very brief inter-set rest periods (30 s) has reported greater hyper- trophic and strength responses following IHRT compared to work-matched nor- moxic training (Manimmanakorn et al., 2013a,b). However, another study using similar exercise loads (30% 1RM), yet longer inter-set rest periods (60 s) and a greater hypoxic stress (FIO2 = 12%) has observed no additive beneﬁts of IHRT (Friedmann et al., 2003). While conﬂicting, these ﬁndings may indicate The addition of systemic hypoxia to resistance training has previously resulted in signiﬁcantly enhanced hypertrophic and strength responses to both low- load (20% 1-repetition maximum; 1RM) (Manimmanakorn et al., 2013a,b) and moderate-load (70% 1RM) (Nishimura et al., 2010) resistance training. While research into intermittent hypoxic resis- tance training (IHRT) is in its infancy, some studies have reported conﬂicting results, which is likely due to differ- ing research methodologies. In a recent review, it has been suggested that many of the potential mechanisms underpinning OPINION ARTICLE published: 13 October 2014 doi: 10.3389/fphys.2014.00397 Intermittent hypoxic resistance training: does it provide added beneﬁt? OPINION ARTICLE published: 13 October 2014 doi: 10.3389/fphys.2014.00397 Brendan R. Scott 1*, Katie M. Slattery 1,2 and Ben J. Dascombe 1 1 Applied Sports Science and Exercise Testing Laboratory, Faculty of Science and Information Technology, University of Newcastle, Ourimbah, NSW, Australia 2 New South Wales Institute of Sport, Sydney, NSW, Australia rcise Testing Laboratory, Faculty of Science and Information Technology, University of Newcastle, Ourimbah, NSW, Australia port, Sydney, NSW, Australia INTRODUCTION muscle adaptations to BFR training and IHRT are linked to the muscular oxygena- tion status and degree of metabolic stress associated with exercise (Scott et al., 2014). The purpose of this paper is to brieﬂy sum- marize the adaptive responses that have been reported following both low- and moderate-load IHRT and to highlight key areas of concern for IHRT methodology, including the level of hypoxia used and the degree of metabolic stress imposed during exercise. that for IHRT using low-loads, both the duration of inter-set rest periods and the level of hypoxia affect the adaptive responses. Methods to enhance the adaptive responses to resistance training are of great interest to clinical and athletic populations alike. Altering the muscular environment by restricting oxygen avail- ability during resistance exercise has been shown to induce favorable physiological adaptations. An acute hypoxic stimu- lus during exercise essentially increases reliance on anaerobic pathways, aug- menting metabolic stress responses, and subsequent hypertrophic processes (Scott et al., 2014). Hypoxic strategies during resistance exercise were originally investi- gated using blood ﬂow restriction (BFR) methods (Takarada et al., 2000), whereby a cuff is applied proximally to a limb to partially limit arterial inﬂow while occlud- ing venous outﬂow from the working muscles. Another method that has been investigated more recently is performing resistance exercise in systemic hypoxia, by means of participants breathing a hypoxic air mixture. p Nishimura et al. (2010) used moderate- load resistance training (70% 1RM) combined with moderate-level hypoxia (FIO2 = 16%) and a relatively brief inter-set rest period (60 s), demonstrat- ing enhanced hypertrophic, and strength responses following IHRT compared to the equivalent training in normoxia. However, research using similar exer- cise loads (70% 1RM or 10RM) and levels of hypoxia (FIO2 = 14.4–15%) in conjunction with longer inter-set rest periods (90–120 s) has not found additive hypertrophic or strength ben- eﬁts for IHRT (Ho et al., 2014b; Kon et al., 2014). Furthermore, a study that employed moderate-load exercise (10RM) and brief inter-set rest (60 s), but a high- level of hypoxia (FIO2 = 12.7%) noted signiﬁcantly greater hypertrophic, but not strength adaptations, following IHRT compared to normoxic training (Kurobe et al., 2014). When considering the avail- able evidence, it appears that moderate hypoxia in conjunction with relatively brief inter-set rest periods (and sub- sequently increased metabolic stress) are paramount in enhancing muscular development via IHRT. www.frontiersin.org METABOLIC STRESS This is a probable differ- ence between these two resistance train- ing methods. Although further research is needed to examine the dose-response relationship between the level of hypoxia during IHRT and subsequent muscle It is possible that for enhanced mus- cle hypertrophy and strength development from IHRT, the level of hypoxia may fol- low a hermetic relationship, meaning that some beneﬁcial acute responses to hypoxia may be attenuated if the level of hypoxia is too high. Muscle function during IHRT using high-levels of hypoxia could be impacted by the presence of a “central gov- ernor,” which was popularized by Noakes et al. (2001). This theory postulates that the degree of motor unit recruitment by the central nervous system is determined by the brain’s need to protect itself and the body, ensuring survival, and maintenance of integrity during and following exercise (Millet et al., 2009). As such, a hypoxia- mediated reduction in central drive may occur above a certain hypoxic threshold. Amann et al. (2006) have hypothesized that oxygen supply affects the regulation of motor output to ensure that muscular fatigue does not exceed a critical threshold. A mechanism by which hypoxia may alter energetic metabolism during resis- tance exercise may by slowing the rate of phosphocreatine (PCr) recovery between sets. Resynthesis of PCr occurs primar- ily by oxidative processes, and is therefore sensitive to manipulations of oxygen avail- ability (Haseler et al., 1999). It is possible that relatively brief inter-set rest periods will result in subsequent sets beginning with a lower PCr concentration when training in hypoxia, placing greater stress on anaerobic glycolysis, and consequently increasing the accumulation of metabo- lites. This may account for the ﬁndings of no added hypertrophic or strength beneﬁts following IHRT in previous research that has employed inter-set rest periods of 90– 120 s (Ho et al., 2014b; Kon et al., 2014). Indeed, Ho et al. (2014a) have reported no differences in blood lactate concen- trations between hypoxic and normoxic IHRT groups following 5 sets of 15 repeti- tions of squats (30% 1RM) with 90 s inter- set rest. Therefore, it is likely that inter-set rest periods of 90 s or longer are sufﬁ- cient to attenuate any hypoxia-mediated rise in metabolic stress, limiting the poten- tial anabolic effects of hypoxia. LEVEL OF HYPOXIA The hypoxic stimulus used in previ- ous IHRT investigations has ranged from FIO2 = 12–16%, or investigators have adjusted the hypoxic stimulus to main- tain arterial oxygen saturation at 80%. October 2014 \| Volume 5 \| Article 397 \| 1 www.frontiersin.org www.frontiersin.org Hypoxic resistance training Scott et al. Importantly, no two investigations have employed the same hypoxic stimulus, which has led to differing physiological responses, and difﬁculty in directly com- paring the ﬁndings of each study. It has been established that the level of hypoxia or altitude an individual is exposed to has a dose-response relationship on subse- quent markers of endurance performance (Chapman et al., 2014), and it is logical that a similar relationship may exist for muscular development following IHRT. in this regard, given that venous out- ﬂow is occluded by the BFR stimulus and metabolites therefore cannot be re- distributed away from the exercising limb. However, venous outﬂow is maintained during IHRT, meaning that metabolites can enter circulation and be distributed to other parts of the body. Logic there- fore dictates that for augmented metabolic stress during IHRT, inter-set rest periods should be short enough to ensure that a hypoxia-mediated metabolic stress is still present within the muscles during each subsequent set. activation, current evidence suggests that beneﬁcial muscular adaptations are only possible when moderate-level hypoxia is employed in conjunction with brief inter- set rest periods (Nishimura et al., 2010; Manimmanakorn et al., 2013a,b). Frontiers in Physiology \| Striated Muscle Physiology METABOLIC STRESS The degree of metabolic stress associated with resistance training has been proposed as an important regulator of subsequent adaptive muscular responses (Schoenfeld, 2013). Intramuscular hypoxia dur- ing resistance exercise likely increases the reliance on anaerobic metabolism (Kawada, 2005), which accelerates the production of metabolites. Indeed, researchers have previously demonstrated that IHRT (FIO2 = 13%) using both low- loads (5 sets of 14 repetitions at 50% 1RM with 60 s inter-set rest) (Kon et al., 2012) and moderate-loads (5 sets of 10 repeti- tions at 70% 1RM with 60 s inter-set rest) (Kon et al., 2010) can increase metabolic stress, measured via blood lactate concen- tration, when compared to the equivalent exercise in normoxia. As such, increased metabolic stress may be a primary mecha- nism underpinning augmented muscular responses to hypoxic resistance training methods. p y muscular development following IHRT. It is possible that for enhanced mus- cle hypertrophy and strength development from IHRT, the level of hypoxia may fol- low a hermetic relationship, meaning that some beneﬁcial acute responses to hypoxia may be attenuated if the level of hypoxia is too high. Muscle function during IHRT using high-levels of hypoxia could be impacted by the presence of a “central gov- ernor,” which was popularized by Noakes et al. (2001). This theory postulates that the degree of motor unit recruitment by the central nervous system is determined by the brain’s need to protect itself and the body, ensuring survival, and maintenance of integrity during and following exercise (Millet et al., 2009). As such, a hypoxia- mediated reduction in central drive may occur above a certain hypoxic threshold. Amann et al. (2006) have hypothesized that oxygen supply affects the regulation of motor output to ensure that muscular fatigue does not exceed a critical threshold. It has been previously reported that exercising in systemic hypoxia can induce changes in cerebral oxygenation, which in turn may limit incremental exercise per- formance (Subudhi et al., 2007). Altered cerebral oxygenation may play a reg- ulatory role during IHRT using high- level hypoxia, effectively attenuating any increases in muscle activation that may occur using moderate-level hypoxia. This cerebral mechanism is unlikely to be present during BFR exercise, where the hypoxic environment is localized to the limb being trained and signiﬁcant eleva- tions in muscle activation are regularly reported (Takarada et al., 2000; Yasuda et al., 2009). METABOLIC STRESS Schoenfeld (2013) proposed hyper- trophic adaptations to resistance exercise may be mediated by metabolic stress via enhanced muscle activation, up-regulated endocrine responses, greater production of local myokines and reactive oxygen species, and cellular swelling. If hypoxia- mediated increases in metabolic stress are in fact an underlying mechanism for adap- tation to IHRT, it is important to under- stand how best to utilize hypoxic stimuli to augment metabolic stress. If this is not considered during IHRT program design, it is likely that mechanisms downstream from the metabolic stress will not be fur- ther enhanced, and no additive beneﬁts will be observed for IHRT. One key area factor that has received limited research attention is the duration of inter-set recov- ery periods. As highlighted by Bird et al. (2005), the length of inter-set rest periods not only determines the degree of adeno- sine triphosphate-PCr energy recovery, but also the extent to which blood lactate concentrations are elevated. Importantly, IHRT is vastly different to BFR methods It has been previously reported that exercising in systemic hypoxia can induce changes in cerebral oxygenation, which in turn may limit incremental exercise per- formance (Subudhi et al., 2007). Altered cerebral oxygenation may play a reg- ulatory role during IHRT using high- level hypoxia, effectively attenuating any increases in muscle activation that may occur using moderate-level hypoxia. This cerebral mechanism is unlikely to be present during BFR exercise, where the hypoxic environment is localized to the limb being trained and signiﬁcant eleva- tions in muscle activation are regularly reported (Takarada et al., 2000; Yasuda et al., 2009). This is a probable differ- ence between these two resistance train- ing methods. Although further research is needed to examine the dose-response relationship between the level of hypoxia during IHRT and subsequent muscle While some research has previously demonstrated increased metabolic stress following IHRT (Kon et al., 2010, 2012), conﬂicting data have been recently reported (Kurobe et al., 2014; Ho et al., 2014a). Kurobe et al. (2014) observed no differences between hypoxic and normoxic training groups performing 3 sets of 10 repetitions of elbow extensions (10RM) with 60 s inter-set rest. These conﬂict- ing results may be partly explained by the differences in exercise volume, as the studies by Kon et al. METABOLIC STRESS (2010, 2012) both employed 5 sets of two exercises, whereas October 2014 \| Volume 5 \| Article 397 \| 2 Frontiers in Physiology \| Striated Muscle Physiology \| 2 Hypoxic resistance training Scott et al. Kurobe et al. (2014) employed a single exercise for 3 sets. In addition, the study by Ho et al. (2014a) used very long inter-set recovery periods, especially considering the low exercise loads prescribed, which likely attenuated any hypoxia-mediated increase in metabolic stress. It is possible that hypoxic conditions facilitate only a small increase in metabolic stress, which requires more than 3 sets with short rest intervals between to provide a signiﬁcant accumulated effect. However, this expla- nation remains speculative and requires further research. Kurobe et al. (2014) employed a single exercise for 3 sets. In addition, the study by Ho et al. (2014a) used very long inter-set recovery periods, especially considering the low exercise loads prescribed, which likely attenuated any hypoxia-mediated increase in metabolic stress. It is possible that hypoxic conditions facilitate only a small increase in metabolic stress, which requires more than 3 sets with short rest intervals between to provide a signiﬁcant accumulated effect. However, this expla- nation remains speculative and requires further research. performance in netball athletes. J. Sci. Med. Sport 16, 337–342. doi: 10.1016/j.jsams.2012.08.009 performance in netball athletes. J. Sci. Med. Sport 16, 337–342. doi: 10.1016/j.jsams.2012.08.009 Amann, M., Eldridge, M. W., Lovering, A. T., Stickland, M. K., Pegelow, D. F., and Dempsey, J. A. (2006). Arterial oxygenation inﬂuences cen- tral motor output and exercise performance via effects on peripheral locomotor muscle fatigue in humans. J. Physiol. (Lond). 575, 937–952. doi: 10.1113/jphysiol.2006.113936 Manimmanakorn, A., Manimmanakorn, N., Taylor, R., Draper, N., Billaut, F., Shearman, J. P., et al. (2013b). Effects of resistance training combined with vascular occlusion or hypoxia on neuromus- cular function in athletes. Eur. J. Appl. Physiol. 113, 1767–1774. doi: 10.1007/s00421-013-2605-z Millet, G. Y., Aubert, D., Favier, F. B., Busso, T., and Benoit, H. (2009). Effect of acute hypoxia on cen- tral fatigue during repeated isometric leg contrac- tions. Scand. J. Med. Sci. Sports 19, 695–702. doi: 10.1111/j.1600-0838.2008.00823.x Bird, S. P., Tarpenning, K. M., and Marino, F. E. (2005). Designing resistance training programmes to enhance muscular ﬁtness: a review of the acute programme variables. Sports Med. 35, 841–851. doi: 10.2165/00007256-200535100-00002 Chapman, R. F., Karlsen, T., Resaland, G. K., Ge, R. L., Harber, M. P., Witkowski, S., et al. (2014). CONCLUSIONS Friedmann, B., Kinscherf, R., Borisch, S., Richter, G., Bartsch, P., and Billeter, R. (2003). Effects of low-resistance/high-repetition strength training in hypoxia on muscle structure and gene expression. Pﬂügers Arch. 446, 742–751. doi: 10.1007/s00424- 003-1133-9 When considering the current research into IHRT, it is clear that methodological inconsistencies between studies may have caused the conﬂicting results. While there is currently a limited body of research examining IHRT, it appears that adaptive responses may be inﬂuenced by the level of hypoxia and the inter-set rest periods (and subsequent alterations in metabolic stress) used. We propose that for IHRT to elicit greater muscular adaptations than the equivalent training in normoxia, exer- cise protocols should be designed to pro- vide a substantial metabolic stimulus, with particular care being taken to implement relatively brief rest periods between each set. However, it should also be acknowl- edged that a range of other factors could inﬂuence adaptive responses to IHRT. For example, neuroplasticity may be altered by long periods of intermittent hypoxic expo- sures, and could therefore play a role in adaptation to extended IHRT programs. Nonetheless, the purpose of this paper is to provoke thought amongst scientists regarding how resistance training program design may be manipulated in conjunc- tion with systemic hypoxia to enhance adaptive responses. Given the available evidence, we suggest that inter-set rest periods should be very brief for low-load exercise (∼20–30% 1RM; ∼30 s) and brief for moderate-load exercise (∼70% 1RM; ∼60 s). Furthermore, it may also be important to ensure that a moderate, rather than high, level of hypoxia is used (FIO2 = ∼14–16%). Schoenfeld, B. J. (2013). Potential mechanisms for a role of metabolic stress in hypertrophic adap- tations to resistance training. Sports Med. 43, 179–194. doi: 10.1007/s40279-013-0017-1 Scott, B. R., Slattery, K. M., Sculley, D. V., and Dascombe, B. J. (2014). Hypoxia and resistance exercise: a comparison of localized and sys- temic methods. Sports Med. 44, 1037–1054. doi: 10.1007/s40279-014-0177-7 Haseler, L. J., Hogan, M. C., and Richardson, R. S. (1999). Skeletal muscle phosphocreatine recovery in exercise-trained humans is dependent on O2 availability. J. Appl. Physiol. 86, 2013–2018. Ho, J. Y., Huang, T. Y., Chien, Y. C., Chen, Y. C., and Liu, S. Y. (2014a). Effects of acute expo- sure to mild simulated hypoxia on hormonal responses to low-intensity resistance exercise in untrained men. Res. Sports Med. 22, 240–252. doi: 10.1080/15438627.2014.915834 Subudhi, A. W., Dimmen, A. C., and Roach, R. C. CONCLUSIONS (2007). Effects of acute hypoxia on cerebral and muscle oxygenation during incremental exercise. J. Appl. Physiol. 103, 177–183. doi: 10.1152/jap- plphysiol.01460.2006 Ho, J. Y., Kuo, T. Y., Liu, K. L., Dong, X. Y., and Tung, K. (2014b). Combining normobaric hypoxia with short-term resistance training has no additive ben- eﬁcial effect on muscular performance and body composition. J. Strength Cond. Res. 28, 935–941. doi: 10.1519/JSC.0000000000000289 Takarada, Y., Nakamura, Y., Aruga, S., Onda, T., Miyazaki, S., and Ishii, N. (2000). Rapid increase in plasma growth hormone after low-intensity resis- tance exercise with vascular occlusion. J. Appl. Physiol. 88, 61–65. Yasuda, T., Brechue, W. F., Fujita, T., Shirakawa, J., Sato, Y., and Abe, T. (2009). Muscle activation during low-intensity muscle contractions with restricted blood ﬂow. J. Sports Sci. 27, 479–489. doi: 10.1080/02640410802626567 Kawada, S. (2005). What phenomena do occur in blood ﬂow-restricted muscle? Int. J. KAATSU Train. Res. 1, 37–44. doi: 10.3806/ijktr.1.37 Kon, M., Ikeda, T., Homma, T., Akimoto, T., Suzuki, Y., and Kawahara, T. (2010). Effects of acute hypoxia on metabolic and hormonal responses to resistance exercise. Med. Sci. Sports Exerc. 42, 1279–1785. doi: 10.1249/MSS.0b013e3181ce61a5 Conﬂict of Interest Statement: The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. Kon, M., Ikeda, T., Homma, T., and Suzuki, Y. (2012). Effects of low-intensity resistance exer- cise under acute systemic hypoxia on hormonal responses. J. Strength Cond. Res. 26, 611–617. doi: 10.1519/JSC.0b013e3182281c69 Received: 07 August 2014; paper pending published: 09 September 2014; accepted: 24 September 2014; published online: 13 October 2014. Kon, M., Ohiwa, N., Honda, A., Matsubayashi, T., Ikeda, T., Akimoto, T., et al. (2014). Effects of sys- temic hypoxia on human muscular adaptations to resistance exercise training. Physiol. Rep. 2:e12033. doi: 10.14814/phy2.12033 Citation: Scott BR, Slattery KM and Dascombe BJ (2014) Intermittent hypoxic resistance training: does it provide added beneﬁt? Front. Physiol. 5:397. doi: 10.3389/fphys.2014.00397 This article was submitted to Striated Muscle Physiology, a section of the journal Frontiers in Physiology. Kurobe, K., Huang, Z., Nishiwaki, M., Yamamoto, M., Kanehisa, H., and Ogita, F. (2014). Effects of resis- tance training under hypoxic conditions on mus- cle hypertrophy and strength. Clin. Physiol. Funct. Imaging doi: 10.1111/cpf.12147. [Epub ahead of print]. Copyright © 2014 Scott, Slattery and Dascombe. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). METABOLIC STRESS Deﬁning the “dose” of altitude training: how high to live for optimal sea level performance enhancement. J. Appl. Physiol. 116, 595–603. doi: 10.1152/jap- plphysiol.00634.2013 Nishimura, A., Sugita, M., Kato, K., Fukuda, A., Sudo, A., and Uchida, A. (2010). Hypoxia increases mus- cle hypertrophy induced by resistance training. Int. J. Sports Physiol. Perform. 5, 497–508. Noakes, T. D., Peltonen, J. E., and Rusko, H. K. (2001). Evidence that a central governor regulates exercise performance during acute hypoxia and hyperoxia. J. Exp. Biol. 204, 3225–3234. CONCLUSIONS The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licen- sor are credited and that the original publication in this journal is cited, in accordance with accepted aca- demic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Copyright © 2014 Scott, Slattery and Dascombe. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licen- sor are credited and that the original publication in this journal is cited, in accordance with accepted aca- demic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Manimmanakorn, A., Hamlin, M. J., Ross, J. J., Taylor, R., and Manimmanakorn, N. (2013a). Effects of low-load resistance training combined with blood ﬂow restriction or hypoxia on muscle function and ACKNOWLEDGMENTS ACKNOWLEDGMENTS This manuscript was not supported by any funding. This manuscript was not supported by any funding. October 2014 \| Volume 5 \| Article 397 \| 3 www.frontiersin.org
https://openalex.org/W3162351593	https://www.nature.com/articles/s41598-021-89824-x.pdf	English	null	Tropical Indo-Pacific SST influences on vegetation variability in eastern Africa	Scientific reports	2,021	cc-by	7,730	Tropical Indo‑Pacific SST influences on vegetation variability in eastern Africa In‑Won Kim1,2, Malte F. Stuecker3, Axel Timmermann1,2, Elke Zeller1,2, Jong‑Seong Kug4, So‑Won Park4 & Jin‑Soo Kim5 OPEN Mechanisms by which tropical Pacific and Indian Ocean sea surface temperatures (SST) influence vegetation in eastern Africa have not been fully explored. Here, we use a suite of idealized Earth system model simulations to elucidate the governing processes for eastern African interannual vegetation changes. Our analysis focuses on Tanzania. In the absence of ENSO-induced sea surface temperature anomalies in the Tropical Indian Ocean (TIO), El Niño causes during its peak phase negative precipitation anomalies over Tanzania due to a weakening of the tropical-wide Walker circulation and anomalous descending motion over the Indian Ocean and southeastern Africa. Resulting drought conditions increase the occurrence of wildfires, which leads to a marked decrease in vegetation cover. Subsequent wetter La Niña conditions in boreal winter reverse the phase in vegetation anomalies, causing a gradual 1-year-long recovery phase. The 2-year-long vegetation decline in Tanzania during an ENSO cycle can be explained as a double-integration of the local rainfall anomalies, which originate from the seasonally-modulated ENSO Pacific-SST forcing (Combination mode). In the presence of interannual TIO SST forcing, the southeast African precipitation and vegetation responses to ENSO are muted due to Indian Ocean warming and the resulting anomalous upward motion in the atmosphere. Natural fluctuations in Africa’s vegetation are affected by rainfall variability1–7. Especially the Sahel, eastern Africa, and southern Africa show large interannual variations in terrestrial productivity, which can be attributed to year-to-year changes in water stress8. Linking African terrestrial rainfall variability to large-scale climate vari- ability might result in potential predictability of associated vegetation changes. Especially the El Niño-Southern Oscillation (ENSO) is an important climate driver of interannual rainfall variability in parts of Africa. On aver- age, El Niño events cause droughts in southern Africa and enhanced precipitation and corresponding floods in eastern Africa9. An associated vegetation response can be identified by the relationship between ENSO and the Normalized Difference Vegetation Index (NDVI) over both these regions10–14. In contrast, over the Sahel, the relationship between ENSO and NDVI is weak15. Interannual vegetation changes over eastern Africa show a nonlinear relationship with rainfall variability and a strong dependency on land cover type is observed2. Glob- ally, an asymmetric relationship between net primary production and rainfall is observed for grasslands16. These examples highlight that the vegetation response to climate factors is often modulated by nonlinear land processes. Wildfires can play an important role in these interannual vegetation changes through climate-fire-vegetation interactions17,18. www.nature.com/scientificreports www.nature.com/scientificreports Tropical Indo‑Pacific SST influences on vegetation variability in eastern Africa In‑Won Kim1,2, Malte F. Stuecker3, Axel Timmermann1,2, Elke Zeller1,2, Jong‑Seong Kug4, So‑Won Park4 & Jin‑Soo Kim5 OPEN For wet savannas in Africa, an increase in rainfall-induced fuel moisture leads to a decrease in the burned area7, while for dry savannas, an increase in moisture promotes biomass as a role of fuels, thereby facilitating more fires19. The roles of climate and fire in African vegetation dynamics were further highlighted in a series of fire-off experiments20,21, and the vegetation model experiments with fire modules realistically repre- sented vegetation compared to the fire-off experiments.fi p g g g pt y p Wildfires can play an important role in these interannual vegetation changes through climate-fire-vegetation interactions17,18. For wet savannas in Africa, an increase in rainfall-induced fuel moisture leads to a decrease in the burned area7, while for dry savannas, an increase in moisture promotes biomass as a role of fuels, thereby facilitating more fires19. The roles of climate and fire in African vegetation dynamics were further highlighted in a series of fire-off experiments20,21, and the vegetation model experiments with fire modules realistically repre- sented vegetation compared to the fire-off experiments.fi g pif p In addition to potential nonlinear effects caused by vegetation types and fire activity mentioned above, obser- vations and model experiments reveal a spatial asymmetry in atmospheric response over Africa between El Niño and La Niña22. Furthermore, El Niño and La Niña events also influence the pattern and seasonal evolution of NDVI in eastern Africa in an asymmetric way12. In addition, nonlinear ENSO teleconnections to Africa might 1Center for Climate Physics, Institute for Basic Science, Busan, Republic of Korea. 2Pusan National University, Busan, Republic of Korea. 3Department of Oceanography and International Pacific Research Center, School of Ocean and Earth Science and Technology, University of Hawaiʻi at Mānoa, Honolulu, HI, USA. 4Division of Environmental Science and Engineering, Pohang University of Science and Technology, Pohang, Republic of Korea. 5Department of Evolutionary Biology and Environmental Studies, University of Zurich, Zurich, Switzerland. email: iwkimi@pusan.ac.kr \| https://doi.org/10.1038/s41598-021-89824-x Scientific Reports \| (2021) 11:10462 www.nature.com/scientificreports/ Figure. 1. Composite differences of SST anomalies (shading; unit: K), and 200 hPa velocity potential anomalies in D(0)JF(1) (contours; unit: m2 s−1, scaled by 106) between El Niño and La Niña events for the Periodic experiment (a), the Pacific experiment (b), the Tropics experiment (c), HadISST1 and ERA5 reanalysis (d): El Niño events: 1982, 1987, 1991, 1997, 2002, 2009; La Niña events: 1984, 1988, 1999, 2000, 2007, 2010 for the Pacific experiment, the Tropics experiment, and the observations. Tropical Indo‑Pacific SST influences on vegetation variability in eastern Africa In‑Won Kim1,2, Malte F. Stuecker3, Axel Timmermann1,2, Elke Zeller1,2, Jong‑Seong Kug4, So‑Won Park4 & Jin‑Soo Kim5 OPEN further be affected by ENSO-induced asymmetric sea surface temperature (SST) responses over the Atlantic and Indian Oceans. For instance, SST variability over the southern Atlantic and tropical Pacific Ocean have a negative relationship with rainfall over the Sahel10. Indian Ocean Dipole (IOD) events, typically accompanied by ENSO23, positively correlate with eastern African rainfall during the short rainy season24,25. Although the aforementioned studies have demonstrated the impacts of ENSO on African vegetation based on observations, we still lack a deeper understanding on the involved physical processes, including the role of Although the aforementioned studies have demonstrated the impacts of ENSO on African vegetation based on observations, we still lack a deeper understanding on the involved physical processes, including the role of wildfires. Moreover, the direct ENSO impacts (i.e., caused by SST anomalies in the tropical Pacific Ocean) are difficult to delineate from indirect ENSO impacts (i.e., caused by ENSO-induced SST anomalies in the Indian Ocean) based on observations alone. Therefore, in this study, we investigate the vegetation response in sub- Saharan Africa to both direct and indirect ENSO forcing through a series of targeted model experiments and compare them to the observations. Tropical Indo‑Pacific SST influences on vegetation variability in eastern Africa In‑Won Kim1,2, Malte F. Stuecker3, Axel Timmermann1,2, Elke Zeller1,2, Jong‑Seong Kug4, So‑Won Park4 & Jin‑Soo Kim5 OPEN Negative values (dashed lines) indicate anomalous upward motion and positive values (solid lines) indicate anomalous downward motion. Stippling indicates a statistically significant difference at the 95% significance level for the 200 hPa velocity potential. Fig. 1 was generated using NCAR Command Language Version 6.5.0 (http://dx.doi.org/10.5065/D6WD3XH5). Figure. 1. Composite differences of SST anomalies (shading; unit: K), and 200 hPa velocity potential anomalies in D(0)JF(1) (contours; unit: m2 s−1, scaled by 106) between El Niño and La Niña events for the Periodic experiment (a), the Pacific experiment (b), the Tropics experiment (c), HadISST1 and ERA5 reanalysis (d): El Niño events: 1982, 1987, 1991, 1997, 2002, 2009; La Niña events: 1984, 1988, 1999, 2000, 2007, 2010 for the Pacific experiment, the Tropics experiment, and the observations. Negative values (dashed lines) indicate anomalous upward motion and positive values (solid lines) indicate anomalous downward motion. Stippling indicates a statistically significant difference at the 95% significance level for the 200 hPa velocity potential. Fig. 1 was generated using NCAR Command Language Version 6.5.0 (http://dx.doi.org/10.5065/D6WD3XH5). further be affected by ENSO-induced asymmetric sea surface temperature (SST) responses over the Atlantic and Indian Oceans. For instance, SST variability over the southern Atlantic and tropical Pacific Ocean have a negative relationship with rainfall over the Sahel10. Indian Ocean Dipole (IOD) events, typically accompanied by ENSO23, positively correlate with eastern African rainfall during the short rainy season24,25. further be affected by ENSO-induced asymmetric sea surface temperature (SST) responses over the Atlantic and Indian Oceans. For instance, SST variability over the southern Atlantic and tropical Pacific Ocean have a negative relationship with rainfall over the Sahel10. Indian Ocean Dipole (IOD) events, typically accompanied by ENSO23, positively correlate with eastern African rainfall during the short rainy season24,25. Although the aforementioned studies have demonstrated the impacts of ENSO on African vegetation based on observations, we still lack a deeper understanding on the involved physical processes, including the role of wildfires. Moreover, the direct ENSO impacts (i.e., caused by SST anomalies in the tropical Pacific Ocean) are difficult to delineate from indirect ENSO impacts (i.e., caused by ENSO-induced SST anomalies in the Indian Ocean) based on observations alone. Therefore, in this study, we investigate the vegetation response in sub- Saharan Africa to both direct and indirect ENSO forcing through a series of targeted model experiments and compare them to the observations. Results h l The Walker circulation response to ENSO. To investigate the drivers of the vegetation response over Africa to Indo-Pacific SST anomalies, we first focus on the tropical large-scale atmospheric circulation and its interannual variations. The position and strength of the Walker circulation are closely coupled to SST anomalies in the tropical Pacific. Both the “Periodic” and the “Pacific” experiments (SST anomalies are only prescribed in the tropical Pacific) show pronounced Walker circulation changes between El Niño and La Niña events with anomalous ascending motion over the eastern Pacific region (and corresponding upper-level divergence) and anomalous descending motion (and corresponding upper-level convergence) during the peak ENSO phase of December–January–February [D(0)JF(1)] (Fig. 1a,b). Importantly, the edge of the descending motion extends to the African continent in the two experiments. In contrast, the “Tropics” experiment shows that the center of the descending motion shifts toward the Maritime Continent, inducing weaker subsidence around the Indian and Atlantic Ocean, accompanying tropical Indian Ocean (TIO) warming (Fig. 1c). This large-scale circulation response is similar to what is seen for the observations (Fig. 1d). The TIO warming pattern seen in Fig. 1c,d is largely forced by El Niño and then is prolonged for several months after the El Niño event due to the so-called capacitor effect26,27. The pattern of large-scale atmospheric anomalies in the Tropics experiment (Fig. 1c) is more consistent with the observations (Fig. 1d) than the Periodic and Pacific experiments (Fig. 1a, b). This suggests that TIO warming affects the change of the large-scale atmospheric circulation around the African continent related to ENSO, as suggested previously28. The response of rainfall, vegetation, and wildfire over Africa to ENSO. To understand first the symmetric (i.e., linear) response to El Niño and La Niña events, we show El Niño minus La Niña composites. Observed composite differences between El Niño and La Niña events display pronounced positive precipitation anomalies over the Horn of Africa and negative anomalies over Southern Africa in D(0)JF(1) (Fig. 2j). The three experiments (Periodic, Pacific, and Tropics) reproduce these anomalies reasonably well (Fig. 2a,d,g, Table S1). Interestingly, the Periodic and the Pacific experiments exhibit a 50 % D(0)JF(1) rainfall reduction over Tanza- nia for the El Niño minus La Niña composite (Fig. 2a,d) and an accompanying negative Net Primary Production https://doi.org/10.1038/s41598-021-89824-x Scientific Reports \| (2021) 11:10462 \| www.nature.com/scientificreports/ Figure. 2. Results h l Composite differences (unit: %) between El Niño and La Niña events for precipitation (PRCP) anomalies in D(0)JF(1), net primary production (NPP) anomalies in JFM(1), and Leaf Area Index (LAI) anomalies in MJJ(2) for the Periodic experiment (a–c), the Pacific experiment (d–f), the Tropics experiment (g–i), as well as PRCP and LAI for the observations (j, l). Stippling indicates a statistically significant difference at the 95% significance level. Black box shows the surrounding Tanzania region (2–13°S, 28–42°E). The lead-lag cross-correlation between the Niño3.4 index and LAI anomalies over Tanzania for the Periodic experiment (yellow), the Pacific experiment (blue), and the Tropics experiment (red) (k). Fig. 2 was generated using NCAR Command Language Version 6.5.0 (http://dx.doi.org/10.5065/D6WD3XH5). Figure. 2. Composite differences (unit: %) between El Niño and La Niña events for precipitation (PRCP) anomalies in D(0)JF(1), net primary production (NPP) anomalies in JFM(1), and Leaf Area Index (LAI) anomalies in MJJ(2) for the Periodic experiment (a–c), the Pacific experiment (d–f), the Tropics experiment (g–i), as well as PRCP and LAI for the observations (j, l). Stippling indicates a statistically significant difference at the 95% significance level. Black box shows the surrounding Tanzania region (2–13°S, 28–42°E). The lead-lag cross-correlation between the Niño3.4 index and LAI anomalies over Tanzania for the Periodic experiment (yellow), the Pacific experiment (blue), and the Tropics experiment (red) (k). Fig. 2 was generated using NCAR Command Language Version 6.5.0 (http://dx.doi.org/10.5065/D6WD3XH5). (NPP) anomaly during January–February–March of the decaying ENSO year [JFM(1)] (Fig. 2b,e). Mostly, the stronger NPP anomalies prevail over grassland and deciduous forest (herbaceous cover) along with a larger precipitation response, in accordance with previous studies2. In contrast, the observations and the more realis- tic Tropics experiment show only very weak rainfall, NPP, and Leaf Area Index (LAI) anomalies over Tanzania (Fig. 2g,h,i,j,l). Agreeing with other studies29, this finding suggests that tropical Indian or Atlantic Ocean SST anomalies might play an important role in muting the direct Pacific response over this region. We hypothesize specifically that the negligible observed rainfall response over Tanzania in the observations can be attributed to compensating effects from the direct Pacific effect and the El Niño-related Indian Ocean warming effect on the Walker circulation (Fig. 1b,c).h g This hypothesis is further supported by the lead-lag cross-correlation relationship between ENSO and LAI anomalies in the Periodic and the Pacific experiments (Fig. 2k). Scientific Reports \| (2021) 11:10462 \| Results h l According to this analysis, ENSO is leading LAI anomalies in Tanzania by about one year in these two experiments, whereas no statistically significant correla- tion can be found in the Tropics experiment. ENSO negatively correlates with LAI over Tanzania at a maximum lag of 16-months (R = 0.49, p < 0.00001) in the Pacific and 18-months (R = 0.58, p < 0.00001) in the Periodic experiments (Fig. 2k). In contrast, for the Tropics experiment, the correlation is not significant (R = 0.06, p = 0.11) (Fig. 2k). Regarding the LAI response to ENSO at this 16-18 months lag (that is, in May-June-July in year 2 after the ENSO event peak time: MJJ(2)), we find larger negative anomalies over Tanzania in the Pacific and the Periodic experiments (Fig. 2c,f), while they are much weaker anomalies in the Tropics experiment and the observations (Fig. 2i,l). ( g ) To further support our hypothesis, we compare CLM4 NPP and LAI changes to those obtained with the offline vegetation model BIOME4 which has been forced by the CESM1.2.2 monthly output of the sensitivity experiments (see Data and Experiments). The simulated annual mean NPP response over Africa to ENSO rep- resented by the CLM4 agrees well with BIOME4 (Fig. S1-2). However, negative anomalies of annual mean LAI over Tanzania persist in CLM4 for much longer than in the BIOME4 experiments (Fig. S3-4), which is related to the fact that the BIOME4 model does not integrate the climate forcings of the previous months, but finds an equilibrium solution for each monthly forcing, in contrast to the CLM4.i q y g Moreover, over Tanzania, the delayed response to ENSO is also found in the CLM4 in wildfire activity (Fig. 3). The periodic experiment shows negative anomalies in burned area over Tanzania in D(0)JF(1) and statistically insignificant differences in the Pacific and Tropics experiments, as well as in the observations. However, the Scientific Reports \| (2021) 11:10462 \| https://doi.org/10.1038/s41598-021-89824-x www.nature.com/scientificreports/ Figure. 3. Composite differences of burned area (unit: %) between El Niño and La Niña events for the Periodic experiment, the Pacific experiment, the Tropics experiment, and the observations for D(0)JF(1) (a–d) and SON(1) (e–h): El Niño events: 1997, 2002, 2009; La Niña events: 1999, 2000, 2007, 2010 for observations. Stippling indicates a statistically significant difference at the 95% significance level. Fig. 3 was generated using NCAR Command Language Version 6.5.0 (http://dx.doi.org/10.5065/D6WD3XH5). Figure. 3. Results h l Composite differences of burned area (unit: %) between El Niño and La Niña events for the Periodic experiment, the Pacific experiment, the Tropics experiment, and the observations for D(0)JF(1) (a–d) and SON(1) (e–h): El Niño events: 1997, 2002, 2009; La Niña events: 1999, 2000, 2007, 2010 for observations. Stippling indicates a statistically significant difference at the 95% significance level. Fig. 3 was generated using NCAR Command Language Version 6.5.0 (http://dx.doi.org/10.5065/D6WD3XH5). Figure. 4. Time evolution over Tanzania (2–13°S, 28–42°E) for the Periodic experiment (a), the Pacific experiment (b) and the Tropics experiment (c): Niño3.4 index (solid yellow/blue/red line; unit: K), C-mode index (dashed gray line; unit: K), Tropical Indian Ocean (TIO) SST anomalies (solid gray line; unit: K), precipitation anomalies (solid yellow/blue/red line; unit: mm/day), reconstructed precipitation anomalies (dashed gray line), burned area (solid yellow/blue/red line; unit: fraction), reconstructed burned area anomalies (dashed gray line), LAI anomalies (solid yellow/blue/red line; unit: m2/m2), and reconstructed LAI anomalies (dashed gray line). Transparent shading indicates ± 1 standard deviation. Fig. 4 was generated using NCAR Command Language Version 6.5.0 (http://dx.doi.org/10.5065/D6WD3XH5). Figure. 4. Time evolution over Tanzania (2–13°S, 28–42°E) for the Periodic experiment (a), the Pacific experiment (b) and the Tropics experiment (c): Niño3.4 index (solid yellow/blue/red line; unit: K), C-mode index (dashed gray line; unit: K), Tropical Indian Ocean (TIO) SST anomalies (solid gray line; unit: K), precipitation anomalies (solid yellow/blue/red line; unit: mm/day), reconstructed precipitation anomalies (dashed gray line), burned area (solid yellow/blue/red line; unit: fraction), reconstructed burned area anomalies (dashed gray line), LAI anomalies (solid yellow/blue/red line; unit: m2/m2), and reconstructed LAI anomalies (dashed gray line). Transparent shading indicates ± 1 standard deviation. Fig. 4 was generated using NCAR Command Language Version 6.5.0 (http://dx.doi.org/10.5065/D6WD3XH5). Figure. 4. Time evolution over Tanzania (2–13°S, 28–42°E) for the Periodic experiment (a), the Pacific Figure. 4. Time evolution over Tanzania (2–13 S, 28–42 E) for the Periodic experiment (a), the Pacific experiment (b) and the Tropics experiment (c): Niño3.4 index (solid yellow/blue/red line; unit: K), C-mode index (dashed gray line; unit: K), Tropical Indian Ocean (TIO) SST anomalies (solid gray line; unit: K), precipitation anomalies (solid yellow/blue/red line; unit: mm/day), reconstructed precipitation anomalies (dashed gray line), burned area (solid yellow/blue/red line; unit: fraction), reconstructed burned area anomalies (dashed gray line), LAI anomalies (solid yellow/blue/red line; unit: m2/m2), and reconstructed LAI anomalies (dashed gray line). Transparent shading indicates ± 1 standard deviation. Fig. Results h l 4 was generated using NCAR Command Language Version 6.5.0 (http://dx.doi.org/10.5065/D6WD3XH5). index (dashed gray line; unit: K), Tropical Indian Ocean (TIO) SST anomalies (solid gray line; unit: K), precipitation anomalies (solid yellow/blue/red line; unit: mm/day), reconstructed precipitation anomalies (dashed gray line), burned area (solid yellow/blue/red line; unit: fraction), reconstructed burned area anoma (dashed gray line), LAI anomalies (solid yellow/blue/red line; unit: m2/m2), and reconstructed LAI anomalie (dashed gray line). Transparent shading indicates ± 1 standard deviation. Fig. 4 was generated using NCAR Command Language Version 6.5.0 (http://dx.doi.org/10.5065/D6WD3XH5). Periodic and the Pacific experiments show a 10–20 % increase in burned area over Tanzania in September- October-November in year 1 after ENSO event peak time [SON(1)] (Fig. 3a-d), whereas the observations and the Tropics experiment show statistically insignificant differences (Fig. 3e-h). Periodic and the Pacific experiments show a 10–20 % increase in burned area over Tanzania in September- October-November in year 1 after ENSO event peak time [SON(1)] (Fig. 3a-d), whereas the observations and the Tropics experiment show statistically insignificant differences (Fig. 3e-h). Combination mode‑driven rainfall response over Tanzania. The temporal evolution of the rainfall response over Tanzania to ENSO shows a rapid transition during the peak phase of both El Niño and La Niña in both the Periodic and the Pacific experiments, but not in the Tropics experiment (Fig. 4). The rainfall response is particularly pronounced in the former during the peak phase of ENSO in D(0)JF(1), which is also the climato- logical wet season (Fig. 4, Fig. S5). This illustrates the tight coupling between climatological conditions and the imposed ENSO signal. To further understand the distinct atmospheric response to ENSO in the absence of TIO SST anomalies, we hypothesize that the precipitation response over Tanzania to ENSO is driven by the seasonally modulated interannual ENSO variability, which is referred to as a Combination mode (C-mode)30. According to this simple model, the precipitation anomalies can be written as https://doi.org/10.1038/s41598-021-89824-x Scientific Reports \| (2021) 11:10462 \| www.nature.com/scientificreports/ P∗(t) = αENSO(t) + βENSO(t) · cos (ωat), (1) where α and β are the regression coefficients for the ENSO and theoretical C-mode predictors, and ωa the frequency of the annual cycle. The phase of the annual cycle has a minimum value in July and a maximum in January. ENSO variability here is characterized by the Niño3.4 SST anomalies. dB(t) (2) Appropriate parameters values derived from the CESM1.2.2 and CLM4 output are given in Table S2. In the Periodic experiment, the reconstruction of burned area response over Tanzania captures the simulated temporal evolution reasonably well (R = 0.82, p < 0.00001), suggesting that the burned area response can be determined essentially by the time integral of the direct ENSO effect and the C-mode term. Previous studies support the notion that the lagged response of wildfire activity in some areas can be linked to the integrated effect of antecedent precipitation anomalies7,34. In the Periodic experiment, less rainfall over Tanzania during the wet season [D(0)JF(1)] and successive dry season promote a lagged response in burned area in SON(1) (Fig. S5). Subsequently, LAI anomalies over Tanzania slowly develop after the peak of El Niño and are prolonged until the following La Niña event. Especially, the peak of negative anomalies in LAI occurs during the mature La Niña phase in December-January–February in year 2 [DJF(2)], in spite of the maximum rainfall anomalies during this time (Fig. 4, Fig. S5). The vegetation response to climate factors also depends on vegetation resist- ance and resilience35. The vegetation carbon pool in CLM4 is governed by new growth, litterfall, mortality, and fire processes. The resilience of vegetation response can be represented by the process of new growth of post-fire vegetation. Accordingly, we hypothesize that the LAI response can be largely explained by the integrated effect of burned area (from Eq. 2), where L represents temporal variation of LAI, and µ2 is about 8 month−1 as an inverse damping time scale (characterizing vegetation resilience): (3) dL(t) dt = −µ2L(t) −θ2B(t). (3) The parameters values in equation 3 were calculated by optimizing the first-order autoregressive (AR(1)) model (Tables S2, S3). According to this simplified double-integration model (Eqs. 1–3), the LAI response over Tanza- nia correlates highly with simulated LAI anomalies in the Periodic forcing experiment (R = 0.72, p < 0.00001), indicating that the lagged and prolonged vegetation response to ENSO is explained by vegetation resilience and the integrated effect of wildfire activity. Similar double-integration models have been introduced to explain also the emergence of low-frequency marine biogeochemical variability36. Results h l One can also include a white noise precipitation forcing, but since we consider ensemble mean properties in a linear model, the noise forcing is not necessary to understand the temporal evolution. The time-series in the Periodic and Pacific experiments show that the reconstructions of precipitation anomalies over Tanzania via the C-mode equation reproduce the seasonally varying simulated rainfall response to ENSO well (Periodic: R = 0.64, p < 0.00001; Pacific: R = 0.65, p < 0.00001) (Fig. 4). The simulated La Niña response is somewhat reduced as compared to the El Niño rainfall anomaly. This is reminiscent of an atmospheric nonlinearity to otherwise symmetric SST forcing. Role of wildfires in the vegetation response to ENSO. In d d fi l d d d h Role of wildfires in the vegetation response to ENSO. In the absence of TIO warming in the Peri- odic and Pacific experiments, El Niño-induced drying increases the occurrence of fires, which is manifest in the prolonged positive anomalies in burned area lasting for about one year after the peak of El Niño (Fig. 4a,b). In CLM4, the probability of fire occurrence is parameterized by fuel availability and near-surface soil moisture conditions31. The annual burned area is driven by fire seasonal length and fire resistance depends on plant func- tional type (PFT)32,33. In addition, the water balance for soil moisture is controlled by precipitation, evaporation, and runoff, etc., but precipitation is the main driver for eastern Africa. To investigate the dynamical linkage between precipitation and wildfire responses to ENSO, we hypothesize that changes in burned area B, are driven by precipitation variability P. Here we choose P* as the ENSO-reconstructed precipitation anomaly from Eq. (1). We assume in its simplest linearized form that the burned area does not depend on the available vegetation, which allows us to introduce a fixed mean recovery timescale ( µ1−1 ), in which the burned area can regrow. The simplified linearized model then reads: (2) dB(t) dt = −µ1B(t) −θ1P∗(t). Data and experiments Ob ti W d p Observations. We used precipitation data from Global Precipitation Analysis Products of the Global Pre- cipitation Climatology Centre (GPCC)43 and University of East Anglia’s Climate Research Unit (CRU)44, 200 hPa wind from European Centre for Medium-Range Weather Forecasts (ECMWF) reanalysis generation 5 (ERA5)45, and SST from the Hadley Centre Sea Ice and Sea Surface Temperature data set version 1 (HadISST1)46. To char- acterize observed vegetation changes, we utilized leaf area index (LAI) data derived from the Global Inventory Modeling and Mapping Studies (GIMMS) Normalized Difference Vegetation Index (NDVI3g) for the period 1982 to 201147. The monthly Global Fire Emissions Database version 4 (GFEDv4)48 was used to characterize the 1994–2014 wildfire activity. Model and experiments. We conducted a suite of atmospheric general circulation model (AGCM) exper- iments with the Community Earth System Model (CESM 1.2.2) using the Community Atmosphere Model ver- sion 4.0 (CAM4)49 coupled to the Community Land Model version 4.0 (CLM4)33,50 with active Carbon-Nitrogen (CN) biogeochemistry (https://svn-ccsm-models.cgd.ucar.edu/cesm1/release_tags/cesm1_2_2/).h g y p g g The model, which uses a horizontal resolution of approximately 1-degree, was spun up until the carbon and nitrogen pools were equilibrated to a 1957-2016 SST climatology boundary forcing and present-day greenhouse gas concentrations. We then performed four different types of AGCM model experiment ensembles to investi- gate the vegetation response over sub-Saharan Africa to interannual tropical SST variability starting from these equilibrated initial conditions. The ensemble simulations were performed with different initial conditions. These were generated by applying small perturbations of round-off error magnitude to the initial temperature field using the CESM namelist variable “pertlim”51. g p First, a control experiment (CTRL) was carried out with a repeating global climatological SST forcing for the period 1957–2016 using a 3-member ensemble. The CTRL roughly reproduces the observed annual pre- cipitation (PRCP) and LAI climatological patterns over Africa (Fig. S7, Table S5). The simulated seasonal mean precipitation over Africa region against observations are comparable to CMIP5 models52 (Table S4). The CTRL simulation poorly simulates annual burned area over sub-Saharan Africa, but captures the spatial pattern of the annual burned area over eastern Africa reasonably well (Table S5).i y To illustrate the impact of observed ENSO variability, a “Pacific” experiment was conducted by adding the observed SST anomalies over the tropical eastern Pacific (15°S–15°N, 180°-90°W) for the period 1957–2016 to the climatology with a 10-member ensemble. References 1. Camberlin, P., Martiny, N., Philippon, N. & Richard, Y. Determinants of the interannual relationships between remote sensed photosynthetic activity and rainfall in tropical Africa. Remote Sens. Environ. 106, 199–216 (2007). Discussion and conclusions In this study, we explored how vegetation in the southeastern part of Africa changes in response to interannual ENSO variability through a series of model experiments. Focusing on Tanzania, we found that, in the absence of TIO variability, the rapid transition of precipitation anomalies during ENSO events is determined by the interac- tion between ENSO and the annual cycle of rainfall over Tanzania (the so-called C-mode). After the occurrence of El Niño, the pronounced decrease in rainfall over Tanzania leads to an enhancement in burned area with a time delay, thereby prolonging a marked vegetation decrease for 2 years. This response can be explained by the integrated effect of wildfire (double integrated effect of precipitation) and vegetation resilience through an idealized dynamical model, which explains the AGCM results reasonably well. In contrast, the BIOME4 offline experiment forced with the data from the CESM1.2.2 sensitivity experiment does not capture this effect, (Figs. S2, S4), because it calculates the equilibrium for each monthly forcing and does not account for the time history.i q y g y However, for the Tropics experiment (as well as in the observations) we do not find strong ENSO-related anomalies in precipitation, wildfire, and vegetation over Tanzania. This is because TIO warming during El Niño events compensates the rainfall response to ENSO over Tanzania (Fig. 4c) by weakening the anomalous https://doi.org/10.1038/s41598-021-89824-x Scientific Reports \| (2021) 11:10462 \| www.nature.com/scientificreports/ References 1. Camberlin, P., Martiny, N., Philippon, N. & Richard, Y. Determinants of the interannual relationships between remote sensed photosynthetic activity and rainfall in tropical Africa. Remote Sens. Environ. 106, 199–216 (2007). www.nature.com/scientificreports/ atmospheric subsidence (Fig. 1b,c). This offset response is consistent with the opposite impact between Indian Ocean Basin-wide mode (IOBM) and ENSO on seasonal rainfall variability over Africa discussed in previous studies37. The IOD is another primary climate factor, which can affect rainfall and vegetation variability over eastern Africa2,37–39, but the IOD impact to eastern Africa peaks in September–November [SON(0)] (Fig. S6). This is too early to cause major precipitation and vegetation anomalies in Tanzania (Fig. 4, Fig. S6). h A caveat to be mentioned here is that the current land carbon models have limitations in realistically simu- lating climate-fire-vegetation interactions. In particular, the CLM4 in our model-set-up does not consider the human impact of wildfire activity (e.g., ignition, suppression, etc.), which implies that the model could overesti- mate the wildfire response to climate conditions. Other complications may arise from the fact that the interannual response of vegetation to climate variability in the observations may further depend on the vegetation structure2 and spatial fragmentation, which are not considered in CLM4. p g Despite the modeling caveats, our study provides new conceptual insights into the physical mechanisms of climate-fire-vegetation interactions over eastern Africa. The double-integration model (Eqs. 1–3) proves to be useful in understanding time-variability and leads and lags between climate forcing, fire, and vegetation responses. It can also be applied to understand the effect of projected future changes in ENSO and TIO40–42 on fire and vegetation in eastern Africa. References 1. Camberlin, P., Martiny, N., Philippon, N. & Richard, Y. Determinants of the interannual relationships between remote sensed photosynthetic activity and rainfall in tropical Africa. Remote Sens. Environ. 106, 199–216 (2007). www.nature.com/scientificreports/ The relationship of the El Niño-Southern oscillation to African rainfall. Int. J. Climatol. J. R. Meteorol Soc. 17, 117–135 (1997). ( ) 0. Camberlin, P., Janicot, S. & Poccard, I. Seasonality and atmospheric dynamics of the teleconnection between African rainfall and tropical sea-surface temperature: Atlantic vs. ENSO. Int. J. Climatol. J. R. Meteorol. Soc. 21, 973–1005 (2001). 1. Anyamba, A., Tucker, C. J. & Mahoney, R. From El Niño to La Niña: Vegetation response patterns over east and southern Africa during the 1997–2000 period. J. Clim. 15, 3096–3103 (2002).f g p 12. Philippon, N., Martiny, N., Camberlin, P., Hoffman, M. & Gond, V. Timing and patterns of the ENSO signal in Africa over the last 30 years: Insights from normalized difference vegetation index data. J. Clim. 27, 2509–2532 (2014). f 3. Anyamba, A., Glennie, E. & Small, J. Teleconnections and interannual transitions as observed in African Vegetation: 2015–2017 Remote Sensing 10, 1038 (2018). g 4. Plisnier, P., Serneels, S. & Lambin, E. Impact of ENSO on East African ecosystems: A multivariate analysis based on climate and remote sensing data. Glob. Ecol. Biogeogr. 9, 481–497 (2000). g g g 15. Philippon, N., Martiny, N. & Camberlin, P. Forecasting the vegetation photosynthetic activity over the Sahel: A model output statistics approach. Int. J. Climatol. J. R. Meteorol. Soc. 29, 1463–1477 (2009). pp Y., Fang, J., Ma, W. & Wang, W. Relationship between variability in aboveground net primary production and precipitation b l l d G ph R L tt 35 L23710 htt //d i /10 1029/2008GL035408 (2008) 16. Yang, Y., Fang, J., Ma, W. & Wang, W. Relationship between variability in aboveground net primary production and preci in global grasslands. Geophys. Res. Lett. 35, L23710. https://doi.org/10.1029/2008GL035408 (2008).i g g p y p g 17. Ryan, K. C. Vegetation and wildland fire: Implications of global climate change. Environ. Int. 17, 169–178 (1991). i 18. Chikamoto, Y., Timmermann, A., Stevenson, S., DiNezio, P. & Langford, S. Decadal predictability of soil water, vegetation, and wildfire frequency over North America. Clim. Dyn. 45, 2213–2235 (2015).i i q y y 9. Archibald, S., Roy, D. P., van Wilgen, B. W. & Scholes, R. J. What limits fire? An examination of drivers of burnt area in Southern Africa. Glob. Change Biol. 15, 613–630 (2009).hi g 0. Bond, W. J., Woodward, F. I. & Midgley, G. F. The global distribution of ecosystems in a world without fire. New Phytol. www.nature.com/scientificreports/ & Oleson, K. The Community Land Model’s dynamic global vegetation model (CLM- 32. Levis, S., Bonan, G., Vertenstein, M. & Oleson, K. The Community Land Model’s dynamic global vegetation mod Technical description and user’s guide. NCAR Tech. Note TN-459+ IA 50. https://doi.org/10.5065/D6P26W36 32. Levis, S., Bonan, G., Vertenstein, M. & Oleson, K. The Community Land Model’s dynamic global vegetation model (CLM-DGVM): Technical description and user’s guide. NCAR Tech. Note TN-459+ IA 50. https://doi.org/10.5065/D6P26W36 (2004). h Technical description and user’s guide. NCAR Tech. Note TN-459+ IA 50. https://doi.org/10.5065/D6P26W36 (2004). g g B., Feddema, J., Vertenstein, M. & Kluzek, E. Technical description 33. Oleson, K. W., Bonan, G. B., Feddema, J., Vertenstein, M. & Kluzek, E. Technical description of an urban Community Land Model (CLMU). NCAR, Boulder. https://doi.org/10.5065/D6K35RM9 (2010).i 3. Oleson, K. W., Bonan, G. B., Feddema, J., Vertenstein, M. & Kluzek, E. Technical description of an urban parameterization for the Community Land Model (CLMU). NCAR, Boulder. https://doi.org/10.5065/D6K35RM9 (2010). l h l d ldfi h d 4. Westerling, A. L., Gershunov, A., Brown, T. J., Cayan, D. R. & Dettinger, M. D. Climate and wildfire in the western United States Bull. Am. Meteor. Soc. 84, 595–604 (2003). 5. De Keersmaecker, W. et al. A model quantifying global vegetation resistance and resilience to short-term climate anomalies and their relationship with vegetation cover. Glob. Ecol. Biogeogr. 24, 539–548 (2015). g g g 6. Di Lorenzo, E. & Ohman, M. D. A double-integration hypothesis to explain ocean ecosystem response to climate forcing. Proc Natl. Acad. Sci. 110, 2496–2499 (2013).i 37. Preethi, B., Sabin, T., Adedoyin, J. & Ashok, K. Impacts of the ENSO Modoki and other tropical Indo-Pacific climate-drivers on African rainfall. Sci. Rep. 5, 16653 (2015). 8. Williams, C. & Hanan, N. ENSO and IOD teleconnections for African ecosystems: Evidence of destructive interference between climate oscillations. Biogeosciences 8, 27 (2011).fl g 39. MacLeod, D., Graham, R., O’Reilly, C., Otieno, G. & Todd, M. Causal pathways linking different flavours of ENSO with the Gr Horn of Africa short rains Atmospheric Science Letters 22 e1015 https //doi org/10 1002/asl 1015 (2021) 9. MacLeod, D., Graham, R., O’Reilly, C., Otieno, G. & Todd, M. Causal pathways linking different flavours of ENSO with the Greater Horn of Africa short rains. Atmospheric Science Letters 22, e1015. https://doi.org/10.1002/asl.1015 (2021).f 40. Zheng, X.-T., Xie, S.-P. & Liu, Q. www.nature.com/scientificreports/ www.nature.com/scientificreports/ 2. Hawinkel, P. et al. Vegetation response to precipitation variability in East Africa controlled by biogeographical factors. J. Geophys Res. Biogeosci. 121, 2422–2444 (2016).fi 3. Musau, J., Patil, S., Sheffield, J. & Marshall, M. Spatio-temporal vegetation dynamics and relationship with climate over East Africa Hydrol. Earth Syst. Sci. Discuss. [preprint]. https://doi.org/10.5194/hess-2015-502 (2016).i y y p p p g 4. Archibald, S., Nickless, A., Govender, N., Scholes, R. J. & Lehsten, V. Climate and the inter-annual variability of fire in southern 4. Archibald, S., Nickless, A., Govender, N., Scholes, R. J. & Lehsten, V. Climate and the inter-annual variability of fire in southern Africa: A meta-analysis using long-term field data and satellite-derived burnt area data. Glob. Ecol. Biogeogr. 19, 794–809 (2010). 4. Archibald, S., Nickless, A., Govender, N., Scholes, R. J. & Lehsten, V. Climate and the inter-annual variability of fire in southern Africa: A meta-analysis using long-term field data and satellite-derived burnt area data. Glob. Ecol. Biogeogr. 19, 794–809 (2010). 5. Andela, N. & Van Der Werf, G. R. Recent trends in African fires driven by cropland expansion and El Nino to La Nina transition. Nat. Clim. Change 4, 791–795 (2014). . c ba d, S., N c ess, ., Gove de , N., Sc o es, R. J. & e ste , V. C ate a d t e te a ua va ab ty oi e sout e Africa: A meta-analysis using long-term field data and satellite-derived burnt area data. Glob. Ecol. Biogeogr. 19, 794–809 (2010) i 5. Andela, N. & Van Der Werf, G. R. Recent trends in African fires driven by cropland expansion and El Nino to La Nina transition Nat. Clim. Change 4, 791–795 (2014). Nat. Clim. Change 4, 791–795 (2014). 6. Papagiannopoulou, C. et al. Vegetation anomalies caused by antecedent precipitation in most of the world. Environ. Res. Lett. 12, (2017) g 6. Papagiannopoulou, C. et al. Vegetation anomalies caused by antecedent precipitation in most of the world. Environ. Res. Lett. 12 (2017).i 7. Zubkova, M., Boschetti, L., Abatzoglou, J. T. & Giglio, L. Changes in fire activity in Africa from 2002 to 2016 and their potentia drivers. Geophys. Res. Lett. 46, 7643–7653 (2019). p y 8. Williams, C. A. et al. Interannual variability of photosynthesis across Africa and its attribution. J. Geophys. Res. Biogeosci. 113 G04015. https://doi.org/10.1029/2008JG000718 (2008).h p g 9. Nicholson, S. E. & Kim, J. Data and experiments Ob ti W d A “Tropics” experiment was forced with SST anomalies over the whole tropics (15°S–15°N) for the period 1957–2016 to investigate the response to other modes of pantropical SST variability in addition to ENSO with a 3-member ensemble. An idealized “Periodic” experiment was designed to investigate the response to symmetric ENSO variability (see for instance Stuecker et al.53). The regressed ENSO SST anomaly pattern over the tropical eastern Pacific with an idealized sinusoidal 2.5 years periodicity was added to the observed SST climatology (1957–2016) and the experiment was run for 100 years with a 3-member ensem- ble. The climate response in all perturbation experiments is defined relative to the control experiment climate. Outside the tropical SST perturbation regions, the SST is the same as in the CTRL simulation. p p g To further determine the robustness of the vegetation response to tropical SST-driven climate, we performed offline simulations with dynamic and equilibrium global vegetation model, BIOME4, at a 0.5-degree resolution54, which can simulate biogeography of vegetation. BIOME4 was forced under present-day atmospheric CO2 concen- trations (367 ppm) using monthly climatological values of minimum surface temperature, surface temperature, cloudiness, and precipitation obtained from the CESM1.2.2 “Pacific”, “Tropics”, and “Periodic” experiments (see supplementary figures S2, S4). Received: 11 January 2021; Accepted: 30 April 2021 Received: 11 January 2021; Accepted: 30 April 2021 Scientific Reports \| (2021) 11:10462 \| https://doi.org/10.1038/s41598-021-89824-x www.nature.com/scientificreports/ 165 525–538 (2005). 21. Scheiter, S. & Higgins, S. I. Impacts of climate change on the vegetation of Africa: An adaptive dynamic vegetation modelling approach. Glob. Change Biol. 15, 2224–2246 (2009). 2. Frauen, C., Dommenget, D., Tyrrell, N., Rezny, M. & Wales, S. Analysis of the nonlinearity of El Niño-Southern Oscillation tel econnections. J. Clim. 27, 6225–6244 (2014). 23. Stuecker, M. F. et al. Revisiting ENSO/Indian Ocean dipole phase relationships. Geophys. Res. Lett. 44, 2481–2492 (2017). Wenhaji Ndomeni, C., Cattani, E., Merino, A. & Levizzani, V. An observational study of the variability of East African rainfall with f d l Q l S ( ) 24. Wenhaji Ndomeni, C., Cattani, E., Merino, A. & Levizzani, V. An observational study of the variability of East African rainfall respect to sea surface temperature and soil moisture. Q. J. R. Meteorol. Soc. 144, 384–404 (2018).f 25. Wolff, C. et al. Reduced interannual rainfall variability in East Africa during the last ice age. Science 333, 743–747 (2011). 26 Xi S P l I di O i ff I d P ifi li d i h f ll i El Niñ J Cli 2 f 26. Xie, S.-P. et al. Indian Ocean capacitor effect on Indo–western Pacific climate during the summer following El Niño. J. Cli 730–747 (2009). 27. Cai, W. et al. Pantropical climate interactions. Science 363, eeav4236. https://doi.org/10.1126/science.aav4236 (2019).l 28. Liu, W., Cook, K. H. & Vizy, E. K. Influence of Indian Ocean SST regionality on the East African short rains. Clim. Dyn 4991–5011 (2020).h 9. Latif, M., Dommenget, D., Dima, M. & Grötzner, A. The role of Indian Ocean sea surface temperature in forcing East African rainfall anomalies during December–January 1997/98. J. Clim. 12, 3497–3504 (1999). g y 0. Stuecker, M. F., Timmermann, A., Jin, F.-F., McGregor, S. & Ren, H.-L. A combination mode of the annual cycle and the El Niño/ Southern Oscillation. Nat. Geosci. 6, 540–544 (2013).hhii 31. Thonicke, K., Venevsky, S., Sitch, S. & Cramer, W. The role of fire disturbance for global vegetation dynamics: Coupling fire into a Dynamic Global Vegetation Model. Glob. Ecol. Biogeogr. 10, 661–677 (2001).h y g g g 2. Levis, S., Bonan, G., Vertenstein, M. & Oleson, K. The Community Land Model’s dynamic global vegetation model (CLM-DGVM) Technical description and user’s guide NCAR Tech Note TN 459+ IA 50 https://doi org/10 5065/D6P26W36 (2004) y g g g 32. Levis, S., Bonan, G., Vertenstein, M. www.nature.com/scientificreports/ www.nature.com/scientificreports/ ardel, S. et al. A new grid for the IFS. ECMWF Newsl. 146, 23–28 ( g 6. Rayner, N. et al. Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century. J. Geophys. Res. Atmosph. 108, 4407. https://doi.org/10.1029/2002JD002670 (2003). y y g 47. Zhu, Z. et al. Global data sets of vegetation leaf area index (LAI) 3g and fraction of photosynthetically active radiation (FPAR) 3g derived from global inventory modeling and mapping studies (GIMMS) normalized difference vegetation index (NDVI3g) for the period 1981 to 2011. Remote Sens. 5, 927–948 (2013). p ( ) 8. Randerson, J., Van der Werf, G., Giglio, L., Collatz, G. & Kasibhatla, P. Global Fire Emissions Database, Version 4.1 (GFEDv4) ORNL DAAC. https://doi.org/10.3334/ORNLDAAC/1293 (2015).h 9. Neale, R. B. et al. The mean climate of the Community Atmosphere Model (CAM4) in forced SST and fully coupled experiments J. Clim. 26, 5150–5168 (2013). 0. Lawrence, D. M. et al. Parameterization improvements and functional and structural advances in version 4 of the Community Land Model. J. Adv. Model. Earth Syst. 3, M03001, https://doi.org/10.1029/2011MS00045 (2011).h y p g 1. Kay, J. E. et al. The Community Earth System Model (CESM) large ensemble project: A community resource for studying climate change in the presence of internal climate variability. Bull. Am. Meteorol. Soc. 96, 1333–1349 (2015).i 2. Zebaze, S., Jain, S., Salunke, P., Shafiq, S. & Mishra, S. K. Assessment of CMIP5 multimodel mean for the historical climate o Africa. Atmos. Sci. Lett. 20, 926 (2019).i , ( ) 53. Stuecker, M. F., Jin, F.-F., Timmermann, A. & McGregor, S. Combination mode dynamics of the anomalous northwest Pacific anticyclone. J. Clim. 28, 1093–1111 (2015). 3. Stuecker, M. F., Jin, F.-F., Timmermann, A. & McGregor, S. Combination mode dynamics of the anomalous northwest Pacific anticyclone. J. Clim. 28, 1093–1111 (2015). y 4. Kaplan, J. O. et al. Climate change and Arctic ecosystems: 2. Modeling, paleodata‐model comparisons, and future projections. J Geophys. Res. Atmosph. 108, 8171. https://doi.org/10.1029/2002JD002559 (2003). Acknowledgements g I.-W. Kim, A. Timmermann, and E. Zeller were supported by the Institute for Basic Science (project code IBS- R028-D1). I.-W. Kim thanks K. Rodgers for comments and editing of the manuscript. M. F. Stuecker thanks Abigail Swann, Marysa Laguë, and Keith Oleson for discussions on CLM. J.-S. Kug and S.-W. Park were supported by NRF-2018R1A5A1024958. J.-S. Kim was supported by University of Zurich Research Priority Programme “Global Change and Biodiversity” (URPP GCB). This is SOEST publication 11337 and IPRC contribution 1520. Competing interests h p g The authors declare no competing interests. www.nature.com/scientificreports/ Response of the Indian Ocean basin mode and its capacitor effect to global warming. J. Clim. 24, 6146–6164 (2011). 1. Chu, J.-E. et al. Future change of the Indian Ocean basin-wide and dipole modes in the CMIP5. Clim. Dyn. 43, 535–551 (2014). 42. Tao, W. et al. Interdecadal modulation of ENSO teleconnections to the Indian Ocean Basin Mode and their relationship under global warming in CMIP5 models. Int. J. Climatol. 35, 391–407 (2015). g g 43. Schneider, U. et al. GPCC’s new land surface precipitation climatology based on quality-controlled in situ data and its role in quantifying the global water cycle. Theoret. Appl. Climatol. 115, 15–40 (2014). q y g g yh pp ( ) 4. Harris, I., Jones, P. D., Osborn, T. J. & Lister, D. H. Updated high-resolution grids of monthly climatic observations–the CRU TS3 10 Dataset. Int. J. Climatol. 34, 623–642 (2014). https://doi.org/10.1038/s41598-021-89824-x https://doi.org/10.1038/s41598-021-89824-x Scientific Reports \| (2021) 11:10462 \| Author contributions I.-W.K. performed the analysis and wrote the initial draft of the manuscript. A.T. and M.F.S. designed the study. M.F.S. conducted the idealized model experiments. E. Z. performed the BIOME4 model experiments and the analysis. All authors discussed the results and reviewed the manuscript. Additional informationh Supplementary Information The online version contains supplementary material available at https://doi.org/ 10.1038/s41598-021-89824-x. Correspondence and requests for materials should be addressed to I.-W.K. Reprints and permissions information is available at www.nature.com/reprints. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. © The Author(s) 2021 https://doi.org/10.1038/s41598-021-89824-x Scientific Reports \| (2021) 11:10462 \|
https://openalex.org/W2802791838	https://liu.diva-portal.org/smash/get/diva2:1300697/FULLTEXT01	English	null	Donor lymphocyte infusions in adolescents and young adults for control of advanced pediatric sarcoma	Oncotarget	2,018	cc-by	5,266	Sebastian J. Schober1, Irene von Luettichau1, Angela Wawer1, Maximilian Steinhauser1, Christoph Salat2, Wolfgang Schwinger3, Marek Ussowicz4, Petar Antunovic5, Luca Castagna6, Hans-Jochem Kolb1, Stefan E.G. Burdach1,7,* and Uwe Thiel1,* 1Department of Pediatrics and Children’s Cancer Research Center, TUM School of Medicine, Technical University of Munich, Kinderklinik München Schwabing, 80804 Munich, Germany 1Department of Pediatrics and Children’s Cancer Research Center, TUM School of Medicine, Technical University of Munich, Kinderklinik München Schwabing, 80804 Munich, Germany 2Medical Center for Hematology and Oncology Munich MVZ, 80639 Munich, Germany 2Medical Center for Hematology and Oncology Munich MVZ, 80639 Munich, Germany 3Department of Pediatrics, Medical University of Graz, A-8036 Graz, Austria 4Department of Pediatric Oncology, Hematology and Bone Marrow Transplantation, Wroclaw Medical University, 50-368 Wroclaw, Poland 4Department of Pediatric Oncology, Hematology and Bone Marrow Transplantation, Wroclaw Medical University, 50-368 Wroclaw, Poland 5Department of Hematology and Regional Tumor Registry, University Hospital Linköping, 581 85 Department of Oncology and Hematology, IRCCS Humanitas Cancer Center, Humanitas University, 6Department of Oncology and Hematology, IRCCS Humanitas Cancer Center, Humanitas University, 20089, Milan, Italy 7CCC München-Comprehensive Cancer Center, DKTK German Cancer Consortium Munich, 80336 Munich, Germany These authors share senior authorship 7CCC München-Comprehensive Cancer Center, DKTK German Cancer Consortium Munich, 80336 Munich, Germany These authors share senior authorship Correspondence to: Sebastian J. Schober, email: sebastianjohannes.schober@tum.de Keywords: donor lymphocyte infusion; allogeneic stem cell transplantation; Ewing sarcoma; rhabdomyo and transplantation Published: April 27, 2018 Copyright: Schober et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License 3.0 (CC BY 3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Copyright: Schober et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License 3.0 (CC BY 3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Copyright: Schober et al. This is an open-access article distributed under the terms of the Creative Commons BY 3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original au www.oncotarget.com www.oncotarget.com Donor lymphocyte infusions in adolescents and young adults for control of advanced pediatric sarcoma Sebastian J. Schober1, Irene von Luettichau1, Angela Wawer1, Maximilian Steinhauser1, Christoph Salat2, Wolfgang Schwinger3, Marek Ussowicz4, Petar Antunovic5, Luca Castagna6, Hans-Jochem Kolb1, Stefan E.G. Burdach1,7,* and Uwe Thiel1,* onclusions: DLI after allo-SCT may control advanced pediatric sarcoma in AYAs ontrollable toxicity. Graft-versus-host disease after DLI 2 patients (ES#4 and RMS#4) developed acute°II– °IV GvHD after DLI and patient ES#4 suffered from extensive chronic GvHD for 108 days. Patient RMS#3 developed limited chronic GvHD after allo-SCT but not after DLI. GvHD symptoms were controlled with steroids, mycophenolat mofetil, basiliximab/etanercept, repetitive extracorporal photopheresis (ECP) and third- party mesenchymal stem cells. Altogether high doses of donor lymphocytes were relatively well tolerated and did not lead to fatal complications (Tables 1 and 2). Whereas the presence of a graft-versus-leukemia effect has been shown, a hypothesized graft-versus- tumor (GvT) effect after allo-SCT in solid pediatric tumors has not yet been proven [5–7]. In ES, an inflammatory environment may have tumorigenic and immunosuppressive effects [8]. Allogeneic stem cell transplantation (allo-SCT) may abrogate disease- maintaining homeostasis and induce a favorable antitumor immunity [9]. However, it may also cause life-threatening GvHD abolishing any beneficial effect of this approach [10]. We believe that allo-SCT has the potential to render ES and RMS susceptible to cellular immunotherapy due to the recognition of foreign histocompatibility antigens and an adjuvant effect of inflammatory responses in alloimmune reactions. Indeed, the application of tumor- redirected TCR-transgenic T cells in two ES patients with haploidentical transplants did not cause GvHD and led to a partial tumor regression in at least one patient [11]. However, due to the Human Leukocyte Antigen (HLA)- restriction of TCR-transgenic T cells, this treatment is currently limited to HLA-A2 positive ES patients. For all other patients, non-specific DLI may constitute an option. Median overall survival after allo-SCT and median post-relapse survival Altogether, 7 out of 8 patients died due to progressive disease. Median OS after allo-SCT of all patients was 27.5 months. Median OS after allo-SCT of ES patients was 36.5 months and 22 months for RMS patients. Median OS after allo-SCT for patients receiving haploidentical grafts was 40 months versus 10 months for patients receiving HLA-matched grafts. Median time to death after most recent relapse (PRS) of all patients was 13 months. Median time to death after most recent relapse for patients receiving haploidentical grafts was 23 months versus 3 months for patients receiving HLA-matched grafts. To our knowledge, the role of non-specific DLI after allo-SCT in ES and RMS patients has not yet been studied systematically. ABSTRACT Background: Allogeneic stem cell transplantation (allo-SCT) and donor lymphocyte infusions (DLI) may induce a graft-versus-tumor effect in pediatric sarcoma patients. Here, we describe general feasibility, toxicity and efficacy of DLI after allo-SCT. Results: 4 of 8 patients responded. ES#4 had stable disease (SD) for 9 months after DLI and RMS#4 partial response for 8 months with combined hyperthermia/ chemotherapy. In ES#4, DLI led to SD for 6 months and reverted residual disease before allo-SCT into complete remission. After DLI, ES#4 and RMS#4 developed acute GvHD (°III–°IV), ES#4 also developed chronic GvHD. 5 patients including ES#4 lived longer than expected. Median survival after allo-SCT was 2.3 years, post-relapse survival (PRS) was 13 months. Off note, HLA-mismatched DLI were associated with a trend towards increased survival after allo-SCT and increased PRS compared to HLA-matched DLI (23 versus 3 months). Materials and Methods: We studied eight adolescents and young adults (AYAs) with advanced Ewing sarcoma (ES#1-4) and rhabdomyosarcoma (RMS#1-4) who received DLI. Escalating doses ranged from 2.5 x 104 to 1 x 108 CD3+ cells/kg body weight. AYAs were evaluated for response to DLI, graft-versus-host disease (GvHD) and survival. Conclusions: DLI after allo-SCT may control advanced pediatric sarcoma in AYAs with controllable toxicity. www.oncotarget.com Oncotarget 22741 INTRODUCTION analysis, responses to combination treatments including DLI were neither associated with the underlying disease nor a higher or lower frequency of certain matched or mismatched HLA-haplotypes (data not shown). Of note, 3 patients with responding clinical disease had received haploidentical grafts and DLI (ES#2, ES#4, RMS#4). Patients with advanced Ewing sarcoma (ES), defined as ≥ 2 bone metastases, and/or bone marrow involvement or relapse ≤ 2 years after diagnosis have poor prognoses [1]. Intensified therapy regimens including high-dose chemotherapy (HDC), involved field irradiation, autologous and allogeneic stem cell transplantation (auto- and allo-SCT) [2] could improve survival in some of these patients, excluding patients with bone marrow (BM) infiltration who do not survive irrespective of therapy [3]. Advanced rhabdomyosarcoma (RMS) patients with metastatic disease at diagnosis or those with recurrent disease have 5-year-survival rates not exceeding 30% [4]. Individual outcomes ES#1 was in complete remission (CR) after allo-SCT for 13 months until he suffered from a multifocal relapse. Relapsed disease was treated with irinotecan/temozolomide/ temsirolimus alternating with cyclophosphamide/topotecan chemotherapy, involved field irradiation and 6 repetitive DLI with up to 5 × 106 CD3+ cells/kg body weight. He showed no signs of GvHD after DLI. Tumor lesions did not respond to any treatment and progressed until DOD. ES#2 was in CR after allo-SCT for 16 months until a distant pulmonary metastasis was confirmed. Relapsed disease was treated with 7 DLI with up to 1 × 107 CD3+ cells/kg. Prior to each DLI, preparative hyperthermia with concurrent ifosfamide/ carboplatin/etoposide chemotherapy in reduced dosages (mini-ICE) was administered. He had stable pulmonary disease for 7 months after DLI without any signs of GvHD (Supplementary Figure 1). The patient suffered from tumor progression after 12 months (7 months after 1st DLI) and received molecular targeted therapies in combination with Disease outcome after DLI Abbreviations: allo-SCT = allogeneic stem cell transplantation; auto-SCT = autologous stem cell transplantation; CR = complete remission; DLI = donor lymphocyte infusion; DOD = death of disease; ES = Ewing sarcoma; GvHD = graft-versus-host disease: OS = overall survival; PD = progressive disease; PR = partial remission; RFS = relapse-free survival; SD = stable disease; UK = unknown Abbreviations: allo-SCT = allogeneic stem cell transplantation; auto-SCT = autologous stem cell transplantation; CR = complete remission; DLI = donor ly isease; ES = Ewing sarcoma; GvHD = graft-versus-host disease: OS = overall survival; PD = progressive disease; PR = partial remission; RFS = relapse UK = unknown. salvage chemotherapy regimens (vorinostat/paclitaxel/ vincristine), which led to a course of SD for 5 months. Afterwards, the patient became refractory to other combinatory treatments (vorinostat/sorafenib/tretinoin and rapamycin/dasatinib/temozolomide) and died of pulmonary disease progression (DOD). ES#3 had treatment-refractory PD and did not develop GvHD despite DLI (twice) with 1 × 106 CD3+ cells/kg until she died of disease progression (DOD; Supplementary Figure 2). ES#4 had residual disease at the time of allo-SCT, which was reverted into CR after 4 DLI with up to 3 × 105 CD3+ cells/kg. She developed extensive chronic GvHD for 108 days and stayed in CR for 19 months. Distant, localized relapse was treated with weekly DLI (60 times) with up to 3 × 105 CD3+/kg in combination with local irradiation. This therapy regimen led to a course of SD for 6 months without any signs of GvHD. PD was again treated with 5 DLI containing up to 2 × 106 CD3+ cells/kg in combination with weekly bevacizumab and local irradiation. Treatment was discontinued because of acute GvHD °III– °IV (gut, liver). Symptoms were controlled with steroids, mycophenolat mofetil, basiliximab and repetitive ECP. After GvHD control, disease progressed until DOD. million units in total) in-between the 5th and 6th DLI were administered. At the time of data assessment, he was alive in CR for 97 months without any signs of GvHD. RMS#3 developed acute GvHD (°III) and limited chronic GvHD after allo-SCT. Despite break of immune tolerance, she relapsed and suffered from refractory PD after 10 months, which was treated with irradiation and one infusion of 1 × 108 CD3+/kg donor lymphocytes. Chronic GvHD was absent before DLI. The patient then became treatment-refractory and died of disease progression (DOD). Disease outcome after DLI RMS#4 stayed in CR for 18 months after allo-SCT until she relapsed and received one dose of 1 × 106 CD3+/kg donor lymphocytes upfront, as well as hyperthermia/mini-ICE. Thereafter, she developed °II–°III acute skin GvHD, which was controlled with methylprednisolone. The hyperthermia/chemotherapy regimen was then continued and the patient experienced good PR of pancreatic as well as multiple lymph node metastases (Figure 1). PR lasted for 8 months until disease progressed leading to DOD. Individual survival data for all patients is shown in Figure 2. Disease outcome after DLI 4 out of 8 patients had a favorable clinical outcome after treatment modalities including DLI. Combination therapy of DLI and hyperthermia/chemotherapy were associated with SD for 9 months (ES#2, Supplementary Figure 1) and in patient RMS#4 with partial remission lasting 8 months (Figure 1A, 1B). Furthermore, DLI in patient ES#4 reverted residual disease before allo-SCT into complete remission and led to a course of SD lasting 6 months under repetitive DLI. Patient RMS#2 remained in CR for 97 months after his post-transplant treatment for relapsed disease including seven doses of DLI following surgical resection and salvage chemotherapy. In this www.oncotarget.com Oncotarget 22742 Table 1: Patient characteristics and individual outcomes Patient Number Gender Age at Diagnosis Primary at Diagnosis/ Stage Molecular Genetics Bone marrow Involvement at Diagnosis Reason for Allo-SCT Auto-SCT prior to Allo- SCT Age at Allo-SCT Disease Stage at Allo-SCT Hyperthermia before/ after DLI for Relapse Tx Disease Course after DLI (months) RFS after Allo-SCT (months) OS after Allo-SCT (months) Post-relapse Survival (months) Status at last Follow-up ES#1 m 15 multifocal EWS-FLI1 (type 1) Yes multifocal disease Yes 16 CR No PD 13 (relapse) 24 11 DOD ES#2 m 15 multifocal EWS-FLI1 (type 1) Yes multifocal disease Yes 16 CR Yes SD (7 m) 16 (relapse) 49 33 DOD ES#3 f 17 multifocal EWS-FLI1 (type 2) Yes multifocal disease Yes 18 PD No PD 0 (never in CR) 5 N.A. DOD ES#4 f 17 local EWS-FLI1 (type 1) No multifocal relapse Yes 21 residual disease No CR (9 m) after 4 DLI; SD (6 m) under repetitive DLI 19 (relapse) 62 43 DOD RMS#1 f 15 IV UK UK Relapse after initial treatment No 19 CR No PD 6 8 2 DOD RMS#2 m 14 IV embryonal No No CR No 14 CR No CR 28 97 69 Alive in CR RMS#3 f 25 IV alveolar Yes No CR Yes 26 CR No PD 10 13 3 DOD RMS#4 f 14 IV alveolar Yes No CR Yes 14 CR Yes PR (8 m) 18 31 13 DOD Abbreviations: allo-SCT = allogeneic stem cell transplantation; auto-SCT = autologous stem cell transplantation; CR = complete remission; DLI = donor lymphocyte infusion; DOD = death of disease; ES = Ewing sarcoma; GvHD = graft-versus-host disease: OS = overall survival; PD = progressive disease; PR = partial remission; RFS = relapse-free survival; SD = stable disease; UK = unknown. DISCUSSION RMS#1 did not show any signs of GvHD despite 2 DLI with 1 × 107 CD3+/kg. She relapsed 6 months after allo- SCT and died 2 months later due to disease progression. RMS#2 was in CR for 28 months after allo-SCT before he relapsed. He was then treated with surgery and chemotherapy according to the CWS 96 relapse protocol. Consecutively, 7 escalating doses of donor lymphocytes with up to 1 × 108 CD3+/kg in combination with interleukin-2 (25 Patients with advanced ES and RMS may become eligible for allo-SCT to induce a hypothesized GvT effect. This approach, however, may result in life- threatening toxicities and GvHD. In the past, reduced intensity conditioning (RIC) was implemented to reduce HDC-associated toxicity and to facilitate a GvT effect in those patients. However, reduced toxicity was bought www.oncotarget.com Oncotarget 22743 with higher relapse rates leading to equal OS compared to HDC-based regimen and the role of allo-SCT to induce a GvT effect remained unclear [1]. Most pediatric tumors are considered poorly immunogenic compared to e.g. melanoma or lung cancer, correlating with low response rates to immune-checkpoint inhibitors [12, 13]. Further immune-evasive factors, such as low or varying HLA class I expression in-between patients and tumor sites, immune-modulatory pro-neoplastic effects of the tumor-microenvironment (TME) or high frequencies of regulatory T cells (Tregs) have been described [14–17]. Thus, strategies to render those tumors susceptible to immunotherapeutic approaches are needed. We hypothesize that allo-SCT does not constitute an endpoint but a starting point of immunotherapy in a subgroup of patients yet to be defined. Several reports such as the use of adjuvant tumor lysate-pulsed dendritic cell vaccination in patients after standard antineoplastic treatment [18] as well as the use of natural killer (NK) cells raise hope that ES may become targetable with immunotherapy [19, 20]. In the present analysis, 4 out of 8 patients responded to treatments including DLI. Clinical responses in patient ES#2 and RMS#4 were achieved with additional hyperthermia and mini-ICE. Apart from conversion to CR (after allo-SCT) and SD with weekly DLI in patient ES#4, DLI-associated tumor responses were observed in combination approaches (e.g. hyperthermia, mini-ICE). 5 out of 8 patients lived longer than expected compared to historical study groups, including ES #4 with chronic GvHD [1, 9, 21]. Compared to survival data from non-transplanted patients with similar risk profiles [21], the patients in this study fared better. DISCUSSION Post-relapse survival for all patients was 13 months, for ES patients 33 months and for RMS patients 8 months. Without CR2, median time to death in the study population of Ferrari et al. was 6 months [21]. Therefore, we hypothesize that there is a benefit in overall survival for the subgroup of high-risk patients in comparison to standard of care therapy that includes HDC as well. This survival benefit appears even more obvious when the subgroup of HLA-mismatched patients, who received DLI, is considered separately (40 months of median overall survival after allo-SCT for haploidentical transplanted patients versus 10 months for HLA-matched transplanted patients). In accordance, the post-relapse survival for haploidentical transplanted patients is higher than for those having received an HLA-matched graft (median time to death: 23 months versus 3 months), possibly originating from a stronger alloreactivity of donor cells against tumor alloantigens in a HLA-mismatched versus HLA-identical setting. Figure 1: Positron emission tomography–computed tomography (PET-CT) of patient RMS#4 shows reduction of pathological [18F] fluorodeoxyglucose (FDG)-uptake of pancreatic tumour mass under combinatory treatment including DLI, hyperthermia and mini-ICE chemotherapy indicating partial tumor regression (A: before DLI, B: eight months after DLI). Figure 1: Positron emission tomography–computed tomography (PET-CT) of patient RMS#4 shows reduction of pathological [18F] fluorodeoxyglucose (FDG)-uptake of pancreatic tumour mass under combinatory treatment including DLI, hyperthermia and mini-ICE chemotherapy indicating partial tumor regression (A: before DLI, B: eight months after DLI). www.oncotarget.com Oncotarget 22744 Table 2: Graft and DLI-related information Table 2: Graft and DLI-related information Patient Number Allo-SCT Conditioning HLA match GvHD Treatment/ Prophylaxis Onset of GvHD after allo-SCT Acute GvHD after Allo-SCT Chronic GvHD after Allo-SCT Day of DLI post allo-SCT DLI for Relapse/ Progression after Allo-SCT (CD3+ per kg/BW) Onset of GvHD after DLI Acute GvHD after DLI Chronic GvHD after DLI ES#1 FLU/MEL/ OKT3/TT haploidentical (mother) STER/CSA d+34 Stage I–II (skin) None d+485, d+566, d+639, d+691, d+716 Six times; 2.5, 5, 10, 25, 100 and 500 × 104 None None None ES#2 FLU/MEL/ OKT3/TT haploidentical; class I matched (father) CSA d+63 Stage I (skin) None 7 times; in- between d+642 and d+824 Seven times; 2.5, 5, 10, 25, 100, 500 and 1.000 × 104 None None None ES#3 FLU/MEL/ OKT3/TT HLA identical match (sibling) CSA None None N.A. DISCUSSION d+70, d+104 Twice; 10 × 104 and 100 × 104 None None None ES#4 FLU/MEL/ OKT3/TT haploidentical (sibling) BASX, ECP,STER/ MMF, OKT3 None None None N.A. after 1st four DLIs; 64 times in-between d+606 and d+917, and six times d+1081 and d+1133 74 times; 4 times to consolidate allo-SCT up to 30 × 104, 64 times up to 30 × 104 and six times up to 200 × 104 N.A. after 1st 4 DLI, d+63 after 1st DLI/ d+10 after last six DLI twice; Stage II–III (skin) after 1st four DLIs; Stage III–IV (intestine, liver) after last six DLIs up to 200 × 104 Extended RMS#1 CRBPL/ MEL/TT HLA identical match (sibling) UK None None None N.A. twice; 1.000 × 104 None None N.A. RMS#2 BU/CTX/ TT HLA identical match (sibling) CSA None None None N.A. Seven times; 10, 30, 50, 100, 250, 500 and 1.000 × 104 None None None RMS#3 CTX/FLU HLA identical match (sibling) Yes, but UK N.A. Stage III Limited N.A. once; 10.000 × 104 None None N.A. RMS#4 FLU/MEL/ OKT3/TT haploidentical (father) ETN, MSC, STER/ MMF, OKT3 d+49 Stage IV (intestine) Non d+576 once; 100 × 104 d+53 (not biopsy- proven, but responsive to treatment) Stage II–III (skin) None Abbreviations: allo-SCT = allogeneic stem cell transplantation; BASX = basiliximab; BU = busulfan; BW = body weight; CSA = cyclosporine A; CTX = cyclophosphamide; CRBPL = carboplatin; d = day; DLI = donor lymphocyte infusion; ECP = extracorporal photopheresis; ES = Ewing sarcoma; ETN = etanercept; FLU = fludarabin; GvHD = graft-versus-host disease; HLA = heuman leukocyte antigen; MEL = melphalan; MMF = mycophenolat mofetil; MSC = mesenchymal stem cells; N.A. = not assessable; OKT3 = anti-CD3; RMS = rhabdomyosarcoma; STER = steroids; TT = thiotepa; UK = unknown. Abbreviations: allo-SCT = allogeneic stem cell transplantation; BASX = basiliximab; BU = busulfan; BW = body weight; CSA = cyclosporine A; CTX = cyclophosphamide; CRBPL = carboplatin; d = day; DLI = donor lymphocyte infusion; ECP = extracorporal photopheresis; ES = Ewing sarcoma; ETN = etanercept; FLU = fludarabin; GvHD = graft-versus-host disease; HLA = heuman leukocyte antigen; MEL = melphalan; MMF = mycophenolat mofetil; MSC = mesenchymal stem cells; N.A. = not assessable; OKT3 = anti-CD3; RMS = rhabdomyosarcoma; STER = steroids; TT = thiotepa; UK = unknown. DISCUSSION Eliciting a pro-inflammatory environment using allo-SCT may lead to enhanced phagocytic, NK cell as well as T cell activity [22]. In this regard, there is supporting evidence that alloreactive NK cells play a role in controlling minimal residual disease (MRD) in leukemia and solid pediatric tumors and are discussed to GvHD symptoms after DLI were absent in 6 out of 8 patients despite high doses of up to 1 × 108 CD3+ T cells/ kg. At the time of treatment, 4 DLI-non-responders had active tumor burden increasing in size, suggesting disease- triggered immune-evasive mechanisms or even a general immune-suppressive host environment. Figure 2: Individual survival (in months) after allogeneic stem cell transplantation is depicted for ES (A) and RMS patients (B). Abbreviations: allo-SCT, allogeneic stem cell transplantation; CR, complete remission; DOD, death of disease; ES, Ewing’s sarcoma patient; OS, overall survival; RFS, relapse-free survival; RMS, rhabdomyosarcoma patient. Patients ES #1/#2/#4 and RMS #4 received HLA-mismatched, haploidentical grafts. Patients ES #3 and RMS #1/#2/#3 received HLA-matched grafts from sibling donors. Figure 2: Individual survival (in months) after allogeneic stem cell transplantation is depicted for ES (A) and RMS patients (B). Abbreviations: allo-SCT, allogeneic stem cell transplantation; CR, complete remission; DOD, death of disease; ES, Ewing’s sarcoma patient; OS, overall survival; RFS, relapse-free survival; RMS, rhabdomyosarcoma patient. Patients ES #1/#2/#4 and RMS #4 received HLA-mismatched, haploidentical grafts. Patients ES #3 and RMS #1/#2/#3 received HLA-matched grafts from sibling donors. www.oncotarget.com Oncotarget 22745 survival (RFS). Overall survival after allo-SCT (OS) was calculated from the time of allo-SCT until time of death. Post-relapse survival (PRS) was calculated from the time of most recent relapse after allo-SCT until death of disease (DOD). Expected median survival after allo-SCT is referred to most recent published data, i.e. 12 months for both ES [21] and RMS patients [9]. All patients and their legal guardians signed informed consent prior to treatment which was approved by the institutional review boards according to precepts established by the Helsinki Conference Declaration. prolong survival after haploidentical allo-SCT compared to a HLA-matched setting [19]. A recent analysis showed even lower rates of chronic GvHD rates and similar survival outcomes of haploidentical allo-SCT compared to matched sibling or unrelated donor transplants by implementing post-transplantation cyclophosphamide [23]. In the present analysis, DLI after allo-SCT are a feasible therapy option for a subgroup of advanced ES and RMS patients. DISCUSSION Of particular importance, GvHD rates and extent were controllable and no patient died of complications. Interpretation of the present data is limited due to low patient numbers and patient heterogeneity. Nevertheless, tumor control was observed in a number of patients. ACKNOWLEDGMENTS We would like to thank all patients and their families as well as treating physicians and nurses for their contribution to this study. Data preparation and analysis Information on patients’ pretreatment is elaborated in the Supplementary Materials. We re-evaluated data of eight patients (14–26 years) from previous publications with advanced ES and RMS, who received DLI after allo-SCT between 1997 and 2011 [3, 9]. Median follow-up after allo-SCT was 27.5 months. Inclusion criteria were diagnosis of ES or RMS (all subtypes, respectively), DLI after allo-SCT and non- participation in ongoing prospective trials. Diagnosis and relapse were confirmed by histopathology. Tumor size was monitored with CT or MR-based imaging. Disease specific chromosomal translocations were detected by molecular-genetic testing. As we describe retrospective data of a small heterogeneous group we did not perform statistical significance tests. Graft source, HLA-match and GvHD prophylaxis/treatment Consecutively, our data suggest that high-risk pediatric ES and RMS patients when considered for an allo-SCT, should receive haploidentical grafts and additional pre-emptive DLI, in case acute GvHD symptoms are absent. 7 out of 8 patients received CD3/CD19-depleted allografts from peripheral blood (PB). Only RMS#2 received a bone marrow (BM) allograft. 4 out of 8 patients (ES#3, RMS#1, #2 and #3) received HLA-matched grafts from sibling-donors. Patients ES#1, #2, #4 and RMS#4 received HLA-mismatched grafts from family members (Table 2). Conditioning regimens, GvHD prophylaxis and treatment are also listed in Table 2. Patients did not receive immunosuppressive therapy during or after DLI. When patients developed GvHD symptoms (°II), DLI were discontinued and immunosuppression was started (compare Table 2). If the allo-SCT setting has the capacity to cause immune-susceptibility in bulky tumors, there is hope to cure allo-transplanted ES and RMS patients with high-risk relapse profiles in CR using pre-emptive DLI. CONFLICTS OF INTEREST S. Burdach has ownership interest in PDL biopharma and holds intellectual property (EU and US patents) in gene expression analysis. The other authors declare no potential conflicts of interest. S. Burdach has ownership interest in PDL biopharma and holds intellectual property (EU and US patents) in gene expression analysis. The other authors declare no potential conflicts of interest. REFERENCES rhabdomyosarcoma: a retrospective assessment. Br J Cancer. 2013; 109:2523–32. https://doi.org/10.1038/ bjc.2013.630. 1. Thiel U, Wawer A, Wolf P, Badoglio M, Santucci A, Klingebiel T, Basu O, Borkhardt A, Laws HJ, Kodera Y, Yoshimi A, Peters C, Ladenstein R, et al. No improvement of survival with reduced- versus high-intensity conditioning for allogeneic stem cell transplants in Ewing tumor patients. Ann Oncol. 2011; 22:1614–21. https://doi.org/10.1093/ annonc/mdq703. 10. 10. Baird K, Fry TJ, Steinberg SM, Bishop MR, Fowler DH, Delbrook CP, Humphrey JL, Rager A, Richards K, Wayne AS, Mackall CL. Reduced-intensity allogeneic stem cell transplantation in children and young adults with ultrahigh-risk pediatric sarcomas. Biol Blood Marrow Transplant. 2012; 18:698–707. https://doi.org/10.1016/j. bbmt.2011.08.020. 2. Burdach S, Meyer-Bahlburg A, Laws HJ, Haase R, van Kaik B, Metzner B, Wawer A, Finke R, Gobel U, Haerting J, Pape H, Gadner H, Dunst J, et al. High-dose therapy for patients with primary multifocal and early relapsed Ewing‘s tumors: results of two consecutive regimens assessing the role of total-body irradiation. J Clin Oncol. 2003; 21:3072– 8. https://doi.org/10.1200/JCO.2003.12.039. 11. Thiel U, Schober SJ, Einspieler I, Kirschner A, Thiede M, Schirmer D, Gall K, Blaeschke F, Schmidt O, Jabar S, Ranft A, Alba Rubio R, Dirksen U, et al. Ewing sarcoma partial regression without GvHD by chondromodulin-I/HLA- A*02:01-specific allorestricted T cell receptor transgenic T cells. Oncoimmunology. 2017; 6:e1312239. https://doi.org/ 10.1080/2162402X.2017.1312239. 3. Thiel U, Wawer A, von Luettichau I, Bender HU, Blaeschke F, Grunewald TG, Steinborn M, Roper B, Bonig H, Klingebiel T, Bader P, Koscielniak E, Paulussen M, et al. Bone marrow involvement identifies a subgroup of advanced Ewing sarcoma patients with fatal outcome irrespective of therapy in contrast to curable patients with multiple bone metastases but unaffected marrow. Oncotarget. 2016; 7:70959–68. https://doi.org/10.18632/ oncotarget.10938. 12. Merchant MS, Wright M, Baird K, Wexler LH, Rodriguez- Galindo C, Bernstein D, Delbrook C, Lodish M, Bishop R, Wolchok JD, Streicher H, Mackall CL. Phase I Clinical Trial of Ipilimumab in Pediatric Patients with Advanced Solid Tumors. Clin Cancer Res. 2016; 22:1364–70. https:// doi.org/10.1158/1078-0432.CCR-15-0491. 13. Rizvi NA, Hellmann MD, Snyder A, Kvistborg P, Makarov V, Havel JJ, Lee W, Yuan J, Wong P, Ho TS, Miller ML, Rekhtman N, Moreira AL, et al. Cancer immunology. Mutational landscape determines sensitivity to PD-1 blockade in non-small cell lung cancer. Science. 2015; 348:124–8. https://doi.org/10.1126/science.aaa1348. 4. Stevens MC. Treatment for childhood rhabdomyosarcoma: the cost of cure. Lancet Oncol. 2005; 6:77–84. https://doi. org/10.1016/S1470-2045(05)01733-X. 5. FUNDING Disease was staged according to WHO classifications. Complete remission (CR) was defined as the absence of detectable disease, progressive disease (PD) as at least 25% increase in tumor volume of previous lesions despite treatment. Partial remission (PR) was defined as at least 50% tumor volume reduction. Residual disease included both PD and PR. Stable disease (SD) was defined as neither PD, PR nor CR. GvHD was graded using the Glucksberg criteria. The period of CR after allo-SCT until re-occurrence of any local or metastatic disease was termed relapse-free The study was supported by grants to SB from Wilhelm Sander-Stiftung (2006.109.1), Else- Kröner-Fresenius-Stiftung (P31/08//A123/07) and the Deutsche Kinderkrebsstiftung (DKS 2010.07); SB and UT received grants from the Federal Ministry of Education and Research Germany (TranSarNet 01GM1104B; TranSarNet 01GM0869). IvL received a grant from Cura Placida Children’s Cancer Research Foundation. www.oncotarget.com Oncotarget 22746 REFERENCES Koscielniak E, Gross-Wieltsch U, Treuner J, Winkler P, Klingebiel T, Lang P, Bader P, Niethammer D, Handgretinger R. Graft-versus-Ewing sarcoma effect and long-term remission induced by haploidentical stem- cell transplantation in a patient with relapse of metastatic disease. J Clin Oncol. 2005; 23:242–4. https://doi. org/10.1200/JCO.2005.05.940. 14. Berghuis D, de Hooge AS, Santos SJ, Horst D, Wiertz EJ, van Eggermond MC, van den Elsen PJ, Taminiau AH, Ottaviano L, Schaefer KL, Dirksen U, Hooijberg E, Mulder A, et al. Reduced human leukocyte antigen expression in advanced-stage Ewing sarcoma: implications for immune recognition. J Pathol. 2009; 218:222–31. https://doi. org/10.1002/path.2537. 6. Ohta H, Hashii Y, Yoshida H, Kusuki S, Tokimasa S, Yoneda A, Fukuzawa M, Inoue N, Hara J, Kusafuka T, Ozono K. Allogeneic hematopoietic stem cell transplantation against recurrent rhabdomyosarcoma. J Pediatr Hematol Oncol. 2011; 33:e35–8. https://doi.org/10.1097/ MPH.0b013e3181e7ddc5. 15. Brinkrolf P, Landmeier S, Altvater B, Chen C, Pscherer S, Rosemann A, Ranft A, Dirksen U, Juergens H, Rossig C. A high proportion of bone marrow T cells with regulatory phenotype (CD4+CD25hiFoxP3+) in Ewing sarcoma patients is associated with metastatic disease. Int J Cancer. 2009; 125:879–86. https://doi.org/10.1002/ijc.24461. 7. Kolb HJ. Graft-versus-leukemia effects of transplantation and donor lymphocytes. Blood. 2008; 112:4371–83. https:// doi.org/10.1182/blood-2008-03-077974. 16. Sorbye SW, Kilvaer T, Valkov A, Donnem T, Smeland E, Al-Shibli K, Bremnes RM, Busund LT. Prognostic impact of lymphocytes in soft tissue sarcomas. PLoS One. 2011; 6:e14611. https://doi.org/10.1371/journal.pone.0014611. 8. Lissat A, Joerschke M, Shinde DA, Braunschweig T, Meier A, Makowska A, Bortnick R, Henneke P, Herget G, Gorr TA, Kontny U. IL6 secreted by Ewing sarcoma tumor microenvironment confers anti-apoptotic and cell- disseminating paracrine responses in Ewing sarcoma cells. BMC Cancer. 2015; 15:552. https://doi.org/10.1186/ s12885-015-1564-7. 17. Joyce JA, Fearon DT. T cell exclusion, immune privilege, and the tumor microenvironment. Science. 2015; 348:74– 80. https://doi.org/10.1126/science.aaa6204. 18. Merchant MS, Bernstein D, Amoako M, Baird K, Fleisher TA, Morre M, Steinberg SM, Sabatino M, Stroncek DF, Venkatasan AM, Wood BJ, Wright M, Zhang H, et al. Adjuvant Immunotherapy to Improve Outcome in High- Risk Pediatric Sarcomas. Clin Cancer Res. 2016; 22:3182– 91. https://doi.org/10.1158/1078-0432.CCR-15-2550. 9. Thiel U, Koscielniak E, Blaeschke F, Grunewald TG, Badoglio M, Diaz MA, Paillard C, Prete A, Ussowicz M, Lang P, Fagioli F, Lutz P, Ehninger G, et al. Allogeneic stem cell transplantation for patients with advanced www.oncotarget.com Oncotarget 22747 patients with Ewing sarcoma. Pediatr Blood Cancer. 2015; 62:994–9. https://doi.org/10.1002/pbc.25388. 19. REFERENCES Schlegel P, Feuchtinger T, Nitschke-Gerard C, Seidel UJ, Lang AM, Kyzirakos C, Teltschik HM, Ebinger M, Schumm M, Koscielniak E, Handgretinger R, Lang P. Favorable NK cell activity after haploidentical hematopoietic stem cell transplantation in stage IV relapsed Ewing‘s sarcoma patients. Bone Marrow Transplant. 2015; 50:S72–6. https:// doi.org/10.1038/bmt.2015.100. patients with Ewing sarcoma. Pediatr Blood Cancer. 2015; 62:994–9. https://doi.org/10.1002/pbc.25388. 22. Bouchlaka MN, Redelman D, Murphy WJ. Immunotherapy following hematopoietic stem cell transplantation: potential for synergistic effects. Immunotherapy. 2010; 2:399–418. https://doi.org/10.2217/imt.10.20. 23. Martinez C, Gayoso J, Canals C, Finel H, Peggs K, Dominietto A, Castagna L, Afanasyev B, Robinson S, Blaise D, Corradini P, Itala-Remes M, Bermudez A, et al. Post-Transplantation Cyclophosphamide-Based Haploidentical Transplantation as Alternative to Matched Sibling or Unrelated Donor Transplantation for Hodgkin Lymphoma: A Registry Study of the Lymphoma Working Party of the European Society for Blood and Marrow Transplantation. J Clin Oncol. 2017; 35:3425–32. https:// doi.org/10.1200/JCO.2017.72.6869. 20. Lang P, Pfeiffer M, Muller I, Schumm M, Ebinger M, Koscielniak E, Feuchtinger T, Foll J, Martin D, Handgretinger R. Haploidentical stem cell transplantation in patients with pediatric solid tumors: preliminary results of a pilot study and analysis of graft versus tumor effects. Klin Padiatr. 2006; 218:321–6. https://doi. org/10.1055/s-2006-942256. 21. Ferrari S, Luksch R, Hall KS, Fagioli F, Prete A, Tamburini A, Tienghi A, DiGirolamo S, Paioli A, Abate ME, Podda M, Cammelli S, Eriksson M, et al. Post-relapse survival in www.oncotarget.com Oncotarget 22748
https://openalex.org/W3082798739	https://www.nature.com/articles/s41698-020-00131-6.pdf	English	null	Mendelian randomization analyses of genetically predicted circulating levels of cytokines with risk of breast cancer	npj precision oncology	2,020	cc-by	6,075	INTRODUCTION with complex traits, bring the wide usage of MR to a new climax22,23. Recently, a comprehensive MR analysis regarding genetically determined levels of circulating cytokines and risk of stroke gave us a good demonstration24. It was implemented leveraging the most comprehensive GWAS which evaluated 41 cytokines and growth factors in 8293 healthy subjects of Finnish ancestry25, and the largest GWAS meta-analysis on stroke and stroke subtypes to date (MEGASTROKE)26. Breast cancer (BC) remains the main cause of cancer death and the second most common female cancer in western countries, although the death rate of BC has dropped by 40% from 1989 to 20161,2. Inﬂammatory mechanisms have been implicated in genesis, development and metastasis of BC3,4, and ﬁnally the prognosis and recurrence of BC5–7. Cytokines have indispensable functions in pathology of inﬂammation8,9 and could serve as the potential targets of anti-inﬂammatory intervention of BC. yp ( ) Through checking the UKB SNP-Heritability Browser, we found the SNP-h2 for BC was 0.144. Hereby, to clarify whether genetically predicted circulating levels of cytokines are associated with risk of BC, we implemented two-sample MR analyses using data from this most comprehensive GWAS on cytokines25 and the largest BC GWAS from the Breast Cancer Association Consortium (BCAC) with totally 122,977 BC cases and 105,974 healthy controls27. Totally 24 cytokines with available data, including Eotaxin, growth-regulated oncogene-a (GROa), interleukin-12, p70 (IL12p70), interleukin-13 (IL13), interleukin-16(IL16), monocyte chemoattractant protein-1/ C-C motif chemokine ligand 2 (MCP1/CCL2), macrophage migra- tion inhibitory factor (MIF), macrophage inﬂammatory protein-1 beta/C-C motif chemokine ligand 4 (MIP1b/CCL4), and stem cell factor (SCF), beta nerve growth factor (bNGF), cutaneous T-cell attracting (CTACK), hepatocyte growth factor (HGF), interleukin-17 (IL17), interleukin-18 (IL18), interleukin-2 receptor, alpha subunit (IL2ra), interferon gamma-induced protein 10 (IP10), macrophage colony-stimulating factor (MCSF), monokine induced by interferon-gamma (MIG), platelet derived growth factor BB (PDGFbb), regulated on activation, normal T cell expressed and secreted (RANTES), stem cell growth factor beta (SCGFb), tumor necrosis factor-beta (TNFb), TNF-related apoptosis inducing ligand (TRAIL), and vascular endothelial growth factor (VEGF), were evaluated for their associations with BC risk. We predicted the y The role of cytokines in BC carcinogenesis has been widely explored by observational studies and yielded equivocal results, with suggestive associations between risk of BC and abnormal levels of cytokines, including tumor necrosis factor-alpha (TNF-α), interleukin-6 (IL6), C-reactive protein (CRP), transforming growth factor β (TGF-β)10–13, etc. www.nature.com/npjprecisiononcology 1The second clinical college, Chongqing Medical University, Chongqing, China. 2Department of Breast and Thyroid Surgery, Daping Hospital, Third Military Medical University, Chongqing, China. 3Department of Epidemiology, College of Preventive Medicine, Third Military Medical University, Chongqing, China. 4Student Brigade, College of Basic Medicine, Third Military Medical University, Chongqing, China. 5Department of Anaesthesia, The ﬁrst afﬁliated hospital of Third Military medical University, Chongqing, China. 6Banan People’s hospital of Chongqing, Chongqing, China. 7College of public health, Southwest medical University, Luzhou, China. 8These authors contributed equally: Shen Li, Yan Xu, Yao Zhang. ✉email: xymacq@hotmail.com Mendelian randomization analyses of genetically predicted circulating levels of cytokines with risk of breast cancer Shen Li1,8, Yan Xu2,8, Yao Zhang3,8, Lili Nie4, Zhihua Ma5, Ling Ma6, Xiaoyu Fang7 and Xiangyu Ma3✉ To determine whether genetically predicted circulating levels of cytokines are associated with risk of overall breast cancer (BC), estrogen receptor (ER)-positive and ER-negative BC, we conducted two-sample MR analyses using data from the most comprehensive genome-wide association studies (GWAS) on cytokines in 8293 Finnish participants and the largest BC GWAS from the Breast Cancer Association Consortium (BCAC) with totally 122,977 BC cases and 105,974 healthy controls. We systematically screened 41 cytokines (of which 24 cytokines have available instruments) and identiﬁed that genetically predicted circulating levels (1-SD increase) of MCP1 (OR: 1.08; 95% CIs: 1.03–1.12; P value: 3.55 × 10−4), MIP1b (OR: 1.02; 95% CIs: 1.01–1.04; P value: 2.70 × 10−3) and IL13 (OR: 1.06; 95% CIs: 1.03–1.10; P value: 3.33 × 10−4) were signiﬁcantly associated with increased risk of overall BC, as well as ER-positive BC. In addition, higher levels of MIP1b and IL13 were also signiﬁcantly associated with increased risk of ER-negative BC. These ﬁndings suggest the crucial role of cytokines in BC carcinogenesis and potential of targeting speciﬁc inﬂammatory cytokines for BC prevention. npj Precision Oncology (2020) 4:25 ; https://doi.org/10.1038/s41698-020-00131-6 Published in partnership with The Hormel Institute, University of Minnesota RESULTS Overview outlines of the cytokines with BC risk y In the current study, we evaluated whether genetically predicted circulating levels of cytokines are associated with risk of overall BC, ER-positive and ER-negative BC. Through ﬁltering the threshold of signiﬁcance (p < 5 × 10−8), false discovery rate (FDR < 5%), F- statistics (>10), and linkage disequilibrium (LD; r2 < 0.1), 229 independent SNPs were ﬁnally selected as the proxy of circulating levels of 24 cytokines. Figure 1 and Supplementary Table 1 presented the number of SNPs included in the genetic instruments, along with R2, F-statistics for the instruments, and the results of MR analyses of circulating levels of 24 cytokine with risk of BC and subgroups. F-statistics for their respective genetic instruments ranged from 29 to 636, suggesting that our analyses were unlikely to suffer from weak instrument bias. Among them, nine cytokines, including Eotaxin, GROa, IL12p70, IL13, IL16, MCP1, MIF, MIP1b, and SCF, showed signiﬁcant associations with risk of BC in either total participants or subgroups (Fig. 1). Higher levels of eight cytokines was associated with increased risk of BC, while higher level of MIF was associated with decreased risk of BC. Even following Bonferroni correction for testing multiple cytokines (p < 0.05/48 = 1.1 × 10−3), three cytokines, including MCP1, MIP1b and IL13, still showed statistically signiﬁcant associations with risk of total BC or subtype BC. Detection of possible pleiotropy effect To remove the potential inﬂuence of the possible pleiotropy effect, a series of measures have been adopted. At the stage of instrument construction, SNPs those were associated with levels of more than one cytokine were removed. At the stage of data analysis, ﬁrst there was no indication for directional pleiotropy effects as assessed by the MR-Egger intercept for MCP1, MIP1b, and IL13 (P value: 0.200, 0.112, and 0.224, respectively). Second, MR-PRESSO analyses were conducted and revealed no potential outliers for MCP1 and IL13 (P value: 0.533 and 0.350, respectively). One outlier variant (rs7621046) was detected for MIP1b. Following exclusion of SNP rs7621046 in the restrictive MR analysis, genetically raised MIP1b was still associated with higher risk of any BC (OR: 1.02; 95% CIs: 1.01–1.03; P value: 1.80 × 10−3), while Cochran Q test for the heterogeneity didn’t reject the null hypothesis (Cochran P value: 0.151). RESULTS Further, comprehensive lookup of the MR-Base, PhenoScanner and the GWAS catalog didn’t reveal signiﬁcant associations of the selected variants in current study with other traditional BC risk factors, including height, obesity, age of menarche and menopause. Genetically predicted circulating levels of MCP1, MIP1b, and IL13 and risk of BC Published in partnership with The Hormel Institute, University of Minnesota INTRODUCTION However, these relationships between inﬂammatory cytokines and BC risk identiﬁed in observational studies could be susceptible to confounding factors, limited follow-up time, small sample size and the reverse causation, which might mislead the conclusions14. For example, contrary to previous studies, plasma CRP level was not found to be signiﬁcantly associated with BC risk in the Women’s Health Study with 27,919 participants15. In addition, both serum levels of IL6 and TNF-α were associated with either increased or decreased risk of BC in different samples16–19. Hence, the potential causality of individual cytokines in determining BC risk remains elusive. One more robust method being widely for causal inference is Mendelian randomization (MR), which could overcome the limitations of observational research20,21. MR uses genetic variants as instrumental variables, avoiding being misled by confounding factors, limited follow-up time, small sample size and the reverse causation of different traits14,20. Genome-wide association studies (GWAS), which have identiﬁed thousands of variants associated Published in partnership with The Hormel Institute, University of Minnesota S Li et al. 2 circulating cytokine levels using a set of estimated SNP effects based on a GWAS to identify protein quantitative trait loci (pQTL). Estrogen receptor (ER) status is important as a BC prognostic and predictive biomarker, and could also affect the medical decision of hormonal therapy or other treatments28. Thus, we also conducted stratiﬁed analyses of ER-positive and ER-negative BC. heterogeneity in the MCP1 and IL13 associations with BC didn’t reject the null hypothesis (Cochran P value: 0.494 and 0.5 respectively). However, signiﬁcant heterogeneity was detected MIP1b (Cochran P value: 0.010). We also ran the MR analy including the pleiotropic SNPs in the sensitivity analyses, a found results were not changed for IL13 and MIP1b. Genetica Fig. 1 Mendelian randomization analyses of circulating cytokine and growth factor levels with risk of BC and subgroups. Fig. 1 Mendelian randomization analyses of circulating cytokine and growth factor levels with risk of BC and subgroups. Fig. 1 Mendelian randomization analyses of circulating cytokine and growth factor levels with risk of BC and subgroups. Fig. circulating cytokine levels using a set of estimated SNP effects based on a GWAS to identify protein quantitative trait loci (pQTL). Estrogen receptor (ER) status is important as a BC prognostic and predictive biomarker, and could also affect the medical decision of hormonal therapy or other treatments28. Thus, we also conducted stratiﬁed analyses of ER-positive and ER-negative BC. INTRODUCTION heterogeneity in the MCP1 and IL13 associations with BC risk didn’t reject the null hypothesis (Cochran P value: 0.494 and 0.582, respectively). However, signiﬁcant heterogeneity was detected for MIP1b (Cochran P value: 0.010). We also ran the MR analysis including the pleiotropic SNPs in the sensitivity analyses, and found results were not changed for IL13 and MIP1b. Genetically raised MCP1 (OR: 1.06; 95% CIs: 1.02–1.10; P value: 0.001) was still associated with increased BC risk, basing on 14 instrumental SNPs. During the bidirectional MR, results showed no evidence of BC causally inﬂuencing the levels of MCP1 (OR: 1.00; 95% CIs: 0.95–1.04; P value: 0.885), MIP1b (OR: 1.01; 95% CIs: 0.97–1.06; P value: 0.573) and IL13 (OR: 1.03; 95% CIs: 0.96–1.11; P value: 0.380). heterogeneity in the MCP1 and IL13 associations with BC risk didn’t reject the null hypothesis (Cochran P value: 0.494 and 0.582, respectively). However, signiﬁcant heterogeneity was detected for MIP1b (Cochran P value: 0.010). We also ran the MR analysis including the pleiotropic SNPs in the sensitivity analyses, and found results were not changed for IL13 and MIP1b. Genetically raised MCP1 (OR: 1.06; 95% CIs: 1.02–1.10; P value: 0.001) was still associated with increased BC risk, basing on 14 instrumental SNPs. During the bidirectional MR, results showed no evidence of BC causally inﬂuencing the levels of MCP1 (OR: 1.00; 95% CIs: 0.95–1.04; P value: 0.885), MIP1b (OR: 1.01; 95% CIs: 0.97–1.06; P value: 0.573) and IL13 (OR: 1.03; 95% CIs: 0.96–1.11; P value: 0.380). npj Precision Oncology (2020) 25 DISCUSSION Our systematic retrieval only identiﬁed limited npj Precision Oncology (2020) 25 S Li et al. 3 Fig. 2 Causal estimates of selected cytokines (MCP1, MIP1b and IL13) on risk of overall, ER-positive and ER-negative breast cancer. Inverse-variance weighted (IVW) method was used as the primary method for the MR analyses to test the potential causal associations between Circulating Levels of selected cytokines and BC risk. Causal estimates express the change in odds ratio (OR) per standard deviation (SD) increment in cytokine concentration. Error bars indicate 95% conﬁdence intervals. Fig. 2 Causal estimates of selected cytokines (MCP1, MIP1b and IL13) on risk of overall, ER-positive and ER-negative breast cancer. Inverse-variance weighted (IVW) method was used as the primary method for the MR analyses to test the potential causal associations between Circulating Levels of selected cytokines and BC risk. Causal estimates express the change in odds ratio (OR) per standard deviation (SD) increment in cytokine concentration. Error bars indicate 95% conﬁdence intervals. human BC carcinogenesis48. Further research is required to decipher the biological pathways underpinning these associations. observational studies reporting association between MCP1 level and BC risk. Some conclusions were null, possibly because our study had increased statistical power and previous reports might be false negative29–33. Some studies reported positive correlation between MCP1 level and BC risk, however, the conclusions were still underpowered or possibly biased due to the inherent defect of traditional observational studies34–39. Among the potentially modiﬁable cytokine factors, MIP1b which has the most instru- mental SNPs, contributes to higher BC risk (OR: 1.02; 95% CIs: 1.01–1.04; P value: 2.70 × 10−3), especially in ER-negative BC cases (OR: 1.04; 95% CIs: 1.02–1.06; P value: 7.69 × 10−4). MIP1b plays a central role in the recruitment of regulatory T cells to B cells and antigen-presenting cells, and failure to do this would cause autoimmune activation40. IL13 is another cytokine associated with higher BC risk, especially in ER-negative BC cases (OR: 1.08; 95% CIs: 1.02–1.15; P value: 7.69 × 10−3). Autocrine IL13 plays an important role in the pathophysiology of BC, through inhibiting estrogen-induced proliferation and favoring acquisition of breast cancer cell differentiation markers41. Cytokines are vital to many biological processes, and its serum concentration is tightly regulated. We additionally found suggestive evidence for sig- niﬁcant associations between Eotaxin, GROa, IL12p70, IL16, MIF, SCF, and BC risk. DISCUSSION The results of primary analyses for genetically predicted circulat- ing levels (1-SD increase) of MCP1, MIP1b, and IL13 and risk of BC were presented in Fig. 2. In the primary analyses, genetically raised MCP1 (OR: 1.08; 95% CIs: 1.03–1.12; P value: 3.55 × 10−4), MIP1b (OR: 1.02; 95% CIs: 1.01–1.04; P value: 2.70 × 10−3) and IL13 (OR: 1.06; 95% CIs: 1.03–1.10; P value: 3.33 × 10−4) were associated with increased BC risk. For ER-positive BC cases, genetically raised MCP1 (OR: 1.08; 95% CIs: 1.03–1.13; P value: 2.70 × 10−3), MIP1b (OR: 1.02; 95% CIs: 1.00–1.04; P value: 0.012) and IL13 (OR: 1.05; 95% CIs: 1.01–1.09; P value: 0.024) were consistently associated with increased BC risk. For ER-negative BC cases, genetically raised MIP1b (OR: 1.04; 95% CIs: 1.02–1.06; P value: 7.69 × 10−4) and IL13 (OR: 1.08; 95% CIs: 1.02–1.15; P value: 7.69 × 10−3) were consistently associated with increased BC risk. In alternative analyses (Fig. 3), there was also strong evidence for associations of genetically raised levels of MCP1, MIP1b, and IL13 with BC risk, except for the MR-Egger method, which was mainly used to detect the possible pleiotropy effect. Cochran Q test for the In this two-sample MR analysis with the largest GWAS datasets, we systematically screened 41 cytokines and identiﬁed that geneti- cally predicted circulating levels (1-SD increase) of MCP1, MIP1b, and IL13 were all signiﬁcantly associated with increased risk of overall BC and ER-positive BC. In addition, higher levels of MIP1b and IL13 were also signiﬁcantly associated with increased risk of ER-negative BC. No obvious pleiotropy effect was detected. Besides, we also found suggestive evidence (which didn’t reaching the Bonferroni-corrected threshold) for associations between Eotaxin, GROa, IL12p70, IL16, MIF, SCF, and BC risk. To our knowledge, this study should be the largest and most compre- hensive MR assessment of associations between genetically inﬂammatory cytokines and BC risk to date. Genetically raised MCP1 was associated with increased overall BC risk (OR: 1.08; 95% CIs: 1.03–1.12; P value: 3.55 × 10−4), and ER-positive BC cases (OR: 1.08; 95% CIs: 1.03–1.13; P value: 2.70 × 10−3). Our systematic retrieval only identiﬁed limited Genetically raised MCP1 was associated with increased overall BC risk (OR: 1.08; 95% CIs: 1.03–1.12; P value: 3.55 × 10−4), and ER-positive BC cases (OR: 1.08; 95% CIs: 1.03–1.13; P value: 2.70 × 10−3). Published in partnership with The Hormel Institute, University of Minnesota DISCUSSION Among them, Eotaxin, GROa, IL12p70, IL16, and SCF took part in the regulation of inﬂammatory pathways in BC carcinogenesis, and contributed to higher BC risk42–45. Interest- ingly, genetically raised level of MIF was associated with decreased risk of BC (OR: 0.88; 95% CIs: 0.81–0.95; P value: 1.9 × 10−3), especially in ER-negative BC cases (OR: 0.84; 95% CIs: 0.72–0.97; P value: 0.018). This was contradictory to the results of observational studies, which showed higher MIF level in BC patients46,47. Possible reason might be the dual role of MIF in human BC carcinogenesis . Further research is required to decipher the biological pathways underpinning these associations. The methodological strength should be that we implemented the most recent and comprehensive dataset for cytokine levels and the largest available GWAS dataset for BC risk. The sample size in our analysis affords us greater power to detect causal associations with BC risk and allows us to more accurately estimate effect magnitudes. Indeed all the post-hoc powers for the primary associations of MCP1, MIP1b, and IL13 in overall BC, ER-positive BC and ER-negative BC have reached more than 90.0% at a signiﬁcance level of 0.05. In addition, the post-hoc powers for the suggestive associations of Eotaxin (81.3%), GROa (76.0%), IL12p70 (71.2%%), IL16 (84.2%), MIF (91.6%) has also reached more than 70.0%. The interpretation and generalizability of the study ﬁndings are also affected by several limitations. First should be the limited cytokine instruments. To pursue the homogeneity of research cytokine variables, the instrument selection was based on a single GWAS. Although we implemented the most comprehensive and largest GWAS on cytokines, several cytokines being implicated in BC risk previously, such as CRP, TGF-β, were still missed in current analysis. Second, we didn’t obtain reliable genetic instruments for 17 cytokines as no SNPs meeting the strict ﬁltering procedures. Third, we assumed the sex difference of cytokine was small in the current study, as no sex speciﬁc results were reported. Strictly speaking, the MR analyses in the study were testing the associations between female cytokines and female BCs. Fourth, we didn’t validate our ﬁndings in an independent GWAS, as there were no GWAS with sample sizes that match the BACA study (including 122,977 BC cases). Fifth, collider bias remains an issue in MR analysis. Selection of cytokine instruments y The circulating cytokine levels were predicted using a set of estimated SNP effects based on a GWAS to identify pQTL. The full list of the 41 cytokines was provided in Supplementary Table 225. For each cytokine, the single nucleotide polymorphisms (SNPs) were ﬁltered according to the following procedures: (1) a genome-wide threshold of signiﬁcance (p < 5 × 10-8) and a signiﬁcance threshold of a FDR < 5% were adopted. This caused totally 7262 SNPs were selected. (2) To avoid the pleiotropic effect, 809 SNPs those were associated with levels of more than one cytokine were removed, leaving 6,453 SNPs for further step. (3) then, 5983 SNPs were identiﬁed to be available in the BCAC dataset. (4) We computed F-statistics to quantify the strength of the selected instruments, and all were well above the threshold of F statistics > 10 typically recommended for MR analyses. (5) Linkage disequilibrium (LD; r2 < 0.1 in the European 1000 G reference panel), which retaining SNPs with the lowest p-value as independent instrument, were exerted to remove the superposition effect of correlated SNPs using LDlink50. Finally, we identiﬁed 229 SNPs not in LD and to be signiﬁcantly associated with circulating cytokine levels. These instruments related to the circulating levels of 24 cytokines, whereas for 17 cytokines no SNPs meeting the above conditions (Supplementary table 1). Conclusively, using a two-sample MR approach, we ﬁnd evidence that higher genetically predicted circulating level of MCP1, MIP1b, and IL13 are associated with increased risk of overall BC and ER-positive BC. These ﬁndings suggest the crucial role of cytokines in BC carcinogenesis and potential of targeting speciﬁc inﬂammatory cytokines for BC prevention. Further research is necessary to assess the viability of these cytokine biomarkers as drug targets for BC prevention and treatment. Nonetheless, these ﬁndings have potential implications for changing the inﬂamma- tion status of the general women to reduce the burden of BC and improve female health. Data source For the genetic instruments of the cytokines, the summary statistics were taken from the most comprehensive and largest GWAS for cytokines, which genotyped up to 8293 Finnish participants from three independent population cohorts: the Cardiovascular Risk in Young Finns Study, FINRISK1997, and FINRISK200225. For BC, we obtained the data from the largest BC consortia—BCAC, which provided summary association statistics for overall (105,974 controls and 122,977 BC cases), ER-positive (69,501 BC cases) and ER-negative BC (21,468 BC cases)27. The consortia results included the OncoArray (45,494 controls and 61,282 BC cases) and iCOGS Statistical analyses Published in partnership with The Hormel Institute, University of Minnesota DISCUSSION Sixth, we still can’t remove the possible pleiotropy effect (vertical or horizontal pleiotropy) that might be concealed by small sample size or the Published in partnership with The Hormel Institute, University of Minnesota npj Precision Oncology (2020) 25 S Li et al. S Li et al. 4 Fig. 3 Alternative analyses (penalized IVW, robust IVW, penalized robust IVW, MR-Egger, simple median, weighted median, and penalized weighted median method) for causal estimates of selected cytokines (MCP1, MIP1b and IL13) on risk of breast cancer. Causal estimates express the change in odds ratio (OR) per standard deviation (SD) increment in cytokine concentration. Error bars indicate 95% conﬁdence intervals. Fig. 3 Alternative analyses (penalized IVW, robust IVW, penalized robust IVW, MR-Egger, simple median, weighted median, and penalized weighted median method) for causal estimates of selected cytokines (MCP1, MIP1b and IL13) on risk of breast cancer. Causal estimates express the change in odds ratio (OR) per standard deviation (SD) increment in cytokine concentration. Error bars indicate 95% conﬁdence intervals. datasets (42,892 controls and 46,785 BC cases), and the combined results from eleven additional GWAS (17,588 controls and 14,910 BC cases). To minimize the ancestral mismatch, current analyses are restricted to women of European ancestry only. All studies included rigorous quality control, imputation to the 1000 Genomes Project panel. As all analyses were based on summary statistics and not individual-level data, no ethical approval was required. small number of SNP instruments, although we have used a series of methods, including MR-Egger, MR-PRESSO analyses, and checking whether the SNPs used as instruments have any known pleiotropic effects in curated genotype to phenotype databases. However, its impact is likely to be less than other biases, and it could be substantial only when the effects of the risk factor and confounders on selection are particularly large49. Sixth, the results from the MR-Egger analyses for MCP1, MIP1b, and IL13 were insigniﬁcant, which indicates the possibility of dependent on exclusion restriction. REFERENCES 29. Shen, J. et al. Metabolic hormones and breast cancer risk among Mexican American Women in the Mano a Mano Cohort Study. Sci. Rep. 9, 9989 (2019). 1. DeSantis, C. E. et al. Breast cancer statistics, 2019. CA Cancer J. Clin. 69, 438–451 (2019). 1. DeSantis, C. E. et al. Breast cancer statistics, 2019. CA Cancer J. Clin. 69, 438–451 (2019). 30. Chen, C. et al. Circulating galectins -2, -4 and -8 in cancer patients make important contributions to the increased circulation of several cytokines and chemokines that promote angiogenesis and metastasis. Br. J. Cancer 110, 741–752 (2014). 2. Global Burden of Disease Cancer, C. et al. Global, Regional, and National Cancer Incidence, Mortality, Years of Life Lost, Years Lived With Disability, and Disability- Adjusted Life-Years for 29 Cancer Groups, 1990 to 2017: A Systematic Analysis for 2. Global Burden of Disease Cancer, C. et al. Global, Regional, and National Cancer Incidence, Mortality, Years of Life Lost, Years Lived With Disability, and Disability- Adjusted Life-Years for 29 Cancer Groups, 1990 to 2017: A Systematic Analysis for the Global Burden of Disease Study. JAMA Oncol. 5, 1749–1768 (2019). the Global Burden of Disease Study. JAMA Oncol. 5, 1749–1768 31. Wang, J., He, Q., Shao, Y. G. & Ji, M. Chemokines ﬂuctuate in the progression of primary breast cancer. Eur. Rev. Med. Pharmacol. Sci. 17, 596–608 (2013). 3. Diakos, C. I., Charles, K. A., McMillan, D. C. & Clarke, S. J. Cancer-related inﬂam- mation and treatment effectiveness. Lancet Oncol. 15, e493–e503 (2014). 32. Touvier, M. et al. Association between prediagnostic biomarkers of inﬂammation and endothelial function and cancer risk: a nested case-control study. Am. J. Epidemiol. 177, 3–13 (2013). 4. Wellenstein, M. D. et al. Loss of p53 triggers WNT-dependent systemic inﬂam- mation to drive breast cancer metastasis. Nature 572, 538–542 (2019). 5. Fillon, M. Changes in inﬂammation and insulin pathways mediate the association between resistance training and breast cancer survival. CA Cancer J. Clin. 68, 175–177 (2018). 33. Potter, S. M. et al. Systemic chemokine levels in breast cancer patients and their relationship with circulating menstrual hormones. Breast Cancer Res. Treat. 115, 279–287 (2009). 6. Cole, S. W. Chronic inﬂammation and breast cancer recurrence. J. Clin. Oncol. 27, 3418–3419 (2009). 34. Cohen, E. N. et al. Elevated serum levels of sialyl Lewis X (sLe(X)) and inﬂam- matory mediators in patients with breast cancer. Breast Cancer Res. Treat. 176, 545–556 (2019). 7. S Li et al. 5 effect sizes were in SD-scaled units in Ahola-Olli’s report25. Causal estimates are thus presented per genetically predicted SD, and a log- linear association with odds of BC is implicit across the range of intermediate risk exposure. In current study, inverse-variance weighted (IVW) method was used as the primary method for the MR analyses to test the potential causal associations between circulating levels of selected cytokines and BC risk. In the absence of directional pleiotropy, it provides robust causal estimates. Further, penalized IVW, robust IVW, penalized robust IVW, MR-Egger, simple median, weighted median, and penalized weighted median method were used for alternative analyses. Among these methods, MR-Egger allows free estimation of the intercept, and a statistically signiﬁcant intercept term implies the presence of unbalanced pleiotropy and causal estimates in MR Egger are less precise than those in IVW. Meanwhile, we also conducted bidirectional MR to test the potential causal effect of BC (Supplementary Table 3 presents SNPs that were used as instruments for breast cancer) on cytokine levels. To detect the possible pleiotropy effect, we ﬁrst tested whether the intercept from MR-Egger regression differed from zero, which provided evidence of directional pleiotropy. Second, the MR-Pleiotropy Residual Sum and Outlier (MR- PRESSO) was used to identify and correct for potential outliers52. Third, we looked up the MR-Base, PhenoScanner database and the GWAS catalog for potential associations of the selected variants in our study with other BC risk factors. Post-hoc power calculations were conducted for our IVW analyses using an online Mendelian randomization power calculation tool (https://sb452.shinyapps.io/power/)53. All statistical analyses were con- ducted using the R (version 3.6.3) or SAS 9.4 statistical software. All P values are two-tailed, and P < 0.05 was considered signiﬁcant. 12. Baharlou, R. et al. Reduced levels of T-helper 17-associated cytokines in the serum of patients with breast cancer: indicators for following the course of disease. Cent. Eur. J. Immunol. 41, 78–85 (2016). 13. Ivanovic, V. et al. Elevated plasma levels of transforming growth factor-beta 1 (TGF-beta 1) in patients with advanced breast cancer: association with disease progression. Eur. J. Cancer 39, 454–461 (2003). 14. Smith, G. D. & Ebrahim, S. ‘Mendelian randomization’: can genetic epidemiology contribute to understanding environmental determinants of disease? Int. J. Epi- demiol. 32, 1–22 (2003). 15. Zhang, S. M. et al. C-reactive protein and risk of breast cancer. J. Natl Cancer Inst. 99, 890–894 (2007). 16. Wang, H. Reporting summary 23. Porcu, E. et al. Mendelian randomization integrating GWAS and eQTL data reveals genetic determinants of complex and clinical traits. Nat. Commun. 10, 3300 (2019). Further information on research design is available in the Nature Research Reporting Summary linked to this article. 24. Georgakis, M. K. et al. Genetically determined levels of circulating cytokines and risk of stroke. Circulation 139, 256–268 (2019). DATA AVAILABILITY 25. Ahola-Olli, A. V. et al. Genome-wide Association Study Identiﬁes 27 loci inﬂuen- cing concentrations of circulating cytokines and growth factors. Am. J. Hum. Genet. 100, 40–50 (2017). All summary genetic association data used in this study are available online, Cytokine GWAS and BCAC. 26. Malik, R. et al. Multiancestry genome-wide association study of 520,000 subjects identiﬁes 32 loci associated with stroke and stroke subtypes. Nat. Genet. 50, 524–537 (2018). Received: 6 February 2020; Accepted: 4 August 2020; Received: 6 February 2020; Accepted: 4 August 2020; 27. Michailidou, K. et al. Association analysis identiﬁes 65 new breast cancer risk loci. Nature 551, 92–94 (2017). 28. Berry, D. A. et al. Estrogen-receptor status and outcomes of modern che- motherapy for patients with node-positive breast cancer. JAMA 295, 1658–1667 (2006). & Yang, X. Association between serum cytokines and progression of breast cancer in Chinese population. Medicine 96, e8840 (2017). breast cancer in Chinese population. Medicine 96, e8840 (2017 17. Ma, Y. et al. IL-6, IL-8 and TNF-alpha levels correlate with disease stage in breast cancer patients. Adv. Clin. Exp. Med. 26, 421–426 (2017). cancer patients. Adv. Clin. Exp. Med. 26, 421–426 (2017). 18. Heikkila, K. et al. Associations of circulating C-reactive protein and interleukin-6 with cancer risk: ﬁndings from two prospective cohorts and a meta-analysis. Cancer Causes Control 20, 15–26 (2009). 19. Yoon, Y. S., Kwon, A. R., Lee, Y. K. & Oh, S. W. Circulating adipokines and risk of obesity related cancers: a systematic review and meta-analysis. Obes. Res. Clin. Pract. 13, 329–339 (2019). 20. Emdin, C. A., Khera, A. V. & Kathiresan, S. Mendelian randomization. JAMA 318, 1925–1926 (2017). 21. Lawlor, D. A., Harbord, R. M., Sterne, J. A., Timpson, N. & Davey Smith, G. Men- delian randomization: using genes as instruments for making causal inferences in epidemiology. Stat. Med. 27, 1133–1163 (2008). 22. Zhao, Q., Chen, Y., Wang, J. & Small, D. S. Powerful three-sample genome-wide design and robust statistical inference in summary-data Mendelian randomiza- tion. Int. J. Epidemiol. 48, 1478–1492 (2019). Published in partnership with The Hormel Institute, University of Minnesota Statistical analyses In current study, R2 representing the proportion of variance in a risk factor explained by the genetic instrument, and F-statistic representing the strength of the association between the genetic instrument and levels of the risk factor were calculated51. For cytokine exposures, we scaled MR estimates per standard deviation (SD) difference of the risk factor, as all Published in partnership with The Hormel Institute, University of Minnesota Published in partnership with The Hormel Institute, University of Minnesota npj Precision Oncology (2020) 25 Correspondence and requests for materials should be addressed to X.M. 48. Verjans, E. et al. Dual role of macrophage migration inhibitory factor (MIF) in human breast cancer. BMC Cancer 9, 230 (2009). Reprints and permission information is available at http://www.nature.com/ reprints 49. Gkatzionis, A. & Burgess, S. Contextualizing selection bias in Mendelian rando- mization: how bad is it likely to be? Int J. Epidemiol. 48, 691–701 (2019). Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afﬁliations. 50. Machiela, M. J. & Chanock, S. J. LDlink: a web-based application for exploring population-speciﬁc haplotype structure and linking correlated alleles of possible functional variants. Bioinformatics 31, 3555–3557 (2015). 51. Burgess, S., Thompson, S. G. & Collaboration, C. C. G. Avoiding bias from weak instruments in Mendelian randomization studies. Int. J. Epidemiol. 40, 755–764 (2011). 52. Verbanck, M., Chen, C. Y., Neale, B. & Do, R. Detection of widespread horizontal pleiotropy in causal relationships inferred from Mendelian randomization between complex traits and diseases. Nat. Genet. 50, 693–698 (2018). Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons. org/licenses/by/4.0/. 53. Burgess, S. Sample size and power calculations in Mendelian randomization with a single instrumental variable and a binary outcome. Int. J. Epidemiol. 43, 922–929 (2014). REFERENCES Pierce, B. L. et al. Elevated biomarkers of inﬂammation are associated with reduced survival among breast cancer patients. J. Clin. Oncol. 27, 3437–3444 (2009). 35. Lubowicka, E. et al. Plasma chemokine CCL2 and its receptor CCR2 concentra- tions as diagnostic biomarkers for breast cancer patients. BioMed. Res. Int. 2018, 2124390 (2018). 8. Mantovani, A., Dinarello, C. A., Molgora, M. & Garlanda, C. Interleukin-1 and related cytokines in the regulation of inﬂammation and immunity. Immunity 50, 778–795 (2019). 36. Comen, E. et al. TNF is a key cytokine mediating neutrophil cytotoxic activity in breast cancer patients. NPJ Breast Cancer 2, 16009 (2016). 9. Ouyang, W., Rutz, S., Crellin, N. K., Valdez, P. A. & Hymowitz, S. G. Regulation and functions of the IL-10 family of cytokines in inﬂammation and disease. Annu. Rev. Immunol. 29, 71–109 (2011). 37. Narita, D. et al. Altered levels of plasma chemokines in breast cancer and their association with clinical and pathological characteristics. Neoplasma 63, 141–149 (2016). 10. Gunter, M. J. et al. Circulating adipokines and inﬂammatory markers and post- menopausal breast cancer risk. J. Natl. Cancer Inst. 107, djv169 (2015). 38. Dehqanzada, Z. A. et al. Assessing serum cytokine proﬁles in breast cancer patients receiving a HER2/neu vaccine using Luminex technology. Oncol. Rep. 17, 687–694 (2007). 11. Agnoli, C. et al. Biomarkers of inﬂammation and breast cancer risk: a case-control study nested in the EPIC-Varese cohort. Sci. Rep. 7, 12708 (2017). npj Precision Oncology (2020) 25 Published in partnership with The Hormel Institute, University of Minnesota S Li et al. 6 CA148065, X01HG007492), Cancer Research UK (C1287/A10118, C1287/A16563, C1287/A10710) and The European Union (HEALTH-F2-2009-223175 and H2020 633784 and 634935). All studies and funders are listed in Michailidou et al.27. 39. Lebrecht, A. et al. Monocyte chemoattractant protein-1 serum levels in patients with breast cancer. Tumour Biol. 25, 14–17 (2004). 40. Bystry, R. S., Aluvihare, V., Welch, K. A., Kallikourdis, M. & Betz, A. G. B cells and professional APCs recruit regulatory T cells via CCL4. Nat. Immunol. 2, 1126–1132 (2001). AUTHOR CONTRIBUTIONS 41. Aspord, C. et al. Breast cancer instructs dendritic cells to prime interleukin 13- secreting CD4+ T cells that facilitate tumor development. J. Exp. Med. 204, 1037–1047 (2007). Research conception and design: M.X., L.S., X.Y., and Z.Y. contribute equally to this manuscript, and should be considered as “co-ﬁrst author”. Analyzed the data: M.X., L.S., X.Y., and Z.Y. Contributed to the writing of the manuscript: M.X., L.S., X.Y., and Z.Y. Contributed to the collection of data: L.S., X.Y., Z.Y., N.L., M.Z., M.L., and F.X. All authors have read, and conﬁrm that they meet, ICMJE criteria for authorship. 42. Kawaguchi, K. et al. Alteration of speciﬁc cytokine expression patterns in patients with breast cancer. Sci. Rep. 9, 2924 (2019). 43. Wolfson, E., Solomon, S., Schmukler, E., Goldshmit, Y. & Pinkas-Kramarski, R. Nucleolin and ErbB2 inhibition reduces tumorigenicity of ErbB2-positive breast cancer. Cell Death Dis. 9, 47 (2018). COMPETING INTERESTS 44. Perez-Rivas, L. G. et al. Serum protein levels following surgery in breast cancer patients: a protein microarray approach. Int. J. Oncol. 41, 2200–2206 (2012). The authors declare no competing interests. 45. Yan, R. et al. SCF(JFK) is a bona ﬁde E3 ligase for ING4 and a potent promoter of the angiogenesis and metastasis of breast cancer. Genes Dev. 29, 672–685 (2015). 46. Fersching, D. M. et al. Apoptosis-related biomarkers sFAS, MIF, ICAM-1 and PAI-1 in serum of breast cancer patients undergoing neoadjuvant chemotherapy. Anticancer Res. 32, 2047–2058 (2012). ADDITIONAL INFORMATION Supplementary information is available for this paper at https://doi.org/10.1038/ s41698-020-00131-6. 47. Avalos-Navarro, G. et al. Circulating soluble levels of MIF in women with breast cancer in the molecular subtypes: relationship with Th17 cytokine proﬁle. Clin. Exp. Med. 19, 385–391 (2019). Correspondence and requests for materials should be addressed to X.M. Correspondence and requests for materials should be addressed to X.M. Published in partnership with The Hormel Institute, University of Minnesota © The Author(s) 2020 ACKNOWLEDGEMENTS This study were supported by a grant from Science foundation for outstanding young investigator of Third Military medical University (Grant to Dr Xiangyu Ma). The breast cancer genome-wide association analyses were supported by the Government of Canada through Genome Canada and the Canadian Institutes of Health Research, the ‘Ministère de l’Économie, de la Science et de l’Innovation du Québec’ through Genome Québec and grant PSR-SIIRI-701, The National Institutes of Health (U19 © The Author(s) 2020 Published in partnership with The Hormel Institute, University of Minnesota Published in partnership with The Hormel Institute, University of Minnesota npj Precision Oncology (2020) 25
https://openalex.org/W4300160117	https://europepmc.org/articles/pmc5932083?pdf=render	English	null	Spillover modes in multiplex games: double-edged effects on cooperation, and their coevolution	arXiv (Cornell University)	2,017	cc-by	8,354	www.nature.com/scientificreports www.nature.com/scientificreports www.nature.com/scientificreports Spillover modes in multiplex games: double-edged effects on cooperation and their coevolution Received: 26 January 2018 Accepted: 11 April 2018 Published: xx xx xxxx Tommy Khoo1, Feng Fu 1,2 & Scott Pauls1 In recent years, there has been growing interest in studying games on multiplex networks that account for interactions across linked social contexts. However, little is known about how potential cross- context interference, or spillover, of individual behavioural strategy impact overall cooperation. We consider three plausible spillover modes, quantifying and comparing their effects on the evolution of cooperation. In our model, social interactions take place on two network layers: repeated interactions with close neighbours in a lattice, and one-shot interactions with random individuals. Spillover can occur during the learning process with accidental cross-layer strategy transfer, or during social interactions with errors in implementation. Our analytical results, using extended pair approximation, are in good agreement with extensive simulations. We find double-edged effects of spillover: increasing the intensity of spillover can promote cooperation provided cooperation is favoured in one layer, but too much spillover is detrimental. We also discover a bistability phenomenon: spillover hinders or promotes cooperation depending on initial frequencies of cooperation in each layer. Furthermore, comparing strategy combinations emerging in each spillover mode provides good indication of their co-evolutionary dynamics with cooperation. Our results make testable predictions that inspire future research, and sheds light on human cooperation across social domains. The ubiquity of cooperation in human societies and nature is a puzzling phenomenon1–4. At first glance, cooper- ation seems unlikely: cooperators incur cost in providing benefits to others, while opportunistic individuals can reap rewards without returning the favour5. Nonetheless, cooperation can arise in structured populations through the mechanism of network reciprocity6–8. This basic observation drove deeper investigations into reciprocity in structured populations9–12.t Real world networks are often interdependent, where a small perturbation to one network can trigger a chain of events that results in cataclysmic effects on both networks13. Taking the importance of such interconnectedness into account has led to a recent boom in the study of multiplex networks14,15. In the same vein, evolutionary games on multiplex networks are attracting increasing attention (we refer readers to16 for a review).f Various mechanisms have been proposed to associate different evolutionary games taking place on otherwise disjoint networks. One utility function approach incorporates payoffs accumulated across games on different networks into each strategic decision17,18. An alternative approach allows strategic behaviour to be transmitted from one setting to another through peer influence and social learning19,20. 1Department of Mathematics, Dartmouth College, Hanover, NH, 03755, USA. 2Department of Biomedical Data Science, Geisel School of Medicine, Dartmouth College, Hanover, NH, 03755, USA. Correspondence and requests for materials should be addressed to T.K. (email: tommy.za.khoo@gmail.com) Received: 26 January 2018 Accepted: 11 April 2018 Published: xx xx xxxx Scientific RePorTS \| (2018) 8:6922 \| DOI:10.1038/s41598-018-25025-3 Spillover modes in multiplex games: double-edged effects on cooperation and their coevolution g g pl g In this paper, we draw inspiration from empirical results21–25, which suggest that norms and heuristics culti- vated during repeated interactions could “spill over” to affect decision making in one-shot situations. 21 24 26 In two experiments21,24, when subjects first participate in the iterated prisoner’s dilemma (IPD)26, they observed greater cooperation and prosocial behaviour in subsequent one-shot games. A similar increase in proso- cial behaviour follows a repeated public goods game with conditions favourable for cooperation25. Relatedly, cooperation levels rose when switching from an IPD with a large continuation probability to one with a small continuation probability22, as well as when switching from playing an IPD with a fixed partner to playing an IPD in which every iteration was played with a random partner23. y y The hypothesized explanation21,27,28 for these phenomena is that repeated interactions foster coopera- tive heuristics or norms in participants which then affected subsequent one-shot games4,29. Although recent 1Department of Mathematics, Dartmouth College, Hanover, NH, 03755, USA. 2Department of Biomedical Data Science, Geisel School of Medicine, Dartmouth College, Hanover, NH, 03755, USA. Correspondence and requests for materials should be addressed to T.K. (email: tommy.za.khoo@gmail.com) Scientific RePorTS \| (2018) 8:6922 \| DOI:10.1038/s41598-018-25025-3 1 www.nature.com/scientificreports/ Figure 1. Spillover modes in multiplex games. In (A) On the bottom layer, individuals play the iterated prisoner’s dilemma on a square lattice with periodic boundary. On the top layer, they play the one shot prisoner’s dilemma on a random regular network with degree four. Individuals compare payoffs with neighbours within the layer, to decide whether to adopt a neighbour’s strategy in that layer. In addition, there is a probability of using one of the three spillover modes instead, to adopt strategies from another layer. For (B) neighbour imitation mode, the individual compares payoff with neighbours on the same layer, and applies adopted strategies to the opposite layer. In (C) self comparison mode, the individual compares her normalized payoffs on each layer, and decides whether to implement her strategy from one layer in another. For (D) context interference mode, the individual is susceptible to temporary interference from mistaking the context of the interaction, which results in temporarily using strategies from the opposite layer. Figure 1. Spillover modes in multiplex games. In (A) On the bottom layer, individuals play the iterated prisoner’s dilemma on a square lattice with periodic boundary. Spillover modes in multiplex games: double-edged effects on cooperation and their coevolution On the top layer, they play the one shot prisoner’s dilemma on a random regular network with degree four. Individuals compare payoffs with neighbours within the layer, to decide whether to adopt a neighbour’s strategy in that layer. In addition, there is a probability of using one of the three spillover modes instead, to adopt strategies from another layer. For (B) neighbour imitation mode, the individual compares payoff with neighbours on the same layer, and applies adopted strategies to the opposite layer. In (C) self comparison mode, the individual compares her normalized payoffs on each layer, and decides whether to implement her strategy from one layer in another. For (D) context interference mode, the individual is susceptible to temporary interference from mistaking the context of the interaction, which results in temporarily using strategies from the opposite layer. endeavours27,30,31 studied spillover using an evolutionary framework, systematic exploration of spillover mech- anisms through the lens of games on multiplex networks are still lacking, with the exception of a recent work32.i endeavours27,30,31 studied spillover using an evolutionary framework, systematic exploration of spillover mech- anisms through the lens of games on multiplex networks are still lacking, with the exception of a recent work32.i We fill this gap by modelling spillover as strategy interference between layers on a multiplex network (see Fig. 1A). Instead of the unstructured populations seen in the experiments, our n individuals participate in two layers of interactions with population structure that emulate recurring close proximity and distant one-shot contacts. We represent recurring close proximity contacts using a square periodic lattice on the bottom layer where agents play m rounds of the IPD with their neighbours using two possible strategies C or D. Here, C refers to tit-for-tat (TFT), where individuals cooperates on the first iteration and play the opponent’s previous strategy for future iterations, and D stands for always defecting (ALLD) during every iteration. While other strategies are possible, we focus on these two classic strategies33,34 for simplicity. If two players play C, they both receive a payoff of m(b − c). If they both play D, they both receive 0. Scientific RePorTS \| (2018) 8:6922 \| DOI:10.1038/s41598-018-25025-3 www.nature.com/scientificreports/ www.nature.com/scientificreports/ (C) or defect (D). Two cooperators will both receive a payoff of b − c, while two defectors both receives 0. If one cooperates and the other defects, the cooperator gets a payoff of −c while the defector gets b.i p p g p yf g We propose three modes for spillover. In the first, which we call neighbour imitation spillover (NIS) mode Fig. 1B, individuals on one layer may imitate the strategic behaviour (C or D) of a neighbour on the opposite layer. In the second, self comparison spillover (SCS) mode Fig. 1C, individuals compare their payoffs between layers and learn from their experience. Finally, for the third context interference spillover (CIS) mode, illustrated in Fig. 1D, individuals may make a temporary mistake and apply their strategic behaviour from another layer. A parameter p determines the frequency of spillover occurrences in each case, and hence is a proxy for the strength of the spillover effect.f pf Although the strategy sets differ in the IPD and PD, we allow cooperative (C) and defective (D) behaviour to be transmitted through spillover in order to model experimental observations. Here, we do not account for the fine detail interactions between games32 but instead focus on the coarse grain interface of behavioural norms and social heuristics acquired through two different social contexts. Furthermore, we model repeated interactions with close proximity contacts as occurring with greater frequency than distant one-shot contacts, and hence we compare the m rounds IPD to the one round PD. Finally, we only use intra-layer payoffs in this part of our inves- tigation in order to keep the two social contexts distinct, as well as to restrict cross context interference to occur only through our three spillover modes.h y g p Through these three modes, we amalgamate key ideas from prior work. NIS and SCS encapsulates the notion of individuals making a mistake in learning a potentially suboptimal strategy from a different social setting19,20, either through their own experience or by interacting with others. CIS captures the idea of individuals making implementation mistakes due to confounding two different social settings27,30,31. Results We start our exploration by examining how the strength of the spillover effect impacts cooperation. To accom- plish this, we produced simulation and pair approximation results for a range of parameter combinations. We present these in Fig. 2, with individual plots for each mode (NIS Fig. 2A, SCS Fig. 2B and CIS Fig. 2C).h The most striking feature of Fig. 2 is the existence of an optimal value of p which maximizes the average coop- eration level in the multiplex network. Average cooperation initially increases with p, before reaching the opti- mal value and subsequently plummeting. This demonstrates a double-edged effect of spillover: a little spillover between the two social settings allows cooperators on the repeated local interaction layer to exert their influence on the distant one-shot contact layer and provides a boost to overall cooperation. On the other hand, a spillover effect that is too strong leads to too much influence by defectors on the distant one-shot contacts layer, and is deleterious to cooperation. p We note that conditions that are overly favourable or hostile to cooperation will lead to one layer overwhelm- ing the other and consequently a rapid monotonic rise or decline in cooperation, instead of an intermediate optimal value of p (see Supplementary Fig. S1). Similarly, when both layers are analysed with the same game and parameters (both IPD or both PD), the overall conditions are too hostile for cooperation to thrive in the presence of spillover, and average cooperation in the multiplex decline with any increase in p (see Supplementary Figs S2 and S3).f Another feature of Fig. 2 is that CIS appears to be more resilient to this double-edge effect than NIS and SCS as cooperation levels for CIS tend to be higher than the other two. This is due to the fact that while implementation mistakes occur under CIS, payoff comparison always occurs on the layer on which the strategies are adopted. Hence, individuals have an easier time learning correct strategies, and implementation mistakes need not have a prolonged impact. On the other hand, under NIS and SCS, individuals directly learn and adopt strategies across layers, resulting in mistakes that have larger, long term repercussions on cooperation.i Next, we examine fine details for several parameter combinations from Fig. 2A–C highlighted with squares. These microscopic details (Fig. 2D–F) show the proportions of individuals for each possible strategy combination for the top and bottom layer. www.nature.com/scientificreports/ We also note that our spillover mechanisms model mistakes occurring between distinct network games, as opposed to random errors modelled by mechanisms such as weak selection and mutation.i y Our main finding is that cooperation depends subtly on the strength p and the initial level of cooperation on both the layers - a double-edged effect where different combinations encourage or discourage cooperation. These results expanded upon previous work containing an alternative formulation of NIS20, which used two one-shot games to investigate neighbour imitation and found that there is an intermediate optimal frequency for cooper- ation in one of the games, in the case of a well-mixed population. We also advance their results for unstructured populations by using pair approximation36 to derive analytical solutions that incorporate population structure, and demonstrate the effectiveness of our solutions with extensive simulations. These solutions allow us to study both the macroscopic overall cooperation level, as well as microscopic details regarding strategy combinations. p p p g g gy Finally, we analyse the three spillover modes as they coevolve with cooperation in the presence of mutation. Our findings suggest that transient implementation mistakes (CIS) outperforms mechanisms under which indi- viduals might learn and retain suboptimal strategies (NIS, SCS). On the other hand, when we allow repeated local interactions to play a larger role during spillover conditions become more conducive to cooperation, making it less punishing to make learning mistakes with long lasting impact. This effect allows more deliberate mechanisms that promotes cooperation on both layers, like NIS and SCS, to thrive. Scientific RePorTS \| (2018) 8:6922 \| DOI:10.1038/s41598-018-25025-3 Spillover modes in multiplex games: double-edged effects on cooperation and their coevolution If one plays C and the other plays D, the C player gets a payoff of −c while the D player gets b.f p y g On the top layer, the same agents, randomly connected to four different agents during every run, play the (one-shot) prisoner’s dilemma35 (PD) as a proxy for distant one-shot contacts. Here, agents choose to cooperate Scientific RePorTS \| (2018) 8:6922 \| DOI:10.1038/s41598-018-25025-3 2 www.nature.com/scientificreports/ Figure 2. Double-edged effects of spillover. (A–C) Shows simulation and pair approximation results for average proportion of cooperators for each spillover mode (A) neighbour imitation, (B) self comparison and (C) context interference. As shown in more details by (D–F), analytical and simulation results are in good agreement at low frequencies of spillover p. For some parameter combinations, a double-edge effect of spillover on cooperation reveals itself in the form of an initial increase in cooperation with p, before a subsequent decline after an optimal p. (D–F) presents microscopic details of strategy profile proportions associated with each spillover mode, taken from (A–C) at parameter combinations ( ) . Self comparison has the largest proportion of individuals cooperating on both layers (green line), while context interference produces the largest proportion of individuals cooperating on the bottom layer and defecting on the top layer (yellow line). Parameters: n = 3600, m = 4, β = 0.2, b = 1, α = 0.5. Simulations: 6 × 106 time steps, averaged over 100 runs. p has step size 0.01 from p = 0 to 0.4, and step size 0.1 otherwise. Figure 2. Double-edged effects of spillover. (A–C) Shows simulation and pair approximation results for average proportion of cooperators for each spillover mode (A) neighbour imitation, (B) self comparison and (C) context interference. As shown in more details by (D–F), analytical and simulation results are in good agreement at low frequencies of spillover p. For some parameter combinations, a double-edge effect of spillover on cooperation reveals itself in the form of an initial increase in cooperation with p, before a subsequent decline after an optimal p. (D–F) presents microscopic details of strategy profile proportions associated with each spillover mode, taken from (A–C) at parameter combinations ( ) . Self comparison has the largest proportion of individuals cooperating on both layers (green line), while context interference produces the largest proportion of individuals cooperating on the bottom layer and defecting on the top layer (yellow line). Parameters: n = 3600, m = 4, β = 0.2, b = 1, α = 0.5. Simulations: 6 × 106 time steps, averaged over 100 runs. p has step size 0.01 from p = 0 to 0.4, and step size 0.1 otherwise. and almost all D on the top layer. www.nature.com/scientificreports/ As we saw above, in CIS, players learn strategies more easily compared to NIS and SCS, so individuals can learn the optimal strategy of playing D on the bottom layer and playing C on the top layer. y On the other hand, SCS (Fig. 2E) stands out as having the highest proportion of individuals who are coopera- tors on both layers, with NIS (Fig. 2D) coming in a close second. In both cases, individuals are adopting spillover strategies that their neighbours or themselves have been successfully using within the opposite layer. This leads to a higher level of cooperation in the one-shot PD layer due to individuals learning and retaining the suboptimal strategy of playing C on that layer.f gy p y g y Figure 3 further illuminates the differences in microscopic details between the three spillover modes for a subset of the parameter combinations in Fig. 2. Figure 3A–C shows SCS having the highest Xc c, while Fig. 3D–F shows CIS having the highest Xc d. We get a clearer view of how NIS differs from the rest, with higher Xd d and Xd c (Fig. 3G–L). Supplementary Figs S4 and S5 extends Fig. 3 for a range of p from 0 to 0.40. In our previous results, we have initialised individual strategies C or D on both layers uniformly at random. But how will spillover behave when the initial probability of being a cooperator varies on each layer? We address this question by exhaustively exploring the parameter space using pair approximation as illustrated by Fig. 4 (NIS Fig. 4A, SCS Fig. 4B, CIS Fig. 4C).i g g g Here, we discover a bistability phenomenon. For a fixed frequency of spillover p, the parameter space is par- titioned into two distinct regions. Depending on the initial proportion of cooperators on each network layer, the spillover effect can either help or hinder cooperation, as shown in more detailed plots by both simulation and pair approximation (Fig. 4D–I). These detailed plots are also reproduced with p = 0 in Supplementary Fig. S6 in order to highlight the impact of spillover. This bistability phenomenon has potential social policy implications: if the proportion of cooperators in one setting can be actively raised to a sufficient level, spillover can promote overall levels of cooperation. p As shown in Fig. Scientific RePorTS \| (2018) 8:6922 \| DOI:10.1038/s41598-018-25025-3 Results For simplicity, we let Xb a be the proportion of individuals playing strategy a on the top layer and b on the bottom layer. Figure 2D–F demonstrates that for small values of p, there is excellent agreement between simulation and our pair approximation results at even the microscopic level (detailed equations are presented in the SI). The most outstanding feature is that context interference spillover mode has much higher Xc d than the other modes, as seen in Fig. 2F. This happens because, the parameter combinations in Fig. 2 leads to almost all C on the bottom layer Scientific RePorTS \| (2018) 8:6922 \| DOI:10.1038/s41598-018-25025-3 3 www.nature.com/scientificreports/ www.nature.com/scientificreports/ 4, the number of initial cooperators on the bottom layer has a larger impact on whether spillover hinders or helps cooperation than the number of initial cooperators on the top layer. An example of this Scientific RePorTS \| (2018) 8:6922 \| DOI:10.1038/s41598-018-25025-3 4 www.nature.com/scientificreports/ Figure 3. Microscopic characteristics of spillover modes. Figure compares proportion of individual top-b strategy combinations of each spillover mode for the parameters in Fig. 2, with different scales on the verti axes. Each spillover mode differs in their microscopic characteristics. In (A–C), self comparison has the hi proportion of individuals cooperating on both layers, Xc c. In (D–F), context interference has the highest proportion with the payoff maximizing strategy profile Xc d. While in (G–L), neighbour imitation has the la Xd d and Xd c proportions. Parameters: n = 3600, m = 4, b = 1, β = 0.2, α = 0.5. Simulations: 6 × 106 time st averaged over 100 runs. c is the cost of cooperation. Figure 3. Microscopic characteristics of spillover modes. Figure compares proportion of individual top-bottom strategy combinations of each spillover mode for the parameters in Fig. 2, with different scales on the vertical axes. Each spillover mode differs in their microscopic characteristics. In (A–C), self comparison has the highest proportion of individuals cooperating on both layers, Xc c. In (D–F), context interference has the highest proportion with the payoff maximizing strategy profile Xc d. While in (G–L), neighbour imitation has the largest Xd d and Xd c proportions. Parameters: n = 3600, m = 4, b = 1, β = 0.2, α = 0.5. Simulations: 6 × 106 time steps, averaged over 100 runs. c is the cost of cooperation. can be seen in Fig. 4A, where at p = 0.3, when the number of initial cooperators on the top layer is close to zero, sufficient number of initial cooperators on the bottom layer can still result in a spillover effect that helps coopera- tion. In contrast, for close to zero initial cooperators on the bottom layer, no amount of initial cooperators on the top layer will result in a beneficial effect of spillover on cooperation. if So far, we have studied these three spillover modes separately. Next, we will compare them when all three modes are present and are potentially competing with each other. www.nature.com/scientificreports/ We initialise each individual with a spillover mode uniformly at random and allow the modes to coevolve with strategy while at the same time introducing mutation. Here, the imitation of spillover modes is decided using the sum of payoffs across both layers in order to isolate the effect of natural selection on spillover modes, and to avoid conflating natural selection with cross context interference. Figure 5 shows our results for various p. g p Each bar in Fig. 5A indicates the proportion of individuals with each of the three spillover mode. For all of the choices of p in Fig. 5A, the highest is CIS, followed by SCS and then NIS, with this trend becoming more prominent as p increases. We explain the dominance of CIS for this parameter combination by the crucial role of Scientific RePorTS \| (2018) 8:6922 \| DOI:10.1038/s41598-018-25025-3 5 www.nature.com/scientificreports/ Figure 4. Bistability phenomenon. Spillover can hinder or promote cooperation depe conditions. (A–C) illustrates, for each frequency of spillover p, the initial level of coope in zero cooperation at equilibrium, under each spillover mode. In all cases, for a fixed p of initial cooperators on each network layer beyond which the spillover effect switches cooperation. The initial proportion of cooperators required for spillover effect to give r becomes more demanding as p increases. Context interference is shown to be most res self comparison, while neighbour imitation is the most vulnerable. (D–I) demonstrate phenomenon for selected parameter combinations ( )  with both simulation and pair a Parameters: n = 3600, m = 4, c = 0.35, b = 1, β = 0.2, α = 0.5. Data for (A–C) starts a initial cooperators for both layers. Simulations in (D–I) 6 × 106 time steps, averaged o Figure 4. Bistability phenomenon. Spillover can hinder or promote cooperation depending on initial conditions. (A–C) illustrates, for each frequency of spillover p, the initial level of cooperators that would result in zero cooperation at equilibrium, under each spillover mode. In all cases, for a fixed p, there is a threshold level of initial cooperators on each network layer beyond which the spillover effect switches to working in favor of cooperation. The initial proportion of cooperators required for spillover effect to give rise to cooperation becomes more demanding as p increases. Context interference is shown to be most resilient to this, followed by self comparison, while neighbour imitation is the most vulnerable. www.nature.com/scientificreports/ Figure 5 shows the proportion of each spillover mode, for low values of p, when spillover modes coevolve with strategy in the presence of mutation. Self comparison (red) is second place in (A) despite having the most individuals who cooperate on both layers. Context interference (yellow) has the largest proportion due to having the most individuals playing the payoff maximizing combination of C on the bottom layer and D on the top layer. Neighbour imitation (blue) accounts for the least proportion of the population. (B) shows that cooperation in the top layer increases with p but at the expense of cooperation in the bottom layer. Parameters: n = 400, m = 4, c = 0.35, β = 0.2, b = 1. Mutation rate μ = 10−4, 7.2 × 109 total time steps, combined from at most 7 runs for both (A and B). individuals playing D on the top prisoner’s dilemma layer, and C on the bottom iterated prisoner’s dilemma layer (Fig. 3). When p is small, the top layer is approximately at full defection, while the bottom layer is approximately at full cooperation. So, playing D on top and C on the bottom offers the highest total payoff on average. As we saw in Figs 2F and 3, CIS shows a relative abundance of individuals with this type of mixed strategy, providing a convincing explanation for the supremacy of CIS in the competition. SCS mode produces a relatively high proportion of individuals cooperating on both layers, which might intu- itively be what one would desire in such a social system. However, this tendency results in a loss of individuals playing the optimal combination of strategies (Xc d), leading to a loss of competitiveness when pitted against other spillover modes, under the parameter combination of Fig. 5A. The dominance of CIS appears to be robust when the cost of cooperation c was lowered to c = 0.30, 0.25 and 0.20 (see Supplementary Fig. S7). However, at c = 0.20, this dominance is reduced. Conditions are favourable for cooperation, so we have high equilibrium proportions of cooperators on both layers (Supplementary Fig. S8A). In this situation, Xc c has a higher payoff than Xc d and hence, NIS and SCS become more competitive. This increase in competitiveness of the strategy that cooperates on both layers can be overcome by an increase in p (Supplementary Fig. S7A for p = 0.05). www.nature.com/scientificreports/ However, this does not happen when cooperation on both levels are high enough, and CIS once again loses its advantage (Supplementary Figs S7A and S8A, show the value c = 0.20, p = 0.1).l g y g p In Fig. 6, we consider an additional parameter α ∈ (0, 1) which governs the relative influence of the two layers. When spillover occurs under NIS and SCS, α is the probability that an individual on the top layer is chosen to possibly learn a strategy from the bottom layer. Under CIS, α is the probability that an individual uses her bottom layer strategy during spillover. We make these definitions so that parameter α consistently refers to how strong an influence the bottom IPD layer has when spillover occurs.i l We find that higher α promotes cooperation in general (Fig. 6C–E) and this could potentially alter the results of coevolution. As shown in Fig. 6A, it is possible for SCS to be favoured by selection instead of CIS when we set α = 0.95. This happens because there is a much higher level of cooperation in both layers when α is high (Figs 5B, 6B and S9B) which allows Xc c to be more competitive than Xc d. We show comparison of the cooperation level on both layers over time for both the α = 0.5 and α = 0.95 scenarios described in Supplementary Figs S10 and S11. www.nature.com/scientificreports/ (D–I) demonstrates the bistability phenomenon for selected parameter combinations ( )  with both simulation and pair approximation. Parameters: n = 3600, m = 4, c = 0.35, b = 1, β = 0.2, α = 0.5. Data for (A–C) starts at 0.05 and ends at 0.95 initial cooperators for both layers. Simulations in (D–I) 6 × 106 time steps, averaged over 200 runs. 6 Scientific RePorTS \| (2018) 8:6922 \| DOI:10.1038/s41598-018-25025-3 www.nature.com/scientificreports/ Figure 5. Coevolution of spillover modes and cooperation. Figure 5 shows the proportion of each spillover mode, for low values of p, when spillover modes coevolve with strategy in the presence of mutation. Self comparison (red) is second place in (A) despite having the most individuals who cooperate on both layers. Context interference (yellow) has the largest proportion due to having the most individuals playing the payoff maximizing combination of C on the bottom layer and D on the top layer. Neighbour imitation (blue) accounts for the least proportion of the population. (B) shows that cooperation in the top layer increases with p but at the expense of cooperation in the bottom layer. Parameters: n = 400, m = 4, c = 0.35, β = 0.2, b = 1. Mutation rate μ = 10−4, 7.2 × 109 total time steps, combined from at most 7 runs for both (A and B). Figure 5. Coevolution of spillover modes and cooperation. Figure 5 shows the proportion of each spillover Figure 5. Coevolution of spillover modes and cooperation. Figure 5 shows the proportion of each spillover mode, for low values of p, when spillover modes coevolve with strategy in the presence of mutation. Self comparison (red) is second place in (A) despite having the most individuals who cooperate on both layers. Context interference (yellow) has the largest proportion due to having the most individuals playing the payoff maximizing combination of C on the bottom layer and D on the top layer. Neighbour imitation (blue) accounts for the least proportion of the population. (B) shows that cooperation in the top layer increases with p but at the expense of cooperation in the bottom layer. Parameters: n = 400, m = 4, c = 0.35, β = 0.2, b = 1. Mutation rate μ = 10−4, 7.2 × 109 total time steps, combined from at most 7 runs for both (A and B). Figure 5. Coevolution of spillover modes and cooperation. Discussion Our results generate testable hypotheses that can inspire future research. Several experiments21–24 in the literature had participants play the iterated prisoner’s dilemma and then switch to various versions of one-shot games. All of these cases reported that the iterated prisoner’s dilemma, with conditions that favour cooperation, had a positive effect on cooperation level in the subsequent one-shot game. f q g Using setups similar to the experiments, one of the spillover mode could be incorporated to test for the existence of an optimal frequency of spillover p. For instance, NIS could be implemented experimentally by periodically hiding or revealing the strategies of neighbours on a layer. Similarly, CIS could be implemented by occasionally hiding or mislabelling the layers. At the same time, this can be used to test if spillover modes differ in the proportion of individuals playing each of the four possible top-bottom strategy combinations (Fig. 3), which offers a novel method for comparing and categorizing the myriad of spillover mechanisms that are possible. Scientific RePorTS \| (2018) 8:6922 \| DOI:10.1038/s41598-018-25025-3 7 www.nature.com/scientificreports/ Figure 6. Uneven influence of multiplex layers in spillover process. Figure 6 shows the effect of varying parameter α ∈ (0, 1) which governs the relative influence of the two multiplex layers. (A) shows that as α increases, self comparison spillover mode can have the highest proportion instead. A high α allows repeated interactions in the bottom layer to have a large influence and creates more favourable conditions for cooperation in the top layer, as shown by (B). This allows for individuals cooperating on both layers, which is plentiful under SCS, to become competitive during coevolution. (C–E) shows how cooperation in the multiplex increases with α under each spillover mode. Context interference is seen to be most resilient against the effect of spillover. Parameters for (A,B) n = 400, m = 4, b = 1, c = 0.35, β = 0.2, p = 0.05. Mutation rate μ = 10−3, at least 109 total time steps, combined from at most 10 runs. Parameters for (C–E) m = 4, b = 1, c = 0.35, β = 0.2, p = 0.05, generated using pair approximation. Figure 6. Uneven influence of multiplex layers in spillover process. Figure 6 shows the effect of varying parameter α ∈ (0, 1) which governs the relative influence of the two multiplex layers. Discussion (A) shows that as α increases, self comparison spillover mode can have the highest proportion instead. A high α allows repeated interactions in the bottom layer to have a large influence and creates more favourable conditions for cooperation in the top layer, as shown by (B). This allows for individuals cooperating on both layers, which is plentiful under SCS, to become competitive during coevolution. (C–E) shows how cooperation in the multiplex increases with α under each spillover mode. Context interference is seen to be most resilient against the effect of spillover. Parameters for (A,B) n = 400, m = 4, b = 1, c = 0.35, β = 0.2, p = 0.05. Mutation rate μ = 10−3, at least 109 total time steps, combined from at most 10 runs. Parameters for (C–E) m = 4, b = 1, c = 0.35, β = 0.2, p = 0.05, generated using pair approximation. Finally, our findings regarding the bistability that arise from varying the initial levels of cooperation on each layer (Fig. 4) could be leveraged to promote cooperation through spillover. As was done in a recent human behav- iour experiment37, mixing automated bots with human subjects can lead to the desired cooperation level. Finally, our findings regarding the bistability that arise from varying the initial levels of cooperation on each layer (Fig. 4) could be leveraged to promote cooperation through spillover. As was done in a recent human behav- iour experiment37, mixing automated bots with human subjects can lead to the desired cooperation level. www.nature.com/scientificreports/ Then, the focal individual does payoff comparison and strategy updating on the non-focal layer. However, if she decides to copy this neighbour’s strategy, the strategy is instead applied to focal layer, as illustrated by Fig. 1B. Self Comparison. If spillover occurs under SCS, we choose the top layer to be the focal layer with proba- bility α. Otherwise, with probability 1 − α, we choose the bottom layer. The focal individual then does payoff comparison and strategy updating with herself on the layer opposite to the focal layer, as shown in Fig. 1C. This means that there is a chance for the individual’s strategy on the focal layer to be replaced with her strategy on the non-focal layer. However, during SCS, her payoff on the bottom IPD layer is normalized by dividing the payoff matrix throughout by the number of game iterations. The payoff matrix used for SCS is,    − −    C D C D b c c m b m 0 (5) (5) Context Interference. Under CIS, individuals always update their strategy. However, the procedure for doing so is modified. During strategy updating, both the focal individual and the randomly chosen neighbour independently has a probability p of experiencing context interference. If one of them does, she has an independ- ent probability α of using her bottom layer strategy for all parts of the strategy updating procedure. Otherwise, with probability 1 − α, she uses her top layer strategy. These definitions are made so that parameter α consistently refers to how strong an influence the bottom IPD layer has when spillover occurs. This modified strategy updating procedure is shown in Fig. 1D. Coevolution and mutation. Figure 5 was generated by subjecting the three spillover modes to co-evolution and mutation. Individuals initially are assigned a spillover mode at random. We then proceed with two distinct phases: the regular phase and subsequently the coevolution phase. During the regular phase, at each discrete time step, a focal individual is chosen at random, then according to her spillover mode, strategy updating or spillover is carried out as described above. During the coevolution phase, a focal individual and a focal layer are chosen uni- formly at random. The focal individual then picks a neighbour on the focal layer at random. www.nature.com/scientificreports/ www.nature.com/scientificreports/ Let ∈ s {[1, 0] , [0, 1] } i k T T be the strategy of individual i on layer k = {T, B}, and Mk be the payoff matrix of the game on layer k = {T, B}. Then, the total payoff of an individual i on layer k is given by,  ∑ = ∈ P s M s ( ) , (3) i k j i k T k j k i k (3) where i k  is the neighbourhood of i on layer k. i g y At each discrete time step, we randomly choos h d d l d h h i At each discrete time step, we randomly choose a focal individual. Under NIS and SCS, with probability 1 − p, this individual updates her strategy. Otherwise, with probability p, spillover occurs. Spillover results in the trans- mission of cooperative (C) or defective (D) behaviour across layers, these behaviours correspond to the strategies C or D in both the IPD and PD games. Under CIS, this individual always updates her strategy using a modified procedure described below. Strategy updating. If individual i chooses to update her strategy, the top layer is chosen as the focal layer with probability 1 2. Otherwise, the bottom layer is chosen as the focal layer. Next, one of her neighbour j on the focal layer is picked at random. Then, the probability that i copies the strategy of j on the focal layer k is given by the Fermi equation38,39, → = + β − − F s s e ( ) 1 1 , (4) j k i k P P ( ) j k i k (4) where parameter β determines the intensity of selection, and P P , i k j k are the total payoffs within the layer k, of the focal individual i and the neighbour j respectively. where parameter β determines the intensity of selection, and P P , i k j k are the total payoffs within the layer k, of the focal individual i and the neighbour j respectively. Neighbour Imitation. If spillover occurs under NIS, we choose the top layer to be the focal layer with prob- ability α. Otherwise, with probability 1 − α, we choose the bottom layer. A neighbour on the layer opposite to the focal layer is chosen randomly. Methods In our model, n individuals are placed on two network layers. The top layer, T, is a random regular network of degree four, regenerated with every run, while the bottom layer, B, is a two dimensional lattice with periodic boundaries. Individuals play a special version of the prisoner’s dilemma game, known as the donation game, with their neighbours on the top layer: they are initially a cooperator C = [1, 0]T or defector D = [0, 1]T with equal probability. The payoff matrix MT that we use for prisoner’s dilemma is, =   − − . M b c c b 0 (1) T (1) where b is the benefit of cooperation, while c is the cost of cooperation. On the bottom layer, individuals play the iterated prisoner’s dilemma game with their partners, where the game is repeated m times. They initially start with the strategy C or D with equal probability. In this case, C refers to tit-for-tat (TFT), where the individual cooper- ates on the first iteration and plays the opponent’s previous strategy for future iterations. Here, D stands for always defecting (ALLD) during every iteration. The payoff matrix MB for this is, =    − −  . M m b c c b ( ) 0 (2) B (2) 8 Scientific RePorTS \| (2018) 8:6922 \| DOI:10.1038/s41598-018-25025-3 www.nature.com/scientificreports/ Their total payoffs on the multiplex network is then calculated by summing up their total payoffs across both layers. Payoff comparison is then done using the Fermi equation, which gives the probability of the focal individual copying the neighbour’s spillover mode. These two phases are run for 104 time steps each, and then repeated in the same order (regular then coevolution phase) until the desired total number of time steps is achieved when summed across all phases. In both the regular and coevolution phases whenever strategy update is successful there is a probability μ of p ) p p In both the regular and coevolution phases, whenever strategy update is successful, there is a probability μ of mutation. When mutation occurs, strategy or spillover mode is selected at random, instead of copied. Analytical solutions. We derived analytical solutions for each of the spillover mode, taking into account population structure, using extended pair approximation. We refer readers to the SI for the pair approximation equations and their details. Scientific RePorTS \| (2018) 8:6922 \| DOI:10.1038/s41598-018-25025-3 9 www.nature.com/scientificreports/ Data availability. The datasets and code supporting this article have been uploaded as part of the supple- mentary material. References Evolutionary games on multilayer networks: a colloquium. Eur. Phys. J. B 88(5), 124 (2015). 17. Wang, Z., Szolnoki, A. & Perc, M. Optimal interdependence between networks for the evolution of cooperation. Sci. Rep. 3, (2013). ( ) 18. Wang, Z., Szolnoki, A. & Perc, M. Evolution of public cooperation on interdependent networks: The impact of biased utility functions. EPL 97(4), 48001 (2012). 9. Jiang, L. L., Li, W. J. & Wang, Z. Multiple effect of social influence on cooperation in interdependent network games. Sci. Rep. 5 14657 (2015). J g, , , J g, pfl p p k g R p , 14657 (2015). 20 Santos M D Dorogovtsev S N & Mendes J F F Biased imitation in coupled evolutionary games in interdependent networks Sci ( ) 0. Santos, M. D., Dorogovtsev, S. N. & Mendes, J. F. F. Biased imitation in coupled evolutionary games in interdependent networks. Sci Rep. 4, 4436 (2014). p ( ) 21. Peysakhovich, A. & Rand, D. G. Habits of virtue: Creating norms of cooperation and defection in the laboratory. Manag. Sci. 62(3), 631–647 (2015). 22. Arechar, A. A., Kouchaki, M. & Rand, D. G. Examining Spillovers between Long and Short Repeated Prisoner’s Dilemma G Played in theLaboratory. Games. 9(1), 5 (2018).f y y 23. Duffy, J. & Ochs, J. Cooperative behavior and the frequency of social interaction. GEB 66(2), 785–812 (2009).i 24. Fréchette, G. R. & Yuksel, S. Infinitely repeated games in the laboratory: Four perspectives on discounting and random termina Exp. Econ. 20(2), 279–308 (2017). p 25. Stagnaro, M. N., Arechar, A. A. & Rand, D. G. From good institutions to generous citizens: Top-down incentives to coop promote subsequent prosociality but not norm enforcement. Cognition 167, 212–254 (2017).hh 6. Axelrod, R. The evolution of strategies in the iterated prisoner’s dilemma. The Dynamics Of Norms. 1–16 (Cambridge University Press 1987). ) 7. Bear, A. & Rand, D. G. Intuition, deliberation, and the evolution of cooperation. Proc. Natl. Acad. Sci. USA 113(4), 936–941 (2016) 8. Rand, D. G. et al. Social heuristics shape intuitive cooperation. Nat. Commun. 5, 3677 (2014).h D. G. Intuition, deliberation, and the evolution of cooperation. Proc 28. Rand, D. G. et al. Social heuristics shape intuitive cooperation. Nat. Commun. 5, 3677 (2014).h 29. Fudenberg, D. & Maskin, E. The folk theorem in repeated games with discounting or with incomplete information. Econometrica 54(3), 533–554 (1986). 0. Bear, A., Kagan, A. References & Rand, D. G. Co-evolution of cooperation and cognition: the impact of imperfect deliberation and context sensitive intuition. Proc. Royal Soc. B 284(1851), 20162326 (2017). y 1. Jagau, S. & van Veelan, M. A general evolutionary framework for the role of intuition and deliberation in cooperation. Nat. Hum Behav. 1(8), 0152 (2017). ( ) ( ) 32. Reiter, J. G., Hilbe, C., Rand, D. G., Chatterjee, K. & Nowak, M. A. Crosstalk in concurrent repeated games impedes direct reciprocity and requires stronger levels of forgiveness. Nat. Commun. 9(1), 555 (2018). q g g 3. Imhof, L. A., Fudenberg, D. & Nowak, M. A. Evolutionary cycles of cooperation and defection. Proc. Natl. Acad. Sci. USA 102(31) 10797–10800 (2005). 4. Toupo, D. F., Rand, D. G. & Strogatz, S. H. Limit Cycles Sparked by Mutation in the Repeated Prisoner’s Dilemma. Int. J. Bifurc Chaos 24(12), 1430035 (2014).l ( ) ( ) 35. Rapoport, A. & Chammah, A. M. Prisoner’s Dilemma: A Study In Conflict And Cooperation (University of Michigan Press 1965).fh 35. Rapoport, A. & Chammah, A. M. Prisoner’s Dilemma: A Study In Conflict And Cooperation (University of Michigan Press 1965). 36. Van Baalen, M. Pair approximations for different spatial geometries. The Geometry Of Ecological Interactions: Simplifying Spatial 5. Rapoport, A. & Chammah, A. M. Prisoner’s Dilemma: A Study In Conflict And Cooperation (University of Michigan Press 1965). 6. Van Baalen, M. Pair approximations for different spatial geometries. The Geometry Of Ecological Interactions: Simplifying Spatia Complexity. 359–387 (Cambridge University Press, 2000). Complexity. 359–387 (Cambridge University Press, 2000). 37. Shirado, H. & Christakis, N. A. Locally noisy autonomous agents improve global human coordination in network experiments. Nature 545(7654), 370 (2017). 38. Traulsen, A., Semmann, D., Sommerfeld, R. D., Krambeck, H. J. & Milinski, M. Human strategy updating in evolutionary games. Proc. Natl. Acad. Sci. USA 107(7), 2962–2966 (2010). ( ) ( ) 39. Szabó, G. & Töke, C. Evolutionary prisoner’s dilemma game on a square lattice. Phys. Rev. E 58(1), 69 (1998) References References 1. Axelrod, R. & Hamilton, W. D. The evolution of cooperation. Science 211(4489), 1390–1396 (1981).h ,h p Q ( ), ( ) 3. Levin, S. A. Learning to live in a global commons: socioeconomic challenges for a sustainable environment. Ecol. Res. 21(3), 328–333 (2006). ) and, D. G. & Nowak, M. A. Human cooperation. Trends Cogn Sci. 17(8), 413–425 (2013).h ( ) 4. Rand, D. G. & Nowak, M. A. Human cooperation. Trends Cogn Sci. 17(8), 413–425 (2013).h p g 5. Hardin, G. The Tragedy of the Commons. Science 162(3859), 1243–1248 (1968). h g y 6. Nowak, M. A. Five rules for the evolution of cooperation. Science 314(5805), 1560–1563 (2006). g y A. Five rules for the evolution of cooperation. Science 314(5805), 1 p ( ) ( ) 7. Lieberman, E., Hauert, C. & Nowak, M. A. Evolutionary dynamics on graphs. Nature 433(7023), 312 (2005). y y g p 8. Santos, F. C. & Pacheco, J. M. A new route to the evolution of cooperation. J. Evol. Biol. 19(3), 726–733 (2006). p 9. Szabó, G. & Fath, G. Evolutionary games on graphs. Phys. Rep. 446(4–6), 97–216 (2007). abó, G. & Fath, G. Evolutionary games on graphs. Phys. Rep. 446(4– ath, G. Evolutionary games on graphs. Phys. Rep. 446(4–6), 97–216 y g g p y p 0. Ohtsuki, H., Hauert, C., Lieberman, E. & Nowak, M. A. A simple rule for the evolution of cooperation on graphs and socia networks. Nature 441(7092), 502–505 (2006). 1. Santos, F. C. & Pacheco, J. M. Scale-free networks provide a unifying framework for the emergence of cooperation. Phys. Rev. Lett 95(9), 098104 (2005). ( ) ( ) 12. Perc, M. & Szolnoki, A. Social diversity and promotion of cooperation in the spatial prisoner’s dilemma game. Phys. Rev. E 77(1), 011904 (2008). ( ) 3. Buldyrev, S. V., Parshani, R., Paul, G., Stanley, H. E. & Havlin, S. Catastrophic cascade of failures in interdependent networks. Nature 464(7291), 1025–1028 (2010). ( ) ( ) 14. Kivelä, M. et al. Multilayer networks. J. Complex Net. 2(3), 203–271 (2014).h . Kivelä, M. et al. Multilayer networks. J. Complex Net. 2(3), 203–27 14. Kivelä, M. et al. Multilayer networks. J. Complex Net. 2(3), 203–271 (2014).h y p 15. Boccaletti, S. et al. The structure and dynamics of multilayer networks. Phys. Rep. 544(1), 1–122 (2014). h 16. Wang, Z., Wang, L., Szolnoki, A. & Perc, M. Acknowledgementsh g The work in this paper was supported by the Dartmouth Faculty Start-up Fund to F.F. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. g The work in this paper was supported by the Dartmouth Faculty Start-up Fund to F.F. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Scientific RePorTS \| (2018) 8:6922 \| DOI:10.1038/s41598-018-25025-3 Author Contributions T.K., F.F. and S.P. conceived the model. F.F. and T.K. derived the analytical solutions. T.K. wrote the code and analysed the results. F.F. and S.P. supervised the research. All authors reviewed the manuscript. T.K., F.F. and S.P. conceived the model. F.F. and T.K. derived the analytical solutions. T.K. wrote the code and analysed the results. F.F. and S.P. supervised the research. All authors reviewed the manuscript. Additional Information Supplementary information accompanies this paper at https://doi.org/10.1038/s41598-018-25025-3. Scientific RePorTS \| (2018) 8:6922 \| DOI:10.1038/s41598-018-25025-3 10 www.nature.com/scientificreports/ Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not per- mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. © The Author(s) 2018 Scientific RePorTS \| (2018) 8:6922 \| DOI:10.1038/s41598-018-25025-3 11
https://openalex.org/W4388758195	https://www.periodicojs.com.br/index.php/hs/article/download/1736/1525	Portuguese	null	FUNCTIONAL PRO GROUP FOR SENIOR ADULTS	Health and Society	2,023	cc-by	2,207	FUNCTIONAL PRO GROUP FOR SENIOR ADULTS Lucas Ordozgoite Coelho Joaquim da Silva1 Mateus de Almeida Ribeiro2 Resumo: Como sabemos hoje em dia vemos um público muito mais novo nas academias, crossfit, fun cionais e entre outros meios de manter a saúde em prática. e observamos que hoje o número de idosos vem crescendo pra voltar a fazer parte da maioria que praticar atividade saudável, muitos acham que não preci sam e acham até mesmo desnecessário, mas pra essa faixa etária é de extrema importância que eles não fi quem parados, pois pode ser prejudicial à saúde deles, exemplos: Doenças cardíacas e AVCs entre outros. O funcional tem o intuito de preparar o idoso para as atividades do dia a dia, prevenindo-o contra lesões, retardando o envelhecimento e evitando doenças crônicas como hipertensão, obesidades e diabetes, comuns na terceira idade. esse treinamento surgiu como finalidade trabalhar todas as fun cionalidades do corpo humano. o funcional na areia ajuda principalmente em Coordenação Motora, Flexibilidade, Equilíbrio, membros superiores e inferiores e tronco. Palavras-chave: Funcional, saúde, exercício físico Palavras-chave: Funcional, saúde, exercício físico Abstract: As we know, nowadays we see a much younger audience in gyms, crossfit, functional and other means of maintaining health in practice. and we observe that today the number of elderly peo ple is growing to become part of the majority who practice healthy activities, many think they don’t need it and even think it’s unnecessary, but for this age group it is extremely important that they don’t 1 B h l d Ed ã Fí i d U i id d Nil Li 1 Bacharelado em Educação Física da Universidade Nilton Lins 2 Bacharelado em Educação Física da Universidade Nilton Lins 85 ISSN: 2763-5724 / Vol. 03 - n 06 - ano 2023 ISSN: 2763-5724 / Vol. 03 - n 06 - ano 2023 85 stand still, as it can be harmful to their health, examples: Heart disease and stroke, among others. The functional aims to prepare the elderly for daily activities, preventing them from injuries, delaying aging and avoiding chronic diseases such as hypertension, obesity and diabetes, common in old age. This training arose with the aim of working on all the functionalities of the human body. Functioning in the sand helps mainly with Motor Coordination, Flexibility, Balance, upper and lower limbs and trunk. Keywords: Functional, health, physical exercise INTRODUÇÃO Como sabemos hoje em dia vemos um público muito mais novo nas academias, crossfit, fun cionais e entre outros meios de manter a saúde em prática. e observamos que hoje o número de idosos vem crescendo pra voltar a fazer parte da maioria que praticar atividade saudável, muitos acham que não precisam e acham até mesmo desnecessário, mas pra essa faixa etária é de extrema importância que eles não fiquem parados, pois pode ser prejudicial à saúde deles, exemplos: Doenças cardíacas e AVCs entre outros. O funcional tem o intuito de preparar o idoso para as atividades do dia a dia, prevenindo-o contra lesões, retardando o envelhecimento e evitando doenças crônicas como hipertensão, obesida des e diabetes, comuns na terceira idade. esse treinamento surgiu como finalidade trabalhar todas as funcionalidades do corpo humano. o funcional na areia ajuda principalmente em Coordenação Motora, Flexibilidade, Equilíbrio, membros superiores e inferiores e tronco. A qualidade de vida da população idosa está relacionada com suas atividades cotidianas e da manutenção ou ganho de auto nomia no seu meio familiar (DARNTON-HILL, 1995). Existem vários tipos de funcionais. O funcional que estamos abordando é atividades com o peso do corpo sem ter que precisar pesos por fora, montando circuitos com cones, chapéu francês, ISSN: 2763-5724 / Vol. 03 - n 06 - ano 2023 86 ISSN: 2763-5724 / Vol. 03 - n 06 - ano 2023 86 bambolês, cordas, Nossa primeira experiência profissional com funcional na areia e via todos os dias os idosos curiosos pra fazer e ao mesmo tempo com medo, e achando que era mais pra jovens ou atletas de alto rendimento, explicava o quanto ia ajudar eles principalmente depois da pandemia vieram esses problemas respiratórios, vim com o objetivo de mostrar a importância pro grupo da terceira idade e mostrar a eles como é excencial essa prática de atividade física. FUNCIONAL – DA ORIGEM O termo “funcional” pode ser entendido como:a)Referente à função ou ao desempenho des ta; b) Concernente a funções orgânicas vitais ou à sua realização; c) Diz-se daquilo que é capaz de cumprir com eficiência seus fins utilitários;d)é utilizada também como adjetivo particular ou relativo àsfunções biológicas ou psíquicas. Ou seja, a aplicação correta do termo “funcional” deve estar associada ou se relacionar às funções do sistema psico-bioló gico humano, com eficácia e respeitando as funções citadas. Todos os treinamentos, sem exceção, devem ter por objetivo o desenvolvimento de alguma variável de funcionalidade. O treinamento da variável funcional não depende de nenhum tipo de equi pamento e ou determinado tipo de exercício (no entanto, parece que somente recebem o mérito de ser “funcional” o treinamento que apresentatais características). (GONZALES ,2020) A prescrição de treinamento funcional deve fornecer a adequada “dose” de exercícios frente àspossibilidades de resposta ao estimulo e garantir adaptações ótimas em relação aos critérios de eficácia e funcionalidade. Ao eleger determinado exercício, levando-se em consideração a perspectiva da funcionali dade, não significa que se realizou “treinamento funcional”, porque está diante de aspectos distintos. ISSN: 2763-5724 / Vol. 03 - n 06 - ano 2023 87 Um treinamento para ser considerado funcional deve contemplar exercícios selecionados tendo como critério a sua funcionalidade e isto só é possível atendendo às cinco variáveis distintas da funcionali dade5: a) frequência adequada dos estímulos de treinamento; b) volume em cada uma das sessões; c) a intensidade adequada; d)densidade, ou seja, ótima relação entre duração do esforço e a pausa de re cuperação); e) organizaçãometodológica das tarefas. O manejo adequado das variáveis supracitadas permitirá uma eficaz consecução dosobjetivos pretendidos na melhora ou manutenção da capacidade funcional do sistema psico-biológico. ( GONZALES, 2020) A proposta do treinamento funcional carece de critérios de aplicação e progressão baseada em fundamentos do treinamento desportivo.Além disso, não se tem ciência deuma abordagem que integreametodologiado treinamento funcional aostradicionais programas de condicionamento físico saudável. Outro aspecto que parece ser utilizado de modo superficial é o da “naturalidade”, ou seja,que algumas tarefas motoras aplicadas a alguns indivíduos são mais”naturais”que outras, que parecem não possuir esta característica. Quando se busca definição do termo natural, encontra-se o seguinte conceito: “pertencente ou relativo à natureza conforme a qualidade e propriedade das coisas”. FUNCIONAL – DA ORIGEM Seguin do tal conceituação será mais natural o exercício que atenda às propriedades bio- psicológicas, em contrapartida, será menos natural aqueles que não o façam. Outro aspecto a ser considerado quanto à “naturalidade” está associado à”ausência de artifícios”, ou seja, deveriam ser utilizados exercícios realizados sem implemento. Por fim, é comum afirmar que é mais natural aquilo que esteja ligado aos aspectos de vida do indivíduo. (GONZALES, 2020) A longevidade progressiva é óbvia e inevitável. As famílias necessitam se conscientizar de que o idoso permanecerá junto a elas por anos, pois, com o evidente envelhecimento populacional, cada vez mais os lares têm a possibilida de da presença de idosos. Preocupadas com a manutenção dos laços e papéis das pessoas de mais idade com seus familiares, objetivamos indagar junto às ido ISSN: 2763-5724 / Vol. 03 - n 06 - ano 2023 88 sas pertencentes a um grupo de terceira idade, valorizar a interação recíproca entre idoso e família e identificar, por meio de entrevistas grupais, como compartilham suas vivências no seio familiar e de acordo com Beck, Gonzáles e Colomé, “o envelhecimento é um processo cumulativo, irreversível, universal, não patológico, de deterioração de um organismo maduro e que pode incapacitar o indiví duo para o desenvolvimento de algumas atividades” ( BECK, GONZALES 2020) As aulas de funcional são divididas em aquecimento, parte principal e relaxamento, realiza das na piscina e com duração de aproximadamente de 50 a 60 minutos. No desenvolvimento da aula pode-se fazer uso de músicase ritmos juntamente com os exercícios (SANTOS, 2021, p. 11). A música é um recurso didático bem interessante para os professores utilizarem durante suas aulas, principalmente com a população idosa. Isso ocorre porque sugere aos indivíduos a expressão de emoção através de gestos. Ao lidar com o público idoso é importante que as aulas sejam estrutu radas de forma a proporcionar atividades motivadoras, visando o desempenho das tarefas diárias e o cumprimento das recomendações médicas. É fundamental que essas atividades sejam adaptadas às necessidades e limitações dos idosos, incentivando aautonomia e a qualidade de vida. No geral, o funcional são altamente benéficos para os idosos, melhorando a qualidade de vida, a saúde física e emocional. Problema de Pesquisa Porque os Idosos não procuram Fazer atividades físicas?. Como o funcional. Objetivos Identificar através de revisão de literatura os principais motivos do diminui-o do grupo da terceira idade praticar tão pouco essa modalidade que é o Funcional ISSN: 2763-5724 / Vol. 03 - n 06 - ano 2023 90 Justificativa Não é nada positivo que os idosos ficam em casa sem fazer nada, assistindo Tv e jogando conversa fora. é essencial que eles se movam, assim se torna possível que eles consigam se motivar 89 ISSN: 2763-5724 / Vol. 03 - n 06 - ano 2023 ISSN: 2763-5724 / Vol. 03 - n 06 - ano 2023 ISSN: 2763-5724 / Vol. 03 - n 06 - ano 2023 89 muito mais. e o Treinamento Funcional é uma ótima maneira de colocar isso em prática. Um ido so que queira manter sua independência e viver bem deve se manter ativo. Ainda, se levarmos em consideração que boa parte dos acidentes relacionados à terceira idade está relacionada à falta de mobilidade física, fica clara a importância da atividade física. O sedentarismo entre outras doenças prejudica demais no dia a dia quando alguém envelhece sem manter suas capacidades funcionais, o que infelizmente é o caso da maioria das pessoas, ele fica exposto a uma série de riscos. de riscos o funcional vem ajudar principalmente na parte cardiovascular, alguns idosos sentem-se cansados em excesso e têm dificuldades para realizar alguma tarefa por falta de resistência física. Isso está rela cionado aos problemas cardiovasculares que vão surgindo com o tempo. Os benefícios do funcional para os Idosos: Melhora o equilíbrio, aumenta a flexibilidade, Cria resistência física, aumenta a força muscular, Melhora da socialização, Temos que ter total cuidado o idoso no treinamento no funcional na areia, exemplos quando for trabalhar verificar se a areia está apropriada para fazer o treino, pra não ter nenhuma sujeira, e acabar furando o pé de algum aluno, molhar de vez em quando pra areia não ficar muito seca e dura e prejudicar algum aluno. Com essa atividade, o indivíduo da terceira idade consegue recuperar suas capacidades funcionais, ficando menos exposto a riscos de lesão e quedas durante seu cotidiano. Nes se sentido, o Treinamento Funcional consegue auxiliar o idoso a recuperar a mobilidade e o equilíbrio e melhorar a qualidade de vida deles, graças a nós profissionais de educação física. B) Apontar os benefícios do funcional para o grupo da 3ª idade B) Apontar os benefícios do funcional para o grupo da 3ª idade CONSIDERAÇÕES FINAIS O presente trabalho investigou, por meio de uma revisão literária, aspectos significativos dos processos de mudança que estão ocorrendo, especialmente em relação às questões demográficas e às demandas da população conhecida como terceira idade. Identificou-se através deste estudo que, o envelhecimento é caracterizado como o conjunto de mudanças no organismo que ocorrem ao longo do tempo, afetando diferentes aspectos do corpo, tais como sua estrutura, funcionamento psicológico e capacidades físicas. Essas transformações afetam principalmente as pessoas idosas. Em virtude dos fatos mencionados neste estudo, foi possível observar que a prática do fun cional traz benefícios para a saúde física dos idosos. No geral, a importância desta pesquisa está no entendimento de como o funcional pode contribuir para a promoção da saúde desta classe, visto que os idosos representam um grupo crescente na sociedade brasileira. Espera-se que os resultados pos sam subsidiar a elaboração de políticas públicas e práticas de promoção de saúde para essa população, bem como contribuir para o desenvolvimento de novas pesquisas sobre o tema. Referências bibliográficas MEDLINE, BioMed Central, SciELO, Scholar Google e SportDiscus. Os trabalhos selecionados fo ram publicados durante o período de 2020 a 2023. YabuutiP. L. K., JesusG. de M. de, BurattiA., Bas saniG. A., CastroH., PereiraJ. dá S., NakamotoJ. M., GushikenE. S., FilhoC. B., & GonçalvesI. de O. (2019). ISSN: 2763-5724 / Vol. 03 - n 06 - ano 2023 91 Objetivos específicos A) Conceituar o que é a modalidade Funcional B) Apontar os benefícios do funcional para o grupo da 3ª idade MENESES, YÚLA PIRES DA SILVEIRA FONTENELE DE. Hidroginástica: treinamento e qua lidade de vida – Parnaíba, PI: Acadêmica Editorial, 2021. LIMA, F.A.S. Relação entre memória de trabalho e capacidade física: um estudo com idosos prati cantes de hidroginástica. 2021. 45f. Dissertação (Mestrado em Educação Física) -Centro de Ciências da Saúde, Universidade Federal do Rio Grande do Norte, Natal, 2021. MENESES, YÚLA PIRES DA SILVEIRA FONTENELE DE. Hidroginástica: treinamento e qua lidade de vida – Parnaíba, PI: Acadêmica Editorial, 2021. ISSN: 2763-5724 / Vol. 03 - n 06 - ano 2023 92
https://openalex.org/W3198707284	https://europepmc.org/articles/pmc8443541?pdf=render	English	null	The Nazeris fauna of the Nanling Mountain Range, China (Coleoptera, Staphylinidae, Paederinae)	ZooKeys	2,021	cc-by	6,595	ZooKeys 1059: 117–133 (2021) doi: 10.3897/zookeys.1059.72240 https://zookeys.pensoft.net ZooKeys 1059: 117–133 (2021) doi: 10.3897/zookeys.1059.72240 https://zookeys.pensoft.net ZooKeys 1059: 117–133 (2021) doi: 10.3897/zookeys.1059.72240 https://zookeys.pensoft.net Keywords Flightless, leaf litter, new species, Oriental Region, rove beetles, taxonomy Abstract Fourteen species of Nazeris Fauvel, 1873 are reported for the Nanling Mountain Range, China. Four of them are described as new: N. xingmini Lin & Hu, sp. nov. (Guangdong, Jiangxi), N. huaiweni Lin & Hu, sp. nov. (Guangdong), N. meihuaae Lin & Hu, sp. nov. (Guangdong, Jiangxi) and N. lichongi Lin & Hu, sp. nov. (Hunan). An identification key to the Nazeris species reported for the Nanling Mountains, and a map showing their distribution are provided. http://zoobank.org/74BFDC46-8DE9-424C-B168-5F03CF818C3C Citation: Lin X-B, Hu J-Y (2021) The Nazeris fauna of the Nanling Mountain Range, China (Coleoptera, Staphylinidae, Paederinae). ZooKeys 1059: 117–133. https://doi.org/10.3897/zookeys.1059.72240 Copyright Xiao-Bin Lin, Jia-Yao Hu. This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. The Nazeris fauna of the Nanling Mountain Range, China (Coleoptera, Staphylinidae, Paederinae) Xiao-Bin Lin1, Jia-Yao Hu1 1 Department of Biology, College of Life Sciences, Shanghai Normal University, 100 Guilin Road, 1st Educa tional Building 423-A Room, Shanghai, 200234 China Corresponding author: Jia-Yao Hu (hujiayao@shnu.edu.cn) emic editor: Adam Brunke \| Received 27 July 2021 \| Accepted 14 August 2021 \| Published 8 September 202 Introduction The speciose paederine genus Nazeris Fauvel, 1873 previously included 191 species and one subspecies in China. The genus is distinguished from other paederines particularly by the morphology of the aedeagus, which has a pair of dorso-lateral apophyses (Assing 118 Xiao-Bin Lin & Jia-Yao Hu / ZooKeys 1059: 117–133 (2021) 2009). All the known species of the genus are micropterous and flightless. Therefore, most of them have very narrow distributions. The Nanling Mountain Range in central and eastern China, extend more than 500 km through Guangxi, Hunan, Guangdong and Jiangxi provinces, with several peaks of more than 1500 m. To the present, fifteen Nazeris species have been described from Nanling Mountain Range (Assing 2014, 2016; Hu and Li 2017; Hu et al. 2018a; Hu and Qiao 2019). During recent field trips in the Nanling Mountains, many specimens of Nazeris were collected. Among them, four new species were found. In the present paper, we describe the new species and provide illustrations of their major diagnostic features. Material and methods The type material is deposited in the Insect Collection of the Shanghai Normal University, Shanghai, China (SNUC). The dissected body parts were mounted in Euparal on plastic slides. The habitus photos were taken using a Canon 7D camera. The photos of the ster nites and aedeagi were taken using a Canon G9 camera mounted on an Olympus CX31 microscope. The original map source was obtained from http://www.simplemappr.net, an on-line tool for creating maps that can be freely used for publications and presentations. Measurements Body length: measured from the anterior margin of the labrum to the apex of the abdomen Length of forebody: measured from anterior margin of labrum to the posterior margin of the elytra Eye length: longitudinal length of eye in dorsal view Postocular length: measured from posterior margin of eye to posterior constriction of head in dorsal view Head width: width of head across (and including) eyes Head length: measured from clypeal anterior margin to posterior constriction of head Pronotum width: maximal width of pronotum Pronotum length: measured in midline from front margin to posterior margin Width of elytra: combined width of elytra at posterior margin Length of elytra: measured from apex of scutellum to posterior margin. Body length: measured from the anterior margin of the labrum to the apex of the abdomen Body length: measured from the anterior margin of the labrum to the apex of the abdomen Length of forebody: measured from anterior margin of labrum to the posterior margin of the elytra Results Nazeris alatus Hu & Li, 2017 Fig. 27 Nazeris gaoleii Hu, Luo & Li, 2018 Fig. 27 Non-type material examined. China: Guangdong Prov.: Shaoguan, Nanling N. R.: 1 ♂, 1 ♀, 24°56'38"N, 112°59'31"E, 1316–1575 m, 29.vi.2020, Xia, Zhang, Yin and Lin leg.; 2 ♀♀, Guangdong Diyifeng, 24°55'29.62"N, 112°59'31.42"E, 1538– 1784 m, 28.vi.2020, Xia, Zhang, Yin and Lin leg.; 1 ♂, Xiaohuangshan, 24°53'58"N, 113°01'27"E, 1,425 m, 23.viii.2020, sifted, Zhong Peng leg.; Hunan Prov.: Yizhang, Mangshan N. R.: 3 ♂♂, 3 ♀♀, Mengkengshi, 24°55'10"N, 112°58'37"E, 1625 m, 28.viii.2020, sifted, Zhong Peng leg.; 1 ♂, Jiangjunzhai, 24°57'03"N, 112°55'37"E, 1220 m, 27.viii.2020, sifted, Zhong Peng leg. (SNUC). Non-type material examined. China: Guangdong Prov.: Shaoguan, Nanling N. R.: 1 ♂, 1 ♀, 24°56'38"N, 112°59'31"E, 1316–1575 m, 29.vi.2020, Xia, Zhang, Yin and Lin leg.; 2 ♀♀, Guangdong Diyifeng, 24°55'29.62"N, 112°59'31.42"E, 1538– 1784 m, 28.vi.2020, Xia, Zhang, Yin and Lin leg.; 1 ♂, Xiaohuangshan, 24°53'58"N, 113°01'27"E, 1,425 m, 23.viii.2020, sifted, Zhong Peng leg.; Hunan Prov.: Yizhang, Mangshan N. R.: 3 ♂♂, 3 ♀♀, Mengkengshi, 24°55'10"N, 112°58'37"E, 1625 m, 28.viii.2020, sifted, Zhong Peng leg.; 1 ♂, Jiangjunzhai, 24°57'03"N, 112°55'37"E, 1220 m, 27.viii.2020, sifted, Zhong Peng leg. (SNUC). g g g Comparative notes. Nazeris gaoleii is very similar to N. jiaweii Hu, Liu & Li, 2018b in general appearance and aedeagal characters, but can be separated by the nar rower posterior excision of male sternite VIII (Hu et al. 2018a: 179, fig. 21), and much shorter dorso-lateral apophyses of the aedeagus (Hu et al. 2018a: 179, fig. 22).h i Distribution and habitat data. The species is known from Nanling in northern Guangdong and Mangshan in southern Hunan (Fig. 27). The specimens were col lected by sifting leaf litter at altitudes of 1220–1820 m. Nazeris alatus Hu & Li, 2017 Fig. 27 Non-type material examined. China: Guangxi Prov.: Guilin, Huaping N. R.: 5 ♂♂, 4 ♀♀, nr. Guangfu Peak, 24°33'36.57"N, 109°55'40.81"E, ca 1800 m, 22.iv.2021, sifted, 119 Nazeris fauna of the Nanling Mountain Range Yin, Zhang, Pan and Shen leg.; 4 ♂♂, 3 ♀♀, Yunxi Valley, 25°34'00.62"N, 109°56'19.59"E, 1460–1550 m, 23.iv.2021, sifted, Yin, Zhang, Pan and Shen leg. (SNUC). Yin, Zhang, Pan and Shen leg.; 4 ♂♂, 3 ♀♀, Yunxi Valley, 25°34'00.62"N, 109°56'19.59"E, 1460–1550 m, 23.iv.2021, sifted, Yin, Zhang, Pan and Shen leg. (SNUC). Yin, Zhang, Pan and Shen leg.; 4 ♂♂, 3 ♀♀, Yunxi Valley, 25°34'00.62"N, 109°56'19.59"E, 1460–1550 m, 23.iv.2021, sifted, Yin, Zhang, Pan and Shen leg. (SNUC). g g Comparative notes. Nazeris alatus is very similar to N. yanzhuqii Hu & Qiao, 2019 in general appearance, but can be separated by the midline of the pronotum with short and narrow impunctate elevation posteriorly (Hu and Li 2017: 337, fig. 15); by the shorter and narrower ventral process of the aedeagus, and by the narrower dorso- lateral apophyses of the aedeagus (Hu and Li 2017: 337, fig. 18). i Distribution and habitat data. The species is known only from Huaping in northeast Guangxi (Fig. 27). The specimens were collected by sifting leaf litter at alti tudes of 1700–1800 m. Nazeris latilobatus Assing, 2016 Fig. 27 Non-type material examined. China: Guangxi: Xing’an, Mao’ershan N. R.: 1 ♂, 25°52'29.52"N, 110°28'20.01"E, 528 m, 25.viii.2020, Chong Li leg.; 2 ♂♂, 25°30'15.72"N, 110°25'50.87"E, 1900–2040 m, 27.viii.2020, Lu Qiu leg.; 20 ♂♂, 16 ♀♀, Antangping, 25°54'44.07"N, 110°27'37.68"E, 1660 m, 6–7.v.2021, sifted, Yin, Zhang, Pan and Shen leg.; 1 ♂, botanical garden, 25°53'03.83"N, 110°29'13.53"E, 1160 m, 8.v.2021, sifted, Yin, Zhang, Pan and Shen leg. (SNUC). Comparative notes. Nazeris latilobatus is similar to N. qini Hu & Li, 2012 from Dayaoshan in external and the male sexual characters, but can be separat ed by the deeper posterior excision of the male sternite VIII (Assing 2016: 307, fig. 9), and the much broader apex of the aedeagal ventral process (Assing 2016: 307, fig. 11). Comparative notes. Nazeris latilobatus is similar to N. qini Hu & Li, 2012 from Dayaoshan in external and the male sexual characters, but can be separat ed by the deeper posterior excision of the male sternite VIII (Assing 2016: 307, fig. 9), and the much broader apex of the aedeagal ventral process (Assing 2016: 307, fig. 11). Distribution and habitat data. The species is known only from Mao’ershan in northeast Guangxi (Fig. 27). The specimen was collected by sifting leaf litter at alti tudes of 450–2040 m. Nazeris huapingensis Hu & Li, 2017 Fig. 27 Non-type material examined. China: Guangxi Prov.: Guilin, Huaping N. R.: 6 ♂♂, 9 ♀♀, nr. Guangfu Peak, 24°33'36.57"N, 109°55'40.81"E, ca 1800 m, 22.iv.2021, sifted, Yin, Zhang, Pan and Shen leg.; 8 ♂♂, 11 ♀♀, Yunxi Valley, 25°34'00.62"N, 109°56'19.59"E, 1460–1550 m, 23.iv.2021, sifted, Yin, Zhang, Pan and Shen leg. (SNUC). Comparative notes. Nazeris huapingensis is very similar to N. obtortus Assing, 2016 from the same locality in general appearance and separated only by aedeagal characters: the apex of the ventral process in ventral view is much broader; apices of the dorso-lateral apophyses rounder and broader (Hu and Li 2017: 336, fig. 13). 120 Xiao-Bin Lin & Jia-Yao Hu / ZooKeys 1059: 117–133 (2021) Xiao-Bin Lin & Jia-Yao Hu / ZooKeys 1059: 117 120 Distribution and habitat data. The species is kno northeast Guangxi (Fig. 27). The specimens were collected tudes of 1300–1800 m. Distribution and habitat data. The species is known only from Huaping in northeast Guangxi (Fig. 27). The specimens were collected by sifting leaf litter at alti tudes of 1300–1800 m. Nazeris obtortus Assing, 2016 Non-type material examined. China: Guangxi Prov.: Guilin, Huaping N. R.: 1 ♂, 3 ♀♀, Hongtan, 25°36'15"N, 109°57'35"E, 820–950 m, 24.iv.2021, sifted, Yin, Zhang, Pan and Shen leg. (SNUC). Comparative notes. Nazeris obtortus is quite similar to N. huapingensis and sepa rated only by aedeagal characters: the apex of the ventral process and apices of the dorso-lateral apophyses are much narrower (Assing 2016: 309, fig. 16). Distribution and habitat data. The species is known only from Huaping in northeast Guangxi (Fig. 27). The specimens were collected by sifting leaf litter at alti tudes of 820–1200 m. Nazeris rubidus Hu, Luo & Li, 2018 Fig. 27 Non-type material examined. China: Guangdong Prov.: Shaoguan, Nanling N. R.: 5 ♂♂, 7 ♀♀, 24°56'38"N, 112°59'31"E, 1316–1575 m, 29.vi.2020, Xia, Zhang, Yin and Lin leg.; ♂♂, 43 ♀♀, 24°55'43.67"N, 113°0'58.50"E, 1,020 m, 27.vi.2020, Xia, Zhang, Yin and Lin leg.; 2 ♂♂, 5 ♀♀, Xiaohuangshan, 24°53'58"N, 113°01'27"E, 1425 m, 23.viii.2020, sifted, Zhong Peng leg. (SNUC). g g g Comparative notes. Nazeris rubidus is very similar to N. huapingensis in general appearance and aedeagal characters, but can be separated by the following combina tion of characters: the posterior excision of the male sternite VIII is wider (Hu et al. 2018a: 176, fig. 11); the apex of the ventral process of the aedeagus is widely rounded in ventral view (Hu et al. 2018a: 176, fig. 12); the dorso-lateral apophyses is nearly straight in lateral view (Hu et al. 2018a: 176, fig. 13).h i Distribution and habitat data. The species is known from Nanling in northern Guangdong and Mangshan in southern Hunan (Fig. 27). The specimens were col lected by sifted leaf litter at altitudes of 700–1820 m. Nazeris nanlingensis Hu, Luo & Li, 2018 Fig. 27 Non-type material examined. China: Guangdong Prov.: Shaoguan, Nanling N. R.: 2 ♂♂, 1 ♀, 24°56'38"N, 112°59'31"E, 1316–1575 m, 29.vi.2020, Xia, Zhang, Yin and Lin leg.; 2 ♂♂, 2 ♀♀, Xiaohuangshan, 24°53'58"N, 113°01'27"E, 1425 m, 23.viii.2020, sifted, Zhong Peng leg.; Hunan Prov.: Yizhang County, Mangshan N. R.: 2 ♂♂, 1 ♀, Mengkengshi, 24°55'10"N, 112°58'37"E, 1625 m, 28.viii.2020, sifted, Zhong Peng leg.; 6 ♂♂, 5 ♀♀, Jiangjunzhai, 24°57'03"N, 112°55'37"E, 1220 m, 27.viii.2020, sifted, Zhong Peng leg. (SNUC). Non-type material examined. China: Guangdong Prov.: Shaoguan, Nanling N. R.: 2 ♂♂, 1 ♀, 24°56'38"N, 112°59'31"E, 1316–1575 m, 29.vi.2020, Xia, Zhang, Yin and Lin leg.; 2 ♂♂, 2 ♀♀, Xiaohuangshan, 24°53'58"N, 113°01'27"E, 1425 m, 23.viii.2020, sifted, Zhong Peng leg.; Hunan Prov.: Yizhang County, Mangshan N. R.: 2 ♂♂, 1 ♀, Mengkengshi, 24°55'10"N, 112°58'37"E, 1625 m, 28.viii.2020, sifted, Zhong Peng leg.; 6 ♂♂, 5 ♀♀, Jiangjunzhai, 24°57'03"N, 112°55'37"E, 1220 m, 27.viii.2020, sifted, Zhong Peng leg. (SNUC). Comparative notes. Nazeris nanlingensis is very similar to N. rubidus Hu, Luo & Li, 2018a from the same locality in general appearance and aedeagal characters, but can be separated by the smaller forebody size; less dense punctation of the head and pronotum (Hu et al. 2018a: 177, fig. 14); a wider ventral process and slenderer dorso-lateral apophyses of the aedeagus in ventral view (Hu et al. 2018a: 177, fig. 17). i Distribution and habitat data. The species is known from Nanling in northern Guangdong and Mangshan in southern Hunan (Fig. 27). The specimens were col lected by sifting leaf litter at altitudes of 1100–1850 m. Nazeris fauna of the Nanling Mountain Range 121 Nazeris rugosus Hu & Qiao, 2019 Fig. 27 Non-type material examined. China: Guangxi Prov.: Xing’an, Mao’ershan N. R.: 1 ♂, 25°51'57.56"N, 110°24'46.19"E, 2100 m, 5.v.2021, bamboo, broad-leaved bush, sift ed, Yin, Zhang, Pan and Shen leg.; 1 ♂, Lijiangyuan, 25°53'32.64"N, 110°25'41.68"E, 1990–2030 m, 6.v.2021, sifted, Yin, Zhang, Pan and Shen leg. (SNUC). Comparative notes. Nazeris rugosus is distinguished from all the known species of Nazeris from the Nanling Mountains by the microsculpture covering the head, pronotum and abdomen (Hu and Qiao 2019: 436, figs 18–20), and by the distinctive shape of the ae deagus, particularly the apically narrowed ventral process (Hu and Qiao 2019: 436, fig. 23). Xiao-Bin Lin & Jia-Yao Hu / Zo 122 Distribution and habitat data. The sp northeast Guangxi (Fig. 27). The specimen tudes of 1990–2100 m. Nazeris yanzhuqii Hu & Qiao 2019 122 Xiao-Bin Lin & Jia-Yao Hu / ZooKeys 1059: 117–133 (2021) Distribution and habitat data. The species is known only from Mao’ershan in northeast Guangxi (Fig. 27). The specimen was collected by sifting leaf litter at alti tudes of 1990–2100 m. Nazeris yanzhuqii Hu & Qiao, 2019 Fig. 27 Non-type material examined. China: Guangxi Prov.: Xing’an, Mao’ershan N. R.: 1 ♀, Lijiangyuan, 25°53'32.64"N, 110°25'41.68"E, 1990–2030 m, 6.v.2021, sifted, Yin, Zhang, Pan and Shen leg. (SNUC). Comparative notes. Nazeris yanzhuqii is most similar to N. alatus in general ap pearance and aedeagal characters, but can be separated by the impunctate elevation of the pronotum very narrow or absent (Hu and Qiao 2019: 438, fig. 35), by the longer and wider ventral process of the aedeagus in ventral view, with much smaller basal laminae, and by the wider dorso-lateral apophyses of the aedeagus (Hu and Qiao 2019: 438, fig. 38).h Distribution and habitat data. The species is known only from Mao’ershan in northeast Guangxi (Fig. 27). The specimen was collected by sifting leaf litter at alti tudes of 1940–2140 m. Distribution and habitat data. The species is known only from Mao’ershan in northeast Guangxi (Fig. 27). The specimen was collected by sifting leaf litter at alti tudes of 1940–2140 m. Nazeris xingmini Lin & Hu, sp. nov. http://zoobank.org/C2BCCA84-E6E7-430C-BEE6-D7A6FA183751 Figs 1, 5–9, 27 Nazeris yuyimingi Hu & Qiao, 2019 Fig. 27 Fig. 27 Non-type material examined. China: Guangxi Prov.: Xing’an, Mao’ershan N. R.: 1 ♀, nr. Antangping, 25°54'44.07"N, 110°27'37.68"E, 1660 m, 7.v.2021, sifted, Yin, Zhang, Pan and Shen leg. (SNUC). Comparative notes. Nazeris yuyimingi is similar in general appearance and aedea gal characters to N. chenyanae Hu & Li, 2017, but can be separated by the shallowly emarginate male sternite VII (Hu and Qiao 2019: 437, fig. 26), by the narrower ven tral process and the wider apex of the dorso-lateral apophyses of the aedeagus in ventral view (Hu and Qiao 2019: 437, fig. 28). i Distribution and habitat data. The species is known only from Mao’ershan in northeast Guangxi (Fig. 27). The specimen was collected by sifting leaf litter at alti tudes of 1143–1660 m. Nazeris xingmini Lin & Hu, sp. nov. http://zoobank.org/C2BCCA84-E6E7-430C-BEE6-D7A6FA183751 Figs 1, 5–9, 27 Nazeris xingmini Lin & Hu, sp. nov. Xiao-Bin Lin & Jia-Yao H 124 Xiao-Bin Lin & Jia-Yao Hu / ZooKeys 1059: 117–133 (2021) Distribution and habitat data. The species is known from Chebaling in northern Guangdong and Jiulianshan in southern Jiangxi (Fig. 27). The specimens were col lected by sifting leaf litter at altitudes of 357–587 m. Comparative notes. The new species is very similar to N. inaequalis Assing, 2014 in general appearance and separated only by the aedeagal characters: the apex of the ventral process is symmetric in ventral view (Fig. 8); dorso-lateral apophyses extending beyond the apex of the ventral process. Etymology. The species is named in honor of Xing-Min Wang (South China Ag ricultural University) who helped a lot during our collection in Nanling. Nazeris huaiweni Lin & Hu, sp. nov. http://zoobank.org/2C276CF2-9719-40ED-A290-B50052066D15 Figs 2, 10–14, 27 Nazeris huaiweni Lin & Hu, sp. nov. http://zoobank.org/2C276CF2-9719-40ED-A290-B50052066D15 Figs 2, 10–14, 27 Type material. Holotype: China: ♂: “China: Guangdong, Shaoguan, Nanling N. R., 24°56'38"N, 112°59'31"E, 1316–1575 m, 29.vi.2020, Xia, Zhang, Yin and Lin leg.” (SNUC). Paratypes: 2 ♀♀, same data, except “Ruyuan, Nanling, nr. Ruyang, Xiao huangshan, 24°53'44.7"N, 113°1'26.9"E, 1270–1570 m, 2021.v.02, Hu, Lin, Zhou and Li leg.” (SNUC). g Description. Body length 6.2–6.8 mm; forebody length 3.2–3.3 mm. Body (Fig. 2) dark brown; antennae and legs yellowish brown. Head (Fig. 10) approximately as long as wide; punctation very dense, moderately coarse, distinctly umbilicate, interstices lacking microsculpture; postocular portion ap proximately twice as long as eye length. Pronotum (Fig. 10) 1.12–1.21 times as long as wide, 0.93–1.02 times as long and 0.83–0.84 times as broad as head; punctation non-umbilicate, moderately dense and as coarse as that of head; midline posteriorly with short and very narrow impunctate elevation; interstices lacking microsculpture. Elytra (Fig. 10) 0.67–0.76 times as long as wide, 0.69–0.71 times as long and 1.03–1.12 times as broad as pronotum; punctation slightly denser and coarser than that of pronotum; interstices lacking microsculpture. Abdomen with punctation dense and rather coarse on tergites III–V, dense and less coarse on tergite VI, moderately dense and fine on tergites VII–VIII; interstices lacking microsculpture. Male. Sternite VII (Fig. 11) with posterior margin truncate at middle. Sternite VIII (Fig. 12) with triangular posterior excision. Nazeris xingmini Lin & Hu, sp. nov. Nazeris xingmini Lin & Hu, sp. nov. http://zoobank.org/C2BCCA84-E6E7-430C-BEE6-D7A6FA183751 Figs 1, 5–9, 27 Type material. Holotype: China: ♂: “China: Guangdong Prov., Shixing County, Chebaling N. R., 24°43'22"N, 114°15'22"E, 357 m, 19.viii.2020, Liang Tang leg.” Nazeris fauna of the Nanling Mountain Range 123 Figures 1–4. Male habitus of Nazeris spp 1 N. xingmini 2 N. huaiweni 3 N. meihuaae 4 N. lichongi. Scale bars: 1.0 mm. Figures 1–4. Male habitus of Nazeris spp 1 N. xingmini 2 N. huaiweni 3 N. meihuaae 4 N. lichongi. Scale bars: 1.0 mm. (SNUC). Paratypes: 1 ♂, same data as holotype; 1 ♂, 2 ♀♀, “China: Jiangxi Prov., Longnan County, Jiulianshan N. R., 24°30'59.23"N, 114°24'52.98"E, alt. 587 m, 16.viii.2020, Liang Tang leg.” (SNUC). (SNUC). Paratypes: 1 ♂, same data as holotype; 1 ♂, 2 ♀♀, “China: Jiangxi Prov., Longnan County, Jiulianshan N. R., 24°30'59.23"N, 114°24'52.98"E, alt. 587 m, 16.viii.2020, Liang Tang leg.” (SNUC). g g g Description. Body length 6.4–7.5 mm; forebody length 3.2–3.4 mm. g g Body (Fig. 1) dark brown; legs yellowish brown; antennae dark brown to light brown. Head (Fig. 5) 1.02–1.12 times as long as wide; punctation very dense, moderately coarse, non-umbilicate, interstices lacking microsculpture; postocular portion approxi mately 1.6–2.1 times as long as eye length. g g Pronotum (Fig. 5) 1.18–1.22 times as long as wide, 0.95–1.02 times as long and 0.83–0.91 times as broad as head; punctation non-umbilicate, moderately dense and as coarse as that of head; midline posteriorly with short and very narrow impunctate elevation; interstices lacking microsculpture. Elytra (Fig. 5) 0.59–0.67 times as long as wide, 0.50–0.57 times as long and 0.96– 1.09 times as broad as pronotum; punctation as dense as, and slightly coarser than that of pronotum; interstices lacking microsculpture. Abdomen with punctation dense and rather coarse on tergites III–V, dense and less coarse on tergite VI, moderately dense and fine on tergites VII–VIII; interstices lacking microsculpture. Male. Sternite VII (Fig. 6) with posterior margin truncate at middle. Sternite VIII (Fig. 7) with wide triangular posterior excision. Aedeagus (Figs 8, 9) well scle rotized; with ventral process narrowed near middle in ventral view, with U-shaped excision at apex in ventral view, with pair of wing-like basal laminae ventrally; dorso-lateral apophyses moderately strong, distinctly curved in ventral view, curved dorsally and slightly widened at apices in lateral view, extending beyond apex of ventral process. Nazeris xingmini Lin & Hu, sp. nov. Aedeagus (Figs 13, 14) with ventral process gradually narrowed in apicad half, with acute apex in ventral view or lateral view, with pair of wing-like basal laminae ventrally; dorso-lateral apophyses moder ately slender, slightly widened near middle and apices in ventral view, not reaching apex of ventral process. Distribution and habitat data. The species is known only from Nanling in north ern Guangdong (Fig. 27). The specimens were collected by sifted leaf litter at altitudes of 1270–1575 m. Nazeris fauna of the Nanling Mountain Range 125 Figures 5–9. Nazeris xingmini 5 forebody 6 male sternite VII 7 male sternite VIII 8 aedeagus in ventral i 9 d i l t l i S l b 1 0 (5) 0 5 (6 9) Figures 5–9. Nazeris xingmini 5 forebody 6 male sternite VII 7 male sternite VIII 8 aedeagus in ventral view 9 aedeagus in lateral view. Scale bars: 1.0 mm (5); 0.5 mm (6–9). 5 9 N i i i i 5 f b d 6 l i VII 7 l Figures 5–9. Nazeris xingmini 5 forebody 6 male sternite VII 7 male sternite VIII 8 aedeagus in ventral view 9 aedeagus in lateral view. Scale bars: 1.0 mm (5); 0.5 mm (6–9). Comparative notes. The new species is very similar to N. divisus Hu & Li, 2015 in general appearance, but can be separated by the wider and shallower posterior excision of male sternite VIII (Fig. 12), by the ventral process with acute apex (Fig. 13), and by the wider dorso-lateral apophyses of aedeagus (Fig. 13). Etymology. The species is named in honor of Huai-Wen Wang (Administration of Nanling National Nature Reserve) who helped a lot during our collection in Nanling. Nazeris meihuaae Lin & Hu, sp. nov. http://zoobank.org/B62BB393-C4B6-41EE-96A7-2B3E49F0C50A Figs 3, 15–21, 27 Type material. Holotype: China: ♂: “China: Guangdong Prov., Shixing County, Che baling N. R., 24°40'41.82"N, 114°10'20.42"E, 1067 m, 20.viii.2020, Liang Tang leg.” (SNUC). Paratypes: 5 ♂♂, 18 ♀♀, same data as holotype; 3 ♂♂, 8 ♀♀, same data, ex cept “872 m, 20.viii.2020”; 2 ♀♀, “China: Guangdong Prov., Shixing County, Chebaling Type material. Holotype: China: ♂: “China: Guangdong Prov., Shixing County, Che baling N. R., 24°40'41.82"N, 114°10'20.42"E, 1067 m, 20.viii.2020, Liang Tang leg.” (SNUC). Body (Fig. 3) reddish brown; antennae and legs yellowish brown. g Description. Body length 4.1–4.8 mm; forebody length 2.2–2.6 mm. N. R., 24°40'58"N, 114°10'14"E, 468–870 m, 24.vi.2020, Xia, Zhang, Yin and Lin leg.”; 6 ♂♂, 3 ♀♀, “China: Jiangxi Prov., Longnan County, Jiulianshan N. R., 24°30'10"N, 114°26'35"E, 795 m, 18.viii.2020, Liang Tang leg.”; 1 ♂, 2 ♀♀, “China: Jiangxi Prov., Longnan County, Jiulianshan N. R., 24°30'10.43"N, 114°26'35.28"E, leaf litter, sifted, 1253 m, 17.viii.2020, Liang Tang leg.”; 5 ♂♂, 1 ♀, “China: Jiangxi Prov., Longnan County, Jiulianshan, Huangniushi, 24°31'22.7"N, 114°25'3.6"E, 600–1000 m, 10.v.2021, C.-L. Zhou & C. Li leg.”; 3 ♀♀, “China: Jiangxi Prov., Longnan County, Jiulianshan, summit of Huangniushi, 24°30'53"N, 114°26'6.72"E, 1000–1230 m, 12.v.2021, Zhou and Li leg.”; 1 ♂, “China: Jiangxi Prov., Longnan County, Jiulianshan, summit of Huangniushi, 24°30'53"N, 114°26'6.72"E, 1,000–1,230 m, 12.v.2021, Zhou and Li leg.” (SNUC). g Description. Body length 4.1–4.8 mm; forebody length 2.2–2.6 mm. Body (Fig 3) reddish brown; antennae and legs yellowish brown Description. Body length 4.1–4.8 mm; forebody length 2.2–2.6 mm Body (Fig. 3) reddish brown; antennae and legs yellowish brown. Nazeris xingmini Lin & Hu, sp. nov. Paratypes: 5 ♂♂, 18 ♀♀, same data as holotype; 3 ♂♂, 8 ♀♀, same data, ex cept “872 m, 20.viii.2020”; 2 ♀♀, “China: Guangdong Prov., Shixing County, Chebaling Xiao-Bin Lin & Jia-Yao Hu / ZooKeys 1059: 117–133 (2021) 126 Figures 10–14. Nazeris huaiweni 10 forebody 11 male sternite VII 12 male sternite VIII 13 aedeagus in ventral view 14 aedeagus in lateral view. Scale bars: 1.0 mm (10); 0.5 mm (11–14). Xiao-Bin Lin & Jia-Yao Hu / ZooKeys 1059: 117–133 (2021) 126 Xiao Bin Lin & Jia Yao Hu / ZooKeys 1059: 117 133 (2021) 126 Figures 10–14. Nazeris huaiweni 10 forebody 11 male sternite VII 12 male sternite VIII 13 aedeagus in ventral view 14 aedeagus in lateral view. Scale bars: 1.0 mm (10); 0.5 mm (11–14). 4 h 0 f b d l 2 l Figures 10–14. Nazeris huaiweni 10 forebody 11 male sternite VII 12 male sternite VIII 13 aedeagus in ventral view 14 aedeagus in lateral view. Scale bars: 1.0 mm (10); 0.5 mm (11–14). N. R., 24°40'58"N, 114°10'14"E, 468–870 m, 24.vi.2020, Xia, Zhang, Yin and Lin leg.”; 6 ♂♂, 3 ♀♀, “China: Jiangxi Prov., Longnan County, Jiulianshan N. R., 24°30'10"N, 114°26'35"E, 795 m, 18.viii.2020, Liang Tang leg.”; 1 ♂, 2 ♀♀, “China: Jiangxi Prov., Longnan County, Jiulianshan N. R., 24°30'10.43"N, 114°26'35.28"E, leaf litter, sifted, 1253 m, 17.viii.2020, Liang Tang leg.”; 5 ♂♂, 1 ♀, “China: Jiangxi Prov., Longnan County, Jiulianshan, Huangniushi, 24°31'22.7"N, 114°25'3.6"E, 600–1000 m, 10.v.2021, C.-L. Zhou & C. Li leg.”; 3 ♀♀, “China: Jiangxi Prov., Longnan County, Jiulianshan, summit of Huangniushi, 24°30'53"N, 114°26'6.72"E, 1000–1230 m, 12.v.2021, Zhou and Li leg.”; 1 ♂, “China: Jiangxi Prov., Longnan County, Jiulianshan, summit of Huangniushi, 24°30'53"N, 114°26'6.72"E, 1,000–1,230 m, 12.v.2021, Zhou and Li leg.” (SNUC). Nazeris fauna of the Nanling Mountain Range 127 Figures 15–21. Nazeris meihuaae (15–19 specimen from Chebaling 20–21 specimen from Jiulianshan) 15 forebody 16 male sternite VII 17 male sternite VIII 18, 20 aedeagus in ventral view 19, 21 aedeagus in lateral view. Scale bars: 1.0 mm (15); 0.5 mm (16–21). Nazeris fauna of the Nanling Mountain Range Nazeris fauna of the Nanling Mountain Range 127 Figures 15–21. Nazeris meihuaae (15–19 specimen from Chebaling 20–21 specimen from Jiulianshan) 15 forebody 16 male sternite VII 17 male sternite VIII 18, 20 aedeagus in ventral view 19, 21 aedeagus in lateral view. Scale bars: 1.0 mm (15); 0.5 mm (16–21). Head (Fig. Nazeris xingmini Lin & Hu, sp. nov. 15) 0.97–1.03 times as long as wide; punctation very dense, moderately coarse, distinctly umbilicate and partly confluent, interstices lacking microsculpture; postocular portion approximately 1.5–2.1 times as long as eye length. Head (Fig. 15) 0.97–1.03 times as long as wide; punctation very dense, moderately coarse, distinctly umbilicate and partly confluent, interstices lacking microsculpture; postocular portion approximately 1.5–2.1 times as long as eye length. Pronotum (Fig. 15) 1.05–1.23 times as long as wide, 0.91–1.10 times as long and 0.83–0.87 times as broad as head; punctation non-umbilicate, moderately dense and as coarse as that of head; midline posteriorly with short and very narrow impunctate elevation; interstices lacking microsculpture. Elytra (Fig. 15) 0.61–0.75 times as long as wide, 0.54–0.66 times as long and 0.97–1.10 times as broad as pronotum; punctation as dense as, and slightly coarser than that of pronotum; interstices lacking microsculpture. Abdomen with punctation dense and rather coarse on tergites III–V, dense and less coarse on tergite VI, moderately dense and fine on tergites VII–VIII; interstices lacking microsculpture. Male. Sternite VII (Fig. 16) with posterior margin shallowly emarginate in the middle. Sternite VIII (Fig. 17) with wide triangular posterior excision. Aedeagus (Figs 18–21) with ventral process short, widened near middle in ventral view, with pair of Xiao-Bin Lin & Jia-Yao Hu / ZooKeys 1059: 117–133 (2021) 128 finger-like basal laminae ventrally; dorso-lateral apophyses distinctly curved and wid ened in apical third in ventral view, extending beyond apex of ventral process. Distribution and habitat data. The species is known from Chebaling in northern Guangdong and Jiulianshan in southern Jiangxi (Fig. 27). The specimens were col lected by sifting leaf litter at altitudes of 468–1253 m.h Comparative notes. This species is very similar in general appearance and aedeagal characters to N. pengzhongi Hu & Li, 2015, but can be separated by the finger-like basal laminae of the ventral process and the longer dorso-lateral apophyses of the aedeagus (Figs 18, 20). The new species is also similar in general appearance to N. rubidus and N. nan lingensis, but can be separated by the distinctly longer laminae of the ventral process and the apically wider dorso-lateral apophyses of the aedeagus (Figs 18, 20). Compared with the holotype from Chebaling, Guangdong (Figs 18, 19), the specimens from Jiulianshan, Jiangxi (Figs 20, 21) display a slightly shorter ventral process and slightly narrower apices of the dorso-lateral apophyses of the aedeagus. Nazeris xingmini Lin & Hu, sp. nov. Based on the similar general appearance and male sternites, these aedeagal differences are treated as intraspecific variation.h fi Etymology. The species is named in honor of Mei-Hua Xia, who collected some of the type specimens. Nazeris lichongi Lin & Hu, sp. nov. http://zoobank.org/8DFC4E1C-104B-47C5-9511-5CB6B1C0662D Figs 4, 22–27 Type material. Holotype: China: ♂: “China: Hunan Prov., Yongzhou County, Du pangling N. R., 25°26'12.45"N, 111°20'23.29"E, 448 m, 29.viii.2020, sifted, Chong Li leg.” (SNUC). g Description. Body length 4.7 mm; forebody length 2.4 mm. Body (Fig. 4) reddish brown; antennae and legs yellowish brown. Head (Fig. 22) 0.97 times as long as wide; punctation very dense, moderately coarse, distinctly umbilicate and partly confluent, interstices lacking microsculpture; postocular portion approximately 1.6 times as long as eye length. Pronotum (Fig. 22) 1.17 times as long as wide, as long as and 0.83 times as broad as head; punctation non-umbilicate, moderately dense and as coarse as that of head; midline posteriorly with short and very narrow impunctate elevation; interstices lack ing microsculpture. Elytra (Fig. 22) 0.77 times as long as wide, 0.66 times as long and as broad as pro notum; punctation as dense as, and slightly coarser than that of pronotum; interstices lacking microsculpture. Abdomen with punctation dense and rather coarse on tergites III–V, dense and less coarse on tergite VI, moderately dense and fine on tergites VII–VIII; interstices lacking microsculpture. Male. Sternite VII (Fig. 23) with posterior margin shallowly emarginate in the middle. Sternite VIII (Fig. 24) with triangular posterior excision. Aedeagus (Figs 25, Nazeris fauna of the Nanling Mountain Range 129 Figures 22–26. Nazeris lichongi 22 forebody 23 male sternite VII 24 male sternite VIII 25 aedeagus in ventral view 26 aedeagus in lateral view. Scale bars: 1.0 mm (22); 0.5 mm (23–26). Nazeris fauna of the Nanling Mountain Range 129 Nazeris fauna of the Nanling Mountain Range 129 Figures 22–26. Nazeris lichongi 22 forebody 23 male sternite VII 24 male sternite VIII 25 aedeagus in ventral view 26 aedeagus in lateral view. Scale bars: 1.0 mm (22); 0.5 mm (23–26). 26) with broad ventral process, slightly widened in apical half, with round apex in ventral view, with pair of heart-like basal laminae; dorso-lateral apophyses slender, dis tinctly curved in ventral view, curved dorsally and slightly widened at apices in lateral view, extending beyond apex of ventral process. Distribution and habitat data. The species is known only from Dupangling in southern Hunan (Fig. 27). Nazeris xingmini Lin & Hu, sp. nov. The specimen was collected by sifting leaf litter at an altitude of 448 m.h The new species is similar in general appearance and aedeagal characters to N. rubidus and N. nanlingensis, but can be separated by the slightly dorsally curved dorso-lateral apophyses of the aedeagus in lateral view (Fig. 26), and by the heart-like basal laminae of the ventral process (Fig. 25). Etymology. The species is named in honor of Chong Li, who collected some of the type specimens. Xiao-Bin Lin & Jia-Yao Hu / ZooKeys 1059: 117–133 (2021) 130 Figure 27. Map showing the distribution of Nazeris in Nanling Mountain Range A N. inaequalis B N. rubidus C N. nanlingensis D N. gaoleii E N. huaiweni F N. obtortus G N. huapingensis H N. alatus I N. exilis J N. chenyanae K N. latilobatus L N. maoershanus M N. rugosus N N. yuyimingi O N. biacumi natus P N. yanzhuqii Q N. xingmini R N. meihuaae S N. lichongi. Figure 27. Map showing the distribution of Nazeris in Nanling Mountain Range A N. inaequalis B N. rubidus C N. nanlingensis D N. gaoleii E N. huaiweni F N. obtortus G N. huapingensis H N. alatus I N. exilis J N. chenyanae K N. latilobatus L N. maoershanus M N. rugosus N N. yuyimingi O N. biacumi natus P N. yanzhuqii Q N. xingmini R N. meihuaae S N. lichongi. Key to Nazeris species in Nanling mountain range 1 Head with non-umbilicate punctation (Fig. 5).............................................2 – Head with umbilicate punctation (Fig. 15)..................................................6 2 Body reddish brown, abdomen with fine microsculpture on all tergites......... ..................................................................... N. gaoleii Hu Luo & Li, 2018 – Body dark brown, abdomen lacking microsculpture....................................3 3 Pronotum with inconspicuous or lacking impunctate elevation in posterior half (Hu and Qiao 2019: 440, fig. 35); forebody length at most 2.9 mm....... ...................................................................N. yanzhuqii Hu & Qiao, 2019 – Pronotum with narrow impunctate elevation in posterior half (Fig. 5); fore body length at least 3.1 mm.........................................................................4 4 Ventral process of the aedeagus distinctly asymmetrical, dorso-lateral apophy ses not reaching apex of ventral process (Assing, 2014: 26, fig. 58)................ .......................................................................... N. inaequalis Assing, 2014 – Ventral process of the aedeagus symmetrical, dorso-lateral apophyses extend ing beyond apex of ventral process...............................................................5 5 Dorso-lateral apophyses of aedeagus moderately strong, with widened apex (Figs 8, 9).....................................................................N. xingmini sp. nov. Nazeris xingmini Lin & Hu, sp. nov. – Dorso-lateral apophyses of aedeagus slender, with acute apex (Hu and Li 2017: 337, figs 18, 19)...........................................N. alatus Hu & Li, 2017 Key to Nazeris species in Nanling mountain range Nazeris fauna of the Nanling Mountain Range 131 6 Body dark brown (Figs 1, 2), body length at least 6.1 mm, forebody length at least 3.2 mm................................................................................................7 – Body reddish brown (Figs 3, 4), body length at most 6.0 mm, forebody length at most 3.0 mm.........................................................................................11 7 Apex of ventral process of aedeagus divided into two branches in ventral view (Hu and Li 2017: 338, fig. 23)....................................................................8 – Apex of ventral process of aedeagus not divided into two branches in ventral view.............................................................................................................9 8 Sternite VII with posterior margin weakly protruding at middle (Hu and Li 2017: 338, fig. 21); ventral process of aedeagus with thin apical branches (Hu and Li 2017: 338, figs 23, 24); dorso-lateral apophyses of aedeagus slightly curved in lateral view (Hu and Li 2017: 338, fig. 24)...................N. exilis Hu & Li, 2017 – Sternite VII with posterior margin truncate at middle (Hu and Qiao 2019: 438, fig. 31); ventral process of aedeagus with wide apical branches (Hu and Qiao 2019: 438, figs 33, 34); dorso-lateral apophyses of aedeagus straight in lateral view (Hu and Qiao 2019: 438, fig. 34)............................................... .............................................................N. biacuminatus Hu & Qiao, 2019 9 Sternite VIII with rounded triangular posterior excision (Fig. 12); dorso-lat eral apophyses of aedeagus not reaching apex of ventral process (Fig. 13)....... .................................................................................... N. huaiweni sp. nov. – Sternite VIII with sharp, V-shaped posterior excision (Hu and Qiao 2019: 437, fig. 27); dorso-lateral apophyses of aedeagus extending slightly beyond apex of ventral process (Hu and Qiao 2019: 437, fig. 28)..........................10 10 Male sternite VII shallowly emarginate in the middle (Hu and Qiao 2019: 437, fig. 26); dorso-lateral apophyses of aedeagus widened near apex in ventral view (Hu and Qiao 2019: 437, fig. 28)........N. yuyimingi Hu & Qiao, 2019 – Male sternite VII not emarginate in the middle (Hu and Li 2017: 340, fig. 26); dorso-lateral apophyses of aedeagus not widened near apices in ventral view (Hu and Li 2017: 340, fig. 28)...............N. chenyanae Hu & Li, 2017 11 Head and pronotum with fine microsculpture (Hu and Qiao 2019: 436, figs 18, 19)............................................................N. Nazeris xingmini Lin & Hu, sp. nov. rugosus Hu & Qiao, 2019 – Head and pronotum lacking microsculpture..............................................12 12 Dorso-lateral apophyses of aedeagus extending to same level as apex of ventral process (Hu and Qiao 2019: 434, fig. 10)..........N. latilobatus Assing, 2016 – Dorso-lateral apophyses of aedeagus extending distinctly beyond apex of ven tral process.................................................................................................13 13 Ventral process of aedeagus nearly triangular, with narrow apex in ventral view (Assing 2016: 309, fig. 16).............................N. obtortus Assing, 2016 – Ventral process of aedeagus broad, with wide apex in ventral view.............14 14 Basal laminae of ventral process of aedeagus very long, more than half length of ventral process (Figs 18, 20).....................................N. meihuaae sp. nov. – Basal laminae of ventral process of aedeagus very short, much less than half length of ventral process.............................................................................15 Xiao-Bin Lin & Jia-Yao Hu / ZooKeys 1059: 117–133 (2021) 15 Ventral process of aedeagus with round apex in ventral view......................16 – Ventral process of aedeagus with truncate or emarginate apex in ventral view ..................................................................................................................17 16 Dorso-lateral apophyses of aedeagus curved ventrally in lateral view (Hu et al. 2018a: 176, fig. 13); ventral process with wing-like basal laminae (Hu et al. 2018a: 176, fig. 12)................................... N. rubidus Hu, Luo & Li, 2018 – Dorso-lateral apophyses of aedeagus curved dorsally in lateral view (Fig. 26); ventral process with heart-like basal laminae (Fig. 25).....N. lichongi sp. nov. 17 Ventral process of aedeagus in ventral view distinctly widened in apical half (Hu et al. 2018a: 177, fig. 17)............N. nanlingensis Hu, Luo & Li, 2018 – Ventral process of aedeagus in ventral view narrowed in apical half (Hu and Li 2017: 336, fig. 13).................................................................................18 18 Apex of ventral process of aedeagus nearly truncate in ventral view (Hu and Li 2017: 336, fig. 13); apices of dorso-lateral apophyses roundly widened in ventral view (Hu and Li 2017: 336, fig. 13)................................................... ...................................................................N. huapingensis Hu & Li, 2017 – Apex of ventral process of aedeagus with small semi-circular emargination in ventral view (Hu and Qiao 2019: 435, fig. 15); apices of dorso-lateral apo physes not widened in ventral view (Hu and Qiao 2019: 435, fig. 15)........... ..............................................................N. maoershanus Hu & Qiao, 2019 16 Acknowledgements We thank Chong Li, Zhong Peng, Liang Tang, Mei-Hua Xia, Jia-Min Yin, Wen-Xuan Zhang and Cheng-Lin Zhou (Shanghai, China) for collecting specimens, and Xing- Min Wang (South China Agricultural University) and Huai-Wen Wang (Administra tion of Nanling National Nature Reserve) for assisting us during our trip to Nanling. We are also most grateful to two anonymous reviewers for their helpful comments on an earlier version of the manuscript. References Assing V (2009) A revision of the Western Palaearctic species of Nazeris Fauvel, 1873 (Coleop tera: Staphylinidae: Paederinae). Deutsche Entomologische Zeitschrift 56(1): 109–131. https://doi.org/10.1002/mmnd.200900010 Assing V (2014) A revision of Nazeris. IV. New species from China, Taiwan, and Thailand, and additional records (Coleoptera: Staphylinidae: Paederinae). Stuttgarter Beiträge zur Naturkunde A Neue Serie 7: 11–32. Assing V (2016) A revision of Nazeris VIII. Five new species from China and additional records (Coleoptera: Staphylinidae: Paederinae). Linzer Biologische Beiträge 48(1): 301–315. Nazeris fauna of the Nanling Mountain Range 133 Fauvel A (1873) Faune Gallo-Rhénane ou species des insectes qui habitent la France, la Bel gique, la Hollande, le Luxembourg, la Prusse Rhénane, la Nassau et la Valais avec tableaux synoptiques et planches gravées. Tome 3. Livraison 4. Le Blanc-Hardel, Caen, 215–390. Hu JY, Li LZ, Zhao MJ (2012) Six new species of Nazeris Fauvel (Coleoptera, Staphylinidae, Paederinae) from Guangxi, South China. Zootaxa 3399: 35–44. https://doi.org/10.11646/ zootaxa.3399.1.3 Hu JY, Li LZ (2015) The Nazeris fauna of the Luoxiao Mountain Range, China. (Coleoptera, Staphylinidae, Paederinae). PLoS ONE 10(7): 1–21. https://doi.org/10.1371/journal. pone.0131942 Hu JY, Li LZ (2017) Four new species of Nazeris Fauvel in Guangxi, China (Coleoptera, Staphylinidae, Paederinae). Zootaxa 4312(2): 333–342. https://doi.org/10.11646/ zootaxa.4312.2.8 Hu JY, Luo YT, Li LZ (2018a) New species and record of Nazeris Fauvel in southern Chi na (Coleoptera, Staphylinidae, Paederinae). Zootaxa 4429(1): 173–180. https://doi. org/10.11646/zootaxa.4429.1.10 Hu JY, Liu YX, Li LZ (2018b) Two new species of Nazeris Fauvel in the Luoxiao Moun tain Range, China (Coleoptera, Staphylinidae, Paederinae). Zootaxa 4370(2): 180–188. https://doi.org/10.11646/zootaxa.4370.2.6 Hu JY, Qiao YJ (2019) Five new species of Nazeris Fauvel in Guangxi, China (Coleoptera, Staphyli nidae, Paederinae). Zootaxa 4543(3): 431–441. https://doi.org/10.11646/zootaxa.4543.3.8
https://openalex.org/W4250944668	https://www.qeios.com/read/16AAH7/pdf	English	null	HPV16-E7-T512	Definitions	2,020	cc-by	174	Qeios · Definition, February 2, 2020 Open Peer Review on Qeios HPV16-E7-T512 National Cancer Institute Qeios ID: 16AAH7 · https://doi.org/10.32388/16AAH7 Open Peer Review on Qeios Source National Cancer Institute. HPV16-E7-T512. NCI Thesaurus. Code C151958. Source National Cancer Institute. HPV16-E7-T512. NCI Thesaurus. Code C151958. A transcription activator-like effector nucleases (TALEN)-edited plasmid targeting human papillomavirus (HPV) type 16 (HPV16) epitope E7, with potential antineoplastic activity. Upon administration of TALEN-edited HPV16 E7 T512, the TALEN targets and binds to a specific site on genomic HPV16 E7, and cleaves the DNA sequences encoding E7. This causes double-strand DNA (dsDNA) breaks within the viral DNA, which prevents the transcription and translation of E7. Inhibition of E7 expression restores and increases the expression of tumor suppressor genes, such as retinoblastoma 1 (RB1), and promotes downstream signaling. Altogether, TALEN-based HPV editing increases cell cycle arrest, induces apoptosis, reduces viral replication and load, eliminates HPV and inhibits tumor cell proliferation of HPV-driven cancer cells. HPV16 E7 is an oncogene that is vital to viral function and HPV infection, and promotes carcinogenesis. 1/1
https://openalex.org/W1996599185	https://europepmc.org/articles/pmc3828297?pdf=render	English	null	Structural Insights into Saccharomyces cerevisiae Msh4–Msh5 Complex Function Using Homology Modeling	PloS one	2,013	cc-by	9,842	Introduction directly and connects with domain I through the uncharacterized domain II. These domains are also conserved in the MutSa and MutSb homologs. MutSc has homology with domains II, III, IV and V but lacks the N terminal domain I. Absence of domain I is expected to result in a large single channel of dimensions 70630 A˚ and inability to bind mismatch DNA during replication [1,3]. Instead the Msh4/5 proteins serve as pro-crossover factors during meiotic recombination. Physical, biochemical, genetic and cyto- logical studies have illuminated several aspects of Msh4–Msh5 function in meiotic crossing over as outlined below. The MutS homodimer in bacteria is involved in the repair of mismatches that occur during DNA replication [1]. The MutS homologs in eukaryotes form heterodimeric complexes with each other except Msh1. MutSa (Msh2–Msh6) heterodimeric complex is required for repair of mismatches and small (1–2 base) insertion/deletion (in/del) loops that arise during DNA replication [2]. The MutSb (Msh2–Msh3) complex repairs some single base in/del loops and loops that are two bases or larger [2]. MutSc (Msh4–Msh5) does not participate in repair of mismatches or in/ del loops [3]. Instead this complex plays a critical role in ensuring meiotic crossover formation and segregation of homologous chromosome pairs [3,4,5,6]. In Saccharomyces cerevisiae meiotic crossovers are initiated by the programmed introduction of ,140–170 DNA double strand breaks (DSBs) by the Spo11 protein in combination with accessory factors [8,9,10]. Physical assays performed in S. cerevisiae have provided molecular details into the sequence of events during repair of DSBs into crossover products [11,12,13,14,15]. DSBs are processed by endo and exonucleases to produce 39 single stranded DNA [16,17,18]. Dmc1 and Rad51 proteins form nucleoprotein filaments on the 39 single stranded DNA and catalyse strand invasion into homologous duplex DNA [19,20]. The nascent strand invasion matures into a single end invasion intermediate (SEI). For DSBs that are repaired as interfering crossovers, the SEI intermediate is thought to be stabilized by the Msh4–Msh5 The MutS homodimer has the shape of an oval disk with two channels of dimensions ,30620 and ,40620 A˚ with DNA passing through the larger channel [1,7]. Each subunit of the MutS protein comprises of five structural domains (Figure 1A). Domains I and IV bind mismatch DNA and the domain V contains ATP/ADP nucleotide binding sites. Domain I is also involved in mismatch recognition using the conserved Phe-X-Glu motif [2]. Abstract The Msh4–Msh5 protein complex in eukaryotes is involved in stabilizing Holliday junctions and its progenitors to facilitate crossing over during Meiosis I. These functions of the Msh4–Msh5 complex are essential for proper chromosomal segregation during the first meiotic division. The Msh4/5 proteins are homologous to the bacterial mismatch repair protein MutS and other MutS homologs (Msh2, Msh3, Msh6). Saccharomyces cerevisiae msh4/5 point mutants were identified recently that show two fold reduction in crossing over, compared to wild-type without affecting chromosome segregation. Three distinct classes of msh4/5 point mutations could be sorted based on their meiotic phenotypes. These include msh4/5 mutations that have a) crossover and viability defects similar to msh4/5 null mutants; b) intermediate defects in crossing over and viability and c) defects only in crossing over. The absence of a crystal structure for the Msh4–Msh5 complex has hindered an understanding of the structural aspects of Msh4–Msh5 function as well as molecular explanation for the meiotic defects observed in msh4/5 mutations. To address this problem, we generated a structural model of the S. cerevisiae Msh4–Msh5 complex using homology modeling. Further, structural analysis tailored with evolutionary information is used to predict sites with potentially critical roles in Msh4–Msh5 complex formation, DNA binding and to explain asymmetry within the Msh4–Msh5 complex. We also provide a structural rationale for the meiotic defects observed in the msh4/5 point mutations. The mutations are likely to affect stability of the Msh4/5 proteins and/or interactions with DNA. The Msh4–Msh5 model will facilitate the design and interpretation of new mutational data as well as structural studies of this important complex involved in meiotic chromosome segregation. Citation: Rakshambikai R, Srinivasan N, Nishant KT (2013) Structural Insights into Saccharomyces cerevisiae Msh4–Msh5 Complex Function Using Homology Modeling. PLoS ONE 8(11): e78753. doi:10.1371/journal.pone.0078753 Editor: Freddie SalsburyJr, Wake Forest University, United States of America Editor: Freddie SalsburyJr, Wake Forest University, United States of America Received June 30, 2013; Accepted September 20, 2013; Published November 14, 2013 Received June 30, 2013; Accepted September 20, 2013; Published November 14, 2013 Copyright: 2013 Rakshambikai et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: KTN is supported by IISER-TVM intramural funds and a Wellcome Trust-DBT India Alliance Intermediate Fellowship. Abstract This research is supported by Department of Biotechnology, Government of India. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: ns@mbu.iisc.ernet.in Introduction More generally the availability of a model for Msh4–Msh5 structure will facilitate the design of new mutational studies of the complex, interpretation of MSH4/5 polymorphism data in populations and a mechanistic understanding of Msh4–Msh5 function in meiotic crossing over. Consistent with the physical studies, genetic and biochemical data support the role of the Msh4–Msh5 complex in meiotic crossover formation. S. cerevisiae msh4D, msh5D mutants have strong defects in meiotic crossing over (2.5 fold decrease), spore viability (30–40%) and disjunction of homologous chromosomes [3,4,24]. Mutations of these genes in male and female mice cause chromosome pairing and synapsis defects and result in sterility [25,26,27]. In humans, non-disjunction of homologous chromo- somes during meiosis is associated with infertility, and congenital birth defects (such as Down syndrome) [28]. Biochemical studies have shown that the hMSH4–hMSH5 complex specifically binds to Holliday junction DNA and its progenitors that are key intermediates during crossover formation [21]. hMSH4–hMSH5 is thought to form multiple sliding clamps on these substrates and stabilize them. Biochemical data also suggest that the hMSH4 protein interacts with the MutL homologs hMLH1 and hMLH3 [29,30]. These data are supported by cell biological observations in mammals that suggest a subset of the Msh4–Msh5 complexes stabilizing Holliday junctions interact with the Mlh1, Mlh3 proteins [29,30,31,32,33]. The Mlh1 foci on pachytene chromo- somes are known to correspond to future crossover sites [34,35,36]. The Msh4–Msh5, Mlh1–Mlh3 complexes are part of the major crossover pathway in S. cerevisiae and mammals. A smaller subset of crossovers in these organisms is made through the Mus81-Mms4 pathway [24,37,38,39]. The central role played by the Msh4– Msh5 complex in meiotic crossing over encouraged a detailed mutational study of these proteins for meiotic crossover and chromosome segregation defects [6]. It was observed that the Msh5 protein is more sensitive to mutations and msh5 mutants showed more severe phenotypes compared to msh4 mutants. msh4/ 5 mutations were classified into three types based on crossover frequency and spore viability. These include a) mutations with spore viability and crossover frequency similar to that of msh4/5 Introduction These include residues at the putative interface of the Msh4–Msh5 complex, residues that may be involved in DNA binding and double mutations that may serve as compensatory mutations. Such information is useful to predict incompatibilities between segre- gating polymorphisms in MSH4 and MSH5 genes in populations. More generally the availability of a model for Msh4–Msh5 structure will facilitate the design of new mutational studies of the complex, interpretation of MSH4/5 polymorphism data in populations and a mechanistic understanding of Msh4–Msh5 function in meiotic crossing over. complex and form double Holliday junctions (dHJ) by capture of the second DSB end. Resolution of dHJ into crossovers is facilitated by Msh4–Msh5 in association with other repair factors [21,22,23]. These functions of Msh4–Msh5 complex are summa- rized in a simple model by Snowden et al., [21]. ‘‘null’’ mutations, b) mutations with intermediate defects in spore viability and crossing over and c) mutations with only crossover defects. Interestingly, mutations in equivalent positions in Msh4 and Msh5 ATPase and DNA binding domains were observed that had asymmetric effects on crossover frequency and spore viability. The aim of this study is to provide a structural basis for understanding Msh4–Msh5 function as well as molecular expla- nation for each of these msh4/5 mutations. As no crystal structure is available for the Msh4–Msh5 complex, homology modeling was used to generate a structural model for this complex using the hMSH2–hMSH6 crystal structure as the template. Homology modeling has proved to be useful in a number of cases where crystal structures are not available for a protein [40,41]. The modeling studies suggest that the msh4/5 mutations result in meiotic defects by two mechanisms: by affecting stability of the Msh4/5 proteins or interaction of the Msh4–Msh5 complex with the DNA. The model has not only been used to explain the structural basis of the meiotic defects observed in the mutations but also to propose further mutations that may be analyzed. These include residues at the putative interface of the Msh4–Msh5 complex, residues that may be involved in DNA binding and double mutations that may serve as compensatory mutations. Such information is useful to predict incompatibilities between segre- gating polymorphisms in MSH4 and MSH5 genes in populations. Introduction The DNA and nucleotide binding domains are connected by domain III. Domain III connects with domain IV November 2013 \| Volume 8 \| Issue 11 \| e78753 1 PLOS ONE \| www.plosone.org Yeast Msh4-Msh5 Complex Figure 1. Structure of E. coli MutS homodimer and a model of the S. cerevisiae Msh4–Msh5 complex. A) E. coli MutS homodimer showing five domains. Domain I is colored in magenta, domain II is colored in blue, domain III in green, domain IV in red and domain V in yellow. The connecting regions are colored in black. B) Cartoon representation of the modeled complex of S. cerevisiae Msh4 and Msh5. Msh4 is colored in purple and Msh5 in magenta. doi:10.1371/journal.pone.0078753.g001 Figure 1. Structure of E. coli MutS homodimer and a model of the S. cerevisiae Msh4–Msh5 complex. A) E. coli MutS homodimer showing five domains. Domain I is colored in magenta, domain II is colored in blue, domain III in green, domain IV in red and domain V in yellow. The connecting regions are colored in black. B) Cartoon representation of the modeled complex of S. cerevisiae Msh4 and Msh5. Msh4 is colored in purple and Msh5 in magenta. doi:10.1371/journal.pone.0078753.g001 ‘‘null’’ mutations, b) mutations with intermediate defects in spore viability and crossing over and c) mutations with only crossover defects. Interestingly, mutations in equivalent positions in Msh4 and Msh5 ATPase and DNA binding domains were observed that had asymmetric effects on crossover frequency and spore viability. The aim of this study is to provide a structural basis for understanding Msh4–Msh5 function as well as molecular expla- nation for each of these msh4/5 mutations. As no crystal structure is available for the Msh4–Msh5 complex, homology modeling was used to generate a structural model for this complex using the hMSH2–hMSH6 crystal structure as the template. Homology modeling has proved to be useful in a number of cases where crystal structures are not available for a protein [40,41]. The modeling studies suggest that the msh4/5 mutations result in meiotic defects by two mechanisms: by affecting stability of the Msh4/5 proteins or interaction of the Msh4–Msh5 complex with the DNA. The model has not only been used to explain the structural basis of the meiotic defects observed in the mutations but also to propose further mutations that may be analyzed. Results and Discussion Homology modeling of the Msh4–Msh5 complex Crystal structures of bacterial MutS, and eukaryotic MutSa and MutSb complexes are available [1,42,43,44]. The structural information is useful for providing explanations for the phenotypic effect of mutations in these proteins. It also enables prediction of important residues and domains that may compromise the protein function if mutated. We built a homology model of S. cerevisiae November 2013 \| Volume 8 \| Issue 11 \| e78753 PLOS ONE \| www.plosone.org 2 Yeast Msh4-Msh5 Complex Table 1. In/dels in Msh4 and Msh5 that could not be modeled. Protein In/del number Residues part of in/dels Msh4 1 1–58 2 172–174 3 272–275 4 688–696 5 846 6 901 Msh5 1 115 2 204 3 264 4 449 5 510 6 607–610 7 678–686 8 768–774 9 837–878 doi:10.1371/journal.pone.0078753.t001 Table 1. In/dels in Msh4 and Msh5 that could not be modeled. Protein In/del number Residues part of in/dels Msh4 1 1–58 2 172–174 3 272–275 4 688–696 5 846 6 901 Msh5 1 115 2 204 3 264 4 449 5 510 6 607–610 7 678–686 8 768–774 9 837–878 doi:10.1371/journal.pone.0078753.t001 Table 1. In/dels in Msh4 and Msh5 that could not be modeled. Msh4–Msh5 based on alignment with the hMSH2–hMSH6 complex (PDB code 2o8b) as the template. Alignments between the templates and targets obtained from automatic programs are considered unsatisfactory. This is because, automatic programs tend to introduce breaks in the middle of regular secondary structural elements, or align hydrophobic residues in the target with solvent exposed residues in the template among other reasons. Extensive manual intervention was therefore required to arrive at a high quality alignment suitable for the comparative modeling. Reasons for choosing hMSH2–hMSH6 complex as the template are discussed below. p The choice of using hMSH2–hMSH6 as the template was intentional based on the rigorous analysis of the quality of alignment between Msh4 & hMSH2, Msh4 & hMSH6, Msh5 & hMSH2, Msh5 & hMSH6 and Msh4 & hMSH3 pairs. Poor sequence identity of the order of 20% amongst all pairs meant that templates could not be decided purely based on sequence identity. Therefore all possible alignments were assessed to decide the template. Figure S1 provides a structure based alignment between Msh4, hMSH3 and hMSH6. This alignment has been generated by first structurally aligning hMSH3 and hMSH6 and then aligning Msh4 to this alignment. Results and Discussion From the alignment, considering insertions and deletions (in/dels) unique to hMSH3 or hMSH6 individually, hMSH3 shows more in/dels than hMSH6 when both proteins are aligned with Msh4. Particularly one of the deletions in hMSH3 which spans to about 25 residues is problematic considering that ab-initio modeling of such long stretches is likely to be rather inaccurate. In addition, a number of observations suggest recognition and repair of mispairs as well as general DNA binding occurs through conserved mechanism in hMSH2– hMSH6 and MutS, but is substantially different in case of hMSH2–hMSH3 [42,45,46,47,48,49,50]. Thus, we decided to use the hMSH2–hMSH6 template for modeling Msh4–Msh5 struc- ture. We modeled Msh4 with hMSH6 as the template and Msh5 with hMSH2 as the template. doi:10.1371/journal.pone.0078753.t001 residues between Msh4 and Msh5 and to analyze the interaction of the complex with Holliday junction DNA. These are discussed in further detail below. Asymmetry in the Msh4–Msh5 complex. Subunits of the MutS and MutSa complexes show asymmetry for mismatch binding and ATP hydrolysis [1,42,45,46,52,53,54,55]. A similar functional asymmetry has also been observed for the Msh4–Msh5 complex in the ATPase and DNA binding domains [6,56]. For example, both the ATP binding mutant alleles, msh4 G639A and msh5 G648A have spore viability similar to msh4/5D mutants (Table S3) [6]. But the ATP hydrolysis domain mutant msh4 R676W has wild-type spore viability while the equivalent mutation in msh5 R685W shows null phenotype. Similarly mutant alleles in the DNA binding domain, msh4 N532A, Y485A, L493A, and L553A have spore viabilities of 89, 95, 75 and 95%, respectively compared to equivalent mutations in msh5 D527A (30%), Y480A (67%), V488A (40%), and L548A (50%) respectively. Reasons for their asymmetric phenotypes are outlined below. The ATP binding mutations msh4 G639A and msh5 G648A are poorly tolerated in both Msh4 and Msh5 because of structural constraint. The Glycine residues have positive Q values which is not comfortably adopted by non-Glycine residues. This is also indicated by the high conservation of these residues. The ATP hydrolysis residue, Msh5 R685 is involved in main chain hydrogen bonding to stabilize the b sheet as shown in Figure 2A and is relatively more crucial than Msh4 R676. The msh5 R685W mutation is therefore poorly tolerated compared to the equivalent mutation msh4 R676W. In the DNA binding domain, Msh5 D527 is a solvent exposed residue and hence leads to instability when mutated to a hydrophobic residue such as Alanine. Results and Discussion The Msh5 Y480 in the DNA binding domain is involved in aromatic interactions as shown in Figure 2B which is not satisfied when mutated to Alanine. There are no such strong constraints for Msh4 N532 and Y485. Mutations of these Msh4 residues are therefore tolerated better. The hMSH4–hMSH5 complex specifically binds to Holliday junction DNA compared to linear ds DNA or other branched DNA structures [21]. Asymmetry in the DNA and ATP binding domains of the Msh4–Msh5 complex might reflect The curated alignment obtained has been provided in Figure S2. The decision on the choice of the template was influenced by the quality of alignments which are discussed below. It was observed that there existed more cases of in/dels within regular secondary structures in case of hMSH2 as template for Msh4 and hMSH6 as template for Msh5. The assessment of the alignment involved the use of structural environments around the sequence. Also, features such as solvent accessibility and secondary structure of residues were considered. For example, buried apolar residues replaced by buried polar residues and exposed polar residues being replaced by exposed apolar residues were commonly seen when Msh4 was aligned with hMSH2 and Msh5 with hMSH6 as the template. So, if the templates are swapped in modeling Msh4 and Msh5, not only was the quality of alignment poor, the modeled structure had large number of short contacts and collapsed during the energy-minimization steps. Two non-bonded atoms are said to be in short contact if their inter- atomic distance is too short in comparison to the classic contact criteria proposed by Ramachandran and coworkers [51]. The modeled structure of Msh4–Msh5 complex is represented as a cartoon in Figure 1B. The residues that could not be modeled due to in/dels in the alignment are indicated in Table 1. Correspon- dence between amino acid position in the Msh4, Msh5 protein sequences and the modeled structure are shown in Tables S1 and S2. Structural insights from the Msh4–Msh5 model Structural insights from the Msh4–Msh5 model The Msh4–Msh5 model was used to address asymmetry of the Msh4 and Msh5 subunits within the complex, to map interface PLOS ONE \| www.plosone.org PLOS ONE \| www.plosone.org November 2013 \| Volume 8 \| Issue 11 \| e78753 3 Yeast Msh4-Msh5 Complex Figure 2. Molecular interactions of Msh5 residues involved in ATP hydrolysis and DNA binding. A) Side chain interactions of R685 with Q705 and D250 stabilize the b-sheet preceding R685. B) Aromatic-aromatic interactions between Y480, Y486, Y530 and Y534. Interactions are shown in dashed green lines. doi:10.1371/journal.pone.0078753.g002 Figure 2. Molecular interactions of Msh5 residues involved in ATP hydrolysis and DNA binding. A) Side chain interactions of R685 with Q705 and D250 stabilize the b-sheet preceding R685. B) Aromatic-aromatic interactions between Y480, Y486, Y530 and Y534. Interactions are shown in dashed green lines. doi:10.1371/journal.pone.0078753.g002 Design of compensatory mutations. The Msh4–Msh5 model structure can be used to predict mutational changes that are compensatory. For example, in the putative interface region, Msh4 K819, D283 and K284 are involved in ionic interactions with Msh5 D732, H264 and D269 respectively. In principle if these residues are mutated such that the overall interaction is retained, for example K819D and D732K, the phenotype is expected to be close to the wild type. The ongoing efforts are directed towards design and generation of such mutants which will further our understanding of sequence-structure-function relation- ship of Msh4–Msh5 complex. different roles for Msh4 and Msh5 in recognition and binding of Holliday junction DNA. Binding to Holliday junctions. Volume of the cavity in the Msh4–Msh5 modeled structure was calculated to be 16676 A˚ 3. The images of the cavity are shown in Figure 3. We also estimated the volume of the Holliday junction in square planar conformation to be 6228 A˚ 3. The Holliday junction is known to take up stacked conformations in the presence of metal ions [57,58]. The volume measurements of the central cavity of the Msh4–Msh5 complex are consistent with the dimensions for a square planar geometry or other conformations of the Holliday junction. However it is not possible to decisively conclude the exact conformation or nature of binding of the Holliday junction to the Msh4–Msh5 complex on the basis of these studies. The probability of various Msh4/5 residues to bind to the DNA based on prediction by the Multi- VORFFIP (MV) server is indicated in Table S4 [59]. Structural insights from the Msh4–Msh5 model A probability of greater than 0.7 indicates a higher chance of being able to interact with the DNA. Out of thirty seven residues in Msh4 and ten residues in Msh5 showing a probability greater than 0.7, eight residues in Msh4 were mapped to be in the DNA binding domain. None of the residues having probability of greater than 0.7 in Msh5 map to the DNA binding domain. These results also suggest differences between Msh4 and Msh5 in DNA binding. DNA binding residues were also predicted on the basis of a structure based sequence alignment using Msh proteins from human and yeast. The DNA binding residues known in the literature and that predicted by the Multi-VORFFIP (MV) server were compared [42,44,59]. The two categories of residues are marked in Figure S3. Among the residues mentioned in literature, Msh4 H73 has a relatively high probability of 0.66 while other residues show a lower probability. In addition, two residues, Msh5 L155 and Msh5 I254 are conserved across the MSH family further confirming their role in DNA binding. From the residues predicted by the MV server, most residues are not well conserved and therefore seem to be specific to the Msh4–Msh5 complex. However, two residues, Msh5 D104 and Msh5 N347 are conserved in the MSH family and hence are likely to be important for DNA binding. Structural interpretation of msh4/5 mutant data Panel B shows the surface representation of the model. doi:10.1371/journal.pone.0078753.g003 Figure 3. Msh4–Msh5 model highlighting the binding cavity for the Holliday junction. Panel A shows the cartoon representation and Panel B shows the surface representation of the model. doi:10.1371/journal.pone.0078753.g003 c) Residue involved in aromatic-aromatic and cation-pi inter- actions interactions, cation-pi interactions, ionic interactions, hydrogen bonding and buried or solvent exposed residues. A significant proportion of interactions involve hydrogen bonding of main chain and side chain atoms or side chain and side chain atoms. A detailed information on the residues that constitute this network has been indicated in Table S3. W298 is involved in an aromatic-aromatic interaction with F445 in Msh5 as shown in Figure 5B. In the case of msh5 W298A, Alanine has an aliphatic side chain which cannot participate in such an interaction and hence affects stability. The Msh5 W298 is also involved in a cation pi interaction with Msh5 R312 (Figure 5C) which will be lost when Tryptophan is mutated to Alanine. Structural interpretation of msh4/5 mutant data The msh4/5 mutations are likely to cause meiotic defects by three main modes. The mutation may disrupt the structural integrity of local regions and hence affect the overall stability of the Msh4–Msh5 complex. The mutation may disrupt the interaction between the Msh4 and Msh5 proteins and prevent complex formation. Finally, the mutation may affect DNA binding by the Msh4/5 proteins. Twenty seven msh4/5 mutations cause significant meiotic defects (Table S3) [6]. The position of these residues in the Msh4–Msh5 complex has been indicated in Figure 4. From the Msh4–Msh5 homology model, seventeen of these mutations are predicted to affect structural stability of the individual proteins and hence that of the overall complex (Table 3). Six msh4/5 mutations are predicted to disrupt the interaction of the Msh4–Msh5 complex with DNA or destabilize the local structure around the DNA binding region. None of the mutations lie in the Msh4– Msh5 interface region. Meiotic defects of four msh4/5 mutations (msh5 D76A, D532A, D539A and msh4 L493A) could not be explained with the Msh4–Msh5 modeled structure. Yeast-two- hybrid analysis suggests sixteen mutations disrupt the Msh4–Msh5 complex (Table S3) [6]. However, for msh4 E276A, msh5 G648A and R685A although the mutations have effects on local stability it does not affect the interaction as indicated by the yeast-two-hybrid data (Table S3). Structural explanations for meiotic defects observed in individual msh4/5 point mutations are provided below. Msh4–Msh5 interface residues. Most probable Msh4– Msh5 interface residues were identified as discussed in Materials and Methods. They include residues within the ATPase domain in Msh4 (745–820 aa) and Msh5 (815–845 aa). The information of the residues that are involved in interaction with these residues and the nature of interaction have been indicated in Table 2. Null Mutations. Nine mutations, two in MSH4 and seven in MSH5 (msh4 D139A, G639A and msh5 W298A, D433A, V488A, D527A, G648A, Y661A, R685W) have meiotic defects similar to msh4/5D. In our model these involve residues stabilizing a-helical regions, residues part of the left handed a helical region of the Ramachandran map, residues involved in aromatic-aromatic November 2013 \| Volume 8 \| Issue 11 \| e78753 PLOS ONE \| www.plosone.org 4 Yeast Msh4-Msh5 Complex Figure 3. Msh4–Msh5 model highlighting the binding cavity for the Holliday junction. Panel A shows the cartoon representation and Panel B shows the surface representation of the model. doi:10.1371/journal.pone.0078753.g003 Figure 3. Msh4–Msh5 model highlighting the binding cavity for the Holliday junction. a) Residues stabilizing a-helical regions Msh4 D139 and Msh5 D433 are involved in stabilizing a- helical regions. The Msh4 D139 serves as an N-cap residue to stabilize a helix four residues downstream. In the case of Msh5, N430 is not a good initiator of the a-helix and hence the D433 stabilizes the structure by means of a hydrogen bond between side chain of D433 and main chain amide of N430 as shown in Figure 5A. Therefore, mutation to Alanine will disrupt these interactions thereby disrupting integrity of the local structure. d) d) Residue involved in ionic interactions Residue involved in ionic interactions The side chain of Msh5 R685 forms a salt bridge with side chain of Msh5 D250. In addition, R685 lies before a region of insertion in Msh5. The side chain of R685 is also involved in hydrogen bonding with main chain O of Q696 as shown in Figure 5D. The residues that form a part of this insertion (T688 to Q696) have high propensity to form a b-sheet. A mutation to Alanine disrupts this network of interaction and causes destabilization. The side chain of Msh5 R685 forms a salt bridge with side chain of Msh5 D250. In addition, R685 lies before a region of insertion in Msh5. The side chain of R685 is also involved in hydrogen bonding with main chain O of Q696 as shown in Figure 5D. The residues that form a part of this insertion (T688 to Q696) have high propensity to form a b-sheet. A mutation to Alanine disrupts this network of interaction and causes destabilization. b) Residues with conformations in left handed a-helical region of the Ramachandran map e) Residues that are buried or solvent exposed Residues that are buried or solvent exposed Msh5 G648 and Msh4 G639 are two residues with a positive ø dihedral angle. These angles are accommodated only in the case of Glycine due to the lack of side chain. When these residues are mutated to Alanine with this combination of ø and Y angles the residues experience short contacts involving their side chains and hence destabilize the structure. Msh5 G648 and Msh4 G639 are two residues with a positive ø dihedral angle. These angles are accommodated only in the case of Glycine due to the lack of side chain. a) Residues stabilizing a-helical regions There are nine such mutations, four in MSH4 and five in MSH5 (msh4 Y143A, F194A, R456A, L493A and msh5 R436A, Y480A, D532A, L548A, D680A). These residues are involved in aromatic-aromatic interactions, cation-pi interactions and ionic interactions. A few residues are involved in stabilizing the DNA binding region. However, in two cases msh4 L493A and msh4 D532A no explanation could be provided on the basis of the modeled structure. This is mainly due to the high sequence variation between the hMSH2–hMSH6 template and the Msh4–Msh5 model in this region. Msh4 Y143, F194 and Msh5 Y480 are involved in aromatic- aromatic interactions with surrounding aromatic residues within a distance of 6 A˚ as shown for Y480 in Figure 6A. These interactions are disrupted when mutated to Alanine. Msh5 L548 is involved in tight packing which is lost when mutated to Alanine which has a smaller side chain. Msh4 R456 and Msh5 R436 are proximal to the DNA binding region as shown for R456 in Figure 6B. Hence mutation to Alanine will affect these interactions and stability of binding. The Msh5 D680 is involved in salt bridge formation with the side chains of K681 and K716 which is lost in the msh5 D680A mutant. Mutations with only crossover defect. These mutations are deviant from the wild type only with respect to recombination frequency and have been described previously as msh4/5-t mutations. There are 9 such mutations distributed in both MSH4 and MSH5 (msh4 E276A, F491A, N532A, R676W and msh5 D76A, D250A, S416A, Y486A, D539A). These residues are involved in aromatic-aromatic interactions (Msh5 Y486), cation-pi interactions (Msh4 F491) and ionic interactions (Msh4 R676, Msh5 D250) with surrounding residues in 6 A˚ radius. Msh4 E276 and N532 are involved in a tight packing which may be disturbed doi:10.1371/journal.pone.0078753.t002 g p g y Figure 4. Representation of twenty seven msh4/5 mutations on the model of the Msh4–Msh5 complex. Msh4 is coloured in violet and Msh5 is coloured in grey. Msh4/5 residues whose mutations cause null phenotype, intermediate defects in crossing over and viability or only crossover defects are represented as blue, green and red spheres respectively. 1–10 indicate Msh4 residues and 11–27 are Msh5 residues. doi:10.1371/journal.pone.0078753.g004 Figure 4. Representation of twenty seven msh4/5 mutations on the model of the Msh4–Msh5 complex. Msh4 is coloured in violet and Msh5 is coloured in grey. a) Residues stabilizing a-helical regions When these residues are mutated to Alanine with this combination of ø and Y angles the residues experience short contacts involving their side chains and hence destabilize the structure. Msh5 V488 and Y661 are buried and also very tightly packed amongst the surrounding residues as shown for V488 in Figure 5E. This stabilization is disturbed when the residue is mutated to an Alanine as it creates a void in the region. Msh5 D527 is a solvent exposed residue. A local destabilization is caused when this is mutated to a hydrophobic residue such as Alanine. November 2013 \| Volume 8 \| Issue 11 \| e78753 PLOS ONE \| www.plosone.org 5 Yeast Msh4-Msh5 Complex Table 2. Residues predicted to occur at the interface of the Msh4–Msh5 heterodimer. Protein Residue number Protein Interacting residue Msh4 H746 Msh5 K726, T728, I730 Msh4 I753 Msh5 L737 Msh4 I793 Msh5 F737 Msh4 V795 Msh5 F699, L700 Msh4 I799 Msh5 L700, A707 Msh4 P802 Msh5 L710 Msh4 I804 Msh5 L710, I740 Msh4 I811 Msh5 L736, I740 Msh4 A818 Msh5 I845 Msh4 K819 Msh5 D732 Msh5 H264 Msh4 D283 Msh5 D269 Msh4 K284 Msh5 S819 Msh4 D722 Msh5 G821 Msh4 D722 Msh5 A825 Msh4 M723 Msh5 V827 Msh4 K724 Msh5 C828 Msh4 G690,K724 Msh5 L830 Msh4 M693, A697, L700 Msh5 I834 Msh4 L700, A729, V730 Msh5 A838 Msh4 V726, A729 Msh5 L841 Msh4 F724, L728, I753 Msh5 I845 Msh4 A818 doi:10.1371/journal.pone.0078753.t002 Table 2. Residues predicted to occur at the interface of the Msh4–Msh5 heterodimer. Mutations with intermediate defects in crossing over and viability. In terms of severity of phenotypes, these set of mutations second the null mutations. There are nine such mutations, four in MSH4 and five in MSH5 (msh4 Y143A, F194A, R456A, L493A and msh5 R436A, Y480A, D532A, L548A, D680A). These residues are involved in aromatic-aromatic interactions, cation-pi interactions and ionic interactions. A few residues are involved in stabilizing the DNA binding region. However, in two cases msh4 L493A and msh4 D532A no explanation could be provided on the basis of the modeled structure. This is mainly due to the high sequence variation between the hMSH2–hMSH6 template and the Msh4–Msh5 model in this region. Mutations with intermediate defects in crossing over and viability. In terms of severity of phenotypes, these set of mutations second the null mutations. Conclusions The Msh4–Msh5 complex plays an important role in different stages of the meiotic recombination pathway. In the absence of a crystal structure for this complex, we built a homology model of the S. cerevisiae Msh4–Msh5 complex. The modeling studies suggest that Msh4 is most likely functionally similar to hMSH6 of the hMSH2–hMSH6 complex and likewise Msh5 is similar to hMSH2. The model also explains the functional asymmetry between Msh4 and Msh5 with respect to ATP and DNA binding mutations [6,21]. Together these observations imply distinct roles for the Msh4 and Msh5 subunits in the recognition and binding of DNA substrates analogous to the distinct role of subunits in the MutS and MutSa complexes. The volume measurements of the cavity formed by the Msh-Msh5 complex reveals it is sufficient in size to bind an unfolded Holliday junction. The model also predicts possible interface residues and DNA binding residues whose mutations are likely to affect the function of the Msh4– Msh5 complex. S. cerevisiae Msh4 and Msh5 proteins have been Mutations affecting interaction with DNA Mutations affecting interaction with DNA msh4 D139A, Y143A, Y194A, E276A, F491A, G639A, R676W msh5 D250A, W298A, D433A, Y480A, D527A, L548A, G648A, Y661A, D680A, R685W R456A, N532A S416A, R436A, Y480A, V488A analyzed by mutational studies [6]. The model of the Msh4–Msh5 heterodimer provides structural explanations for msh4/5 mutations affecting crossover frequency and spore viability. The model can also facilitate the design of new mutational studies, design of structure based inhibitors of the Msh4–Msh5 complex as well as predict the functional impact of polymorphisms in the MSH4, MSH5 genes. Such studies will be useful for understanding the mechanism of crossover formation by the Msh4–Msh5, Mlh1– Mlh3 pathway [11,22,60]. upon mutation to Alanine which has a smaller side chain. The Msh5 S416 is a residue proximal to DNA and hence may affect the stability of binding. Structural explanation could not be provided for the meiotic defects observed in mutations msh5 D76A and D539A. a) Residues stabilizing a-helical regions Msh4/5 residues whose mutations cause null phenotype, intermediate defects in crossing over and viability or only crossover defects are represented as blue, green and red spheres respectively. 1–10 indicate Msh4 residues and 11–27 are Msh5 residues. doi:10.1371/journal.pone.0078753.g004 November 2013 \| Volume 8 \| Issue 11 \| e78753 PLOS ONE \| www.plosone.org 6 Yeast Msh4-Msh5 Complex Materials and Methods This step further adjusts the side chain conformation by sampling various rotamers for each residue that correspond to lowest energy. Further, conformations of specific residues were manually refined to remove short contacts. The best structure was chosen on the basis of lowest energy and statistical parameters such as the DOPE Score. Intra and inter protein interactions were identified using the PIC (Protein Interactions Calculator) server [71]. PIC has standard algorithms encoded for identifying various kinds of interactions such as hydrogen bonding, van der Waals interaction and salt bridge. As it is locally developed in our group and easily available we used it in the current work. The interface residues identified were further pruned on the basis of the hMSH4– hMSH5 interface residues [21]. The Multi-VORFFIP (MV) server was used to predict DNA binding residues [58]. MV has been used in the current work as it is the state of the art method that was shown to be highly sensitive and quite effective compared to many other methods in predicting functional residues in proteins [58]. DNA binding residues in hMSH2–hMSH6 template were also considered to map residues proximal to DNA binding site. A structure based multiple sequence alignment of Msh1, Msh2, Msh3, Msh4, Msh5 from yeast was constructed using the EXPRESSO server and formatted using the ESPript server [72,73]. The h MSH2, hMSH3, hMSH6 sequences were used as reference since structures are available for the same. The DNA binding residues described in the literature were mapped on the alignment to obtain equivalences in Msh4 and Msh5 [42,44]. Likewise the residues predicted to have high probability of DNA binding were also analyzed for conservation in the structure based multiple sequence alignment of human and yeast MSH proteins. Residue solvent accessibility was calculated using NACCESS program which is the most commonly used method over a long time for calculating solvent accessibility [74]. [ ] The model of Msh4–Msh5 complex was built using MOD- ELLER (version 9.10) auto-model program with added energy optimization steps [67]. We preferred using MODELLER over other comparative modeling methods as MODELLER can accept the sequence alignment between the template and target from the user (which is extremely important in the current modeling work) and also it has in-built sensitive approaches as described below to maximize the accuracy of structural models generated. Materials and Methods The model of Msh4–Msh5 complex thus obtained was subjected to energy minimization using FoldX program which is one of the widely-used and highly effective programs for energy optimization [68,69,70]. This step further adjusts the side chain conformation by sampling various rotamers for each residue that correspond to lowest energy. Further, conformations of specific residues were manually refined to remove short contacts. The best structure was chosen on the basis of lowest energy and statistical parameters such as the DOPE Score. obtained from fold prediction and threading algorithms PHYRE and I-TASSER [61,62,63]. It is well known from the CASP (Critical Assessment of protein Structure Prediction) experiments that PHYRE and I-TASSER perform quite well, in general, compared to most other methods. PHYRE generates a profile (sequence finger-print) of the family of the query sequence and secondary structures of the sequence are predicted. The profile is then searched against a fold library to choose an appropriate template onto which the sequence is threaded. I-TASSER builds models on the basis of multiple-threading alignments and iterative template fragment assembly simulations. The use of hMSH2– hMSH6 as template was validated by both servers with e-values better than 10210 suggesting high confidence and that both complexes are likely to adopt the same fold. In addition, structural alignments of the S. cerevisiae Msh4, Msh5 protein sequences were also obtained from Bioinfo metaserver or 3D-Jury which uses the alignment information that is predicted consistently by other reliable servers [64]. We used such metaservers as they employ multiple methods and provide consensus and consistent results which are likely to be more accurate than the results from individual methods. The final alignment used for structural modeling is the result of refinement of alignments obtained from PHYRE, I-TASSER and 3D-Jury servers which were manually scrutinized for consistency with respect to conservation, secondary structure and solvent accessibility at various residue positions. The S. cerevisiae Msh2–Msh6 complex is also known to bind Holliday junction structures which further justifies its use as a template [65,66]. 3-D models of Msh4 and Msh5 are generated in the bound forms. The models were superimposed on the template structures which are available in the complex form. The model of Msh4–Msh5 complex thus obtained was subjected to energy minimization using FoldX program which is one of the widely-used and highly effective programs for energy optimization [68,69,70]. Materials and Methods Two possible templates (hMSH2–hMSH6 and hMSH2– hMSH3) are available for the modeling of Msh4–Msh5 complex [42,44]. The model of the MSH4–MSH5 complex was built using the crystal structure of hMSH2–hMSH6 complex as the template (Figure S4). Since the sequence identity between the target and template is only in the order of 20%, the choice of template complex was validated using structure based alignment methods. As obtaining accurate alignment is quite difficult if the sequence similarity between the target and template is low, we used multiple algorithms for the alignment. We also considered structural environment such as solvent accessibility, secondary structures and hydrogen bonding in the manual analysis and refinement of the alignment. We started off with considering alignments Figure 5. Molecular interactions of Msh4/5 residues whose mutations cause msh4/5D phenotype. A) Side chain interaction of Msh5 D433 with amide of N430. B) Aromatic-aromatic interaction between Msh5 W298 and F445. C) Cation pi interaction involving Msh5 W298 and R312. D) Ionic interaction between Msh5 R685 and D250 and hydrogen bonding between Msh5 R685 and Q696. Interactions are shown by dashed green lines. E) Tight packing of the Msh5 V488 residue. doi:10.1371/journal.pone.0078753.g005 Figure 5. Molecular interactions of Msh4/5 residues whose mutations cause msh4/5D phenotype. A) Side chain interaction of Msh5 D433 with amide of N430. B) Aromatic-aromatic interaction between Msh5 W298 and F445. C) Cation pi interaction involving Msh5 W298 and R312. D) Ionic interaction between Msh5 R685 and D250 and hydrogen bonding between Msh5 R685 and Q696. Interactions are shown by dashed green lines. E) Tight packing of the Msh5 V488 residue. doi:10.1371/journal.pone.0078753.g005 November 2013 \| Volume 8 \| Issue 11 \| e78753 PLOS ONE \| www.plosone.org 7 Yeast Msh4-Msh5 Complex Figure 6. Molecular interactions of Msh4/5 residues whose mutations cause intermediate defects in crossing over and viability. A) Aromatic-aromatic interactions between Msh5 Y480, Y486, Y530 and Y534. B) Proximity of Msh4 R456 to DNA. Interactions are shown by dashed green lines. doi:10.1371/journal.pone.0078753.g006 Figure 6. Molecular interactions of Msh4/5 residues whose mutations cause intermediate defects in crossing over and viability. A) Aromatic-aromatic interactions between Msh5 Y480, Y486, Y530 and Y534. B) Proximity of Msh4 R456 to DNA. Interactions are shown by dashed green lines. doi:10.1371/journal.pone.0078753.g006 3-D models of Msh4 and Msh5 are generated in the bound forms. The models were superimposed on the template structures which are available in the complex form. p (PDF) We are thankful to Prof. Eric Alani for discussions and comments on the manuscript. Figure S3 Structure based sequence alignment of Msh1–6 and hMSH2, hMSH3, hMSH6. Highly conserved positions are highlighted in red, positions with conservative substitutions are highlighted in a blue box with residues marked in red. DNA binding residues identified from literature survey are highlighted in a yellow box. Residues predicted by Multi- References 1. Obmolova G, Ban C, Hsieh P, Yang W (2000) Crystal structures of mismatch repair protein MutS and its complex with a substrate DNA. Nature 407: 703– 710. 1. Obmolova G, Ban C, Hsieh P, Yang W (2000) Crystal structures of mismatch repair protein MutS and its complex with a substrate DNA. Nature 407: 703– 710. 13. Borner GV, Kleckner N, Hunter N (2004) Crossover/noncrossover differenti- ation, synaptonemal complex formation, and regulatory surveillance at the leptotene/zygotene transition of meiosis. Cell 117: 29–45. 2. Kunkel TA, Erie DA (2005) DNA mismatch repair. Annu Rev Biochem 74: 681–710. p yg 14. Schwacha A, Kleckner N (1995) Identification of double Holliday junctions as intermediates in meiotic recombination. Cell 83: 783–791. 15. Allers T, Lichten M (2001) Differential timing and control of noncrossover and crossover recombination during meiosis. Cell 106: 47–57. 3. Ross-Macdonald P, Roeder GS (1994) Mutation of a meiosis-specific MutS homolog decreases crossing over but not mismatch correction. Cell 79: 1069– 1080. g 16. Garcia V, Phelps SE, Gray S, Neale MJ (2011) Bidirectional res 16. Garcia V, Phelps SE, Gray S, Neale MJ (2011) Bidirectional resection of DNA double-strand breaks by Mre11 and Exo1. Nature 479: 241–244. 4. Hollingsworth NM, Ponte L, Halsey C (1995) MSH5, a novel MutS homolog, facilitates meiotic reciprocal recombination between homologs in Saccharomy- ces cerevisiae but not mismatch repair. Genes Dev 9: 1728–1739. double-strand breaks by Mre11 and Exo1. Nature 479: 241–244. 17. Neale MJ, Pan J, Keeney S (2005) Endonucleolytic processin 17. Neale MJ, Pan J, Keeney S (2005) Endonucleolytic processing of protein-linked DNA double-strand breaks. Nature 436: 1053–1057. protein-linked DNA double-strand breaks. Nature 436: 1053–105 5. Novak JE, Ross-Macdonald PB, Roeder GS (2001) The budding yeast Msh4 protein functions in chromosome synapsis and the regulation of crossover distribution. Genetics 158: 1013–1025. 18. Mimitou EP, Symington LS (2009) DNA end resection: many nucleases make light work. DNA Repair 8: 983–995. 19. Bishop DK, Park D, Xu L, Kleckner N (1992) DMC1: a meiosis-specific yeast homolog of E. coli recA required for recombination, synaptonemal complex formation, and cell cycle progression. Cell 69: 439–456. 6. Nishant KT, Chen C, Shinohara M, Shinohara A, Alani E (2010) Genetic analysis of baker’s yeast Msh4–Msh5 reveals a threshold crossover level for meiotic viability. PLoS Genetics 6: e1001083. 20. Shinohara A, Ogawa H, Ogawa T (1992) Rad51 protein involved in repair and recombination in S. cerevisiae is a RecA-like protein. Author Contributions Conceived and designed the experiments: RR NS KTN. Performed the experiments: RR. Analyzed the data: RR NS KTN. Wrote the paper: RR NS KTN. Materials and Methods MOD- ELLER generates 3-D models of a protein on the basis of known 3-D structures of one or more proteins which are known to be related to the target. Structural restraints for model building are generated using the template structure(s) and expressed in terms of probability density functions. In the current work, structural models of Msh4 and Msh5 were separately generated using appropriate templates, as discussed already, with template structures available in the complexed form. This ensures that the Homologues of Msh4 and Msh5 of not necessarily known 3-D structure were obtained by use of PSI-BLAST queried against the UNIPROT-SPROT database [75,76]. Multiple sequence align- ments were performed using the MAFFT-LINSI program [77]. The alignment was used to compute extent of conservation of November 2013 \| Volume 8 \| Issue 11 \| e78753 PLOS ONE \| www.plosone.org 8 Yeast Msh4-Msh5 Complex VORFFIP (MV) server to have high probability of DNA binding are highlighted in a black box. (PDF) residues mutated using entropic method with Scorecons server [78]. The homology search and alignment approaches used in the current work have been benchmarked in our group and are well- known to be highly sensitive and accurate. Figure S4 Protocol used to model Msh4–Msh5 struc- ture. The sequences were submitted to three structure prediction servers and the alignment obtained from all these were compiled and manually curated to get final alignment which was provided to MODELER to build the structure. The structure was then energy minimized and checked for stereo-chemical quality (removal of short contacts). (PDF) Figure S4 Protocol used to model Msh4–Msh5 struc- ture. The sequences were submitted to three structure prediction servers and the alignment obtained from all these were compiled and manually curated to get final alignment which was provided to MODELER to build the structure. The structure was then energy minimized and checked for stereo-chemical quality (removal of short contacts). (PDF) Volume measurements were made using the 3V server, which is known to use a robust algorithm to determine the size of cavities. We used 3V to calculate the size of the cavity formed by the Msh4–Msh5 heterodimer and also to calculate the volume of Holliday junction DNA [79]. Table S1 Sequence to structure mapping for Msh4. (XLSX) Table S2 Sequence to structure mapping for Msh5. (XLSX) Table S3 Explanation for meiotic defects observed in msh4/5 mutants based on the homology model. (DOCX) Figure S2 Alignment of Msh4 and Msh5 amino acid sequences with the hMSH2–hMSH6 complex (PDB code 2o8b). The alignment was used to model the Msh4–Msh5 complex structure. Msh4/5 residues whose mutations cause null phenotype, intermediate defects in crossing over and viability or only crossover defects are shown in blue, green and red boxes respectively. Table S4 Probability of each residue to bind/stabilize DNA. Supporting Information Figure S1 Alignment of hMSH6, hMSH3 and Msh4 sequences. 2o8bB corresponds to the hMSH6 from the hMSH2–hMSH6 complex and 3THY corresponds to hMSH3 from the hMSH2–hMSH3 complex. Highlighted in yellow are the deletions that are unique to hMSH3. Table S1 Sequence to structure mapping for Msh4. (XLSX) References Nat Rev Mol Cell Biol 5: 937–944. 59. Segura J, Jones PF, Fernandez-Fuentes N (2012) A holistic in silico approach to predict functional sites in protein structures. Bioinformatics 28: 1845–1850. 34. Baker SM, Plug AW, Prolla TA, Bronner CE, Harris AC, et al. (1996) Involvement of mouse Mlh1 in DNA mismatch repair and meiotic crossing over. Nat Genet 13: 336–342. 60. De Muyt A, Jessop L, Kolar E, Sourirajan A, Chen J, et al. (2012) BLM helicase ortholog Sgs1 is a central regulator of meiotic recombination intermediate metabolism. Mol Cell 46: 43–53. 35. Barlow AL, Hulten MA (1998) Crossing over analysis at pachytene in man. Eur J Hum Genet 6: 350–358. 61. Kelley LA, Sternberg MJ (2009) Protein structure prediction on the Web: a case study using the Phyre server. Nat Protoc 4: 363–371. 36. Cole F, Kauppi L, Lange J, Roig I, Wang R, et al. (2012) Homeostatic control of recombination is implemented progressively in mouse meiosis. Nat Cell Biol 14: 424–430. y g y 62. Roy A, Kucukural A, Zhang Y (2010) I-TASSER: a unified platform for automated protein structure and function prediction. Nat Protoc 5: 725–738. 37. de los Santos T, Hunter N, Lee C, Larkin B, Loidl J, et al. (2003) The Mus81/ Mms4 endonuclease acts independently of double-Holliday junction resolution to promote a distinct subset of crossovers during meiosis in budding yeast. Genetics 164: 81–94. 63. Zhang Y (2008) I-TASSER server for protein 3D structure prediction. BMC Bioinformatics 9: 40. 64. Von Grotthuss M, Pas J, Wyrwicz L, Ginalski K, Rychlewski L (2003) Application of 3D-Jury, GRDB, and Verify3D in fold recognition. Proteins 53: 418–423. 38. Guillon H, Baudat F, Grey C, Liskay RM, de Massy B (2005) Crossover and noncrossover pathways in mouse meiosis. Mol Cell 20: 563–573. 65. Alani E, Lee S, Kane MF, Griffith J, Kolodner RD (1997) Saccharomyces cerevisiae MSH2, a mispaired base recognition protein, also recognizes Holliday junctions in DNA. J Mol Biol 265: 289–301. 39. Holloway JK, Booth J, Edelmann W, McGowan CH, Cohen PE (2008) MUS81 generates a subset of MLH1-MLH3-independent crossovers in mammalian meiosis. PLoS Genetics 4: e1000186. j J 66. Marsischky GT, Lee S, Griffith J, Kolodner RD (1999) ‘Saccharomyces cerevisiae MSH2/6 complex interacts with Holliday junctions and facilitates their cleavage by phage resolution enzymes. J Biol Chem 274: 7200–7206. 40. Marti-Renom MA, Stuart AC, Fiser A, Sanchez R, Melo F, et al. References Cell 69: 457–470. 7. Jiricny J (2000) Mismatch repair: the praying hands of fidelity. Curr Biol 10: R788–790. 21. Snowden T, Acharya S, Butz C, Berardini M, Fishel R (2004) hMSH4–hMSH5 recognizes Holliday Junctions and forms a meiosis-specific sliding clamp that embraces homologous chromosomes. Mol Cell 15: 437–451. 8. Keeney S, Giroux CN, Kleckner N (1997) Meiosis-specific DNA double-strand breaks are catalyzed by Spo11, a member of a widely conserved protein family. Cell 88: 375–384. 22. Nishant KT, Plys AJ, Alani E (2008) A mutation in the putative MLH3 endonuclease domain confers a defect in both mismatch repair and meiosis in Saccharomyces cerevisiae. Genetics 179: 747–755. 9. Bergerat A, de Massy B, Gadelle D, Varoutas PC, Nicolas A, et al. (1997) An atypical topoisomerase II from Archaea with implications for meiotic recombination. Nature 386: 414–417. 23. Zakharyevich K, Tang S, Ma Y, Hunter N (2012) Delineation of joint molecule resolution pathways in meiosis identifies a crossover-specific resolvase. Cell 149: 334–347. 10. Maleki S, Neale MJ, Arora C, Henderson KA, Keeney S (2007) Interactions between Mei4, Rec114, and other proteins required for meiotic DNA double- strand break formation in Saccharomyces cerevisiae. Chromosoma 116: 471– 486. 24. Argueso JL, Wanat J, Gemici Z, Alani E (2004) Competing crossover pathways act during meiosis in Saccharomyces cerevisiae. Genetics 168: 1805–1816. 11. Zakharyevich K, Ma Y, Tang S, Hwang PY, Boiteux S, et al. (2010) Temporally and biochemically distinct activities of Exo1 during meiosis: double-strand break resection and resolution of double Holliday junctions. Mol Cell 40: 1001–1015. 25. Kneitz B, Cohen PE, Avdievich E, Zhu L, Kane MF, et al. (2000) MutS homolog 4 localization to meiotic chromosomes is required for chromosome pairing during meiosis in male and female mice. Genes Dev 14: 1085–1097. 12. Hunter N, Kleckner N (2001) The single-end invasion: an asymmetric intermediate at the double-strand break to double-holliday junction transition of meiotic recombination. Cell 106: 59–70. 26. Edelmann W, Cohen PE, Kneitz B, Winand N, Lia M, et al. (1999) Mammalian MutS homologue 5 is required for chromosome pairing in meiosis. Nat Genet 21: 123–127. November 2013 \| Volume 8 \| Issue 11 \| e78753 9 PLOS ONE \| www.plosone.org Yeast Msh4-Msh5 Complex 51. Ramachandran GN, Sasisekharan V (1968) Conformation of polypeptides and proteins. Adv Protein Chem 23: 283–438. 27. de Vries SS, Baart EB, Dekker M, Siezen A, de Rooij DG, et al. References (1999) Mouse MutS-like protein Msh5 is required for proper chromosome synapsis in male and female meiosis. Genes Dev 13: 523–531. 52. Lamers MH, Perrakis A, Enzlin JH, Winterwerp HH, de Wind N, et al. (2000) The crystal structure of DNA mismatch repair protein MutS binding to a G x T mismatch. Nature 407: 711–717. , , J , p , , ( ) The crystal structure of DNA mismatch repair protein MutS binding to a G x T mismatch. Nature 407: 711–717. 28. Nagaoka SI, Hassold TJ, Hunt PA (2012) Human aneuploidy: mechanisms and new insights into an age-old problem. Nat Rev Genet 13: 493–504. 53. Martik D, Baitinger C, Modrich P (2004) Differential specificities and simultaneous occupancy of human MutSalpha nucleotide binding sites. J Biol Chem 279: 28402–28410. 29. Santucci-Darmanin S, Neyton S, Lespinasse F, Saunieres A, Gaudray P, et al. (2002) The DNA mismatch-repair MLH3 protein interacts with MSH4 in meiotic cells, supporting a role for this MutL homolog in mammalian meiotic recombination. Hum Mol Genet 11: 1697–1706. 54. Bjornson KP, Modrich P (2003) Differential and simultaneous adenosine di- and triphosphate binding by MutS. J Biol Chem 278: 18557–18562. 30. Santucci-Darmanin S, Walpita D, Lespinasse F, Desnuelle C, Ashley T, et al. (2000) MSH4 acts in conjunction with MLH1 during mammalian meiosis. FASEB J 14: 1539–1547. 55. Antony E, Khubchandani S, Chen S, Hingorani MM (2006) Contribution of Msh2 and Msh6 subunits to the asymmetric ATPase and DNA mismatch binding activities of Saccharomyces cerevisiae Msh2–Msh6 mismatch repair protein. DNA Repair 5: 153–162. 31. Kolas NK, Cohen PE (2004) Novel and diverse functions of the DNA mismatch repair family in mammalian meiosis and recombination. Cytogenet Genome Res 107: 216–231. 56. Snowden T, Shim KS, Schmutte C, Acharya S, Fishel R (2008) hMSH4– hMSH5 adenosine nucleotide processing and interactions with homologous recombination machinery. J Biol Chem 283: 145–154. 32. Kolas NK, Svetlanov A, Lenzi ML, Macaluso FP, Lipkin SM, et al. (2005) Localization of MMR proteins on meiotic chromosomes in mice indicates distinct functions during prophase I. J Cell Biol 171: 447–458. 57. Lilley DM (2000) Structures of helical junctions in nucleic acids. Q Rev Biophys 33: 109–159. 33. Svetlanov A, Cohen PE (2004) Mismatch repair proteins, meiosis, and mice: understanding the complexities of mammalian meiosis. Exp Cell Res 296: 71– 79. 58. Liu Y, West SC (2004) Happy Hollidays: 40th anniversary of the Holliday junction. References (2000) Comparative protein structure modeling of genes and genomes. Annu Rev Biophys Biomol Struct 29: 291–325. 67. Eswar N, Webb B, Marti-Renom MA, Madhusudhan MS, Eramian D, et al. (2007) Comparative protein structure modeling using MODELLER. In: John E . Coligan et al editors Current protocols in protein science. Chapter 2: Unit 2.9. p y 41. Sanchez R, Sali A (1998) Large-scale protein structure modeling of the Saccharomyces cerevisiae genome. Proc Natl Acad Sci USA 95: 13597–13602. 42. Warren JJ, Pohlhaus TJ, Changela A, Iyer RR, Modrich PL, et al. (2007) Structure of the human MutSalpha DNA lesion recognition complex. Mol Cell 26: 579–592. 68. Guerois R, Nielsen JE, Serrano L (2002) Predicting changes in the stability of proteins and protein complexes: a study of more than 1000 mutations. J Mol Biol 320: 369–387. 43. Tseng Q, Orans J, Hast MA, Iyer RR, Changela A, et al. (2011) Purification, crystallization and preliminary X-ray diffraction analysis of the human mismatch repair protein MutSbeta. Acta Crystallogr Sect F Struct Biol Cryst Commun 67: 947–952. 69. Schymkowitz JW, Rousseau F, Martins IC, Ferkinghoff-Borg J, Stricher F, et al. (2005) Prediction of water and metal binding sites and their affinities by using the Fold-X force field. Proc Natl Acad Sci USA 102: 10147–10152. 70. Schymkowitz J, Borg J, Stricher F, Nys R, Rousseau F, et al. (2005) The FoldX web server: an online force field. Nucleic Acids Res 33: W382–388. . Schymkowitz J, Borg J, Stricher F, Nys R, Rousseau F, et al. (2005) 44. Gupta S, Gellert M, Yang W (2012) Mechanism of mismatch recognition revealed by human MutSbeta bound to unpaired DNA loops. Nat Struct Mol Bio 19: 72–78. 71. Tina KG, Bhadra R, Srinivasan N (2007) PIC: Protein Interactions Calculator. Nucleic Acids Res 35: W473–476. 45. Bowers J, Sokolsky T, Quach T, Alani E (1999) A mutation in the MSH6 subunit of the Saccharomyces cerevisiae MSH2–MSH6 complex disrupts mismatch recognition. J Biol Chem 274: 16115–16125. 72. Armougom F, Moretti S, Poirot O, Audic S, Dumas P, et al. (2006) Expresso: automatic incorporation of structural information in multiple sequence alignments using 3D-Coffee. Nucleic Acids Res 34: W604–608. 46. Drotschmann K, Yang W, Brownewell FE, Kool ET, Kunkel TA (2001) Asymmetric recognition of DNA local distortion. Structure-based functional studies of eukaryotic Msh2–Msh6. J Biol Chem 276: 46225–46229. g g 73. References Gouet P, Courcelle E, Stuart DI, Metoz F (1999) ESPript: analysis of multiple sequence alignments in PostScript. Bioinformatics 15: 305–308. sequence alignments in PostScript. Bioinformatics 15: 305–30 74. Hubbard SJ, Thornton JM (1993) NACCESS, Computer Program. Department of Biochemistry and Molecular Biology, University College London. Al h l S dd S h ff AA h h l G d 47. Dufner P, Marra G, Raschle M, Jiricny J (2000) Mismatch recognition and DNA-dependent stimulation of the ATPase activity of hMutSalpha is abolished by a single mutation in the hMSH6 subunit. J Biol Chem 275: 36550–36555. of Biochemistry and Molecular Biology, University College L 75. Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, et al. (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res 25: 3389–3402. 48. Lee SD, Surtees JA, Alani E (2007) Saccharomyces cerevisiae MSH2–MSH3 and MSH2–MSH6 complexes display distinct requirements for DNA binding domain I in mismatch recognition. J Mol Biol 366: 53–66. 76. UniProt Consortium (2012) Reorganizing the protein space at the Universal Protein Resource (UniProt). Nucleic Acids Res 40: D71–75. 49. Goldfarb T, Alani E (2005) Distinct roles for the Saccharomyces cerevisiae mismatch repair proteins in heteroduplex rejection, mismatch repair and nonhomologous tail removal. Genetics 169: 563–574. 77. Katoh K, Misawa K, Kuma K, Miyata T (2002) MAFFT: a novel method for rapid multiple sequence alignment based on fast Fourier transform. Nucleic Acids Res 30: 3059–3066. g 50. Surtees JA, Alani E (2006) Mismatch repair factor MSH2–MSH3 binds and alters the conformation of branched DNA structures predicted to form during genetic recombination. J Mol Biol 360: 523–536. 78. Valdar WS (2002) Scoring residue conservation. Proteins 48: 227–241 79. Voss NR, Gerstein M (2010) 3V: cavity, channel and cleft volume calculator and extractor. Nucleic Acids Res 38: W555–562. November 2013 \| Volume 8 \| Issue 11 \| e78753 PLOS ONE \| www.plosone.org 10
https://openalex.org/W2991494685	http://publikationen.ub.uni-frankfurt.de/files/70017/1-s2.0-S0370269320303683-main.pdf	English	null	Global baryon number conservation encoded in net-proton fluctuations measured in Pb–Pb collisions at <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" altimg="si1.svg"><mml:msqrt><mml:mrow><mml:msub><mml:mrow><mml:mi>s</mml:mi></mml:mrow><mml:mrow><mml:mi mathvariant="normal">NN</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:msqrt><mml:mo linebreak="goodbreak" linebreakstyle="after">=</mml:mo><mml:mn>2.76</mml:mn></mml:math> TeV	Physics letters. B	2,020	cc-by	14,179	a b s t r a c t Experimental results are presented on event-by-event net-proton ﬂuctuation measurements in Pb–Pb collisions at √sNN = 2.76 TeV, recorded by the ALICE detector at the CERN LHC. These measurements have as their ultimate goal an experimental test of Lattice QCD (LQCD) predictions on second and higher order cumulants of net-baryon distributions to search for critical behavior near the QCD phase boundary. Before confronting them with LQCD predictions, account has to be taken of correlations stemming from baryon number conservation as well as ﬂuctuations of participating nucleons. Both effects inﬂuence the experimental measurements and are usually not considered in theoretical calculations. For the ﬁrst time, it is shown that event-by-event baryon number conservation leads to subtle long-range correlations arising from very early interactions in the collisions. Article history: Received 6 May 2020 Received in revised form 8 June 2020 Accepted 13 June 2020 Available online 18 June 2020 Editor: L. Rolandi © 2020 European Organization for Nuclear Research. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Funded by SCOAP3. lam behavior for the second cumulants of net-baryons, while the fourth cumulants of net-baryons from LQCD are signiﬁcantly below the corresponding Skellam baseline [11,12]. Phase transitions in strongly interacting matter can be ad- dressed by investigating the response of the system to exter- nal perturbations via measurements of ﬂuctuations of conserved charges such as baryon number or electric charge [1,2]. At LHC en- ergies there would be, for vanishing light quark masses (u and d quarks), a temperature-driven second order phase transition be- tween a hadron gas and a quark–gluon plasma [3]. For realistic quark masses this transition becomes a smooth cross over [4,5]. However, because of the small masses of the current quarks, one can still probe critical phenomena at LHC energies (vanishing baryon chemical potential) as reported in [6]. Indeed, recent LQCD calculations [4,5] exhibit a rather strong signal for the existence of a pseudo-critical chiral temperature of about 156 MeV. Moreover, this pseudo-critical temperature is in agreement with the chemi- cal freeze-out temperature as extracted by the analysis of hadron multiplicities [7,8] measured by the ALICE experiment. This implies that the strongly interacting matter created in central collisions of Pb nuclei at LHC energies freezes out very near the chiral phase transition line. Contents lists available at ScienceDirect Contents lists available at ScienceDirect Contents lists available at ScienceDirect Global baryon number conservation encoded in net-proton ﬂuctuations measured in Pb–Pb collisions at √sNN = 2.76 TeV .ALICE Collaboration ⋆ ⋆E-mail address: alice -publications @cern .ch. κ1(nB) = nB=−∞ nB P(nB) = ⟨nB⟩ https://doi.org/10.1016/j.physletb.2020.135564 0370-2693/© 2020 European Organization for Nuclear Research. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Funded by SCOAP3. Physics Letters B 807 (2020) 135564 Physics Letters B 807 (2020) 135564 0.1016/j.physletb.2020.135564 20 European Organization for Nuclear Research. Published by Elsevier B.V. This is an open access article under the CC BY license ommons.org/licenses/by/4.0/). Funded by SCOAP3. a b s t r a c t Two forward scintillator arrays (V0) are located on either side of the interaction point and cover the pseudorapidity intervals 2.8 < η < 5.1 and −3.7 < η < −1.7 [18]. A minimum-bias trigger condition is deﬁned by requiring a combination of hits in the two innermost layers of the ITS and coincidence in both V0 detectors. The event centrality is selected based on the signal amplitudes in the V0 detectors [18]. In the upper panel of Fig. 1 the centrality dependence of the eﬃciency-corrected second cumulants of net-protons is compared with the sum of the mean multiplicities (ﬁrst cumulants) of pro- tons and antiprotons. Also included are the ﬁrst and second cu- mulants of protons and antiprotons. The eﬃciency correction for the cumulants is performed by using proton and anti-proton eﬃ- ciencies in analytic formulas derived in [28,29] assuming binomial eﬃciency losses. The characteristics of the ALICE detector response and applied analysis procedure ensures that this assumption is ful- ﬁlled. The accuracy of the correction procedure was estimated to be on the percent level and is included in the systematic uncer- tainties. We note that possible corrections for volume ﬂuctuations such as discussed in [30,31] were not applied to the data since, at LHC energies, second cumulants of net protons, our main observ- able, are free from such effects [32]. The detectors used for track reconstruction are the Time Projec- tion Chamber (TPC) [19] and the Inner Tracking System (ITS) [20]. In order to keep the tracking eﬃciency as high as possible, only the TPC detector was used for particle identiﬁcation, while pre- cise vertex determination was performed with the ITS detector. The track selection criteria used are described in Section 3.1 of [21]. Charged particle tracks with at least 80 out of maximum of 159 speciﬁc energy loss (dE/dx) samples in the TPC were used in this analysis for the best particle identiﬁcation. Moreover, in or- der to suppress contributions of secondary particles from weak decays, the distance-of-closest-approach (DCA) of the extrapolated track to the primary collision vertex was taken to be less than 2 cm along the beam direction. In the transverse plane, a more re- strictive and transverse momentum (pT) dependent DCA cut of less than (0.018 cm + 0.035p−1.01 T ) with pT in GeV/c, was imposed [22]. The subsample approach was chosen to estimate the statisti- cal uncertainties of the reconstructed cumulants [21,33]. a b s t r a c t The IM counts proxies of particle multiplicities W j event-by-event (2) We note that, the ﬁrst and second cumulants correspond to the expected value and the variance of net-baryon distribution, re- spectively. The second cumulant can be represented as a sum of the corresponding cumulants for single baryons and antibaryons plus the correlation term for the joint probability distributions of baryons and antibaryons P(nB, nB) W j = n i=1 ρ j(xi) ρ(xi) , ρ(xi) = j ρ j(xi), (5) (5) where j stands for a particle type, xi denotes the measured values of dE/dx for a given track i and ρ j(x) is the inclusive dE/dx dis- tribution of particle type j within a speciﬁed phase space bin. The summation in Eq. (5) runs over all selected n tracks in the given event. The pure and mixed moments of the W j distributions were calculated by averaging over all events, leading to the moments of the true multiplicity distributions. κ2(nB) = κ2(nB) + κ2(nB) −2 nBnB −⟨nB⟩ nB , (3) (3) (nB) = κ2(nB) + κ2(nB) −2 nBnB −⟨nB⟩ nB , where the mixed moment nBnB is deﬁned as where the mixed moment nBnB is deﬁned as nBnB = ∞ nB=0 ∞ nB=0 nBnBP(nB,nB). (4) (4) The IM is based on input of moments of W j distributions and inclusive dE/dx ﬁts in bins of momentum and pseudorapidity. The dE/dx distributions were ﬁt with generalized Gaussian functions, taking into account non-Gaussian components of the experimen- tal dE/dx distributions. The ﬁts of inclusive distributions of dE/dx were performed separately for positively and negatively charged particles in 20 MeV/c momentum and 0.1 units of η bins. Equation (3) shows that, for vanishing correlations between baryons and antibaryons ( nBnB = ⟨nB⟩ nB ), the second cumulant of net-baryons is exactly equal to the sum of the corresponding second cumulants for baryons and antibaryons. The data presented below were obtained by analyzing about 13 × 106 minimum-bias (cf. [16] for deﬁnition) Pb–Pb events at a center-of-mass energy per nucleon–nucleon pair of √sNN = 2.76 TeV recorded by the ALICE detector [17] in the year 2010. a b s t r a c t Hence, the singularities arising from the second or- der phase transition can be captured by measuring ﬂuctuations of conserved charges such as the net-baryon number. Evaluated within the framework of the Hadron Resonance Gas (HRG), net- baryon distributions coincide with the Skellam distribution, which is the probability distribution of the difference of two random vari- ables, each generated from statistically independent Poisson distri- butions [9,10]. In fact, at the pseudo-critical temperature of 156 MeV, similar to the HRG framework, LQCD also predicts a Skel- Conserved quantities of course ﬂuctuate only in sub-regions of the available total phase space of the reaction. In statistical mechanics they are, hence, analyzed within the Grand Canonical Ensemble (GCE) [13] formulation, where only the average num- ber of net-baryons are conserved. In order to compare theoretical calculations within the GCE, such as the HRG [7] and LQCD [4,5], with experimental results, the data are analyzed in different accep- tance windows by imposing selection criteria on rapidity and/or transverse momentum of the detected particles. Indeed, if the se- lected acceptance window is too small, possible dynamical cor- relations will be washed out and the measured ﬂuctuations will approach the Poisson limit [14], implying that net-baryons will be distributed according to the Skellam distribution. Recently the effects of limited acceptance were studied [15]. There, it was investigated under which conditions net-baryon ﬂuc- tuations depend on the size of the acceptance. An obvious case is ﬂuctuations caused by correlations due to baryon number con- servation. To identify these and other long-range correlations it is interesting to perform the experimental analysis as a function of the acceptance size. The analysis is set up by providing the necessary deﬁnitions. Given the number of baryons (nB) and antibaryons (nB), the ﬁrst and second cumulants of the net-baryon probability distribution P(nB), with nB = nB −nB, are deﬁned as κ1(nB) = ∞ nB=−∞ nB P(nB) = ⟨nB⟩, (1) (1) ALICE Collaboration / Physics Letters B 807 (2020) 135564 2 κ2(nB) = ∞ nB=−∞ (nB −⟨nB⟩)2 P(nB) = (nB −⟨nB⟩)2 . (2) deal eﬃciently with the overlapping dE/dx distributions of pro- tons, kaons, pions and electrons considered in the present analysis. Their speciﬁc probability distributions were obtained by unfolding the moments of the measured multiplicity distributions for each particle species. a b s t r a c t In order to evaluate systematic uncertainties stemming from track selec- tion criteria, the applied selection ranges were varied around their nominal values. Other sources of systematic uncertainties originate from the parameters of the ρ j(xi) functions, entering Eq. (5). The latter was estimated by hypothesis testing using the Kolmogorov- Smirnov (K-S) statistics. For this purpose the parameters of the ρ j(xi) functions were varied by testing the null hypothesis that measured dE/dx distributions and ﬁt functions are similar within a given signiﬁcance level of 10% (cf. [21,33] for details). The ﬁnal systematic uncertainties on cumulants were computed by adding in quadrature the maximum systematic variations from individual contributions. T Since the energy loss of charged particles in the gas volume of the TPC depends explicitly on the particle momentum (p), the analysis was performed in momentum space. Correspondingly, the particle identiﬁcation (PID) procedure consists of correlating par- ticle momentum with the speciﬁc energy loss dE/dx. This allows the event-by-event ﬂuctuation analysis to be performed with high overall reconstruction eﬃciency of about 80% for protons, almost independent of the collision centrality. The latter is important be- cause small eﬃciencies induce Poisson ﬂuctuations. To ensure the best possible dE/dx resolution, the phase space coverage was re- stricted to 0.6 < p < 1.5 GeV/c and \|η\| < 0.8 for the present anal- ysis. The corresponding pT range at \|η\| = 0.8 is about 0.45 < pT < 1.12 GeV/c, which gradually approaches the used momentum range towards midrapidity. Moreover, a differential analysis is provided as function of η in the range η = 0.2 to 1.6. By their deﬁnition, cumulants are extensive quantities, i.e., are proportional to the system volume. This also explains the central- ity dependence of all cumulants, presented in the upper panel of Fig. 1. To remove the system size dependence, normalized cumu- lants R1 and R2 are introduced as R1 = κ2(np −np)/ < np + np >, R2 = κ2(np)/ < np > . (6) (6) In the middle and bottom panels of Fig. 1, deviations from unity are visible for both R1 and R2. Moreover, the amount of deviation for R2 is about twice as large compared to that of R1. The cumulants of net-protons were then reconstructed using the Identity Method (IM) [21,23–27]. a b s t r a c t This agreement is expected because of the small acceptance window as discussed above. For η > 0.8, deviations from the Skellam distribution are observed. The amount of deviation is small but signiﬁcant and in good agreement with the prediction assuming global baryon number conservation. The observed devi- ation is therefore consistent with the assumption of global baryon number conservation, i.e. conservation within the full phase space. On the other hand, local baryon number conservation may induce additional correlations between protons and antiprotons, which would lead to a further reduction of the measured κ2(np − np) [35]. In Fig. 2 the data are compared to the predictions from an analysis of effects of local baryon number conservation for dif- ferent values of correlation width ycorr between protons and an- tiprotons. Within the experimental uncertainties the data are best described with the assumption of global baryon number conserva- tion, which corresponds to the correlation width ycorr = 2\|ybeam\| but, within one standard deviation (1.56 for the last point at η = 1.6), the data are also consistent with a large correlation width of ycorr = 5 [35]. We ﬁnd that for ycorr = 4.5, with a 5% signiﬁcance level, the last point is not consistent with the experi- mental data. The results from the HIJING event generator (cf. blue solid line in Fig. 2), which can be described with ycorr = 2, and from a recent study reported in [41] would imply much stronger correlations between protons and antiprotons, not consistent with Inspection of Fig. 2 shows that, for small pseudorapidity ranges of \|η\| < 0.4 corresponding to η < 0.8, the experimentally mea- sured net-proton distributions closely follow a Skellam distribu- tion. This agreement is expected because of the small acceptance window as discussed above. For η > 0.8, deviations from the Skellam distribution are observed. The amount of deviation is small but signiﬁcant and in good agreement with the prediction assuming global baryon number conservation. The observed devi- ation is therefore consistent with the assumption of global baryon number conservation, i.e. conservation within the full phase space. On the other hand, by construction, for vanishing net-proton numbers, R1 should not contain any contributions from volume ﬂuctuations, i.e., the values of R1 obtained from the model should be consistent with unity [32]. In fact at LHC energies R1 becomes a strongly intensive quantity [34]. a b s t r a c t This method is designed to In order to shed light on these observations, the results are compared with predictions from a model constructed recently [32], ALICE Collaboration / Physics Letters B 807 (2020) 135564 3 3 Fig. 1. Measured second cumulants of net-proton distributions (red solid boxes) compared with the sum of the mean multiplicities (open squares). The second cu- mulants of single proton and antiproton distributions are presented with the ﬁlled and open circles, respectively. The ﬁrst cumulants of protons and antiprotons are hardly distinguishable because of the nearly equal mean numbers of protons and antiprotons at LHC energy. In the middle and bottom panels the normalized cu- mulants R1 and R2 are presented. The band visible in the bottom panel is the prediction for R2 in the presence of volume ﬂuctuations [32]. Fig. 2. Pseudorapidity dependence of the norm protons R1. Global baryon number conservation i dashed lines represent the predictions from the conservation [35]. The blue solid line, represents generator. non-central collisions, baryon transport taken into account, which is rather mo avoid the model dependence, the com for the central events and in the estim ter only produced baryons are used. In baryons are used in the pseudorapidity ported in [16,38–40]. Next, using HIJIN estimates were obtained for the total a in full phase space. The average numb i t th d ﬁiti f ( f E (7)) Fig. 1. Measured second cumulants of net-proton distributions (red solid boxes) compared with the sum of the mean multiplicities (open squares). The second cu- mulants of single proton and antiproton distributions are presented with the ﬁlled and open circles, respectively. The ﬁrst cumulants of protons and antiprotons are hardly distinguishable because of the nearly equal mean numbers of protons and antiprotons at LHC energy. In the middle and bottom panels the normalized cu- mulants R1 and R2 are presented. The band visible in the bottom panel is the prediction for R2 in the presence of volume ﬂuctuations [32]. Fig. 2. Pseudorapidity dependence of the normalized second cumulants of net- protons R1. Global baryon number conservation is depicted as the pink band. The dashed lines represent the predictions from the model with local baryon number conservation [35]. The blue solid line, represents the prediction using the HIJING generator. Fig. 2. Pseudorapidity dependence of the normalized second cumulants of net- protons R1. a b s t r a c t Global baryon number conservation is depicted as the pink band. The dashed lines represent the predictions from the model with local baryon number conservation [35]. The blue solid line, represents the prediction using the HIJING generator. non-central collisions, baryon transport to mid-rapidity has to be taken into account, which is rather model dependent. In order to avoid the model dependence, the comparison is performed only for the central events and in the estimate of the alpha parame- ter only produced baryons are used. In doing so, the number of baryons are used in the pseudorapidity range of \|η\| < 0.5 as re- ported in [16,38–40]. Next, using HIJING and AMPT simulations, estimates were obtained for the total average number of baryons in full phase space. The average number of protons np entering into the deﬁnition of α (cf. Eq. (7)) was taken from the current analysis for each pseudorapidity interval. Finally, using these val- ues of α, the pink band in Fig. 2 is calculated with Eq. (7). The ﬁnite width of the band reﬂects the difference between predictions of the two event generators. Fig. 1. Measured second cumulants of net-proton distributions (red solid boxes) compared with the sum of the mean multiplicities (open squares). The second cu- mulants of single proton and antiproton distributions are presented with the ﬁlled and open circles, respectively. The ﬁrst cumulants of protons and antiprotons are hardly distinguishable because of the nearly equal mean numbers of protons and antiprotons at LHC energy. In the middle and bottom panels the normalized cu- mulants R1 and R2 are presented. The band visible in the bottom panel is the prediction for R2 in the presence of volume ﬂuctuations [32]. in which participant ﬂuctuations are included following the analy- sis of the ALICE centrality selection [18]. Within uncertainties, the model predictions are fully consistent with the measured R2 val- ues, lending support to the interpretation that volume ﬂuctuations are at the origin of the observed deviation. This is also supported by the observation of a small structure observed in the 10–20% centrality class, where, compared to the ﬁrst two centrality classes, the centrality bin width is doubled. Inspection of Fig. 2 shows that, for small pseudorapidity ranges of \|η\| < 0.4 corresponding to η < 0.8, the experimentally mea- sured net-proton distributions closely follow a Skellam distribu- tion. a b s t r a c t It should be further noted that, for ALICE Collaboration / Physics Letters B 807 (2020) 135564 4 Education, Culture, Sports, Science and Technology (MEXT) and Japan Society for the Promotion of Science (JSPS) KAKENHI, Japan; Consejo Nacional de Ciencia y Tecnología (CONACYT), through Fondo de Cooperación Internacional en Ciencia y Tecnología (FON- CICYT) and Dirección General de Asuntos del Personal Academico (DGAPA), Mexico; Nederlandse Organisatie voor Wetenschappelijk Onderzoek (NWO), Netherlands; The Research Council of Norway, Norway; Commission on Science and Technology for Sustainable Development in the South (COMSATS), Pakistan; Pontiﬁcia Uni- versidad Católica del Perú, Peru; Ministry of Science and Higher Education and National Science Centre, Poland; Korea Institute of Science and Technology Information and National Research Foun- dation of Korea (NRF), Republic of Korea; Ministry of Education and Scientiﬁc Research, Institute of Atomic Physics and Ministry of Research and Innovation and Institute of Atomic Physics, Romania; Joint Institute for Nuclear Research (JINR), Ministry of Education and Science of the Russian Federation, National Research Centre Kurchatov Institute, Russian Science Foundation and Russian Foun- dation for Basic Research, Russia; Ministry of Education, Science, Research and Sport of the Slovak Republic, Slovakia; National Re- search Foundation of South Africa, South Africa; Swedish Research Council (VR) and Knut & Alice Wallenberg Foundation (KAW), Sweden; European Organization for Nuclear Research, Switzerland; Suranaree University of Technology (SUT), National Science and Technology Development Agency (NSDTA) and Oﬃce of the Higher Education Commission under NRU project of Thailand, Thailand; Turkish Atomic Energy Agency (TAEK), Turkey; National Academy of Sciences of Ukraine, Ukraine; Science and Technology Facilities Council (STFC), United Kingdom; National Science Foundation of the United States of America (NSF) and United States Department of Energy, Oﬃce of Nuclear Physics (DOE NP), United States of America. the experimental data. We note here that correlations arising from baryon or charge conservation have also been analyzed in the framework of balance functions [42,43]. Such an analysis could also shed interesting light on global vs. local baryon conservation. From the present results it is concluded that effects due to lo- cal baryon number conservation are not large, if present at all in second cumulants of net-protons. The large correlation length ob- served in the data implies that the normalized second cumulant R1 is determined by collisions in the very early phase of the Pb–Pb in- teraction [44]. Acknowledgements The ALICE Collaboration would like to thank all its engineers and technicians for their invaluable contributions to the construc- tion of the experiment and the CERN accelerator teams for the outstanding performance of the LHC complex. The ALICE Collab- oration gratefully acknowledges the resources and support pro- vided by all Grid centres and the Worldwide LHC Computing Grid (WLCG) collaboration. The ALICE Collaboration acknowledges the following funding agencies for their support in building and run- ning the ALICE detector: A. I. Alikhanyan National Science Labora- tory (Yerevan Physics Institute) Foundation (ANSL), State Commit- tee of Science and World Federation of Scientists (WFS), Armenia; Austrian Academy of Sciences, Austrian Science Fund (FWF): [M 2467-N36] and Nationalstiftung für Forschung, Technologie und Entwicklung, Austria; Ministry of Communications and High Tech- nologies, National Nuclear Research Center, Azerbaijan; Conselho Nacional de Desenvolvimento Cientíﬁco e Tecnológico (CNPq), Fi- nanciadora de Estudos e Projetos (Finep), Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) and Universidade Fed- eral do Rio Grande do Sul (UFRGS), Brazil; Ministry of Education of China (MOEC), Ministry of Science & Technology of China (MSTC) and National Natural Science Foundation of China (NSFC), China; Ministry of Science and Education and Croatian Science Foun- dation, Croatia; Centro de Aplicaciones Tecnológicas y Desarrollo Nuclear (CEADEN), Cubaenergía, Cuba; The Ministry of Education, Youth and Sports of the Czech Republic, Czech Republic; The Dan- ish Council for Independent Research Natural Sciences, the Villum Fonden and Danish National Research Foundation (DNRF), Den- mark; Helsinki Institute of Physics (HIP), Finland; Commissariat à l’Énergie Atomique (CEA), Institut National de Physique Nucléaire et de Physique des Particules (IN2P3) and Centre National de la Recherche Scientiﬁque (CNRS) and Région des Pays de la Loire, France; Bundesministerium für Bildung und Forschung (BMBF) and GSI Helmholtzzentrum für Schwerionenforschung GmbH, Ger- many; General Secretariat for Research and Technology, Ministry of Education, Research and Religions, Greece; National Research De- velopment and Innovation Oﬃce, Hungary; Department of Atomic Energy, Government of India (DAE), Department of Science and Technology, Government of India (DST), University Grants Com- mission, Government of India (UGC) and Council of Scientiﬁc and Industrial Research (CSIR), India; Indonesian Institute of Sciences, Indonesia; Centro Fermi - Museo Storico della Fisica e Centro Studi e Ricerche Enrico Fermi and Istituto Nazionale di Fisica Nucle- are (INFN), Italy; Institute for Innovative Science and Technology, Nagasaki Institute of Applied Science (IIST), Japanese Ministry of References [1] V. Koch, Hadronic ﬂuctuations and correlations, in: R. Stock (Ed.), Relativis- tic Heavy Ion Physics, 2010, http://materials .springer.com /lb /docs /sm _lbs _978 - 3 -642 -01539 -7 _20. [2] STAR Collaboration, L. Adamczyk, et al., Energy dependence of moments of net- proton multiplicity distributions at RHIC, Phys. Rev. Lett. 112 (2014) 032302, arXiv:1309 .5681 [nucl -ex]. [3] M.A. Stephanov, QCD phase diagram and the critical point, Prog. Theor. Phys. Suppl. 153 (2004) 139–156, arXiv:hep -ph /0402115 [hep -ph], Int. J. Mod. Phys. A 20 (2005) 4387. [4] A. Bazavov, et al., The chiral and deconﬁnement aspects of the QCD transition, Phys. Rev. D 85 (2012) 054503, arXiv:1111.1710 [hep -lat]. [5] S. Borsanyi, Z. Fodor, J.N. Guenther, S.K. Katz, K.K. Szabo, A. Pasztor, I. Portillo, C. Ratti, Higher order ﬂuctuations and correlations of conserved charges from lattice QCD, J. High Energy Phys. 10 (2018) 205, arXiv:1805 .04445 [hep -lat]. [6] B. Friman, F. Karsch, K. Redlich, V. Skokov, Fluctuations as probe of the QCD phase transition and freeze-out in heavy ion collisions at LHC and RHIC, Eur. Phys. J. C 71 (2011) 1694, arXiv:1103 .3511 [hep -ph]. [7] A. Andronic, P. Braun-Munzinger, K. Redlich, J. Stachel, Decoding the phase structure of QCD via particle production at high energy, Nature 561 (7723) (2018) 321–330, arXiv:1710 .09425 [nucl -th]. [8] A. Andronic, P. Braun-Munzinger, B. Friman, P.M. Lo, K. Redlich, J. Stachel, The thermal proton yield anomaly in Pb-Pb collisions at the LHC and its resolution, Phys. Lett. B 792 (2019) 304–309, arXiv:1808 .03102 [hep -ph]. [9] K. Redlich, Probing QCD chiral cross over transition in heavy ion collisions with ﬂuctuations, Cent. Eur. J. Phys. 10 (2012) 1254–1257, arXiv:1207.2610 [hep -ph]. [9] K. Redlich, Probing QCD chiral cross over transition in heavy ion collisions with ﬂuctuations Cent Eur J Phys 10 (2012) 1254 1257 arXiv:12072610 [hep ph] [10] J.G. Skellam, The frequency distribution of the difference between two Poisson variates belonging to different populations, J. R. Stat. Soc. A 109 (3) (1946) 296. [11] HotQCD Collaboration, A. Bazavov, et al., Chiral crossover in QCD at zero and non-zero chemical potentials, Phys. Lett. B 795 (2019) 15–21, arXiv:1812 .08235 [hep -lat]. 6 [12] A. Bazavov, et al., The QCD equation of state to O(μ6 B) from Lattice QCD, Phys. Rev. D 95 (5) (2017) 054504, arXiv:1701.04325 [hep -lat]. [13] L.D. Landau, E.M. Lifshitz, Statistical Physics, Pergamon Press, 1980. [14] A. Bzdak, V. a b s t r a c t We note that long range rapidity correlations were investigated in other contexts in [45,46]. The search for critical behavior, as predicted for higher cumulants of net-baryon distri- butions [12,47], will be the topic of future investigations. Declaration of competing interest The authors declare that they have no known competing ﬁnan- cial interests or personal relationships that could have appeared to inﬂuence the work reported in this paper. a b s t r a c t The origin for the deviation of the measured R1 values from unity must therefore be beyond the vol- ume ﬂuctuations scenario. To further understand these differences, the acceptance dependence is studied. On the other hand, local baryon number conservation may induce additional correlations between protons and antiprotons, which would lead to a further reduction of the measured κ2(np − np) [35]. In Fig. 2 the data are compared to the predictions from an analysis of effects of local baryon number conservation for dif- ferent values of correlation width ycorr between protons and an- tiprotons. Within the experimental uncertainties the data are best described with the assumption of global baryon number conserva- tion, which corresponds to the correlation width ycorr = 2\|ybeam\| but, within one standard deviation (1.56 for the last point at η = 1.6), the data are also consistent with a large correlation width of ycorr = 5 [35]. We ﬁnd that for ycorr = 4.5, with a 5% signiﬁcance level, the last point is not consistent with the experi- mental data. The results from the HIJING event generator (cf. blue solid line in Fig. 2), which can be described with ycorr = 2, and from a recent study reported in [41] would imply much stronger correlations between protons and antiprotons, not consistent with The analysis is performed in eight different pseudorapidity in- tervals from \|η\| < 0.1 up to \|η\| < 0.8 in steps of 0.1. The obtained normalized second cumulants R1 of net-protons are presented in Fig. 2. Again the data are below unity, with the deviation linearly increasing with increasing acceptance. Such a behavior was predicted based on the assumption of global baryon number conservation [32,36,37], which induces cor- relations between protons and antiprotons leading to the following dependence on the acceptance factor α R1 = 1 −α, (7) (7) R1 = 1 −α, where α = np / N4π B with np and N4π B referring to the mean number of protons inside the acceptance and the mean number of baryons in full phase space. References [38] ALICE Collaboration, B.B. Abelev, et al., K 0 S and production in Pb-Pb collisions at √sNN = 2.76 TeV, Phys. Rev. Lett. 111 (2013) 222301, arXiv:1307.5530 [nucl - ex]. [22] ALICE Collaboration, K. Aamodt, et al., Suppression of charged particle produc- tion at large transverse momentum in central Pb-Pb collisions at √sNN = 2.76 TeV, Phys. Lett. B 696 (2011) 30–39, arXiv:1012 .1004 [nucl -ex]. [39] ALICE Collaboration, B.B. Abelev, et al., Multi-strange baryon production at mid-rapidity in Pb-Pb collisions at √sNN = 2.76 TeV, Phys. Lett. B 728 (2014) 216–227, arXiv:1307.5543 [nucl -ex], Erratum: Phys. Lett. B 734 (2014) 409. [ ] l dl h h l f ﬂ [23] M. Gazdzicki, K. Grebieszkow, M. Mackowiak, S. Mrowczynski, Identity method to study chemical ﬂuctuations in relativistic heavy-ion collisions, Phys. Rev. C 83 (2011) 054907, arXiv:1103 .2887 [nucl -th]. [40] P. Braun-Munzinger, A. Kalweit, K. Redlich, J. Stachel, Confronting ﬂuctuations of conserved charges in central nuclear collisions at the LHC with predictions from Lattice QCD, Phys. Lett. B 747 (2015) 292–298, arXiv:1412 .8614 [hep -ph]. [24] M.I. Gorenstein, Identity method for particle number ﬂuctuations and correla- tions, Phys. Rev. C 84 (2011) 024902, arXiv:1106 .4473 [nucl -th], Erratum: Phys. Rev. C 97 (2) (2018) 029903. [41] C.A. Pruneau, Role of baryon number conservation in measurements of ﬂuctu- ations, Phys. Rev. C 100 (3) (2019) 034905, arXiv:1903 .04591 [nucl -th]. [25] A. Rustamov, M.I. Gorenstein, Identity method for moments of multiplicity dis- tribution, Phys. Rev. C 86 (2012) 044906, arXiv:1204 .6632 [nucl -th]. [42] S. Pratt, Correlations and ﬂuctuations: a summary of Quark Matter 2002, Nucl. Phys. A 715 (2003) 389–398, arXiv:nucl -th /0308022 [nucl -th]. [26] T. Anticic, et al., Phase-space dependence of particle-ratio ﬂuctuations in Pb + Pb collisions from 20 A to 158 A GeV beam energy, Phys. Rev. C 89 (5) (2014) 054902, arXiv:1310 .3428 [nucl -ex]. [43] ALICE Collaboration, B. Abelev, et al., Charge correlations using the balance function in Pb-Pb collisions at √sNN = 2.76 TeV, Phys. Lett. B 723 (2013) 267–279, arXiv:1301.3756 [nucl -ex]. [27] M. Arslandok, A. Rustamov, Tidentity module for the reconstruction of the moments of multiplicity distributions, Nucl. Instrum. Methods A 946 (2019) 162622, http://www.sciencedirect .com /science /article /pii /S0168900219311222. [44] A. Dumitru, F. Gelis, L. McLerran, R. Venugopalan, Glasma ﬂux tubes and the near side ridge phenomenon at RHIC, Nucl. Phys. References Koch, Acceptance corrections to net baryon and net charge cumu- lants, Phys. Rev. C 86 (2012) 044904, arXiv:1206 .4286 [nucl -th]. [15] P. Braun-Munzinger, A. Rustamov, J. Stachel, Experimental results on ﬂuctua- tions of conserved charges confronted with predictions from canonical ther- modynamics, Nucl. Phys. A 982 (2019) 307–310, arXiv:1807.08927 [nucl -th]. ALICE Collaboration / Physics Letters B 807 (2020) 135564 5 5 [16] ALICE Collaboration, B. Abelev, et al., Centrality dependence of π, K, p produc- tion in Pb-Pb collisions at √sNN = 2.76 TeV, Phys. Rev. C 88 (2013) 044910, arXiv:1303 .0737 [hep -ex]. [32] P. Braun-Munzinger, A. Rustamov, J. Stachel, Bridging the gap between event- by-event ﬂuctuation measurements and theory predictions in relativistic nu- clear collisions, Nucl. Phys. A 960 (2017) 114–130, arXiv:1612 .00702 [nucl -th]. [17] ALICE Collaboration, K. Aamodt, et al., The ALICE experiment at the CERN LHC, J. Instrum. 3 (2008) S08002. [33] T. Anticic, et al., Phase-space dependence of particle-ratio ﬂuctuations in Pb + Pb collisions from 20 A to 158 A GeV beam energy, Phys. Rev. C 89 (5) (2014) 054902, arXiv:1310 .3428 [nucl -ex]. [18] ALICE Collaboration, B. Abelev, et al., Centrality determination of Pb-Pb col- lisions at √sNN = 2.76 TeV with ALICE, Phys. Rev. C 88 (4) (2013) 044909, arXiv:1301.4361 [nucl -ex]. [34] M.I. Gorenstein, M. Gazdzicki, Strongly intensive quantities, Phys. Rev. C 84 (2011) 014904, arXiv:1101.4865 [nucl -th]. [35] P. Braun-Munzinger, A. Rustamov, J. Stachel, The role of the local conservation laws in ﬂuctuations of conserved charges, arXiv:1907.03032 [nucl -th]. [19] J. Alme, et al., The ALICE TPC, a large 3-dimensional tracking device with fast readout for ultra-high multiplicity events, Nucl. Instrum. Methods A 622 (2010) 316–367, arXiv:1001.1950 [physics .ins -det]. [36] S. Mrowczynski, Measuring charge ﬂuctuations in high-energy nuclear colli- sions, Phys. Rev. C 66 (2002) 024904, arXiv:nucl -th /0112007 [nucl -th]. [ ] d k h k k b d h l [20] G. Dellacasa, et al., ALICE Collaboration, ALICE technical design report of the inner tracking system (ITS), CERN-LHCC-99-12, 1999. [37] A. Bzdak, V. Koch, V. Skokov, Baryon number conservation and the cumulants of the net proton distribution, Phys. Rev. C 87 (1) (2013) 014901, arXiv:1203 . 4529 [hep -ph]. 0 [21] ALICE Collaboration, S. Acharya, et al., Relative particle yield ﬂuctuations in Pb-Pb collisions at √sNN = 2.76 TeV, Eur. Phys. J. C 79 (3) (2019) 236, arXiv:1712 .07929 [nucl -ex]. References A 810 (2008) 91–108, arXiv: 0804 .3858 [hep -ph]. [28] A. Bzdak, V. Koch, Acceptance corrections to net baryon and net charge cumu- lants, Phys. Rev. C 86 (2012) 044904, arXiv:1206 .4286 [nucl -th]. [45] A. Capella, A. Krzywicki, Unitarity corrections to short range order: long range rapidity correlations, Phys. Rev. D 18 (1978) 4120. [29] A. Bzdak, V. Koch, Local eﬃciency corrections to higher order cumulants, Phys. Rev. C 91 (2) (2015) 027901, arXiv:1312 .4574 [nucl -th]. [46] V.V. Vechernin, Forward-backward correlations between multiplicities in win- dows separated in azimuth and rapidity, Nucl. Phys. A 939 (2015) 21–45, arXiv:1210 .7588 [hep -ph]. [30] T. Sugiura, T. Nonaka, S. Esumi, Volume ﬂuctuation and multiplicity correlation on higher-order cumulants, Phys. Rev. C 100 (4) (2019) 044904, arXiv:1903 . 02314 [nucl -th]. [47] G.A. Almasi, B. Friman, K. Redlich, Baryon number ﬂuctuations in chiral ef- fective models and their phenomenological implications, Phys. Rev. D 96 (1) (2017) 014027, arXiv:1703 .05947 [hep -ph]. [31] X. Luo, N. Xu, Search for the QCD critical point with ﬂuctuations of conserved quantities in relativistic heavy-ion collisions at RHIC: an overview, Nucl. Sci. Tech. 28 (8) (2017) 112, arXiv:1701.02105 [nucl -ex]. ALICE Collaboration S. Acharya 141, D. Adamová 94, A. Adler 74, J. Adolfsson 80, M.M. Aggarwal 99, G. Aglieri Rinella 33, M. Agnello 30, N. Agrawal 10,53, Z. Ahammed 141, S. Ahmad 16, S.U. Ahn 76, A. Akindinov 91, M. Al-Turany 106, S.N. Alam 141, D.S.D. Albuquerque 122, D. Aleksandrov 87, B. Alessandro 58, H.M. Alfanda 6, R. Alfaro Molina 71, B. Ali 16, Y. Ali 14, A. Alici 10,26,53, A. Alkin 2, J. Alme 21, T. Alt 68, L. Altenkamper 21, I. Altsybeev 112, M.N. Anaam 6, C. Andrei 47, D. Andreou 33, H.A. Andrews 110, A. Andronic 144, M. Angeletti 33, V. Anguelov 103, C. Anson 15, T. Antiˇci´c 107, F. Antinori 56, P. Antonioli 53, R. Anwar 125, N. Apadula 79, L. Aphecetche 114, H. Appelshäuser 68, S. Arcelli 26, R. Arnaldi 58, M. Arratia 79 I.C. Arsene 20, M. Arslandok 103, A. Augustinus 33, R. Averbeck 106, S. Aziz 61, M.D. Azmi 16, A. Badalà 55, Y.W. Baek 40, S. Bagnasco 58, X. Bai 106, R. Bailhache 68, R. Bala 100, A. Baldisseri 137, M. Ball 42, S. Balouza 104, R. Barbera 27, L. Barioglio 25, G.G. Barnaföldi 145, L.S. Barnby 93, V. Barret 134, P. Bartalini 6, K. Barth 33, E. Bartsch 68, F. Baruffaldi 28, N. Bastid 134, S. Basu 143, G. Batigne 114, B. Batyunya 75, D. Bauri 48, J.L. Bazo Alba 111, I.G. Bearden 88, C. Bedda 63, N.K. Behera 60, I. Belikov 136, A.D.C. Bell Hechavarria 144, F. Bellini 33, R. Bellwied 125, V. Belyaev 92, G. Bencedi 145, S. Beole 25, A. Bercuci 47, Y. Berdnikov 97, D. Berenyi 145, R.A. Bertens 130, D. Berzano 58, M.G. Besoiu 67, L. Betev 33, A. Bhasin 100, I.R. Bhat 100, M.A. Bhat 3, H. Bhatt 48, B. Bhattacharjee 41, A. Bianchi 25, L. Bianchi 25, N. Bianchi 51, J. Bielˇcík 36, J. Bielˇcíková 94, A. Bilandzic 104,117, G. Biro 145, R. Biswas 3, S. Biswas 3, J.T. Blair 119, D. Blau 87, C. Blume 68, G. Boca 139, F. Bock 33,95, A. Bogdanov 92, S. Boi 23, L. Boldizsár 145, A. Bolozdynya 92, M. Bombara 37, G. Bonomi 140, H. Borel 137, A. Borissov 92,144, H. Bossi 146, E. Botta 25, L. Bratrud 68, P. Braun-Munzinger 106, M. Bregant 121, M. Broz 36, E.J. Brucken 43, E. Bruna 58, G.E. Bruno 105, M.D. Buckland 127, D. Budnikov 108, H. Buesching 68, S. ALICE Collaboration Bufalino 30, O. Bugnon 114, P. Buhler 113, P. Buncic 33, Z. Buthelezi 72,131, J.B. Butt 14, J.T. Buxton 96, S.A. Bysiak 118, D. Caffarri 89, S. Acharya 141, D. Adamová 94, A. Adler 74, J. Adolfsson 80, M.M. Aggarwal 99, G. Aglieri Rinella 33, M. Agnello 30, N. Agrawal 10,53, Z. Ahammed 141, S. Ahmad 16, S.U. Ahn 76, A. Akindinov 91, M. Al-Turany 106, S.N. Alam 141, D.S.D. Albuquerque 122, D. Aleksandrov 87, B. Alessandro 58, H.M. Alfanda 6, R. Alfaro Molina 71, B. Ali 16, Y. Ali 14, A. Alici 10,26,53, A. Alkin 2, J. Alme 21, T. Alt 68, L. Altenkamper 21, I. Altsybeev 112, M.N. Anaam 6, C. Andrei 47, D. Andreou 33, H.A. Andrews 110, A. Andronic 144, M. Angeletti 33, V. Anguelov 103, C. Anson 15, T. Antiˇci´c 107, F. Antinori 56, P. Antonioli 53, R. Anwar 125, N. Apadula 79, L. Aphecetche 114, H. Appelshäuser 68, S. Arcelli 26, R. Arnaldi 58, M. Arratia 79, I.C. Arsene 20, M. Arslandok 103, A. Augustinus 33, R. Averbeck 106, S. Aziz 61, M.D. Azmi 16, A. Badalà 55, Y.W. Baek 40, S. Bagnasco 58, X. Bai 106, R. Bailhache 68, R. Bala 100, A. Baldisseri 137, M. Ball 42, S. Balouza 104, R. Barbera 27, L. Barioglio 25, G.G. Barnaföldi 145, L.S. Barnby 93, V. Barret 134, P. Bartalini 6, K. Barth 33, E. Bartsch 68, F. Baruffaldi 28, N. Bastid 134, S. Basu 143, G. Batigne 114, B. Batyunya 75, D. Bauri 48, J.L. Bazo Alba 111, I.G. Bearden 88, C. Bedda 63, N.K. Behera 60, I. Belikov 136, A.D.C. Bell Hechavarria 144, F. Bellini 33, R. Bellwied 125, V. Belyaev 92, G. Bencedi 145, S. Beole 25, A. Bercuci 47, Y. Berdnikov 97, D. Berenyi 145, R.A. Bertens 130, D. Berzano 58, M.G. Besoiu 67, L. Betev 33, A. Bhasin 100, I.R. Bhat 100, M.A. Bhat 3, H. Bhatt 48, B. Bhattacharjee 41, A. Bianchi 25, L. Bianchi 25, N. Bianchi 51, J. Bielˇcík 36, J. Bielˇcíková 94, A. Bilandzic 104,117, G. Biro 145, R. Biswas 3, S. Biswas 3, J.T. Blair 119, D. Blau 87, C. Blume 68, G. Boca 139, F. Bock 33,95, A. Bogdanov 92, S. Boi 23, L. Boldizsár 145, A. Bolozdynya 92, M. Bombara 37, G. Bonomi 140, H. Borel 137, A. Borissov 92,144, H. Bossi 146, E. Botta 25, L. Bratrud 68, P. Braun-Munzinger 106, M. ALICE Collaboration Hippolyte 136, B. Hohlweger 104, D. Horak 36, A. Hornung 68, S. Hornung 106, R. Hosokawa 15,133, P. Hristov 33, C. Huang 61, C. Hughes 130, P. Huhn 68, T.J. Humanic 96, H. Hushnud 109, L.A. Husova 144, N. Hussain 41, S.A. Hussain 14, D. Hutter 38, J.P. Iddon 33,127, R. Ilkaev 108, M. Inaba 133, G.M. Innocenti 33, M. Ippolitov 87, A. Isakov 94, M.S. Islam 109, M. Ivanov 106, V. Ivanov 97, V. Izucheev 90, B. Jacak 79, N. Jacazio 53, P.M. Jacobs 79, S. Jadlovska 116, J. Jadlovsky 116, S. Jaelani 63, C. Jahnke 121, M.J. Jakubowska 142, M.A. Janik 142, T. Janson 74, M. Jercic 98, O. Jevons 110, M. Jin 125, F. Jonas 95,144, P.G. Jones 110, J. Jung 68, M. Jung 68, A. Jusko 110, P. Kalinak 64, A. Kalweit 33, V. Kaplin 92, S. Kar 6, A. Karasu Uysal 77, O. Karavichev 62, T. Karavicheva 62, P. Karczmarczyk 33, E. Karpechev 62, A. Kazantsev 87, U. Kebschull 74, R. Keidel 46, M. Keil 33, B. Ketzer 42, Z. Khabanova 89, A.M. Khan 6, S. Khan 16, S.A. Khan 141, A. Khanzadeev 97, Y. Kharlov 90, A. Khatun 16, A. Khuntia 118, B. Kileng 35, B. Kim 60, B. Kim 133, D. Kim 147, D.J. Kim 126, E.J. Kim 73, H. Kim 17,147, J. Kim 147, J.S. Kim 40, J. Kim 103, J. Kim 147, J. Kim 73, M. Kim 103, S. Kim 18, T. Kim 147, T. Kim 147, S. Kirsch 38,68, I. Kisel 38, S. Kiselev 91, A. Kisiel 142, J.L. Klay 5, C. Klein 68, J. Klein 58, S. Klein 79, C. Klein-Bösing 144, M. Kleiner 68, A. Kluge 33, M.L. Knichel 33, A.G. Knospe 125, C. Kobdaj 115, M.K. Köhler 103, T. Kollegger 106, A. Kondratyev 75, N. Kondratyeva 92, E. Kondratyuk 90, J. Konig 68, P.J. Konopka 33, L. Koska 116, O. Kovalenko 84, V. Kovalenko 112, M. Kowalski 118, I. Králik 64, A. Kravˇcáková 37, L. Kreis 106, M. Krivda 64,110, F. Krizek 94, K Krizkova Gajdosova 36 M Krüger 68 E Kryshen 97 M Krzewicki 38 A M Kubera 96 V Kuˇcera 60 A. Caliva 106, E. Calvo Villar 111, R.S. Camacho 44, P. Camerini 24, A.A. Capon 113, F. Carnesecchi 10,26, p aron 137, J. Castillo Castellanos 137, A.J. Castro 130, E.A.R. Casula 54, F. ALICE Collaboration Kisiel 142, J.L. Klay 5, C. Klein 68, J. Klein 58, S. Klein 79, C. Klein-Bösing 144, M. Kleiner 68, A. Kluge 33, M.L. Knichel 33, A.G. Knospe 125, C. Kobdaj 115, M.K. Köhler 103, T. Kollegger 106, A. Kondratyev 75, N. Kondratyeva 92, E. Kondratyuk 90, J. Konig 68, P.J. Konopka 33, L. Koska 116, O. Kovalenko 84, V. Kovalenko 112, M. Kowalski 118, I. Králik 64, A. Kravˇcáková 37, L. Kreis 106, M. Krivda 64,110, F. Krizek 94, k d 36 68 h 97 k 38 b 96 ˇ 60 ALICE Collaboration / Physics Letters B 807 (2020) 135564 6 6 ALICE Collaboration / Physics Letters B 807 (2020) 135564 A. Caliva 106, E. Calvo Villar 111, R.S. Camacho 44, P. Camerini 24, A.A. Capon 113, F. Carnesecchi 10,26, R. Caron 137, J. Castillo Castellanos 137, A.J. Castro 130, E.A.R. Casula 54, F. Catalano 30, C. Ceballos Sanchez 52, P. Chakraborty 48, S. Chandra 141, W. Chang 6, S. Chapeland 33, M. Chartier 127, S. Chattopadhyay 141, S. Chattopadhyay 109, A. Chauvin 23, C. Cheshkov 135, B. Cheynis 135, V. Chibante Barroso 33, D.D. Chinellato 122, S. Cho 60, P. Chochula 33, T. Chowdhury 134, P. Christakoglou 89, C.H. Christensen 88, P. Christiansen 80, T. Chujo 133, C. Cicalo 54, L. Cifarelli 10,26, F. Cindolo 53, J. Cleymans 124, F. Colamaria 52, D. Colella 52, A. Collu 79, M. Colocci 26, M. Concas 58,ii, G. Conesa Balbastre 78, Z. Conesa del Valle 61, G. Contin 24,127, J.G. Contreras 36, T.M. Cormier 95, Y. Corrales Morales 25, P. Cortese 31, M.R. Cosentino 123, F. Costa 33, S. Costanza 139, P. Crochet 134, E. Cuautle 69, P. Cui 6, L. Cunqueiro 95, D. Dabrowski 142, T. Dahms 104,117, A. Dainese 56, F.P.A. Damas 114,137, M.C. Danisch 103, A. Danu 67, D. Das 109, I. Das 109, P. Das 85, P. Das 3, S. Das 3, A. Dash 85, S. Dash 48, S. De 85, A. De Caro 29, G. de Cataldo 52, J. de Cuveland 38, A. De Falco 23, D. De Gruttola 10, N. De Marco 58, S. De Pasquale 29, S. Deb 49, B. Debjani 3, H.F. Degenhardt 121, K.R. Deja 142, A. Deloff 84, S. Delsanto 25,131, D. Devetak 106, P. Dhankher 48, D. Di Bari 32, A. Di Mauro 33, R.A. Diaz 8, T. Dietel 124, P. Dillenseger 68, Y. Ding 6, R. Divià 33, D.U. ALICE Collaboration Dixit 19, Ø. Djuvsland 21, U. Dmitrieva 62, A. Dobrin 33,67, B. Dönigus 68, O. Dordic 20, A.K. Dubey 141, A. Dubla 106, S. Dudi 99, M. Dukhishyam 85, P. Dupieux 134, R.J. Ehlers 146, V.N. Eikeland 21, D. Elia 52, H. Engel 74, E. Epple 146, B. Erazmus 114, F. Erhardt 98, A. Erokhin 112, M.R. Ersdal 21, B. Espagnon 61, G. Eulisse 33, D. Evans 110, S. Evdokimov 90, L. Fabbietti 104,117, M. Faggin 28, J. Faivre 78, F. Fan 6, A. Fantoni 51, M. Fasel 95, P. Fecchio 30, A. Feliciello 58, G. Feoﬁlov 112, A. Fernández Téllez 44, A. Ferrero 137, A. Ferretti 25, A. Festanti 33, V.J.G. Feuillard 103, J. Figiel 118, S. Filchagin 108, D. Finogeev 62, F.M. Fionda 21, G. Fiorenza 52, F. Flor 125, S. Foertsch 72, P. Foka 106, S. Fokin 87, E. Fragiacomo 59, U. Frankenfeld 106, U. Fuchs 33, C. Furget 78, A. Furs 62, M. Fusco Girard 29, J.J. Gaardhøje 88, M. Gagliardi 25, A.M. Gago 111, A. Gal 136, C.D. Galvan 120, P. Ganoti 83, C. Garabatos 106, E. Garcia-Solis 11, K. Garg 27, C. Gargiulo 33, A. Garibli 86, K. Garner 144, P. Gasik 104,117, E.F. Gauger 119, M.B. Gay Ducati 70, M. Germain 114, J. Ghosh 109, P. Ghosh 141, S.K. Ghosh 3, P. Gianotti 51, P. Giubellino 58,106, P. Giubilato 28, P. Glässel 103, D.M. Goméz Coral 71, A. Gomez Ramirez 74, V. Gonzalez 106, P. González-Zamora 44, S. Gorbunov 38, L. Görlich 118, S. Gotovac 34, V. Grabski 71, L.K. Graczykowski 142, K.L. Graham 110, L. Greiner 79, A. Grelli 63, C. Grigoras 33, V. Grigoriev 92, A. Grigoryan 1, S. Grigoryan 75, O.S. Groettvik 21, F. Grosa 30, J.F. Grosse-Oetringhaus 33, R. Grosso 106, R. Guernane 78, M. Guittiere 114, K. Gulbrandsen 88, T. Gunji 132, A. Gupta 100, R. Gupta 100, I.B. Guzman 44, R. Haake 146, M.K. Habib 106, C. Hadjidakis 61, H. Hamagaki 81, G. Hamar 145, M. Hamid 6, R. Hannigan 119, M.R. Haque 63,85, A. Harlenderova 106, J.W. Harris 146, A. Harton 11, J.A. Hasenbichler 33, H. Hassan 95, D. Hatzifotiadou 10,53, P. Hauer 42, S. Hayashi 132, S.T. Heckel 68,104, E. Hellbär 68, H. Helstrup 35, A. Herghelegiu 47, T. Herman 36, E.G. Hernandez 44, G. Herrera Corral 9, F. Herrmann 144, K.F. Hetland 35, T.E. Hilden 43, H. Hillemanns 33, C. Hills 127, B. ALICE Collaboration Bregant 121, M. Broz 36, E.J. Brucken 43, E. Bruna 58, G.E. Bruno 105, M.D. Buckland 127, D. Budnikov 108, H. Buesching 68, S. Bufalino 30, O. Bugnon 114, P. Buhler 113, P. Buncic 33, Z. Buthelezi 72,131, J.B. Butt 14, J.T. Buxton 96, S.A. Bysiak 118, D. Caffarri 89, , J , , , , , A.D.C. Bell Hechavarria 144, F. Bellini 33, R. Bellwied 125, V. Belyaev 92, G. Bencedi 145, S. Beole 25, A. Bercuci 47, Y. Berdnikov 97, D. Berenyi 145, R.A. Bertens 130, D. Berzano 58, M.G. Besoiu 67, L. Betev 33, A. Bhasin 100, I.R. Bhat 100, M.A. Bhat 3, H. Bhatt 48, B. Bhattacharjee 41, A. Bianchi 25, L. Bianchi 25, N. Bianchi 51, J. Bielˇcík 36, J. Bielˇcíková 94, A. Bilandzic 104,117, G. Biro 145, R. Biswas 3, S. Biswas 3, J.T. Blair 119, D. Blau 87, C. Blume 68, G. Boca 139, F. Bock 33,95, A. Bogdanov 92, S. Boi 23, L. Boldizsár 145, A. Bolozdynya 92, M. Bombara 37, G. Bonomi 140, H. Borel 137, A. Borissov 92,144, H. Bossi 146, E. Botta 25, L. Bratrud 68, P. Braun-Munzinger 106, M. Bregant 121, M. Broz 36, E.J. Brucken 43, E. Bruna 58, G.E. Bruno 105, M.D. Buckland 127, D. Budnikov 108, H. Buesching 68, S. Bufalino 30, O. Bugnon 114, P. Buhler 113, P. Buncic 33, Z. Buthelezi 72,131, J.B. Butt 14, J.T. Buxton 96, S.A. Bysiak 118, D. Caffarri 89, 6 ALICE Collaboration / Physics Letters B 807 (2020) 135564 A. Caliva 106, E. Calvo Villar 111, R.S. Camacho 44, P. Camerini 24, A.A. Capon 113, F. Carnesecchi 10,26, R. Caron 137, J. Castillo Castellanos 137, A.J. Castro 130, E.A.R. Casula 54, F. Catalano 30, C. Ceballos Sanchez 52, P. Chakraborty 48, S. Chandra 141, W. Chang 6, S. Chapeland 33, M. Chartier 127, S. Chattopadhyay 141, S. Chattopadhyay 109, A. Chauvin 23, C. Cheshkov 135, B. Cheynis 135, V. Chibante Barroso 33, D.D. Chinellato 122, S. Cho 60, P. Chochula 33, T. Chowdhury 134, P. Christakoglou 89, C.H. Christensen 88, P. Christiansen 80, T. Chujo 133, C. Cicalo 54, L. Cifarelli 10,26, F. Cindolo 53, J. Cleymans 124, F. Colamaria 52, D. Colella 52, A. Collu 79, M. Colocci 26, M. Concas 58,ii, G. Conesa Balbastre 78, Z. Conesa del Valle 61, G. Contin 24,127, J.G. Contreras 36, T.M. Cormier 95, Y. Corrales Morales 25, P. ALICE Collaboration Cortese 31, M.R. Cosentino 123, F. Costa 33, S. Costanza 139, P. Crochet 134, E. Cuautle 69, P. Cui 6, L. Cunqueiro 95, D. Dabrowski 142, T. Dahms 104,117, A. Dainese 56, F.P.A. Damas 114,137, M.C. Danisch 103, A. Danu 67, D. Das 109, I. Das 109, P. Das 85, P. Das 3, S. Das 3, A. Dash 85, S. Dash 48, S. De 85, A. De Caro 29, G. de Cataldo 52, J. de Cuveland 38, A. De Falco 23, D. De Gruttola 10, N. De Marco 58, S. De Pasquale 29, S. Deb 49, B. Debjani 3, H.F. Degenhardt 121, K.R. Deja 142, A. Deloff 84, S. Delsanto 25,131, D. Devetak 106, P. Dhankher 48, D. Di Bari 32, A. Di Mauro 33, R.A. Diaz 8, T. Dietel 124, P. Dillenseger 68, Y. Ding 6, R. Divià 33, D.U. Dixit 19, Ø. Djuvsland 21, U. Dmitrieva 62, A. Dobrin 33,67, B. Dönigus 68, O. Dordic 20, A.K. Dubey 141, A. Dubla 106, S. Dudi 99, M. Dukhishyam 85, P. Dupieux 134, R.J. Ehlers 146, V.N. Eikeland 21, D. Elia 52, H. Engel 74, E. Epple 146, B. Erazmus 114, F. Erhardt 98, A. Erokhin 112, M.R. Ersdal 21, B. Espagnon 61, G. Eulisse 33, D. Evans 110, S. Evdokimov 90, L. Fabbietti 104,117, M. Faggin 28, J. Faivre 78, F. Fan 6, A. Fantoni 51, M. Fasel 95, P. Fecchio 30, A. Feliciello 58, G. Feoﬁlov 112, A. Fernández Téllez 44, A. Ferrero 137, A. Ferretti 25, A. Festanti 33, V.J.G. Feuillard 103, J. Figiel 118, S. Filchagin 108, D. Finogeev 62, F.M. Fionda 21, G. Fiorenza 52, F. Flor 125, S. Foertsch 72, P. Foka 106, S. Fokin 87, E. Fragiacomo 59, U. Frankenfeld 106, U. Fuchs 33, C. Furget 78, A. Furs 62, M. Fusco Girard 29, J.J. Gaardhøje 88, M. Gagliardi 25, A.M. Gago 111, A. Gal 136, C.D. Galvan 120, P. Ganoti 83, C. Garabatos 106, E. Garcia-Solis 11, K. Garg 27, C. Gargiulo 33, A. Garibli 86, K. Garner 144, P. Gasik 104,117, E.F. Gauger 119, M.B. Gay Ducati 70, M. Germain 114, J. Ghosh 109, P. Ghosh 141, S.K. Ghosh 3, P. Gianotti 51, P. Giubellino 58,106, P. Giubilato 28, P. Glässel 103, D.M. Goméz Coral 71, A. Gomez Ramirez 74, V. Gonzalez 106, P. González-Zamora 44, S. Gorbunov 38, L. Görlich 118, S. Gotovac 34, V. Grabski 71, L.K. ALICE Collaboration Graczykowski 142, K.L. Graham 110, L. Greiner 79, A. Grelli 63, C. Grigoras 33, V. Grigoriev 92, A. Grigoryan 1, S. Grigoryan 75, O.S. Groettvik 21, F. Grosa 30, J.F. Grosse-Oetringhaus 33, R. Grosso 106, R. Guernane 78, M. Guittiere 114, K. Gulbrandsen 88, T. Gunji 132, A. Gupta 100, R. Gupta 100, I.B. Guzman 44, R. Haake 146, M.K. Habib 106, C. Hadjidakis 61, H. Hamagaki 81, G. Hamar 145, M. Hamid 6, R. Hannigan 119, M.R. Haque 63,85, A. Harlenderova 106, J.W. Harris 146, A. Harton 11, J.A. Hasenbichler 33, H. Hassan 95, D. Hatzifotiadou 10,53, P. Hauer 42, S. Hayashi 132, S.T. Heckel 68,104, E. Hellbär 68, H. Helstrup 35, A. Herghelegiu 47, T. Herman 36, E.G. Hernandez 44, G. Herrera Corral 9, F. Herrmann 144, K.F. Hetland 35, T.E. Hilden 43, H. Hillemanns 33, C. Hills 127, B. Hippolyte 136, B. Hohlweger 104, D. Horak 36, A. Hornung 68, S. Hornung 106, R. Hosokawa 15,133, P. Hristov 33, C. Huang 61, C. Hughes 130, P. Huhn 68, T.J. Humanic 96, H. Hushnud 109, L.A. Husova 144, N. Hussain 41, S.A. Hussain 14, D. Hutter 38, J.P. Iddon 33,127, R. Ilkaev 108, M. Inaba 133, G.M. Innocenti 33, M. Ippolitov 87, A. Isakov 94, M.S. Islam 109, M. Ivanov 106, V. Ivanov 97, V. Izucheev 90, B. Jacak 79, N. Jacazio 53, P.M. Jacobs 79, S. Jadlovska 116, J. Jadlovsky 116, S. Jaelani 63, C. Jahnke 121, M.J. Jakubowska 142, M.A. Janik 142, T. Janson 74, M. Jercic 98, O. Jevons 110, M. Jin 125, F. Jonas 95,144, P.G. Jones 110, J. Jung 68, M. Jung 68, A. Jusko 110, P. Kalinak 64, A. Kalweit 33, V. Kaplin 92, S. Kar 6, A. Karasu Uysal 77, O. Karavichev 62, T. Karavicheva 62, P. Karczmarczyk 33, E. Karpechev 62, A. Kazantsev 87, U. Kebschull 74, R. Keidel 46, M. Keil 33, B. Ketzer 42, Z. Khabanova 89, A.M. Khan 6, S. Khan 16, S.A. Khan 141, A. Khanzadeev 97, Y. Kharlov 90, A. Khatun 16, A. Khuntia 118, B. Kileng 35, B. Kim 60, B. Kim 133, D. Kim 147, D.J. Kim 126, E.J. Kim 73, H. Kim 17,147, J. Kim 147, J.S. Kim 40, J. Kim 103, J. Kim 147, J. Kim 73, M. Kim 103, S. Kim 18, T. Kim 147, T. Kim 147, S. Kirsch 38,68, I. Kisel 38, S. Kiselev 91, A. ALICE Collaboration Catalano 30, 52 48 141 6 33 , J , J , , , C. Ceballos Sanchez 52, P. Chakraborty 48, S. Chandra 141, W. Chang 6, S. Chapeland 33, M. Chartier 127, 141 109 23 135 135 J J eballos Sanchez 52, P. Chakraborty 48, S. Chandra 141, W. Chang 6, S. Chapeland 33, p y y , p y y , , , y , V. Chibante Barroso 33, D.D. Chinellato 122, S. Cho 60, P. Chochula 33, T. Chowdhury 134, P. Christakoglou 89, 88 80 133 54 10 26 53 , , , , , e 69, P. Cui 6, L. Cunqueiro 95, D. Dabrowski 142, T. Dahms 104,117, A. Dainese 56, q j g K.R. Deja 142, A. Deloff 84, S. Delsanto 25,131, D. Devetak 106, P. Dhankher 48, D. Di Bari 32, A. Di Mauro 33, R A Di 8 T Di l 124 P Dill 68 Y Di 6 R Di ià 33 D U Di i 19 Ø Dj l d 21 ALICE Collaboration / Physics Letters B 807 (2020) 135564 7 ALICE Collaboration / Physics Letters B 807 (2020) 135564 A.B. Kurepin 62, A. Kuryakin 108, S. Kushpil 94, J. Kvapil 110, M.J. Kweon 60, J.Y. Kwon 60, Y. Kwon 147, S.L. La Pointe 38, P. La Rocca 27, Y.S. Lai 79, R. Langoy 129, K. Lapidus 33, A. Lardeux 20, P. Larionov 51, E. Laudi 33, R. Lavicka 36, T. Lazareva 112, R. Lea 24, L. Leardini 103, J. Lee 133, S. Lee 147, F. Lehas 89, S. Lehner 113, J. Lehrbach 38, R.C. Lemmon 93, I. León Monzón 120, E.D. Lesser 19, M. Lettrich 33, P. Léva X. Li 12, X.L. Li 6, J. Lien 129, R. Lietava 110, B. Lim 17, V. Lindenstruth 38, S.W. Lindsay 127, C. Lippmann 1 M.A. Lisa 96, V. Litichevskyi 43, A. Liu 19, S. Liu 96, W.J. Llope 143, I.M. Lofnes 21, V. Loginov 92, C. Loizide P. Loncar 34, X. Lopez 134, E. López Torres 8, J.R. Luhder 144, M. Lunardon 28, G. Luparello 59, Y. Ma 39, A. Maevskaya 62, M. Mager 33, S.M. Mahmood 20, T. Mahmoud 42, A. Maire 136, R.D. Majka 146, M. Malaev 97, Q.W. Malik 20, L. Malinina 75,iii, D. Mal’Kevich 91, P. Malzacher 106, G. Mandaglio 55, V. Manko 87, F. Manso 134, V. Manzari 52, Y. Mao 6, M. Marchisone 135, J. ALICE Collaboration Mareš 66, G.V. Margagliotti 24 A. Margotti 53, J. Margutti 63, A. Marín 106, C. Markert 119, M. Marquard 68, N.A. Martin 103, P. Martinengo 33, J.L. Martinez 125, M.I. Martínez 44, G. Martínez García 114, M. Martinez Pedreira 33, S. Masciocchi 106, M. Masera 25, A. Masoni 54, L. Massacrier 61, E. Masson 114, A. Mastroserio 52,138, A.M. Mathis 104,117, O. Matonoha 80, P.F.T. Matuoka 121, A. Matyja 118, C. Mayer 118, M. Mazzilli 52, M.A. Mazzoni 57, A.F. Mechler 68, F. Meddi 22, Y. Melikyan 62,92, A. Menchaca-Rocha 71, C. Mengke 6, E. Meninno 29,113, M. Meres 13, S. Mhlanga 124, Y. Miake 133, L. Micheletti 25, D.L. Mihaylov 104, K. Mikhaylov 75,91, A. Mischke 63,i, A.N. Mishra 69, D. Mi´skowiec 106, A. Modak 3, N. Mohammadi 33, A.P. Mohanty 63, B. Mohanty 85, M. Mohisin Khan 16,iv, C. Mordasini 104, D.A. Moreira De Godoy 144, L.A.P. Moreno 44, I. Morozov 62, A. Morsch 33, T. Mrnjavac 33, V. Muccifora 51, E. Mudnic 34, D. Mühlheim 144, S. Muhuri 141, J.D. Mulligan 79, M.G. Munhoz 121, R.H. Munzer 68, H. Murakami 132, S. Murray 124, L. Musa 33, J. Musinsky 64, C.J. Myers 125, J.W. Myrcha 142, B. Naik 48, R. Nair 84, B.K. Nandi 48, R. Nania 10,53, E. Nappi 52, M.U. Naru 14, A.F. Nassirpour 80, C. Nattrass 130, R. Nayak 48, T.K. Nayak 85, S. Nazarenko 108, A. Neagu 20, R.A. Negrao De Oliveira 68, L. Nellen 69, S.V. Nesbo 35, G. Neskovic 38, D. Nesterov 112, L.T. Neumann 142, B.S. Nielsen 88, S. Nikolaev 87, S. Nikulin 87, V. Nikuli F. Noferini 10,53, P. Nomokonov 75, J. Norman 78,127, N. Novitzky 133, P. Nowakowski 142, A. Nyanin 87, J. Nystrand 21, M. Ogino 81, A. Ohlson 80,103, J. Oleniacz 142, A.C. Oliveira Da Silva 121,130, M.H. Oliver 14 C. Oppedisano 58, R. Orava 43, A. Ortiz Velasquez 69, A. Oskarsson 80, J. Otwinowski 118, K. Oyama 81, Y. Pachmayer 103, V. Pacik 88, D. Pagano 140, G. Pai´c 69, J. Pan 143, A.K. Pandey 48, S. Panebianco 137, P. Pareek 49,141, J. Park 60, J.E. Parkkila 126, S. Parmar 99, S.P. Pathak 125, R.N. Patra 141, B. Paul 23,58, H. P T. Peitzmann 63, X. Peng 6, L.G. Pereira 70, H. Pereira Da Costa 137, D. Peresunko 87, G.M. Perez 8, E. Perez Lezama 68, V. Peskov 68, Y. Pestov 4, V. ALICE Collaboration Xu 6, S. Yalcin 77, K. Yamakawa 45, S. Yang 21, S. Yano 137, Z. Yin 6, H. Yokoyama 63, I.-K. Yoo 17, J.H. Yoon 60, S. Yuan 21, A. Yuncu 103, V. Yurchenko 2, V. Zaccolo 24, A. Zaman 14, C. Zampolli 33, H.J.C. Zanoli 63, N. Zardoshti 33, A. Zarochentsev 112, P. Závada 66, N. Zaviyalov 108, H. Zbroszczyk 142, M. Zhalov 97, S. Zhang 39, X. Zhang 6, Z. Zhang 6, V. Zherebchevskii 112, D. Zhou 6, Y. Zhou 88, Z. Zhou 21, J. Zhu 6,106, Y. Zhu 6, A. Zichichi 10,26, M.B. Zimmermann 33, G. Zinovjev 2, N. Zurlo 140 Y. Sibiriak 87, S. Siddhanta 54, T. Siemiarczuk 84, D. Silvermyr 80, G. Simatovic 89, G. Simonetti 33,104, R. Singh 85, R. Singh 100, R. Singh 49, V.K. Singh 141, V. Singhal 141, T. Sinha 109, B. Sitar 13, M. Sitta 31, T.B. Skaali 20, M. Slupecki 126, N. Smirnov 146, R.J.M. Snellings 63, T.W. Snellman 43,126, C. Soncco 111, J. Song 60,125, A. Songmoolnak 115, F. Soramel 28, S. Sorensen 130, I. Sputowska 118, J. Stachel 103, I. Stan 67, P. Stankus 95, P.J. Steffanic 130, E. Stenlund 80, D. Stocco 114, M.M. Storetvedt 35, L.D. Stritto 29, A.A.P. Suaide 121, T. Sugitate 45, C. Suire 61, M. Suleymanov 14, M. Suljic 33, R. Sultanov 91, M. Šumbera 94, S. Sumowidagdo 50, S. Swain 65, A. Szabo 13, I. Szarka 13, U. Tabassam 14, G. Taillepied 134, J. Takahashi 122, G.J. Tambave 21, S. Tang 6,134, M. Tarhini 114, M.G. Tarzila 47, A. Tauro 33, G. Tejeda Muñoz 44, A. Telesca 33, C. Terrevoli 125, D. Thakur 49, S. Thakur 141, D. Thomas 119, F. Thoresen 88, R. Tieulent 135, A. Tikhonov 62, A.R. Timmins 125, A. Toia 68, N. Topilskaya 62, M. Toppi 51, F. Torales-Acosta 19, S.R. Torres 9,120, A. Triﬁro 55, S. Tripathy 49, T. Tripathy 48, S. Trogolo 28, G. Trombetta 32, L. Tropp 37, V. Trubnikov 2, W.H. Trzaska 126, T.P. Trzcinski 142, B.A. Trzeciak 63, T. Tsuji 132, A. Tumkin 108, R. Turrisi 56, T.S. Tveter 20, K. Ullaland 21, E.N. Umaka 125, A. Uras 135, G.L. Usai 23, A. Utrobicic 98, M. Vala 37, N. Valle 139, S. Vallero 58, N. van der Kolk 63, L.V.R. van Doremalen 63, M. van Leeuwen 63, P. Vande Vyvre 33, D. Varga 145, Z. Varga 145, M. Varga-Kofarago 145, A. ALICE Collaboration Petráˇcek 36, M. Petrovici 47, R.P. Pezzi 70, S. Piano 59, M. Pikna 13, P. Pillot 114, O. Pinazza 33,53, L. Pinsky 125, C. Pinto 27, S. Pisano 10,51, D. Pistone 55, M. Płosko ´n 79, M. Planinic 98, F. Pliquett 68, J. Pluta 142, S. Pochybova 145,i, M.G. Poghosyan 95, B. Polichtchouk 90, N. Poljak 98, A. Pop 47, H. Poppenborg 144, S. Porteboeuf-Houssais 134, V. Pozdniako S.K. Prasad 3, R. Preghenella 53, F. Prino 58, C.A. Pruneau 143, I. Pshenichnov 62, M. Puccio 25,33, J. Putschke 143, R.E. Quishpe 125, S. Ragoni 110, S. Raha 3, S. Rajput 100, J. Rak 126, A. Rakotozaﬁndrabe 1 L. Ramello 31, F. Rami 136, R. Raniwala 101, S. Raniwala 101, S.S. Räsänen 43, R. Rath 49, V. Ratza 42, I. Ravasenga 30,89, K.F. Read 95,130, A.R. Redelbach 38, K. Redlich 84,v, A. Rehman 21, P. Reichelt 68, F. Reidt 33, X. Ren 6, R. Renfordt 68, Z. Rescakova 37, J.-P. Revol 10, K. Reygers 103, V. Riabov 97, T. Richert 80,88, M. Richter 20, P. Riedler 33, W. Riegler 33, F. Riggi 27, C. Ristea 67, S.P. Rode 49, M. Rodríguez Cahuantzi 44, K. Røed 20, R. Rogalev 90, E. Rogochaya 75, D. Rohr 33, D. Röhrich 21, P.S. Rokita 142, F. Ronchetti 51, E.D. Rosas 69, K. Roslon 142, A. Rossi 28,56, A. Rotondi 139, A. Roy 49, P. Roy 109, O.V. Rueda 80, R. Rui 24, B. Rumyantsev 75, A. Rustamov 86, E. Ryabinkin 87, Y. Ryabov 97, A. Rybicki 118, H. Rytkonen 126, O.A.M. Saarimaki 43, S. Sadhu 141, S. Sadovsky 90, K. Šafaˇrík 36, S.K. Saha 141, B. Sahoo 48, P. Sahoo 48,49, R. Sahoo 49, S. Sahoo 65, P.K. Sahu 65, J. Saini 141, S. Sakai 133, S. Sambyal 100, V. Samsonov 92,97, D. Sarkar 143, N. Sarkar 141, P. Sarma 41, V.M. Sarti 104, M.H.P. Sas 63, E. Scapparone 53, B. Schaefer 95, J. Schambach 119, H.S. Scheid 68, C. Schiaua 47, R. Schicker 103, A. Schmah 103, C. Schmidt 106, H.R. Schmidt 102, M.O. Schmidt 103, M. Schmidt 102, N.V. Schmidt 68,95, ALICE Collaboration / Physics Letters B 807 (2020) 135564 8 Y. Sibiriak 87, S. Siddhanta 54, T. Siemiarczuk 84, D. Silvermyr 80, G. Simatovic 89, G. Simonetti 33,104, R. Singh 85, R. Singh 100, R. Singh 49, V.K. Singh 141, V. Singhal 141, T. Sinha 109, B. Sitar 13, M. ALICE Collaboration Sitta 31, T.B. Skaali 20, M. Slupecki 126, N. Smirnov 146, R.J.M. Snellings 63, T.W. Snellman 43,126, C. Soncco 111, J. Song 60,125, A. Songmoolnak 115, F. Soramel 28, S. Sorensen 130, I. Sputowska 118, J. Stachel 103, I. Stan 67, P. Stankus 95, P.J. Steffanic 130, E. Stenlund 80, D. Stocco 114, M.M. Storetvedt 35, L.D. Stritto 29, A.A.P. Suaide 121, T. Sugitate 45, C. Suire 61, M. Suleymanov 14, M. Suljic 33, R. Sultanov 91, M. Šumbera 94, S. Sumowidagdo 50, S. Swain 65, A. Szabo 13, I. Szarka 13, U. Tabassam 14, G. Taillepied 134, J. Takahashi 122, G.J. Tambave 21, S. Tang 6,134, M. Tarhini 114, M.G. Tarzila 47, A. Tauro 33, G. Tejeda Muñoz 44, A. Telesca 33, C. Terrevoli 125, D. Thakur 49, S. Thakur 141, D. Thomas 119, F. Thoresen 88, R. Tieulent 135, A. Tikhonov 62, A.R. Timmins 125, A. Toia 68, N. Topilskaya 62, M. Toppi 51, F. Torales-Acosta 19, S.R. Torres 9,120, A. Triﬁro 55, S. Tripathy 49, T. Tripathy 48, S. Trogolo 28, G. Trombetta 32, L. Tropp 37, V. Trubnikov 2, W.H. Trzaska 126, T.P. Trzcinski 142, B.A. Trzeciak 63, T. Tsuji 132, A. Tumkin 108, R. Turrisi 56, T.S. Tveter 20, K. Ullaland 21, E.N. Umaka 125, A. Uras 135, G.L. Usai 23, A. Utrobicic 98, M. Vala 37, N. Valle 139, S. Vallero 58, N. van der Kolk 63, L.V.R. van Doremalen 63, M. van Leeuwen 63, P. Vande Vyvre 33, D. Varga 145, Z. Varga 145, M. Varga-Kofarago 145, A. Vargas 44, M. Vasileiou 83, A. Vasiliev 87, O. Vázquez Doce 104,117, V. Vechernin 112, A.M. Veen 63, E. Vercellin 25, S. Vergara Limón 44, L. Vermunt 63, R. Vernet 7, R. Vértesi 145, L. Vickovic 34, Z. Vilakazi 131, O. Villalobos Baillie 110, A. Villatoro Tello 44, G. Vino 52, A. Vinogradov 87, T. Virgili 29, V. Vislavicius 88, A. Vodopyanov 75, B. Volkel 33, M.A. Völkl 102, K. Voloshin 91, S.A. Voloshin 143, G. Volpe 32, B. von Haller 33, I. Vorobyev 104, D. Voscek 116, J. Vrláková 37, B. Wagner 21, M. Weber 113, S.G. Weber 144, A. Wegrzynek 33, D.F. Weiser 103, S.C. Wenzel 33, J.P. Wessels 144, J. Wiechula 68, J. Wikne 20, G. Wilk 84, J. Wilkinson 10,53, G.A. Willems 33, E. Willsher 110, B. Windelband 103, M. Winn 137, W.E. Witt 130, Y. Wu 128, R. ALICE Collaboration Vargas 44, M. Vasileiou 83, A. Vasiliev 87, O. Vázquez Doce 104,117, V. Vechernin 112, A.M. Veen 63, E. Vercellin 25, S. Vergara Limón 44, L. Vermunt 63, R. Vernet 7, R. Vértesi 145, L. Vickovic 34, Z. Vilakazi 131, O. Villalobos Baillie 110, A. Villatoro Tello 44, G. Vino 52, A. Vinogradov 87, T. Virgili 29, V. Vislavicius 88, A. Vodopyanov 75, B. Volkel 33, M.A. Völkl 102, K. Voloshin 91, S.A. Voloshin 143, G. Volpe 32, B. von Haller 33, I. Vorobyev 104, D. Voscek 116, J. Vrláková 37, B. Wagner 21, M. Weber 113, S.G. Weber 144, A. Wegrzynek 33, D.F. Weiser 103, S.C. Wenzel 33, J.P. Wessels 144, J. Wiechula 68, J. Wikne 20, G. Wilk 84, J. Wilkinson 10,53, G.A. Willems 33, E. Willsher 110, B. Windelband 103, M. Winn 137, W.E. Witt 130, Y. Wu 128, R. Xu 6, S. Yalcin 77, K. Yamakawa 45, S. Yang 21, S. Yano 137, Z. Yin 6, H. Yokoyama 63, I.-K. Yoo 17, J.H. Yoon 60, S. Yuan 21, A. Yuncu 103, V. Yurchenko 2, V. Zaccolo 24, A. Zaman 14, C. Zampolli 33, H.J.C. Zanoli 63, N. Zardoshti 33, A. Zarochentsev 112, P. Závada 66, N. Zaviyalov 108, H. Zbroszczyk 142, M. Zhalov 97, S. Zhang 39, X. Zhang 6, Z. Zhang 6, V. Zherebchevskii 112, D. Zhou 6, Y. Zhou 88, Z. Zhou 21, J. Zhu 6,106, Y. Zhu 6, A. Zichichi 10,26, M.B. Zimmermann 33, G. Zinovjev 2, N. Zurlo 140 A.A.P. Suaide 121, T. Sugitate 45, C. Suire 61, M. Suleymanov 14, M. Suljic 33, R. Sultanov 91, M. Šumbera 94, S. Sumowidagdo 50, S. Swain 65, A. Szabo 13, I. Szarka 13, U. Tabassam 14, G. Taillepied 134, J. Takahashi 122, G.J. Tambave 21, S. Tang 6,134, M. Tarhini 114, M.G. Tarzila 47, A. Tauro 33, G. Tejeda Muñoz 44, A. Telesca 33, C. Terrevoli 125, D. Thakur 49, S. Thakur 141, D. Thomas 119, F. Thoresen 88, R. Tieulent 135, A. Tikhonov 62, A.R. Timmins 125, A. Toia 68, N. Topilskaya 62, M. Toppi 51, F. Torales-Acosta 19, S.R. Torres 9,120, 1 A.I. Alikhanyan National Science Laboratory (Yerevan Physics Institute) Foundation, Yerevan, Armenia 1 A.I. ALICE Collaboration Alikhanyan National Science Laboratory (Yerevan Physics Institute) Foundation, Yerevan, Armenia 2 2 Bogolyubov Institute for Theoretical Physics, National Academy of Sciences of Ukraine, Kiev, Ukraine 3 Bose Institute, Department of Physics and Centre for Astroparticle Physics and Space Science (CAPSS), Kolkata, India 3 Bose Institute, Department of Physics and Centre for Astroparticle Physics and Space Science (CAPSS), Kolkata, India 4 4 Budker Institute for Nuclear Physics, Novosibirsk, Russia 5 California Polytechnic State University, San Luis Obispo, CA, United St 5 California Polytechnic State University, San Luis Ob 6 Central China Normal University, Wuhan, China 7 Centre de Calcul de l’IN2P3, Villeurbanne, Lyon, France entro de Aplicaciones Tecnológicas y Desarrollo Nuclear (CEADEN), Hava 9 Centro de Investigación y de Estudios Avanzados (CINVESTAV), Mexico City and Mérida, Mexico 10 entro de Investigación y de Estudios Avanzados (CINVESTAV), Mexico Cit 10 Centro Fermi – Museo Storico della Fisica e Centro Studi e Ricerche ‘Enrico Fermi’, Rome, Italy 0 Centro Fermi – Museo Storico della Fisica e Centro Studi e Ricerche ‘En 11 Chicago State University, Chicago, IL, United States 12 China Institute of Atomic Energy, Beijing, China 13 Comenius University Bratislava, Faculty of Mathematics, Physics and Informatics, Bratislava, Slovakia 14 13 Comenius University Bratislava, Faculty of Mathematics, Physics an 14 COMSATS University Islamabad, Islamabad, Pakistan 15 Creighton University, Omaha, NE, United States 15 Creighton University, Omaha, NE, United States 16 Department of Physics, Aligarh Muslim University, Aligarh, India 7 Department of Physics, Pusan National University, Pusan, Republic of 18 Department of Physics, Sejong University, Seoul, Republic of Korea 18 Department of Physics, Sejong University, Seoul, Republic of Korea 19 Department of Physics, University of California, Berkeley, CA, United States 20 20 Department of Physics, University of Oslo, Oslo, Norway 21 Department of Physics and Technology, University of Bergen, Bergen, Norway 21 Department of Physics and Technology, University of Bergen, Bergen, Norway 21 Department of Physics and Technology, University of Ber 22 Dipartimento di Fisica dell’Università ‘La Sapienza’ and Sezione INFN, Rome, p p 23 Dipartimento di Fisica dell’Università and Sezione INFN, Cagliari, Italy 23 Dipartimento di Fisica dell’Università and Sezione INFN, Cagliari, Italy 24 Dipartimento di Fisica dell’Università and Sezione INFN, Trieste, Italy 4 Dipartimento di Fisica dell’Università and Sezione INFN, Trieste, Italy 25 Dipartimento di Fisica dell’Università and Sezione INFN, Turin, Italy 25 Dipartimento di Fisica dell’Università and Sezione INFN, Turin, Italy 26 Dipartimento di Fisica e Astronomia dell’Università and Sezione INFN, Bologna, Italy 26 Dipartimento di Fisica e Astronomia dell’Università and Sezione INFN, Bologna, Italy 27 Dipartimento di Fisica e Astronomia dell’Università and Sezione INFN, Catania, Italy 7 Dipartimento di Fisica e Astronomia dell’Università and Sezione INFN, Catania, Italy 28 Dipartimento di Fisica e Astronomia dell’Università and Sezione INFN, Padova, Italy 29 Dipartimento di Fisica ‘E.R. ALICE Collaboration Caianiello’ dell’Università and Gruppo Collegato INFN, Sale 29 Dipartimento di Fisica ‘E.R. Caianiello’ dell’Università and Gruppo Collegato INFN, Salerno, Italy 30 Dipartimento DISAT del Politecnico and Sezione INFN, Turin, Italy 30 Dipartimento DISAT del Politecnico and Sezione INFN, Turin, Italy p , , y 31 Dipartimento di Scienze e Innovazione Tecnologica dell’Università del Piemonte Orientale and INFN Sezione di Torino, Alessandria, Italy 32 31 Dipartimento di Scienze e Innovazione Tecnologica dell’Università d 32 Dipartimento Interateneo di Fisica ‘M. Merlin’ and Sezione INFN, Bari, Italy 32 Dipartimento Interateneo di Fisica ‘M. Merlin’ and Sezione INFN, Bari, Italy 33 European Organization for Nuclear Research (CERN), Geneva, Switzerland 34 Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, University of Split, Split, Cr 35 35 Faculty of Engineering and Science, Western Norway University of Applied Sciences, Bergen, Norway 35 Faculty of Engineering and Science, Western Norway University of 36 Faculty of Nuclear Sciences and Physical Engineering, Czech Technical University in Prague, Prague, Czech Republic 37 Š 36 Faculty of Nuclear Sciences and Physical Engineering, Czec 37 Faculty of Science, P.J. Šafárik University, Košice, Slovakia 37 Faculty of Science, P.J. Šafárik University, Košice, Slovakia 37 Faculty of Science, P.J. ALICE Collaboration Šafárik University, Košice, Slovakia y f , J f y, , 38 Frankfurt Institute for Advanced Studies, Johann Wolfgang Goethe-Universität Frankfurt, Frankfurt, Germany 39 y f J f y 38 Frankfurt Institute for Advanced Studies, Johann Wolfgang 38 Frankfurt Institute for Advanced Studies, Johann Wolfgang 39 Fudan University, Shanghai, China 39 Fudan University, Shanghai, China ALICE Collaboration / Physics Letters B 807 (2020) 135564 9 40 Gangneung-Wonju National University, Gangneung, Republic of Korea 40 Gangneung-Wonju National University, Gangneung, Republic of Korea g g j y g g p 41 Gauhati University, Department of Physics, Guwahati, India 41 Gauhati University, Department of Physics, Guwahati, India y p f y 42 Helmholtz-Institut für Strahlen- und Kernphysik, Rheinische Friedrich-Wilhelms-Universität Bonn, Bonn, Germany 42 Helmholtz-Institut für Strahlen- und Kernphysik, Rhe 42 Helmholtz-Institut für Strahlen- und Kernphysik, Rheinisch 43 Helsinki Institute of Physics (HIP), Helsinki, Finland 43 Helsinki Institute of Physics (HIP), Helsinki, Finland 44 High Energy Physics Group, Universidad Autónoma de Pue 44 High Energy Physics Group, Universidad Autónoma de Puebla, Puebla, Mexic 45 45 Hiroshima University, Hiroshima, Japan y J p 46 Hochschule Worms, Zentrum für Technologietransfer und Telekommunikation (ZTT), Worms, Germany Worms, Zentrum für Technologietransfer und Telekommunikation (ZTT), W f g f ( 47 Horia Hulubei National Institute of Physics and Nuclear Engineering, Bucharest, Ro 47 Horia Hulubei National Institute of Physics and Nuclear Engin 47 Horia Hulubei National Institute of Physics and Nuclear Engi National Institute of Physics and Nuclear Engineering, Bucharest, Roma 48 Indian Institute of Technology Bombay (IIT), Mumbai, India 48 Indian Institute of Technology Bombay (IIT), Mumbai, India 49 Indian Institute of Technology Indore, Indore, India 49 Indian Institute of Technology Indore, Indore, India 50 Indonesian Institute of Sciences, Jakarta, Indonesia 50 Indonesian Institute of Sciences, Jakarta, Indonesia 51 INFN, Laboratori Nazionali di Frascati, Frascati, Italy 51 INFN, Laboratori Nazionali di Frascati, Frascati, Italy 52 INFN, Sezione di Bari, Bari, Italy 53 INFN, Sezione di Bologna, Bologna, Italy 53 INFN, Sezione di Bologna, Bologna, Italy 54 INFN, Sezione di Cagliari, Cagliari, Italy 55 INFN, Sezione di Catania, Catania, Italy 56 INFN, Sezione di Padova, Padova, Italy 57 INFN, Sezione di Roma, Rome, Italy 58 INFN, Sezione di Torino, Turin, Italy 59 INFN, Sezione di Trieste, Trieste, Italy 60 Inha University, Incheon, Republic of Korea O), Institut National de Physique Nucléaire et de Physique des Particules (IN2P3/CNRS), Université de Paris-Sud, Université Paris-Saclay, Orsay, y, , p f 61 Institut de Physique Nucléaire d’Orsay (IPNO), Institut National de Physique Nucléaire et de Physique des Particules (IN2P3/CNRS), Université de Paris-Sud, Université Paris-Saclay, Orsay, France 62 y p f 61 Institut de Physique Nucléaire d’Orsay (IPNO), Institut National de Physique Nucléaire et de Physique des Particules (IN2P3/CNRS), Univ France 62 Institute for Nuclear Research, Academy of Sciences, Moscow, Russia 62 Institute for Nuclear Research, Academy of Sciences, Moscow, Russia 63 Institute for Subatomic Physics, Utrecht University/Nikhef, Utrecht, Ne 64 Institute of Experimental Physics, Slovak Academy of Sciences, Košice, Slovakia 64 Institute of Experimental Physics, Slovak Academy of Sciences, Košice, Slovakia 65 Institute of Physics, Homi Bhabha National Institute, Bhubaneswar, India 65 Institute of Physics, Homi Bhabha National Institute, Bhubaneswar, India 66 Institute of Physics of the Czech Academy of Sciences, 67 Institute of Space Science (ISS), Bucharest, Romania f p ( ) 68 Institut für Kernphysik, Johann Wolfgang Goethe-Universität Frankfurt, Frankfurt, Germany 68 Institut für Kernphysik, Johann Wolfgang Goethe-Universität Frankfurt, Frankfurt, Germany 68 Institut für Kernphysik, Johann Wolfgang Goethe-Un Instituto de Ciencias Nucleares, Universidad Nacional Autónoma de Méx 70 Instituto de Física, Universidade Federal do Rio Grande do Sul (UFRGS), Porto Alegre, Brazil 70 Instituto de Física, Universidade Federal do Rio Grande do Sul (UFRGS), Porto Alegre, Brazil Instituto de Física, Universidade Federal do Rio Grande do Sul (UFRGS), P ( ) g 71 Instituto de Física, Universidad Nacional Autónoma de México, Mexico City, Mexico 71 Instituto de Física, Universidad Nacional Autónoma de México, Mexico City, Mex 72 iThemba LABS, National Research Foundation, Somerset West, Sout 73 Jeonbuk National University, Jeonju, Republic of Korea 73 Jeonbuk National University, Jeonju, Republic of Korea J y J j p f 74 Johann-Wolfgang-Goethe Universität Frankfurt Institut für Informatik, Fachbereich Informatik und Mathematik, Frankfurt, Germany J y J j p f 74 Johann-Wolfgang-Goethe Universität Frankfurt Institut für Informatik, Fachbereich Informatik und Mathematik, Frankfurt, Germany 75 74 Johann-Wolfgang-Goethe Universität Frankfurt Institut 75 Joint Institute for Nuclear Research (JINR), Dubna, Russia J f (J ) 76 Korea Institute of Science and Technology Information, Daejeon, Republic of Korea J f (J ) 76 Korea Institute of Science and Technology Information, Daejeon, Republic of Korea 76 Korea Institute of Science and Technology Information, Dae f 77 KTO Karatay University, Konya, Turkey 78 Laboratoire de Physique Subatomique et de Cosmologie, Université Grenoble-Alpes, CNRS-IN2P3, Grenoble, France 78 Laboratoire de Physique Subatomique et de Cosmologie, Université Grenoble-Alpes, CNRS-IN2 78 Laboratoire de Physique Subatomique et de Cosmologie, Université 79 Lawrence Berkeley National Laboratory, Berkeley, CA, United States 79 Lawrence Berkeley National Laboratory, Berkeley, CA, United States 80 Lund University Department of Physics, Division of Particle Physics, Lund, Sweden 80 Lund University Department of Physics, Division of Particle P 80 Lund University Department of Physics, Division of Par 81 Nagasaki Institute of Applied Science, Nagasaki, Japan 81 Nagasaki Institute of Applied Science, Nagasaki, Japan 82 Nara Women’s University (NWU), Nara, Japan y ( ) J p 83 National and Kapodistrian University of Athens, School of Science, Department of Physics, Athens, G 83 National and Kapodistrian University of Athens, Sch p y f 84 National Centre for Nuclear Research, Warsaw, Poland 84 National Centre for Nuclear Research, Warsaw, Poland f , , 85 National Institute of Science Education and Research, Homi Bhabha National Institute, Jatni, India 86 85 National Institute of Science Education and Researc 85 National Institute of Science Education and Research, Homi Bhabha National Institute, Jatni, India 86 National Nuclear Research Center, Baku, Azerbaijan 87 National Research Centre Kurchatov Institute, Moscow, Russia 87 National Research Centre Kurchatov Institute, Moscow, Russia 88 88 Niels Bohr Institute, University of Copenhagen, Copenhagen, Denmark 88 Niels Bohr Institute, University of Copenhagen, Copenhagen, Denmark 89 Nikhef, National institute for subatomic physics, A 89 Nikhef, National institute for subatomic physics, Amsterdam, Nethe 89 Nikhef, National institute for subatomic physics, Amsterdam, Neth 90 NRC Kurchatov Institute IHEP, Protvino, Russia 91 NRC «Kurchatov Institute» - ITEP, Moscow, Russia 92 NRNU Moscow Engineering Physics Institute, Moscow, Russia 92 NRNU Moscow Engineering Physics Institute, Moscow, Russia 93 Nuclear Physics Group, STFC Daresbury Laboratory, Daresbury 93 Nuclear Physics Group, STFC Daresbury Laboratory, Daresbury, United Kingdom 94 Nuclear Physics Institute of the Czech Academy of Sciences, ˇRež u 95 Oak Ridge National Laboratory, Oak Ridge, TN, United States 96 Ohio State University, Columbus, OH, United States 96 Ohio State University, Columbus, OH, United States 97 Petersburg Nuclear Physics Institute, Gatchina, Russi g y 98 Physics department, Faculty of science, University of Zagreb, Zagreb, Croatia 98 Physics department, Faculty of science, University of Zagreb, Zagreb, Croatia 98 Physics department, Faculty of science, University of y p y f y f g 99 Physics Department, Panjab University, Chandigarh, India 100 Physics Department, University of Jammu, Jammu, India 101 Physics Department, University of Rajasthan, Jaipur, India 101 Physics Department, University of Rajasthan, Jaipur, India 102 Physikalisches Institut, Eberhard-Karls-Universität Tübingen, Tübingen, Germany 102 Physikalisches Institut, Eberhard-Karls-Universität Tübingen, Tübingen, Germany 3 Physikalisches Institut, Ruprecht-Karls-Universität Heidelberg, Heidelb 104 Physik Department, Technische Universität München, Munich, Germany 104 Physik Department, Technische Universität München, Munich, Germany 105 Politecnico di Bari, Bari, Italy 106 Research Division and ExtreMe Matter Insti 106 Research Division and ExtreMe Matter Institute EMMI, GSI Helmholt 107 Rudjer Boškovi´c Institute, Zagreb, Croatia 108 Russian Federal Nuclear Center (VNIIEF), Sarov, Russia 110 School of Physics and Astronomy, University of Birmingham, Birmingham, United Kingdom 110 School of Physics and Astronomy, University of Birmingham, Birmingham, United Kingdom 111 Sección Física, Departamento de Ciencias, Pontiﬁcia Universidad Católica del Perú, Lim 112 111 Sección Física, Departamento de Ciencias, Pontiﬁcia Universidad C 112 St. ALICE Collaboration Petersburg, Russia g y g 113 Stefan Meyer Institut für Subatomare Physik (SMI), Vienna, Austria 113 Stefan Meyer Institut für Subatomare Physik (SMI), Vienna, Austri 114 SUBATECH, IMT Atlantique, Université de Nantes, CNRS-IN2P3, Nantes, Franc 114 SUBATECH, IMT Atlantique, Université de Nantes, CNRS-IN2P3, Na 115 Suranaree University of Technology, Nakhon Ratchasima, Thailand 115 Suranaree University of Technology, Nakhon Ratchasima, Thailand 115 Suranaree University of Technology, Nakhon Ratc 116 Technical University of Košice, Košice, Slovakia 117 Technische Universität München, Excellence Cluster ‘Universe’, Mu 117 Technische Universität München, Excellence Cluster ‘Universe’, Munich, Germ 118 118 The Henryk Niewodniczanski Institute of Nuclear Physics, Polish Academy of Sciences, Cracow, Polan 118 The Henryk Niewodniczanski Institute of Nuclear Physics, Polish Academy of Sciences, Cracow, Poland ALICE Collaboration / Physics Letters B 807 (2020) 135564 10 119 The University of Texas at Austin, Austin, TX, United States 120 Universidad Autónoma de Sinaloa, Culiacán, Mexico 121 Universidade de São Paulo (USP), São Paulo, Brazil 122 Universidade Estadual de Campinas (UNICAMP), Campinas, Brazil 123 Universidade Federal do ABC, Santo Andre, Brazil 124 University of Cape Town, Cape Town, South Africa 125 University of Houston, Houston, TX, United States 126 University of Jyväskylä, Jyväskylä, Finland 127 University of Liverpool, Liverpool, United Kingdom 128 University of Science and Technology of China, Hefei, China 129 University of South-Eastern Norway, Tonsberg, Norway 130 University of Tennessee, Knoxville, TN, United States 131 University of the Witwatersrand, Johannesburg, South Africa 132 University of Tokyo, Tokyo, Japan 133 University of Tsukuba, Tsukuba, Japan 134 Université Clermont Auvergne, CNRS/IN2P3, LPC, Clermont-Ferrand, France 135 Université de Lyon, Université Lyon 1, CNRS/IN2P3, IPN-Lyon, Villeurbanne, Lyon, France 136 Université de Strasbourg, CNRS, IPHC UMR 7178, F-67000 Strasbourg, France 137 Université Paris-Saclay Centre d’Etudes de Saclay (CEA), IRFU, Départment de Physique Nucléaire (DPhN), 138 Università degli Studi di Foggia, Foggia, Italy 139 Università degli Studi di Pavia, Pavia, Italy 140 Università di Brescia, Brescia, Italy 141 Variable Energy Cyclotron Centre, Homi Bhabha National Institute, Kolkata, India 142 Warsaw University of Technology, Warsaw, Poland 143 Wayne State University, Detroit, MI, United States 144 Westfälische Wilhelms-Universität Münster, Institut für Kernphysik, Münster, Germany 145 Wigner Research Centre for Physics, Budapest, Hungary 146 Yale University, New Haven, CT, United States 147 Yonsei University, Seoul, Republic of Korea 119 The University of Texas at Austin, Austin, TX, United States 120 Universidad Autónoma de Sinaloa, Culiacán, Mexico 121 Universidade de São Paulo (USP), São Paulo, Brazil 122 Universidade Estadual de Campinas (UNICAMP), Campinas, Brazil 123 Universidade Federal do ABC, Santo Andre, Brazil 124 University of Cape Town, Cape Town, South Africa 125 University of Houston, Houston, TX, United States 126 University of Jyväskylä, Jyväskylä, Finland 127 University of Liverpool, Liverpool, United Kingdom 128 University of Science and Technology of China, Hefei, China 129 University of South-Eastern Norway, Tonsberg, Norway 130 University of Tennessee, Knoxville, TN, United States 131 University of the Witwatersrand, Johannesburg, South Africa 132 University of Tokyo, Tokyo, Japan 133 University of Tsukuba, Tsukuba, Japan 134 Université Clermont Auvergne, CNRS/IN2P3, LPC, Clermont-Ferrand, 135 Université de Lyon, Université Lyon 1, CNRS/IN2P3, IPN-Lyon, Villeur 136 Université de Strasbourg, CNRS, IPHC UMR 7178, F-67000 Strasbour 137 Université Paris-Saclay Centre d’Etudes de Saclay (CEA), IRFU, Dépa 138 Università degli Studi di Foggia, Foggia, Italy 139 Università degli Studi di Pavia, Pavia, Italy 140 Università di Brescia, Brescia, Italy 141 Variable Energy Cyclotron Centre, Homi Bhabha National Institute, K 142 Warsaw University of Technology, Warsaw, Poland 143 Wayne State University, Detroit, MI, United States 144 Westfälische Wilhelms-Universität Münster, Institut für Kernphysik, 145 Wigner Research Centre for Physics, Budapest, Hungary 146 Yale University, New Haven, CT, United States 147 Yonsei University, Seoul, Republic of Korea 119 The University of Texas at Austin, Austin, TX, United States 120 Universidad Autónoma de Sinaloa, Culiacán, Mexico 121 Universidade de São Paulo (USP), São Paulo, Brazil 122 Universidade Estadual de Campinas (UNICAMP), Campinas, Brazil 123 Universidade Federal do ABC, Santo Andre, Brazil 132 University of Tokyo, Tokyo, Japan , , y 141 Variable Energy Cyclotron Centre, Homi Bhabha National Institute, Kolkata, India 144 Westfälische Wilhelms-Universität Münster, Institut für Kernphysik, Münster, Germany 145 45 Wigner Research Centre for Physics, Budapest, Hungary 146 Yale University, New Haven, CT, United States 14 147 Yonsei University, Seoul, Republic of Korea i Deceased. ALICE Collaboration Petersburg State University, St. ALICE Collaboration ii Dipartimento DET del Politecnico di Torino, Turin, Italy. ii Dipartimento DET del Politecnico di Torino, Turin, Italy. University, D.V. Skobeltsyn Institute of Nuclear, Physics, Moscow, Ru iv Department of Applied Physics, Aligarh Muslim University, Aligarh, India. v Institute of Theoretical Physics, University of Wroclaw, Poland. v Institute of Theoretical Physics, University of Wroclaw, Poland.
W2109267573.txt	https://zenodo.org/records/2279475/files/article.pdf	de		Ueber intraabdominalen Druck	Archiv für Gynäkologie	1,912	public-domain	13,130	Ueber intraabdominalen Druck. Von L. Kaiser, Fraaenarzt in Amsterdam. (Mit 14 Textfigurem) In vorliegender Arbeit erlaube ieh mir, die Ergebnisse einiger Versuehe fiber den intraabdominalen Druek~ sowie die Sehlfisse~ zu welehen sie reich geffihrt haben, mitzuteilen. Herrn Professor H. A. L o r e n t z in Leiden 7 der meine Betraehtungen veto phySikalisehen Standpunkte naehzuprfifen die Giite hatt% und Herrn Prof. Dr. G. van R y n b e r k in kmsterdam~ der mir erlaubte~ inseinem Laboratorium zu arbei~en, spreehe ieh bier meinen herz!leben Dank aus. Der Besehreibung meiner Versuehe und den dadureh veranlassten Betraehtungen mfge eine kurze Uebersieht fiber die auf diesem Gebiete bereits gemaeh~en Beobaehtungen nnd eine Aus~inandersetzung der jetzt vielfaeh herrsehenden knsichten vorangesehiek[ werden. Im aahre 1865 publizierte W. B r a u n e im Zentralblatt der medizinisehen Wissensehaften und spgter in einer Sehrift ,Die Obersehenkelvene des Nensehen und ihre Beziehungen usw. ~ die Resultate seiner Untersuehungen fiber den Druek~ unter dem die Contenta des Mastdarmes stehen. B r a u n e verfuhr bei seinen Versuehen in folgender Weise: Er fiihrte einen Sehlaueh ein, der mit einer graduierten Glasrfhre in Verbindung stand, ftillte die Rfhre mit Wasser, und indem er den Wasserstcand ablas~ bestimmte er den Druek, unter dem das in den Mastdarm gelaufene Wasser sieh befand. In dieser Weise kam er zu den folgenden Sehltissen: 1. Die Druekhfhe betrug bei aufreehter Stellung etwa 40 era. Dies kam der Entfernung zwisehen dem Proeessus xiphoideus und tier Ampulla ree~i nahezu gleieh. 302 t( ai s er, Ueber in~raabdominalen Druek. 2. Sie blieb sich, yon kleinen Sehwankungen abgesehen, bei versehiedenen Individuen derselben K6rpergr6sse gleieh. 3. Sie blieb ungef~thr dieselbe~ aueh wenn man bis 2 Liter Wasser in den Darm Iaufen liess. 4. Sie ~nderte sieh wesentlieh, wenn man dem K6rper eine horizontale Lage ~ab. 5. Unter dem Gefiihle des Dr~ngens traten, ohne dass willkiirlieh etwas dazu beigetragen wurde, Bewegungen der Wassers~iule auf~ wobei sieh die Druekh6he bis fiber 60 em steigerte. 6. Dureh willkiirlieh herbeigefiihrte Bauehpresse vermoehte man die Wassers~tule his fiber 100 em zu erheben. Spgter (1872) hat Sehatz im Arehiv Nr Gynttkologie eine ausfiihrliehe Abhandlung fiber denselben Gegenstand ver~ffentlieht~ in der er zunXehst B r a u n e s Ansieht, dass der gelegentlieh vorkommende negative (unteratmosph/trisehe) Druek in der Vena eruralis einen negativen Druek in der Bauehh6hle beweisen solle, grfindlieh widerlegt. Uebrigens ergaben seine eigenen Beobaehtungen genau dieselben l{esultate wie die Messungen B r a u n es. S c h a t z hat aueh den Druek in der Harnblase gemessen~ wobei er folgende Resultate erhielt: Bei den versehiedenen K6rperlagen stieg das Wasser in der Glasr6hre his zur H6he der bezeiehneten XSrperteile. Aufrechte HMtung - - Niveau auf der H6he des Proeessus xiphoideus. Horizontale Lage - - Niveau unter dem Nabel. Lage auf der Seite - - Niveau in der Mitre zwischen der naeh oben gewandten Spina ant. sup. os. ilei und dem Nabel. [,age auf Hgnden nnd Knien - - Niveau ungefghr auf der H6he tier Vorderseite tier Lendenwirbelk6rper. Welter mass Sehatz den Druek im l{eetum, sowohl bei aufreeh~ stehendem 7 wie aueh bei naeh vorn oder hinten gebeugtem Oberk6rper. Die yon ibm el'haltenen l{esultate sind in naehstehender Figur zusammengestellt~ in der die Linien die Riehtung des Oberk6rpers bezeiehnen (und zwar denke man sieh den Kopf, wo die Zahlen skehen)~ w~ihrend die Zahlen den Rektaldruek an~,eben. K ai s e r, Ue~oer ~ n t r a ~ a b d o m i n a l e n Druck. 30 3 3~ 4 t4 Kniehandlage Rfiekenlage +B (Zu bemerken ist, class man nur bei aktiver Bewegung die grosse Erh~hung bei nach hinten gebeugtem Rumple sieht). gurde der g6rper in gestreckter Haltung in versehiedenen Richtungen aufgestellt, so erhielt er obige Ziffern. Scha~z hat eine grosse Zahl derartiger 5Iessungen ausgefiihrt und die Resul{ate in dieser sehr iibersiehtliehen Weise zusammengestellt. Um die zahlreichen ]~rgebnisse besser im Ged~tehtnis behalten zu k6nnen~ sagt er noch~ class der Druck an einer bestimmten Stelle in einer gewissen Haltung beinahe immerfort genau so gross sei, wie er in einer mit gasser geftillten BauehhOhle sein wiirde. S e h w e r d t hat im Jahre 1896 in seiner AbMndlung iiber Enteroprose, die Druckverhgltnisse im Abdomen einem eingehenden Studium unkerworfen. Aus den yon ihm mitgeteilten Beobachtungen gebe ieh hier eine Zahlenreihe, welehe die yon ihm bei verschiedenen Personen gefundenen Druckgr/Sssen im geetmn beim Stehen und beim Liegen betriftt. Stehen Liegen 39 em Wasser 28 . ,, 30 . . . . 31 . . . . 21 ,, ,, 24 em Wasser 1 5 ,, . 15 . . . . 18 . . . . 1 6 ,, . 30 28 26 26 22 24 21 19 16 18 . . . . . . . . . . . . ,, . . . . . ,, . ,, ,, ,, ,, ,, . ,, ,, Ausserdem ist ein Versuch von S c h w e r d t zu erwghnen, bei dem ein kr'Mtiger 5Iann im )lagen einen Druck yon 11: in der Blase yon 29 und im Rectum ,,'on 39 cm zeigte. Als der Mann sich hinlegte, stieg der Druek im 5Iagen his 291/2 em~ im Reetnm f~el er his 11 cm. 304 Kaiser, Ueber ini:raabdominalenDruek. H 6 r m a n n , der mit einem gew6hnliehen W~ssermanometer mass, das er r naehdem er Luft in das Rectum geblasen hatte, mittels eines Schlauches mit einer Rektalsonde verband, land bet stehender Lage Zahlen, die yon 18 his 34 em Wasser variierten und im Durchschni~t 22~2 betrugen. W e i s k e r beobaehtete bet einem Manne einen Rektaldruek yon 29--33 cm; bet einem andern 22--24 cm; die Zahlen ~nderten sieh nieht, wenn er die Patienten grosse Mengen Wasser trinken liess. Bet stehender Lage land R i c h a r d K n o r r e]nen Blasendruek yon 30 cm, bet liegender yon 10--15 cm. Dass der Druck in der ttarnblase Dis zu ether F~llung yon 250 ccm~ 13--15 eta Wasser (helm Stehen e~was mehr) bet riigt, crw/ihnt Z w a a r d e m a k e r . O. Ktistner suehte mit cinem Irrigator und einem Schlauch den Druck im Rectum zu ermitteln, er bestimmte n~tmlich die H/She, in d e r e r den Irrigator halten musste: damit das Wasser weder in das Rectum hinein, noch aus demselben herauslief. Im Stehen land er 40 era, im Liegen 15 era. Der geiz des einstr6menden Wassers wird wahrscheinlieh K0ntraktioneo des Rectums ausgel6st haben, die, wie wir sp~tter erSrtern werden: Einfluss auf die Beobaehtungsresultate haben k6nnen. B r a u n e leitee aus seinen Messungen des Rektaldruckes und aus tier beobachtetea Atemschwankung der Venen des Beines, im Zusammenhang mit dem gelegentlich yon ihm konstatierten unteratmosph~trisehen Druck in der Vena eruralis: ab, dass wXhrend der Ruhe tiberhaup[ kein Druek (tiberatmosphiirischer Druck) in dem Bauehe herrsehe; nur w/thrend tier Bewegung sollen sieh die Muskeln der Bauchwand kontrahieren und den Bauchinhalt zusammenpressen, was sich aueh dutch das Ansehwelien der Venen der Unterextremit~ten zeigte. Wit' erwiihnten schon, dass die Beobachtungen S c h a t z ' s dieselben Resultate wie die 3Iessungen B r a u n e s ergaben; S c h a t z geht aber yon ether Voraussetzung aus~ die ieh nicht fiir richtig halte. Er glaubt n~tmlieh, dass eine regelmitssige Verteilung des Druekes naeh hydrostatischen Gesetzen nut dann im Bauche bestehen k6nne, wenn tier Darmkanal fortwiihrend ganz mit Fltissigkeit gefiillt w'//re, in Wirkliehkeit seien abet die aufeinandergesehiehteten Darmschlingen mit ihrem ha lbfliissigen Inhalt wohl niemals geschmeidig genug, um den ganzen Bauchinhalg im grossen und ganzen~ was die Uebertragung des Druekes betrifft, als eine kontinuierliche halbfl/issige Masse betrachten z~ kiSnnen. Kaiser, Ueber intraaSdominalen Druck. 305 Indes dr/ingt sich uns bei genauer Betrachtung der yon S e h a t z erhaltenen gesultate tier Gedanke auf~ dass man es hier mit einem, im wesentlichen yon dem Gewichte der Eingeweide ahh~ngigen hydrostatisehen Druek zu tun habe. S c h a t z verwirR diese Auffassung, weil er sieh einen a i r anderen Einfliisse tiberherrschenden~gleichmgssigen ~Spannungsdruck ~< vorsteiIL Er Nhrt alle bei einem Wechsel der KSrperstetlung auftretenden DruekXnderungen zuriiek auf den gexnderten Kontraktionszustand der Bauchmuskeln, die den 0berkSrper aufreeht auf dem Becken zu erhalten haben (genau so wie die Streekseile den Mast eines Sehiffes), wobei versehiedene Neigungen des KSrpers einen versehiedenen Spannungsgrad der Muskeln erfordern. Auch babe das Gewieht der Bauehdeeken und des Brustkorbes dann und wann einen Eiufluss. Der in der ganzen BauchhShle herrsehende uniforme ,,Spannungsdruek" sei zwar in den versehiedenen Positionen versehieden~ abet jedesmal nut abhgngig yon tier Spannung und tier Last der Bauch- und Brustwand; das Gewieht der Eingeweide babe nur geringen Einfluss. Wenden wir uns jetzt den Ansi&ten spgterer Autoren zu~ die diesen Einfluss nieht so gering anschlagen. S c h w e r d t unterseheidet einen, dureh die elastisehe Spannung der Bauehmuskeln und des Zwerchfells hervorgebrachten ,~Spannungsdruck~: der gleiehmXssig tiber die ganze BauchhShle verteilt sein soll, und einen sieh dazu addierenden, yon dem Gewicht des Bauchinhaltes herriihrenden ~Belastungsdruck"~ der nicht gleichmgssig verteilt ist~ sonderu im unteren Teil des Bauchraames (oberhalb des Beekenbodens) am grSssten und im oberen Tell (unter dem Diaphragma) gleich Null ist. Zu diesen beiden Ursachen des intraabdominaten Druekes kommt nach S e h w e r d t noch eine dritte. Infolge einer bei den G/~rungsprozessen im Digestionskanal auftretenden Gasentwieklung kann n~tmlich ein ,,Aufbls entstehen. Allerdings ist mir die Unabh~ngigkeit dieses letzteren vom Spannungsdruek nieht reeht Mar geworden. Jedenfalls ist durch die Unterscheidung yon Belastungsdruck nnd Spannungsdruek das uns besch~ftigende Problem um einen Schritt weitergeNhrt worden. Es wird jetzt klar, dass es nut yon der GrSsse des Spannnngsdruekes abh~ingt~ ob der intraabdominale Druek stellenweise unteratmosphXrisch (negativ) wird. 306 if a i s er, Ueber intr~abdominalea Druck. Von jeher ist dann und warm eine Beo[~aehtnng yon einem gelegentlieh sich in der Bauehh~Shle zeigenden negativen Druek mitgeteilt worden. Sehatz hebt hervoq dass tiefe Inspirationen und Lagerung auf der Seite im Woehenbette NegativitXt des Bauel~druekes hervorruNn kSnne, aueh erwghn~ er das yon ibm wiederholt beobaehtete Einstramen yon Luft in die Geb'armutterh/Shle wfihrend der Einfiihrung eines Zangenl6ffets. Unzweideutiger ist die yon S e h a t z erw~thnte Nitteilung H a m e r n j k s : der in ,,Das Herz und seine Bewegungen c' sagf: Wenn man einem l(e~daver die Bauehh/Jhle 6ffnet~ so verkleinert sigh der Brustkorb noeh weiter~ wie alas an einem an tier Trachea befestigten 3lanometer sichtbar ist. Wolkow und D e l i t z i n fanden~ dass bei h'angenden Leiehen naeh Aufheben des hermetisehen Versehlusses des Bauehes das Zwerehfell vide Zentimeter in die H/She riiekte. Weisker land im 5Iagen oft einen negativen Druek 7 eine Beobaehtung, die _~Ioritz und S c h r e i b e r bei aufrechter Stellung bestgtigten, w~ihrend sie bei liegender Haltung oder beim auf dem I(opfe Stehen einen positiven Druek im Magen fanden. H e g a r besehreibt ausftihrlieh~ wie in der yon Sims eingeftthrten I(niehandlage der Druek in tier Vagina und dem Rectum stark negativist. Beim t(atheterisieren in der Riiekenlage sind Bauehdruek und Blasentonus oft nieht imstande 7 den Urin nut wenige Zentimeter aufsteigen zu ]assen, ja Luftaspiration in die Blase ist beobaehtet worden (Odebrecht). P a w l i k - K e l l y s Luftkystoskopie setzt negativen Druek roraus und bei Garrulitas Vulvae und Ruetus hysterieus wird Luft in die Vagina bezw. in den ~'Iagen aspiriert. Aueh in dec T r e n d e l e n b u r g s e h e n Lagerung erfahren wir, wie die Last der Baueheingeweide die Verteilung des Druekes in der Bauchh/Shle beeinflussL~ und wir sehen subatmosphgrisehen Drock im Operadonsgebiete. knsser dureh seine ~Iessungen bei Menschen hat H O r m a n n dutch Punktion dutch die Bauehdeoke hindureh mit einem an ein Manometer verbundenen Troikart das stellenweise Bestehen yon negativem Druek in der Bauchhahle bei Tieren dargetan. Obsehon aueh noch andere Forseher einen subatmosph/irisehen Druck zu konstatieren Gelegenheit hatten~ herrseht doeh im all- Kaiser, Ueber intraabdominalenDruck. 307 gemeinen die A.nsieht~ dass ein positiver Druek unter dem Di~phragma tlegel ist. Ira Zentralblatt fiir Gyn~ikologie 1902 sag~ H a g e n - T h o r n : ,~Jedenfalls besteht in der Ba-nehh{Shle ein bestgndiger allgemeiner Druekwechsel (dureh Atmung und Bewegung) wie in allen anderen K/Srperh{Shlen, und zwar fiir gew~Shnliehein positiver Druek, weleherbei gewissen UffJs~:gnden auch negativ werden kann. ~ Yon K o s s m a n n und R. Neyer wird kurz naehher in der nXmliehen Zeitschrift der negative Druck ein Phantasiegespinst genannt. Als yon entseheidender Bedeutung werden meistens die Beobaehtungen J. R o s e n t h a l s angef~hrt. Diese wurden abet zu dem Zweeke unternommen, den intrathorae~len Druek dutch Messungen im Oesophagus zu bestimmen und nur nebenbei ergab sieh aueh die CarSsse des Druckes im Nagen. Bei zwei Patienten Leubes, die nach oft wiederholter EinNhrung der i'~Iagensonde dieselbe ohne g~irg'en oder Erbrechen ertrugen, sah l~osenthal, wenn er den eingef~ihrten Schlaueh mit einem WassermanomeCer verband~ jedesmal eine sehr charakteristisehe Erseheinung. So lange das tPenster des Schlauehes sich. im Oesophagus befand, zeigte das Wasser im o ffenen Schenkel einen negativen Druek an, wurde der Sehlaueh abet weiter ejngefiihrt, so ~nderte sich der Druck, sobald die Cardia iibersehritten wurde und das Sondenfensterin den Nagen kam; dan~ stieg alas gasser im offenen Sehenkel des ~'Ianometers plgtzlieh. hiSher als im gesehlossenen. Ueber die Stellung der Patienten, iiber die ArC nnd Weis% wie die reflektorischen Wtirgbewegungen yon ihnen unterdriick[ wurden 7 finden wit keine Angaben, wohl bemerkC R o s e n t h a l , dass der UnCersehied des Druckes ober- und unterhalb der Cardia~ bei tiefen Inspirationen am deutliehsten war. Jedenfalls war yon einem normglen Verhalten bier nieht dfe 7Rede. g e l l i n g , der die Verh~ltnisse in der Bauehh6hle sehr deutlich beschreibg~ legC Naehdruek darauf, dass wegen tier aufreehCen Haltung die stafisehen Verh'altnisse beim 5Iensehen ganz andere als bei den u sind. Er sagt ungdghr folgendes: Die Leibesh/Shle (Brust- unc~ BauchhiShle zusammen) besteht aus zwei Teilen~ einer Ober- uncl einer LTnterhglfCe~ die zwar nicht gesehieden sind~ aber gesondert: 308 Kai s er, Uebor in~raabdominalenDruek. betraohtet werden miissen~ denn die Oberh~tlfte wird yon dem starren unnachgiebigen Brustkasten, die Unterh~lfte grSsstenteils yon weichen, muskulSsen W/tnden umgeben. Ringsum auf diesen WXnden ruht der atmosphXrische Druek. Die Oberhglfte wird wieder in zwei Teile geteilt durch die Kuppe des Zwerehfelles, deren hSchster Punkt gewShnlieh sieh in HShe der Brustwarze befindet. Die in dem Brustkasten enthaltenen Lungen sind gedehnt und mtissen wegen der hermetisehen Versehliessung des Pleuraraumes an dig InnenflXehe des Brustkorbes angesehmiegt bleiben~ obsehon sie mit /(raft sieh zusammenzuziehen suehen. Da die PleurM~Ohle ringsum gesehlossen und luftleer ist, wird cture]~ diesen elastisehen gug des Lungengewebes eine Saugkraft auf den Brustkasten und auf das Zwerehfell ausgetibt~ die dieses emporzuziehen und jenen einzuziehen such~. Bei allen i~Iensehen erleidet das Zwerehfell fortw~ihrend, aber nieht immer in gleiehem Masse diesen Zug. l~'iathes, der diese Betrachtungen Kellings g~tnzlieh zu den seinigen maeht~ kntipft daran Ansehauungen 5her das Entstehen tier Enteroptose. Er sagt: Da der atmosphXrische Druek ringsum auf die Bauehwand lastet und die Bauehh6hle hermetiseh versehlossen ist, wodureh die Eingeweide dem Zwerehfel[ folgen m~issen, so besteht eine Neigung, die Eingeweide in den Brustkasten hineinzusehieben, Bei einer kr~f~igen Entwieklung des Brustkastens nnd einer normalen Elastizit~it des Lungengewebes kann also ein betr/iehtlieher Tell der Last der Baueheingeweide dutch die Thoraxaspiration gehoben werden und haben die Bauehdeeken nnr einen kleinen Tell dieser Last zu wagen. Stiller, obsehon im Ganzen den Be~raehtungen yon 5lathes beistimmend, ist der Meinung, dass dieser auf die ohne Zweifel wiehtige Z~gkraft tier Lunge im normalen Zus~ande und 3eren Naohlass als Ursaohe der Enteroptose, zuviel @ewioht, legt. Aueh bei andern sieh mit diesem Gegenstande beseh~fftigenden t~'orsehern haben die yon Kelling und N a t h e s vertretenen Ansiehten nut geringen Anklang gefunden. Die yon Henke gegebene Besehreibung des Bauehraumes und tier Vertei[ung der versehiedenen Organe in demselben stimmt meiner ~leinung naeh ganz gut mit dieser Auffassung Kellings iiberein und mSge hier der ttauptsaehe naek eingesehaltet werden. g ai s er, Ueber intraabdominalen Druek. 309 Der Bauehraum ist normaliter an einzelnen Stellen nur wenig fief. Die vordere Bauehwand befindet sieh oberhalb des Nabels in kurzer Entfernung yon der Wirbels~ule und oberhalb des Beekeneingangs reehts und links ganz nahe dem Psoasrande. An diesen engen Stellen ist bei sehlanken Personen nut 1--11/2 em Zwisehenraum~ sodass der Bauehraum abgeteilt ist in einen oberen (unterdiaphragmatischen) Baum~ zwei seitliehe R~ume und in den Unterbaueh. Die D/inndarmsehlingungen sind derartig gruppiere, dass sie die verengten Stellen nur ein oder zwei Nal p~ssieren und sind weiter gr/Ssstenteils im Un~erbauehraum enthalten. In den wenig Raum darbietenden Seitenteilen befinden sich haupts~ehlieh die auf- und absteigenden Sehenkel des Colons, w~ihrend Leber, 5Iagen und Nilz den grossen oberen Bauehraum einnehmen. Ieh babe reich dureh den Gedanken f(ihren lassen, dass der Bauehinhalt in erster Ann~iherung als eine fliissige Masse vom spezifisehen Gewieht eins betraehtet werden kann. Auf die Griinde~ die sieh fiir diesen Satz anfiihren lassen, komme ieh welter unten zuriick. Nimmt man denselben an~ so geniigt offenbar die Kenntnis des Druekes in einem Punkte, um ihn an allen Stellen angeben zn kSnnen, and zwar muss der Druek unter allen Umst~tnden iu allen Punkten einer horizontalen Ebene den gleiehen Weft haben und naeh oben hin regelm~ssig abnehmen, in dem Masse, dass f~ir zwei horizontale Ebenen, die 1 em yon einander entfernt sind, die Druckdifferenz 1 em Wasser betr~tgt. Es ist indes nieht aus dem Auge zu verlieren, dass in dieser Weise nur die Aenderungen des Drucks yon Punkt zu Punkt bekannt werden; es is~ ja klar, dass derselbe an allen Stellen um den gleiehen Betrag erhSht oder erniedrigt werden kann, ohne dass dadurch das Gleichgewicht der Fliissigkeit gestCrt wird. Will man den absolu~en Wert des Druckes ins Auge fassen~ so sind andere Betrachtungen erforderlich. Namentlich finder dann ein Satz, der sich au[ das Gleiehgewieht einer zwei gXume yon einander trennenden Wand bezieht, oft niitzliehe Anwendung. Werden nS~mlieh auf die beiden Seiten Drueke yon ungleioher GrCsse ausge~bt, so erfordert das Gleiehgewieht, dass in der Wand selbst Kr~fte wirksam sind, die fiir jeden Tell derselben eine naeh tier Seite des.grSsseren Druckes bin geriehtete Resultierende ergeben. Dabei kann man auf die Intensit~it tier Kraftwirkungen in tier Wand, oder den ,Tonus ~: derselben schliessen, sobdd man die ~3[0 l{aiser, Uebor iatraabdominalen Druek. Differenz des beiderseitigea Drucke kennt. Umgekehrt k6nnen wir aueh, sobald der Tonus vollstiindig bekannt ist, daraus die Druekdifferenz~ und also~ wenn uns der eine Druek bekannt Jst, die Gr6sse des anderen Druckes ableiten. Was nun diese ]etztere Frage anbelangt, so liegen die Verh/~ltnisse am einfachsten bei einer dtinnen elastisehen Membran, etwa einer gespannten Gummihaut. Bei einer solehen wird die Riehtung der auf einen kleinen Teil wirkenden, aus den Spannungen resultierenden Kraft sofort dutch die Kr~mmung angezeigt; sie ist naeh der konkaven Seite bin geriehtet und auf dieser Seite-muss also der hShere Druek bes[ehen. Entsteht 7 wio das bei lebenden Geweben vorkommt, der Tonus nicht ~ms gewiJhnIiehen elastischen Kr'gften~ sondern a.us 3[uskelkonfiraktion, so kann m~n sieh in einfaehen FXllen noeh Xhnlieher Sehlussfolgerungen bedienen. Im allgemeinen ist dann aber Vorsieht geboten, da eine vollst'andige Besehreibung des Spannungszustandes die genaue Kenntnis des Kontraktionsgrades und der R!ehtung aller in Wirkung tretenden 31uskelbandel erfordern wiirde. Es braueht tibrigens kaum gesagt zu werden, dass die Verh'altnisse im lebenden K/Srper sehr verwiekelt sind. Eine erseh6pfende Bespreehung der DruekverMltnisse im Abdomen hgtte auf eine grosse Zahl yon Faktoren Rtieksieht zu nehmen. Man denke nut an den /iusseren atmosphgrisehen Druek 7 an die Nuskulatur der Bauehwand~ an die Unregelm~ssigkeiten, die daraus entstehen~ dass in Wirkliehkeit die Eingeweide einen gewissen Grad yon Festigkeit haben. Ferner an den Einfluss der bei der G'arung des Darminhaltes auftretenden Gasentwieklungen, und an die Druekgnderungen, die dureh willktirliehe oder unwillkiirliehe Bewegungen (Dr/ingen, Atembewegungen, Pefistaltik) hervorgerufen werden. Wie man sieht, sind die Erseheinnngen/iusserst kompliziert. Es ist gewiss merkwiirdig, wie sieh im gesunden Organismus dot eine Tei[ des (~nzen an den andern a.npasst. WSs der Druek im Bauehraum naeh oben hin regelmgssig abnimmt, hat jede Wand, die 5.ussere sowohl wie jede innere Seheidewand gerade diejenige Gestalt und denjenigen Spannungszustand, die fiir das Gleiehgewieht erforderlieh sind, und dabei besteh~ z. B. im Innern tier Blutgef/isse an jeder Stel]e gerade ein soleher Druek~ dass die Zirkulation ungest6rt vet sieh gehen kann. Bevor wit nun von den hier in Erinnerung gebraehten Prinzipien Kaiser, Ueber intraabdominalen Dru{'~k. 311 einige Anwendungen maehen~ m6ge noeh darauf hingewiesen werden~ dass dig Frage n~eh den U r s a e h e n einer tats~ehlich bestehenden Druekverteilung im ~llgemeinen versehiedene, gleieh bereehtigte, wenn aueh night gleich naheliegende Antworten zul~isst. Is~ z. B. eine vertikale~ unten mit einem Gummiballon versehene GlasrOhre his zu einer gewissen H/She mit g a s s e r gefiillt, so kann man die Ursache des im Innern des Bgllons bestehenden Druekes naeh Belieben in tier Wirkung der Sehwerkra~N auf die Wassers~iul% in dem Dehnungszustand der Gummiwand~ oder in dem inneren Zustande der eingesehlossenen Fl~issigkeit selbst erblicken. Das eine ist ebenso richtig wie das andere; worauf es ankommt, ist nut dies, dass eine bestimm~e WasserhiShe. eine bestimmte Zusammendrttekung des Wassers im Ballon~ und eine bestimmte Ausdehnung der Wand zu gleieher ZeiC bestehen und sigh weehselseitig bedingen. In 5;hnlieher ~Veise kann man aueh einen in der Bauchh6hle bestehenden Druek, je nach dem Gesichtspunkie~ auf den man sieh stellt~ betrachten als hervorgebraeht dureh den Zustand der Bauchwand oder des Zwerehfells~. dureh die Lage des K6rpers, durch das Gewicht der Eingeweide~ dutch eine Atembewegung oder eine Gasentwicklung; mitunter wird man den Druek in Teile zerlegen, deren jeden man einer eigenen Ursache zusehreibt. Fehler sind hierbei nieht zu befiirchten, wenn man nur die allgemeinen Bedingungen des Gleichgewichts im Auge beh'al~, und die angestellte Ueberlegung braueht uns aLlch nicht davon abzuhalten, im Interesse der Klarheit und AnschaulichkeiC dig beobachteten Erseheinungen auf bestimmte Ursaehen zuriiekzuNhren, g i r d in die soeben genannte Glasr/Shre eine weitere Wassermenge hineingegossen, so wird man die Druckerh~Shung der gr/Ssseren I-I{She tier Wassers/i.ule zuschreiben. Dagegen wird man~ wenn das g~sser dureh Zusammendrracknng des Ballons in die H6he getrieben wiirde, in erster Linie an die /iussere zusammendrtiekende I~raft denken. Man k~nn versuehen, sich ohne direkte 5Iessungen eine Vorstellung yon dem Drueke im Abdomen zu verschaffen~ indem man entweder yon den in der Brusth{Shle bestehenden Verh~i.ltnissen oder yon dem Zustande tier J3auchwand ausgeht. Was den ersteren g e g betrifft, so m{Sge zun~tchst daran erinnert werden~ dass das Gungengewebe sich immeq sogar w~hrend der Inspiration zusammenzuziehen sueht, und dass also der Druek in tier Brusth/Shle ausserhalb tier Lunge niedriger sein muss als der Druck in Bronchien und Alveolen. g i r wollen diesen letzteren 312 1(ai s e r, Ueber in~raabdominalen Druck mit p - ] - a bezeichnen, we p den atmosphg.rischen Druek bedeutet, und die GrSsse a, die positi~T oder negativ sein kann, anzeigt, um wieviel der Druek gr6sser oder kleiner als der atmosph/irisehe Druek ist. Ist nun b die genannte Differenz, so hat man fiir den Druek im Pleuraraume p+a.--b oder c = a - - b~ wenn wir das immer vorkommendep fortlassen und gleiehsam den atmosph~irisehen Druek als Nulipunkf, yon dem ab wir den Drnek reehnen, wghlen. W~hrend tier Inspiration ist nun a negativ; dann hat jedenfalls c das gleiehe Vorzeiehen. Bei der Exspiration dagegen ist a positiv. Indes daft man, wie es seheint, annehmen, dass der Wert dieses Gliedes yon b iibertroffen wird, sodass c aneh jetzt negativ wird. Oberhalb des Zwerehfetls bestgnde demnaeh stets ein negativer Druek, und zwar d~irfte derselbe~ naeh den 3Iessungen versehiedener Physiologen zu nrteilen~ einer Wasserh5he yon 8 bis 10 em gleiehkommen und w~ihrend der Inspiration noeh betr~iehtlieh gr6sser seinl). Hieraus liesse sieh nun ohne weiteres der Druek im oberen Teile der BanehhShle ableiten, wenn man den Zustand des Zwerehfells kennte. Ist dieses vollkommen sehlaff, ohne jede Kontraktion, so wird aneh an der Unterseite der negative Druek e bestehen. In gewissem Sinne kSnnte man dann sagen (vgl. K e l l i n g und Mathes), dass die Kraft~ mit der das Lungengewebe sieh zusammenzuziehen oder die Kraft, mit der die Brnstwand sieh zu erweitern sueht (welehe letztere Kraft die Zusammenziehung der Lungen verhindert), die Eingeweid6 der BanehhShle zu einem grSsseren oder kleineren Teile t r g g t . Wit wollen hierzu bemerken, dass ein negativer Druek e = 10 em Wasser fiir die ganze OberflXehe des Zwerehfells, (wenn wir diese auf 400 qem ansehlagen d(irfen), eine Kraft yon 4 kg ergibt. Nimmt man an, dass c bei tiefer Inspiration einen ~dermal grSsseren Weft haben kann, 1) Dieser negative Druck muss in allen Teilen des Pleuraraumes bestehen; er w[irde aueh herrschen in einer fi[issigen Schicht zwisehen beiden PleurabHittern, wenn eine solche vorhanden w~re. Sind die beiden Pleurablotter ohne Zwisehenschicht direkt miteinander in Berfihrung~ so kann man sagen, dass die Kraft% mit denen sie aufeinander wirken~ dieselbe GrSsse und Richtung haben, wie wenn in einer sie trennenden Fl[issigkeit der Druck r best~inde. t( ais e r, Ueber intraabdomin~len Druck. 31 3 so kommt eine Kraft yon 16 kg heraus. Was das Gewieht der Eingeweide betrifft, so hat man fiir die aus dem KiSrper entfernte Leber etwa 1,5 kg gefunden. Es fragt sieh jetzt, welchen Einfluss die Kontraktionen des Zwer&fells haben. Der Zweretffellmuskel~ der sieh einerseits am Ceritrum tendineum~ andrerseits an dem Rippenrand und an den Wirbelk6rpern inseriert, hat bei seiner Kontraktion das Bestreben, die naeh oben konvexe W61bung des Diaphragmas abzuflaehen und zun/i.ehst, w~hrend die Rippen den relativ festen Ansatzpunkt bilden, die Eingeweide des Bauehes naeh unten und vorne zu bewegen. Wenn der Druekuntersehied unterhalb und oberhalb des Diaphragmas gross wird, nimmt der Widerstand, den das Centrnm tendineum beim Tiefersinken erleidet~ zu und wird as mehr fixiert, sodass es dahin kommt, dass die Rippen verhgltnismgssig noeh am leiehtesten zu bewegen sind. Dann haben die Kontraktionen des Mnskels das gesultat, class der Brustkorb gehoben und erweitert wird. Ob das Zwerehfell wirklieh tiefer tritt~ oder ob umgekehrt bei fixiertem Zwerehfell der Brustkasten gehoben wird, hgngt also ab von dem Untersehied des Druekes unterhalb und oberhalb des Zwerehfells. Immerhin darf man abet sagen, dass die Kontraktionen das Zwerehfell naeh unten zu bewegen streben. Ist K die Kraft pro Flgeheneinheit, mit der dies gesehieht, so ergibt si& fiir den Druek im oberen Tell der Bauehh6hle K@c eine GrSsse, die~ je naeh dem Werte yon K 7 ioositiv, oder, ebenso wie c, negativ sein kann. Obgleich der jetzt versuehte Weg die M6gliehkeit eines negativen intraabdominalen Druekes erkennen lXsst, kann er uns doeh nieht zu einer entseheidenden Bestimmung fiihren. Ebensowenig kann man dazu gelangen, wenn man yon der Gestalt der Bauehwand ausgeht, dazu liegen aueh bier die Verh~ltnisse zu verwiekelt. Im allgemeinen diirfte, wie mir seheint~ gelten, dass ein im ganzen konvex gestalteter~ die Brust und das Beeken ringsum iiberragender Baueh, wie man einen solehen besonders bei korpulenten Mg.nnern mittleren Alters antrifft~ auf iiberatmosph~risehen Druek hinweist. Sieht man doeh, wie tier Bauch beim Anwenden der Bauehpresse aueh eine sieh der Kugelform n~hernde Gestalt annimmt. A r c h i v fi~r Gyn~ikologie. Bd. 96. H. 2. 21 3l4 l~a i s er, Ueber in~r~abdominalenDruek. Wenn andel'erseits dutch enorme Zusammenziehung des Darmkanals (z. B. bei 5Ieningitis) oder bei sonstiger Verminderung des Bauchinhaltes (z. B. bei Ph[hisis) die ganze Bauehwand kahnf/irmig eingesunken ist~ kann man auf eine naeh ausw~rts geriehtete Spannung der Bauehdeeke~ und also auf einen un[erafmosphgrisehen Druck im Innern sehliessen Aus dem Gesag~en geht hervor, dass es zur Bestimmung des intraabdominalen Druekes eigener Nessungen bedarf. Selbstverst~ndlieh k6nnen diese nur in der Weise staltfinden~ dass mail das Nessinstrument in eine yon aussen zug/~ngliehe Organh6hle hineinfiihrt~ und es ist Mar, dass man dann immer, um den Druck in der Bauchh6hle zu erhalten~ den gemessenen Wer~ wegen des Tonus tier I-I~ihlenwand zu korrigieren hat. Als HShlungen kommen der 3lagen~ die Harnblase und der ARer in Betraeht. Was den ersteren be~rifft, so hat Kelling oinige ~}Iessungen ausgefiihrt; er erw~hnt als Ninimumdruck 6 his 8 em Wasser. Bei einem 3'Ianne mit }lagenfis~el fand er den Druek abh/ingig yon dem Ptillungsgrad 6 - 1 8 em. Diese Bestimmungen haben indessen fiir unsern Zweck wenig Bodeu[ung. Die weehselnde Spannung des Magens~ der sieh dutch eine Zusammenziehung (Peristole) dem jeweiligen Inha][e f~g~ und die,sen durch wellonartige intermittierende Kon~raktionen (Peristaltik) welter bef6rdert, wird wahrseheinlieh auch grosse individuelle Versehiedenheiten darbielen (Vielesser, Magenatonie) und 5brigens yon dem Fiillungsgrad abhXngig sein. Der Druck~ unter dem der Inhalt der Harnblase steht~ ist yon Sehatz und anderen systematiseh bestimmt worden; die Ergebnisse dieser Beoba&tungen erw~hnte ich sehon teilweise, Die ttarnblase ha~ nieht nut eine yon dem Fiillungsgrad abh~ngige~ meist geringe Wandspannung~ sondern es treten bin und wieder Kontraktionen dot Blasenmuskulatur (Harndrang) auf. W~hrend einer Periode der Ersehlaffung wird also der Harn sieh unter einem Drueke befinden 7 weleher tier Druekver~eilung in der Bauehh6hle nahezu entsprieh~ w~thrend einer Kontraktion is~ der Druck soviel h6her, als dem Kontraktionsgrade tier Harnblasenmuskulatur en~sprieht. Die Druckbestimmungen vieler Forsd~er haben im After stattgeNnden, und Nr diese Zwecke bietet diese OrganhShle grosse VorCeilc. Erstens ist die Ampulla reeti selbst fiir gr6ssere Instrumente leicht zugXnglich. Zweitens wird das eingeffihrte Instrument meistens gut vertragen, und es bringt nieht wie das FAnffihren eines Nagensehlauches oder eines Katheters eine merklidle Vet- Kaiser, Uober iatraabdominalen Druek. 315 gnderung in den Spannungszuscand des Zwerehfells und dee Bauchwand und inf01gedessen in die Druekverhg,ltnisse des Bauehes. Die eingefiihrte Rekta]sonde sehliesst ein ruhiges ganz normales Verhalten nieht aus. Wie hoc,h dieser Vorteil zu veiansehlagen ist~ erf/ihrt man erst~ wenn man sieh mit dergleiehen Untersuehungen besehgftigt. Sehon das Bewusstsein~ untersueht zu werden, hat Einfluss auf den Spannungsgrad der Bauehmuskeln und des Diaphragmas. Ieh babe daher w/ihrend des Messens immer dureh Gesprgehe oder Lektiire die Auflnerksamkeit des Untersuehten abzulenken gesueht. Meiner Erfahrung n a e h kann man die G-r6sse des Rektaldruekes, wie er unter normalen Umst/tnden ist~ nur dann genau bestimmen, wenn man Gelegenheit hat~ die ngmliehe Person wiederholt zu untersuehen. Das Ausser20---19-te-- -- is- l? .... it- -- to-- -- 8-- -- e gew6hnliehe der Untersuehung verliert sich bald, die Messung geht sehnell und akurat yon statten, und am Ende variieren die gefundenen Zahlen nur sehr wenig7 obgleieh die unkon~rollierbare weehselnde Spannung des Rectums immerhin eine Variation yon einigen Zentimetern bedingen wird. Die das Rectum umhttllenden 5Iuskeln haben zwar Perioden yon merkbarer Aktivit~,t~ die sehon ~us anderen Griinden vermieden werden miissen, aber ebeiaso wie die ttarnblasenmuskulatur aueh lange Perioden yon relativ-er Iluhe. Dass Luft zur Fiillung der Sonde und zur Uebertragung des Druekes verwendet wird, ist ein nieht zu unterseMtzender Vorteil. L u e i a n i und P. B e r t sind ebenso wie A. t t o g g e dem t~ek~aldruck genau nachgegangen. Beim Anwenden yon P. B e r t s und yon L u e i a n i s Sonde wird eine einen kleinen Luftraum einsehliessende qummimembran ins geotum eingefiihrK Kontrolle ~1" 316 K ai s e r, Ueber intraabdominalen Druek. fiber den F/illungsgrad desselben scheint mir w~ihrend des Experimentes unm~Sglieh. A. H o g g e Nll~ das yon ihm verwendete Prgservativkondom ganz mit Luft an. Ein so betr~ehtliehes Luftvelum maeht das Registrieren weniger abh/~ngig yon der Genauigkeit der Einstellung des Manometers, bei meinen Versuehen abet reagierte die Rektalmuskulatur dutch Kontraktionen auf jede Dehnnng. Zu genaueren l~{essungen sind also nut die Sonden mit kleinem Luftvolum zu verwenden, besonders wenn Kontrolle fiber die nieht dureh Skybala verhinderte freie Entfaltnng der Nembran mOglieh ist. gel meinen Messungen wandte ieh einen aus einem 12 em langen GlasrShrehen a bestehenden Explorateur an~ der an einem Ende zwei 2 em voneinanderliegende Verbreiterungen yon je 1 em im Durehmesser hat. Zwisehen diesen beiden Ansehwellungen ist das R/Shrehen mehrfaeh durehlSehert. Ueber das verdiekte Ende und fiber die seitlichen Oeffm~ngen wird ein Pingerkondom gestfilpt~ clas dann fest um das R~hrehen zugebunden wird (b). Der Explorateur steht mittels einer langen engen Gummir/~hre mit einem Manometer in Verbindung. Das Einfiihren des mit dem Kondom versehenen Explorateurs in den After geht sehr leieht~ und die im t{ondom enthaltene Luff wird beim Passieren des Sehliessmuskels fast v611ig hinausgetrieben. Das Manometer besteht aus einer zweisehenkeligen sehr dt~nnen Glasr6hre (BarometerrShre); im kurzen Sehenkel betindet sieh eine zylindrische Erweiterung c, welehe soviel Luft enthalten kann~ als ni3tig ist, um das Kondom geniigend zu ftillen, nieht soviel aber~ dass irgend eine Spannung in der Gummimembran entstehen kann. Der andere Sehenkel ist mehr als ein hMbes Meter lang und kalibriert. Zwisehen den beiden Sehenkeln ist ein drittes RShrehen seitlieh angebraeht~ und hierauf steht eine kleine mit Wasser geNllte Spritze. Naehdem tier Explorateur eingeffihrt worden ist, wird dureh die Spritze soviel Wasser hineingetrieben, dass die in c enthaltene Luft entweieht und das Kondom sieh entfaltet. Das weiter eingeffihrte Wasser steigt indessen im langen Sehenkel. Ist das Wasserniveau im kurzen Sehenkel bis 0 gestiegen~ so gibt das Niveau im langen Sehenkel direk~ den Druek an~ unter welehem das im Kondom befindliehe Luftquantum steht. Will man sieh davon (iberzeugen~ dass nichts die Entfaltung Kaiser, Ueber intrgabdominalen Druck. 317 des Kondoms verhindert~ so braucht man nur durch ein Hin- und Herbewegen des Stempels der Spritze das Wasser im kurzen Sehenkel eine gewisse HShe auf- oder absteigen zu lassen und darauf zu achten, ob das Niveau im langen Schenkel genau ebenso viele Zentimeter auf- und absteigt. Um die Atmungsschwankungen des intrarektalen Druekes bequem verfolgen zu kGnnen~ braehte ich das Wasserniveau im kurzen Sehenkel in den obersten Teil der Erweiterung~ also 2 em unter 0. Man hat dann den Vorteil, dass dieser Punkt sehr wenig variiert, w/ihrend das Niveau in der langen, engen R@re grGssere Be~ wegungen macht. Es mGge noch bemerkt werden, dass die Kapillarwirkung in der engen RShre meines Manometers das Wasser gerade um 2 cm hebt, und dass also bei der genannten Einstellung im kurzen Schenkel an der langen R6hre direkt die mit Riicksieht auf diese Wirkung korrigierte SteighGhe abgelesen wird. Dieser Apparat hat sieh als sehr brauehbar und zuverl//ssig erwiesen. Der P~illungszustand und die Beweglichkeit der eingeNhrten Gummiblase lassen sieh fortwghrend kon~rollieren, wS;hrend bei den kleinen Volumsehwankungen der eingefiihrten geringen Luftmenge jeder Reiz und jede Dehnung des Rectums ausgesehlossen ist. Meiner Erfahrung naeh kann ich den Worten C. Hasses: ~Die Messungen im Mastdarm sind mit so vielen Fehlerquellen b ehaftet, dass denselben irgend ein wissenschaftlieher Wert nicht beizumessen ist", nieht beistimmen. In der unten stehenden Tabelle gebe ich einige der yon mir erhMtenen Zahlen. Es sei hierbei hervorgehoben~ dass diese Messungen alte w~hrend der Ruhe und meistens bei Erwachsenen stattfanden, und dass unter den yon mir untersuchten M~innern kein einziger dicker Mann sich befand. Die unter ~Bauchh(ihe" angegebenen Zahlen gebeu die Entfernung der Brustwarzenhorizontale his zum oberen Sehambeinrande an, welehe Entfernung der HGhe der Bauehh~hle yon der Kuppe des Zwerehfelles bis zur Ampulla reeti nahezu gleichkommen wird. Bei No. 4 trat wghrend des Messens nieht tier gewiinsehte Grad der Unbefangenheit ein, die Messungen No. 13 und No. 18 an Kindern misslangen g/inzlich. Die gefundenen Zahlen stimmen sehr gut mit den Ergebnissen tier Messungen anderer Forscher iiberein. Sic sind im allgemeinen etwas niedriger, was uns abet nicht wundern kann~ wenn wir erw~gen~ dass die bei anderen Methoden 318 I~ai s e r, Ueber intraabdominalen Druek. Baueh hShe 1! Nr. Bauehwand ~ = ~ Bemerkungen m~ssig sehlaff m~ssig resistent sehlaff resistent sehlaff resistent miissig resistent mgssig resistent mgssig sehlaff mRssig resistent resistent 34 24 30 3~ 34 26 21 23 30 22,5 21,5 sehlaff schlafi resistent resistent ~9 ehirurg, grankheit gesund Puer.pera 11. Tag gesund Enterop• gesund gesund (Vaginaldruek) gesund Hernie gesund gesund Enteroptose Nephreptose gesund gesund gesund gesund gesund gesund gesund gesund I 1 3 4 5 6 8 9 10 11 12 14 15 16 17 19 20 21 22 23 24 40 36 35 18 43 47 25 42 48 40 16 18 aS 18 14 43 18 15 21 59 21 42,5 39,5 36 l0 31 42 37 35 38 36 37 36 30 35 w. 35 m. 33,5 m. 35 w. 31 m. 37 m. m. i g3 m. ]I 44 m. m~ssig resistent resistent resistent resistent mii,ssig sehlaff resistent 31 26 31 35 26 20,5 31 40 26,5 mehr als bei der meinigen vorkommenden Fehlerquellen derark sind~ dass zu hohe Werte gefunden werden m[issen. Man denke z. B. an dis unbemerkte Spannung der eingefiihrioen Gummiblase bei deren ungleichmgssiger Enffaltung dureh den Druck eyentueller Skybala, sowie an die durch Reizung des Rectums bewirkten Kontraktionen desselben oder der Bauchmuskeln. Die Registrierung der Schwankungen dureh die Atmung und gelegentlich durch die Peristaltik findet meiner Erfahrung naeh am besten so statt~ dass man naeh genauer Messung des Rektaldruekes den im After verbleibenden Explorateur mitt;Is eines engen starken Cxummischlauehes mit einem Wassermanometer mit weiterer tlShre in Verbindung bringt. Itierauf erhSht man den Druek in dem Exploraieur und dem damit verbundenen Schenkel des Ma.nometers durch Einblasen yon Luft~ his das Manometer genau den gemessenen gektaldruek anzeigt. Das dutch die gespiration verursaehte Auf- und kbsteigen des Wassers (welches bei leieht angestrengtem Atmen 1/2--11/2 cm betr/igt) lSsst sieh mittels eines kleinen, mit dem noeh offenen Sehenkel des Manometers in Verbindung gebraehien Mareysehen Tambours gut registrieren. I{ ai s e r, Usher intraabdominalen Druck. 319 Wenn das Messen und Einsgellen nich-t einen gewissen Grad der Genauigkeit hatte, besonders bei zu niedriger Einstellung: erhglt man nut teilweise gesohriebene t(urvcn. Die Trggheit des Wasssrmanometers hat ihren Einfluss auf die Kurvsn. Kurz anhaltends Druckschwankungen werdea zu ktein~ etwas l~inger dauerndo zu gross verzsichnsk Usbsraus schnelle Druckschwankungen (die aber im Rectum nicht vorkommen) wtirden nur ganz geringe Bewegungen veranlassen. Das Hin- und Wisderpendeln des Wassers in der RShre nach einer grossen Bswegung der Fl~issigkeitssXuls zeigt sich (bei unsersm Instrument nach ungef~hr 1/2 Sekunds) dutch sine zweite Bewegung des Hebels. Die beiden Kurvcn auf der linken Seite yon Kurve A geben an~ wis dieselben Druckschwankungen sich verschiedenartig zeigen~ je nachdem sic, a unmittelbar oder b mittels des eJngeschaltstsn Manometers~ registriert wurden. Die Kurve A gibt untsn R d die Rektaldruckschwankung wieder, w/~hrend die beobachtete Person sich in ruhiger vertikaler Th. i: Rd. A. Schwgnkung des Rekta/druckes bci g[cfcm Attach im Stehcn. Nanometcr auf 32 cm eingestollk Stellung befand und angestrsngt a~mete. Die B'ewegung des unteren Thoraxabsehnittes Th. inf. ist mit Mareys Pneumographen aufgsnommen. Die horizontale Linie zeigt den Druck im After~ 32 cm, worauf eingestellt wurds~ an. Die Druckschwankungsn betragen 1 bis hSchs~ens 1~/~ cm. Auf dsr horizontalsn Geraden ist die Zsit in Sskunden angsgeben. Bei I(urve B bei einer zweiten Person~ auch bei lsich~ anges~rengtsm Atmen in aufrechter Haltung aufgenommsn, war das Manometer auf 26 cm~ bei der Aufnahme yon Kurve C bei dsrselben (zweRen) Person in ]iegender Haltung ant 12 cm eingestellt. Die Kurven sind bier rsproduziert, nur um den Grad der 820 t{ ~is or, Ueber intraabdominalen Druek. @enauigkeit des l~'Iessens zu zeigen. Oebrigens wird yon ihnen nnr beil'autlg die Redo sein. Wie wit sehon hervorgehoben, reizt die Sonde die gektalwand sehr wenig~ 16st also nur selt.en merkbare Zusammenziehungen des t{eetums aus. Perioden x~on peris{altiseher Wirksamkeit wurden aueh gelegentlieh bei den untersuohten Personen angetroffen. Sie haben mitunter zum Abbreehen der Beobaehtung Anlass gegeben. Um iiber die GriSsse der dureh I~ontraktion der gektalmuskulatur verursachten Druekerh{Shung wenigstens ein oberfl~teh- Th. in I~d. B. Schwankung des Rcktaldr~tckes beim Stchon. Manometer at.tf 26 cm eingesto~tL Th. inl Rd. C. Sehwankung des Rektaldruckes im Liegea. 3{anomol,er auf 12 cm eingestell'c. liehes Urteil zu erhalten~ habo ich~ dis ieh in einer Periode yon peristaltiseher Wirksamkeit den Draok bestimmte, die Mmungsschwankungen auf ein Minimum reduziert~ indem ieh die untersuehte Parson oberfl~tehlieh atmen liess~ und habe dann jede Viertelminute den Druek notiert. Die yon mir bei dieser Beobaehtung, die wegen eintretenden St{lhldranges abgebroohen werden mtlsste, gefundenen Zahlen bewegten sigh zwischen 32 und 40 em~ wghrend einmal vodibergehend 45 em gemessen wurde. Die dutch gewShnliehes Atmen hervorgebraehten Drueksehwankungen sind im allgemeinen viol geringer; wie wit schon l~Mser, Ueber intra~bdominalen Druek. 321 sagten, betragen sie meis~ x/2 bis hSchstens 11/2 em~ and bei foreiertem Ein- und Ausatmen bei einem Kn~ben war die griiss[e yon mir beobaehtete Sehwankung 3 bis 5 era. Wie die Explorateurballons verzeiehnet aueh unsere gektalsonde die dutch Zusammenziehung tier Rektalmuskulatur verursaehten Drueksehwankungen. Kurve D gibt oben die Bewegung des unteren Thoraxabsehnittes, darunter die Bewegung des Abdomens, beide mit P. Berts Thoraeographen aufgenommen, und drittens die Sehwankung des l~ektaldruekes wider. Die Kurve wurde in horizontaler Stellung aufgenommen, der gemessene Rektaldruek war 28 em (abnorm hoeh), die Zeit ist wieder in Sekunden auf der horizontalen Geraden angegeben. Die kleinen Wellen sind dureh dan Blutstrom veranlasst, die Atmung war sehr oberflgehlieh. Th. i~ abd. R,d. D. Sehwanl~Ttmgdes ]Zektaldruckes durch I{ontraktionen des Rectums. Manometer auf 28 em cinges~ellt. Boi ]iegcnder HaIl:ung. WDohrend bei unseren lefzten Betrachtungen nut yon den Drueksehwankungen verschiedener Art die gede war, muss jetzt daran erinnert werden, dass~ wenn man den absoluten Wert des intraabdominalen Druekes zu kennen wiinseht, eine Korrektion wegen des ~geetaltonus~ d . h . wegen des Spannungszustandes der Rektalwand erforderlich ist. Ein Tell des gemessenen Druekes wird, kann man sagen, dureh diesen Tonus hervorgebraeht. Was nun die Gr6sse dieses Teiles betrifft, so geht aus der Kleinheit~ der gefundenen Zahlen (Kurve C im Liegen 12 em) ohne weiteres hervor, dass er nieht sehr gross sein kann. Andererseits hat Kelling einen Druekuntersehied yon 4 bis 6 em zwisehen der Innen- und Aussenseite einer isolierten Diinndarmsehlinge eines ttundes konsiatier~. 322 Kaiser, Ueber intr~abdominalen Draek. Mit Rticksicht auf diesen Versueh, auf die beobachteten peristalfisehen Sehwankungen und schliesslioh auf den Umstand, dass aueh in jenen Fgllen 7 we ieh die niedrigsten Werte erhielt, immerhin nooh ein gewisser Terms vorhanden gewesen sein kann (woven der gelegentlieh vorkommende Meteorismus Zeugnis gibt), wird es, wie mir schein~, in Ermangdung eines J3esseren gereo,hfferLigt sein~ die dure,h dan Tonus des Rectums bewirkte lokale Druekerl~Shung auf wenigstens 6 em Wasser (wghrend peristaltiseher Rnhe) zu veransehlagen. Die gofundenen Zahlen sind also s/i.mtlieh um diesen Betrag zu rermindern. Es verstehr sieh wohl yon selbse, dass man auoh den Druek in anderen OrganhUhlen naoh meiner Methode bestimmen kann. Dureh die Freundliehkeit des Herrn Dr. van C a m p e n hatte ich Gelegenheit~ mit meinem Manometer und einer gew/Shnliehen Sonde yon L n e i a n i , dessen Kautsehukblase 1 ecm Luft enthalten konnte, den Druek im Magen eines Kranken zu messen, dem man wegen Oesophagusstriktur eine Schr~ig~stel angelegt hatte. Der Magen wurde sorgf'altigst gespiilC und geleert. Als der Explorateur eingefahrt war, befand sieh das mit der Gummiblase versehene Fenster auf dem Niveau einer Stelle der KUrperoberfl~che, die sieh in der Mamillarlinie~ 19 em unterhalb des Untem'andes der 4. Ripp% befindek Bei aufreehter Stetlung war der Druek in der inspiratorischen Pause - - 2,5 era; in der exspiratorischen Pause - - 1 era. Genau derselbe negative Druck zeigte sich bei giederholung des Messens. Der Mann hatte~ obsehon 53 Oa.hre alt, eine straffe Bauchwand, war mager und zeigte einen costo-abdominalen Atmungstypus. Befunde bei Operierten sind abet immer n~r mit u zu verwenden. Es m/Sge jetzt einer der yon mir untersuehten l?~tlle etwas Mher betraehtet werden. Bei einer kr/fftigen jungen Prau (Nr. 12), deren BauehhUhe (Sehambein-Brustwarze) 37 em mass, betrug in ruhiger aufreehCcer Stellung der intrarektale Druek in der expiratorisehen Pause 21:5 em W&SSOP. Die Zahl ist naeh dam soeben Gesagten um wenigstens 6 em zu vermindern. Kaiser, Ueber in~raaSdominalen Druck. 323 Nimm~ man an, dass in der BauchhShle der Druck nach oben bin wie in einer fliissigen Masse abnimmt, so kommt man zu dem Sehluss I dass es auf einer H~he yon 21.5 em unter der Kuppe des Zwerehfells ein Nive~u gab~ we der Druek dem der Atnnosph~re gleich war. In dem fiber diesem Nullniveau liegenden Tell (der Ht~lfte der BauehhShle) herrsehte ein negativer (unteratmosph~triseher' Schema der Ver~eitung des Druckcs bei Patientin No. 12. Druck~ in dem unter dem indifferenten oder Nullniveau tiegenderi Tell ein naeh unten zunehmender positiver. Druck. Es ist rain beaeh/:enswer~, dass man bei nicht gar zu dicken Bauchdeeken durch Palpation die 0bere Zone y o n der unteren unterseheiden und die HShe des indifferenten Niveaus ungefithr bestimmen kann. Es ~ist mir das in manchen PS~llen gelungen~ and 324 g ai s er, Ueber in~raabdominalen Druek. auch andere Aerzte konnten, nachdem sie einmal da,rauf aufmerksam gemacht waren~ dasselbe beobachten. Diesem Falle, wo resistente Bauehdecken und gut entwickelter Brustkasten einen Zustand sichern, den man als normal betrach~en kann, stelle ich Nr. 5 den Fall einer Frau mit sehlaffem Bauch und deutlicher Enteroptose ( L a n d a u s e h e Form) gegeniiber. Schema der Vedoilung des Druekes bei Patien~in No. 5. Bei einer Bauchh6he yon 31 em mass ich hier einen Rektaldruek yon 34 em Wasser, w/ihrcnd yon peristaltiseher Wirksamkeit nichts zu bemerken war. ttier befand sich also, wenn wir wieder 6 cm ffir den Rektaltonus subtrahieren~ das Nullniveau 28 cm oberhalb des Schambeinrandes. Es ist ganz gut miiglich, dass es unmittelbar unter dem t(ai s er, Ueber intra~bdominalenDruek. 325 Zwerehfell lag, das sieh niel~t un!el: der 4~. Rippe, wie die Zeiehnung angibt, sondern Viel niedriger (Tiefstand) befunden haben mag. Von den gezeiehneten K6rperumrissen~ die~ soviel ieh beurteilen kann~ gut auf die beiden ]~'~lle passen, habe ieh den ersten dem Werke yon Stratz ~Die Sch~nheit des weibliehen K6rpers ~, und den zweiten tier Abhandlung yon Mathes ,,Ueber Enteroptose :~ entnommen. Was den Untersehied zwisehen den beobaehteten Personen anbelangt, so ist noeh folgendes hinzuzufiigen. Im ersten Fall Nr. 12 sind die Bauehmuskeln dureh ihren Tonus und dureh'ihre Insertion an dem den Bauehumfang iiberragenden Brustkorb imstande~ w/ihrend die untere Bauehh~lfte eingedriiekt wird~ den oberen, den Rippenbogen am n~ehs~en liegemen Teil der Bauehwand naeh aussen zu ziehen~ was notwendig ist, wenn im Innern ein negativer Druek bestehen soll. Im zweiten Falle is~ wegen des verringerten Umfanges der unteren Thorax/Sffnung nur eine geringe naeh ausw~rts geriehtete Spannung m/Jglieh. Wahrseheinlieh finder der im ausgedehnten Unterbaueh herrsehende hohe Druek~ nieht in dem Muskeltonus einen geniigenden Widerstand~ sondern wird ein grosser Tell des Gewiehtes der Eingeweide dureh den, leider naeh und naeh naehgebenden bindegewebigen Teil der Bauehwand getragen. Die Annahme, dass bei einem bestimmten Grad der Dehnung der Bauehwand die Faseien anfangen in Spannung zu geraten~ ist in Einklang mit den Beobaehtungen yon Quineke~ Weisker und H~rmann~ wobei sieh ergab, dass sine Volumzunahme des Bauehes his zu einer gewissen tt6he ohne Einfluss auf den intraabdominalen Druek war; wurde diese Grenze (z. B. 2 L.) aber iibersehritten~ so trat Erhghung des Abdominaldruekes auf. Dass die untere Thoraxapertur 7 indem sie dem naeh unten geriehteten Zuge der Bauehwand naehgibt, noeh enger wird~ als sie bei der paralytischen Thoraxform sehon sein wiirde~ ist ohne weiteres klar. Auf diese Weise erkl/~re ieh mir die geringe Tiefe des Oberbauehes bei enteroptotisehen Frauen. Wit wollen jetzt des ngheren die Frage ins Auge fassen, 1. ob alas spezifisehe Gewicht des gesamten Bauehinhaltes dem des Wassers ungef~hr gleieh sei, 2. ob die Baueheingeweide dutch BXnder an den geringen zur Druekiibertragung erforderliehen Bewegungen verhindert werden und 3. ob es schwer deformierbare Organe gebe~ 326 ~aiser, Uebor inSraabdominalen Druck. die einen so betriichtliehen [['eil der Bauehhijhle eimmhmen, dass sic die regelmgssige Verteilung des Druekes verhindern k6nnen. Die wenig volumin6sen festeren K6rper in der Bauehh0hle, Uterus, Skybala, k6nnen nur ganz lokale Vergnderungen hervorrufen. Als solehe sind wahrseheinlieh die Hgmorrhoiden zti betraehten~ deren Entsteben u. a, mit dem l~ngeren Verweilen yon festeren Kotmassen im Rectum im Zusammenhang zu stehen seheint. 1. Was das spezifisehe Gewlcht betrifft, gebe ieh bier die yon Vierordt gefundenen Za.hlen (naeh Z w a a r d e m a k e r ) . Spez. Gewich~ der Leber . . . . . , Milz . . . . . Driisen . ii ,, dos P a n k r e a s . . . . . , Fe~tes . . . . . , Uterus . . . . . . . . . . Muskelgewebes ,, . . . . elas~. G e w e b e s ,. ,, ,, B l u t e s . . ~ 1,072 1,058 = 1,0114 = 1,046 -= 0,971 1,052 = 1,041 = 1,122 ~- 1,058 Man sieht, dass diese versehiedenen GrSssen sieh innerhalb sehr enger Grenzen halten. Der Diinndarm hat einen diinnbreiigen Inha.lt~ erst im Dickdarm f~ngt die Eindiekung an. Das spezifisehe Gewicht der Fgees ist, je nach deren Wassergehalt, etwas grSsser oder etwas kleiner als alas des Wassers. Wenn man nun den meist geringen Gasgehalt tier Gedgrme in Reehnung zieht, wird man L u d w i g beistimmen: wenn er alas spezifisehe Gewieht des gesamten Bauehinhaltes gleich eins setzt. Ausserdem weist das Verhalten der freien Flt~ssigkeitsergtisse in der Bauehh6hle im allgemeinen nicht auf eine grosse Differenz in dem spezifisehen GewiehC des Bauchinhaltes und der Fltissigkeit hin. Nur wenn die Ged~rme dureh den Meteorismus leiehter werden~ nehmen sie immer die hSchste Stel[e des Bauehes ein, sonst ist man oft erstaunt, wie die Fliissigkeit sieh zwisehen den Organen des Bguches versteckt, und sieht man Abseesse sieh im Bauehraume ausbreiten~ a ls w~tren sie nicht dem Einflusse der Schwerkraft unterworfen. Die bei periappendikularem Abscess vorkommenden sekund~tren Abscesse unter dem Zwerehfell entstehen teiiweise dutch Fortleitung lungs tier Biutbahn~ teilweise dutch Fortkrieehen dureh das Nesocolon und den retroperitonealen Raum: aber es gibt aueh sieher konstatier~e F~tlle, we tier Eiter in tier Palte zwisehen Colo~ und Peritoneum parietal% der Wirkung der Sehwerkraft entgegen~ im freien Bauchraum his zur Leber aufstieg. Kaiser, Ueber ingraabdominalen Druck. 327 2. ]Die friiher vielfach wMtverbreitete AnsicM~ dass die Organe des Bauches an dem Peritoneum aufgehangt seien wie die Xleider an einem Kleiderh~nger~ hat man Mlm~hlieh mebr verlassen. Die grosse Dehnbarkeit des Peritoneums spricht schon hiergegen, und ebensowenig wird die elasCische Traktion der Gef~ssc die Eingeweide auf die Dauer auf ihrem Platz zurfiekhalten k~nnen. Schatz und S c h w e r d t sind der 5Ieinung, dass ein Teil (i/s) des Gewiehtes der Organe yon den B~ndern, ein Teil (Vs) veto intraabdominalen Drucke getragen wird. Quincke und Meltzing sagen, die B~tnder bestimmen den Platz der Eingeweide, aber der Bauchdruck tr~gt die Last derselben. Naeh Kelling ruhen in aufrechter Stellung die Organe auf den darunterliegenden wie auf einem Kissen. Des Auftreten yon H~ngebauch sei nur abhgngig yon der Sehlaffheit der Bauchwgnde, meint ~lathes. Mann~ der die Ansichten yon Winkler und L a n d a u teilt, sagt kurz und gut, wfirden alle Bgnder durehgesehnitten, so blieben die Eingeweide noch auf ihrem Platze, nut ein stark abweichendes spezifisches Gewieht wiirde zu einer Versehiebung AMass geben k6nnen. S i m m o n d s konkludiert aus seinen Beobachtungen: Ist der Druck yon unten her unzulitnglich, so sinkt alles, was sinken kann, und van der Hoeven, aus dessen Arbeit ich diese Worte S i m m o n d s zitier% sagt etwas welter, Coloptose, Enteroptose, Gastroptos% Nephroptose, ttepatoptose usw. entstehen alle infolge ErscMaffung des Tonus der Bauchwgnde. Wenckebach hebt den Einfluss des pathologisch verminderten Bauchinhal~es hervor~ eine bisher nieht gen~igend gewfirdigte, gelegentlich mitwirkende Ursache. In den znr Druekiibertragung n~tigen ganz geringen Bewegungen werden also die Bauchorgane wohl nieht verhindert werden, well ftir die Atmung und far jede Aenderung der Position leichte Beweglichkeit und Deformierbarkeit des Bauehinhaltes absolute Bedingung ist. Dass, wenn die Verschiebungen gr/Sssere pathologische Dimensionen annehmen~ die Bgnder wohl, sei es aueh nut, bis sie sieh dem neuen Zustand angepasst haben, Einfluss iiben, ist selbstverstgndlieh. Schon seit Glenard haben alle Aerzte, obgleich sic vielleieht iiber die Entstehungsweise der Enteroptose versehiedene Ansichten hegten~ dutch straffe~ elastisehe Umsehntirung des Unterleibes 328 tiais or, Ueber in~raabdominatenDruek. danach gestrebt, das verlorene normaie Verhalten im Bauche wiederherzustellen. glastische Gurte (Glenard), I-Ieftptlaster-Verb/tnde (Rose, Clemm~ Weismann)~ auf den Bauch gedrtiekte Metallplatten (Watts~ Rotgans) und mi~ Luftkissen versehene Aluminiumplatten (Wenekebaeh) sind mit ]~rfolg angewendet worden, und, wenn die Patienten dureh die vermehrte Eindriiekung des LTnterbauehes, naeh Anlegen der Bauchbinde sieh auffallend besser ftihlen, so liegt der Gedanke sehr nahe, dass sie vorher an Ermangelung dieser Eindriickung erkrankt waren. Sohade ist es nar, dass keine dieser Binden auch in der oberen Banehh~,Ifte den verlorenen Wandtonus zu ersetzen vermag. 3. a) Bei der Palpation des normalen Bauehes bekommt man den Eindruck, dass, abgesehen yon den Nieren, die normalerweise nieht znm Bauchinhalt geh6ren, und dem Uterus, yon dem schon die Redo war~ die tibrigen Bauehorgane sieh nut sohwer oder nieht gesondert abtasten lassen. Der in dem schlaffen Gedgrme enthal{ene diinnbreiige lnhal{, die bewegliehen Darmwgnde sel bst mi~ dem von reiehlichem Fett geffillten Netz und Gekr/Sse f~illen alle offenen Stellen im Bauchraume. Die Blase ist nut wtthrend einer gontraktion der Blasenmuskulatur einigermassen prall, sonst lieg~ sie schlaff auf dem Boden der Bauehh/Shle darnieder. Dass der Nagen den zu dieken Inhalt dureh eigene Sekretion zu einer diinnbreiigen Nasse umformt~ wird uns jederzeit beim Erbreehen dargetan. Nut die m/tehtige Leber k~Snnte die Gieichm/issigkei~ g/inzlieh zerst/Sren. Sehen wir daher zu, wie es w~ihrend des Lebens um die Deformierbarkeit dieses Organes steht. Dass der Rand einer normalen Leber so sehwer durehzutasten ist, sprieht sehon nicht fiir starre gonsistenz. C. H~sse ist iiberraseht, dass die normale Leber als ein steifes Organ hingestellt wird, wghrend doeh die weehselnde Nutftille und die Sekretionsznstgnde dafiir spreehen, dass das Leberparenehym in sieh beweglieh ist~ und dass somit das ganze Organ Formver~inderungen erleiden kann. Wtihrend der Respiration bleibt die Leber dem sieh abflaehenden Zwerehfell angesehmiegt, wghrend die untere Thoraxapertnr sieh erweitert. Hieraus folgt, dass die Leber bei der Inspiration nieht nur tiefer sinkt, sondern aueh brei~er wird~ nnd dass die Kriimmung der oberon Leberflgehe ab- 32 9 K ai s e r, Ueber in~raabdomin~len Druek. nimmt. Ueberhaupt bekommt maa~ wenn man die aus dem K6rper entfernten Organe betraehtet~ yon ihrer Steifigkeit eine ungenaue Vorstellung. Die Abkiihlung, die Aenderungen des Blur- und Lymphgehaltes, die eventuellen @erinnungen kgnnen n~.mlich naeh dem Tode einen ganz anderen Zustand hervorrufen; sehon ~usserst geringe Untersehiede im Wassergehalt oder in der Temperatur bringen bei den aufgequollenen Kolloiden sehr grosse Ver~inderungen tier physisehen Besehaffenheit zustande. Bekanntlieh nimmt die Xonsistenz des Pet~gewebes bei geringer Temperaturerniedrigsng schnell zu. Die RSntgenaufnahmen S e hii rm a y er s zeigen aufs deutliehste, dass die Leber sieh ganz wie eine halbfliissige Masse dem oberen Bauchraume bei dessen Gestaltsver~inderungen genau Ngt. Mit Hinsieht auf den enormen Blutgehalt (2973 pCt. der ganzen Blutmenge) hat Wenckebaeh ganz recht, wenn er die Leber mit einem Sehwamme vergleicht. Alles zusammengenommen seheint mir in der Tat die Annahme gereehtfertigt zu sein, der Bauehinha.lt ist welch und geschmeidig genug, um genau so wie eine halbfliissige Masse den Druek fortpflanzen zu kSnnen. Obschon, wie R. Meyer mit Reeht bemerkt, die eigene Gestalt der verschiedenen Organe eine gewisse Ungleiehm~tssigkeit in der Yerteilung des Druekes verursachen muss, wird aber die den meisten Bauehorganen eigene leiehte Deformierbarkeit, die, wie wir sahen, w/thrend des Lebens wahrseheinlieh sehr viel grSsser ist, als sie, aus dem I~Srper entfernt~ vermuten I~;sst, zur Folge haben~ dass diese Ungleichmgssigkeit zwisehen sehr engen Grenzen beschrXnkt bleibt. Uns interessieren iibrigens aueh in erster Linie nieht die kleinen Abweichungen, sondern vielmehr die allgemeinen Gesetzd fiir die Verteilung des in~raabdominalen Druekes. b) In dieser ttinsieht sei bier in Erinnerung gebrachg, dass 7 wenn bei einem (niederen oder hSheren) Organismus die zur Ern/ihrung und zur Regelung des Stoffweehsels nStigen SaftstrSmungen iiber grSssere Abs~tnde stattfinden miissen, diese StrSmungen in pr~iformier~en Kan/ilen ihren Weg finden. D a n u n die WXnde dieser Kan~;le nicht start sind, so ist die Beschaffenheit des umgebenden Mediums yon grosset Bedeutung. Regelm/issige DurehstrSmung ist nur mSglieh, wenn nieht Kaliber und Gestalt der GeNsse under dem Einflusse tier ge/inderten Wirkung der Schwerkraft bei jedem Stellungswechsel des KSrpers sieh ver~;ndern. Obgleich bisher die Einfliisse tier Temperatur, die osmotischen Eigensehaften usw. vorArchiv far Gynfikologie. Bd. 96. I:I. 2, ~ 330 t{aiser, Ueber incraa.bddminalenDruck. wiegend in Be~craeht gezogen worden sind. sind auch namentlieh fiir die Zirkulatidn die mit dem spez. @ewicht zusammenMngenden DruekverhXltnisse sehr wich~ig. (Nan braueht hierbei noch nieht einmal a.n die ungeheure Lt~nge des aufrecht gehenden Iguanodonten zu denken, deren Skelette jtings~ im Br~issler geologisehen Museum aufgestellt worden sind). Die am einfaehsten gebauten: niederen Organismen. bei welehen die inneren S~fte nnd Str6mungen wemger als bei den h~heren Formen yon /tusseren Einfltissen unabh/tngig sind. k6nnen nut in einer Umgebung leben, die keinen tiefgreifenden Einfluss auf die inneren Vorg~tnge hat. Wghrend bei den Wasserbewohnern keine besonderen Einriehtungen das die Zirkulation f6rdernde hydrostatisehe Gleiehgewieht aufreeht zu erhalten brauehen, finder man solehe bei den ausserhalb des Wassers lebenden Tieren und zwar in umso gri3sserer Entwickelung und Vollkommenheit, je h6her die Bedeutung einer regelmXssigen Zirkulation ist. Als Beispiel der in Rede stehenden Wirkungen erw~thne ieh hier die Untersnehungen L e o n Hills, die auf dem IV. International.on Physiologen-Kongress in Cambridge mitgeteilt wurden: ~Ein Aal oder eine Natter wird ausgestreekt auf ein Brett geheftet und alas Herz freigelegt. Stellt man das Tier senkreebt~ den Kopf nach oben, so wird das Herz naeh wenigen Schl~gen blutleer. Streieht man dann den Ki~rper des Tieres vom Schwanzende naeh oben zu~ so wird das Herz wieder volt. Stellt man das Xopfende naeh unten, so schwillt das Herz prall an, bis es yon dem beim Aal besonders festen Herzbeutel zusammengehalten wird. Die Venen tier aufreeht gehaltenen Sehlange enthalten nur in den unteren zwei Dritteilen des K6rpers Blur. SCellt man denselben Versueh an, indem man den K/3rper des Tieres in gasser taueht, so wird alas Herz nieht blutleer. Der Xussere Druek des Wassers h~;lt einem Teile der Blurs/rule alas Gleiehgewieht. a Aueh bei einem Kaninehen, wie bei der Sehlange zeigte sieh in der aufreehten Stellung der Schutz der Zirkulation unzuliinglieh. Beim Aal, der in seinem Elemente verweilt, bleibt~ welehe Stellung er aueh einnehmen m6ge~ die Differenz des Druekes ausserhalb und innerhalb seiner C~eft~sse unbeeinflusst. Das n~mliehe, am meisten zweckdienliehe Nittel zur Wahrung dieses Gleiehgewiehts trifft man aueh beim Cerebrospinalsystem an. Die in der sehlaffen Dura ma~er eingesehlossenes die gef/tssfiihrende Pia mater samt der weiehen Oerebrospinalmasse. urn- K a i s e r , Ueber intraabdominalen Dr~lck. 331 spiilend6 Plfissigkeit verleiht dem GefXssystem des Gehirnes und des Riickenmarkes ein yon der Stellung des I(Srpers gSmz]Seh~unabh~ngiges Gleiehgewicht. Bei den Extremit/iten sind die starken elastischen Wandungen der Schlagadern~ die kontraktilen W~nde der feineren Gef//ss% die Klappen der Venen zum selben Zweck dienlieh; ausserdem aber spielen die, yon straffen Pascien eingehiill~en~ elastiseh-weiehen Muskelmassen~ welehe die Sd~lagadern fast iiberall umgeben 7 wohl mi~ eine Rolle. Auch im Bauche ist eine gleichm~ssige Fortbewegung des Blutes yon hSchster Bedeutung, nit]it allein der Wiehtigkei~ der in ihm enthattenen Organe wegen~ s0ndern auch, weil ein so grosses Blutquantum die Bauchorgane passiert~ dass jede StSrung eine tiefgehende Wirkung auf die ganze Zirkulation ausiibt. Jed0ch fehlen der Vena portarum, ebenso wie der ~Vena Cava und den Venae iliaeae die vielen Klappen~ und die grosse wichtige Vena cava gehsrt ihrer diinnen, muskulSsen Wand wegen zu den ,Veines reeeptives!' yon genaut. Ausserdem hat das Kapillargef~;sssystem der Leber kein kontraktiles Endothel. Alies weist darauf hin~ dass der die Gef/~sse umgebende Bauchinhalt auf analoge Weise, wie eine Fltissigkeit es tun wiirde~ befSrdernd auf die regelm/issige Blutverteilung wirkt und also den n/imliehen Gesetzen unterworfen sein muss wie der Inhalt der GefXsse. W/ire die Druekverteilung nicht die yon mir angenommene~ so miissten unter Umst/~nden StSrungen in der Zirkulation entr stehen. Die Vena cava~ in deren Innern der Blutdruck (und zwar unabh~ngig yon seiner absoluten Grssse) naeh unten zu regelm~issig zunehmen muss, miisste bei unregelmS;ssiger Verteilung des/~usseren Druekes mit ihrer diinnen, nachgiebigen Wand stellenweise ausgedehnt oder zusammengedriickt werden. Wenn man sich hinlegte, so wiirden zwischen den Druek' /~nderungen ausserhalb und innerhalb der Gefs Unterschiede auftreten 7 welche die Zirkulation beeinflussen wiirden. Die Ergebnisse der Messungen und die daran geknttpften Erw~gungen ftihrten ~2ns dazu, in vielen P/~llen die NSgliehkeit des BestehenS eines unteratmosphs Druckes in der oberen Bauchh/i.lfte anzunehmen. Jetzt will ich~ sei es der verwickelten Verh/tltnisse wegen, auch mit dem nStigen Vorbehalt noch einige weitere Betrachtungen 22 * 339 I~2gi s er, Ueber intraabdominalen Drunk. vorbringen~ die auf das Vorhandensein eines solchen Druekes hinwsissn. a) Es wird yon allen ikutoren hervorgehoben, dass man bsi an Entsroptose Leidenden die Aort% Ms ob sin direkt unter dsr Bauchdecke gelegsn w~re: pulsisrsn fithlen kann. l~osengar~ and viele hash ibm haben sogar bsi stehender Haltung siehtba,re Pulsa.tionen tier Aorta beobaehtet~ and diess zeigtsn sich auch gslegentlieh bsi meiner Patientin Nr. 5. Dass die beobaehtsten periodischen Erhsbungen yon den Pulsationen nines in gsringer Tiefe liegendsn Gsf~sses herrt~hren, untsrliegC keinem Zweifel. Yon Wichtigksit ist aSer: wie mir scheint~ die Bemerkung, dass derart~ige, auf eine kleine Fl/iehe besehrgnkts Erhebungen bei einer schlaffen Wand eher als bei einsr stark gespanntsn zu erwarten sind. Bsst~nde im Innern sin betr~chtlicher positiver oder negativer Drunk, so wiirde wohl die demselbsn entsprechsnde Wand dis Sichtbarkeit des Pulses verhindern. Nur in der ntiehs~sn Umgebung des Nuttniveaus werden also dergleiehen Pulsationen sish dsutlich zsigen k~innen; wsnn noch dazu dissss Nivsau zusammentrifft mit sinsr Stelle, wo dutch die geringe Tiefe des Bauchss das Gef~ss sieh nahezu d irskt unter den Bauehdeeken befindet. b) Nngelen hat mittsls eines dicken Stiftss die Bauchdeeksn singedriiekt und versusht, auf disse Weise den intraabdominalen Drunk kennen zu lernen~ die erforderliehe umfangreiche Berechnung hat ihn yon dsr weiteren Durchf/ihrung dieser Methods abgehalten. Schon fr~her ha~ts aneh ich auf nine /ihnliche Weiss vsrsucht, ein Urteil fiber die UnLersehiede in der Spannung der Bauehdecksn auf versehiedenen HOhen zu gewinnen. Mein Apparat bestand aus einem 10 sm langen und 9. cm dieken Stilt, auf welchem nine Millimetsrskala angebraeht war. Derselbe hatte sin Federdynamometer als Handgriff und konnts dutch ein durehl/3shertss Brettehen hindurehgesehobsn werden. Presst man das Brettehen kr/fftig gsgen die Bauehdeeke, nnd driiskt man den Stilt mittels des Dynamometers darch die Oeffnung in dem Brettehsn naeh innsn~ wobsi die Bauehdecken eingesttilpt werden, so kann man jeden Augenblick den Grad tier Einstiilpung and die ausgs~ibts Kraft ablesen. Ieh glanbe nieht: dass man auf diese Weise den intraabdomina[sn Drunk wird bestimmen kGnnen; wohl abet gswinnt man nine 333 K ai s e r , Ueber i n t r a a b d o m i n a l e n Druck. Einsieht in den jeweiligen Spannungsgrad der Baachwamd an versehiedenen Stellen. Wenn man z. B. etwas medianwgrts yon der Mamillarlinie auf einer Hiihe yon 5 i'esp. 10~ 15 oder 20 em oberhalb des Os pubis~ den Stilt nacheinander t/2 , 1, 11h, 2~ 21/2 usw. em naoh innen drOokt und jedesmal die dazu benOtigte Kraft notiert~ so sieht man, dass zwar /iberall bei immer tieferem Eindriieken die Kraft sehnell zunimmt, dass aber diese Zunahme nioht auf jeder HOhe gleieh sehnell gesohieht. In naehfolgender Zeichnung gibt die L/inge der vertihlen Linien jedesmal die Gr/Ssse der angewandten Kraft, die Stelle, wo die Linien eingezeiohnet sind, den Ort und die Tiefe der entstandenen Einstfilpungen an. Gesunder Enteroptotis t h e F r a u ~{ann 8 cm 10 CI]I 15 cni 14 em 20 cm 20 cm Oberhalb Os p u b i s Oberhalb Os pubis 1 " 11/2 " 2 " 21/2 " 3 Tiefo d. Eindrueks I " 11/3 ' 2 " 21/2 . 3 " 31/2 Dass, wie ein Blick ~uf diese Figuren uns lehrt: bei tieferem Eindr0cken die erforderliche Kraft rasch w/ichst~ kann uns nicht wundern, da auch die Oberfl~che der hervorgebrachten kegelf6rmigen gertiefung sohnell zunimmt. Das 5[erkwtirdige aber ist~ dass bei manehen Personen ein 334 X &is er, Ueber in~craabdominalen Druck. grosser Unterschied in dem Grade der Zunahme sich bemerkbar maeht, wenn des Eindrtioken bald in einem tioferi bald in einem hSher gelegenen Punkte der Ban&wand statt~ndet. W~hrend in der linken H~lfte unserer F~gur die vertikalen Linien naeh unten hin rasch an L~nge zunehmen, ist yon einer solchen Zunahme in der rechten H~tlfte h u m etwas zu bemerken. VielMcht weist die geringere Eindrtickbarkeit der Bauohwand des gosunden Mannes in ihrem hSheren Teil darauf hin, dass hier die Wand schon dureh die Spannung der Muskeln naoh aussen gezogen wird, und sie httngt also mit dem negativen Druek in der oberen BauchhShle znsammen. c) Zum Schluss kann auf die Kurven B und C dor t~ektaldru&sehwankungen hingewiesen werden~ die w~hrend des Stehens und des Liegens dureh die Atmung hervorgerufen wurden. Den nXmliehen Unterschied, den unsere Kurven aufweisen, je naehdem sie wXhrend des Stehens oder des Liegens aufgenommen sind, hat aueh A. t t o g g e und zwar regelm'assig wahrnehmen kSnnen. Obgleieh hier darauf nieht weiter eingegangen werden sell, will ieh doeh hervorheben~ dass die Kurven darauf hindeuten~ dass der Atemmeehanismus meistens sieh Xndert, wenn man yon tier vertikaten Stellung in die horizontale iibergeht und umgekehrt. Diese Aenderung beim Siehhinlegen hat ihre Ursaehe wahrseheinlieh in der aufgehobenen Spannung der Bauehdeeke und in der ver~nderten Einwirkung der Sehwerkraft, welehe beide zusammenwirken, nm bei dem erwghnten Stellungsweehsel eine grosse Aenderung (Zunahme) in dem unter dem Diaphragma herrsehenden Druek hervorzurufen. Der Einfluss, den die ElastieitXt der Bauchwand auf den Druek im Abdomen haben kann, l~ss~ sieh dureh ein einfaehes Experiment veransehauliehen. Zwei Gummisehl~uehe mit ungleich starker Wand sind auf 5 versehiedenen HShenniveaus mit je einem Manometer verbunden, die ebenso wie diese ganz mit einer gef~rbten PlCissigkeit gef~ili~ sind. In ihren oberen Teilen befindet sieh je ein Gummiballon yon gleieher Gr6sse. Wird nun in die Ballons in demsdben Masse ]Luft hineingepress~, so fliesst ein Tall der F1/issigkeit aus den oberen Oeffnungen der Manometer hinaus. Jeder Sehlauch t~tsst hierbei eine gleiehe Quantit~t~ Fliissigkeit heraustreten. Oeffnet man, naehdem Ruhe eingetre~en ist, d e n Hahn, so dass die eingepressie Luft wieder entweiehen kann, so Kaiser, Ueber intraabdomina~enDruek. 335 fallen sofort die Niveaus in den 3Ianometern. Bei dem sehl~fferen Sehlauche-ist das Resultat aber ein a~nderes gls bei dem mit st~rkerer Wand. W~;hrend bei diesem die Nanometer zeigen~ dass i m oberen Niveau ein stark negativer, im zweiten ein sehwaeh negativer und im dritten ein sehr kleiner positiver Druck besteht, siehtman beiml s&laffen Sohlattehe nur im obersten Niveau einen sehwach negativen Drucl~, wahrend schon im zweRen der Druek positi~- ist. Der Sehlaueh s=elbsiL zeig~ aber im oberen Tell eine Einschntirung und im unteren Teil eine Erweiterung, wXhrend d e r 9starre~Schlaneh eylindriseh geblieben ist. Diese beiden Sehl/iuehe, deren Untersehied in Wandst~rke nicht einmal sehr gross ist~ geben genau das Verbalten wieder~ das wir be[ den beiden gesonderg hervorgehobenen PXllen Nr. 5 und Nr. 12 angetroffen haben. Der Druek ist im Innern tier Sehl~tuche auf /ihnliehe geise /'erteiR, wie ieh es far diese beiden Personen angen0mmen habe~ die Wirkung dieser inneren Druckverh~;ltnisse auf den sehlaffen Schlauch isg den bei Fall 5 beobachtetenVer~tnderungen der GestaIt des Bauehes vollkommen analog. Das yon mir so genannte Nullniveau wird durch den Stand der Manometer angegeben und befindet sieh genau auf einer H6he, die sieh naeh unseren Messungen aueh gewiss ergeben wiirde. Erinnert man sieh~ dgss der N. obliq, extern, ebenso wie der ~{. rect. abd. ausserhalb des Grustkorbes ~/n den gippen und der 3I. obliq, intern, an der Darmbeinschaufel sieh inserieren, so ist 336 Kaiser, Ueber intraabdominalen Druek. es ersichtlich, dass bei Anspannung diese schr'gg und longitudinal verlaufenden Muskeln und ihre Fascien die Bauchwand nach aussen ziehen, wiihrend nur der an die Innenflgche der Rippen und am Lig. lumbocostale angeheftete g. transv, abd. mit seinen trans~Tersalen Fasern den Bauch ringsum einzuschn/iren verma.g, so k6nnen bei jugendlicher straffer Bauehwand innerhalb des Bauches grosse Differenzen rnit dem atmosph/irischen Druek vorkommen, ohne dass eine deutliche Deformierung des Bauches darauf hinweist. Genau so wie S c h a t z den bei Sehwangeren nieht erh6h~en intraabdominalen Druck dutch die Hervorw61bung tier Mitre der Bauchwand dureh den steifen Uterus erklart~ wobei zur Fiillung der beiderseits des O~erus geIegenen l~tume tier ttbrige Baucbinhalt n~r gerade geniigC, so erkli/re ich mir den subatmosphXrischen Druck in der oberen Bauchh•tfte dadureh, dass die Rippenbog'en die straffe Bauch~and am Einsinken verhindern. Auch ohne krankhaf~en Tonusverlust tritt eine Deformierung des Bauehes ein, sobald die ers~en yon Bauchwanderschlaffung begleite~en Altersver/tnderungen sich zeigen. Auf den beiden bier reproduzierton Pllotographien sieh~ man, wie in dot oberen H~ilfte des Bauches des 47jXhrigen Mannes Nr. 6 und in noch h6herem }Iasse in dem Bauche des 59jghrigen Nannes Nr. 23 eine Einsenkung sichtbar ist, die schwerlich Nuskelkontrak~ionen zugesehrieben werden kann. Als Durchsehnitt arts mehreren genauen Messungen bereehnete ich den Stand des Nullniveaus auf der an der Seite dot Abbildungen verzeichneten HOhe. Die konvex ges~alfbete Porto des Un~erbauches im h6heren Alter hag nut zum Teil ihre Ursache in Fettansammlung nnter der tIaut, zum Teil aber auch in dem dort herrschende~ ttberatmosphgrischen Druck und dem verminderCen Tonus der Bauchwand, wodureh eine Vorw61bung entsfcehen kann. Auf diese Weise enCs~anden bei krankhaftem Tonusverlus~ die yon van der H o e v e n in ,~die Asthenie und die Lageanomalien usw." publizierten Gestal~en. In Zusammenhang gebracht mit unseren Betraehtungen umschreiben wir die Entstehungsweise dieser Formver~;nderungen des Bauches am genauesten, wenn wit sagen: Der atmosph~iriseho Druck~ tier gr6sser isg als der im Innern tier oberen Bauehh~lfte, aber geringer als tier in der unCeren Bauchh/~lfte herrschende Druck, rnacht im oberen Teil eine naeh ausw/~r~s, im unteren Tefl eine hash einw~rts gerichtete Spannung der Bauch- K ~ i s e r , Uther ]ntr~a.bdon~in~len Druck. Nr. 6. int. 23. 337 338 t(aiser, Ueber in~r~abdominalenDruek. deeken unumg'21nglieh. W/~hrend bei normaler Bauehwand der Muskeltonus zu dieser Spannung geniigt, kann dieselbe, sobald dieser Tonus etwas abnimm~ ohne Dehnung; also Einsenkung in der oberen, Vorw61bung in der unteren H~;lfte nieht vorkommen. g o I f li ]h Nachdem wit im vorigen Absehnitt den Einfluss der Elastizi~/it der W/tnde auf die Druekverteilung im Inneren eines Beh~lters betraehtet haben, soll jetzt noeh kurz der Einfluss der Druekverh~ltnisse in einer fliissigen Masse auf die yon ihr umgebenen Gef/~sse besproehen werden. Der Verwiekeltheig des Problems halber m6ge ein einfaehes Beispiel zur Erkl/irung dienen. K a i s e r , Ueber intraabdominalen Druck. 339 hn zylindrischen Beh~flter e f g h ist eine elastisehe g6hre derart befestigt, dass sic eine gestreckte Oese a 1 c k b % bildet. In der Mitre derselben beflndet sieh das seitIiche gohr d, dutch welches man, ebenso wie dureh a I u n d a2, Wasser ein- und ausstrSmen lassen kann. Fliesst bei horizontaler Lage des Apparates Wasser in die 0effnungen al und a2 und aus d, so strSmt genau ebensoviel dureh die Schlinge % b d wie dureh a 1 e d, denn beide sind gleieh, lang und yon derselben Weite. Stellt man aber das Gefs vertikal 7 so s sieh dieses. Ist b naeh unten gekehrt~ dann hat die Wand der RShre daselbst einem Druek zu widerstehen, der nahe um die ttshe des Gef~sses e f g h gr6sser ist, Ms der Druek in c. In der Schlinge b wird die R/3hre also mehr ausgedehnt als in der Schlinge e~ und b lgsst infolgedessen mehr Wasser hindureh. Fiillt man nun das Xussere Gef~tss e f g h mit Wasser an, so wird aueh in ihm der Druek in demselben Masse von oben nach unten hin zunehmen wie in der R/Shre. Unter diesen Umsk~inden haben die beiden Zweige wieder gleiche Weite und werden somit yon gleiehen Wassermengen durehstrSmt. Man denke sich nunmehr den Versueh etwas anders eingeriehtet, man schliesse nErnlich d und lasse das Wasser in a2 hinein-und aus al herausstrSmen. Dabei bleibe das gussere GefXss mit Wasser gefiilk, so dass es, was die in Betraeht kommenden Erseheinungen betrifft, einerlei ist, welehe Lage wit dem Apparate geben. Wir wollen eine vertikale Stellung voraussetzen, aber, wie wir offenbar tun d~irfen, yon der Wirkung der Schwerkraft ggnzlieh absehen. Die Gleichheit des Lumens tier beiden Zweige, die im vorigen Fall bei gefiilltem /iusseren Gef~ss bestand, wird jetzt dadurch gest6rt, dass das Wasser, nachdem es % b k durehs~r6mt hat: noeh den Reibungswiderstand in k e a, tiberwinden muss. Der Druck in der t~Shre wird in b etwas gr6sser Ms in e sein~ sagen wir p -I- a, wenn er in e die Gr~Ssse p hat. Far unsere Anwendung kommt es nun speziell darauf an, dass es yon dem Druck q im Xusseren Wasser abh/~ngt, wie wei~ die g6hre bei b und e sein wird. Zun/iehst sei der Druek q kleiner als p. ErhShen wit ihn dann, wozu der bewegliehe Stempel e f dienen kann, so dass er schliesslieh gleieh p wird: so kann die Wand bei c ganz schlaff 340 Kai s er, Usher in~raabd.ominalenDruck. werden~ wghrend sie bei b infoJge der Druekdifferenz a noch gespannt und ausgedehnt :ist. Dagegen kOnnen wir dureh eine Verminderung des Druckes q bewirken~ dass auch in c eine Ausdehnung der Wand besteht, und es kann vorkomrnen, dass diese bei weiterer VergrSsserung der Druckdifferenz nut noch wenig zunimmt. Dann wird da.s Lt~men be[ b~ obgleich bier eine urn a gr6ssere Druc~kdifferenz besteht, nut wenig wei~er sein als bei c. Unter geeJgneten Umsttinden wird also Steigerung des Druckes im .BeMker den Unterschied in der Weite der beiden Z~veige ver. gr6ssern~ Erniedrigung des tiusseren Druckes aber diesen Untersehied verringern. W~ihrend unsere gessungen gezeigt haben~ dass bei einer asthenischen K/~rperges~al~ im allgemeinen das Nullniveau h6her liegt und tier intraabdominale Drnck h6her ist als bei normalem kr/iftigen K6rperbau~ k6nnen wit nun an der Hand unseres Versuehes eine Einsiehg gewinnen in den Einfluss, den dieser 1)ruckunterschied auf die Verteilung des Blutes in den Bauchorganen hubert muss. Wie das Wasser der beiden @ef~ossschlingen a2 c k trod k b a hat das Blur der Vena por~ae nacheinander zwei Kapillarsysteme zu durehs~r~men. Druckerh~husg im Bauchc wird zum gesul.tat haben, dass die Gefi~sse des ers~eren Gebie~es relagiv mehr ansgedehnt, die des zweiten mehr verengt werden. Bei Verringerung des Druckes tritg dieser im Nachteil des zweiten Gefiissgebietes wirksame Einfluss mehr und mehr zuriick. Bei einer schlaffen Haltung werden also einigermassen analoge Cngleichm~tssigkeiten in der Zirkulation, wie bei dem seinem Elemente entzogenen Aal, auftreten~ die bei strummer Haltung mehr oder weniger verschwinden. Betrachten wir nun Fall No. 12. ttier befinde~ sich ein grosset Tell der Baucheisgeweide in einem Gebiete~ wo negativer Druck herrseht. Auch ohne weitere Ueberlegung wiirdo man schon sagen k6nnen, dass die Leber, das blutreiche, schwammige Organ, dessert KapillargeNsse des kontraktilen Endotbels entbehren, sich in einem Zustande dec Ausdehnung nlit erweiterten KapitlaNefttssen befinden muss. Das dutch die Vena portac aus dem Digestionstractus in die Leber hineinstrSmende Nut tinder nur einen relativ geringen Widerstand~ bevor es in die weitklaffenden Venue subhepaticae hineinkommt, aus denen es dutch die Thoraxaspiration weiterbef6rdert~ werden kann. Kaiser, Ueber irltraabdominalenDruek. 341 In dem Falle der Enteroptotica No. 5 dagegen sind auch bei ganz freier Bewegliehkeit der Bauehorgane die Gef'~sse im oberen Bauehraume verengt. Kommt noeh dazu, dass die (doeh immer etwas Widerstand bietenden) Organe der oberen Banehh~ilfte und vor allem die Leber dureh den atmosph~risehen Druek zusammengepresst werden und demselben nieht ausweiehen k6nnen, so setzen die ~-erengten Leberkapilla.rgef/tsse dem B].ute der Vena portae einen viel grbsseren Widerstand entgegen: Verengerung der Gef~sse des zweiten Bezirkes verursaeht Druekerh/bhung mit BluttiberNllung in der unteren Bauehhiilfte) dem zuerst durehstr6mten Gef/issgebiete. Die bei Asthenie und Enteroptose so oft geahnte und (vergl. Wenekebach. 1. e) einige l~lale naehgewiesene Splanchnic stasis seheint mir in dieser Weise erkl~irlieh zu sein. Gehen wir nun zum Sehlusse den yon den versebiedenen Autoren mit dem asthenischen Habitns in Verbindung gebraehten 14rankheiten nach, so begegnen wit erstens solchen, die dem erh6hgen intraabdominalen Druek bei naehgiebigem Bindegewebe ihr Entstehen verdanken. Tuffier hat auf das von ibm als ~inf6riorit6 physiologique des tissus a charakterisierte Krankheitsbild aufmerksam gemacht, und gerade bei den Personen~ die dieses zeigen und schlaffes Bindegewebe und mangelhaften Tonus haben, steht das Nullniveau hoeh~ und der auf dem Beekenboden lastende Druek ist also hibher' a l s bei Gesunden. Dass Eingeweidebrtiehe, Vorfall tier Seheide und des Rectums gerade bei diesen Personen h/tufig sind~ darf uns also nieht wundern. In zweiter Linie kommt alas grosse Heer der mit tier Enteroptose in Zusammenhang gebraehten~ mehr oder weniger direkt auf Zirkulationsst6rungen beruhenden Krankheiten. Ausser der sehon erwtihnten Splanehnie stasis (seheinbare Aniimie) und den Ern~ihrungss!6rungen hebe ieh hervoq dass die Beeintr~tehtigung der Blutdurehstdbmung der Leber vielleieht mitwirken mag bei der ]~ntstehung vieler bei dem asthenisehen Habitus vorkommenden Krankheiten des Stoffweehsels. Aueh mehr eireumseripte Aeusserungen einer lokalen ZirkulationsstOrung k/bnnen m6glieherweise ihre prim~ire Ursaehe in den gettnderten Druekverh/tltnissen im Abdomen haben. Glenard und Nontennis erw~thnen das 6ftere Zusammentreffen der Enteroptose mi~ Leberinsuffizienz und Cholelithiasis. 342 Kaiser, Ueber intraabdominalen Druck. Der Enferoptose ebenfalls schreibt Meynert einen urs~iehsS;chliehen Zusammenhang mit der Chlorose zu, wiihrend Singer einen solchen Zusammenhang mi~ dem Magengesehwtir und tier ehronischen Obstipation vermutet. D e l b e t besehreib~ eine eigenartige, bei Enteroptose h~ufige Form des Darmkatarrhs. Weiterhin ist as, naehdem Lejars und Tuffier Anastomosierungen zwisehen Nierenkapselvenen und der Vena portae naehgcwiesen haben, nieht unm6glich, dass es Formen der orthostatisehen A!buminurie gibt, die in den besproehenen Verh~;ltnissen ihre Erkl~rung finden m~gen. Dasselbe kann man vertauten von manehen anderen unter den vielen bei Asthenie und Enteroptose vorkommenden Krankheiten, die aufzuzXblen hier nieht am Platze wiire. Literatur. Aubertin~ Albuminurie orthostatique. Semaine mddieale. 1906. Batraud~ J.~ La Neurasthenic g6nitale feminine. Paris 1906. Bert~ P.~ Legons de Physiologic comparge etc. Paris 1870. Braune, Die Oberschenkelvene des Mensehen. Leipzig 1871. Brugh 7 van der~ Nieuwe metingen omtrenit intrapleurale drukking. Leiden 1896. Carnot: P.~ Appendieite et abcbs sousphrdnique. Semaine mgdicale. 1906. Delbet~ Typhlite ptosique. Semaine mddicale. 1905. Du Bois-I~aymond~ Ergebnisse tier Physiologic. 1902. I. Jahrg. tI. Abt. Duch en n e~ Physiologic der Bewegungen. (Deutsche Uebersetzung.) 1885. Engelen~ Bedeutung und Messung des abdominellen Druckes. Deutsche med. Wochenschr. 1911. Faure~ J. L.~ et Siredey~ Trait6 de Gyn6cologie. Paris 1911. Hasse~ C.~ Arch. f. Anatomic, 1886~ 1903~ 1906 und 1907. - - Die Formvergnderungen des menschlichen KSrpers wg,hrend der Atmung. 1870. Hegar~ A.~ Saugph~nomene im Unterleibe. Dieses Arch. Bd. 4. HSrmann~ Intraabdominaler Druck. Dieses Arch. Bd. 75. goeven~ P. C, T. van der~ Die Asthenio und die Lageanomalien usw. Haarlem 1907. Hagen-Thorn~ Was ist intraabdominaler DrucL Centralbl. f. Gym 1902. Hermann, L.~ Intrapleuraler Druck. Arch. f. Phy s. Bd, 30. Henley Grundriss tier Anatomic. 1880. Henke, Der Bauehraum. Arch. f. Anat. 1891. Hill, L. Centralbl. f. Physiol. 1899. Nr. 14. Hogge, A.~ Pression intraabdominale. Arch. de Biologic. 1892. Nr. 12. K ais er, Ueber intraabdominalen Druck. 343 Kelling, Physikalisohe Untersuchtmgen fiber intraabdominale Druckverh~ltnisse. Volkmanns Sammlung. N.F. 144. Knorr, R.~ Cystoskopie des Weibes. Berlin 1908. Kossmann~ Was ist intraabdominaler Drack. Centralbl. f. Gyn. 1902. Ktistner, O., Lehrbueh der Gyn~kologie. Jena 1910. Linossior ot Lcmaine. Orthostatisme etc. Semaine m6dieale. 1903. Luciani~ Physiologic des Mens6hen. (Deutsch.)" 1905: Meynert~ Uebor Chlorose. u Sammlung 115. Montennis, Les D6s6quilibr6s du ventre. Paris 1894. Mann, Ein neuer Beitrag usw. Deutsche med. Woehenschr. 1891. Mathes ~ Uebor Enteroptoser Dieses Arch. Bd. 77. v': Mehring, Leiirbuch der inneren Medizin. Jena 1903. Meyer~ I~., Was ist intraabdominater Druck. Centralbl. f. Gyn. 1902. Odebreoht~ B.eitrag zur Lehro vom intraabdominalen Druck. Berliner ldin. Woehensehr. 1875. g o s e n t h a l , J-., Ueber d~n intrathorakalen Druok. Arch. f. Phys. 1882. Schatz, Physiologic des Wochonbotts. Dieses Arch. Bd. 4 u. 5. Schfirmayer, Fortschritte auf dem Gebiete der RSntgenstrahlen. Bd. 10. ::: Nr: 6. S c h w e r d t , Enteroptose. Deutsche reed. Wochenschr. 1896. Nr. 47 5, 6. Singer, Asthenische und spastisehe Obstipation. 1909. Stl'a~z, Die SchSnheit des weiblichen KSrpers. Stuttgart 1900. Stiller, B:, Die asthenischo Konstituticnskrankheit/ Stuttgart 1907. T e s t u t , Traitg d'Anatomie humaine. Paris 1900. T u f f l e r , Sui" une maladie caraeterisge etc. Semaine mgdic. 1894. V e r s t r a e t e n , C., Modifications do la pressio n intraabd. AnnaIes de la Soc. Mgd. do Gand. 1887--90: Wenck~ebach~ K. F., Ueber pathologisch0 Beziohungen zwischen Kreislauf und Atmung. Volkmanns Samml. 465--466. - - D e r s elb e ~N. Tydschrift v. Geneeskuade. 1907. i. 592/ Weis Smann, Ueber Enteroptose. Mfinchen 1905. Z w a a r d e m a k e r r Leerboek der Physiologic. Haarlem 1911. I.
https://openalex.org/W2811403087	https://europepmc.org/articles/pmc6023125?pdf=render	English	null	Prevalence, characteristic features, and complications associated with the occurrence of unerupted permanent incisors	PloS one	2,018	cc-by	6,519	Chiewee Tan, Manikandan Ekambaram, Cynthia Kar Yung Yiu* Paediatric Dentistry, Faculty of Dentistry, The University of Hong Kong, Pokfulam, Hong Kong SAR, China * ckyyiu@hkucc.hku.hk * ckyyiu@hkucc.hku.hk Editor: James J. Cray, Jr., Ohio State University, UNITED STATES Copyright: © 2018 Tan et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. RESEARCH ARTICLE a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 OPEN ACCESS Citation: Tan C, Ekambaram M, Yiu CKY (2018) Prevalence, characteristic features, and complications associated with the occurrence of unerupted permanent incisors. PLoS ONE 13(6): e0199501. https://doi.org/10.1371/journal. pone.0199501 Editor: James J. Cray, Jr., Ohio State University, UNITED STATES Editor: James J. Cray, Jr., Ohio State University, UNITED STATES Prevalence, characteristic features, and complications associated with the occurrence of unerupted permanent incisors Chiewee Tan, Manikandan Ekambaram, Cynthia Kar Yung Yiu* Abstract a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 This study examined the prevalence, characteristic features, and complications associated with the occurrence of unerupted permanent incisors among children and adolescents attending a university dental teaching hospital. A retrospective review was performed of the clinical records of children and adolescents who attended the Prince Philip Dental Hospital, Hong Kong between 2005 and 2014. All patients who had at least one unerupted permanent incisor tooth were included. A total of 266 subjects with 320 unerupted permanent incisors were identified. The prevalence of unerupted permanent incisors among children and ado- lescents was 2.0%. Permanent maxillary central incisors (70.6%) were the most commonly affected teeth. The most common cause for unerupted incisors were dilacerations (n = 83, 36.7%) for maxillary central incisors; developmental dental anomalies (n = 22; 30.6%) together with unfavorable root development (n = 22; 30.6%) for maxillary laterals incisors; and abnormal tooth/tissue ratio (n = 11, 50.0%) for mandibular incisors. A majority of uner- upted incisors presented with complications the most common being ectopic/displacement/ rotation of the unerupted incisors (46.6%), loss of space (36.9%) and midline shift (27.5%). In conclusion, the causes were distinct for different manifestations of unerupted permanent incisors. As the majority of unerupted incisors presented with complications, a systematic and organized method of history taking as well as clinical and radiographic examinations is mandatory in the diagnosis of unerupted permanent incisors. Introduction Delayed eruption of maxillary permanent incisors necessitates intervention when [1] Delayed eruption of maxillary permanent incisors necessitates intervention when [1] 1. there is eruption of contralateral teeth that occurred greater than six months previously; Data Availability Statement: All relevant data are within the paper and its Supporting Information files. 2. the lower incisors have erupted greater than one year previously and where both maxillary central incisors remain unerupted; 3. or there is deviation from the normal order of eruption (e.g. lateral incisors erupting prior to the central incisors). Funding: There is no funding support for the research. ) A previous study reported that the incidence of unerupted maxillary incisors in the 5 to 12 year-old age group as 0.13% [2] and more prevalent in Asian races [1]. The etiological factors A previous study reported that the incidence of unerupted maxillary incisors in the 5 to 12 year-old age group as 0.13% [2] and more prevalent in Asian races [1]. The etiological factors Competing interests: The authors have declared that no competing interests exist. PLOS ONE \| https://doi.org/10.1371/journal.pone.0199501 June 28, 2018 1 / 14 Prevalence, characteristic features, and complications associated with the unerupted permanent incisors for delayed eruption of maxillary incisors can be classified into two main categories, namely hereditary and environmental factors. Hereditary factors include supernumerary teeth, cleft lip and palate, odontoma, abnormal tooth/tissue ratio, cleidocranial dysostosis, generalized delayed eruption and gingival fibromatosis. Similarly, trauma, early extraction or loss of primary teeth (with or without space loss), retained primary teeth, cystic formation, endocrine abnormalities, and bone disease are environmental factors that may influence eruption of incisors [1]. To date, published studies have focused on case reports of diagnosis and management of unerupted maxillary teeth [3], retrospective evaluation on eruption [4]; treatment protocol [5]; prevalence of various etiological factors and treatment outcome associated with unerupted incisors [6]; as well as association between unerupted incisors and dental anomalies [7]. How- ever, there is no single study on prevalence, characteristic features and complications associ- ated with unerupted permanent incisors among children and adolescents attending the Prince Philip Dental Hospital (PPDH), which is the only university dental teaching hospital in Hong Kong Special Administrative Region, China. Introduction The patient pool in Pediatric Dentistry and Orthodontics (PDO) clinic of PPDH are predominantly those without a referral, followed by referred patients from the School Dental Care Services (a dental health program for almost all primary school children); private general dental practitioners and medical practitioners. Almost all children and adolescents with unerupted permanent incisors are accepted for com- prehensive multidisciplinary treatment in the PDO clinic. Therefore, the aim of this retrospec- tive study was to determine the prevalence, characteristic features and complications associated with the occurrence of unerupted permanent incisors among children and adoles- cents attending a university dental teaching hospital. PLOS ONE \| https://doi.org/10.1371/journal.pone.0199501 June 28, 2018 Materials and methods This retrospective study reviewed hospital records of children and adolescents who had uner- upted permanent incisors and were treated in the PDO clinic at the PPDH, The University of Hong Kong, Hong Kong SAR, between January 2005 and December 2014. The study protocol was approved by the Institutional Review Board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster (UW 15–538). The records were reviewed from 15,987 children and adolescents who had attended the PDO Clinic of the PPDH. All patients who had at least one unerupted permanent incisor that matched the definition of unerupted permanent incisors which was adapted from the clinical guidelines of Royal College of Surgeons of England [1] were included. Those subjects were defined as having unerupted permanent inci- sors when any of the following occurred: • eruption of contralateral incisors that occurred greater than six months previously; • both maxillary central incisors remain unerupted and the lower incisors have erupted greater than one year previously; • there is deviation from the normal sequence of eruption (e.g. lateral incisors erupting prior to the central incisors). Clinical records of those patients were excluded from this study if they presented with • medical complications (metabolic and endocrine disorders, syndromes, orofacial cleft / cra- niofacial malformations etc.) • congenital missing permanent incisors. After obtaining a list of included patients, clinical notes and radiographs were retrieved to obtain the necessary information. A customized data entry form was used to record all relevant PLOS ONE \| https://doi.org/10.1371/journal.pone.0199501 June 28, 2018 2 / 14 Prevalence, characteristic features, and complications associated with the unerupted permanent incisors information of each patient. Two qualified dentists who worked in the PDO clinic of the PPDH were involved in the data collection stage. The assessment process of clinical reports and radiographs was blinded. The inter- and intra-examiner reliability were assessed by using the Cohen’s Kappa statistics. For each patient included in the study, the relevant information that was recorded included: 1. General descriptive data about the patients that includes gender, date of birth, date of diagnosis 2. Type of radiographs used There were good intra-examiner and inter-examiner reliability with Kappa values ranging from 0.85 to 0.87 and 0.80 to 0.83, respectively. Panoramic (OPG), upper anterior occlusal view (UAO) and periapical (PA) view radiographs were the most common combinations of radiographs used for assessment and diagnosis of unerupted incisors (n = 115, 35.9%). Only 46 cases (14.4%) incorporated Cone Beam Computed Tomography (CBCT) in diagnosis and treatment planning of unerupted incisors. General descriptive data A total of 266 subjects with 320 unerupted permanent incisors were identified from the 15,987 subjects in the study, which represented a prevalence of 2.0%. Out of 266 subjects, male sub- jects (n = 139; 52.3%) were slightly higher than female (n = 127; 47.7%). The age of diagnosis ranged from 7.3 years to 13.8 years with a mean age of 10.6 years (S1 Table). Materials and methods Relevant clinical information such as, • The type of radiographs used for assessment and diagnosis of unerupted incisors • Type, number and etiology of unerupted incisors • History of trauma • The types of supernumerary and odontomas associated with unerupted incisors • Any complications associated with the occurrence of unerupted permanent incisors All the data were entered and processed using the Statistical Package for Social Sciences Version 20.0 for Windows1 (SPSS Inc., Chicago, Illinois, USA). Numerical data were summa- rized. Any significant differences and possible associations among variables were compared using Chi-square test. Fisher’s exact test was used when 20% or more of the cells in the table with expected frequencies less than 5. The binomial test was used to determine any difference in between proportion of presence and absence of complications associated with the occur- rence of unerupted permanent incisors. The p-value was set at 0.05. Type and number of unerupted incisors Among the unerupted permanent incisors, maxillary central incisors were the most frequently affected teeth (70.6%), followed by maxillary lateral incisors (22.5%), mandibular lateral inci- sors (4.1%) and mandibular central incisors (2.8%) (Table 1). Most of the patients presented with only one unerupted permanent incisor (n = 220; 68.8%) and 86 patients (26.9%) had two PLOS ONE \| https://doi.org/10.1371/journal.pone.0199501 June 28, 2018 3 / 14 Prevalence, characteristic features, and complications associated with the unerupted permanent incisors Table 1. Prevalence and percentages of different types of unerupted permanent incisors. Types of incisors involved Prevalence Tooth number n (%) Maxillary centrals (n = 226; 70.6%) 1.41% #8 104 (32.5) #9 122 (38.1) Maxillary laterals (n = 72; 22.5%) 0.45% #7 41 (12.8) #10 31 (9.7) Mandibular centrals (n = 9; 2.8%) 0.06% #24 7 (2.2) #25 2 (0.6) Mandibular laterals (n = 13; 4.1%) 0.08% #23 5 (1.6) #26 8 (2.5) Total 2.00% 320 (100.0) https://doi.org/10.1371/journal.pone.0199501.t001 Table 1. Prevalence and percentages of different types of unerupted permanent incisors. unerupted permanent incisors involved. There were only 6 (1.9%) (as shown in Fig 1) and 8 (2.5%) patients that presented with three and four unerupted permanent incisors, respectively. unerupted permanent incisors involved. There were only 6 (1.9%) (as shown in Fig 1) and 8 (2.5%) patients that presented with three and four unerupted permanent incisors, respectively. PLOS ONE \| https://doi.org/10.1371/journal.pone.0199501 June 28, 2018 unerupted permanent incisors were dilacerations (n = 88, 27.5%), followed by supernumerary teeth (n = 66, 19.1%) and ectopic position of tooth bud (n = 52, 16.3%). For the unerupted max- illary central incisors, the most common causes were dilacerations (n = 83, 36.7%) (Fig 2A–2C), followed by supernumerary teeth (Fig 3A–3C) (n = 49; 21.7%) and ectopic position of tooth bud (n = 32; 14.2%). For the unerupted maxillary lateral incisors, the three most common causes were delayed root development (n = 21; 29.2%); malformed/microdontic/underdevel- oped incisors (n = 20; 27.8%) and ectopic position of tooth bud (n = 15; 20.8%). The most com- mon cause of unerupted mandibular incisors was crowding (abnormal tooth/tissue ratio), which included 5 cases (38.5%) for mandibular lateral incisors and 6 cases (66.7%) for mandibu- lar central incisors. Many of the cases presented with one underlying etiology of unerupted per- manent incisors (n = 243; 75.9%); whereas 64 cases (20%) and 7 cases (2.2%) presented with a combination of two and three causes. Only one case (0.7%) presented with combination of 4 underlying causes which includes compound odontomas and root dilaceration in relation to #9, as well as retained #F and delayed root development of #9 compared with #8. Prevalence, characteristic features, and complications associated with the unerupted permanent incisors Table 2. Identified causes of unerupted permanent incisors. Overall n (%) Maxillary centrals n (%) Maxillary laterals n (%) Mandibular centrals n (%) Mandibular laterals n (%) A. Hereditary causes 1. Supernumerary teeth 61 (19.1) 49 (21.7) 11 (15.3) 0 (0.0) 1 (7.7) 2. Abnormal tooth/tissue ratio (crowding) 29 (9.1) 7 (3.1) 11 (15.3) 6 (66.7) 5 (38.5) 3. Developmental dental anomalies 29 (9.1) 6 (2.6) 22 (30.6) 0 (0.0) 1 (7.7) a. Malformed/microdontic/ underdeveloped 23 (7.2) 3 (1.3) 20 (27.8) 0 (0.0) 0 (0.0) b. Regional odontodysplasia 6 (1.9) 3 (1.3) 2 (2.8) 0 (0.0) 1 (7.7) 4. Odontomas 21 (6.6) 19 (8.4) 1 (1.4) 1 (11.1) 0 (0.0) 5. Generalized delayed eruption 4 (1.3) 2 (0.9) 2 (2.8) 0 (0.0) 0 (0.0) B. Environment causes 6. Dilacerations 88 (27.5) 83 (36.7) 5 (7.0) 0 (0.0) 0 (0.0) a. Root dilacerations 33 (10.3) 30 (13.3) 3 (4.2) 0 (0.0) 0 (0.0) b. Crown dilacerations 31 (9.7) 29 (12.8) 2 (2.8) 0 (0.0) 0 (0.0) c. Crown and root dilacerations 24 (7.5) 24 (10.6) 0 (0.0) 0 (0.0) 0 (0.0) 7. Ectopic position of tooth bud 52 (16.3) 32 (14.2) 15(20.8) 2 (22.2) 3 (23.1) 8. Retained primary teeth 38 (11.9) 23 (10.2) 10 (13.9) 1 (11.1) 4 (30.8) 9. Unfavorable root development 37 (11.6) 14 (6.2) 22 (30.6) 0 (0.0) 1 (7.7) a. Delayed root development 31 (9.7) 9 (4.0) 21 (29.2) 0 (0.0) 1 (7.7) b. Arrested root development 6 (1.9) 5 (2.2) 1 (1.4) 0 (0.0) 0 (0.0) 10. Obstruction by adjacent teeth 15 (4.7) 5 (2.2) 10 (13.9) 0 (0.0) 0 (0.0) 11. Early extraction or loss of primary teeth 14 (4.4) 14 (6.2) 0 (0.0) 0 (0.0) 0 (0.0) a. With space loss 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) b. Without space loss 14 (4.4) 14 (6.2) 0 (0.0) 0 (0.0) 0 (0.0) 12. Pathological changes—Cystic formation etc. 5 (1.6) 5 (2.2) 0 (0.0) 0 (0.0) 0 (0.0) 13. Ankyloses/ impaction of primary teeth 2 (0.6) 0 (0.0) 2 (2.8) 0 (0.0) 0 (0.0) 14. Ankyloses of permanent teeth 1 (0.3) 0 (0.0) 0 (0.0) 1 (11.1) 0 (0.0) https://doi.org/10.1371/journal.pone.0199501.t002 Table 2. Identified causes of unerupted permanent incisors. Etiology of unerupted incisors There were in total 14 causes of unerupted permanent incisors that were identified and catego- rized under either “hereditary factors” or “environmental factors” (Table 2). There were five cases (1.6%) with unknown/unspecified etiology. Overall, the three most common causes for Fig 1. Panoramic radiograph of a 14-year-old girl (with no history of dental trauma) presented with root dilaceration of #9, unerupted #8 with enlarged follicle, unerupted #7 and #10 with delayed root development; as well as an impacted #22. h //d i /10 1371/j l 0199501 001 Fig 1. Panoramic radiograph of a 14-year-old girl (with no history of dental trauma) presented with root dilaceration of #9, unerupted #8 with enlarged follicle, unerupted #7 and #10 with delayed root development; as well as an impacted #22. https://doi.org/10.1371/journal.pone.0199501.g001 PLOS ONE \| https://doi.org/10.1371/journal.pone.0199501 June 28, 2018 4 / 14 PLOS ONE \| https://doi.org/10.1371/journal.pone.0199501 June 28, 2018 Types of supernumerary and odontomas associated with unerupted incisors Among the type of supernumerary teeth associated with unerupted permanent incisors, the conical type of supernumerary tooth (n = 34, 55.7%) was significantly higher in number when compared with tuberculate (n = 12, 19.7%), supplemental supernumerary teeth (n = 10, 16.4%) and unspecified (n = 5, 8.2%) types (p<0.001). Conversely, compound odontomas (n = 17, 81.0%) had a significant higher proportion when compared with complex odontomas (n = 0, 0.0%) (p<0.001). Associated complications with the occurrence of unerupted incisors Overall, a significantly higher number of unerupted incisors presented with associated compli- cations (n = 243; 75.9%) than those without complications (n = 77; 24.1%) (p<0.001). The complications associated with the unerupted permanent incisors can be categorized under general dental aspect, associated with unerupted incisors, associated with adjacent teeth/struc- tures (Table 3). The unerupted maxillary central incisors were found to be most frequently associated with ectopic/displacement/rotation of the unerupted permanent incisor itself (n = 111; 49.1%), space loss (n = 93; 41.2%) and midline shift (n = 61; 27.0%). For the uner- upted maxillary lateral incisors, ectopic/ displacement of the unerupted incisor itself (n = 30; 41.7%) and its adjacent teeth/structures (n = 19; 26.4%) were the two most common complica- tions associated with the occurrence of unerupted permanent incisors. A majority of uner- upted mandibular permanent incisors presented with space loss (n = 8; 88.9%) and midline shift (n = 8; 88.9%). History of dental/ maxillofacial trauma Among all the 14 causes of unerupted permanent incisors, dilacerations (p<0.001), ectopic position of tooth bud (p = 0.012) and ankylosed/impacted primary teeth (p = 0.039) had signif- icant association with positive history of trauma. 5 / 14 PLOS ONE \| https://doi.org/10.1371/journal.pone.0199501 June 28, 2018 Prevalence, characteristic features, and complications associated with the unerupted permanent incisors PLOS ONE \| https://doi.org/10.1371/journal.pone.0199501 June 28, 2018 6 / 1 6 / 14 PLOS ONE \| https://doi.org/10.1371/journal.pone.0199501 June 28, 2018 PLOS ONE \| https://doi.org/10.1371/journal.pone.0199501 June 28, 2018 Prevalence, characteristic features, and complications associated with the unerupted permanent incisors Prevalence, characteristic features, and complications associated with the unerupted permanent incisors Fig 2. a. Upper anterior occlusal view of a 12-year-old girl who presented with a crown and root dilacerations of #8 as well as positive history of traumatic injury to the primary incisor when she was 4-year-old. b. Intra-operative photograph of the same subject which shows the orientation of the dilacerated crown. c. Postoperative photograph of the same subject with the extracted #8 and its associated tooth follicle. Fig 2. a. Upper anterior occlusal view of a 12-year-old girl who presented with a crown and root dilacerations of #8 as well as positive history of traumatic injury to the primary incisor when she was 4-year-old. b. Intra-operative photograph of the same subject which shows the orientation of the dilacerated crown. c. Postoperative photograph of the same subject with the extracted #8 and its associated tooth follicle. https://doi.org/10.1371/journal.pone.0199501.g002 https://doi.org/10.1371/journal.pone.0199501.g002 https://doi.org/10.1371/journal.pone.0199501.g002 PLOS ONE \| https://doi.org/10.1371/journal.pone.0199501 June 28, 2018 PLOS ONE \| https://doi.org/10.1371/journal.pone.0199501 June 28, 2018 Discussion Occlusal view of the same subject showing talon’s cusps and dens invaginatus on palatal surface of erupted #10 (black arrow) after six months of surgical removal of the inverted mesiodens between #8 and #9 Fig 3. a. Periapical radiograph of an 8-year-old boy who presented with an unerupted #8 due to the inverted mesiodens between #8 and #9. b. Frontal view of the same subject with erupting #7 and #8 after one month of surgical removal of the inverted mesiodens between #8 and #9. c. Occlusal view of the same subject showing talon’s cusps and dens invaginatus on palatal surface of erupted #10 (black arrow) after six months of surgical removal of the inverted mesiodens between #8 and #9. https://doi.org/10.1371/journal.pone.0199501.g003 dimensional positioning of the unerupted teeth and adjacent structures [13]. As a general prin- ciple, the recommendations from clinical guidelines of the Royal College of Surgeons of England [1] and the American Academy of Pediatric Dentistry [14] should be adopted during radiographic assessment of unerupted permanent incisors. Findings from the present study showed diversities in the hereditary and environmental causes for each specific type of unerupted permanent incisors (Table 2). Trauma in the pri- mary dentition has the potential to cause eruption disturbances of permanent incisors, espe- cially maxillary central incisors as shown by this study. Acute trauma to the primary dentition can cause dilaceration of the long axis of the permanent successor [15]. Any trauma during odontogenesis can affect the morphogenic stages of dental development and malformations, such as partial or complete arrest of root formation which occur during root formation [16]. Development of odontoma-like malformations in the permanent dentition caused by intru- sion of a primary incisor has also been reported [16, 17]. Therefore, the clinician should advise the parents regarding the possible complications following traumatic dental injuries of the pri- mary dentition. Regular follow-up as recommended by the International Association of Dental Traumatology (IADT) guidelines [18] is also mandatory so that early detection and treatment of possible severe developmental disturbances can be carried out. In addition, the results of this study also identified that the etiologies of unerupted incisors may also occur in isolation or in combination, specifically for maxillary permanent lateral inci- sors, which were most commonly affected by unfavorable root development (extremely delayed/arrested root development) and developmental dental anomalies (malformed/micro- dontic/underdeveloped) (Fig 4) [19, 20]. Discussion The demographic data in our study showed that male subjects (n = 139; 52.3%) were slightly higher than female (n = 127; 47.7%) with a mean age of 10.6 years. This is in accordance with other retrospective studies, which also reported higher prevalence of unerupted maxillary cen- tral incisors in males with a mean age of 10.6 and 9.4 years [5, 8]. The contributing factors for unerupted incisors among male patients could be explained by a greater prevalence of super- numerary teeth in males [9, 10] and also possible involvement of sexual chromosomes in the etiology of tooth eruption disturbances [8]. In the present study, the most common combination of radiographs that were used in assessment and management of unerupted permanent incisors were panoramic radiograph, UAO view and PA view radiographs. This is like another retrospective study [10], in which a combination of more than one film was used in 95% of cases to enable localization of the uner- upted teeth using the parallax method. A study by Jacobs [11] suggested that a panoramic radiograph with an anterior occlusal radiograph is the preferred combination of radiographs to localize unerupted mandibular anterior teeth. Radiographs may be justified when there are abnormal incisor relationships, localization of tooth position is necessary to formulate a treat- ment plan, and where a reasonable expectation that pathology exists on clinical grounds [12]. Cone Beam Computed Tomography (CBCT) could be used to provide valuable information in qualitative analysis of dento-osseous structures, morphological alterations and exact tri- PLOS ONE \| https://doi.org/10.1371/journal.pone.0199501 June 28, 2018 7 / 14 Prevalence, characteristic features, and complications associated with the unerupted permanent incisors Prevalence, characteristic features, and complications associated with the unerupted permanent incisors PLOS ONE \| https://doi.org/10.1371/journal.pone.0199501 June 28, 2018 PLOS ONE \| https://doi.org/10.1371/journal.pone.0199501 June 28, 2018 8 / 14 Prevalence, characteristic features, and complications associated with the unerupted permanent incisors Prevalence, characteristic features, and complications associated with the unerupted permanent incisors Fig 3. a. Periapical radiograph of an 8-year-old boy who presented with an unerupted #8 due to the inverted mesiodens between #8 and #9. b. Frontal view of the same subject with erupting #7 and #8 after one month of surgical removal of the inverted mesiodens between #8 and #9. c. Discussion A 9 years old male subject who initially presented on 15th January 2011, with malformed and delayed root development on #10 (white arrow). #11 (black arrow) had erupted in advance as compared to #10 on a review dated on 1st August 2012. #10 had finally erupted clinically on the subsequent review dated on 21st November 2012 (after 22 months of observation). https://doi.org/10.1371/journal.pone.0199501.g004 https://doi.org/10.1371/journal.pone.0199501.g004 Along with maxillary incisors, this study also addressed the occurrence and key characteris- tic features of 22 cases of unerupted mandibular incisors which were rarely reported in the lit- erature likely due to its low prevalence. This study showed that causes for most unerupted mandibular incisors were due to crowding (abnormal tooth tissue ratio), followed by retained primary incisors and ectopic position of tooth buds. This condition was also often complicated with further space loss and midline shift as shown in Fig 5. Therefore, recognition of potential crowding during primary dentition and early mixed dentition by the clinician is the first important step in reducing the occurrence of unerupted mandibular incisors. Likewise, lack of space in the primary dentition during routine dental examination may raise the concern of future crowding in the permanent dentition [21]. Moreover, any over-retained primary tooth should be extracted to prevent obstruction to the eruption of the permanent successor [1]. Findings of the present study revealed that complications associated with the occurrence of unerupted permanent incisors can have an effect on the rest of the dentition (Fig 5), such as loss of space [22, 23] and midline shift [24, 25]. Other reported complications, which were also well documented in several published studies, include enlarged follicle or cystic changes of unerupted incisors [26, 27] or association with supernumerary teeth [28, 29], root resorption and early exfoliation of adjacent teeth. Complications may also be associated with ectopic/ dis- placement/ rotation of the impacted incisor itself [30, 31] or the adjacent structures/ teeth [32]. For instance, a microdont maxillary lateral incisor displaying delayed development may inhibit Along with maxillary incisors, this study also addressed the occurrence and key characteris- tic features of 22 cases of unerupted mandibular incisors which were rarely reported in the lit- erature likely due to its low prevalence. This study showed that causes for most unerupted mandibular incisors were due to crowding (abnormal tooth tissue ratio), followed by retained primary incisors and ectopic position of tooth buds. Discussion Surgical removal should only be performed when proven complications, such as obstruction of eruption, displacement or root resorption of adjacent teeth, or pathological cystic changes have occurred. Table 3. Complications associated with the occurrence of unerupted permanent incisors. Overall n (%) Maxillary centrals n (%) Maxillary laterals n (%) Mandibular centrals n (%) Mandibular laterals n (%) 1. General dental aspects a. Loss of space 118 (36.90) 93 (41.2) 12 (16.7) 8 (88.9) 8 (88.9) b. Midline shift 88 (27.50) 61 (27.0) 12 (16.7) 8 (88.9) 8 (88.9) 2. Associated with unerupted incisors a. Ectopic/displacement/rotation 149 (46.60) 111 (49.1) 30 (41.7) 2 (22.2) 2 (22.2) b. Enlarged follicle 53 (16.6) 37 (16.4) 10 (13.9) 5 (55.6) 1 (7.7) c. Cystic changes 5 (1.60) 5 (2.2) 0 (0.0) 0 (0.0) 0 (0.0) 3. Associated with adjacent teeth/structures a. Ectopic/displacement/rotation 83 (25.90) 58 (25.7) 19 (26.4) 1 (11.1) 1 (11.1) b. Enlarged follicle 3 (0.9) 2 (0.9) 0 (0.0) 1 (11.1) 0 (0.0) c. Root resorption 5 (1.60) 0 (0.0) 2 (2.8) 0 (0.0) 0 (0.0) d. Early exfoliation 4 (1.30) 0 (0.0) 1 (1.4) 0 (0.0) 0 (0.0) e. Overeruption of opposing incisors 8 (2.50) 7 (3.1) 1 (1.4) 0 (0.0) 0 (0.0) 4. No complication 77 (24.10) 57 (25.2) 18 (25.0) 0 (0.0) 0 (0.0) htt //d i /10 1371/j l 0199501 t003 le 3. Complications associated with the occurrence of unerupted permanent incisors. Table 3. Complications associated with the occurrence of unerupted permanent incisors. PLOS ONE \| https://doi.org/10.1371/journal.pone.0199501 June 28, 2018 PLOS ONE \| https://doi.org/10.1371/journal.pone.0199501 June 28, 2018 9 / 14 Prevalence, characteristic features, and complications associated with the unerupted permanent incisors Fig 4. A 9 years old male subject who initially presented on 15th January 2011, with malformed and delayed root development on #10 (white arrow). #11 (black arrow) had erupted in advance as compared to #10 on a review dated on 1st August 2012. #10 had finally erupted clinically on the subsequent review dated on 21st November 2012 (after 22 months of observation). Fig 4. A 9 years old male subject who initially presented on 15th January 2011, with malformed and delayed root deve arrow) had erupted in advance as compared to #10 on a review dated on 1st August 2012. #10 had finally erupted clinically November 2012 (after 22 months of observation). Fig 4. PLOS ONE \| https://doi.org/10.1371/journal.pone.0199501 June 28, 2018 Discussion This condition was also often complicated with further space loss and midline shift as shown in Fig 5. Therefore, recognition of potential crowding during primary dentition and early mixed dentition by the clinician is the first important step in reducing the occurrence of unerupted mandibular incisors. Likewise, lack of space in the primary dentition during routine dental examination may raise the concern of future crowding in the permanent dentition [21]. Moreover, any over-retained primary tooth should be extracted to prevent obstruction to the eruption of the permanent successor [1]. p p p Findings of the present study revealed that complications associated with the occurrence of unerupted permanent incisors can have an effect on the rest of the dentition (Fig 5), such as loss of space [22, 23] and midline shift [24, 25]. Other reported complications, which were also well documented in several published studies, include enlarged follicle or cystic changes of unerupted incisors [26, 27] or association with supernumerary teeth [28, 29], root resorption and early exfoliation of adjacent teeth. Complications may also be associated with ectopic/ dis- placement/ rotation of the impacted incisor itself [30, 31] or the adjacent structures/ teeth [32]. For instance, a microdont maxillary lateral incisor displaying delayed development may inhibit PLOS ONE \| https://doi.org/10.1371/journal.pone.0199501 June 28, 2018 10 / 14 Prevalence, characteristic features, and complications associated with the unerupted permanent incisors Fig 5. Panoramic radiograph of an 8-year-old boy showing the complications associated with the occurrence of an unerupted and ectopic #26. https://doi org/10 1371/journal pone 0199501 g005 n 8-year-old boy showing the complications associated with the occurrence of an unerupted and ectopic #26. Fig 5. Panoramic radiograph of an 8-year-old boy showing the complications associated with the occurrence of an unerupted and ectopic #26. Fig 5. Panoramic radiograph of an 8-year-old boy showing the complications associated with the occurrence of an unerupted and ectopic #26. https://doi.org/10.1371/journal.pone.0199501.g005 Fig 5. Panoramic radiograph of an 8-year-old boy showing the complications associated with the occurrence of an unerupted and ectopic #26. https://doi.org/10.1371/journal.pone.0199501.g005 eruption of maxillary central incisor(s) and maxillary canine(s) (Fig 6). In a case series, Koba- yashi et al. [19] reported that developmental anomalies and immaturity involving adjacent per- manent lateral incisors could be associated with eruption failure of maxillary permanent central incisors. Discussion However, the immature tooth germ does not change the eruptive direction of the tooth as much as do the presence of odontomas and supernumerary teeth, even if the developing tooth germ is positioned close to the unerupted tooth [19]. Nevertheless, early detection of unerupted permanent incisors would be essential to prevent the impact of the above-mentioned complications, reduce the treatment complexity and thus improve the clini- cal outcome. Due to the retrospective nature of the present study, the major limitation was information bias which could have led to missing data for analysis and interpretation. This was also com- pounded by the lack of completeness and standardized method of history taking, clinical and radiographic examinations in the information gathering process. PLOS ONE \| https://doi.org/10.1371/journal.pone.0199501 June 28, 2018 Conclusions 1. The prevalence of unerupted permanent incisors among children and adolescents attending a university dental teaching hospital was 2.0%. 1. The prevalence of unerupted permanent incisors among children and adolescents attending a university dental teaching hospital was 2.0%. 2. The most common etiologies for unerupted incisors were dilacerations for maxillary central incisors; developmental dental anomalies together with unfavorable root development for maxillary laterals incisors; and abnormal tooth/tissue ratio for mandibular incisors. 3. The majority of unerupted incisors presented with complications, which can have an effect on the remaining dentition (i.e. space loss and midline shift), as well as on the impacted incisor itself and/or its adjacent structures/teeth. 3. The majority of unerupted incisors presented with complications, which can have an effect on the remaining dentition (i.e. space loss and midline shift), as well as on the impacted incisor itself and/or its adjacent structures/teeth. 11 / 14 PLOS ONE \| https://doi.org/10.1371/journal.pone.0199501 June 28, 2018 Prevalence, characteristic features, and complications associated with the unerupted permanent incisors Fig 6. Panoramic radiograph of an 8-year-old boy showing the unerupted #10 with delayed root development (white arrow), which also acted as an obstruction to the eruption of #9 (black arrow) and #11 (dotted arrow). -old boy showing the unerupted #10 with delayed root development (white arrow), which also acted as an obstruction to dotted arrow). Fig 6. Panoramic radiograph of an 8-year-old boy showing the unerupted #10 with delayed root development (white arrow), which also acted as an obstruction to the eruption of #9 (black arrow) and #11 (dotted arrow). https://doi.org/10.1371/journal.pone.0199501.g006 Supporting information S1 Table. Original data of 266 subjects with unerupted permanent incisors from clinical record and radiographs. (XLSX) Conceptualization: Manikandan Ekambaram. Conceptualization: Manikandan Ekambaram. Writing – original draft: Chiewee Tan. Writing – original draft: Chiewee Tan. Writing – review & editing: Manikandan Ekambaram, Cynthia Kar Yung Yiu. References 1. Yaqoob O, O’Neill J, Patel S, Seehra J, Bryant C, Noar J, et al. Management of unerupted maxillary inci- sors. 2016. 1. Yaqoob O, O’Neill J, Patel S, Seehra J, Bryant C, Noar J, et al. Management of unerupted maxillary inci- sors. 2016. 12 / 14 PLOS ONE \| https://doi.org/10.1371/journal.pone.0199501 June 28, 2018 Prevalence, characteristic features, and complications associated with the unerupted permanent incisors 2. MacPhee C. The incidence of erupted supernumerary teeth in consecutive series of 4000 school chil- dren. Br Dent J. 1935; 58:59–60. 2. MacPhee C. The incidence of erupted supernumerary teeth in consecutive series of 4000 school chil- dren. Br Dent J. 1935; 58:59–60. 3. Ayers E, Kennedy D, Wiebe C. Clinical recommendations for management of mesiodens and unerupted permanent maxillary central incisors. European archives of paediatric dentistry: official journal of the European Academy of Paediatric Dentistry. 2014; 15(6):421–8. Epub 2014/07/06. https://doi.org/10. 1007/s40368-014-0132-1 PMID: 24994110. 4. Leyland L, Batra P, Wong F, Llewelyn R. A retrospective evaluation of the eruption of impacted perma- nent incisors after extraction of supernumerary teeth. The Journal of clinical pediatric dentistry. 2006; 30(3):225–31. Epub 2006/05/11. PMID: 16683671. 5. Lygidakis NN, Chatzidimitriou K, Theologie-Lygidakis N, Lygidakis NA. Epsilonvaluation of a treatment protocol for unerupted maxillary central incisors: retrospective clinical study of 46 children. Eur Arch Paediatr Dent. 2015; 16(2):153–64. Epub 2014/11/06. https://doi.org/10.1007/s40368-014-0150-z PMID: 25370386. 6. Betts A, Camilleri GE. A review of 47 cases of unerupted maxillary incisors. Int J Paediatr Dent. 1999; 9 (4):285–92. Epub 2000/05/18. PMID: 10815587. 7. Bartolo A, Camilleri A, Camilleri S. Unerupted incisors—characteristic features and associated anoma- lies. Eur J Orthod. 2010; 32(3):297–301. Epub 2009/09/12. https://doi.org/10.1093/ejo/cjp094 PMID: 19745002. 8. Topkara A, Sari Z. Impacted teeth in a turkish orthodontic patient population: prevalence, distribution and relationship with dental arch characteristics. European journal of paediatric dentistry: official journal of European Academy of Paediatric Dentistry. 2012; 13(4):311–6. Epub 2012/12/29. PMID: 23270290. 9. Mitchell L, Bennett TG. Supernumerary teeth causing delayed eruption—a retrospective study. British journal of orthodontics. 1992; 19(1):41–6. Epub 1992/02/01. PMID: 1562577. 10. Mason C, Azam N, Holt RD, Rule DC. A retrospective study of unerupted maxillary incisors associated with supernumerary teeth. The British journal of oral & maxillofacial surgery. 2000; 38(1):62–5. Epub 2000/04/28. https://doi.org/10.1054/bjom.1999.0210 PMID: 10783451. 11. Jacobs SG. Radiographic localization of unerupted mandibular anterior teeth. American Journal of Orthodontics and Dentofacial Orthopedics. 2000; 118(4):432–8. https://doi.org/10.1067/mod.2000. 108783 PMID: 11029740 12. Taylor NG, Jones AG. PLOS ONE \| https://doi.org/10.1371/journal.pone.0199501 June 28, 2018 References Are anterior occlusal radiographs indicated to supplement panoramic radiogra- phy during an orthodontic assessment? Br Dent J. 1995; 179(10):377–81. Epub 1995/11/25. PMID: 8519560. 13. Jeremias F, Fragelli CM, Mastrantonio SD, Dos Santos-Pinto L, Dos Santos-Pinto A, Pansani CA. Cone-beam computed tomography as a surgical guide to impacted anterior teeth. Dental research jour- nal. 2016; 13(1):85–9. Epub 2016/03/11. https://doi.org/10.4103/1735-3327.174723 PMID: 26962322; PubMed Central PMCID: PMCPMC4770477. 14. Guideline on prescribing dental radiographs for infants, children, adolescents, and persons with special health care needs. Pediatric dentistry. 2012; 34(5):189–91. Epub 2012/12/06. PMID: 23211908. 15. Topouzelis N, Tsaousoglou P, Pisoka V, Zouloumis L. Dilaceration of maxillary central incisor: a litera- ture review. Dent Traumatol. 2010; 26(5):427–33. Epub 2010/09/14. https://doi.org/10.1111/j.1600- 9657.2010.00915.x PMID: 20831640. 16. Gungormus M, Yolcu U, Aras MH, Halicioglu K. Simultaneous occurrence of compound odontoma and arrested root formation as developmental disturbances after maxillofacial trauma: a case report. Medi- cina oral, patologia oral y cirugia bucal. 2010; 15(2):e398–400. Epub 2009/09/22. PMID: 19767697. 17. Shaked I, Peretz B, Ashkenazi M. Development of odontoma-like malformation in the permanent denti- tion caused by intrusion of primary incisor—a case report. Dental traumatology: official publication of International Association for Dental Traumatology. 2008; 24(3):e395–7. Epub 2008/05/21. https://doi. org/10.1111/j.1600-9657.2008.00564.x PMID: 18489475. 18. Flores MT, Malmgren B, Andersson L, Andreasen JO, Bakland LK, Barnett F, et al. Guidelines for the management of traumatic dental injuries. III. Primary teeth. Dent Traumatol. 2007; 23(4):196–202. Epub 2007/07/20. https://doi.org/10.1111/j.1600-9657.2007.00627.x PMID: 17635351. 19. Kobayashi H, Taguchi Y, Noda T. Eruption disturbances of maxillary permanent central incisors associ- ated with anomalous adjacent permanent lateral incisors. International journal of paediatric dentistry / the British Paedodontic Society [and] the International Association of Dentistry for Children. 1999; 9 (4):277–84. Epub 2000/05/18. PMID: 10815586. 20. Torres JN, Caracas HC, Bolognese AM, Torres S. Conservative approach for a patient with extreme delay in maxillary lateral incisor development. American journal of orthodontics and dentofacial orthope- dics: official publication of the American Association of Orthodontists, its constituent societies, and the American Board of Orthodontics. 2012; 141(6):773–82. Epub 2012/05/30. https://doi.org/10.1016/j. ajodo.2010.08.019 PMID: 22640679. PLOS ONE \| https://doi.org/10.1371/journal.pone.0199501 June 28, 2018 13 / 14 Prevalence, characteristic features, and complications associated with the unerupted permanent incisors 21. Society. MotCSCotBO. Managing the Developing Occlusion—A guide for dental practitioners: British Orthodontic Society Revised and Updated February 2010. 22. Pavoni C, Franchi L, Lagana G, Baccetti T, Cozza P. PLOS ONE \| https://doi.org/10.1371/journal.pone.0199501 June 28, 2018 References Management of impacted incisors following sur- gery to remove obstacles to eruption: a prospective clinical trial. Pediatric dentistry. 2013; 35(4):364–8. Epub 2013/08/13. PMID: 23930638. 23. Pinho T, Neves M, Alves C. Impacted maxillary central incisor: surgical exposure and orthodontic treat- ment. American journal of orthodontics and dentofacial orthopedics: official publication of the American Association of Orthodontists, its constituent societies, and the American Board of Orthodontics. 2011; 140(2):256–65. Epub 2011/08/02. https://doi.org/10.1016/j.ajodo.2009.11.018 PMID: 21803264. 24. Nuvvula S, Melkote TH, Mohapatra A, Nirmala SV. Impacted mandibular permanent incisors related to supernumerary teeth: a rare condition. Pediatric dentistry. 2012; 34(1):70–3. Epub 2012/02/23. PMID: 22353462. 25. Sant’Anna EF, Marquezan M, Sant’Anna CF. Impacted incisors associated with supernumerary teeth treated with a modified Haas appliance. Am J Orthod Dentofacial Orthop. 2012; 142(6):863–71. Epub 2012/12/01. https://doi.org/10.1016/j.ajodo.2011.07.030 PMID: 23195372. 26. Rohilla M, Namdev R, Dutta S. Dentigerous cyst containing multiple impacted teeth: a rare case report. Journal of the Indian Society of Pedodontics and Preventive Dentistry. 2011; 29(3):244–7. Epub 2011/ 10/12. https://doi.org/10.4103/0970-4388.85834 PMID: 21985882. 27. Jena AK. Management of Multiple Impacted Teeth Associated with a Large Dentigerous Cyst in the Maxilla. Journal of dentistry for children (Chicago, Ill). 2015; 82(3):157–62. Epub 2016/01/06. PMID: 26731252. 28. Sharma D, Garg S, Singh G, Swami S. Trauma-induced dentigerous cyst involving an inverted impacted mesiodens: case report. Dental traumatology: official publication of International Association for Dental Traumatology. 2010; 26(3):289–91. Epub 2010/06/25. https://doi.org/10.1111/j.1600-9657. 2010.00882.x PMID: 20572846. 29. Shun Y. Dentigerous cyst associated with an impacted anterior maxillary supernumerary tooth. Journal of dentistry for children (Chicago, Ill). 2008; 75(1):104–7. Epub 2008/05/29. PMID: 18505659. 30. Ferrazzano GF, Cantile T, Roberto L, Baldares S, Manzo P, Martina R. An impacted central incisor due to supernumerary teeth: a multidisciplinary approach. European journal of paediatric dentistry: official journal of European Academy of Paediatric Dentistry. 2014; 15(2 Suppl):187–90. Epub 2014/08/08. PMID: 25101499. 31. Ferres-Amat E, Maura-Solivellas I, Prats-Armengol J, Ferres-Amat E, Mareque-Bueno J, Ferres-Padro E. Study of the frequency, localisation and morphology of supernumerary teeth in 1960 Spanish non- syndromic paediatric patients. European journal of paediatric dentistry: official journal of European Academy of Paediatric Dentistry. 2015; 16(1):19–23. Epub 2015/03/21. PMID: 25793948. 32. Punde PA, Patil NM, Pawar RL. Unusual presentation of ’u-shaped’ impacted maxillary central incisor with intranasal root: successful surgical management. Ethiopian journal of health sciences. 2014; 24 (3):273–6. Epub 2014/09/04. PMID: 25183935; PubMed Central PMCID: PMCPMC4141232. 14 / 14
https://openalex.org/W2149940066	https://substanceabusepolicy.biomedcentral.com/counter/pdf/10.1186/1747-597X-8-12	English	null	Factors associated with self-assessed increase in tobacco consumption among over-indebted individuals in Germany: a cross-sectional study	Substance abuse treatment, prevention, and policy	2,013	cc-by	6,920	RESEARCH Open Access Open Access Factors associated with self-assessed increase in tobacco consumption among over-indebted individuals in Germany: a cross-sectional study Heiko Rueger1, Heide Weishaar2, Elke B Ochsmann3, Stephan Letzel1 and Eva Muenster1 Correspondence: heiko.rueger@uni-mainz.de 1Institute for Occupational, Social, and Environmental Medicine, University Medical Center of the University of Mainz, Obere Zahlbacher Strasse 67, Mainz D-55131, Germany Full list of author information is available at the end of the article Rueger et al. Substance Abuse Treatment, Prevention, and Policy 2013, 8:12 http://www.substanceabusepolicy.com/content/8/1/12 Rueger et al. Substance Abuse Treatment, Prevention, and Policy 2013, 8:12 http://www.substanceabusepolicy.com/content/8/1/12 Rueger et al. Substance Abuse Treatment, Prevention, and Policy 2013, 8:12 http://www.substanceabusepolicy.com/content/8/1/12 © 2013 Rueger et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2013 Rueger et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Abstract Background: Over-indebtedness is an increasing phenomenon in industrialised nations causing individual hardship and societal problems. Nonetheless, few studies have explored smoking among over-indebted individuals. Background: Over-indebtedness is an increasing phenomenon in industrialised nations causing individual hardship and societal problems. Nonetheless, few studies have explored smoking among over-indebted individuals. Methods: A cross-sectional survey (n=949) on retrospectively assessed changes in tobacco consumption was Background: Over-indebtedness is an increasing phenomenon in industrialised nations causing individual hardship and societal problems. Nonetheless, few studies have explored smoking among over-indebted individuals. Methods: A cross-sectional survey (n=949) on retrospectively assessed changes in tobacco consumption was carried out in 2006 and 2007 among clients of 84 officially approved debt and insolvency counselling centres in Germany (response rate 39.7%). Logistic regressions were performed to explore factors associated with reports of increased smoking after onset of over-indebtedness. Methods: A cross-sectional survey (n=949) on retrospectively assessed changes in tobacco consumption was carried out in 2006 and 2007 among clients of 84 officially approved debt and insolvency counselling centres in Germany (response rate 39.7%). Logistic regressions were performed to explore factors associated with reports of increased smoking after onset of over-indebtedness. Results: 63% of all respondents stated daily or occasional tobacco consumption. Almost one fifth reported an increase in smoking after becoming over-indebted. Females were less likely to report increased smoking than men (aOR 0.66, 95% CI 0.44-0.99) whereas respondents who had been over-indebted for more than 10 years were more likely to report increased smoking than those who had been over-indebted for less than five years (aOR 1.66; 95%-CI 1.00-2.76). The odds of increased smoking were also elevated among those who reported that their families and friends had withdrawn from them as a consequence of their over-indebtedness (aOR 1.82; 95%-CI 1.06-3.14). Conclusions: The study identifies over-indebted individuals and particularly over-indebted men as a high-risk group of smokers. Low levels of social embeddedness/support were associated with a further increase in smoking after becoming over-indebted. Given recent increases of over-indebtedness, the findings highlight the need to develop appropriate public health policies. Keywords: Over-indebtedness, Bankruptcy, Smoking, Financial stress, Social support Keywords: Over-indebtedness, Bankruptcy, Smoking, Financial stress, Social support Keywords: Over-indebtedness, Bankruptcy, Smoking, Financial stress, Social support Study design and participants The study utilized data from the “OI-survey”, a cross- sectional survey on health status and health-related lifestyles in over-indebted individuals using a self- administered German language questionnaire. 949 eligible individuals who attended one of all 84 officially approved debt and insolvency counselling centres in the German Länder (provinces) of Rhineland-Palatinate (n=666) and Mecklenburg-West Pomerania (n=283) and satisfied the recruitment criteria (attended a minimum of two counsel- ling sessions at one of the centres between 2006 and 2007, ≥16 years of age) were enrolled in the study. In order to obtain a sample of unrelated individuals, only one individual per household was interviewed. If more than one individual attended the counselling session, the over-indebted person was interviewed. If both attendees (usually married or unmarried couples) were affected, it was left to the attendees to decide who would complete the questionnaire. The respondent who filled in the questionnaire was asked to report on his or her own smoking habits. The response rate was 39.7%. All 84 of- ficially approved debt and insolvency counselling cen- tres located in the two provinces were included in the survey. In Germany, debt and insolvency counselling of- fered by officially approved counselling centres is free of charge for the attendee. During the first appointment, the counselor assesses whether the attendee meets the eligibility criteria for receiving debt and insolvency counselling of “being in serious financial difficulties”. The fact that demand for debt counselling exceeds sup- ply by far and the rigorous recruitment criteria of the study (participants had to have attended at least two counselling sessions) meant that study participants were almost certainly over-indebted. Further study design details are published elsewhere [11,14,15]. A test for representativeness conducted among the Rhineland- Palatinate sample reported no differences in sex and age compared to official 2006 statistics of all counselling clients. Marital status, nationality and amount of debt, however, revealed statistically significant differences, with non-German citizens and married individuals be- ing slightly underrepresented and single individuals being slightly overrepresented. Outcome measures An individual’s smoking status was surveyed using the item “Have you smoked previously or do you currently smoke?” Participants who reported being daily or occa- sional smokers were classified as smokers. Participants who reported that they used to smoke were classified as ex-smokers. Changes in smoking after onset of over- indebtedness were examined using the item “Has your life- style regarding consumption of tobacco products changed since you have become over-indebted?” Possible answers were “I consume less tobacco than before”, “I consume the same amount of tobacco as before”, “I consume more to- bacco than before” and “Does not apply”. The increase in tobacco consumption after onset of over-indebtedness was chosen as the target category (=1), whereas participants who reported decreased or unchanged tobacco consump- tion after onset of over-indebtedness served as the refe- rence category (=0). Background households describes the situation when, over an ex- tended period, a household’s income is insufficient to service its debts on time despite a reduction of the stan- dard of living [5]. Over the last decades, the number of over-indebted private households and individuals has in- creased. Current estimates suggest that approximately three million German households are affected and reveal a high share of over-indebted households in many other industrialised nations, varying between 2% in Finland and 12% in the United States [6,7]. Over-indebtedness and fi- nancial difficulties are related concepts as they usually affect individuals of lower socio-economic status. Recent studies have found an association with health and health behaviour [8-13], revealing higher prevalences of overweight and obesity [14], back pain [15] and mental disorder [16] in over-indebted individuals in Germany. The current financial and economic crisis has resulted in several European countries having stagnating economic growth, increasing unemployment rates and marked re- ductions in public welfare spending. Possible health con- sequences, including increased suicide rates, problematic health behaviour and mental disorders, are beginning to receive some (public and) academic attention [1-3]. Given that research shows that loss of employment and low income are major drivers of over-indebtedness [4], it can be assumed that the current crisis will lead to an increase in over-indebtedness. Over-indebtedness of individuals or * Correspondence: heiko.rueger@uni-mainz.de 1Institute for Occupational, Social, and Environmental Medicine, University Medical Center of the University of Mainz, Obere Zahlbacher Strasse 67, Mainz D-55131, Germany Full list of author information is available at the end of the article Rueger et al. Substance Abuse Treatment, Prevention, and Policy 2013, 8:12 http://www.substanceabusepolicy.com/content/8/1/12 Page 2 of 8 Ethical approval was obtained from the medical associ- ation of Rhineland-Palatinate, Germany. Research in high-income countries has further sug- gested an association between financial difficulties and smoking cessation, with those experiencing financial stress being less likely to make quit attempts or suc- cessfully quit smoking [17-20]. The study at hand adds to this area of research and examines smoking in over- indebted individuals in Germany and factors associated with changes in tobacco consumption in times of over-indebtedness. Covariates To identify potential risk factors for an increase in smoking, multi- variate binary logistic regressions were performed (Table 2). In a first step (model 1), perceived social sup- port was analysed in combination with the covariates and confounders mentioned above. In a second step (model 2), changes in social embeddedness following onset of over-indebtedness was additionally included in the analysis. Adjusted odds ratios and 95%-confidence intervals were calculated, with the level of statistical significance being set at α≤0.05. Missing data were ex- cluded for bivariate analyses and placed in a separate category for multivariate analyses. SPSS version 17 (SPSS Inc., Chicago, Illinois) statistical software package was used for all analyses. Results Smoking in over-indebted individuals A total of 598 over-indebted individuals (63.0%) reported to either smoke daily or occasionally. Men were signifi- cantly more likely to smoke (67.7%) than women (59.3%, χ2=7.142, df=1, p=.008). 168 participants (17.7%) reported to be ex-smokers. 74.7% (n=572) of all current and former smokers answered the question regarding a possible change in tobacco consumption after onset of over- indebtedness. 39.0% of this subpopulation (n=223, 23.5% of the entire study sample) reported to smoke less, 30.4% (n=174, 18.3% of the entire study sample) reported no changes in tobacco consumption, and 30.6% (n=175, 18.4% of the entire study sample) reported increased smoking. Description of the study sample The characteristics of the subpopulation of over-indebted individuals who reported changes in tobacco consumption after onset of over-indebtedness examined in this study (N=572) are presented in Table 1. About one half of the study sample was female (48.6%) and aged under 40 years (50.3%), with age ran- ging from 18 to 79 years. About one third was divorced, separated or widowed (38.3%) and had undergone more than nine years of schooling (33.6%). One fifth had a net viduals who answered hange in tobacco debtedness (n=572) N % 294 51.4 278 48.6 131 22.9 157 27.4 189 33.0 92 16.1 3 0.5 380 66.4 192 33.6 445 77.8 127 22.2 156 27.3 193 33.7 219 38.3 4 0.7 201 35.1 186 32.5 166 29.0 19 3.3 129 22.6 218 38.1 139 24.3 79 13.8 7 1.2 203 35.5 265 46.3 100 17.5 4 0.7 303 53.0 155 27.1 103 18.0 11 1.9 Statistical methods Initially, smoking in over-indebted individuals and over- all changes in tobacco consumption (i.e. increased, de- creased or unchanged consumption) following onset of over-indebtedness were examined. Covariates For this, the analysis covered the complete sample of 949 participants. In a second step, factors associated with an increase in to- bacco consumption after onset of over-indebtedness were examined. For this analysis, 572 current and former smokers who answered the question regarding a possible change in smoking were investigated. Smoking prevalence and changes in prevalence after onset of over-indebtedness were calculated. To identify potential risk factors for an increase in smoking, multi- variate binary logistic regressions were performed (Table 2). In a first step (model 1), perceived social sup- port was analysed in combination with the covariates and confounders mentioned above. In a second step (model 2), changes in social embeddedness following onset of over-indebtedness was additionally included in the analysis. Adjusted odds ratios and 95%-confidence intervals were calculated, with the level of statistical significance being set at α≤0.05. Missing data were ex- cluded for bivariate analyses and placed in a separate category for multivariate analyses. SPSS version 17 (SPSS Inc., Chicago, Illinois) statistical software package was used for all analyses. Results Smoking in over-indebted individuals A total of 598 over-indebted individuals (63.0%) reported to either smoke daily or occasionally. Men were signifi- cantly more likely to smoke (67.7%) than women (59.3%, χ2=7.142, df=1, p=.008). 168 participants (17.7%) reported to be ex-smokers. 74.7% (n=572) of all current and former smokers answered the question regarding a possible change in tobacco consumption after onset of over- indebtedness. 39.0% of this subpopulation (n=223, 23.5% of the entire study sample) reported to smoke less, 30.4% (n=174, 18.3% of the entire study sample) reported no changes in tobacco consumption, and 30.6% (n=175, 18.4% of the entire study sample) reported increased smoking. Description of the study sample The characteristics of the subpopulation of over-indebted individuals who reported changes in tobacco consumption after onset of over-indebtedness examined in this study (N=572) are presented in Table 1. About one half of the study sample was female (48.6%) and aged under 40 years (50.3%), with age ran- ging from 18 to 79 years. About one third was divorced, separated or widowed (38.3%) and had undergone more than nine years of schooling (33.6%). Covariates In order to measure changes regarding social embedded- ness since the beginning of over-indebtedness, partici- pants were asked whether anything had changed in their family or circle of acquaintances since they started suf- fering from financial hardship. An index including the following three categories was generated: “neither family members nor friends have withdrawn”, “either family members or friends have withdrawn”, and “family mem- bers and friends have withdrawn”. The measurement of perceived social support was carried out using the stan- dard F-SozU, a measuring instrument based on self- assessment of perceived social support. A short version of the scale with 14 items (Cronbach’s Alpha=.94) was used [21]. The resulting scale was stratified a priori into four categories. Health awareness was surveyed by ask- ing respondents about the degree to which they were concerned about their health with the first and last two attributes of a five point categorical scale being sub- sumed. The duration of over-indebtedness was consid- ered in order to measure chronic stress and categorised a priori into “five years or less”, “six to ten years” and “more than ten years”. The question “To what degree do you feel burdened by your debts?” (response categories: “not at all”, “slightly”, “moderately”, “highly” and “very highly”) was used as an additional proxy for an indivi- dual’s perceived levels of stress. Well-established socio-demographic confounders in- cluding age, gender, formal education, net household income and marital status were categorised a priori. The categorisation of these measures is detailed in Table 1. Page 3 of 8 Page 3 of 8 Rueger et al. Substance Abuse Treatment, Prevention, and Policy 2013, 8:12 http://www.substanceabusepolicy.com/content/8/1/12 Rueger et al. Substance Abuse Treatment, Prevention, and Policy 2013, 8:12 http://www.substanceabusepolicy.com/content/8/1/12 Rueger et al. Substance Abuse Treatment, Prevention, and Policy 2013, 8:12 http://www.substanceabusepolicy.com/content/8/1/12 Statistical methods Initially, smoking in over-indebted individuals and over- all changes in tobacco consumption (i.e. increased, de- creased or unchanged consumption) following onset of over-indebtedness were examined. For this, the analysis covered the complete sample of 949 participants. In a second step, factors associated with an increase in to- bacco consumption after onset of over-indebtedness were examined. For this analysis, 572 current and former smokers who answered the question regarding a possible change in smoking were investigated. Smoking prevalence and changes in prevalence after onset of over-indebtedness were calculated. Covariates One fifth had a net Table 1 Descriptive statistics of individuals who answered the question regarding a possible change in tobacco consumption after onset of over-indebtedness (n=572) N % Sex Male 294 51.4 Female 278 48.6 Age 18-30 131 22.9 31-40 157 27.4 41-50 189 33.0 51+ 92 16.1 n.a. 3 0.5 Educational level ≤9 years 380 66.4 > 9 years 192 33.6 Income ≤1500 € 445 77.8 > 1500 € 127 22.2 Marital status Married 156 27.3 Unmarried 193 33.7 Divorced/Separated/Widowed 219 38.3 n.a. 4 0.7 Duration of over-indebtedness ≤5 years 201 35.1 6 to 10 years 186 32.5 > 10 years 166 29.0 n.a. 19 3.3 Social support High 129 22.6 Less than high 218 38.1 More than low 139 24.3 Low 79 13.8 n.a. 7 1.2 Health awareness High 203 35.5 Medium 265 46.3 Low 100 17.5 n.a. 4 0.7 Changes in social embeddedness since onset of over-indebtedness No withdrawal 303 53.0 Friends or relatives withdrawn 155 27.1 Friends and relatives withdrawn 103 18.0 n.a. 11 1.9 Notes: OI-survey. Table 1 Descriptive statistics of individuals who answered the question regarding a possible change in tobacco consumption after onset of over-indebtedness (n=572) Statistical methods Initially, smoking in over-indebted individuals and over- all changes in tobacco consumption (i.e. increased, de- creased or unchanged consumption) following onset of over-indebtedness were examined. For this, the analysis covered the complete sample of 949 participants. In a second step, factors associated with an increase in to- bacco consumption after onset of over-indebtedness were examined. For this analysis, 572 current and former smokers who answered the question regarding a possible change in smoking were investigated. SPSS version 17 (SPSS Inc., Chicago, Illinois) statistical software package was used for all analyses. 38.3 0.7 Rueger et al. Substance Abuse Treatment, Prevention, and Policy 2013, 8:12 http://www.substanceabusepolicy.com/content/8/1/12 Rueger et al. Substance Abuse Treatment, Prevention, and Policy 2013, 8:12 http://www.substanceabusepolicy.com/content/8/1/12 Rueger et al. Statistical methods Substance Abuse Treatment, Prevention, and Policy 2013, 8:12 http://www.substanceabusepolicy.com/content/8/1/12 Page 4 of 8 Page 4 of 8 household income of more than 1500 Euros (22 2%) A since onset of over-indebtedness was reported by a fifth Table 2 Factors associated with increased smoking in over-indebted individuals (reference category: decreased or unchanged tobacco consumption) Increase in smoking Model 1c Model 2cd Row percentages aOR 95% CI aOR 95% CI Sex p=.018, df=1a Male (ref.) 35.0 1 1 Female 25.9 0.66 0.44-0.99 0.67 0.45-0.99 Age p=.393, df=3a 18-30 (ref.) 26.7 1 1 31-40 35.0 1.03 0.57-1.84 1.00 0.56-1.80 41-50 31.7 0.80 0.45-1.44 0.78 0.43-1.42 51+ 27.2 0.63 0.30-1.30 0.62 0.30-1.29 Educational level p=.665, df=570b ≤9 years (ref.) 30.0 1 1 > 9 years 31.8 1.08 0.73-1.62 1.12 0.75-1.68 Income p=.119, df=570b ≤1500 € (ref.) 29.0 1 1 > 1500 € 36.2 1.34 0.84-2.12 1.38 0.86-2.20 Marital status p=.251, df=2a Married (ref.) 32.1 1 1 Unmarried 26.4 0.73 0.42-1.26 0.78 0.45-1.36 Divorced/Separated/Widowed 33.8 1.10 0.68-1.77 1.14 0.70-1.86 Duration of over-indebtedness p=.010, df=551b ≤5 years (ref.) 26.4 1 1 6 to 10 years 29.0 1.05 0.65-1.68 1.04 0.64-1.68 > 10 years 39.2 1.76 1.00-2.73 1.66 1.00-2.76 Social support p<.001, df=563b High (ref.) 22.5 1 1 Less than high 26.6 1.21 0.71-2.06 1.10 0.64-1.89 More than low 36.7 1.88 1.07-3.30 1.51 0.83-2.74 Low 45.6 2.78 1.45-5.26 1.99 0.99-3.98 Health awareness p<.001, df=566b High (ref.) 22.2 1 1 Medium 33.2 1.59 1.03-2.47 1.66 1.07-2.58 Low 42.0 1.98 1.15-3.43 2.15 1.23-3.76 Changes in social embeddedness since onset of over-indebtedness p<.001, df=559b No withdrawal (ref.) 24.1 1 Friends or relatives withdrawn 35.5 1.45 0.91-2.31 Friends and relatives withdrawn 41.7 1.82 1.06-3.14 Notes: OI-survey; a chi2 test; b t-test (Spearman correlation); c multivariate binary logistic regression (n=572); d changes in social embeddedness following onset of over-indebtedness was additionally included in the analysis. Table 2 Factors associated with increased smoking in over-indebted individuals (reference category: decreased or unchanged tobacco consumption) ociated with increased smoking in over-indebted individuals (reference category: decreased or o consumption) ith increased smoking in over-indebted individuals (reference category: decreased or ti ) Notes: OI-survey; a chi2 test; b t-test (Spearman correlation); c multivariate binary logistic regression (n=572); d changes in social embeddedness following onset of over-indebtedness was additionally included in the analysis. Discussion C d Considering that approximately 16 million people smoke and 3 million households are estimated to be affected by over-indebtedness in Germany alone, the absence of studies which have investigated the relationship between smoking and an individual’s financial strain as experienced in a situation of over-indebtedness is staggering. On the basis of a cross-sectional survey conducted among over- indebted individuals in Germany, this study investigates smoking prevalence and factors associated with increased tobacco consumption in over-indebted individuals. The measure for perceived stress revealed high levels of stress among over-indebted individuals and almost no variation: 87% of the study sample reported feeling highly or very highly burdened, suggesting that over- indebtedness placed an extraordinary strain on the af- fected individuals. Accordingly, the analysis presented in Table 2 revealed that the longer the period of over- indebtedness, the more likely individuals were to in- crease tobacco consumption. 39.2% of all respondents who had been over-indebted for more than 10 years reported an increase in smoking, whereas only 26.4% of those who had experienced over-indebtedness for less than 5 years reported the same. This difference could also be seen in the fully adjusted model 2 (aOR 1.66; 95%-CI 1.00-2.76). The hypothesis that duration of over-indebtedness has a significant impact on changes in smoking could thus be confirmed. A striking result is the high smoking prevalence among over-indebted individuals, with 63% of all respondents reporting daily or occasional tobacco consumption. In comparison, smoking prevalence in the general German population is considerably lower (34%) and even the sub- population with the highest tendency to smoke (40- to 59-year-old men of low socio-economic status) are signifi- cantly less likely to smoke (54.8%) than over-indebted individuals [22]. The findings suggest that smoking is par- ticularly prevalent among over-indebted individuals. In addition, smokers seem to change their consumption in times of financial hardship with one third reporting an increase in tobacco consumption after onset of over- indebtedness. Health awareness, social embeddedness and perceived social support (in case of not controlling for so- cial embeddedness) were associated with lower risk of increased tobacco consumption whereas duration of over- indebtedness was associated with higher risk. In line with other research showing higher smoking prevalence among men, male respondents were more likely to smoke and had a significantly higher risk to increase their tobacco consumption during over-indebtedness than women. Statistical methods since onset of over-indebtedness was reported by a fifth of all participants (18.0%), whereas more than one fourth reported that either friends or relatives had with- drawn (27.1%) and about half had experienced no with- drawal (53.0%). household income of more than 1500 Euros (22.2%). A third of the study sample had been over-indebted for ten years or more (29.0%), 35.5% reported high levels of health awareness, and 22.6% reported high levels of so- cial support. The withdrawal of friends and relatives Page 5 of 8 Rueger et al. Substance Abuse Treatment, Prevention, and Policy 2013, 8:12 http://www.substanceabusepolicy.com/content/8/1/12 Rueger et al. Substance Abuse Treatment, Prevention, and Policy 2013, 8:12 http://www.substanceabusepolicy.com/content/8/1/12 Factors associated with increased smoking 45.6% in those reporting little social support (Table 2). This effect remained significant in model 1 but lost sta- tistical significance when controlling for changes in so- cial embeddedness after onset of over-indebtedness (model 2), reflecting the close relationship between both variables. It is possible that previous changes in social embeddedness had an impact on the reporting of per- ceived social support at the time of the survey. According to the bivariate analysis, sex, duration of over- indebtedness, health awareness, perceived social support and changes in social embeddedness since onset of over- indebtedness were significantly associated with the out- come measure of increased smoking (Table 2). Men had a significantly higher risk to increase tobacco consumption during over-indebtedness than women. 35.0% of the men as opposed to 25.9% of the women reported such behavioural change. This effect was maintained when multivariate analysis was applied to the data (Table 2). Other socio-demographic and socio-economic variables, including age, education, income and marital status, showed no significant effects in bivariate or multivariate analyses. Discussion C d Individuals with higher levels of health awareness had a lower risk of increased tobacco consumption after on- set of over-indebtedness. 42.0% of all participants who reported low levels of health awareness reported in- creased smoking, whereas this was the case for only 22.2% of all participants reporting high levels of health awareness. This finding was also depicted in the multi- variate analysis and thereby proved to be independent of other factors investigated (Table 2, model 2). Withdrawal of friends and relatives was correlated with a significantly higher risk of increased smoking. Re- spondents who reported that friends and relatives with- drew from them after onset of over-indebtedness had a 1.82 times higher risk of increased smoking in the fully adjusted model (95%-CI 1.06-3.14) compared to partici- pants who did not report that their friends and relatives withdrew (Table 2, model 2). As expected, changes in social embeddedness were closely related to perceived social support (Spearman correlation=-.412, t=10.657, df=557, p<.001). Moreover, increased tobacco consump- tion was observed in those who experienced lower levels of social support. Percentages of increased smoking va- ried between 22.5% in participants reporting strong and A number of limitations of the study have to be consid- ered when discussing the data. First, the cross-sectional design which does not allow for causal explanations regarding over-indebtedness, prevalence and changes in smoking constitutes one of the major weaknesses of our study. Participants’ responses regarding changes since on- set of over-indebtedness serve as proxies to assess changes over time, but a potential recall bias due to retrospective smoking behavior measures [23] has to be acknowledged. Moreover, the respective questionnaire item directly linked financial problems to possible changes in smoking patterns and might therefore have provoked anticipated Rueger et al. Substance Abuse Treatment, Prevention, and Policy 2013, 8:12 http://www.substanceabusepolicy.com/content/8/1/12 Page 6 of 8 Rueger et al. Substance Abuse Treatment, Prevention, and Policy 2013, 8:12 http://www.substanceabusepolicy.com/content/8/1/12 progressive withdrawal of family members and friends might lead to lower levels of perceived social support in over-indebted individuals. As perceived social support decreases, an increase in tobacco consumption can be expected. Conclusions l Our analysis emphasises the need for health promotion interventions to be tailored to the social context in which over-indebted people live and for policy change to promote health among this group of high-risk indivi- duals. In order to develop their full potential, attempts to decrease smoking prevalence should not only focus on changing individual behaviour but need to incorporate policies which address the social determinants of health. While population-wide tobacco control interventions can reduce the social gradient, targeted interventions are needed to reach particular high-risk groups, including in- dividuals who are burdened by financial strain [32]. In this context, our study shows that over-indebted indivi- duals and particularly over-indebted men constitute a specific target group for health promotion. The fact that almost one fourth of all over-indebted individuals reported that they had managed to reduce smoking since the onset of over-indebtedness suggests that considerable potential exists among this population for smoking re- duction and cessation measures, a finding which should be an encouragement when developing smoking cessation services. This study identifies over-indebted individuals as a high-risk group for detrimental health behaviour and thus corresponds with research relating over-indebtedness and financial difficulties to a number of other health- related behaviours [10,12-15]. We acknowledge that the size of the effects measured in this study is relatively weak. However, the fact that approximately 6 million people are estimated to be affected by over-indebtedness in Germany alone illustrates the magnitude of the problem. Given the anticipated number of affected individuals, even small ef- fects can be assumed to have a large impact on overall public health. The findings thus justify suggestions for policy and practice. As expected, health awareness is associated with a lower risk of increased tobacco consumption, suggesting that individuals who pay more attention to their health are more likely to be concerned about their tobacco con- sumption. The duration of over-indebtedness seems to be a crucial factor when analysing changes in tobacco consumption. One possible explanation for the fact that increased tobacco consumption correlates with duration of over-indebtedness might be that increased tobacco consumption is employed as a mechanism to cope with the financial, emotional and social stress caused by over- indebtedness. This hypothesis is supported by previous research which suggests that disadvantaged smokers perceive smoking as a strategy to deal with daily stress [28-30]. Discussion C d Although the inverse relationship between social embeddedness/perceived social support and to- bacco consumption supports previous research which identifies social support as a protective factor when cop- ing with stress [31], the considerable research gap about social embeddedness/social support and changes in to- bacco consumption and the fact that our findings are based on self-reports highlights the need for further research in this specific area. answers. The fact that almost one third of all respondents reported no changes in smoking, however, suggests that respondents did not seem to feel obliged to report changes into any specific direction. Second, all results are based on self-reports. While previous studies suggest that self-reports of smoking provide valid data [24], the reli- ance on self-reports for all measures may imply limitations for the interpretation of our data. Third, a potential selec- tion bias due to a low response rate and sampling having to be based on consent might have skewed the results. It needs to be noted, however, that low response rates can be expected when conducting research on those of low socio-economic status [25]. One of the difficulties in understanding and tackling the problem of smoking in over-indebted individuals is the lack of knowledge about potential pathways between the two factors. In addition to previous research discussing in- creased stress, lack of social support, and lower rates of successful quit attempts as potential mediators of the cor- relation between low socio-economic status and smoking [26-28], our study points to a number of factors which might explain why many of over-indebted individuals increase their smoking. Conclusions l Policies are needed to ensure that debt counsellors as well as medical professionals are alert to the problem, able to assess the prevalence of smoking among their clients and act appropriately. A response to the problem should include respective training for people working in the field, the development of routine assessments of risky health behaviour among over- indebted individuals and the systematic coordination between debt counselling centres and the health service. Debt counselling centres should further be considered as an appropriate, and perhaps not fully exploited, set- ting to implement intensified and target-specific health promotion interventions. Our analysis suggests that the provision of smoking cessation and other health promo- tion measures alongside debt counselling promises to have a considerable potential for reaching a high-risk group of heavy smokers and could constitute a highly effective health promotion measure. 6. Angele J: Überschuldung – letzter ausweg privatinsolvenz. https://www.destatis. de/DE/Publikationen/STATmagazin/WirtschaftsrechnungenZeitbudget/2008_1/ Wirtschaftsrechnungen2008_1.html. 7. Reifner U, Springeneer H: Die private Überschuldung im internationalen Vergleich – Trends, Probleme, Lösungsansätze. In Schulden-kompass. Edited by Schufa Holding AG. Wiesbaden: Index Electronic Prepress; 2004:160–211. 8. Drentea P, Lavrakas P: Over the limit: the association among health, race and debt. Soc Sci Med 2000, 50:517–529. 9. Grafova I: Your money or your life: managing money, managing health. J Fam Econ Issues 2007, 28:285–303. 10. Jacoby MB: Does indebtedness influence health? A preliminary inquiry. J Law Med Ethics 2002, 30:560–571. 11. Rueger H, Schneider NF, Zier U, Letzel S, Muenster E: Health risks of separated or divorced over-indebted fathers: separation from children and financial distress. Soc Work Health Care 2011, 50:242–256. 12. Butterworth P, Rodgers B, Windsor TD: Financial hardship, socio-economic position and depression: results from the PATH through life survey. Soc Sci Med 2009, 69:229–237. 12. Butterworth P, Rodgers B, Windsor TD: Financial hardship, socio-economic position and depression: results from the PATH through life survey. Soc Sci Med 2009, 69:229–237. 13. Meltzer H, Bebbington P, Brugha T, Jenkins R, McManus S: Personal debt and suicidal ideation. Psych Med 2011, 41:771–778. 13. Meltzer H, Bebbington P, Brugha T, Jenkins R, McManus S: Personal debt and suicidal ideation. Psych Med 2011, 41:771–778. 14. Muenster E, Rueger H, Ochsmann E, Letzel S, Toschke AM: Over-indebtedness as a marker of socioeconomic status and its association with obesity: a cross-sectional study. BMC Publ Health 2009, 9:286. 15. Conclusions l Almost all respondents in our study reported to be burdened by their debts, suggesting higher levels of stress among this population and potentially higher urge to develop coping mechanisms to deal with the situ- ation. On the other hand, the study highlights the role of social embeddedness and support for understanding why over-indebted individuals are particularly likely to smoke and increase their tobacco consumption after on- set of over-indebtedness. The analysis shows that the risk of increased tobacco consumption is higher among individuals who experience lower levels of social embed- dedness and social support. Social embeddedness and social support seem to be crucial factors for coping with financial and emotional strain and constitute protective factors regarding tobacco consumption. Assuming that an individual’s perceived social support correlates with his or her embeddedness in a social structure, the One of the policy implications of this study concerns the co-ordination between sectors which have access to financially burdened individuals. A 2009 UK article, which draws attention to the correlation between debt and men- tal health, suggests “co-ordinated ‘debt care pathways’ Page 7 of 8 Page 7 of 8 Rueger et al. Substance Abuse Treatment, Prevention, and Policy 2013, 8:12 http://www.substanceabusepolicy.com/content/8/1/12 between local health and advice services” [33]. Similarly, the findings of this study indicate that a coordinated re- sponse is needed to tackle the relationship between debt and smoking. Policies are needed to ensure that debt counsellors as well as medical professionals are alert to the problem, able to assess the prevalence of smoking among their clients and act appropriately. A response to the problem should include respective training for people working in the field, the development of routine assessments of risky health behaviour among over- indebted individuals and the systematic coordination between debt counselling centres and the health service. Debt counselling centres should further be considered as an appropriate, and perhaps not fully exploited, set- ting to implement intensified and target-specific health promotion interventions. Our analysis suggests that the provision of smoking cessation and other health promo- tion measures alongside debt counselling promises to have a considerable potential for reaching a high-risk group of heavy smokers and could constitute a highly effective health promotion measure. between local health and advice services” [33]. Similarly, the findings of this study indicate that a coordinated re- sponse is needed to tackle the relationship between debt and smoking. Acknowledgments h k 20. Caleyachetty A, Lewis S, McNeill A, Leonardi-Bee J: Struggling to make ends meet: exploring pathways to understand why smokers in financial difficulties are less likely to quit successfully. Eur J Public Health 2012, 22(Suppl 1):41–48. We thank Dr Amrit Caleyachetty for his very useful comments on an earlier draft of this paper. Authors’ contributions 18. Siahpush M, Yong HH, Borland R, Reid JD, Hammond D: Smokers with financial stress are more likely to want to quit but less likely to try or succeed: findings from the international tobacco control (ITC) four country survey. Addiction 2009, 104:1382–1390. HR analysed the data and drafted and revised the paper. He is guarantor. HW drafted and revised the paper. EBO revised the paper. SL initiated the study and revised the paper. EM initiated the study, designed data collection tools, monitored data collection, and revised the paper. All authors read and approved the final manuscript. 19. Kendzor DE, Businelle MS, Costello TJ, Castro Y, Reitzel LR, Cofta-Woerpel LM: Financial strain and smoking cessation among racially/ethnically diverse smokers. Am J Public Health 2010, 100:702–706. Competing interests Thi k 16. Rueger H, Löffler I, Ochsmann E, Alsmann C, Letzel S: Mental illness and over-indebtedness. Mental illness, social networks and financial strain in over-indebted persons. Psychother Psychosom Med Psychol 2010, 60:250–254. This work was supported by the German Federal Ministry for Science, Further Education, Research and Culture of Rhineland-Palatinate as part of the research cluster “Corporate interdependence and social networks”. The funding organization was not involved in any stage at this research. No non-financial competing interests declared. 17. Siahpush M, Carlin JB: Financial stress, smoking cessation and relapse: results from a prospective study of an Australian national sample. Addiction 2006, 101:121–127. Conclusions l Ochsmann E, Rueger H, Letzel S, Drexler H, Muenster E: Over-indebtedness and its association with the prevalence of back pain. BMC Publ Health 2009, 9:451. Author details 1I f O 21. Fydrich T, Sommer G, Tydecks S, Brähler E: Fragebogen zur sozialen unterstützung (F-SozU): Normierung der Kurzform (K-14). Z Med Psychol 2009, 18:43–48. 1Institute for Occupational, Social, and Environmental Medicine, University Medical Center of the University of Mainz, Obere Zahlbacher Strasse 67, Mainz D-55131, Germany. 2Centre for Population Health Sciences, University of Edinburgh, Teviot Place, Edinburgh EH8 9AG, UK. 3Institute for Occupational and Social Medicine; Medical Faculty, RWTH Aachen University, 1Institute for Occupational, Social, and Environmental Medicine, University Medical Center of the University of Mainz, Obere Zahlbacher Strasse 67, y Mainz D-55131, Germany. 2Centre for Population Health Sciences, University of Edinburgh, Teviot Place, Edinburgh EH8 9AG, UK. 3Institute for 22. Lampert T: Smoking, physical inactivity, and obesity: associations with social status. Deutsches Ärzteblatt Int 2010, 107:1–7. Occupational and Social Medicine; Medical Faculty, RWTH Aachen University, Pauwelstrasse 30, Aachen D-52074, Germany. Occupational and Social Medicine; Medical Faculty, RWTH Aachen University, Pauwelstrasse 30, Aachen D-52074, Germany. 23. Kenkel D, Lillard D, Mathios A: Smoke or Fog? The usefulness ofretrospectively reported information about smoking. Addiction 2003, 98:1307–1313. Received: 9 July 2012 Accepted: 26 December 2012 Published: 13 March 2013 Received: 9 July 2012 Accepted: 26 December 2012 Published: 13 March 2013 24. Vartiainen E, Seppala T, Lillsunde P, Puska P: Validation of self reported smoking by serum cotinine measurement in a community-based study. J Epidemiol Community Health 2002, 56:167–170. 25. Lorant V, Demarest S, Miermans P-J, Van Oyen H: Survey error in measuring socio-economic risk factors of health status: a comparison of a survey and a census. Int J Epidemiol 2007, 36:1292–1299. 29. Grafova I: Financial strain and smoking. J Fam Econ Issues 2011, 32:327–340. References 1. Uutela A: Economic crisis and mental health. Curr Opin Psychiatry 2010, 23:127–30. 1. Uutela A: Economic crisis and mental health. Curr Opin Psychiatry 2010, 23:127–30. 26. Wiltshire S, Bancroft A, Parry O, Amos A: ‘I came back here and started smoking again’: perceptions and experiences of quitting among disadvantaged smokers. Health Educ Res 2003, 18:292–303. 2. Stuckler D, Basu S, Suhrcke M, Coutts A, McKee M: Effects of the 2008 recession on health: a first look at European data. Lancet 2011, 378:124–5. 2. Stuckler D, Basu S, Suhrcke M, Coutts A, McKee M: Effects of the 2008 recession on health: a first look at European data. Lancet 2011, 378:124–5. 27. Kotz D, West R: Explaining the social gradient in smokingcessation: it's not in the trying, but in the succeeding. Tob Control 2009, 18:43–6. 3. Shaw BA, Agahi N, Krause N: Are changes in financial strain associated with changes in alcohol use and smoking among older adults? J Stud Alcohol Drugs 2011, 72:917–925. 3. Shaw BA, Agahi N, Krause N: Are changes in financial strain associated with changes in alcohol use and smoking among older adults? J Stud Alcohol Drugs 2011, 72:917–925. 28. Peretti-Watel P, Constance J: “It’s all we got left”. Why poor smokers are less sensitive to cigarette price increases. Int J Environ Res 2009, 6:608–21. 4. Kempson E: Over-indebtedness in Britain. London: Department of Trade and Industry; 2002. 29. Grafova I: Financial strain and smoking. J Fam Econ Issues 2011, 32:327–340. 29. Grafova I: Financial strain and smoking. J Fam Econ Issues 2011, 32:327–340. 5. European Commission: Towards a common operational European definition of over-indebtedness. Brussels: European Communities; 2008. Page 8 of 8 Page 8 of 8 Rueger et al. Substance Abuse Treatment, Prevention, and Policy 2013, 8:12 http://www.substanceabusepolicy.com/content/8/1/12 30. Pyle SA, Haddock CK, Poston WS, Bray RM, Williams J: Tobacco use and perceived financial strain among junior enlisted in the U.S. Military in 2002. Prev Med 2007, 45:460–463. 31. Thoits PA: Mechanisms linking social ties and support to physical and mental health. J Health Soc Behav 2011, 52:145–161. 32. Marmot M: Fair society. London: Healthy Lives. The Marmot Review; 2010. 33. Jenkins R, Fitch C, Hurlston M, Walker F: Recession, debt and mental health: challenges and solutions. Ment Health Fam Med 2009, 6:85–90. References doi:10.1186/1747-597X-8-12 Cite this article as: Rueger et al.: Factors associated with self-assessed increase in tobacco consumption among over-indebted individuals in Germany: a cross-sectional study. Substance Abuse Treatment, Prevention, and Policy 2013 8:12. 30. Pyle SA, Haddock CK, Poston WS, Bray RM, Williams J: Tobacco use and perceived financial strain among junior enlisted in the U.S. Military in 2002. Prev Med 2007, 45:460–463. 31. Thoits PA: Mechanisms linking social ties and support to physical and mental health. J Health Soc Behav 2011, 52:145–161. 32. Marmot M: Fair society. London: Healthy Lives. The Marmot Review; 2010. 33. Jenkins R, Fitch C, Hurlston M, Walker F: Recession, debt and mental health: challenges and solutions. Ment Health Fam Med 2009, 6:85–90. doi:10.1186/1747-597X-8-12 Cite this article as: Rueger et al.: Factors associated with self-assessed increase in tobacco consumption among over-indebted individuals in Germany: a cross-sectional study. Substance Abuse Treatment, Prevention, and Policy 2013 8:12. doi:10.1186/1747-597X-8-12 Cite this article as: Rueger et al.: Factors associated with self-assessed increase in tobacco consumption among over-indebted individuals in Germany: a cross-sectional study. Substance Abuse Treatment, Prevention, and Policy 2013 8:12. Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit
https://openalex.org/W4206292776	https://journals.us.edu.pl/index.php/RIAS/article/download/405/255	English	null	Calls for Papers and Contributions	Science-fiction studies	2,020	cc-by-sa	4,420	Review of International American Studies № 1, VOL. 1 SEPTEMBER 2006 ISSN 19912773 EDITORINCHIEF Michael Boyden RIAS IT AND GRAPHICS TEAM ASSOCIATE EDITOR: Paweł Jędrzejko IT ADVISORS: Tomasz Adamczewski and Wojciech Liber GRAPHIC DESIGNER: Karolina Wojdała DTP ADVISOR: Michał Derda-Nowakowski EDITORIAL BOARD Theo D’haen, Anders Olsson, Liam Kennedy, Sieglinde Lemke, Giorgio Mariani, Ian Tyrrell, Helmbrecht Breinig, Rosario Faraudo, Djelal Kadir TYPESETTING: Ex Machina Editions, Poland e-mail: info@exmachina.pl Review of International American Studies (RIAS) is the electronic journal of the International American Studies Association, the only worldwide, independent, non-governmental association of American Stud- i RIAS f h l b l k f i i l h l h d d f A EDITORINCHIEF Michael Boyden Review of International American Studies (RIAS) is the electronic journal of the International American Studies Association, the only worldwide, independent, non-governmental association of American Stud- ies. RIAS serves as agora for the global network of international scholars, teachers, and students of Amer- ica as hemispheric and global phenomenon. RIAS is published three times a year: in September, January and May by IASA with the institutional support of the University of Silesia in Katowice lending server space to the IASA websites and the electronic support of the Soft For Humans CMS Designers. Subscription rates for RIAS are included along with the Association’s annual dues as speciﬁed in the “Membership” section of the Association’s website (www.iasaweb.org). All topical manuscripts should be directed to the Editor via online submission forms available at RIAS website (www.iasa-rias.org). General correspondence and matters concerning the functioning of RIAS should be addressed to RIAS Editor-in-Chief: Michael Boyden Harvard University Department of English and American Literature and Language Barker Center 12 Quincy Street Cambridge, MA 02138 USA e-mail: michael.boyden@iasa-rias.org For the RIAS cover, we used a fragment of ‘Abstraction’, a work by Automaton [http://fickr.com/people/sbai/], licensed under Creative Commons Attribution-NonCommercial-ShareAlike 2.0 Volume 1, Number 1 2 Calls for Papers and Contributions Venue Venue Faculty of Letters, University of Lisbon Faculty of Letters, University of Lisbon Deadlines — 300–word abstracts and proposals for thematic workshops to be submitted no later than 31 December 2006. — Notifcations of acceptance will be sent out no later than 28 February 2007 INTERNATIONAL AMERICAN STUDIES ASSOCIATION IASA 3RD WORLD CONGRESS, 2023 SEPTEMBER 2007 TRANS/AMERICAN, TRANS/OCEANIC, TRANS/LATION If 1492 marks the advent of modernity, this congress will investigate the implications of the Co- lumbian exchange on the development of culture and identity in the Americas. As a result of the exchange of seeds, plants, animals, the exchange of languages and transplantation of peoples, particularly the extraordinary reach of the African slave trade, the subsequent arrival of peoples from Asia, and the impact of violence against Native peoples in the New World, the Americas have been a particularly fruitful site for exploring the meaning of modernity. We welcome comparative papers apropos of the congress title that explore themes across national geographies in the Americas, across the Atlantic and Pacifc spaces of intercontinental contact, or across language traditions in the Americas. We also welcome papers focused on par- ticular nationalities, including the United States, that help to illuminate the efects and ramifca- tions of a modernity fostered by exploration, conquest, settlement, and globalization. The aim of the congress is to address, among others, the following questions: What kind of entity called ‘America’ is it we study when approached across national, oceanic, or language boundaries? How do we reconcile the liberating potential of hybridity, creolization, or other forms of trans- culturation in light of the histories of forced transplantation and migration and oppression that characterize much American experience? What are the future prospects for an American culture considered in this broad context? What is the role of a globalized American culture produced by the United States thwarting or unwittingly enabling the emergence of new cultural forms? How have the modern media, modern means of transportation, and other means of intercultural communication shifted the meaning of ‘America’ since early colonial contacts? What character- izes sites of resistance to the homogenizing efects of a globalized American culture? As an interdisciplinary organization, IASA welcomes papers and workshops that address these and related questions in the context of analyses of cultural, historical, political, and theoretical material. Fax: +351 217960063 Local Organizing Committee João Ferreira Duarte Helena C. Buescu Maria Teresa Alves Maria Teresa Cid Alexandra Assis Rosa Local Local Faculdade de Letras da Universidade de Lisboa Faculdade de Letras da Universidade de Lisboa Ofcial languages of the congress: English, French, Portuguese, Spanish Ofcial languages of the congress: English, French, Portuguese, Spanish Congress URL: www.iasa2007.eu E-mail address: iasa2007@f.ul.pt Phone: +351 217920085 Congress URL: www.iasa2007.eu E-mail address: iasa2007@f.ul.pt September 2006 45 Prazos Prazos — Até 31 de Decembro de 2006: entrega de resumos de comunicações de 300 palavras e pro- postas de ‘workshops’ temáticos. — Até 28 de Fevereiro de 2007: envio de notifca — Até 28 de Fevereiro de 2007: envio de notifcações de aceitação. International American Studies Association (IASA) 3º Congresso, 20–23 de Setembro, 2007 Trans/americano, trans/oceânico, trans/lativo Se 1492 assinala o advento da modernidade, este congresso examinará as implicações das via- gens de Colombo para o desenvolvimento da cultura e identidade das Américas. Como conse- quência da troca de sementes, plantas e animais, da permuta de línguas e da transplantação de povos, em especial do extraordinário alcance do comércio de escravos em África, da subse- quente chegada dos povos asiáticos e do impacte da violência contra os nativos do Novo Mun- do, as Américas têm constituído um espaço privilegiado para a exploração do sentido da mo- dernidade. Convidamos à apresentação de propostas de comunicação de índole comparatista investigando temas que cruzem as geografas nacionais das Américas, os espaços de contactos intercontinentais tais como o Atlântico e o Pacífco, ou as línguas das Américas e suas tradições. Serão também tidas em consideração comunicações que tratem de nacionalidades específcas, incluindo os EUA, e sejam especialmente relevantes para a clarifcação dos efeitos e ramifca- ções de uma modernidade engendrada pela exploração, conquista, colonização e globalização. É propósito do congresso abordar, entre outras, as seguintes questões: que espécie de entidade é esta chamada ‘América’ que estudamos através de fronteiras nacionais, oceânicas e linguísti- cas? Como conciliar o potencial libertador de formas de transculturação como a hibridização e a crioulização, por exemplo, com as históRIAS de transplantação e migração violentas e de opres- são que caracterizam muita da experiência americana? Que perspectivas se abrem no futuro a uma cultura americana considerada neste contexto alargado? Pode uma cultura americana globalizada produzida pelos EUA funcionar como obstáculo ao surgimento de novas formas cul- turais ou, pelo contrário, involuntariamente facilitar a sua promoção? De que modo os modernos meios de comunicação, de transporte ou de contacto intercultural têm transformado o signif- cado de ‘América’ desde os tempos coloniais? O que caracteriza os lugares de resistência aos efeitos homogeneizadores de uma cultura americana globalizada? Enquanto organização inter- disciplinar, IASA convida à apresentação de comunicações e ‘workshops’ sobre estes problemas, não excluindo outros que com eles se relacionem, no contexto de análises de natureza cultural, histórica, política e teórica. Calls for Papers and Contributions URL do congresso: www./iasa2007.eu Morada electrónica: iasa2007@f.ul.pt Telef.: +351 217920085 Fax: +351 217960063 Comissão Organizadora Local João Ferreira Duarte Helena C. Buescu Maria Teresa Alves Maria Teresa Cid Alexandra Assis Rosa URL do congresso: www./iasa2007.eu Morada electrónica: iasa2007@f.ul.pt Telef.: +351 217920085 Fax: +351 217960063 URL do congresso: www./iasa2007.eu Morada electrónica: iasa2007@f.ul.pt Comissão Organizadora Local João Ferreira Duarte Helena C. Buescu Maria Teresa Alves Maria Teresa Cid Alexandra Assis Rosa International American Studies Association (IASA) 3º Congreso, 20–23 de septiembre de 2007 Trans/americano, trans/oceánico, tra(ns)ducción Si 1492 marca la llegada de la modernidad, este congreso examinará las implicaciones de los via- jes de Colón para el desarrollo de la cultura y de la identidad de las Américas. Como consecuen- cia del intercambio de semillas, plantas y animales, del contacto entre lenguas y del trasplante de poblaciones enteras – especialmente la extraordinaria infuencia del comercio de esclavos africanos, la posterior llegada de población asiática y el impacto de la violencia contra los nati- vos del Nuevo Mundo –, las Américas han venido representando un espacio privilegiado para la exploración del sentido de la modernidad. Invitamos a presentar propuestas de comunicación de naturaleza comparatista, que investiguen temas que crucen las geografías nacionales de las Américas, los espacios de contacto intercontinentales, como el Atlántico o el Pacífco, o las len- guas de las Américas y sus tradiciones. También serán consideradas comunicaciones que se ocu- pen de nacionalidades específcas, incluyendo los EE.UU., y que sean especialmente relevantes para iluminar los efectos y ramifcaciones de una modernidad engendrada por la exploración, conquista, colonización y globalización. El propósito del congreso es abordar, entre otras, las siguientes cuestiones: ¿qué especie de entidad es ésta llamada ‘América’, que estudiamos a tra- vés de fronteras nacionales, oceánicas y lingüísticas? ¿Cómo conciliar el potencial libertador de formas de transculturación como la hibridización o la criollización, por ejemplo, con las histoRIAS de migración y trasplante violentos y de opresión que caracterizan gran parte de la experiencia americana? ¿Qué perspectivas se abren en el futuro a una cultura americana considerada en este contexto amplio? ¿Puede una cultura americana globalizada producida por los EE.UU. funcionar como un obstáculo para la aparición de nuevas formas culturales o, por el contrario, puede faci- litar involuntariamente su promoción? ¿De qué modo los modernos medios de comunicación, de transporte o de contacto intercultural han venido transformando el signifcado de ‘América’ desde los tiempos coloniales? ¿Cuáles son las características de los lugares de resistencia a los efectos homogeneizadores de una cultura americana globalizada? Como organización interdis- ciplinar, IASA les invita a presentar comunicaciones y ‘workshops’ sobre estas cuestiones, sin excluir otras relacionadas con ellas, en el contexto de análisis de naturaleza cultural, histórica, política y teórica. Línguas ofciais do congresso: português, espanhol, francês, inglês. Línguas ofciais do congresso: português, espanhol, francês, inglês. Volume 1, Number 1 46 Lugar Lugar Facultad de Letras de la Universidad de Lisboa September 2006 47 International International American Studies Association (IASA) 3ème Congrès, 20–23 Septembre, 2007 Trans/américain, trans/océanique, trans/latif S’il est vrai que 1492 signale l’avènement de la modernité, ce congrès examinera les incidences des voyages de Christophe Colomb sur le développement de la culture et de l’identité des Amé- riques. L’échange de semences, de plantes et d’animaux, mais aussi de langues, la transplanta- tion de peuples, et notamment l’extraordinaire portée du commerce des esclaves en Afrique, puis l’arrivée de populations asiatiques et l’impact de la violence exercée contre les indigènes du Nouveau Monde, tout cela a fait des Amériques un espace d’exploration privilégié du sens de la modernité. Nous vous invitons à présenter des propositions de communications à carac- tère comparatiste par la recherche de thèmes qui entrecroisent les géographies nationales des Amériques, les espaces de contacts intercontinentaux, tels que l’Atlantique et le Pacifque, ou les langues et les traditions des Amériques. Nous tiendrons compte également de communications traitant de nationalités spécifques, y compris les USA, et apportant une contribution impor- tante à l’élucidation des efets et des ramifcations d’une modernité générée par l’exploration, la conquête, la colonisation et la globalisation. L’objectif du congrès est d’aborder, entre autres, les questions suivantes: quelle sorte d’entité est donc cette «Amérique», que nous étudions au travers de frontières nationales, océaniques et linguistiques? Comment concilier le potentiel li- bérateur de formes de transculturation, comme l’hybridisation et la créolisation, par exemple, avec les violentes histoires de transplantation, de migration et d’oppression, qui caractérisent pour une bonne part l’expérience américaine? Quelles perspectives s’ouvrent-elles à une culture américaine considérée dans ce contexte élargi? Une culture américaine globalisée, produite par les USA, peut-elle fonctionner comme un obstacle au surgissement de nouvelles formes cultu- relles ou au contraire faciliter involontairement leur promotion? Dans quelle mesure les moyens modernes de communication, de transport ou de contact interculturel ont-ils transformé le sens du mot «Amérique» depuis les temps coloniaux? Qu’est-ce qui caractérise les lieux de résistance aux efets homogénéisants d’une culture américaine globalisée? Comme organisation interdis- Volume 1, Number 1 Volume 1, Number 1 48 Langues ofcielles du congrès: français, anglais, espagnol, potugais Langues ofcielles du congrès: français, anglais, espagnol, potugais MELUSINDIA & MELOW: CALL FOR CONTRIBUTIONS MELUS (The Society for the Study of the Multi-Ethnic Literature of the United States India Chap- ter) & MELOW (The Society for the Study of the Multi-Ethnic Literatures of the World) will hold a conference on LITERATURE IN TIMES OF VIOLENCE in Chandigarh, on: 22–24 March 2007 Theme for MELUS India—Literature in Times of Violence: The American Response The Keynote Address will be delivered by Prof. Emory Elliott Distinguished Professor of English University of California Riverside, CA Theme for MELOW—Contemporary World Literatures in Times of Violence This will be the Eighth International Conference of MELUS–INDIA and the Second International Conference of MELOW. The MELOW Conference will dove-tail into the MELUS-India Conference. Calls for Papers and Contributions ciplinaire, IASA invite à présenter des communications et des ateliers sur ces problèmes, sans en exclure d’autres ayant des rapports avec eux, dans le contexte d’analyses de caractère culturel, historique, politique et théorique. Lieu du congrès Faculté des Lettres de l’Université de Lisbonne Délais Délais — Les résumés des communications (300 mots) et les propositions d’ateliers thématiques doi- vent nous parvenir jusqu’au 31 décembre 2006. — Les avis d’acceptation seront envoyés jusqu’au 28 février 2007. Review of International American Studies to the legacy of mass violence and political confict? Does the creative mind buckle under the pres- sures or does it rise above them all to create mournful music? And how does the reader respond to the various tensions that go into the making of great literature? What models are available for understanding these literary responses to the turbulence of the times? Do poetry, fction, dra- ma and flm help us fnd words and images to understand national catastrophe? Can literature narrate mass violence? Does it try to escape violence? Can it be a substitute for violence? Is it a cure or a panacea? p We are looking for papers that discuss the theme, problem, object, or practice of violence. The MELUS–India 2007 Conference will explore these and related issues, taking up diverse genres literature, cinema, theatre, media, popular culture, etc. The focus will be American Litera- ture but papers which cross borders and disciplines are encouraged. 250–word abstracts related to the theme are invited. Tentatively, the conference will be divided into (but not restricted to) the following panels: Tentatively, the conference will be divided into (but not restricted to) the following pan The American Frontier—violence and conquest — The Civil War—Edmund Wilson's Patriotic Gore — Racial Violence—Black Experience — Racial Violence—Black Experience — Domestic Violence—confessional women's poetry Vietnam Experience — Film and Fiction after 9/11 — Violence and Visual Interpretation — Violence and Hollywood Film/Reggae and hate music — Empire and EgalitarianismIssues of History, Empire and Culture in American Literature. — The MELOW 2007 Conference will explore the same issues, focusing on literatures of the world 250–word abstracts are invited related to literature of the last ffty years The MELOW 2007 Conference will explore the same issues, focusing on literatures of the world. 250–word abstracts are invited, related to literature of the last ffty years. The tentative division of panels will be: — Ethnic violence — Gender related violence — Violence of Partitions and Borders — Revolutionary Violence — Violence of Exile and Displacement — Violence and the African Experience — Relationship between boundaries and violence — The Holocaust experience — Violence of exile — For both conferences we are seeking papers that deal with issues related but not conﬁned to race, ethnicity, identity, and gender, in literary texts, ﬁlms, popular culture, media. Com- paratist and interdisciplinary perspectives are encouraged. The major theme— Out of the quarrels with ourselves, if we believe Yeats, literature is created. In fact it is not just the quarrel with ourselves but also the turbulence of the times that is responsible for the pro- duction of literature. These are times of upheaval and violence when one is assaulted physically, emotionally and psychologically from all quarters. Despite the trauma, however, one survives and carries on the best way possible. Artists continue to produce works of art, musicians create music and writers compose their masterpieces. But, one may ask, how does literature respond September 2006 September 2006 49 Calls for Papers and Contributions Main issues to be examined in the paper. — Main issues to be examined in the paper. — Note: Membership of MELUS–India / MELOW is not required at the time when abstracts are sub- mitted. Non-members may join the Association after the acceptance of their papers. Deadlines: Abstracts of Individual papers to be sent by: August 30, 2006. Acceptance of Individual ab- stracts will be dispatched by: Oct 1, 2006. For any clariﬁcation contact: Manju Jaidka (Secretary, MELUS–India, MELOW), email: mjaidka@sify.com or Anil Raina (Treasurer), email: anilraina@glide.net.in. Review of International American Studies 250 word abstracts may be emailed latest by Sept 30, 2006, (as part of the text and NOT as attachment) to mjaidka@sify. com, with a copy to anilraina@glide.net.in Members and non-members of MELUS–India / MELOW may submit abstracts (with the un- derstanding that they will attend the conference). Abstracts received will then be exam- ined by the Conference Committee. Delegates whose abstracts are accepted will be invited to submit individual papers. Abstracts may be submitted on any topic that relates to the themes of MELUS–India or MELOW but only one abstract will be accepted from an individual. All abstracts should include the fol- lowing information: Name, brief CV and contact information of the participant. — Whether the abstract is being proposed for MELUS–India or MELOW. — Title of the proposed paper. — Volume 1, Number 1 50 Puebla, Mexico April 19–22, 2007 www.acla.org Co-Chairs: Lois Parkinson Zamora, Enrique Pérez Castillo, Michael Schuessler Program Chairs: Efraín Kristal, Kathleen Komar Co-Chairs: Lois Parkinson Zamora, Enrique Pérez Castillo, Michael Schuessler Program Chairs: Efraín Kristal, Kathleen Komar Organizing Committee: Margaret Higonnet, Miguel Cabañas, Wendy Faris, Dan Russek, Oscar Fernández Organizing Committee: Margaret Higonnet, Miguel Cabañas, Wendy Faris, Dan Russek, Oscar Fernández Advisory Committee: Djelal Kadir, Silvia Spitta, Adriana Méndez Rodenas, George Handley, Da- vid Damrosch, Christopher Winks, Nancy Worman, Haun Saussey, Tobin Siebers, Peter Connor, Dan Chamberlain, Deborah Cohn Advisory Committee: Djelal Kadir, Silvia Spitta, Adriana Méndez Rodenas, George Handley, Da- vid Damrosch, Christopher Winks, Nancy Worman, Haun Saussey, Tobin Siebers, Peter Connor, Dan Chamberlain, Deborah Cohn Contact information: Lois Parkinson Zamora: lzamora@uh.edu or Michael Schuessler: mschuess@barnard.edu Contact information: ACLA  TRANS, PAN, INTER: CULTURES IN CONTACT AMERICAN COMPARATIVE LITERATURE ASSOCIATION ANNUAL MEETING CALL FOR PAPERS Puebla, Mexico April 19–22, 2007 www.acla.org Trans, Pan, Inter: Cultures in Contact We encourage papers on all comparative topics, not just those dealing with Mexico or the Amer- icas. Below we suggest subtopics but, as usual, individuals may suggest seminar topics of their choosing. Proposals for seminars are to be submitted to the ACLA website by October 1, 2006, and individual paper proposals by November 1, 2006. Crossing Borders and Boundaries of All Kinds — Mythic Subtexts, Modern Texts — Commodities and Cultures — Literary Translation: Textual and Contextual — Across Art Forms and Disciplines: Theory and Practi Comparative Cuisines — Travel Literature — Transculturation, Mestizaje, Creolization — Transatlantic and Transpacifc Encounters — Indigenous Literatures and Languages in Mexico and Beyond — Challenges of Non-Western Cultures to Critical Theory — Comparative Approaches to Literatures of the Americas — Colonialism and Classicism — September 2006 51 Review of International American Studies Caribbean Languages and Literatures Jewish Literature in the Americas Canada and Mexico: “So Far from God, So Close to the United States” Immigration and Exile in/to the Americas US Latino Literatures Latin American Studies and Inter-American Studies Mechanisms of Literary and Cultural Production Borders: US-Mexico, Mexico-Guatemala, Mexico-Belize Performance in Mexico Popular Culture and Literature Ecocritical Approaches in Comparative Context Violence and Testimonial Literature — — — — — — — — — — — — Ecocritical Approaches in Compara — Violence and Testimonial Literature — About Puebla Puebla is located 50 miles southeast of Mexico City at an altitude of 7100 feet, in a broad valley bordered on the west by the Sierra Nevada and its legendary volcanoes Popocatépetl and Iztac- cíhuatl, and on the east by the Sierra Madre Oriental and its snow-capped volcano La Malinche. Puebla preserves a strong colonial favor. Over fve hundred buildings and seventy churches, many in the Baroque style, have been beautifully restored in the city center, which was declared a UNESCO World Heritage Site in 1987. The folk Baroque fourishes in nearby villages, where churches refect the syncretic forms that resulted from the meeting of indigenous and European cultures. Important pre-contact indigenous sites, including the pyramid of Cholula and the cer- emonial center of Cacaxtla, are also nearby and open to the public. Puebla has played a role the Mexican political and economic landscape since its establishment in 1531; here, the Cinco de Mayo battle was won against the invading French in 1862. The city is famous for its cuisine, including mole poblano, chiles en nogada, chalupas poblanas, and for its beautiful Talavera ce- ramics and glazed tiles, which grace tables and embellish façades, fountains and interior patios. For basic information about Puebla, go to: http://www.virtualmex.com/puebla.htm. Culturas en Contacto — Transculturalismo, Panc Interculturalismo Congreso Internacional (2007) American Comparative Literature Association Universidad Autónoma de Puebla Instituto de Ciencias Sociales y Humanidades Puebla, México, Abril 19–22, 2007 Universidad Autónoma de Puebla Instituto de Ciencias Sociales y Humanidades Puebla, México, Abril 19–22, 2007 Coordinadores: Lois Parkinson Zamora, Enrique Pérez Castillo, Michael Schuessler Coordinadores del programa: Efraín Kristal, Kathleen Komar Comité organizador: Margaret Higonnet, Miguel Cabañas, Wendy Faris, Dan Russek, Oscar Fernández Consejo asesor: Djelal Kadir, Silvia Spitta, Adriana Méndez Rodenas, George Handley, David Damrosch, Christopher Winks, Nancy Worman, Haun Saussey, Tobin Siebers, Peter Connor, Dan Chamberlain, Deborah Cohn www.acla.org Coordinadores: Lois Parkinson Zamora, Enrique Pérez Castillo, Michael Schuessler Coordinadores del programa: Efraín Kristal, Kathleen Komar Comité organizador: Margaret Higonnet, Miguel Cabañas, Wendy Faris, Dan Russek, Oscar Fernández Consejo asesor: Djelal Kadir, Silvia Spitta, Adriana Méndez Rodenas, George Handley, David Damrosch, Christopher Winks, Nancy Worman, Haun Saussey, Tobin Siebers, Peter Connor, Dan Chamberlain, Deborah Cohn Volume 1, Number 1 52 Calls for Papers and Contributions Comité local: Guadalupe Grajales P., Silvia Kiczkovsky Y., Ma. Carmen Jiménez R., Gerardo del Rosal, Agustín Grajales P., Lorena Carrillo Comité local: Guadalupe Grajales P., Silvia Kiczkovsky Y., Ma. Carmen Jiménez R., Gerardo del Rosal, Agustín Grajales P., Lorena Carrillo Información, contactar: Lois Parkinson Zamora: lzamora@uh.edu Michael Schuessler: mschuess@barnard.edu Enrique Pérez C. : perezcen@siu.buap.mx Departamento anftrión: Posgrado en Ciencias del Lenguaje , ICSyH Culturas en Contacto Transculturalismo, Panculturalismo, Interculturalismo Se incluirán ponencias sobre todos los temas comparatistas, y de preferencia los que se relacio- nen con México y con América Hispánica, Portuguesa, Anglo y Franco parlante. Abajo se listan algunos temas sugeridos. Además, individualmente se pueden sugerir temas de seminarios y mesas redondas sobre tópicos específcos Las propuestas para seminarios deben enviarse al sitio web del comité de la ACLA antes de Octubre 1, 2006, y las propuestas y resúmenes de ponencias individuales antes de No- viembre 1, 2006. Cruzando fronteras y barreras de todas clases — Subtextos míticos, textos modernos — Artículos de consumo y culturas — Traducción: literaria, textual y contextual — y Al través de formas artísticas y disciplinas: teoría y práctica — Cocinas comparativas — Transculturación, mestizaje y criollización — Encuentros trasatlánticos y traspacífcos — Literaturas y lenguas indígenas de México y el resto de América — El desafío de las culturas no occidentales a la teoría crítica — Transacciones genéricas: constructos culturales y literarios del género, el sexo y la sexualidad — Enfoques comparatistas de las literaturas del continente americano — Enfoques comparatistas de las literaturas del continente americano Colonialismo y clasicismo — — Colonialismo y clasicismo — Lenguajes y literaturas del caribe — La literatura judía del continente americano — México y Canadá: tan lejos de Dios y tan cerca de los Estados Unidos — y j y Inmigración y exilio hacia y dentro del continente americano — Literaturas latinas de los USA Estudios latinoamericanos e interamericanos — — Estudios latinoamericanos e interamericanos — Mecanismos de producción cultural y literaria — Fronteras: México/USA; México/Guatemala; México/Belice — Representación dramática en México — Culturas y literaturas populares — Enfoques ecocríticos en un contexto comparatista — Violencia y literatura testimonial. — Posgrado en Ciencias del Lenguaje THE JOURNAL OF AMERICAN AND CANADIAN STUDIES: CALL FOR CONTRIBUTIONS The Journal of American and Canadian Studies is a peer-reviewed annual, published in bound and online form by the Institute of American and Canadian Studies, Sophia University, Tokyo, and reaches over two thousand readers throughout the world. The Journal seeks scholarly arti- cles from across the discipline(s) of North American studies. Fields of study represented in recent issues include, for example, diplomatic history and foreign policy, African American literature, media studies, and US and Canadian ethnic studies. For submission guidelines and back issues please see the Institute’s home page: http://www.info.sophia.ac.jp/amecana/E2/journal.htm. Review of International American Studies Review of International American Studies Puebla, Pue., México +52 – 2222 29 55 00 exts. 5706, 3122 Puebla, Pue., México +52 – 2222 29 55 00 exts. 5706, 3122 Posgrado en Ciencias del Lenguaje Instituto de Ciencias Sociales y Humanidades Universidad Autónoma de Puebla 2 Oriente 410, Centro September 2006 53 DE/CONSTRUCTION OF ETHNICITY AND NATIONHOOD IN THE AGE OF GLOBALIZATION: CALL FOR CONTRIBUTIONS The age of globalization has witnessed a redefnition of ethnicity as well as the (re)emergence of nationhood, while at the same time witnessing the deconstruction of both. Bi-, multi- and poly-cultural societies more than ever have to confront issues generated from such eforts, which aim at shaping and reshaping identities. While current political developments exacerbate such phenomena, pacts of forgetting and other silent alliances that had been established during the Cold War as well as post—and neo-colonial paradigms are now being interrogated and chal- lenged. Contributions are invited for a volume focusing on these issues to be published by MESEA, follow- ing a symposium organized in September 2005 in Istanbul. The articles should be 5000 to 7000 words in length and follow the MLA style of documentation. Please send abstracts of 100–200 words by 1 October 2006. Complete essays are due 31 January 2007. S d b Gönül Pultar Kültür Arastirmalari Dernegi Inebolu sokak 15/4 Setustu/Kabatas 34427 Istanbul, Turkey phone/fax: + 90 212 292 22 29 e-mail: gpultar@kulturad.org Gönül Pultar Kültür Arastirmalari Dernegi Inebolu sokak 15/4 Setustu/Kabatas 34427 Istanbul, Turkey phone/fax: + 90 212 292 22 29 e-mail: gpultar@kulturad.org 54 Volume 1, Number 1 Volume 1, Number 1
https://openalex.org/W2990183307	https://www.nature.com/articles/s41598-019-54369-7.pdf	English	null	Endoscopic score vs blood cell indices for determining timing of immunomodulator withdrawal in quiescent ulcerative colitis	Scientific reports	2,019	cc-by	7,082	Endoscopic score vs blood cell indices for determining timing of immunomodulator withdrawal in quiescent ulcerative colitis Kazuhiro Takenaka1,3, Keiichi Tominaga 1,3, Mimari Kanazawa1, Koh Fukushi1, Takanao Tanaka1, Akira Kanamori1,2, Takeshi Sugaya1, Kouhei Tsuchida1, Makoto Iijima1, Kenichi Goda1 & Atsushi Irisawa1 While immunomodulators (IMs) are used as key drugs in remission maintenance treatment for ulcerative colitis (UC), there has been no evidence to date for determining monitoring methods and drug withdrawal. Therefore, we examined if a decrease in white blood cell count (WBC) and an elevation in mean cell volume (MCV) could be used as optimization indices and if mucosal healing (MH) could be a rationale for determining the time of IM withdrawal. Subjects were 89 UC patients who were using IMs and for whom clinical remission had been maintained. Those with a Rachmilewitz Clinical Activity Index score of 4 or lower and those with a Mayo endoscopic subscore (MES) of 0 or 1 were defined as MH. The remission maintenance rates of the following comparative groups were examined: an IM continuation group and an IM withdrawal group; an IM continuation group with a WBC of less than 3000 or a MCV of 100 or greater and an IM continuation group with a WBC of 3000 or greater and a MCV of 99 or lower; an IM continuation group of patients for whom MH had been achieved and an IM continuation group of patients for whom MH had not been achieved; and an IM withdrawal group with a MES of 0 and an IM withdrawal group with a MES of 1. A significantly higher remission maintenance rate was observed in the IM continuation group compared to the withdrawal group (p < 0.01). No significant difference was observed between the IM continuation group with a WBC of less than 3000 or a MCV of 100 or greater and the IM continuation group with a WBC of 3000 or greater and a MCV of 99 or lower (p = 0.08). Higher remission maintenance rates were observed in the IM continuation group of patients for whom MH had been achieved compared to the IM continuation group of patients for whom MH had not been achieved (p = 0.03). No significant difference was observed between the IM withdrawal group with MES 0 and the IM withdrawal group with MES 1. (p = 0.48). www.nature.com/scientificreports www.nature.com/scientificreports www.nature.com/scientificreports www.nature.com/scientificreports/ www.nature.com/scientificreports/ Mean age (range) 44.0 (14–81) Female/Male 42/47 Extent of lesion (pancolitis/left-sided colitis/proctitis) 66/23/- Disease type (relapse-remitting/chronic continuous/first attack) 83/9/- Duration of IM treatment (weeks) 48.7 MES (0, 1, 2) 37/42/10 Mean WBC (/µL) 5250 Mean MCV 93.6 IM continued/IM withdrawn 56/33 Table 1. Characteristics of patients eligible for analysis. Table 1. Characteristics of patients eligible for analysis. Table 1. Characteristics of patients eligible for analysis. maintenance rate that exceeds 50% and a large number of reports have stated that IMs are significantly effica- cious at the time of steroid dose-reduction and remission maintenance18–20. Meanwhile, the optimal dosage of IM varies greatly between individual patients due to variation in metabolic disposition and when IMs are adminis- tered. Attention needs to be paid to genetic hair loss, severe myelosuppression, dose- and metabolism-dependent hepatic dysfunction, nausea and fatigue, lymphoproliferative disease resulting from long-term administration, and nonmelanoma skin cancer (NMSC)21–24. An onset of malignant disease as a result of the long-term use of IMs is a serious issue, hence, it is necessary to monitor the condition of UC and consider IM withdrawal where nec- essary. However, it has been reported that relapse rates following IM withdrawal are high, and there is a dilemma between the advantage of withdrawal and the risk of relapse25,26. maintenance rate that exceeds 50% and a large number of reports have stated that IMs are significantly effica- cious at the time of steroid dose-reduction and remission maintenance18–20. Meanwhile, the optimal dosage of IM varies greatly between individual patients due to variation in metabolic disposition and when IMs are adminis- tered. Attention needs to be paid to genetic hair loss, severe myelosuppression, dose- and metabolism-dependent hepatic dysfunction, nausea and fatigue, lymphoproliferative disease resulting from long-term administration, and nonmelanoma skin cancer (NMSC)21–24. An onset of malignant disease as a result of the long-term use of IMs is a serious issue, hence, it is necessary to monitor the condition of UC and consider IM withdrawal where nec- essary. However, it has been reported that relapse rates following IM withdrawal are high, and there is a dilemma between the advantage of withdrawal and the risk of relapse25,26. g p With regard to rheumatoid arthritis, the “treat-to-target” strategy and the concept of “tight control and dis- ease monitoring” have long been recognized; they have also been increasingly accepted in the area of IBD. Results P i ’ Patient’s characteristics. The backgrounds of 89 subjects are shown in Table 1. The mean age was 44 years (14–18 years), and there were 47 males (52.8%) and 42 females (47.2%). Of these, 25.8% (n = 23) were suffering from left-sided UC and 74.2% (n = 66) were suffering from total UC. An endoscopic assessment observed a Mayo endoscopic subscore (MES) of 0 in 37 cases (41.6%), a MES of 1 in 42 cases (47.2%), and a MES of 2 in 10 cases (11.2%). The number of IM withdrawal cases was 33 (37.1%). The inter-rater agreement on endoscopic findings. Inter-rater agreement (multi-rater kappa stat- ics) for the endoscopic findings (MES) was performed. Regarding the agreement rate of endoscopic evaluation between the 2 endoscopists, a high Cohen’s kappa coefficient (κ = 0.75) was obtained, indicating virtually com- plete agreement. Remission maintenance in patient with IM continued or IM withdrawal. Firstly, we comparatively evaluated the rate of remission maintenance between the IM continued and IM withdrawal groups. Although there was no significant difference in the background of the patients between the groups (Table 2), the rate of remission maintenance was higher in the IM continued group than in the withdrawal group (P < 0.01) (Fig. 1). In addition, in patients where IM was continued, there was no significant difference in the rate of remission mainte- nance between the adjusted and non-adjusted patient groups (Table 3, Fig. 2). Remission maintenance in the view point of endoscopic MH. Within the IM continuation group that included 56 cases, a comparative analysis of remission maintenance rates was performed between the MH group (MES 0, 1) and the non-MH group (MES 2). While no significant difference was observed between the 2 groups in terms of patient backgrounds (Table 3), remission maintenance rates in the MH group were sig- nificantly higher than those in the non-MH group (p < 0.01) (Fig. 3). In other words, the analysis found that once MH is achieved, it is possible to maintain remission at a satisfactory level by continuing IMs; however, the continuation of IMs gives rise to problematic adverse events. Considering this viewpoint, we subsequently evaluated remission maintenance after withdrawal of IM in patients where MH had been achieved. www.nature.com/scientificreports/ To date, there have been reports on an investigation into the selection of treatment methods based on endoscopic monitoring for a case of Crohn’s disease (CD) requiring surgical treatment, and an investigation into the dose reduction of 5-aminosalicylic acid (5-ASA) preparations due to endoscopic mucosal healing (MH). Many of these reports suggested that achieving MH is the goal of IBD treatment and that the degree of MH can be an index for selecting a treatment method12,27–29. MH is the obvious optimal target for IBD treatment, and there are many studies on treat to target to achieve the treatment goal28,30,31. On the other hand, there are no randomized controlled trial on treatment withdrawal or drug dose adjustment based on MH. Therefore, as the research gap, in UC cases that achieved MH remains a clinical problem as to whether IMs withdrawal is possible or not. MH may also be an index for IM withdrawal in UC treatment. However, there has been no investigation into whether MH can be an index for IM withdrawal. The aim of the present study was to examine if MH, which is currently considered the goal of UC treatment, can be a rationale for IM withdrawal in UC cases where long-term remis- sion has been achieved. Endoscopic score vs blood cell indices for determining timing of immunomodulator withdrawal in quiescent ulcerative colitis This retrospective study showed that remission maintenance could be firmly obtained by continuing IM administration in case of endoscopic MH; however, MH was not an indicator of IM withdrawal. Ulcerative colitis (UC) is a chronic inflammatory bowel disease (IBD) of unknown etiology. The disease presents with symptoms such as abdominal pain, bloody stool, diarrhea, and weight loss and is characterized by repeated relapse and remission. Because no fundamental therapeutic strategies have been established, the goal of treatment is generally to maintain the remission phase for a long period after remission induction therapy1,2. However, remission maintenance is often difficult. Recently, achievement of mucosal healing (MH) has been recognized as an important factor for remission maintenance, and shown to yield a low relapse rate, reduce hospitalization, result in a low rate of conversion to surgical treatment, reduce healthcare costs, and improve the quality of life of patients3–12. Immunomodulators (IMs), primarily azathioprine, are recommended by guidelines in various countries for the treatment of UC patients who are steroid-dependent or experiencing difficulty with steroid withdrawal13–17. It has been found that the continued oral administration of IMs following remission may lead to a high remission 1Department of Gastroenterology, Dokkyo Medical University, Tochigi, Japan. 2Department of Internal Medicine, Japanese Red Cross Ashikaga Hospital, Tochigi, Japan. 3These authors contributed equally: Kazuhiro Takenaka and Keiichi Tominaga. email: tominaga@dokkyomed.ac.jp Scientific Reports \| (2019) 9:17751 \| https://doi.org/10.1038/s41598-019-54369-7 Results P i ’ In the IM withdrawal group that included 33 cases, 31 had IM withdrawn after MH was confirmed by endoscopy; however, remission maintenance rates in these cases were clearly lower than those of the IM continuation group. Moreover, Scientific Reports \| (2019) 9:17751 \| https://doi.org/10.1038/s41598-019-54369-7 www.nature.com/scientificreports/ IM continued IM withdrawn P value n = 56 n = 33 Mean age 44.0 43.9 NS Female/Male 30/26 17/16 NS Extent of lesion (pancolitis/left-sided colitis/proctitis 45/11/- 21/12/- NS Disease type (relapse-remitting/chronic continuous/first attack) 50/6/- 33/3/- NS Duration of IM treatment (weeks) 57.4 40.2 NS MES (0, 1, 2) 22/26/8 15/16/2 NS Mean WBC (/µL) 5270 5230 NS Mean MCV 93.4 93.7 NS Table 2. Comparison of patient characteristics between the IM continued and IM withdrawn groups. Table 2. Comparison of patient characteristics between the IM continued and IM withdrawn groups. Adjust Non-Adjust P value MH non-MH P value n = 41 n = 15 n = 48 n = 8 Mean age 47.6 47.3 NS 43.9 43.9 NS Female/Male 19/23 7/7 NS 23/25 3/5 NS Extent of lesion (pancolitis/left-sided colitis/proctitis) 32/9/- 13/2/- NS 38/9/- 7/2/- NS Disease type (relapse-remitting/chronic continuous/first attack) 37/4/- 13/2/- NS 42/6 8/-/- NS Duration of IM treatment 48.0 48.8 NS 44.9 45.2 NS Mean WBC (/µL) — — — 5270 5400 NS Mean MCV — — — 93.4 93.8 NS MES (0, 1, 2) 15/20/6 7/6/2 NS — — — Table 3. Comparison of patient characteristics in IM continuous group. Table 3. Comparison of patient characteristics in IM continuous group. Table 3. Comparison of patient characteristics in IM continuous group. Table 3. Comparison of patient characteristics in IM continuous group. Figure 1. Comparison between patients in whom IMs were withdrawn and continued groups revealed a significant difference in remission maintenance rates (p < 0.01), which was the secondary endpoint. Figure 1. Comparison between patients in whom IMs were withdrawn and continued groups revealed a significant difference in remission maintenance rates (p < 0.01), which was the secondary endpoint. a comparative analysis within the IM withdrawal group by pre-withdrawal MES (MES 0, MES 1) observed no significant difference in remission maintenance rates between the groups (p = 0.80) (Table 4, Fig. 4). Discussion Discussion Discussion In treating UC cases with repeated relapses and remissions, IMs as well as 5-ASAs play a significant role in main- taining remission. Specifically, IMs are suitable for treatment of steroid-dependent UC cases where maintenance has not been achieved through 5-ASA monotherapy20. IMs used to treat UC include thiopurine agents and meth- otrexates, and guidelines in Japan, the US, and Europe recommend the use of these drugs when indicated14–18. In treating UC cases with repeated relapses and remissions, IMs as well as 5-ASAs play a significant role in main- taining remission. Specifically, IMs are suitable for treatment of steroid-dependent UC cases where maintenance has not been achieved through 5-ASA monotherapy20. IMs used to treat UC include thiopurine agents and meth- otrexates, and guidelines in Japan, the US, and Europe recommend the use of these drugs when indicated14–18. Scientific Reports \| (2019) 9:17751 \| https://doi.org/10.1038/s41598-019-54369-7 www.nature.com/scientificreports/ Figure 2. Comparison between the IM Adjust and Non-Adjust groups in patients with continued use of IMs revealed a significant difference in remission maintenance rates (p = 0.08), which was the secondary endpoint. Figure 2. Comparison between the IM Adjust and Non-Adjust groups in patients with continued use of IMs if Figure 2. Comparison between the IM Adjust and Non-Adjust groups in patients with continued use of IM revealed a significant difference in remission maintenance rates (p = 0.08), which was the secondary endpoin Figure 3. Comparison between the MH and non-MH groups in patients with continued use of IMs revealed a significant difference in remission maintenance rates (p = 0.01), which was the secondary endpoint. Figure 3. Comparison between the MH and non-MH groups in patients with continued use of IMs revealed a significant difference in remission maintenance rates (p = 0.01), which was the secondary endpoint. MES 0 MES 1 P value n = 15 n = 16 Mean age 43.9 43.8 NS Female/Male 8/7 7/9 NS Extent of lesion (pancolitis/left-sided colitis/proctitis) 11/7/- 8/5/- NS Disease type (relapse-remitting/chronic continuous/first attack) 15/-/- 16/-/- NS Duration of IM treatment (weeks) 41.9 40.5 NS Mean WBC (/µL) 5200 5280 NS Mean MCV 93.5 93.3 NS Table 4. Comparison of patient characteristics between the MES 0 and MES 1 groups. Table 4. Comparison of patient characteristics between the MES 0 and MES 1 groups. Discussion Moreno-Rincón et al.26 reported that the relapse rate within 1 year of IM withdrawal in UC cases where clinical remission had been maintained was 18.88%, and that the rate within 4 years was 40.67%. Furthermore, Fraser AG et al.38 reported that in UC cases where clinical remission had been observed, the proportion of patients for whom remission could be maintained following IM withdrawal was 25.2%. In other words, it was suggested that clinical remission may not constitute a rationale for IM withdrawal. Meanwhile, due to the development of biologics and the advancement of endoscopic devises in recent years, the goal of UC treatment is shifting from clinical remission to MH. In fact, a large number of studies have reported that achieving MH leads to a decrease in relapse rates and the likelihood of subsequent surgical treatment3,7,8,29,39. In this way, the clinical significance of IMs is widely recognized; however, no investigation into IM withdrawal based on MH has been conducted. The present study is the first report on IM withdrawal based on MH.h h p yi p In recent years, the MES has been widely used in MH evaluation. The present investigation also employed the MES in assessing MH. In the present investigation, even though IMs were withdrawn following the achievement of a MES of 0–1, the results showed that remission maintenance rates within 3 and 5 years of withdrawal were low. These results suggested that MH findings can be referred to in determining if IMs should be withdrawn or not. We consider that if the disease activity is so high in a UC patient that they are administered IMs, i.e. a UC patient presenting with steroid-dependency, the initial level of disease activity is high. Hence, IMs will be required for remission maintenance. This may explain why the disease easily relapses following IM withdrawal. We have pre- viously reported that the dosage of 5-ASAs can be reduced if MH is achieved (particularly in cases where the MES is 0). That said, while it is possible to manage the administration of 5-ASAs in anticipation of dose reduction, in other words, while it is possible to perform the maintenance of administration of low-dose 5-ASAs, in managing IM administration, withdrawal instead of dose reduction is anticipated. Discussion MES 0 MES 1 P value n = 15 n = 16 Mean age 43.9 43.8 NS Female/Male 8/7 7/9 NS Extent of lesion (pancolitis/left-sided colitis/proctitis) 11/7/- 8/5/- NS Disease type (relapse-remitting/chronic continuous/first attack) 15/-/- 16/-/- NS Duration of IM treatment (weeks) 41.9 40.5 NS Mean WBC (/µL) 5200 5280 NS Mean MCV 93.5 93.3 NS Table 4. Comparison of patient characteristics between the MES 0 and MES 1 groups. Table 4. Comparison of patient characteristics between the MES 0 and MES 1 groups. Scientific Reports \| (2019) 9:17751 \| https://doi.org/10.1038/s41598-019-54369-7 www.nature.com/scientificreports/ Figure 4. Comparison between the MES 0 and MES 1 groups among patients in whom IMs were withdrawn revealed no significant difference in remission maintenance rates (p = 0.8), which was the primary endpoint. Figure 4. Comparison between the MES 0 and MES 1 groups among patients in whom IMs were withdrawn revealed no significant difference in remission maintenance rates (p = 0.8), which was the primary endpoint. Unfortunately, methotrexates are not funded through the Japanese national health insurance scheme; therefore, only thiopurine agents are used. However, a large number of reports have shown the efficacy of methotrexates32,33. Additionally, it has been suggested that, apart from their use for maintaining remission in UC and CD cases, the concomitant administration of IMs with infliximab suppresses the immunogenicity of infliximab34,35. l pp g yl On the other hand, there are concerns that the long-term administration of IMs may cause myelotoxicity, liver toxicity, infections, and malignancies (lymphoproliferative disorders or NMSC)22–24,36,37. According to a pro- spective observational cohort study from France, the incidence rates of Hodgkin’s lymphoma and non-Hodgkin lymphoproliferative disorder were 0.90 per 1000 patient-years in IM receiving, 0.20 per 1000 patient-years in those who had discontinued, and 0.26 per 1000 patient-years in those who had never received IM. The multivariate-adjusted hazard ratio of lymphoproliferative disorder between patients receiving IM and those who had never received the drugs was 5.28. Median follow-up period of this study was 35 months24. In order to avoid these IM-induced adverse events, it is necessary to consider withdrawing IM treatment; however, IM withdrawal may exacerbate UC symptoms. In fact, there have been a number of reports which show high relapse rates in UC cases following IM withdrawal22–24. www.nature.com/scientificreports/ not considered to be a criterion for withdrawal of IMs. It can be considered that examining relapses following IM withdrawal from a perspective of whether histological healing has been achieved or not may provide different results. Further investigations are required in this aspect. g q p Meanwhile, remission maintenance rates in the IM continuation group were higher than those in the IM with- drawal group. As this result is consistent with existing reports, it is suggested that the patient population of the present study was receiving optimal and appropriate treatment with IMs. We consider that the data underpin the accuracy of the results for the primary endpoint of the present study and are important. Also, while the present investigation suggested that achieving endoscopic MH may lead to sufficient remission maintenance as long as IMs are continued, we consider that this indicates that simply using IMs will not lead to remission maintenance. In other words, it is necessary to appropriately use IMs in patients who have continued to be orally adminis- tered IMs, by assessing treatment effects through regular endoscopic observations and the monitoring of CRP, calprotectin, and occult blood. In the present study, the relapse rate in the group of patients for whom IMs were adjusted by white blood cell count (WBC) and mean cell volume (MCV) was 0%. It is generally recommended that IM treatment is adjusted by measuring 6-thioguanine nucleotide and thiopurine methyltransferase activities. However, this approach is not practical in the actual clinic for the following reasons: there is deviation in values due to the measurement methods; measurement values are not readily available; and these measurements are not funded through the Japanese national health insurance scheme13,44. The results of the present study suggested that determining the applicability of IMs based on WBC and MCV is effective.hi g pp yf The limitations of the present study are: firstly, it was a single-center retrospective study; secondly, the num- ber of cases was small; and thirdly, centralized reading was not adopted for endoscopic observation, and there is discrepancy between assessors to a certain extent. However, since the Cohen’s kappa coefficient for inter-rater agreement in the present study was high at 0.75, we considered that the level of objectiveness was maintained to a certain extent in relation to the above-mentioned third limitation. Methods St d d Study design. The present study is a retrospective cohort study conducted in 2 medical institutions and approved by the ethics committee of each institution, Dokkyo Medical University Hospital and Japanese Red Cross Ashikaga Hospital (approval no. R-22-3J). This study was conducted in accordance with the ethical prin- ciples associated with the Declaration of Helsinki and registered in the University Hospital Medical Network Clinical Trials Registry [UMIN000036506]. The option to opt-out of the study was communicated to the patients via our website with the following message. “Dokkyo Medical University Hospital and Japanese Red Cross Ashikaga Hospital now conducting research using medical data from patients who treated for ulcerative colitis. There will be no additional effort on patients for conducting this study. In addition, we will conduct research in compliance with laws and regulations regarding the protection of patient privacy. If you do not want your medical data to be used in this study, please contact your doctor.”h The primary endpoint was the remission maintenance rate following IM withdrawal indicated by a MES of 0. Secondary endpoints were remission maintenance rates through continued IM administration, remission main- tenance rates in an IM continuation group where MH had been achieved, and remission maintenance rates in an IM continuation group where adjustments were made. Patients. To select eligible patients, we retrospectively reviewed the medical records of 283 UC patients aged between 14 and 81 years who were treated at Dokkyo Medical University Hospital and Japanese Red Cross Ashikaga Hospital between April 2010 and March 2018. Of the 283 cases, a case where remission had not been achieved within 1 year of the oral administration of IMs and a case with a history of anti-TNFα antibody agent administration were excluded. Moreover, in selecting patients, the following definitions were adopted: clinical remission was defined as a Rachmilewitz Clinical Activity Index score of 4 or lower; MH was defined as MES 0 or 1; and the IM Adjust group was to consist of patients with WBC > 3000 (µL) or MCV ≤ 10045. www.nature.com/scientificreports/ It would have been ideal if the present study had been a prospective randomized control trial (RCT); however, the study involved drug withdrawal, and it was difficult to conduct the study as a prospective RCT. Given the limitations of the present study, we believe that the results provided important insights for developing UC treatment strategies going forward.i p p g p g g g g In conclusion, this retrospective study showed that remission maintenance could be firmly obtained by con- tinuing IM administration in case of endoscopic MH; however, MH was not an indicator of IM withdrawal. Discussion Therefore, it is considered that the asso- ciation of 5-ASAs with a MES of 0 (the finding that a MES of 0 may be used as an index for the dose reduction of 5-ASAs) is different from the results of the present study. Additionally, while it has been reported that there are differences between MES 0 and MES 1 in terms of remission maintenance rates and surgery rates2,40–42, we found that it is also important to not only perform an endoscopic mucosal assessment but also to confirm histological healing, in order to ensure remission maintenance43. As the research gap in UC treatment, we studied the clinical issue of IM withdrawal was possible or not in UC cases that achieved MH. However, UC cases that achieve MH or even if MES 0, withdrawal of IMs was made relapse easily. This result indicates that UC cases using IMs are difficult to maintain remission after IMs withdrawal even if MH was achieved. Thus, the achievement of MH is Scientific Reports \| (2019) 9:17751 \| https://doi.org/10.1038/s41598-019-54369-7 www.nature.com/scientificreports/ Scientific Reports \| (2019) 9:17751 \| https://doi.org/10.1038/s41598-019-54369-7 Methods St d d j g p p Regarding administration of 5-ASA, the following 5-ASA preparations were used: a time-dependent ASA preparation (Pentasa®, 2000–4000 mg/d; Kyorin Pharmaceutical Co., Ltd., Tokyo, Japan), a pH-dependent ASA preparation (Asacol®, 2400–3600 mg/d; Zeria Pharmaceutical Co., Ltd., Tokyo, Japan), a pH-dependent MMX® ASA (Lialda®, 2400–4800 mg/d; Mochida Pharmaceutical Co., Ltd., Tokyo, Japan), and salazosulfapyridine (Salazopyrin®, 2000–4000 mg/d; Pfizer Japan Inc., Tokyo, Japan), which are approved in Japan. Regarding IM preparations, azathioprine (Imuran®, 25–75 mg/d; Aspen Japan K.K., Tokyo, Japan) was administered. Endoscopic evaluation. According to the Montreal Classification of UC, the colon was divided into the following 3 segments46,47: (1) Ulcerative proctitis (E1) (Proctitis type): involvement limited to the rectum (i.e., proximal extent of inflammation is distal to the rectosigmoid junction). (2) Left-Sided UC (E2) (Left-sided type): involvement limited to the portion of the colorectum distal to the splenic flexure. (3) Extensive UC (E3) (Pancolitis type): involvement extends proximal to the splenic flexure. yl Endoscopic evaluation was based on the MES29,45, where a MES of 0 (no friability, granularity, and intact vascular pattern) corresponded to normal mucosa and a MES of 1 (mild erythema or decreased vascular pattern) corresponded to healed mucosa. As previously mentioned, MH was defined as a MES of 0 or 1. In addition, a MES of 2 (marked erythema, absent vascular pattern, friability, and erosions) and a MES of 3 (spontaneous bleeding and ulceration) were regarded to correspond to mucosa in the active phase (Fig. 5). Endoscopic findings obtained Scientific Reports \| (2019) 9:17751 \| https://doi.org/10.1038/s41598-019-54369-7 www.nature.com/scientificreports/ www.nature.com/scientificreports/ Figure 5. Endoscopic image of Mayo endoscopic subscore (MES). (A) Endoscopic image of MES 0 (no friability and granularity and intact vascular pattern). (B) Endoscopic image of MES 1 (mild erythema or decreased vascular pattern). (C) Endoscopic image of MES 2 (marked erythema, absent vascular pattern, friability, and erosions). (D) Endoscopic image of MES 3 (spontaneous bleeding and ulceration). Figure 5. Endoscopic image of Mayo endoscopic subscore (MES). (A) Endoscopic image of MES 0 (no friability and granularity and intact vascular pattern). (B) Endoscopic image of MES 1 (mild erythema o decreased vascular pattern). (C) Endoscopic image of MES 2 (marked erythema, absent vascular pattern friability, and erosions). (D) Endoscopic image of MES 3 (spontaneous bleeding and ulceration). Figure 5. Endoscopic image of Mayo endoscopic subscore (MES). (A) Endoscopic image of MES 0 (no friability and granularity and intact vascular pattern). Methods St d d (B) Endoscopic image of MES 1 (mild erythema or decreased vascular pattern). (C) Endoscopic image of MES 2 (marked erythema, absent vascular pattern, friability, and erosions). (D) Endoscopic image of MES 3 (spontaneous bleeding and ulceration). after at least 1 year of clinical remission were evaluated by 2 endoscopists. Each was selected from the 2 medical institutions from those with 10 or more years’ experience and who specialized in IBD, annually treating 200 or more patients with IBD. Endoscopic findings on which the 2 endoscopists agreed were included in analyses. When they disagreed on the evaluation of the findings, cases were scored by assigning 1 point to each of the fol- lowing 3 items: presence of redness, visible vascular pattern, and fine mucosal granularity. When the mean scares calculated by the 2 endoscopists was 2 or higher, patients were determined to have a MES score of 1. To evaluate endoscopic findings, the scores of the colon segments with the most severe inflammation were used. Statistical analyses. Statistical analyses were performed with IBM SPSS Statistics 24® (IBM Japan, Ltd., Tokyo, Japan). The Cohen’s kappa coefficient (κ) was calculated to determine the agreement rate between the 2 endoscopists who evaluated the endoscopic findings. The Pearson χ2 test was performed to compare sex, affected area, endoscopic classification, histological classification, and smoking rate. When the expected value was <5, the Fisher exact test was performed. The Mann-Whitney U test was performed to compare mean age and mean disease duration. To compare the remission maintenance rate, survival curves were generated using the Kaplan-Meier method and the log-rank test was performed. The Cox proportional hazards model was used to identify predictors of clinical relapse. P < 0.05 indicated statistical significance. Received: 28 May 2019; Accepted: 13 November 2019; Published: xx xx xxxx Data availabilityh y The database used for statistical analysis that provided data used to support the findings of this study are restricted by the Hospital Ethical Board in order to protect patient privacy. Data are available for researchers who meet the criteria for access to confidential data. More information is available from Keiichi Tominaga, MD, tominaga@ dokkyomed.ac.jp. Received: 28 May 2019; Accepted: 13 November 2019; Published: xx xx xxxx Received: 28 May 2019; Accepted: 13 November 2019; Published: xx xx xxxx Received: 28 May 2019; Accepted: 13 November 2019; Published: xx xx xxxx Scientific Reports \| (2019) 9:17751 \| https://doi.org/10.1038/s41598-019-54369-7 www.nature.com/scientificreports/ References Nonmelanoma Skin Cancer Risk in Patients With Inflamma Thiopurine Therapy: A Systematic Review of the Literature. Dermatol Surg 44, 469–480 (2018).t p pl y y 22. Hagen, J. W. & Pugliano-Mauro, M. A. Nonmelanoma Skin Cancer Risk in Patients With Inflammatory Bowel Disease Undergoing Thiopurine Therapy: A Systematic Review of the Literature. Dermatol Surg 44, 469–480 (2018).t 22. Hagen, J. W. & Pugliano-Mauro, M. A. Nonmelanoma Skin Cancer Risk in Patients With Inflammatory Bowel Disease Un Thiopurine Therapy: A Systematic Review of the Literature. Dermatol Surg 44, 469–480 (2018).t 23. Khan, N., Abbas, A. M., Lichtenstein, G. R., Loftus, E. V. Jr. & Bazzano, L. A. Risk of lymphoma in patients with ulcerative colitis treated with thiopurines: a nationwide retrospective cohort study. Gastroenterology 145, 1007–1015 (2013).l 4. Beaugerie, L. et al. Lymphoproliferative disorders in patients receiving thiopurines for inflammatory bowel disease: a prospective observational cohort study. Lancet 374, 1617–1625 (2009). y 5. Torres, J. et al. Systematic Review of Effects of Withdrawal of Immunomodulators or Biologic Agents From Patients With Inflammatory Bowel Disease. Gastroenterology 149, 1716–1730 (2015).th nflammatory Bowel Disease. Gastroenterology 149, 1716–1730 (20 l y gy 6. Moreno-Rincón, E. et al. Prognosis of Patients with Ulcerative Colitis in Sustained Remission After Thiopurines Withdrawal Inflamm Bowel Dis 21, 1564–1571 (2015). fl 27. Torres, J., Mehandru, S., Colombel, J. F. & Peyrin-Biroulet, L. Crohn’s disease. Lancet 389, 1741–1755 (2017). C l C h ’ d ft l d d l 28. De Cruz, P. et al. Crohn’s disease management after intestinal re t 29. Takahashi, F. et al. Timing for dose-down of 5-ASA depends on mucosal status with ulcerative colitis. Scand J Gastroenterol 51, 827–834 (2016). 0. Ungaro, R., Colombel, J. F., Lissoos, T. & Peyrin-Biroulet, L. A Treat-to-Target Update in Ulcerative Colitis: A Systematic Review. Am J Gastroenterol 114, 874–883 (2019). 31. Bouguen, G., Levesque, B. G., Pola, S., Evans, E. & Sandborn, W. J. Feasibility of endoscopic assessment and treating to target to achieve mucosal healing in ulcerative colitis. Inflamm Bowel Dis 20, 231–239 (2014). l 2. Miyake, N. et al. Azathioprine is essential following cyclosporine for patients with steroid-refractory ulcerative colitis. World Gastroenterol 21, 254–261 (2015).f 3. Hibi, T., Naganuma, M., Kitahora, T., Kinjyo, F. & Shimoyama, T. Low-dose azathioprine is effective and safe for maintenance o remission in patients with ulcerative colitis. J Gastroenterol 38, 740–746 (2003).l 34. Panaccione, R. et al. Scientific Reports \| (2019) 9:17751 \| https://doi.org/10.1038/s41598-019-54369-7 References Combination therapy with infliximab and azathioprine is superior to monotherapy with either agent in ulcerative colitis. Gastroenterology 146, 392–400 (2014).l l ulcerative colitis. Gastroenterology 146, 392–400 (2014).l gy 35. Colombel, J. F. et al. Infliximab, azathioprine, or combination therapy for Crohn’s disease. N Engl J Med 362, 1383–1395 (2010 36 Ch M t l S f t f thi i th i i fl t b l di l t f ll t d f 3931 ti t I f 35. Colombel, J. F. et al. Infliximab, azathioprine, or combination therapy for Crohns disease. N Engl J Med 362, 1383 1395 (2010). 36. Chaparro, M. et al. Safety of thiopurine therapy in inflammatory bowel disease: long-term follow-up study of 3931 patients. Inflamm Bowel Dis 19, 1404–1410 (2013).l 7. Kirchgesner, J. et al. Risk of Serious and Opportunistic Infections Associated With Treatment of Inflammatory Bowel Diseases Gastroenterology 155, 337–346 (2018).hfil gy 8. Fraser, A. G., Orchard, T. R. & Jewell, D. P. The efficacy of azathioprine for the treatment of inflammatory bowel disease: a 30 year review. Gut 50, 485–489 (2002). 39. Parente, F. et al. Are colonoscopy and bowel ultrasound useful for assessing response to short-term therapy and predicting disease outcome of moderate-to-severe forms of ulcerative colitis?: a prospective study. Am J Gastroenterol 105, 1150–1157 (2010). 40. Manginot, C., Baumann, C. & Peyrin-Biroulet, L. An endoscopic Mayo score of 0 is associated with a lower risk of colectomy than a score of 1 in ulcerative colitis. Gut 64, 1181–1182 (2015). 1. Barreiro-de Acosta, M. et al. Evaluation of the Risk of Relapse in Ulcerative Colitis According to the Degree of Mucosal Healing (Mayo 0 vs 1): A Longitudinal Cohort Study. J Crohns Colitis 10, 13–19 (2016).ih y g y 42. Travis, S. P. et al. Review article: defining remission in ulcerative colitis. Aliment Pharmacol Ther 34, 113–124 (2011). 43. Kanazawa, M. et al. Relationship between endoscopic mucosal healing and histologic inflammation during remission maintenance phase in ulcerative colitis: A retrospective study. Endosc Int Open 7, E568–E575 (2019). y g y 42. Travis, S. P. et al. Review article: defining remission in ulcerative colitis. Aliment Pharmacol Ther 34, 113–124 (2011). 43 K M l R l i hi b d i l h li d hi l i i fl i d i i i i y g y 42. Travis, S. P. et al. References g J ( ) 3. Matsuoka, K. et al. Evidence-based clinical practice guidelines for inflammatory bowel disease. J Gastroenterol 53, 305–353 (2018) p gl y 14. Magro, F. et al. Third European Evidence-based Consensus on Diagnosis and Management of Ulcerative Colitis. Part 1: Definitions, Diagnosis, Extra-intestinal Manifestations, Pregnancy, Cancer Surveillance, Surgery, and Ileo-anal Pouch Disorders. J Crohns Colitis 11, 649–670 (2017).l gh p g gi Diagnosis, Extra-intestinal Manifestations, Pregnancy, Cancer Surveillance, Surgery, and Ileo-anal Pouch Disorders. J Crohns Colitis 11, 649–670 (2017).l 15. Gionchetti, P. et al. Use of corticosteroids and immunosuppressive drugs in inflammatory bowel disease: Clinical practice guidelines of the Italian Group for the Study of Inflammatory Bowel Disease. Dig Liver Dis 49, 604–617 (2017).h l 6. Steinhart, A. H. & Fernandes, A. Clinical practice guidelines for the medical management of nonhospitalized ulcerative colitis: The patient perspective. Can J Gastroenterol Hepatol 29, 294–296 (2015). patient perspective. Can J Gastroenterol Hepatol 29, 294–296 (2015) p 7. Bressler, B. et al. Clinical practice guidelines for the medical management of nonhospitalized ulcerative colitis: the Toronto consensus. Gastroenterology 148, 1035–1058 (2015).fi 18. Gisbert, J. P., Linares, P. M., McNicholl, A. G., Maté, J. & Gomollón, F. Meta-analysis: the efficacy of azathioprine and mercaptopurine in ulcerative colitis. Aliment Pharmacol Ther 30, 126–137 (2009). h ( ) 9. Mantzanis, G. J. et al. A prospective randomized observer-blind 2-year trial of azathioprine monotherapy versus azathioprine and olsalazine for the maintenance of remission of steroid- dependent ulcerative colitis. Am J Gastroenterol 99, 1122–1128 (2004). date in medical therapy of ulcerative colitis. J Clin Gastroenterol 34, 20. Katz, S. Update in medical therapy of ulcerative colitis. J Clin Gastroenterol 34, 397–407 (2002). . Katz, S. Update in medical therapy of ulcerative colitis. J Clin Gas p py 21. Kakuta, Y. et al. NUDT15 codon 139 is the best pharmacogenetic marker for predicting thiopurine-induced severe adverse in Japanese patients with inflammatory bowel disease: a multicenter study. J Gastroenterol 53, 1065–1078 (2018).l y 21. Kakuta, Y. et al. NUDT15 codon 139 is the best pharmacogenetic marker for predicting thiopurine-induced severe in Japanese patients with inflammatory bowel disease: a multicenter study J Gastroenterol 53 1065 1078 (2018) p g p g p in Japanese patients with inflammatory bowel disease: a multicenter study. J Gastroenterol 53, 1065–1078 (2018).l J p pl y y 22. Hagen, J. W. & Pugliano-Mauro, M. A. References References 1. Kornbluth, A. & Sachar, D. B. Practice Parameters Committee of the American College of Gastroenterology. Ulcerative colitis practice guidelines in adults: American College of Gastroenterology, Practice Parameters Committee. Am J Gastroenterol 105, 501–523 (2010). 1. Kornbluth, A. & Sachar, D. B. Practice Parameters Committee of the American College of Gastroenterology. Ulcerative colitis practice guidelines in adults: American College of Gastroenterology, Practice Parameters Committee. Am J Gastroenterol 105, 501–523 (2010). ( ) 2. D’Haens, G. et al. A review of activity indices and efficacy end points for clinical trials of medical therapy in adults with ulcerative colitis. Gastroenterology 132, 763–786 (2007).l 2. D’Haens, G. et al. A review of activity indices and efficacy end points for clinical trials of medical therapy in adults with ulcerative colitis. Gastroenterology 132, 763–786 (2007).l gy 3. Frøslie, K. F., Jahnsen, J., Moum, B. A. & Vatn, M. H. IBSEN Group. Mucosal healing in inflammatory bowel disease: results from a Norwegian population-based cohort. Gastroenterology 133, 412–422 (2007). g p p gy 4. Truelove, S. C. & Richards, W. C. Biopsy studies in ulcerative colitis. Br Med J 1, 1315–1318 (1956).l Neurath, M. F. & Travis, S. P. Mucosal healing in inflammatory bow 6. Colombel, J. F. et al. Early mucosal healing with infliximab is associated with improved longterm clinical outcomes in ulcerative colitis. Gastroenterology 141, 1194–1201 (2011).i gy 7. Meucci, G. et al. Prognostic significance of endoscopic remission in patients with active ulcerative colitis treated with oral and topical mesalazine: a prospective, multicenter study. Inflamm Bowel Dis 18, 1006–1010 (2012).ti l 8. Ardizzone, S. et al. Mucosal healing predicts late outcomes after the first course of corticosteroids for newly diagnosed ulcerative colitis. Clin Gastroenterol Hepatol 9, 483–489 (2011). p 9. Gibson, P. R. et al. Relationship between disease severity and quality of life and assessment of health care utilization and cost for ulcerative colitis in Australia: a cross-sectional, observational study. J Crohns Colitis 8, 598–606 (2014).tl 0. Sandborn, W. J. et al. Colectomy rate comparison after treatment of ulcerative colitis with placebo or infliximab. Gastroenterology 137, 1250–1260 (2009).hfl 11. Feagan, B. G. et al. The effects of infliximab therapy on health-related quality of life in ulcerative colitis patients. Am J Gastroenterol 102, 794–802 (2007).hlh ( ) 12. Peyrin-Biroulet, L. et al. Selecting Therapeutic Targets in Inflammatory Bowel Disease (STRIDE): Determining Therapeutic Goals for Treat-to-Target. Am J Gastroenterol 110, 1324–1338 (2015). Competing interestsh p g The authors declare no competing interests. References Review article: defining remission in ulcerative colitis. Aliment Pharmacol Ther 34, 113–124 (2011). 43. Kanazawa, M. et al. Relationship between endoscopic mucosal healing and histologic inflammation during remission maintenance avis, S. P. et al. Review article: defining remission in ulcerative coliti www.nature.com/scientificreports/ 4. Simsek, M., Meijer, B., Mulder, C. J. J., van Bodegraven, A. A. & de Boer, N. K. H. Analytical Pitfalls of Therapeutic Drug Monitoring of Thiopurines in Patients With Inflammatory Bowel Disease. Ther Drug Monit 39, 584–588 (2017). 44. Simsek, M., Meijer, B., Mulder, C. J. J., van Bodegraven, A. A. & de Boer, N. K. H. Analytical Pitfalls of Therapeutic Drug Monitoring of Thiopurines in Patients With Inflammatory Bowel Disease. Ther Drug Monit 39, 584–588 (2017). h l d l ( l l d) l h l h f l l hlh 5. Rachmilewitz, D. Coated mesalazine (5-aminosalicylic acid) versus sulphasalazine in the treatment of active ulcerative colitis: a randomised trial. BMJ 298, 82–86 (1989). a do sed t a . J 98, 8 86 ( 989). 46. Schroeder, K. W., Tremaine, W. J. & Ilstrup, D. M. Coated oral 5-aminosalicylic acid therapy for mildly to moderately active ulcerative colitis. A randomized study. N Engl J Med 317, 1625–1629 (1987).fi 6. Schroeder, K. W., Tremaine, W. J. & Ilstrup, D. M. Coated oral 5-aminosalicylic acid therapy for mildly to moderately active ulcerative colitis. A randomized study. N Engl J Med 317, 1625–1629 (1987).fi y g 7. Scherl, E. J. et al. Safety and efficacy of a new 3.3 g b.i.d. tablet formulation in patients with mild-to-moderately-active ulcerative colitis: a multicenter, randomized, double-blind, placebo-controlled study. Am J Gastroenterol 104, 1452–1459 (2009). 47. Scherl, E. J. et al. Safety and efficacy of a new 3.3 g b.i.d. tablet formulation in patients with mild-to-moderately-active ulcerative colitis: a multicenter, randomized, double-blind, placebo-controlled study. Am J Gastroenterol 104, 1452–1459 (2009). Author contributions K. Takenaka and K. Tominaga contributed equally to this work. K. Tominaga (corresponding author) designed this study. K. Takenaka, M.K., K.F., T.T., A.K., T.S. and K. Tsuchida analyzed the data. M.I. and K.G. helped to interpret the data. A.I. reviewed this manuscript. All authors approved the final version of the manuscript for submission. Additional information Correspondence and requests for materials should be addressed to K.T. Correspondence and requests for materials should be addressed to K.T. Correspondence and requests for materials should be addressed to K.T. Reprints and permissions information is available at www.nature.com/reprints. Reprints and permissions information is available at www.nature.com/reprints. Reprints and permissions information is available at www.nature.com/reprints. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not per- mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Scientific Reports \| (2019) 9:17751 \| https://doi.org/10.1038/s41598-019-54369-7
https://openalex.org/W2735654485	https://journals.assaf.org.za/index.php/per/article/download/2833/2748	English	null	Learned Staatsrecht from the Heartland of the Rechtsstaat Observations on the Significance of South African-German Interaction in Constitutional Scholarship ?	PER	2,017	cc-by	11,669	∗ Background paper presented at a conference Public Law Themes in South Africa and Germany organised by South African scholars in Public Law under the auspices of the Alexander von Humboldt Foundation’s regional “Humboldt-Kolleg” programme and held at the Stellenbosch Institute for Advanced Study, Stellenbosch from 8-10 September 2005. 2 Lindemans 2002 http://www.pantheon.org/articles/v/valhalla.html 21 Nov: “Valhalla, Hall of the Slain, in Norse mythology is the hall presided over by Odin. This vast hall has five hundred and forty doors. The rafters are spears, the hall is roofed with shields and breast-plates litter the benches. A wolf guards the western door and an eagle hovers over it. It is here that the Valkyries, Odin's messengers and spirits of war, bring half of the heroes that died on the battle fields (the rest go to Freya's hall Folkvang). These heroes, the Einherjar, are prepared in Valhalla for the oncoming battle of Ragnarok. When the battle commences, eight hundred warriors will march shoulder to shoulder out of each door.” ISSN 1727-3781 2005 VOLUME 8 No 1 ISSN 1727-3781 2005 VOLUME 8 No 1 2005 VOLUME 8 No 1 L DU PLESSIS PER/PELJ 2005(8)1 y p ∗∗ Professor of Public Law. University of Stellenbosch. 1 Especially in (substantive) criminal law and the various disciplines of (substantive) private law. y ∗∗ Professor of Public Law. University of Stellenbosch. Observations on the Significance of South African-German Interaction in Constitutional Scholarship∗ L du Plessis∗∗ ; ( ) ; 5 For a helpful overview, relating to the Humboldt contribution in particular, cf Rabie, Van Merwe and Labuschagne 1993 THRHR 608. g 6 One of these scholars, John Dugard, thought that consistent reliance on the Roman-Dutch (common) law could have boosted the safeguarding of basic human rights in apartheid South Africa and that many of the ills of apartheid were attributable to a lack of rigorous reliance on this “rights-friendly” source of South African law: Dugard 1971 SALJ 181 and g ; g 4 For nuanced and insightful discussions of this influence cf Van der Walt 1996 TSAR 521; Van der Walt 1996 TSAR (4) 626; Van der Walt 1992 THRHR 170 4 For nuanced and insightful discussions of this influence cf Van der Walt 1 521; Van der Walt 1996 TSAR (4) 626; Van der Walt 1992 THRHR 170. 3 Cf in this regard Du Plessis Introduction to Law 57-63; Fagan "Roman-Dutch Law" 60-64 g ; g For nuanced and insightful discussions of this influence cf Van der Walt 1996 TSAR (3) ; ( ) ; 5 For a helpful overview, relating to the Humboldt contribution in particular, cf Ra M d L b h 1993 THRHR 608 uanced and insightful discussions of this influence cf Van der Walt 199 Van der Walt 1996 TSAR (4) 626; Van der Walt 1992 THRHR 170. 5 For a helpful overview, relating to the Humboldt contribution in particular, cf R Merwe and Labuschagne 1993 THRHR 608. I During (especially the latter half of) the previous century it was impressed on several generations of law students (mainly but not exclusively) at Afrikaans speaking law faculties in South Africa, to pride themselves on their “principled” legal education.1 Akin to (and indeed associated with) the paranormal knack of “thinking/reasoning like a jurist”, principled legal thinking was not really taught (or learnt), but sustained (like injuries) as a result of exposure to principled law teachers, enhanced by the ambiance of a principle-prone law faculty. In the impressionable, young minds thus shaped Begriffsjurisprudenz was principled legal thinking incarnate, and Germany the Valhalla2 for those forever true to it. 77/150 PER/PELJ 2005(8)1 L DU PLESSIS L DU PLESSIS The so-called purist movement in South Africa,3 whose heyday more or less overlapped with that of the latter-day apartheid regime (1948-1994), bore the torch of principled legal scholarship and jurisprudence. Its adherents preached and promoted – in class but in time also in courtrooms – loyalty to pure Roman- Dutch law, untainted by English legal influence and unperverted by English minded judges’ (mis-)understanding of it. Like the legal systems of the Romano-Germanic or civil-law legal family, Roman-Dutch law – the purists’ source of and guide to principled legal thinking – is of learned Roman-law extraction. This accounts for the purists’ heartfelt empathy with the (German) historical school and, eventually, with the nineteenth century Pandectists who strongly influenced key-facets of the private-law theory taught at Afrikaans speaking law faculties in South Africa.4 Traditional South African-German fellow-feelings in law remained restricted to private law (excluding formal private law), criminal law (excluding criminal procedure), legal history, Roman Law as an academic discipline and perhaps legal philosophy and legal theory too (with international law eventually also creeping into the picture).5 Conspicuously underrepresented on the affinity list, for a long time, was the public law relevant for us at this conference, namely constitutional and administrative law which, in the “old” South Africa, was much more English than Roman-Dutch – as was criminal and civil law of procedure and various branches of commercial law. Constitutional and political reform was at any rate not a foremost concern of prominent purist legal scholars in South Africa. 8 Among which an article by De Waal 1995 SAJHR 1 is foremost. I The handful of pioneers concerned with such reform felt quite comfortable to seek comparative guidance in the constitutional law of the United States of America – the oldest example of a system of modern-day constitutional democracy.6 78/150 L DU PLESSIS PER/PELJ 2005(8)1 In the lively debates foreshadowing, accompanying and assessing the first tottering steps of constitutional democracy in South Africa during the 1990s, legal and constitutional comparison were dominant, and among the jurisdictions sourced for comparative examples Germany (and Canada) suddenly moved to the forefront. A number of South African scholars writing about various aspects of our transition to democracy, professed indebtedness to German sources, but not necessarily because these sources breathe the spirit of principledness.7 For the past ten years or so constitutional scholars have not (re-)turned to “the German example” as source of comparative information with an intensity and enthusiasm paralleling that of the mid-nineties. Most comparative studies referring to Germany8 actually saw the light when South Africa’s transitional or interim Constitution9 was still in force.10 The major object of this conference is to explore with (and for the benefit of) South African scholars, eligible for Von Humboldt stipends, possibilities for postdoctoral research in constitutional and administrative law in Germany. In this background paper I intend pursuing this object in mainly three ways (and not necessarily strictly in the sequence below): • First, to explore the relevance of the German tradition of scholarship for South African legal scholars doing research on matters constitutional. • Secondly, to reflect on possible reasons for strong South African-German affinities in matters constitutional. • Secondly, to reflect on possible reasons for strong South African-German affinities in matters constitutional. Dugard Human Rights and the South African Legal Order 393 397. Dugard was educated at one of the (erstwhile?) bastions of “principled legal thinking” in South Africa, namely the Faculty of Law at the University of Stellenbosch. Dugard Human Rights and the South African Legal Order 393 397. Dugard was educated at one of the (erstwhile?) bastions of “principled legal thinking” in South Africa, namely the y y 7 Francois Venter’s attempt to adapt nineteenth century pandectism’s “principled” theoretical framework for (private-law) subjective rights to use in public law, was the exception and not the rule (cf Venter Publiekregtelike Verhouding) which attracted criticism; cf Van der Vyver "Doctrine of Private-law Rights" 208-209; Wiechers "Publieke Subjektiewe Reg" 270-291. 9 Constitution of the Republic of South Africa 200 of 1993. y y 7 Francois Venter’s attempt to adapt nineteenth century pandectism’s “principled” theoretical framework for (private-law) subjective rights to use in public law, was the exception and not the rule (cf Venter Publiekregtelike Verhouding) which attracted criticism; cf Van der Vyver "Doctrine of Private-law Rights" 208-209; Wiechers "Publieke Subjektiewe Reg" 270-291. 9 Constitution of the Republic of South Africa 200 of 1993. 10 There are, however, examples of more recently dated monographs broadly dealing with comparative issues and in which “the German example” receives its rightful attention; cf eg De Wet Constitutional Enforceability of Economic and Social Rights; Van der Walt Constitutional Property Clauses 121-163; Venter Constitutional Comparison. Dugard Human Rights and the South African Legal Order 393 397. Dugard was educated at one of the (erstwhile?) bastions of “principled legal thinking” in South Africa, namely the Faculty of Law at the University of Stellenbosch. I 8 Among which an article by De Waal 1995 SAJHR 1 is foremost. g y 9 Constitution of the Republic of South Africa 200 of 1993. 10 There are, however, examples of more recently dated monographs broadly dealing with comparative issues and in which “the German example” receives its rightful attention; cf eg De Wet Constitutional Enforceability of Economic and Social Rights; Van der Walt Constitutional Property Clauses 121-163; Venter Constitutional Comparison. 79/150 L DU PLESSIS PER/PELJ 2005(8)1 • Thirdly, to reflect on the consequences of the affinities above-said as they have manifested themselves in South Africa’s two constitutional texts since 1994 and in our constitutional scholarship and jurisprudence. Here I shall mainly devote attention to some developments since the commencement of South Africa’s “final Constitution”11 in 1997. However, at pains to preface rather than pre-empt deliberations on particular issues and themes, this third aspect of the paper will be restricted to bare essentials. As far as the first two aspects of the paper are concerned, I intend sharing a moderate measure of personal experience too, and this will inevitably put an anecdotal spin on my presentation. I mention this principally to forewarn myself that a successful anecdote is one not centred on the person of the anecdotist, but on the illustrative value of the narrative involved. In case I am a failure as an anecdotist let me offer all those who are age-wise still eligible for a Humboldt stipend sound (anecdotal) advice right at the outset: Do not let the opportunity to apply for a Von Humboldt stipend go by default! Not following this advice was a rather regrettable mistake in my own academic career. I have never counted myself among those South African legal scholars revering the conventional version of principled German (or, for that matter, South African) legal thinking. I have always found the legal positivism inherent in such thinking uninteresting, uninspiring and unimaginative. At the same time I have been attracted by – and, indeed, biased towards – the German intellectual tradition in general, especially in legal philosophy and legal theory, my default interests. 80/150 11 The Constitution of the Republic of South Africa 108 of 1996 (hereinafter referred to as the 1996 Constitution). 12 Like Wolf Griechisches Rechtsdenken and Wolf Das Problem der Naturrechtslehre in respect of the first topic and Baur Gott, Recht und Weltliches Regiment im Werke Calvins in respect of the second. 13 Cf eg Du Plessis 1984 THRHR 127 13 Cf eg Du Plessis 1984 THRHR 127. 14 Cf eg Davis, Chaskalson and De Waal "Democracy and Constitutionalism" 1-130. 81/150 13 Cf eg Du Plessis 1984 THRHR 127. 14 Cf eg Davis, Chaskalson and De Waal "Democracy and Constitutionalism" 1-130. respect of the first topic and Baur Gott, Recht und Weltliches Regiment im Werke Calvi in respect of the second. 13 Cf D Pl i 1984 THRHR 127 I German writings on philosophical and legal philosophical issues of interest to me, have mostly left me with a decidedly favourable impression of the quality of German scholarship, and I thank my understanding of topics as diverse as the evolution of early Greek natural law and the legal and political thinking of the church reformer, Jean Calvin, mostly to German sources.12 In 80/150 PER/PELJ 2005(8)1 L DU PLESSIS time (and in a growingly activist vein) some of the work of the Frankfurter Schule (the “Neo-Marxists” – as we used to call them) also shaped aspects of my philosophical thinking beyond repair.13 During the early 1990s the expectation of a new constitutional era in South Africa whetted my appetite for German constitutional law (and constitutional interpretation in particular) largely because of my previously positive encounters with “German learning”. At that time some of the locally published work contemplating a constitutionally reborn South Africa, devoted considerable attention to post-World War II German expertise and experience.14 In the process of actual constitution-making – both during the multi-party negotiations preceding the adoption of the transitional Constitution in 1993 and, later, in the Constitutional Assembly where the final (1996) Constitution took shape – a “German presence” (eventually manifesting itself in both end products) was tangible. There is no single explanation for the consequential German influence on constitution-making and the inception of constitutional democracy in South Africa. As suggested before conventional academic affinities in law do not quite explain this phenomenon (but are also not wholly unrelated to it). Mentally many white Afrikaans speaking South Africans in apartheid South Africa identified with the “Germanic” traits in the German Volksgeist. For a long time German was the third language of preference in many an Afrikaans medium, secondary school. Germanist fellow-feelings among some Afrikaners even engendered concurrence in the Nazi notion of “an Arian Herrnvolk”, remarkably pulling itself up by its bootstraps after a devastating war – against demonic forces that many Afrikaners, at any rate, regarded as hostile to their cause too. I can recall witnessing, even in the mid-1970s, earnest holocaust denial by well- educated, intelligent, Afrikaans speaking colleagues in law. It was, however, not 81/150 L DU PLESSIS PER/PELJ 2005(8)1 these ill-placed, pro-German sentiments that brought the markedly German element to the constitutional negotiating table in South Africa. g 16 The position just sketched, is as it was until 1995 (more or less also the time when the academic isolation ended); cf Rabie, Van der Merwe and Labuschagne 1993 THRHR 610- 613. 15 Rabie, Van der Merwe and Labuschagne 1993 THRHR 622. 82/150 613. 17 Rabie, Van der Merwe and Labuschagne 1993 THRHR 622. 82/150 17 Rabie, Van der Merwe and Labuschagne 1993 THRHR 622. 15 Rabie, Van der Merwe and Labuschagne 1993 THRHR 622. I As apartheid South Africa faced increasing academic isolation, most German institutions involved in academic exchange (universities as well as funding organisations) continued “to provide and keep open channels for comparative research to South African jurists”15 irrespective of race, colour, gender or creed as well as, controversially so, political and institutional affiliation. In the case of Humboldt stipends in law, almost only white, male academics, some approaching and others negotiating midlife, were advantaged by this approach, the only exception being a white female academic in 1986.16 The statistics for other German organisations, supporting academic exchange in law financially, probably did not differ significantly. According to the authors of an article appreciative of the Humboldt Foundation’s contribution to the development of the South African legal system and legal literature (hereinafter “the Humboldt article”) free access to the German academic world, amid increasing academic isolation elsewhere, resulted in “many South African jurists” relying “less heavily on Anglo-American law for comparative analysis” as well as a markedly “positive influence by German law on South African law reform”.17 This last claim will be tested shortly. However, it can safely be assumed that the jurists above-said could not, by themselves, have paved the way for a marked German impact on constitutional developments in South Africa, simply because they lacked the political clout to determine the outcome of decisions during multi-part negotiations (in 1993) and in the Constitutional Assembly (between 1994 and 1996). The “mighty” ANC relied heavily on the input of their German Berater, Hans- Peter Schneider, Director of the Deutsches Institut für Föderalismusforschung e.V. in Hannover, and Professor of Law at the University of Hannover until 82/150 613. 17 Rabie, Van der Merwe and Labuschagne 1993 THRHR 622. PER/PELJ 2005(8)1 L DU PLESSIS 2002.18 Schneider had also been (and I presume still is) a principal adviser to the SPD, the social-democratic ANC’s kindred soul in Germany. The Democratic Party (as it then was19) found a congenial spirited adviser in Ingo von Münch, formerly Professor of law at the University of Hamburg, and active politician in the German FDP. Quite significantly the predominantly black ANC and the more English oriented DP relied more intensely on the input of German advisers than the Afrikaans oriented NP or Freedom Front. Among the South African technical advisers closely involved in constitution- making during the 1990s, there were at least five Humboldt stipendiaries. 18 Cf eg the contributions of Du Plessis "South Africa’s peaceful Revolution" 209-224 and Steyler "Constitutional Promise of Decentralization in Practice" 225-240 in which they reminisce about issues in respect of which Schneider rendered valuable advice. 19 Presently the Democratic Alliance (DA). 20 On 11 September 1995 the former president of South Africa, Nelson Mandela, in his address at a state banquet in honour of former German chancellor, Helmut Kohl, then officially visiting South Africa, for instance said the following: reminisce about issues in respect of which Schneider rendered valua 19 Presently the Democratic Alliance (DA). I Working in concert with German law advisers these South Africans were indeed favourably placed to bring on the “positive influence” of German constitutional law in the law reform in South Africa attendant on constitution-making. It is difficult to determine empirically to what extent this indeed happened, but circumstantial evidence points to the likelihood that German (constitutional) law indeed influenced law reform in South Africa in the manner envisaged by the authors of the Humboldt article. II Predominantly logistic reasons for South African-German interaction in the field of constitutional law and practice have been considered so far. However, what destiny had in store for both countries and both nations, especially during the latter half of the twentieth century, also served to forge an affectionate sense of understanding between them (vast dissimilarities notwithstanding). Germany is an example of a relatively young (post-World War II) democracy from whose experience we in a new South Africa in the making stood to learn a lot.20 The 83/150 PER/PELJ 2005(8)1 L DU PLESSIS constitutions of the vast majority of African states as well as, for instance, the Canadian Charter of Rights and Freedoms of course also date from the post- World War II era. However, the German and South African experiences (put next to each other) stand out because of the matchless sense of urgency that, in both cases, permeated constitution-making and the establishment of constitutional democracy. The “Nicht wieder!” from post–Holocaust Germany reverberated (and indeed inspired action) in post-apartheid South Africa. In both countries far-reaching reconciliation and extensive nation-(re-)building simply had (and still has) to succeed to turn calamities of a shady past into accomplishments of a sunny future. The last step of constitutional significance in the German transition to full democracy, the Wiedervereingung, preceded the first step in the South African transition, the commencement of the transitional Constitution and the first democratic elections, by a mere three and a half years, and both moves were bolstered – or occasioned, some might say – by the same historical event(s): the decline of Communist hegemony in the East Block. South Africans and Germans can furthermore, without compromising the modesty becoming citizens of countries that could (still) have been in tatters, celebrate appreciable achievements along the road of constitutionalism so far – achievements that add to the worthwhileness of Germany and South Africa as examples for constitutional comparison. “We know that the challenges facing South Africa today are in many ways comparable to those that faced Germany after the Second World War. In as much as we benefited immensely from the support of the German people in the struggle against apartheid, we can learn much by drawing on your valuable experiences in reconstruction and development. “We know that the challenges facing South Africa today are in many ways comparable to those that faced Germany after the Second World War. In as much as we benefited immensely from the support of the German people in the struggle against apartheid, we can learn much by drawing on your valuable experiences in reconstruction and development. The German constitution has, for example, become a popular reference point for South African experts. We believe we can learn much from Germany in the field of combating crime. Above all, history has placed economic reconstruction and national reconciliation at the centre of the challenges faced by both countries.” (see ANC 1995 http://www.anc.org.za/ancdocs/history/mandela/1995/sp950911.html 21 Nov). p The German constitution has, for example, become a popular reference point for South African experts. We believe we can learn much from Germany in the field of combating crime. Above all, history has placed economic reconstruction and national reconciliation at the centre of the challenges faced by both countries.” (see ANC 1995 http://www.anc.org.za/ancdocs/history/mandela/1995/sp950911.html 21 Nov). 85/150 21 For a succinct depiction of this style of thinking, cf Kommers Constitutional Jurisprudence 40-41. 22 Kommers Constitutional Jurisprudence 41. 21 For a succinct depiction of this style of thinking, cf Kommers Constitutional Jurisprudence 40-41. III From experience I can say that in the realm of public law scholarship in Germany the phenomenon going by the name “typical German” is rather 84/150 PER/PELJ 2005(8)1 L DU PLESSIS elusive – especially the “typical German intellectual/professor”. Typicalness in German legal scholarship is (typically) associated with the “principled legal thinking” of Begriffsjuriprudenz,21 perceiving the law as a self-contained, rational system of general norms. Legal problems are solved when, through deductive reasoning, a concrete situation is subsumed under a norm appropriate to the exigencies of that type of situation. The state is the source of law and law, in its turn, allegedly rests on an independent foundation of reason and logic. Courts are autonomous institutions that apply the law in a systematic (even mechanistic) way as if it were a system of fixed (and predictable) rules. The “is” and the “ought” of law are markedly distinct, as are “law” and “morality” as well as “law” and “politics”. German public lawyers working with the Basic Law can hardly afford to buy into the conventional paradigm just described. The Basic Law perceives fundamental rights as anterior to the state and “the state’s law” as subject to the objective order of values enshrined in the Basic Law. Law and morality (and law and politics) can therefore not “neatly” be separated. However, to quote Donald Kommers,22 “the approach to judicial reasoning in Begriffsjurisprudenz has outlasted positivism and has had a lasting influence throughout Europe, including Germany . . . German constitutional scholars no less than the justices of the Federal Constitutional Court have made significant attempts to build a theory of judicial decision based on reason and logic.” I think this explains the resoluteness with which even the freest thinkers among the German colleagues in public law I have encountered, can proffer answers to certain questions of law as if the law as it stands is (and can be) uncontentious. The more tentative manner in which South African scholars typically respond when interrogated about the law as it stands, evidences that we are partially made of common law stuff – always ready to defer to any last word that a judicial authority might (for the time-being) have. g y y g 24 Labuschagne 1983 THRHR 422 and Labuschagne 2004 THRHR 43 46. 23 Von Savigny System des heutigen Römischen Rechts 206-262. 25 But also in the civil law tradition in Canada; cf eg Côté The Interpretation of Legislation in Canada 193-350. g g 27 Forsthoff Zur Problematik der Verfassungsauslegung 39-40; Hesse Grundzüge des Verfassungsrechts der Bundesrepublik Deutschland 21; Kommers Constitutional Jurisprudence 42-43; Müller and Christensen Grundlagen Öffentliches Recht 269-297; and Müller 1999 Stell LR 275-276. Brugger 1994 Archiv des Öffentlichen Rechts 1-34 redefines the four methods or techniques of interpretation in a creative manner so as to adapt them to his understanding of the modern-day exigencies of (constitutional) interpretation. 26 Cliteur Inleiding in het Recht 196-202; Labuschagne 2004 THRHR 43 46. 29 Currie and De Waal Bill of Rights Handbook 145-156 rely, for purposes of bill-of-rights interpretation, on an interpretive scheme akin to that of Von Savigny System des heutigen Römischen Rechts 206-262. See also Du Plessis 1998 Acta Juridica 13-16. p 28 Hahlo and Kahn South African Legal System 180. 86/150 23 Von Savigny System des heutigen Römischen Rechts 206-262. 24 Labuschagne 1983 THRHR 422 and Labuschagne 2004 THRHR 43 46. 25 But also in the civil law tradition in Canada; cf eg Côté The Interpretation of Legislation in Canada 193-350. 26 Cliteur Inleiding in het Recht 196-202; Labuschagne 2004 THRHR 43 46. 27 Forsthoff Zur Problematik der Verfassungsauslegung 39-40; Hesse Grundzüge des Verfassungsrechts der Bundesrepublik Deutschland 21; Kommers Constitutional Jurisprudence 42-43; Müller and Christensen Grundlagen Öffentliches Recht 269-297; and Müller 1999 Stell LR 275-276. Brugger 1994 Archiv des Öffentlichen Rechts 1-34 redefines the four methods or techniques of interpretation in a creative manner so as to adapt them to his understanding of the modern-day exigencies of (constitutional) interpretation. 28 Hahlo and Kahn South African Legal System 180. 29 Currie and De Waal Bill of Rights Handbook 145-156 rely, for purposes of bill-of-rights interpretation, on an interpretive scheme akin to that of Von Savigny System des heutigen Römischen Rechts 206-262. See also Du Plessis 1998 Acta Juridica 13-16. 30 Labuschagne 2004 THRHR 43 46. 31 For a summary see Du Plessis 1998 Acta Juridica 8. 23 Von Savigny System des heutigen Römischen Rechts 206 262. 24 Labuschagne 1983 THRHR 422 and Labuschagne 2004 THRHR 43 46. g y 35 My unpublished report on the project was written in Afrikaans: Wetsuitleg in Suid-Afrika die Lig van die Aanvaarding van ‘n Menseregtehandves. ( ) p 33 As manifestation of interpretive purposivism, see Du Plessis (Re-)Interpretation of Statut 115-119. 32 Du Plessis (Re-)Interpretation of Statutes 111-115. L DU PLESSIS L DU PLESSIS PER/PELJ 2005(8)1 • systematic interpretation, as a manifestation of contextualism,32 calling for an understanding of a specific provision in the light of the text or instrument as a whole and of extra-textual indicia; • purposive interpretation33 that sheds light on the possible meanings of a provision with reference to its purpose or ratio, and • historical interpretation situating a provision in the tradition from which it emerged and allowing qualified recourse to information concerning the genesis of the text in which the provision occurs (and concerning the provision itself).34 In short, I think Savigny’s four “methods of interpretation” have found acceptance in German constitutional interpretation because (and simply because) their “reason and logic” (and “pedigree”, one might perhaps add) appeal to constitutional scholars and judges (as jurists). In short, I think Savigny’s four “methods of interpretation” have found acceptance in German constitutional interpretation because (and simply because) their “reason and logic” (and “pedigree”, one might perhaps add) appeal to constitutional scholars and judges (as jurists). 34 In the course of time a fifth “method” was added to the Von Savigny quartet, name comparative interpretation, which facilitates the understanding of a provision, first, in t light of standards of international law and, secondly, in comparison with its counterparts other national legal systems. 87/150 32 Du Plessis (Re-)Interpretation of Statutes 111-115. 33 As manifestation of interpretive purposivism, see Du Plessis (Re-)Interpretation of Statutes 115-119. 34 In the course of time a fifth “method” was added to the Von Savigny quartet, namely comparative interpretation, which facilitates the understanding of a provision, first, in the light of standards of international law and, secondly, in comparison with its counterparts in other national legal systems. 35 My unpublished report on the project was written in Afrikaans: Wetsuitleg in Suid-Afrika in die Lig van die Aanvaarding van ‘n Menseregtehandves. III 85/150 PER/PELJ 2005(8)1 L DU PLESSIS L DU PLESSIS The observations of Kommers also go some way to explain a phenomenon that took me by some surprise when I first came across and reflected on it, and that has ever since intrigued me – (also) because German constitutional scholars take it so much for granted that it is hard to find anyone who cares to explain (let alone justify) it. I refer here to the unquestioned use of FC Von Savigny’s23 four “methods of interpretation” for purposes of constitutional interpretation. Also known as the “Von Savigny quartet”24 these so-called methods were initially designed for the interpretation of pandectaerian Roman law, but in time they have gained acceptance all over the European Continent,25 for the interpretation of codifications of the law, statutes26 and constitutions.27 In the South African context Hahlo and Kahn28 first referred to them and since 1994 they have also met with (extra-judicial) approval.29 These methods or modes of interpretation or, more appropriately, reading strategies – Labuschagne30 speaks of angles of incidence (“invalshoeke”) – modelled on a slightly adapted version of the Savignian model,31 are: • grammatical interpretation concentrating on ways in which the conventions of natural language can assist legal interpretation and can help to limit the many possible meanings of a provision; g g 27 Forsthoff Zur Problematik der Verfassungsauslegung 39-40; Hesse Grundzüge des Verfassungsrechts der Bundesrepublik Deutschland 21; Kommers Constitutional Jurisprudence 42-43; Müller and Christensen Grundlagen Öffentliches Recht 269-297; and Müller 1999 Stell LR 275-276. Brugger 1994 Archiv des Öffentlichen Rechts 1-34 redefines the four methods or techniques of interpretation in a creative manner so as to adapt them to his understanding of the modern-day exigencies of (constitutional) interpretation. p 28 Hahlo and Kahn South African Legal System 180. g y 29 Currie and De Waal Bill of Rights Handbook 145-156 rely, for purposes of bill-of-rights interpretation, on an interpretive scheme akin to that of Von Savigny System des heutigen Römischen Rechts 206-262. See also Du Plessis 1998 Acta Juridica 13-16. 30 Labuschagne 2004 THRHR 43 46. 31 For a summary see Du Plessis 1998 Acta Juridica 8. 31 For a summary see Du Plessis 1998 Acta Juridica 8. 86/150 IV IV Constitutional-law scholarship in Germany is vast, has a long history and accommodates – as is to be expected – diverse and divergent shades and styles of thinking. Endeavouring to discern a mainstream is contentious. Constitutional-law scholarship in Germany is vast, has a long history and accommodates – as is to be expected – diverse and divergent shades and styles of thinking. Endeavouring to discern a mainstream is contentious. By way of example, fairly representative I trust, I wish to give a brief account of my impressions and experiences when I first spent time in Germany some ten years ago, doing comparative research as part of a project entitled Statutory Interpretation in South Africa in the Light of the Adoption of a Bill of Rights.35 My aim was to complete a literature study, reading and collecting materials, making them accessible for later use and compiling a list of available literature By way of example, fairly representative I trust, I wish to give a brief account of my impressions and experiences when I first spent time in Germany some ten years ago, doing comparative research as part of a project entitled Statutory Interpretation in South Africa in the Light of the Adoption of a Bill of Rights.35 My aim was to complete a literature study, reading and collecting materials, making them accessible for later use and compiling a list of available literature 87/150 PER/PELJ 2005(8)1 L DU PLESSIS on my topic. Buzzwords for my study were “constitutional interpretation” (Ver- fassungsinterpretation/-auslegung) and “statutory interpretation” (Gesetzes- interpretation/-auslegung). Normally one would start reading about one’s topic(s) in inclusive works in the field (Staats-/Verfassungsrecht) which, in the German context, means beginning with constitutional law textbooks (Lehrbücher)36 and then, for more detail, moving to Basic Law commentaries.37 For my particular topic(s) it was important also to consult some standard works on legal methodology (Rechtsmethodenlehre).38 As one reads through these general sources, unignorable publications pertinent to more specific aspects of one’s research topic start surfacing – often in the form of collective works in which “the best on the topic” has been included,39 but it can also be in the form of journal articles of restricted scope but lasting effect. 36 Examples are Benda, Maihofer and Vogel (eds) Handbuch des Verfassungsrechts; Von Münch Staatsrecht; Hesse Grundzüge des Verfassungsrechts der Bundesrepublik Deutschland; Arnim Staatslehre der Bundesrepublik Deutschland; Maunz and Zippelius Deutsches Staatsrecht; Stein Staatsrecht. 37 Eg Sachs (ed) Grundgesetz Kommentar; Stern Das Staatsrecht der Bundes-republik Deutschland; Isensee and Kirchhof (eds) Handbuch des Staatsrechts der Bundesrepublik Deutschland; Wassermann (ed) Kommentar zum Grundgesetz. There is even the odd constitutional law casebook: Richter and Schuppert Casebook Verfassungs-recht. Ö pp g 38 Here the work of Müller and Christensen Grundlagen Öffentliches Recht dealing particularly with the methodology of public law, is indispensible. They have recently also published a second volume in which they apply their methodology in the context of European law: Müller and Christensen Europarecht. Classics on legal methodology include Coing Juristische Methodenlehre; Engisch Einführung in das juristische Denken; Larenz Methodenlehre der Rechtswissenschaft; Schmalz Methodenlehre für das Juristische Studium; Zippelius Juristische Methodenlehre. L DU PLESSIS PER/PELJ 2005(8)1 In my field I have also come across work of German colleagues, creative in itself, and with catalytic qualities to unleash creative potential in others too. Examples are: In my field I have also come across work of German colleagues, creative in itself, and with catalytic qualities to unleash creative potential in others too. Examples are: • the work of Peter Häberle exploring the idea of “the constitution as an open process”;42 • Friedrich Müller’s strukturierende Rechtlehre which is post-structuralism incarnate – with German consistency and precision,43 and • Gerhard Robbers’ knack of bringing (in his teaching and writing) constitutional ideas to life through appreciation of music, architecture and the plastic arts.44 I mention these examples so as to confront the stereotyped perception that the work of German constitutional law scholars, though thorough and meticulously systematic, is mostly tedious and uncreative. I mention these examples so as to confront the stereotyped perception that the work of German constitutional law scholars, though thorough and meticulously systematic, is mostly tedious and uncreative. V Let me, in conclusion and as but a curtain raiser to the deliberations that are to follow, briefly take stock of what has happened to provisions of German origin in the South African Constitution – using the 1996 text as point of reference. There is a sense in which the 1996 South African Constitution bears a stronger German resemblance than the 1993 (transitional) Constitution. Section 1 of the former arguably fulfils a role similar to article 20 of the German Basic Law (albeit not in similar terms), namely to characterise the kind of state for which the Constitution provides: Let me, in conclusion and as but a curtain raiser to the deliberations that are to follow, briefly take stock of what has happened to provisions of German origin in the South African Constitution – using the 1996 text as point of reference. There is a sense in which the 1996 South African Constitution bears a stronger Let me, in conclusion and as but a curtain raiser to the deliberations that are to follow, briefly take stock of what has happened to provisions of German origin in the South African Constitution – using the 1996 text as point of reference. There is a sense in which the 1996 South African Constitution bears a stronger German resemblance than the 1993 (transitional) Constitution. 42 Häberle Verfassung als Öffentlicher Prozeß. 43 Müller and Christensen Grundlagen Öffentliches Recht; for a compact version of the strukturierende Rechtslehre cf Müller 1999 Stell LR 269-283. 42 Häberle Verfassung als Öffentlicher Prozeß. 43 Müller and Christensen Grundlagen Öffentliches Recht; for a compact version of the strukturierende Rechtslehre cf Müller 1999 Stell LR 269-283. 44 Robbers "Musik und Verfassung" 197-218. 42 Häberle Verfassung als Öffentlicher Prozeß. Ö IV For instance, the function, status and limits of theories and methods of (constitutional) interpretation (and of theories and methods of interpretation vis-à-vis one another) can, for instance, be problematic.40 In the German context Ernst- Wolfgang Böckenförde has, in the form of two relatively short but widely cited articles, made valuable (pioneering) contributions by duly distinguishing the respective interpretive roles of methods (and principles) of constitutional interpretation, on the one hand, and fundamental rights theories, on the other.41 pp g 38 Here the work of Müller and Christensen Grundlagen Öffentliches Recht dealing particularly with the methodology of public law, is indispensible. They have recently also published a second volume in which they apply their methodology in the context of European law: Müller and Christensen Europarecht. Classics on legal methodology include Coing Juristische Methodenlehre; Engisch Einführung in das juristische Denken; Larenz Methodenlehre der Rechtswissenschaft; Schmalz Methodenlehre für das Juristische Studium; Zippelius Juristische Methodenlehre. ; pp 39 In respect of my research theme Dreier and Schwegmann (eds) Probleme der Verfassungsinterpretation, for instance, stood out. pp 39 In respect of my research theme Dreier and Schwegmann (eds) Probleme der Verfassungsinterpretation, for instance, stood out. S 39 In respect of my research theme Dreier and Schwegmann (eds) Probleme der Verfassungsinterpretation, for instance, stood out. S 40 As Michelman 1995 SAJHR 482-485 shows. 41 Böckenförde 1974 NJW 1529 and Böckenförde 1976 NJW 2089. 41 Böckenförde 1974 NJW 1529 and Böckenförde 1976 NJW 2089. 88/150 L DU PLESSIS 45 "(1) Die Bundesrepublik Deutschland ist ein demokratischer und sozialer Bundesstaat. (2) Alle Staatsgewalt geht vom Volke aus. Sie wird vom Volke in Wahlen und Abstimmungen und durch besondere Organe der Gesetzgebung, der vollziehenden Gewalt und der Rechtsprechung ausgeübt. (3) Die Gesetzgebung ist an die verfassungsmäßige Ordnung, die vollziehende Gewalt und die Rechtsprechung sind an Gesetz und Recht gebunden. (4) Gegen jeden, der es unternimmt, diese Ordnung zu beseitigen, haben alle Deutschen das Recht zum Widerstand, wenn andere Abhilfe nicht möglich ist.“ 46 Cf th di i f th C tit ti l C t’ j i d th l f l i C i 45 (1) Die Bundesrepublik Deutschland ist ein demokratischer und sozialer Bundesstaat. (2) Alle Staatsgewalt geht vom Volke aus. Sie wird vom Volke in Wahlen und Abstimmungen und durch besondere Organe der Gesetzgebung, der vollziehenden Gewalt und der Rechtsprechung ausgeübt. (3) Die Gesetzgebung ist an die verfassungsmäßige Ordnung, die vollziehende Gewalt d di R ht h i d G t d R ht b d , g 46 Cf eg the discussion of the Constitutional Court’s jurisprudence on the rule of law in Currie and De Waal The new Constitutional and Administrative Law 75-81. 45 "(1) Die Bundesrepublik Deutschland ist ein demokratischer und sozialer Bundesstaat. p g g (4) Gegen jeden, der es unternimmt, diese Ordnung zu beseitigen, haben alle Deutschen das Recht zum Widerstand, wenn andere Abhilfe nicht möglich ist.“ L DU PLESSIS Section 1 of the former arguably fulfils a role similar to article 20 of the German Basic Law (albeit not in similar terms), namely to characterise the kind of state for which the Constitution provides: The Republic of South Africa is one, sovereign, democratic state founded on the following values: 89/150 L DU PLESSIS PER/PELJ 2005(8)1 (a) Human dignity, the achievement of equality and the advancement of human rights and freedoms. (b) Non-racialism and non-sexism. (c) Supremacy of the constitution and the rule of law. (d) Universal adult suffrage, a national common voters roll, regular elections and a multi-party system of democratic government, to ensure accountability, responsiveness and openness. p g g (3) Die Gesetzgebung ist an die verfassungsmäßige Ordnung, die vollziehende Gewalt und die Rechtsprechung sind an Gesetz und Recht gebunden. ( ) p (2) Alle Staatsgewalt geht vom Volke aus. Sie wird vom Volke in Wahlen und Abstimmungen und durch besondere Organe der Gesetzgebung, der vollziehenden Gewalt und der Rechtsprechung ausgeübt. Article 20 of the basic law states that: Article 20 of the basic law states that: (1) The Federal Republic of Germany is a democratic and social Federal state. (2) All state authority emanates from the people. It is exercised by the people by means of elections and voting and by separate legislative, executive and judicial organs. (3) Legislation is subject to the constitutional order; the executive and the judiciary are bound by the law. (4) All Germans shall have the right to resist any person seeking to abolish this constitutional order, should no other remedy be possible.45 Of the pentarchy Parteienstaat, Rechtsstaat, Streitbare Demokratie, Sozialstaat and Bundesstaat envisaged in article 20 of the Basic Law, at least the first three are also “appointed” (and I guess one could add “anointed”) by section 1 of the South African Constitution to shape the nature of the polity. What section 1(d) prescribes is certainly a party state and a competitive democracy, while “rule of law” in section 1(c) is the English common law way of saying “Rechtsstaat”, though the two are not exactly synonyms.46 That South Africa is a social state is not really stated in section 1, but it follows by necessary implication from several other constitutional provisions, for example, from section 7(2) that enjoins the state to “promote and fulfil” the rights in the Bill of Rights, from the 90/150 PER/PELJ 2005(8)1 L DU PLESSIS authorisation of affirmative action in section 9(2), from the entrenchment of socio-economic entitlements (albeit restrained) in (amongst others) sections 26 and 27 and last but not the least from the jurisprudence of the Constitutional Court, construing these and other similar provisions.47 It is no secret that for historical reasons prominent South Africans at the helm of constitution-making and, subsequently, government in this country, do not cherish federalist sentiments akin to those of their German counterparts. It is therefore not surprising that section 1 of the South African Constitution does not describe South Africa as a federal state. At the same time South Africa is not anti-federal and can best be described as a co-operative as opposed to a competitive federation.48 A principle that in Germany is meant to counteract the fragmentation of the federation, is invoked in South Africa to impel co-operation among the various spheres of government. 49 See n 20 above. 50 For a reference to this visit, cf ANC 1998 http://www.anc.org.za/people/delange.html 21 Nov (parliamentarian and deputy-minister Johnny de Lange’s curriculum vitae). I am also indebted to Prof Francois Venter of the North-West University, Potchefstroom, who provided me with an oral account of this visit and its consequences. 47 Government of the Republic of South Africa v Grootboom 2000 (11) BCLR 1235 (CC) and Treatment Action Campaign v Minister of Health (1) 2002 (10) BCLR 1033 (CC) 48 Cf th di i f th C tit ti l C t’ j i d th l f l i C i p g ( ) ( ) ( ) 48 Cf eg the discussion of the Constitutional Court’s jurisprudence on the rule of law in Cur and De Waal The new Constitutional and Administrative Law 119-124. 47 Government of the Republic of South Africa v Grootboom 2000 (11) BCLR 1235 (CC) and Treatment Action Campaign v Minister of Health (1) 2002 (10) BCLR 1033 (CC) 48 Cf eg the discussion of the Constitutional Court’s jurisprudence on the rule of law in Currie and De Waal The new Constitutional and Administrative Law 119-124. 49 See n 20 above. Article 20 of the basic law states that: This principle, of course, is the German notion of Bundestreue which, in Chapter 3 of the South African Constitution, goes by the name of “co-operative government”. On an official visit to South Africa in September 199549 former German chancellor, Helmut Kohl, invited South Africa’s constitution-makers to send a delegation to Germany to study aspects of German constitutionalism. A multi-party delegation, consisting mainly of members of the National Assembly and advisers, visited Germany from 8 to 15 January 1996.50 One of the outcomes of this visit was that German ideas on federalism were included in South Africa’s 1996 constitutional text without, however, fully embracing the German system of federalism whose centrifugal impetus is arguably more powerful than that of the South African system. 91/150 47 Government of the Republic of South Africa v Grootboom 2000 (11) BCLR 1235 (CC) and Treatment Action Campaign v Minister of Health (1) 2002 (10) BCLR 1033 (CC) 48 Cf eg the discussion of the Constitutional Court’s jurisprudence on the rule of law in Currie and De Waal The new Constitutional and Administrative Law 119-124. 49 See n 20 above. 50 For a reference to this visit, cf ANC 1998 http://www.anc.org.za/people/delange.html 21 Nov (parliamentarian and deputy-minister Johnny de Lange’s curriculum vitae). I am also indebted to Prof Francois Venter of the North-West University, Potchefstroom, who provided me with an oral account of this visit and its consequences. 91/150 L DU PLESSIS PER/PELJ 2005(8)1 VI VI Anyone conversant with the German Basic Law and the foundational prominence it affords human dignity, will be struck by the statement in section 1(a) of the South African Constitution that human dignity, together with the achievement of equality and the advancement of human rights and freedoms is one of the founding values of the Republic of South Africa as “one, sovereign, democratic state”. References to human dignity also occur in certain pivotal provisions in the Bill of Rights (chapter 2 of the Constitution): • Section 7(1) states that the Bill of Rights “affirms the democratic values of human dignity, equality and freedom”. • Section 7(1) states that the Bill of Rights “affirms the democratic values of human dignity, equality and freedom”. • The general limitation clause, section 36, requires limitations of rights entrenched in the Bill of Rights to comply with the threshold of (amongst others) the extent to which the proposed limitation “is reasonable in an open and democratic society based on human dignity, equality and freedom”.51 • In section 39(1)(a) judicial authorities interpreting the Bill of Rights are enjoined “to promote the values that underlie an open and democratic society based on human dignity, equality and freedom”. The occurrence of the triumvirate of human dignity, equality and freedom in several places in the 1996 Bill of Rights tells where we in South Africa were coming from when we first tried to establish our fledgling constitutional democracy – exactly what article 1 of the Basic Law does with reference to Germany’s unique (and sad) holocaust history. Comparable provisions in South Africa’s transitional Bill of Rights mentioned freedom and equality, but not dignity. In human rights discourse and practice there is a perennial tension between freedom and equality, especially in a society like South Africa where astounding disparities between the haves and have nots still prevail. To the haves, on the one hand, their freedom is a vital item in the arsenal of whatever 51 S 36(1). 51 S 36(1). 51 92/150 PER/PELJ 2005(8)1 L DU PLESSIS empowers them to have ever more. They are particularly mistrustful of government interfering in their affairs. The have nots, on the other hand, demand government intervention to empower them to gain a rightful share in the country’s wealth – and this threatens the position of the haves in whom most of the wealth resides. VI It was therefore eminently sensible of the authors of the 1996 Constitution to come up with a textual strategy that can help negotiate the seemingly insoluble tension between freedom and equality: the inclusion in the constitutional sections referred to above of the value of human dignity with its particular history (also and especially in Germany) curbs the tendency to over-concentrate attention – in an “either . . . or” manner – on the fear of the haves and the plight of the have nots, and it demands deference to the worth and eminence of both the have nots and the haves as dignified human beings. VII 53 The way in which s 8 of the 1996 Constitution has been structured, arguably precludes the necessity of an effort as monumental as Du Plessis v De Klerk to unravel the mystery of horizontal application. L DU PLESSIS L DU PLESSIS L DU PLESSIS Klerk and Another.52 Part of the package of the section 7 deal was the inclusion in the transitional Constitution of section 35(3), the predecessor to section 39(2) of the 1996 Constitution, in an attempt to ensure that the provisions of the Bill of Rights will at least have some “radiating effect” on the interpretation and application of non-constitutional law. The wording of the two consecutive provisions is very similar. That is, however, not the case with sections 7(1) and (2) of the transitional Constitution, and their successors in the 1996 Constitution, sections 8(1)-(3). The latter provisions authorise the direct horizontal application of “[a] provision of the Bill of Rights . . . to the extent that it is applicable taking into account the nature of the right and the nature of any duty imposed by the right”. There has not been a Du Plessis v De Klerk on section 8 (yet),53 but there has in the meantime been a Carmichele v Minister of Safety and Security and Another,54 a hurricane that could well be named Alix Jean after a very brave litigant, that has swept through our existing (common and especially private) law, with a force that initially seemed to have washed away all interest in the vexing (yet neglected) question posed by sections 8(1)- (3) of the Constitution, namely: “Precisely how directly do horizontal provisions of the Bill of Rights apply?” In the meantime, however, in the case of Khumalo and Others v Holomisa55 O’Regan j intimated that direct horizontal application of the Bill of Rights (to natural and juristic persons) is always a possibility, depending on the circumstances of each particular case, but section 8(3) of the Constitution requires any such application of a particular right in the Bill of Rights to be mediated by the common law. “Common law” can be the common law as it stands or the common law as developed by the court if the law as it stands does not adequately cater for the exigencies of the situation under consideration. pp 54 Carmichele v Minister of Safety and Security 2001 (10) BCLR 995 (2001 (4) SA 938) (CC 52 Du Plessis v De Klerk 1996 (5) BCLR 659 (1996 (3) SA 850) (CC). 54 Carmichele v Minister of Safety and Security 2001 (10) BCLR 995 (2001 (4) SA 938) (CC 55 Khumalo v Holomisa 2002 (8) BCLR 771 (2002 5 SA 401) (CC) par 31-34. 52 D Pl i D Kl k 1996 (5) BCLR 659 (1996 (3) SA 850) (CC) y y ( ) ( ( ) 55 Khumalo v Holomisa 2002 (8) BCLR 771 (2002 5 SA 401) (CC) par 31-34. pp 54 Carmichele v Minister of Safety and Security 2001 (10) BCLR 995 (2001 (4) SA 938) (CC). 55 Khumalo v Holomisa 2002 (8) BCLR 771 (2002 5 SA 401) (CC) par 31-34. VII Section 39(2) of the South African Constitution enjoins the judiciary, “[w]hen interpreting any legislation and when developing the common law or customary law” to “promote the spirit, purport and objects of the Bill of Rights”. Looked at superficially, this may seem to be a rather trite (and therefore superfluous) injunction. However, not only does this provision have a fascinating history, but it has also become a kingpin in the evolution of South African law. At the 1993 multi-party negotiations there was a strong sentiment against direct horizontal application of the Bill of Rights. It came from (amongst others and strangely enough) the representatives of the South African Communist Party. After an intense and lively debate in which the German notions of mittelbare and unmittelbare Drittwirkung featured freely and prominently, a compromise was reached and subsequently written into sections 7(1) and (2) of the transitional Constitution. The said section left room for a restrictive understanding of the operation of the Bill of Rights, excluding its direct horizontal effect. This was also how the section was eventually construed in the then landmark (and now almost forgotten) Constitutional Court judgement of Du Plessis and Others v De 93/150 PER/PELJ 2005(8)1 L DU PLESSIS Klerk and Another.52 Part of the package of the section 7 deal was the inclusion in the transitional Constitution of section 35(3), the predecessor to section 39(2) of the 1996 Constitution, in an attempt to ensure that the provisions of the Bill of Rights will at least have some “radiating effect” on the interpretation and application of non-constitutional law. The wording of the two consecutive provisions is very similar. That is, however, not the case with sections 7(1) and (2) of the transitional Constitution, and their successors in the 1996 Constitution, sections 8(1)-(3). The latter provisions authorise the direct horizontal application of “[a] provision of the Bill of Rights . . . to the extent that it is applicable taking into account the nature of the right and the nature of any duty imposed by the right”. There has not been a Du Plessis v De Klerk on section 8 (yet),53 but there has in the meantime been a Carmichele v Minister of Safety and Security and Another,54 a hurricane that could well be named Alix Jean after a very brave litigant, that has swept through our existing (common and especially private) law, with a force that initially seemed to have washed away all interest in the vexing (yet neglected) question posed by sections 8(1)- (3) of the Constitution, namely: “Precisely how directly do horizontal provisions of the Bill of Rights apply?” In the meantime, however, in the case of Khumalo and Others v Holomisa55 O’Regan j intimated that direct horizontal application of the Bill of Rights (to natural and juristic persons) is always a possibility, depending on the circumstances of each particular case, but section 8(3) of the Constitution requires any such application of a particular right in the Bill of Rights to be mediated by the common law. “Common law” can be the common law as it stands or the common law as developed by the court if the law as it stands does not adequately cater for the exigencies of the situation under consideration The Constitutional Court’s judgement in Carmichele was not an interpretation and application of sections 8(1)-(3), but of section 39(2), and it has opened 94/150 PER/PELJ 2005(8)1 L DU PLESSIS floodgates of legal reform through judicial intervention to such an extent that it has probably become necessary to consider strategies to channel the flood. 95/150 56 S v Mhlungu 1995 (7) BCLR 793 (1995 (3) SA 867) (CC) par 59. 57 Cf Du Plessis (Re-)Interpretation of Statutes 29-32. 58 Cf eg Govender v Minister of Safety and Security 2001 (4) SA 273 (SCA). L DU PLESSIS A possible strategy, which our Constitutional Court has invoked without naming it, is what has become famous in Germany (and all of Europe) by the name of subsidiarity. In S v Mhlungu and Others56 subsidiarity “made in South Africa” was verbalised as follows: I would lay it down as a general principle that where it is possible to decide any case, civil or criminal, without reaching a constitutional issue, that is the course which should be followed. I would lay it down as a general principle that where it is possible to decide any case, civil or criminal, without reaching a constitutional issue, that is the course which should be followed. In some of my own writings I have proposed an adaptation of subsidiarity for use in the South African context.57 I leave the detail for another occasion. It is important for the present to note that this is an area for very fruitful co-operation with our German counterparts in public law and probably also European law – given the history of section 39(2) as product of a Drittwirkung debate. In some of my own writings I have proposed an adaptation of subsidiarity for use in the South African context.57 I leave the detail for another occasion. It is important for the present to note that this is an area for very fruitful co-operation with our German counterparts in public law and probably also European law – given the history of section 39(2) as product of a Drittwirkung debate. VIII g ( ) ( ( ) ) ( ) p 57 Cf Du Plessis (Re-)Interpretation of Statutes 29-32. 56 S v Mhlungu 1995 (7) BCLR 793 (1995 (3) SA 867) (CC) par 59. 57 Cf Du Plessis (Re-)Interpretation of Statutes 29-32. 58 Cf G d Mi i f S f d S i 2001 (4) SA 273 (SCA) 56 S v Mhlungu 1995 (7) BCLR 793 (1995 (3) SA 867) (CC) par 59. ( ) p 58 Cf eg Govender v Minister of Safety and Security 2001 (4) SA 273 (SCA 96/150 59 S 33(1)(b). 60 Rabie, Van der Merwe and Labuschagne 1993 THRHR 622. 61 Cf eg Lawrence "From Soweto to Codesa" 1. L DU PLESSIS More directly and exclusively German was the Wesensgehaltgarantie included in the general limitation clause59 in the transitional Bill of Rights. Like their German counterparts South African constitutional scholars and lawyers struggled to get to the heart of the essential content of rights and what added to our misery, on this side of the equator, was the fact that we took over a provision with a singular history and then ignored that history when we wrote in into our transitional Bill of Rights. I have not witnessed a single tear being shed over the demise of the Wesensgehaltgarantie in the 1996 Constitution. VIII The transitional Constitution contained three provisions that were eventually excluded from the 1996 Constitution. Two of them, sections 35(2) and 232(3) prescribed a widely acknowledged interpretive procedure, known as verfassungskonforme Auslegung in the German context, for the interpretation of, respectively, the Bill of Rights and the Constitution as a whole. The omission of these two provisions from the 1996 Constitution did not preclude judicial reliance on the interpretive procedure they (previously) prescribed58 and verfassungskonforme Auslegung is still very much part of our constitutional law as it stands. PER/PELJ 2005(8)1 L DU PLESSIS IX There is much for which we as scholars of and citizens under the South African Constitution can thank our German counterparts. Fortunately they are not in the bad habit of constantly reminding us of it. In the formerly referred to Humboldt article, the authors conclude:60 Over almost three decades, the Alexander von Humboldt Foundation has contributed substantially to the personal development and growth of a number of jurists who have, in turn, exercised a significant influence on both the theory and the practice of South African law. This the Foundation has achieved without endeavouring to transplant German legal doctrine on its South African counterpart. Rather, scholars have been enabled to develop their own legal thought and theory, and so to be part of the evolution of their legal system as they consider appropriate. In my experience a similar spirit permeates interaction between South African and German scholars in matters constitutional. And I do not think that especially during the last ten to fifteen years it has been just a one way traffic: there is indeed much to be learnt from South Africa and especially from our constitutional enterprise of accommodating (and desiring to celebrate) our diversity as an asset. South Africa’s “small miracle”, à la Nelson Mandela,61 can L DU PLESSIS PER/PELJ 2005(8)1 L DU PLESSIS only be sustained if we do not take it for granted – and if we share its fruits with kindred spirits the world over. Humboldt exchanges contribute to a considerable extent to the creation of a supra-national context within which this can be done. 97/150 L DU PLESSIS PER/PELJ 2005(8)1 Bibliography Arnim Staatslehre der Bundesrepublik Deutschland Arnim HH von Staatslehre der Bundesrepublik Deutschland (Vahlen München 1984) Baur Gott, Recht und Weltliches Regiment im Werke Calvins Baur J Gott, Recht und Weltliches Regiment im Werke Calvins (Bouvier Bonn 1965) Benda, Maihofer and Vogel (eds) Handbuch des Verfassungsrechts Benda E, Maihofer W and Vogel HJ (eds) Handbuch des Verfassungsrechts 2nd ed (De Gruyter Berlin 1995) Böckenförde1976 NJW 2089 Böckenförde E-W "Die Methoden der Verfassungsinterpretation - Bestandaufnahme und Kritik" 1976 NJW (29) 2089 Böckenförde 1974 NJW 1529 Böckenförde E-W 1974 "Grundrechtstheorie und Grundrechtsinterpretation" NJW (27) 1529 Brugger 1994 Archiv des Öffentlichen Rechts 1-34 Brugger, W “Konkretisierung des Rechts und Auslegung der Gesetze” 1994 Archiv des Öffentlichen Rechts (119) 1 Cliteur Inleiding in het Recht 196-202 Cliteur PB Inleiding in het Recht (Wolters-Noordhoff Groningen 1992) Coing Juristische Methodenlehre Coing H Juristische Methodenlehre (De Gruyther Berlin 1972) L DU PLESSIS PER/PELJ 2005(8)1 Bibliography PER/PELJ 2005(8)1 L DU PLESSIS Côté The Interpretation of Legislation in Canada 193-350 Côté P-A The Interpretation of legislation in Canada (Yvon Blais Cowansville 1984) Currie and De Waal Bill of Rights Handbook Currie I and De Waal J The Bill of Rights Handbook 5th ed (Juta Kenwyn 2005) Currie and De Waal The new Constitutional and Administrative Law 75-81 Currie I and De Waal J The new Constitutional and Administrative Law Volume 1 (Juta Lansdowne 2001) Davis, Chaskalson and De Waal "Democracy and Constitutionalism" 1-130 Davis D, Chaskalson M and De Waal J "Democracy and constitutionalism: the role of constitutional interpretation" in Van Wyk D et al (eds) Rights and constitutionalism: the new South African legal order (Kenwyn Juta 1994) De Waal 1995 SAJHR 1 De Waal J “A Comparative Analysis of the Provisions of German Origin in the Interim Bill of Rights” 1995 SAJHR (11) 1 De Wet Constitutional Enforceability of Economic and Social Rights De Wet E The Constitutional Enforceability of Economic and Social Rights (The meaning of the German Constitutional Model for South Africa) (Butterworth Durban 1996) Dreier and Schwegmann (eds) Probleme der Verfassungsinterpretation Dreier R and Schwegmann F (eds) Probleme der Verfassungsinterpretation: Dokumentation e. Kontroverse (Nomos Baden- Baden 1976) Du Plessis 1998 Acta Juridica 8 Du Plessis LM “The Jurisprudence of Interpretation and the Exigencies of a new Constitutional Order in South Africa” 1998 Acta Juridica 8 L DU PLESSIS PER/PELJ 2005(8)1 100/150 Du Plessis 1984 THRHR 127 Du Plessis LM “'n Regsorde sonder Norme? Bibliography ('n Verkenning van normrelativerende Tendense in enkele irrasionalistiese Teorieë oor die Reg)“ 1984 THRHR 47 (2) Du Plessis Introduction to Law Du Plessis LM An Introduction to Law 3rd ed (Juta Kenwyn 1999) Du Plessis (Re-)Interpretation of Statutes Du Plessis LM The (Re-)Interpretation of Statutes (Butterworth Durban 2002) Du Plessis "South Africa’s peaceful Revolution" 209-224 Du Plessis LM "South Africa’s peaceful Revolution and its constitutional Safegaurding" in Kramer J and Schubert BG (eds) Verfassungsgebung und Verfassungsreform im In- und Ausland (Nomos Baden-Baden 2005) Dugard 1971 SALJ 181 Dugard J “The Judicial Process, Positivism and Civil Liberty” 1971 SALJ 88 Dugard Human Rights and the South African Legal Order 393, 397 Dugard J Human Rights and the South African Legal Order (University Press Princeton 1978) Engisch Einführung in das juristische Denken Engisch K Einführung in das juristische Denken 8th ed (Kohlhammer Berlin 1989) Fagan "Roman-Dutch Law" 60-64 Fagan E "Roman-Dutch Law in its South African historical Context" in Zimmermann R and Visser D (eds) Southern Cross: Civil Law and Common Law in South Africa (Juta Kenwyn 1996) 33-64 100/150 PER/PELJ 2005(8)1 L DU PLESSIS Forsthoff Zur Problematik der Verfassungsauslegung Forsthoff E Zur Problematik der Verfassungsauslegung (Kohlhammer Stuttgart 1961) Häberle Verfassung als Öffentlicher Prozeß Häberle P Verfassung als Öffentlicher Prozeß: Materialen zu einer Verfassungstheorie der Offenen Gesellschaft (Duncker & Humblot Berlin 1978) Hahlo and Kahn South African Legal System 180 Hahlo HR and Kahn E The South African Legal System and its Background (Juta Cape Town 1973) Hesse Grundzüge des Verfassungsrechts der Bundesrepublik Deutschland Hesse K Grundzüge des Verfassungsrechts der Bundesrepublik Deutschland 19th ed (Müller Heidelberg 1993) Isensee and Kirchhof (eds) Handbuch des Staatsrechts der Bundesrepublik Deutschland Isensee J and Kirchhof P (eds) Handbuch des Staatsrechts der Bundesrepublik Deutschland (Müller Heidelberg 1992) Kommers Constitutional Jurisprudence Kommers DP The Constitutional Jurisprudence of the Federal Republic of Germany 2nd ed (University Press Durham 1997) Labuschagne 2004 THRHR 43 Labuschagne JMT “Die dinamiese Regslandskap, Doelaanpassing en - vervanging en die Geregtigheidswaarde van die Normdop by Regsuitleg en -vorming: Opmerkinge oor die anachronistiese kant van die Stelreël cessante ratione legis, cessat et ipsa lex” 2004 THRHR 67 (1) 43 Labuschagne 2004 THRHR 43 Labuschagne JMT “Die dinamiese Regslandskap, Doelaanpassing en - vervanging en die Geregtigheidswaarde van die Normdop by Regsuitleg en -vorming: Opmerkinge oor die anachronistiese kant van die Stelreël cessante ratione legis, cessat et ipsa lex” 2004 THRHR 67 (1) 43 101/150 L DU PLESSIS PER/PELJ 2005(8)1 L DU PLESSIS PER/PELJ 2005(8)1 102/150 Labuschagne 1983 THRHR 422 Labuschagne JMT “Regsdinamika: Opmerkinge oor die Aard van die Wetgewingsproses” 1983 THRHR 46 (4) Larenz Methodenlehre der Rechtswissenschaft Larenz K Methodenlehre der Rechtswissenschaft 6th ed (Springer Heidelberg 1991) Lawrence "From Soweto to Codesa" 1-12 Lawrence R "Introduction: From Soweto to Codesa" in Friedman S and Atkinson D (eds) The South African Review 7. The Small Miracle. South Africa’s negotiated Settlement (Ravan Press Johannesburg 1994) 1-12 Maunz and Zippelius Deutsches Staatsrecht Maunz T and Zippelius R Deutsches Staatsrecht: Ein Studienbuch 28th ed (Beck München 1991) Michelman 1995 SAJHR 477 Michelman FI “A Constitutional Conversation with Professor Frank Michelman” 1995 SAJHR 477 Müller 1999 Stell LR 269 Müller F “Basic Questions of Constitutional Concretisation” 1999 Stell LR 10 (3) 269 Müller and Christensen Europarecht Müller F and Christensen R Juristische Methodik Band II: Europarecht (Duncker & Humblot Berlin 2003) Müller and Christensen Grundlagen Öffentliches Recht Müller F and Christensen R Juristische Methodik Band I: Grundlagen Öffentliches Recht 8th ed (Duncker & Humblot Berlin 2002) 102/150 PER/PELJ 2005(8)1 L DU PLESSIS Rabie, Van der Merwe and Labuschagne 1993 THRHR 608 Rabie A, Van der Merwe S and Labuschagne JMT “The Contribution of the Alexander von Humboldt Foundation to the Development of the South African Legal System and Literature” 1993 THRHR 56 (4) 608 Rabie, Van der Merwe and Labuschagne 1993 THRHR 608 Rabie A, Van der Merwe S and Labuschagne JMT “The Contribution of the Alexander von Humboldt Foundation to the Development of the South African Legal System and Literature” 1993 THRHR 56 (4) 608 African Legal System and Literature” 1993 THRHR 56 (4) 608 Richter and Schuppert Casebook Verfassungsrecht Richter I and Schuppert GF Casebook Verfassungsrecht 3rd ed (Beck München 1996) Robbers "Musik und Verfassung" 197-218 Robbers G "Musik und Verfassung" in Becker H-J (ed) Interdependenzen zwischen Verfassung und Kultur (Duncker & Humblot Berlin 2003) 197-218 Sachs (ed) Grundgesetz Kommentar Sachs M (ed) Grundgesetz Kommentar 3rd ed (Beck München 2003) Schmalz Methodenlehre für das Juristische Studium Schmalz D Methodenlehre für das Juristische Studium 3rd ed (Nomos Baden-Baden 1992) Stein Staatsrecht Stein E Staatsrecht 14th ed (Mohr Tübingen 1993) Stern Das Staatsrecht der Bundesrepublik Deutschland Stern K Das Staatsrecht der Bundesrepublik Deutschland (Beck München 1994) Steyler "Constitutional Promise of Decentralization in Practice" 225-240 Steyler N "The constitutional Promise of Decentralization in Practice: A Review of 1994-2002" in Kramer J and Schubert BG (eds) Verfassungsgebung und Verfassungsreform im In- und Ausland (Nomos Baden-Baden 2005) 103/150 PER/PELJ 2005(8)1 L DU PLESSIS L DU PLESSIS PER/PELJ 2005(8)1 S v Mhlungu and Others 1995 (7) BCLR 793 (1995 (3) SA 867) (CC) Treatment Action Campaign and Others v Minister of Health and Others (1) 2002 (10) BCLR 1033 (CC) S v Mhlungu and Others 1995 (7) BCLR 793 (1995 (3) SA 867) (CC) Treatment Action Campaign and Others v Minister of Health and Others (1) 2002 (10) BCLR 1033 (CC) L DU PLESSIS PER/PELJ 2005(8)1 105/150 Wassermann (ed) Kommentar zum Grundgesetz Wassermann R (ed) Kommentar zum Grundgesetz für die Bundesrepublik Deutschland (Neuwied Darmstdadt Luchterhand 1984) Wiechers "Publieke Subjektiewe Reg" 270-291 Wiechers M "Die Publieke Subjektiewe Reg" in Strauss SA (ed) Huldigingsbundel vir WA Joubert (Butterworth Durban 1988) Wolf Das Problem der Naturrechtslehre Wolf E Das Problem der Naturrechtslehre (Müller Karlsruhe 1959) Wolf Griechisches Rechtsdenken Wolf E Griechisches Rechtsdenken III I (Frankfurt Klosterman 1954) Zippelius Juristische Methodenlehre Zippelius R Juristische Methodenlehre: Eine Einführung 6th ed (Beck München 1994) Register of legislation Constitution of the Republic of South Africa 1996 Constitution of the Republic of South Africa 200 of 1993 Register of cases Carmichele v Minister of Safety and Security and Another 2001 (10) BCLR 995 (2001 (4) SA 938) (CC) Du Plessis and Others v De Klerk and Another 1996 (5) BCLR 659 (1996 (3) SA 850) (CC) Govender v Minister of Safety and Security 2001 (4) SA 273 (SCA) Government of the Republic of South Africa and Others v Grootboom and Others 2000 (11) BCLR 1235 (CC) Khumalo and Others v Holomisa 2002 (8) BCLR 771 (2002 5 SA 401) (CC) PER/PELJ 2005(8)1 L DU PLESSIS Wassermann (ed) Kommentar zum Grundgesetz Wassermann R (ed) Kommentar zum Grundgesetz für die Bundesrepublik Deutschland (Neuwied Darmstdadt Luchterhand 1984) Wiechers "Publieke Subjektiewe Reg" 270-291 Wiechers M "Die Publieke Subjektiewe Reg" in Strauss SA (ed) Huldigingsbundel vir WA Joubert (Butterworth Durban 1988) Wolf Das Problem der Naturrechtslehre Wolf E Das Problem der Naturrechtslehre (Müller Karlsruhe 1959) Wolf Griechisches Rechtsdenken Wolf E Griechisches Rechtsdenken III I (Frankfurt Klosterman 1954) Zippelius Juristische Methodenlehre Zippelius R Juristische Methodenlehre: Eine Einführung 6th ed (Beck München 1994) Register of legislation Constitution of the Republic of South Africa 1996 Constitution of the Republic of South Africa 200 of 1993 Register of cases Carmichele v Minister of Safety and Security and Another 2001 (10) BCLR 995 (2001 (4) SA 938) (CC) Du Plessis and Others v De Klerk and Another 1996 (5) BCLR 659 (1996 (3) SA 850) (CC) Govender v Minister of Safety and Security 2001 (4) SA 273 (SCA) Government of the Republic of South Africa and Others v Grootboom and Others 2000 (11) BCLR 1235 (CC) Khumalo and Others v Holomisa 2002 (8) BCLR 771 (2002 5 SA 401) (CC) L DU PLESSIS Register of Internet resources ANC 1998 http://www.anc.org.za/people/delange.html 21 Nov African National Congress 1998 Biographies of ANC leaders: Johnny de Lange [Found on Internet] http://www.anc.org.za/people/delange.html [Date of use 21 Nov 2005] ANC 1995 http://www.anc.org.za/ancdocs/history/mandela/1995/sp950911.html 21 Nov African National Congress 1995 ANC documents: History. Presidential Speeches. Speech at State Banquet for the German Chancellor Dr Helmut Kohl [Found on Internet] http://www.anc.org.za/ancdocs/history/mandela/1995/sp950911.html [Date of use 21 Nov 2005] ANC 1995 http://www.anc.org.za/ancdocs/history/mandela/1995/sp950911.html 21 Nov African National Congress 1995 ANC documents: History. Presidential Speeches. Speech at State Banquet for the German Chancellor Dr Helmut Kohl [Found on Internet] http://www.anc.org.za/ancdocs/history/mandela/1995/sp950911.html [Date of use 21 Nov 2005] Lindemans 2002 http://www.pantheon.org/articles/v/valhalla.html 21 Nov Lindemans MF 2002 Valhalla [Found on Internet] http://www.pantheon.org/articles/v/valhalla.html [Date of use 21 Nov 2005] 106/150
https://openalex.org/W2044705396	https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0022011&type=printable	English	null	Subdominant/Cryptic CD8 T Cell Epitopes Contribute to Resistance against Experimental Infection with a Human Protozoan Parasite	PloS one	2,011	cc-by	12,466	Abstract During adaptive immune response, pathogen-specific CD8+ T cells recognize preferentially a small number of epitopes, a phenomenon known as immunodominance. Its biological implications during natural or vaccine-induced immune responses are still unclear. Earlier, we have shown that during experimental infection, the human intracellular pathogen Trypanosoma cruzi restricts the repertoire of CD8+ T cells generating strong immunodominance. We hypothesized that this phenomenon could be a mechanism used by the parasite to reduce the breath and magnitude of the immune response, favoring parasitism, and thus that artificially broadening the T cell repertoire could favor the host. Here, we confirmed our previous observation by showing that CD8+ T cells of H-2a infected mice recognized a single epitope of an immunodominant antigen of the trans-sialidase super-family. In sharp contrast, CD8+ T cells from mice immunized with recombinant genetic vaccines (plasmid DNA and adenovirus) expressing this same T. cruzi antigen recognized, in addition to the immunodominant epitope, two other subdominant epitopes. This unexpected observation allowed us to test the protective role of the immune response to subdominant epitopes. This was accomplished by genetic vaccination of mice with mutated genes that did not express a functional immunodominant epitope. We found that these mice developed immune responses directed solely to the subdominant/cryptic CD8 T cell epitopes and a significant degree of protective immunity against infection mediated by CD8+ T cells. We concluded that artificially broadening the T cell repertoire contributes to host resistance against infection, a finding that has implications for the host-parasite relationship and vaccine development. Citation: Dominguez MR, Silveira ELV, de Vasconcelos JRC, de Alencar BCG, Machado AV, et al. (2011) Subdominant/Cryptic CD8 T Cell Epitopes Contribute to Resistance against Experimental Infection with a Human Protozoan Parasite. PLoS ONE 6(7): e22011. doi:10.1371/journal.pone.0022011 Editor: Georges Snounou, Universite´ Pierre et Marie Curie, FRANCE Editor: Georges Snounou, Universite´ Pierre et Marie Curie, FRANCE Received March 29, 2011; Accepted June 11, 2011; Published July 14, 2011 Copyright: 2011 Dominguez et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: Fundac¸a˜o de Amparo a` Pesquisa do Estado de Sa˜o Paulo (2009/06820-4), The National Institute for Vaccine Technology (INCTV-CNPq), The Millennium Institute for Vaccine Development and Technology (CNPq - 420067/2005-1) and The Millennium Institute for Gene Therapy (Brazil). Subdominant/Cryptic CD8 T Cell Epitopes Contribute to Resistance against Experimental Infection with a Human Protozoan Parasite Mariana R. Dominguez1,2, Eduardo L. V. Silveira1,2¤a, Jose´ Ronnie C. de Vasconce Alencar1,2¤b, Alexandre V. Machado3, Oscar Bruna-Romero4, Ricardo T. Gazzinel Rodrigues1,2* Mariana R. Dominguez1,2, Eduardo L. V. Silveira1,2¤a, Jose´ Ronnie C. de Vasconcelos1,2, Bruna C. G. de Alencar1,2¤b, Alexandre V. Machado3, Oscar Bruna-Romero4, Ricardo T. Gazzinelli4,5,6, Mauricio M. 1 2 1 Centro de Terapia Celular e Molecular (CTCMol), Universidade Federal de Sa˜o Paulo-Escola Paulista de Medicina, Sa˜o Paulo, Brazil, 2 Departamento de Microbiologia, Imunologia e Parasitologia, Universidade Federal de Sa˜o Paulo-Escola Paulista de Medicina, Sa˜o Paulo, Brazil, 3 Centro de Pesquisas Rene´ Rachou, FIOCRUZ, Belo Horizonte, Minas Gerais, Brazil, 4 Departamento de Microbiologia, Instituto de Cieˆncias Biolo´gicas, Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais, Brazil, 5 Departamento de Bioquı´mica e Imunologia, Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais, Brazil, 6 Division of Infectious Disease and Immunology, Department of Medicine, University of Massachusetts Medical School, Worcester, Massachusetts, United States of America Abstract EVLS, RTG and MMR are recipients of fellowships from CNPq. MRD and BCA are recipients of fellowships from FAPESP. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. terests: The authors have declared that no competing interests e * E-mail: mrodrigues@unifesp.br ¤a Current address: Yerkes National Primate Research Center/Emory University, Atlanta, Georgia, United States of America ¤b Current address: Institut Curie U932, Paris, France ¤a Current address: Yerkes National Primate Research Center/Emory University, Atlanta, Georgia, United States of America ¤b Current address: Institut Curie U932, Paris, France PLoS ONE \| www.plosone.org Ethics Statement All experimental procedures were approved by the Ethics Committee for Animal Care of the Federal University of Sa˜o Paulo (Id # CEP 0426/09). In vivo depletion of CD8+ T cells were performed by treating vaccinated A/Sn mice with 53.6.7 MAb. At days 2 and 3 before challenge with trypomastigotes, mice were injected i.p. with a dose of 1 mg of anti-CD8 or control Rat IgG. Seven days after challenge, each mouse received one more dose of 1 mg of anti- CD8 or Rat IgG. The efficacy of depletion of CD8+ spleen cells Immunity to Subdominant/Cryptic Epitopes Immunity to Subdominant/Cryptic Epitopes B10.A or A/Sn mouse was challenged i.p. with a final dose containing 104 or 150 parasites, respectively, in a final volume of 0.2 mL. Parasite development was monitored by counting the number of bloodstream trypomastigotes in 5 mL of fresh blood collected from the tail vein [12]. result of different mechanisms that regulate the formation of the complex of MHC-I-peptide on the surface of antigen presenting cells (APC) such as antigen concentration, stability or epitope availability after processing and translocation to the endoplasmic reticulum, where the MHC-I-peptide complex is assembled to be transported to the APC surface [21–24]. After a stable MHC-I-peptide complex is formed on the surface of the APC, factors related to CD8+ T cells, such as the frequency of precursors, their TCR affinities, their capacity to proliferate in response to antigen and, thus, be incorporated into the pool of responder cells, are factors that shape immunodominance hierar- chies. These factors transcend the MHC restriction element and may create T-cell competition for APCs and other resources, enabling certain CD8+ T cells to dominate and suppress others [23–29]. ii) 59-GGGTCTAGATCAGACCATTTTTAGTTCACC-39. The final PCR product was completely sequenced. The only modifications found were in the nucleotide sequences encoding the immunodominant epitope TEWETGQI. The new sequence encoded the amino acids (AA) TAWETGQA. This plasmid is referred to as pIgSpTAWETGQA. The new gene was also subcloned into the pAdCMV shuttle vector, and the recombinant replication-defective adenovirus human adenovirus 5 was pro- duced by Vectors BioLabs, Philadelphia, USA. This new recom- binant adenovirus is referred to as AdTAWETGQA. Viruses and plasmids were purified as described previously [35–37]. Mice were inoculated intra-muscularly (i.m.) in each tibialis anterioris muscle with 50 mg of plasmid DNA 3 times every 3 weeks. Peptides Peptides were purchased from Genscript (Piscataway, NJ). Purity was as follows: TEWETGQI (95%); PETLGHEI (97.4%); YEIVAGYI (99.40%); TPTAGLVGF (98.6%); GSRNGNDRL (97.1%); ESKSGDAPL (96.1%); HEHNLFGI (98.7%); ESSTP- TAGL (99.1%); ESEPKRPNM (98.7%); VSWGEPKSL (99.2%); YSDGALHLL (97.3%); AESWPSIV (96.5%); and RPNMSRHLF (99.4%). By comparing the specificity of CD8+ T cells of homozygous and heterozygous mouse strains, we observed that the immunodomi- nance that occurs during experimental T. cruzi infection could be exerted not only on epitopes restricted by the same MHC molecules but also, unexpectedly, on the immune response to epitopes restricted by different MHC-I molecules. This phenomenon, termed cross-competition, represents a potent means by which T cells with a certain specificity may become immunodominant [30– 34]. This strong and unusual phenomenon has been shown to be due to T. cruzi infection because following immunization with recombinant adenovirus expressing the same parasite antigens, this pattern of immunodominance was not observed [15]. In the second reaction, they were as follows: Based on these observations, we hypothesized that this compe- tition/immunodomination between T cells of different speci- ficities could be a sophisticated strategy that T. cruzi developed to reduce the breath and magnitude of CD8+ T-cell responses, suppressing the immune responses of these T cells with other specificities in order to escape complete elimination by host effector cells. Thus, we expected that artificially broadening the immune response to include T cells specific for subdominant or cryptic epitopes could favor the host, counteracting the restriction imposed by the infection. Here, we tested this hypothesis by using mice genetically immunized with a mutated form of the amastigote surface protein (asp) - 2 gene in which the immunodominant CD8 T cell epitope is no longer functional. The CD8 T cell-mediated immune response of these mice was directed only to the newly described subdominant/cryptic CD8 T cell epitopes of ASP-2. Even in the absence of an immune response directed to the immunodominant epitope, these mice displayed a significant degree of protective immunity, albeit not as strong as the immune response elicited by the original gene expressing both the immunodominant and the subdominants epitopes. These results are compatible with our hypothesis that artificially broadening the immune response favors the host. Indirectly, we suggest that immunodominance may in fact be a mechanism to establish a chronic infection. In the second reaction, they were as follows: i) 59-GATCCCCGCATCACCGCATGGGAGACGGGA- CAAGCACTCATGATCGTTC-39 ii) 59-GGGTCTAGATCAGACCATTTTTAGTTCACC-39. PCR products were purified, mixed and subjected to a third PCR reaction. Forward and reverse oligonucleotides were respectively PLoS ONE \| www.plosone.org Recombinant plasmids and adenoviruses Plasmid pIgSPCl.9 and the human replication-defective adeno- virus type 5 containing the asp-2 gene were obtained as described previously [35,36]. Mutated asp-2 was generated by a series of PCR reactions using DNA encoding the asp-2 clone 9 gene as a template (Genbank Accession Number: AY186572). In the first reaction, the forward and reverse oligonucleotides were as follows: i) 59-GGGGGTACCATGCTCTCACGTGTTGCT-39; i) 59-GGGGGTACCATGCTCTCACGTGTTGCT-39; ii) 59-GAACGATCATGAGTGCTTGGCCCGTCTCC- CATGCGGTGATGCGGGGATC-39 In the second reaction, they were as follows: Materials and Methods Heterologous prime-boost immunization consisted of priming i.m. with a total of 100 mg of plasmid DNA followed by a dose of viral suspension containing 26108 plaque forming units (pfu) of adenovirus twenty-one days later in the same locations. Immuno- logical assays or challenges were performed 14 days after viral inoculation. Introduction pathology caused by T. cruzi that may occur years after the initial infection in ,30% of infected individuals [6–11]. Thus, understanding how the parasites escape the immune response and persist for such long periods may help us to find new means for interventions against Chagas disease that would improve quality of life for millions of infected individuals in Latin America Recent studies on the CD8+ T-cell immune responses that occur during experimental T. cruzi infection in inbred mouse strains described a surprising immunodominance of certain epitopes expressed by members of a large family of T. cruzi surface antigens named trans-sialidases (TS) [1,5,12–20]. How and why this strong pattern of immunodominance is established is still a matter of debate. In general terms, immunodominance can emerge as a pathology caused by T. cruzi that may occur years after the initial infection in ,30% of infected individuals [6–11]. Thus, understanding how the parasites escape the immune response and persist for such long periods may help us to find new means for interventions against Chagas disease that would improve quality of life for millions of infected individuals in Latin America MHC class Ia-restricted CD8+ T cells are important mediators of the adaptive immune response against infections caused by intracellular microorganisms, including the digenetic intracellular protozoan parasite Trypanosoma cruzi, the causative agent of Chagas disease (American trypanosomiasis). During experimental infec- tion, this T cell subpopulation has been shown to be critical for host survival even when small doses of parasites are used in challenges [1–5]. In spite of the CD8+ T-cell mediated immune response, the parasite survives within the host and establishes a life-long chronic infection. Parasite persistence is considered one of the critical factors in the development of the complex immuno- Recent studies on the CD8+ T-cell immune responses that occur during experimental T. cruzi infection in inbred mouse strains described a surprising immunodominance of certain epitopes expressed by members of a large family of T. cruzi surface antigens named trans-sialidases (TS) [1,5,12–20]. How and why this strong pattern of immunodominance is established is still a matter of debate. In general terms, immunodominance can emerge as a PLoS ONE \| www.plosone.org July 2011 \| Volume 6 \| Issue 7 \| e22011 1 Results During experimental infection of H-2b or H-2a inbred mouse strains with parasites of the Y strain of T. cruzi, two epitopes were identified within the ASP-2 antigen represented by the VNHRFTLV or TEWETGQI peptides. They were recognized by H-2Kb- or H-2Kk-restricted CD8+ cytotoxic T cells, respec- tively [12,14,15,37,38]. ASP-2 is a member of the large family of TS surface antigens and is abundantly expressed only in the intracellular forms (amastigotes) of T. cruzi [39]. Fig. 1A and Table 1. Peptides used in the study. Table 1. Peptides used in the study. Table 1. Peptides used in the study. Peptide AA positions Predicted H-2 restriction TEWETGQI 320–327 Kk PETLGHEI 650–657 Kk YEIVAGYI 140–147 Kk HEHNLFGI 130–137 Kk AESWPSIV 121–128 Kk TPTAGLVGF 488–496 Ld RPNMSRHLF 36–44 Ld GSRNGNDRL 172–179 Ld ESKSGDAPL 69–77 Ld ESEPKRPNM 31–39 Ld ESSTPTAGL 485–493 Ld VSWGEPKSL 239–247 Ld YSDGALHLL 438–446 Ld Peptides were selected by the scores determined by programs available at the sites: http://www.syfpeithi.de/and http://www-bimas.cit.nih.gov/molbio/ hla_bind/. Putative anchor residues are in bold and underlined. doi:10.1371/journal.pone.0022011.t001 Mice and parasites Female 8-week-old H-2a mice (B10.A and A/Sn) were purchased from CEDEME (Federal University of Sa˜o Paulo). Bloodstream trypomastigotes of the Y strain of T. cruzi were obtained from A/Sn mice infected 7–8 days earlier [12]. Each July 2011 \| Volume 6 \| Issue 7 \| e22011 2 Immunity to Subdominant/Cryptic Epitopes Figure 1. Structure and localization of Amastigote Surface Protein-2 (ASP-2). A- Schematic view of the primary structure of T. cruzi ASP-2. B- HeLa cells were infected for 48 h with trypomastigotes of the Y strain. After fixation, indirect immunofluorescence or DAPI staining were performed as described using MAb K22 and imaged using fluorescence microscopy [39]. Bar, 14 mM. doi:10.1371/journal.pone.0022011.g001 Figure 1. Structure and localization of Amastigote Surface Protein-2 (ASP-2). A- Schematic view of the primary structure of T. cruzi ASP-2. B- HeLa cells were infected for 48 h with trypomastigotes of the Y strain. After fixation, indirect immunofluorescence or DAPI staining were performed as described using MAb K22 and imaged using fluorescence microscopy [39]. Bar, 14 mM. doi:10.1371/journal.pone.0022011.g001 before challenge was more than 95% in anti-CD8 treated mice compared to Rat IgG treated ones. before challenge was more than 95% in anti-CD8 treated mice compared to Rat IgG treated ones. (http://faculty.vassar.edu/lowry/VassarStats.html). The LogRank test was used to compare mouse survival rates after challenge with T. cruzi (http://bioinf.wehi.edu.au/software/russell/logrank/). The differences were considered significant when the P value was ,0.05. Immunological T cell assays g y Ex vivo ELISPOT (IFN-c) or in vivo cytotoxic assays were performed exactly as described previously [12,15]. The surface mobilization of CD107a and the intracellular expression of cytokines (IFN-c, TNF-a, IL-2 and IL-10) was evaluated after in vitro culture of splenocytes in the presence or absence of antigenic stimulus. Cells were washed three times in plain RPMI and re- suspended in cell culture medium consisting of RPMI 1640 medium, pH 7.4, supplemented with 10 mM Hepes, 0.2% sodium bicarbonate, 59 mg/l of penicillin, 133 mg/l of streptomycin, and 10% Hyclone fetal bovine sera (Hyclone, Logan, Utah). The viability of the cells was evaluated using 0.2% Trypan Blue exclusion dye to discriminate between live and dead cells. Cell concentration was adjusted to 56106 cells/mL in cell culture medium containing anti-CD28 (2 mg/mL), Brefeldin A (10 mg/ mL), Monensin (5 mg/mL) and FITC-labeled anti-CD107a (Clone 1D4B, 2 mg/mL, BD Pharmingen). In half of the cultures, a final concentration of 10 mM of the VNHRFTLV peptide was added. The cells were cultivated in flat-bottom 96-well plates (Corning) in a final volume of 200 ml in duplicate, at 37uC in a humid environment. After a 20-h incubation, cells were stained for surface markers with Per-CP or PE-labeled anti-CD8 on ice for 20 min. To detect IFN-c, TNF-a? IL-2 and IL-10 by intra-cellular staining (ICS), cells were then washed twice in buffer containing PBS, 0.5% BSA and 2 mM EDTA, fixed in 4% PBS-paraformal- dehyde solution for 10 minutes and permeabilized for 15 minutes in a PBS, 0.1% BSA, 0.1% saponin solution. After being washed twice, cells were stained for intracellular markers using APC or PE-labeled anti-IFN-c (Clone XMG1.2) and? PE- labeled anti- TNF-a (clone MP6-XT22), APC-labeled anti-IL-2 (clone JES6- 5H4) or APC-labeled anti-IL-10 (JES5-16E3) for 20 minutes on ice. Finally, cells were washed twice and fixed in 1% PBS- paraformaldehyde. At least 300,000 cells were acquired on a BD FacsCanto flow cytometer and then analyzed with FlowJo. PLoS ONE \| www.plosone.org Statistical analysis at 0 and 3 weeks, respectively. Control mice were naı¨ve or injected with pcDNA3 followed by pcDNA3/Adb-gal. A- Two weeks after infection or the final immunizing dose, splenic cells were re-stimulated in vitro in the presence of medium or the indicated peptides at a final concentration of 10 mM. The number of splenic IFN-c spot-forming cells (SFC) was estimated by ex vivo ELISPOT assay. B- In We showed previously that the strong pattern of immunodo- minance observed following infection with T. cruzi was not duplicated in mice genetically immunized with a recombinant adenovirus expressing ASP-2 (AdASP-2, ref. 15). Then, we determined whether genetically immunized H-2a mice could present a different pattern of immunodominance. We used a genetic immunization approach which consisted of a heterologous prime-boost regimen using plasmid DNA followed by a recom- binant adenovirus both containing the same asp-2 gene. This protocol provided strong and long lasting protective immunity against experimental infection mediated by CD8+ T cells [37,38]. In these experiments, we used H-2a mice of the A/Sn strain. These mice are highly susceptible to infection with T. cruzi, allowing us to perform protective immunity studies [37,38]. Nevertheless, it is important to mention that the results were similar when we used B10.A mice. After ex vivo stimulation with our synthetic peptides, as expected, IFN-c producing cells were detected following stimulation with the TEWETGQI peptide. In addition to this epitope, two other epitopes (PETLGHEI and YEIVAGYI) induced IFN-c produc- tion by immune cells (Fig. 2A). These peptide-specific IFN-c producing cells were CD8+ T cells as determined by simultaneous staining of intra-cellular IFN-c and the surface marker CD8 (see below). These peptides were recognized by cytotoxic cells in H-2a mice as determined by in vivo cytotoxicity assays using target cells coated with each of these peptides. B10.A mice infected with T. cruzi developed strong in vivo cytotoxicity against target cells coated with the peptide TEWETGQI. In contrast, very limited (if any) in vivo cytotoxicity was observed against target cells coated with peptides PETLGHEI and YEIVAGYI. These results were not due to different kinetics of the immune response because we observed the same results 14 or 28 days after an infectious challenge (Fig. 2B). However, A/Sn (H-2a) mice genetically vaccinated with a heterologous prime-boost vaccination regimen displayed easily detectable in vivo cytotoxic activity against target cells coated with any of these three peptides. Statistical analysis Values were expressed as means 6 SD. These values were compared using one-way ANOVA followed by Tukey’s HSD tests PLoS ONE \| www.plosone.org July 2011 \| Volume 6 \| Issue 7 \| e22011 3 Immunity to Subdominant/Cryptic Epitopes Figure 2. CD8 T-cell epitope identification during immune responses of H-2a mice infected with T. cruzi or genetically vaccinated with asp-2 gene. B10.A mice were infected with 104 T. cruzi blood parasites. A/S mice were immunized with pIgSPCl.9 followed by AdASP-2 and injected i.m. at 0 and 3 weeks, respectively. Control mice were naı¨ve or injected with pcDNA3 followed by pcDNA3/Adb-gal. A- Two weeks after infection or the final immunizing dose, splenic cells were re-stimulated in vitro in the presence of medium or the indicated tid t fi l t ti f 10 M Th b f l i IFN vivo cytotoxic activity was estimated by injecting each mouse with syngeneic CFSE-labeled splenic cells coated with or without 2 mM of the indicated peptide. Results are expressed as means 6 SD of 4 mice per group and are representative of experiments performed at least twice with similar results. Asterisks denote that the number of SFC or in vivo cytotoxicity were significantly higher when compared to naı¨ve or pcDNA3/Adb-gal injected mice (P,0.01, one-way ANOVA). doi:10.1371/journal.pone.0022011.g002 Fig. 1B show, respectively, some of the structural features of ASP-2 antigen and its expression by amastigotes. Fig. 1B show, respectively, some of the structural features of ASP-2 antigen and its expression by amastigotes. To test whether the H-2Kk-restricted epitope TEWETGQI could be an immunodominant epitope, a series of synthetic peptides containing the predicted AA anchor motif for binding to the H-2Kk or H-2Ld alleles of mouse MHC haplotype H-2a was employed. No epitope was predicted to bind to H-2Dd (Table 1). After infection of H-2a mice, antigen-specific IFN-c producing cells could only be detected in the presence of the peptide TEWETGQI (Fig. 2A). We concluded that this epitope was the immunodominant epitope of ASP-2 during infection of H-2a mice with the Y strain of T. cruzi. H-2a mice used in this experiment were B10.A because they are resistant to infection with T. cruzi. Figure 2. CD8 T-cell epitope identification during immune responses of H-2a mice infected with T. cruzi or genetically vaccinated with asp-2 gene. B10.A mice were infected with 104 T. cruzi blood parasites. A/S mice were immunized with pIgSPCl.9 followed by AdASP-2 and injected i.m. Statistical analysis The elimination of target cells coated with peptide TEWETGQI was always stronger than the two others, suggesting a pattern of immunodominance. Figure 2. CD8 T-cell epitope identification during immune responses of H-2a mice infected with T. cruzi or genetically vaccinated with asp-2 gene. B10.A mice were infected with 104 T. cruzi blood parasites. A/S mice were immunized with pIgSPCl.9 followed by AdASP-2 and injected i.m. at 0 and 3 weeks, respectively. Control mice were naı¨ve or injected with pcDNA3 followed by pcDNA3/Adb-gal. A- Two weeks after infection or the final immunizing dose, splenic cells were re-stimulated in vitro in the presence of medium or the indicated peptides at a final concentration of 10 mM. The number of splenic IFN-c spot-forming cells (SFC) was estimated by ex vivo ELISPOT assay. B- In To determine whether other cytokines and/or effector mole- cules could be secreted by peptide-specific T cells, we performed staining to detect surface mobilization of CD107a (a marker for exocytosis) or intra-cellular accumulation of IFN-c, TNF-a, IL-2 July 2011 \| Volume 6 \| Issue 7 \| e22011 PLoS ONE \| www.plosone.org 4 Immunity to Subdominant/Cryptic Epitopes Figure 3. Surface mobilization of CD107a and expression of IFN-c, TNF-a, IL-2 or IL-10 by specific CD8+ T cells from B10.A mice infected with T. cruzi or A/Sn mice immunized with pIgSCl.9/AdASP-2 vaccine. B10.A or A/Sn mice were infected or immunized as described in the legend of Fig. 2. Control mice were either naive mice or mice immunized with pCDNA3/Adb-gal. Twenty one or fourteen days after infection or immunization, respectively, these mice had their splenic cells cultured in the presence of anti-CD107a and anti-CD28, with or without the peptides TEWETGQI, PETLGHEI or YEIVAGYI. After 12 h, cells were stained for CD8, IFN-c, TNF-a, IL-2 and IL-10. Representative analyses (medians) are shown from four mice performed per experiment. A) Example of splenic CD8+ cells from B10.A naive mice cultivated in vitro in the presence of medium alone (Medium) or with the indicated peptides and stained for expression of IFN-c and TNF-a. B, C and D) Examples of splenic CD8+ cells from B10.A infected mice cultivated in vitro in the presence of medium alone (Medium) or with the indicated peptides and stained for expression of: B) IFN-c and TNF-a; C) IFN-c and IL-2; D) IFN-c and IL-10. Statistical analysis The expression of these cytokines was dependent on the infection because they were not detected in cells from naive mice (Fig. 3A). In contrast, a relatively low frequency of CD8+ splenic cells stimulated with peptides PETLGHEI or YEIVAGYI mobilized CD107a to the surface (data not shown) or accumulated intra- cellular IFN-c or TNF-a (Fig. 3B). We were also unable to detect the presence of significant numbers of IL-2 or IL-10 expressing cells in these same cells (Fig. 3C and 3D). Splenic cells from A/Sn mice genetically vaccinated with heterologous prime-boost regimen were stimulated with peptides TEWETGQI, PETLGHEI or YEIVAGYI. The results shows that a large fraction of the CD8+ cells of pIgSPCl.9/AdASP-2 immunized mice at the same time mobilize CD107a to the surface and expressed intra-cellular IFN-c and TNF-a (Fig. 3G to 3J). These cells were therefore multifunctional CD8+ T cells as we have previously described [37]. We were unable to detect the presence of intra-cellular IL-2 or IL-10 in these same cells (data not shown). These results confirmed and extended the ones described in Fig. 2). After challenge, mice immunized with plasmids containing the asp-2 gene (mutated or not) presented significantly lower parasitemia than control mice immunized with pcDNA3 (Fig. 4C). Although the levels of parasitemia were not statistically different when compared to mice vaccinated with the plasmid containing the mutated asp-2 gene, the mortality of these animals was significantly faster (Fig. 4D). We therefore concluded that broadening the CD8+ T cell immune response by vaccination with a plasmid containing epitopes that elicit immune responses to subdominant/cryptic epitopes of ASP-2 could provide some degree of protective immunity. Nevertheless, because protective immunity elicited by vaccination with pIgSPTAWETGQA was not as efficient, the presence of a functional immunodominant epitope was clearly important for effective protective immunity. The description of these two new epitopes allowed us to test whether immunity to the subdominant/cryptic epitope could participate during protective immunity against T. cruzi infection, a phenomenon that has not previously been tested experimentally. For this purpose, we generated a plasmid DNA and a recombinant adenovirus containing a mutated form of the asp-2 gene in which we modified the nucleotides encoding the anchor residues required for the immunodominant TEWETGQI epitope to bind to the H- 2Kk molecule. The mutated gene expressed the AA sequence TAWETGQA, where the alanines (A) replaced glutamic acid (E) or isoleucine (I) of the original epitope. Statistical analysis I and J) Determination of multifunctional CD8+ cells from mice immunized with pIgSPCl.9/ AdASP-2 cultivated in vitro in the presence of the indicated peptides and stained for surface mobilization of CD107a and expression IFN-c and TNF-a. doi:10.1371/journal.pone.0022011.g003 vitro in the presence of medium alone (Medium) or with the indicated peptides and stained for surface mobilization of CD107a and expression IFN-c (panel G) and expression IFN-c and TNF-a (panel H). I and J) Determination of multifunctional CD8+ cells from mice immunized with pIgSPCl.9/ AdASP-2 cultivated in vitro in the presence of the indicated peptides and stained for surface mobilization of CD107a and expression IFN-c and TNF-a. doi:10.1371/journal.pone.0022011.g003 with the immunodominant epitope TEWETGQI or with the subdominant/cryptic epitopes PETLGHEI or YEIVAGYI two weeks after challenge with T. cruzi. We chose this protocol because, as the immune response following plasmid DNA vaccination is usually low, it is easier to detect the anamnestic immune responses after challenge [12]. We observed that all mice immunized with asp-2 genes (mutated or not) presented specific IFN-c producing cells when stimulated with subdominant/cryptic epitopes (PETL- GHEI or YEIVAGYI, Fig. 4A). The immune response was specific because mice immunized with control plasmid pcDNA3 failed to recognize these peptides. However, we detected IFN-c producing cells specific to TEWETGQI only in mice immunized with plasmid pIgSPCl.9. It is noteworthy that the number of cells detected in mice immunized with pIgSPTAWETGQA was similar to the number of cells in pcDNA3 injected mice (Fig. 4A). These numbers reflect cells primed during infection. Analysis of in vivo cytotoxic activity also demonstrated that in mice immunized with the plasmid pIgSPTAWETGQA, the response to the immunodo- minant epitope TEWETGQI was not different from control mice immunized with pcDNA3 (Fig. 4B). These immunological analyses demonstrated that mice immunized with pIgSPTAWETGQA indeed lost the functional immunodominant TEWETGQI response but had an unaltered ability to elicit immune responses to the subdominant epitopes PETLGHEI and YEIVAGYI. with the immunodominant epitope TEWETGQI or with the subdominant/cryptic epitopes PETLGHEI or YEIVAGYI two weeks after challenge with T. cruzi. We chose this protocol because, as the immune response following plasmid DNA vaccination is usually low, it is easier to detect the anamnestic immune responses after challenge [12]. We observed that all mice immunized with asp-2 genes (mutated or not) presented specific IFN-c producing cells when stimulated with subdominant/cryptic epitopes (PETL- GHEI or YEIVAGYI, Fig. 4A). Statistical analysis E and F) Examples of splenic CD8+ cells from mice immunized with pcDNA3/Adb-gal cultivated in vitro in the presence of medium alone (Medium) or with the indicated peptides and stained for surface mobilization of CD107a and expression IFN-c (pane E) or expression of IFN-c and TNF-a (panel F). G and H) Examples of splenic CD8+ cells from mice immunized with pIgSPCl.9/AdASP-2 cultivated in Immunity to Subdominant/Cryptic Epitopes Figure 3. Surface mobilization of CD107a and expression of IFN-c, TNF-a, IL-2 or IL-10 by specific CD8+ T cells from B10.A mice infected with T. cruzi or A/Sn mice immunized with pIgSCl.9/AdASP-2 vaccine. B10.A or A/Sn mice were infected or immunized as described in the legend of Fig. 2. Control mice were either naive mice or mice immunized with pCDNA3/Adb-gal. Twenty one or fourteen days after infection or immunization, respectively, these mice had their splenic cells cultured in the presence of anti-CD107a and anti-CD28, with or without the peptides TEWETGQI, PETLGHEI or YEIVAGYI. After 12 h, cells were stained for CD8, IFN-c, TNF-a, IL-2 and IL-10. Representative analyses (medians) are shown from four mice performed per experiment. A) Example of splenic CD8+ cells from B10.A naive mice cultivated in vitro in the presence of medium alone (Medium) or with the indicated peptides and stained for expression of IFN-c and TNF-a. B, C and D) Examples of splenic CD8+ cells from B10.A infected mice cultivated in vitro in the presence of medium alone (Medium) or with the indicated peptides and stained for expression of: B) IFN-c and TNF-a; C) IFN-c and IL-2; D) IFN-c and IL-10. E and F) Examples of splenic CD8+ cells from mice immunized with pcDNA3/Adb-gal cultivated in vitro in the presence of medium alone (Medium) or with the indicated peptides and stained for surface mobilization of CD107a and expression IFN-c (panel E) or expression of IFN-c and TNF-a (panel F). G and H) Examples of splenic CD8+ cells from mice immunized with pIgSPCl.9/AdASP-2 cultivated in July 2011 \| Volume 6 \| Issue 7 \| e22011 PLoS ONE \| www.plosone.org 5 Immunity to Subdominant/Cryptic Epitopes Immunity to Subdominant/Cryptic Epitopes vitro in the presence of medium alone (Medium) or with the indicated peptides and stained for surface mobilization of CD107a and expression IFN-c (panel G) and expression IFN-c and TNF-a (panel H). Statistical analysis The immune response was specific because mice immunized with control plasmid pcDNA3 failed to recognize these peptides. However, we detected IFN-c producing cells specific to TEWETGQI only in mice immunized with plasmid pIgSPCl.9. It is noteworthy that the number of cells detected in mice immunized with pIgSPTAWETGQA was similar to the number of cells in pcDNA3 injected mice (Fig. 4A). These numbers reflect cells primed during infection. Analysis of in vivo cytotoxic activity also demonstrated that in mice immunized with the plasmid pIgSPTAWETGQA, the response to the immunodo- minant epitope TEWETGQI was not different from control mice immunized with pcDNA3 (Fig. 4B). These immunological analyses demonstrated that mice immunized with pIgSPTAWETGQA indeed lost the functional immunodominant TEWETGQI response but had an unaltered ability to elicit immune responses to the subdominant epitopes PETLGHEI and YEIVAGYI. or IL-10. When we used splenic cells from B10.A mice infected with T. cruzi, we observed that upon stimulation with TE- WETGQI, a large fraction of CD8+ cells mobilize CD107a to the surface (data not shown) and accumulate intra-cellular IFN-c and TNF-a (Fig. 3B). These CD8+ cells were therefore multifunctional CD107a+IFN-c+TNF-a+ (,35%) or IFN-c+TNF-a+ (,40%). We were unable to detect the presence of significant numbers of IL-2 or IL-10 expressing cells in these same samples (Fig. 3C and 3D). The expression of these cytokines was dependent on the infection because they were not detected in cells from naive mice (Fig. 3A). In contrast, a relatively low frequency of CD8+ splenic cells stimulated with peptides PETLGHEI or YEIVAGYI mobilized CD107a to the surface (data not shown) or accumulated intra- cellular IFN-c or TNF-a (Fig. 3B). We were also unable to detect the presence of significant numbers of IL-2 or IL-10 expressing cells in these same cells (Fig. 3C and 3D). S l i ll f A/S i i ll i d i h or IL-10. When we used splenic cells from B10.A mice infected with T. cruzi, we observed that upon stimulation with TE- WETGQI, a large fraction of CD8+ cells mobilize CD107a to the surface (data not shown) and accumulate intra-cellular IFN-c and TNF-a (Fig. 3B). These CD8+ cells were therefore multifunctional CD107a+IFN-c+TNF-a+ (,35%) or IFN-c+TNF-a+ (,40%). We were unable to detect the presence of significant numbers of IL-2 or IL-10 expressing cells in these same samples (Fig. 3C and 3D). Statistical analysis In preliminary experi- ments, we observed that the synthetic peptide TAWETGQA was not recognized by immune cells from genetically vaccinated H-2a mice (data not shown). Details of the plasmid and recombinant adenovirus containing the mutated form of the asp-2 gene are shown in Table 2. We then sought to test the same hypotheses described above using a distinct approach. For that purpose, H-2a mice were primed with plasmid pIgSPCl.9 followed by a booster immuniza- tion with AdASP-2 (heterologous prime- boost regimen). Alterna- tively, mice were primed with plasmid pIgTAWETGQA followed by a booster immunization with AdTAWETGQA. We consider this approach complementary to the one described above because plasmid or adenovirus may used distinct routes for stimulating CD8+ T cells. Initially, we genetically immunized mice with plasmids containing the original gene (pIgSPCl.9), the mutated gene (pIgSPTAWETGQA) or both plasmids simultaneously. Immune responses were estimated by ELISPOT after ex vivo stimulation Immune responses were estimated 14 days after the booster immunization by ELISPOT following ex vivo stimulation with synthetic peptides encoding the immunodominant epitope TE- WETGQI, the subdominant/cryptic epitopes PETLGHEI or YEIVAGYI or the mutated epitope TAWETGQA. We chose this protocol because the immune responses following heterologous prime boost immunization generates strong immune responses that can be easily detected after boosting [37]. We observed that mice immunized with asp-2 genes (mutated or not) had specific IFN-c producing cells when stimulated with peptides PELTHGEI or YEIVAGYI. The fact that these responses were of similar magnitude strongly argued that the expression/immunogenicity of Table 2. Genetic vectors used in the study. Table 2. Genetic vectors used in the study. Table 2. Genetic vectors used in the study. Designation Vector ASP-2 CD8 epitopes pIgSPCl.9 Plasmid TEWETGQI PETLGHEI YEIVAGYI pIgSPTAWETGQA.9 Plasmid TAWETGQA PETLGHEI YEIVAGYI AdASP-2 Adenovirus TEWETGQI PETLGHEI YEIVAGYI AdTAWETGQA Adenovirus TAWETGQA PETLGHEI YEIVAGYI doi:10.1371/journal.pone.0022011.t002 doi:10.1371/journal.pone.0022011.t002 PLoS ONE \| www.plosone.org July 2011 \| Volume 6 \| Issue 7 \| e22011 July 2011 \| Volume 6 \| Issue 7 \| e22011 6 Immunity to Subdominant/Cryptic Epitopes Figure 4. CD8 immune responses and trypomastigote-induced parasitemia and mortality in A/Sn mice genetically immunized with different plasmid DNA containing the asp-2 gene. A/Sn mice were immunized with pcDNA3, pIgSPCl.9, pIgSPTAWETGQA or simultaneously with the last two (pIgSPCl.9 and pIgSPTAWETGQA). Immunization consisted of 3 doses of 100 mg of DNA each given by the i.m. route in the tibialis anterioris three weeks apart. Statistical analysis A- Two weeks after the final immunizing dose, mice were challenged i.p. with 150 bloodstream trypomastigotes. Two weeks after infection, splenic cells were re-stimulated in vitro in the presence of medium only or the indicated peptides at a final concentration of 10 mM. The number of splenic IFN-c spot-forming cells (SFC) was estimated by ex vivo ELISPOT assay. B- In vivo cytotoxic activity was estimated by injecting each mouse with syngeneic CFSE-labeled splenic cells coated with or without 2 mM of the indicated peptide. Results are expressed as means 6 SD of 4 mice per group and are representative of experiments performed at least twice with similar results. Asterisks denote that the number of SFC, or the in vivo cytotoxicity, were significantly higher when compared to SFC found in naı¨ve or pcDNA3/Adb-gal-injected mice (P,0.01). C- Parasitemia for each mouse group is represented as mean 6 SD (n = 5–7). Asterisks denote that mice from groups immunized with pIgSPCl.9 or pIgSPTAWETGQA or both had significantly lower parasitemia (P,0.01) than animals injected with pcDNA3. The curves of parasitemia of animals immunized with pIgSPCl.9 (squares) or pIgSPTAWETGQA (triangles) are superimposed. D- Kaplan-Meier curves for the survival of mouse groups immunized and challenged as described above (n = 5–7). Mice from groups immunized with pIgSPCl.9 or pIgSPCl.9/pIgSPTAWETGQA survived significantly longer than animals injected with pcDNA3 or pIgSPTAWETGQA (P,0.05, in all cases, LogRank test). Mice immunized with pIgSPTAWETGQA also survived longer than animals injected with pcDNA3 (P,0.05). No animals died after the 30th day until they were euthanized. Results are representative of two independent experiments. doi:10.1371/journal.pone.0022011.g004 Figure 4. CD8 immune responses and trypomastigote-induced parasitemia and mortality in A/Sn mice genetically immunized with different plasmid DNA containing the asp-2 gene. A/Sn mice were immunized with pcDNA3, pIgSPCl.9, pIgSPTAWETGQA or simultaneously with the last two (pIgSPCl.9 and pIgSPTAWETGQA). Immunization consisted of 3 doses of 100 mg of DNA each given by the i.m. route in the tibialis anterioris three weeks apart. A- Two weeks after the final immunizing dose, mice were challenged i.p. with 150 bloodstream trypomastigotes. Two weeks after infection, splenic cells were re-stimulated in vitro in the presence of medium only or the indicated peptides at a final concentration of 10 mM. The number of splenic IFN-c spot-forming cells (SFC) was estimated by ex vivo ELISPOT assay. Statistical analysis B- In vivo cytotoxic activity was estimated by injecting each mouse with syngeneic CFSE-labeled splenic cells coated with or without 2 mM of the indicated peptide. Results are expressed as means 6 SD of 4 mice per group and are representative of experiments performed at least twice with similar results. Asterisks denote that the number of SFC, or the in vivo cytotoxicity, were significantly higher when compared to SFC found in naı¨ve or pcDNA3/Adb-gal-injected mice (P,0.01). C- Parasitemia for each mouse group is represented as mean 6 SD (n = 5–7). Asterisks denote that mice from groups immunized with pIgSPCl.9 or pIgSPTAWETGQA or both had significantly lower parasitemia (P,0.01) than animals injected with pcDNA3. The curves of parasitemia of animals immunized with pIgSPCl.9 (squares) or pIgSPTAWETGQA (triangles) are superimposed. D- Kaplan-Meier curves for the survival of mouse groups immunized and challenged as described above (n = 5–7). Mice from groups immunized with pIgSPCl.9 or pIgSPCl.9/pIgSPTAWETGQA survived significantly longer than animals injected with pcDNA3 or pIgSPTAWETGQA (P,0.05, in all cases, LogRank test). Mice immunized with pIgSPTAWETGQA also survived longer than animals injected with pcDNA3 (P,0.05). No animals died after the 30th day until they were euthanized. Results are representative of two independent experiments. doi:10.1371/journal.pone.0022011.g004 Figure 4. CD8 immune responses and trypomastigote-induced parasitemia and mortality in A/Sn mice genetically immunized with different plasmid DNA containing the asp-2 gene. A/Sn mice were immunized with pcDNA3, pIgSPCl.9, pIgSPTAWETGQA or simultaneously with the last two (pIgSPCl.9 and pIgSPTAWETGQA). Immunization consisted of 3 doses of 100 mg of DNA each given by the i.m. route in the tibialis anterioris three weeks apart. A- Two weeks after the final immunizing dose, mice were challenged i.p. with 150 bloodstream trypomastigotes. Two weeks after infection, splenic cells were re-stimulated in vitro in the presence of medium only or the indicated peptides at a final concentration of 10 mM. The number of splenic IFN-c spot-forming cells (SFC) was estimated by ex vivo ELISPOT assay. B- In vivo cytotoxic activity was estimated by injecting each mouse with syngeneic CFSE-labeled splenic cells coated with or without 2 mM of the indicated peptide. Results are expressed as means 6 SD of 4 mice per group and are representative of experiments performed at least twice with similar results. Asterisks denote that the number of SFC, or the in vivo cytotoxicity, were significantly higher when compared to SFC found in naı¨ve or pcDNA3/Adb-gal-injected mice (P,0.01). Statistical analysis Results are expressed as mean 6 SD of 4 mice per group and are representative of experiments performed at least twice with similar results. Asterisks denote that the number of SFC or in vivo cytotoxicity were significantly higher when compared to SFC found in naı¨ve or pcDNA3/Adb-gal injected mice (P,0.01). C- Fourteen days after the last dose, these mice had their splenic cells cultured in the presence of anti-CD28 and Medium or the indicated peptides. After 12 h, cells were stained for CD8, IFN-c and TNF-a. Examples of splenic CD8+ cells from immunized mice. Representative analyses (medians) are shown from four mice performed per experiment. doi:10.1371/journal.pone.0022011.g005 Figure 5. CD8 immune responses in A/Sn mice immunized with asp-2 using the he vaccination regimen. A/Sn mice were primed i.m. with 100 mg of plasmids pcDNA3, pIgSPCl.9 or were boosted i.m. with 26108 pfu Adb-gal, AdASP-2 or AdTAWETGQA. A- Two weeks after the last do presence of medium only or the indicated peptides at a final concentration of 10 mM. The numb estimated by ex vivo ELISPOT assay. B- In vivo cytotoxic activity was estimated by injecting each m coated with or without 2 mM of the indicated peptide. Results are expressed as mean 6 SD o experiments performed at least twice with similar results. Asterisks denote that the number of SFC when compared to SFC found in naı¨ve or pcDNA3/Adb-gal injected mice (P,0.01). C- Fourteen day cells cultured in the presence of anti-CD28 and Medium or the indicated peptides. After 12 h, cells w of splenic CD8+ cells from immunized mice. Representative analyses (medians) are shown from fo doi:10 1371/journal pone 0022011 g005 Figure 5. CD8 immune responses in A/Sn mice immunized with asp-2 using the heterologous DNA prime-adenovirus boost vaccination regimen. A/Sn mice were primed i.m. with 100 mg of plasmids pcDNA3, pIgSPCl.9 or pIgSPTAWETGQA. Three weeks later, these mice were boosted i.m. with 26108 pfu Adb-gal, AdASP-2 or AdTAWETGQA. A- Two weeks after the last dose, splenic cells were re-stimulated in vitro in the presence of medium only or the indicated peptides at a final concentration of 10 mM. The number of splenic IFN-c spot forming cells (SFC) was estimated by ex vivo ELISPOT assay. B- In vivo cytotoxic activity was estimated by injecting each mouse with syngeneic CFSE-labeled splenic cells coated with or without 2 mM of the indicated peptide. Statistical analysis C- Parasitemia for each mouse group is represented as mean 6 SD (n = 5–7). Asterisks denote that mice from groups immunized with pIgSPCl.9 or pIgSPTAWETGQA or both had significantly lower parasitemia (P,0.01) than animals injected with pcDNA3. The curves of parasitemia of animals immunized with pIgSPCl.9 (squares) or pIgSPTAWETGQA (triangles) are superimposed. D- Kaplan-Meier curves for the survival of mouse groups immunized and challenged as described above (n = 5–7). Mice from groups immunized with pIgSPCl.9 or pIgSPCl.9/pIgSPTAWETGQA survived significantly longer than animals injected with pcDNA3 or pIgSPTAWETGQA (P,0.05, in all cases, LogRank test). Mice immunized with pIgSPTAWETGQA also survived longer than animals injected with pcDNA3 (P,0.05). No animals died after the 30th day until they were euthanized. Results are representative of two independent experiments. doi:10.1371/journal.pone.0022011.g004 performed intra-cellular cytokine staining analysis for IFN-c and TNF-a after in vitro peptide stimulation. As depicted in Fig. 5C, we detected CD8+ cells expressing IFN-c and/or TNF-a specific for the TEWETGQI epitope only in mice immunized with pIgSPCl.9/AdASP-2 (Fig. 5C). Similar analyses performed 14 days after an infectious challenge with T. cruzi exhibited the same both genes/antigens were very similar. We could detect IFN-c producing cells specific for the TEWETGQI epitope only in mice immunized with pIgSPCl.9/AdASP-2. In contrast, none of the immunized mice presented IFN-c producing cells when stimulated with the TAWETGQA peptide (Fig. 5A). The analysis of in vivo cytotoxic activity showed a similar picture (Fig. 5B). We also PL July 2011 \| Volume 6 \| Issue 7 \| e22011 PLoS ONE \| www.plosone.org July 2011 \| Volume 6 \| Issue 7 \| e22011 7 Immunity to Subdominant/Cryptic Epitopes Figure 5. CD8 immune responses in A/Sn mice immunized with asp-2 using the heterologous DNA prime-adenovirus boost vaccination regimen. A/Sn mice were primed i.m. with 100 mg of plasmids pcDNA3, pIgSPCl.9 or pIgSPTAWETGQA. Three weeks later, these mice were boosted i.m. with 26108 pfu Adb-gal, AdASP-2 or AdTAWETGQA. A- Two weeks after the last dose, splenic cells were re-stimulated in vitro in the presence of medium only or the indicated peptides at a final concentration of 10 mM. The number of splenic IFN-c spot forming cells (SFC) was estimated by ex vivo ELISPOT assay. B- In vivo cytotoxic activity was estimated by injecting each mouse with syngeneic CFSE-labeled splenic cells coated with or without 2 mM of the indicated peptide. PLoS ONE \| www.plosone.org Statistical analysis Results are expressed as mean 6 SD of 4 mice per group and are representative of experiments performed at least twice with similar results. Asterisks denote that the number of SFC or in vivo cytotoxicity were significantly higher when compared to SFC found in naı¨ve or pcDNA3/Adb-gal injected mice (P,0.01). C- Fourteen days after the last dose, these mice had their splenic cells cultured in the presence of anti-CD28 and Medium or the indicated peptides. After 12 h, cells were stained for CD8, IFN-c and TNF-a. Examples of splenic CD8+ cells from immunized mice. Representative analyses (medians) are shown from four mice performed per experiment. doi:10.1371/journal.pone.0022011.g005 mice injected with pcDNA3/Adb-gal (P,0.01, Fig. 6A). Although most of the mice immunized with pIgSPTAWETGQA/Ad- TAEWTGQA died after challenge, they survived longer than control mice injected with pcDNA3/Adb-gal (Fig. 6B, P,0.01, LogRank test). In parallel, vaccinated and control mice were challenged by the s.c. route. We observed that the majority of the mice immunized with pIgSPTAWETGQA/AdTAEWTGQA survived the infectious challenge (Fig. 6D). We therefore conclu- pattern of response (data not shown). Together, these immuno- logical analyses demonstrated that heterologous prime-boost immunization with pIgSPTAWETGQA/AdTAEWTGQA failed to induce an immune response to the immunodominant epitope TEWETGQI but induced almost unaltered immune responses to the subdominant epitopes PETLGHEI or YEIVAGYI. After challenge, parasitemia in mice immunized with the asp-2 genes (mutated or not) was significantly lower than in control PLoS ONE \| www.plosone.org July 2011 \| Volume 6 \| Issue 7 \| e22011 8 Immunity to Subdominant/Cryptic Epitopes Figure 6. Trypomastigote-induced parasitemia and mortality in A/Sn mice immunized with asp-2 using the heterologous DNA prime-adenovirus boost vaccination regimen. A/Sn mice were immunized as depicted in the legend of Fig. 5. Two weeks after the final immunizing dose, mice were challenged i.p. (Panels A and B) or s.c. (Panels C and D) with 150 bloodstream trypomastigotes. Parasitemia for each mouse group is represented as mean 6 SD (n = 10 or 11). Asterisks denote that mice from groups immunized with pIgSPCl.9/AdASP-2 or pIgSPTAWETGQA/AdTAWTEGQA had significantly lower parasitemia (P,0.01) than pcDNA3/Adb-gal-injected animals. Panels B and D represent Kaplan-Meier curves for survival of the mouse groups immunized and challenged as described above (n = 10 or 11). Mice immunized with pIgSPCl.9/ AdASP-2 survived significantly longer than animals immunized with pIgSPTAWETGQA/AdTAWETGQA or pcDNA3/Adb-gal (P = 0.01 or P,0.01, respectively). Mice immunized with pIgSPTAWETGQA/AdTAWETGQA survived significantly longer than pcDNA3/Adb-gal-injected animals (P,0.01). Statistical analysis (pIgSPCl.9/AdASP-2 or pIgSPTAWETGQA/AdTAEWTGQA) and treated with rat IgG. CD8 depleted mice had their survival time reduced to the same time as control mice which were injected with pcDNA3/Adb-gal. ded that broadening the T cell immune response using a genetic vaccination that elicits immune responses to subdominant/cryptic epitopes of ASP-2 provides protective immunity against infection. Nevertheless, likewise in the case of the plasmid DNA, after an i.p. challenge, in terms of survival, immunization with only genes expressing subdominant epitopes was not as effective as immuni- zation with genes expressing both the dominant and the subdominant epitopes. PLoS ONE \| www.plosone.org Statistical analysis Results are representative of two pooled experiments. No animals died after the 40th day. doi:10.1371/journal.pone.0022011.g006 Figure 6. Trypomastigote-induced parasitemia and mortality in A/Sn mice immunized with asp-2 using the heterologous DNA prime-adenovirus boost vaccination regimen. A/Sn mice were immunized as depicted in the legend of Fig. 5. Two weeks after the final immunizing dose, mice were challenged i.p. (Panels A and B) or s.c. (Panels C and D) with 150 bloodstream trypomastigotes. Parasitemia for each mouse group is represented as mean 6 SD (n = 10 or 11). Asterisks denote that mice from groups immunized with pIgSPCl.9/AdASP-2 or pIgSPTAWETGQA/AdTAWTEGQA had significantly lower parasitemia (P,0.01) than pcDNA3/Adb-gal-injected animals. Panels B and D represent Kaplan-Meier curves for survival of the mouse groups immunized and challenged as described above (n = 10 or 11). Mice immunized with pIgSPCl.9/ AdASP-2 survived significantly longer than animals immunized with pIgSPTAWETGQA/AdTAWETGQA or pcDNA3/Adb-gal (P = 0.01 or P,0.01, respectively). Mice immunized with pIgSPTAWETGQA/AdTAWETGQA survived significantly longer than pcDNA3/Adb-gal-injected animals (P,0.01). Results are representative of two pooled experiments. No animals died after the 40th day. doi:10.1371/journal.pone.0022011.g006 Figure 6. Trypomastigote-induced parasitemia and mortality in A/Sn mice immunized with asp-2 using the heterologous DNA prime-adenovirus boost vaccination regimen. A/Sn mice were immunized as depicted in the legend of Fig. 5. Two weeks after the final immunizing dose, mice were challenged i.p. (Panels A and B) or s.c. (Panels C and D) with 150 bloodstream trypomastigotes. Parasitemia for each mouse group is represented as mean 6 SD (n = 10 or 11). Asterisks denote that mice from groups immunized with pIgSPCl.9/AdASP-2 or pIgSPTAWETGQA/AdTAWTEGQA had significantly lower parasitemia (P,0.01) than pcDNA3/Adb-gal-injected animals. Panels B and D represent Kaplan-Meier curves for survival of the mouse groups immunized and challenged as described above (n = 10 or 11). Mice immunized with pIgSPCl.9/ AdASP-2 survived significantly longer than animals immunized with pIgSPTAWETGQA/AdTAWETGQA or pcDNA3/Adb-gal (P = 0.01 or P,0.01, respectively). Mice immunized with pIgSPTAWETGQA/AdTAWETGQA survived significantly longer than pcDNA3/Adb-gal-injected animals (P,0.01). Results are representative of two pooled experiments. No animals died after the 40th day. doi:10.1371/journal.pone.0022011.g006 (pIgSPCl.9/AdASP-2 or pIgSPTAWETGQA/AdTAEWTGQA) and treated with rat IgG. CD8 depleted mice had their survival time reduced to the same time as control mice which were injected with pcDNA3/Adb-gal. (pIgSPCl.9/AdASP-2 or pIgSPTAWETGQA/AdTAEWTGQA) and treated with rat IgG. CD8 depleted mice had their survival time reduced to the same time as control mice which were injected with pcDNA3/Adb-gal. Discussion The parasitemia for each mouse group is repr that mice from groups immunized with pIgSPCl.9/AdASP-2 or pIgSPTAWETGQA/AdTAWTEGQA and parasitemia (P,0.01) than vaccinated mice treated with anti-CD8 (Panels A and B). Panels C and D re mouse groups immunized and challenged as described above (n = 6). Mice immunized with pIgSPCl Figure 7. CD8 T cell dependence of protective immunity of A/Sn mice immunized with asp-2 using the heterologous DNA prime- adenovirus boost vaccination regimen. A/Sn mice were immunized as described in the legend of Fig. 5. pIgSPTAWETGQA/AdTAWTEGQA had significantly lower parasitemia (P,0.01) than pcDNA3/Adb-gal-injected animals. Before and after challenge, mice were treated as described in Methods section with rat IgG (control) or anti-CD8 MAb. The parasitemia for each mouse group is represented as mean 6 SD (n = 6). Asterisks denote that mice from groups immunized with pIgSPCl.9/AdASP-2 or pIgSPTAWETGQA/AdTAWTEGQA and treated with Rat IgG had significantly lower parasitemia (P,0.01) than vaccinated mice treated with anti-CD8 (Panels A and B). Panels C and D represent Kaplan-Meier curves for survival of the mouse groups immunized and challenged as described above (n = 6). Mice immunized with pIgSPCl.9/AdASP-2 or pIgSPTAWETGQA/AdTAWETGQA and treated with Rat IgG survived significantly longer than vaccinated animals treated with anti-CD8 (P,0.01 in both cases). doi:10.1371/journal.pone.0022011.g007 not substitute completely for CD8+ T cells specific for the immunodominant TEWETGQI epitope. We observed that immunization with plasmids alone or in combination with a recombinant adenovirus that did not express the immunodominant epitope elicited immune responses to the subdominant/cryptic epitopes but failed to provide optimal protective immunity when compared to a plasmid that expresses both the immunodominant and subdominants epitopes (Fig. 4D, 6B and 7D). These results demonstrate that the response to the immunodominant epitope contributes to the immunity elicited by genetic vaccination and is might be required for highly efficient resistance. during infection (Fig. 2), but they also did not stimulate TNF-a, IL-2 or IL-10 secretion by CD8+ T cells (Fig. 3). The precise reason for this strong immunodominant pattern during T. cruzi infection is unknown at present. One possible explanation for this biased immune response could be to provide an advantage to the parasite by avoiding an even higher and broader immune response. This hypothesis is in agreement with earlier studies showing that immunity to subdominant epitopes can provide an important contribution to protective immunity against viral infection [34,40– 43]. Our results also corroborated this hypothesis. Discussion Here, we initially confirmed and extended our previous observation that experimental infection with the human intracel- lular pathogen T. cruzi restricted the repertoire of CD8+ T cells. While immune cells of infected H-2a mice recognized a single immunodominant epitope of ASP-2, cells from mice immunized with recombinant genetic vaccines expressing this same T. cruzi antigen recognized, in addition to the immunodominant epitope, two other subdominant/cryptic epitopes. The sub-dominant epitopes not only failed to elicit IFN-c and in vivo cytotoxicity Finally, to firmly establish that protective immunity was mediated by CD8+ T cells, we performed in vivo depletion experi- ments in mice vaccinated with the pIgSPCl.9/AdASP-2 or pIgSPTAWETGQA/AdTAEWTGQA. Treatment with anti- CD8 MAb renders these mice more susceptible to infection. CD8 depleted mice presented higher parasitemia (Fig. 7A and B) and shorter survival times (Fig. 7C and D) when compared to A/ Sn mice vaccinated with heterologous prime-boost regimen July 2011 \| Volume 6 \| Issue 7 \| e22011 9 Immunity to Subdominant/Cryptic Epitopes Figure 7. CD8 T cell dependence of protective immunity of A/Sn mice immunized with asp-2 using the heterologous DNA prime- adenovirus boost vaccination regimen. A/Sn mice were immunized as described in the legend of Fig. 5. pIgSPTAWETGQA/AdTAWTEGQA had significantly lower parasitemia (P,0.01) than pcDNA3/Adb-gal-injected animals. Before and after challenge, mice were treated as described in Methods section with rat IgG (control) or anti-CD8 MAb. The parasitemia for each mouse group is represented as mean 6 SD (n = 6). Asterisks denote that mice from groups immunized with pIgSPCl.9/AdASP-2 or pIgSPTAWETGQA/AdTAWTEGQA and treated with Rat IgG had significantly lower parasitemia (P,0.01) than vaccinated mice treated with anti-CD8 (Panels A and B). Panels C and D represent Kaplan-Meier curves for survival of the mouse groups immunized and challenged as described above (n = 6). Mice immunized with pIgSPCl.9/AdASP-2 or pIgSPTAWETGQA/AdTAWETGQA and treated with Rat IgG survived significantly longer than vaccinated animals treated with anti-CD8 (P,0.01 in both cases). doi:10.1371/journal.pone.0022011.g007 Immunity to Subdominant/Cryptic Epitopes Figure 7. CD8 T cell dependence of protective immunity of A/Sn mice immunized with a adenovirus boost vaccination regimen. A/Sn mice were immunized as described in the legend significantly lower parasitemia (P,0.01) than pcDNA3/Adb-gal-injected animals. Before and after Methods section with rat IgG (control) or anti-CD8 MAb. Immunity to Subdominant/Cryptic Epitopes T-cell mediated immunity against T. cruzi infection in A/Sn mice [44]. These short proteins contained only the AA 261 to 500 or 261 to 380. In both cases, they did not express the subdominant/ cryptic epitopes. Based on that, we concluded that immunization with the immundominant CD8 epitope alone could provide a high degree of protective immunity even in the absence of the subdominant epitopes [44]. The fact that the immune response directed solely to the immunodominant epitope can provide significant degree of protective immunity against protozoan parasites has been established a long time ago by the use of adoptively transferred T cell clones or heterelogous-prime boost vaccination regimen [45–49]. T-cell mediated immunity against T. cruzi infection in A/Sn mice [44]. These short proteins contained only the AA 261 to 500 or 261 to 380. In both cases, they did not express the subdominant/ cryptic epitopes. Based on that, we concluded that immunization with the immundominant CD8 epitope alone could provide a high degree of protective immunity even in the absence of the subdominant epitopes [44]. The fact that the immune response directed solely to the immunodominant epitope can provide significant degree of protective immunity against protozoan parasites has been established a long time ago by the use of adoptively transferred T cell clones or heterelogous-prime boost vaccination regimen [45–49]. disrupted. However, a more formal demonstration using bone marrow chimeric mice studies is lacking. At the molecular level, this strong immunodominance can be explained by the type of antigen presentation that predominates during T. cruzi infection. In recent studies, important evidence has been provided that subdominant epitopes can only be directly presented by the expressing cells, which might occur in the case of the recombinant adenovirus. However, during indirect priming (cross-priming), these epitopes would be at a disadvantage [52]. T. cruzi may use cross-priming as the dominant route, drastically reducing the priming of subdominant epitopes. In addition to shedding some light on the host-parasite relationship, our results may have important implications for the development of T cell vaccines against parasitic diseases. In our earlier studies, we observed that genetic vaccination with a heterologous prime-boost regimen employing plasmid DNA and recombinant adenovirus elicited strong, long lasting CD8+ T cell- mediated protective immunity against experimental infection in a mouse strain highly susceptible to T. cruzi infection [37,38]. Acknowledgments The authors are in debt with Dr. C. Claser for providing the immunofluorescence picture of Fig. 1. We are also in debt with Drs. Laurent Renia (Singapore Immunology Network), Dr. Chris Ibegbu (Emory Vaccine Center), Dr. Fanny Tzelepis (University of Ottawa) and Dr. Silvia B. Boscardin, for careful reviewing the manuscript. The mechanism operating during T. cruzi infection to restrict the immune response leading to immunodominance has yet to be characterized. We provide initial evidence that it could be explained by T cell competition for APCs by showing that in mice infected simultaneously with two different parasite strains containing different immunodominant epitopes, we could generate maximal responses to both epitopes without immunodominance or competition [15]. Our interpretation was that if the epitopes are presented by different APCs, then the immunodominant pattern is 14. Tzelepis F, Persechini PM, Rodrigues MM (2007) Modulation of CD4+ T cell- dependent specific cytotoxic CD8+ T cells differentiation and proliferation by the timing of increase in the pathogen load. PLoS One 2(4): e393. Author Contributions Conceived and designed the experiments: MRD ELVS JRCdV BCGdA MMR. Performed the experiments: MRD ELVS JRCdV BCGdA. Analyzed the data: MRD ELVS AVM OB-R RTG MMR. Contributed reagents/ materials/analysis tools: AVM OB-R RTG. Wrote the paper: MMR. Conceived and designed the experiments: MRD ELVS JRCdV BCGdA MMR. Performed the experiments: MRD ELVS JRCdV BCGdA. Analyzed the data: MRD ELVS AVM OB-R RTG MMR. Contributed reagents/ materials/analysis tools: AVM OB-R RTG. Wrote the paper: MMR. Immunity to Subdominant/Cryptic Epitopes In their study, they induced simultaneous tolerance to two immunodominant T. cruzi epitopes in a resistant mouse strain. Following infection, an increased susceptibility to infection was observed. Nevertheless, they were still able to control and survive the experimental infection. This protective immunity was possibly mediated by CD8+ T cells specific to subdominant epitopes that substituted for the immunodominant ones. The AA sequences of these subdominant/cryptic epitopes have yet to be identified. Together with our study, they strongly support the notion that the immune responses to both dominant and subdominant/cryptic epitopes can be important for controlling experimental T. cruzi infection in inbred mouse strains. These results are in agreement with the observation with other parasites such as Plasmodium. Tolerance to or removal of the the immunodominant CD8 epitope of P. yoelii led to the development of immunity to CD8 subdominant epitopes as well [50,51]. Finally, our observations may have important implications regarding the basis for the strong immunodominance pattern observed after experimental infection with other intracellular parasites such as Plasmodium, Toxoplasma gondii and Theileria parva [50,51,54-57]. g p g ( ) 15. Tzelepis F, de Alencar BC, Penido ML, Claser C, Machado AV, et al. (2008) Infection with Trypanosoma cruzi restricts the repertoire of parasite-specific CD8+ T cells leading to immunodominance. J Immunol 180: 1737–48. 2. Miyahira Y (2008) Trypanosoma cruzi infection from the view of CD8+ T cell immunity–an infection model for developing T cell vaccine. Parasitol Int 57: 38–48. 1. Tarleton RL (2007) Immune system recognition of Trypanosoma cruzi. Curr Opin Immunol 19: 430–4. 13. Martin DL, Weatherly DB, Laucella SA, Cabinian MA, Crim MT, et al. (2006) CD8+ T-Cell responses to Trypanosoma cruzi are highly focused on strain-variant trans-sialidase epitopes. PLoS Pathog 2(8): e77. Immunity to Subdominant/Cryptic Epitopes Here, we demonstrated that the CD8 T cell-mediated protective immunity observed was directed to three distinct epitopes, two of which are cryptic. The strategy of redirecting immunity to epitopes that are not usually targets of the naturally acquired immune response has been proposed as a possible means to improve immunity against viral infection [40–43]. Recently, this strategy was also proven useful to improve vaccination against T. cruzi infection [53]. g [ ] In earlier studies in which we depleted CD8+ T cells from genetically vaccinated mice, we observed that these mice were unable to control parasitemia and died at the same time as control unvaccinated animals [37]. Therefore, although genetic immuni- zation elicits effector CD4+ T cells, these cells do not account for the protection we observed. This concept was further corroborated with experiments of CD8 T cell-depletion performed here (Fig. 7A to D). Using a different approach, a similar conclusion was also reached by Rosemberg et al., 2010 [17]. In their study, they induced simultaneous tolerance to two immunodominant T. cruzi epitopes in a resistant mouse strain. Following infection, an increased susceptibility to infection was observed. Nevertheless, they were still able to control and survive the experimental infection. This protective immunity was possibly mediated by CD8+ T cells specific to subdominant epitopes that substituted for the immunodominant ones. The AA sequences of these subdominant/cryptic epitopes have yet to be identified. Together with our study, they strongly support the notion that the immune responses to both dominant and subdominant/cryptic epitopes can be important for controlling experimental T. cruzi infection in inbred mouse strains. These results are in agreement with the observation with other parasites such as Plasmodium. Tolerance to or removal of the the immunodominant CD8 epitope of P. yoelii led to the development of immunity to CD8 subdominant epitopes as well [50,51]. In earlier studies in which we depleted CD8+ T cells from genetically vaccinated mice, we observed that these mice were unable to control parasitemia and died at the same time as control unvaccinated animals [37]. Therefore, although genetic immuni- zation elicits effector CD4+ T cells, these cells do not account for the protection we observed. This concept was further corroborated with experiments of CD8 T cell-depletion performed here (Fig. 7A to D). p p p ( g ) Using a different approach, a similar conclusion was also reached by Rosemberg et al., 2010 [17]. 3. Padilla AM, Bustamante JM, Tarleton RL (2009) CD8+ T cells in Trypanosoma cruzi infection. Curr Opin Immunol 21: 385–90. 10. Scharfstein J, Gomes Jde A, Correa-Oliveira R (2009) Mem Inst Oswaldo Cruz. Back to the future in Chagas disease: from animal models to patient cohort studies, progress in immunopathogenesis research 104(Suppl 1): 187–98. g J 16. Bixby LM, Tarleton RL (2008) Stable CD8+ T cell memory during persistent Trypanosoma cruzi infection. J Immunol 181: 2644–50. Discussion Immunity to epitopes that are not commonly recognized during infection (cryptic) provided a significant degree of protective immunity. Nonetheless, immune responses against these cryptic epitopes did In previous studies, we showed that immunization with short proteins in the presence of the TLR9 agonist CpG elicited CD8+ PLoS ONE \| www.plosone.org July 2011 \| Volume 6 \| Issue 7 \| e22011 July 2011 \| Volume 6 \| Issue 7 \| e22011 10 Immunity to Subdominant/Cryptic Epitopes ( pp ) 12. Tzelepis F, de Alencar BC, Penido ML, Gazzinelli RT, Persechini PM, et al. (2006) Distinct kinetics of effector CD8+ cytotoxic T cells after infection with Trypanosoma cruzi in naive or vaccinated mice. Infect Immun 74: 2477–81. 11. Lannes-Vieira J, Silverio JC, Pereira IR, Vinagre NF, Carvalho CM, et al. (2009) Chronic Trypanosoma cruzi-elicited cardiomyopathy: from the discovery to the proposal of rational therapeutic interventions targeting cell adhesion molecules and chemokine receptors–how to make a dream come true. Mem Inst Oswaldo Cruz 104(Suppl 1): 226–35. 9. Gutierrez FR, Guedes PM, Gazzinelli RT, Silva JS (2009) The role of parasite persistence in pathogenesis of Chagas heart disease. Parasite Immunol 31: 673–85. p 4. Boscardin SB, Torrecilhas AC, Manarin R, Revelli S, Rey EG, et al. (2010) Chagas’ disease: an update on immune mechanisms and therapeutic strategies. J Cell Mol Med 14: 1373–84. 5. Junqueira C, Caetano B, Bartholomeu DC, Melo MB, Ropert C, et al. (2010) The endless race between Trypanosoma cruzi and host immunity: lessons for and beyond Chagas disease. Expert Rev Mol Med 12: e29. g 8. Bonney KM, Engman DM (2008) Chagas heart disease pathogenesis: one mechanism or many? Curr Mol Med 8: 510–8. 6. Kierszenbaum F (2005) Where do we stand on the autoimmunity hypothesis of Chagas disease? Trends Parasitol 21: 513–6. 7. Marin-Neto JA, Cunha-Neto E, Maciel BC, Simo˜es MV (2008) Pathogenesis of chronic Chagas heart disease. Circulation 115: 1109–23. Immunity to Subdominant/Cryptic Epitopes Chen W, McCluskey J (2009) Immunodominance and immunodomination: critical factors in developing effective CD8+ T-cell-based cancer vaccines. Adv Cancer Res 95: 203–47. 49. Schneider J, Gilbert SC, Blanchard TJ, Hanke T, Robson KJ, et al. (1998) Enhanced immunogenicity for CD8+ T cell induction and complete protective efficacy of malaria DNA vaccination by boosting with modified vaccinia virus Ankara. Nat Med 4: 397–402. 30. Jenkins MR, Webby R, Doherty PC, Turner SJ (2009) Addition of a prominent epitope affects influenza A virus-specific CD8+ T cell immunodominance hierarchies when antigen is limiting. J Immunol 177: 2917–25. 50. Kumar KA, Sano G, Boscardin S, Nussenzweig RS, Nussenzweig MC, et al. (2006) The circumsporozoite protein is an immunodominant protective antigen in irradiated sporozoites. Nature 444: 937–40. 31. Willis RA, Kappler JW, Marrack PC (2006) CD8 T cell competition for dendritic cells in vivo is an early event in activation. Proc Natl Acad Sci USA 103: 12063–8. 51. Mauduit M, Gru¨ner AC, Tewari R, Depinay N, Kayibanda M, et al. (2009) A role for immune responses against non-CS components in the cross-species protection induced by immunization with irradiated malaria sporozoites. PLoS One 4(11): e7717. 32. Newberg MH, McEvers KJ, Gorgone DA, Lifton MA, Baumeister SH, et al. (2006) Immunodomination in the evolution of dominant epitope-specific CD8+ T lymphocyte responses in simian immunodeficiency virus-infected rhesus monkeys. J. Immunol 176: 319–28. 52. Pavelic V, Matter MS, Mumprecht S, Breyer I, Ochsenbein AF (2009) CTL induction by cross-priming is restricted to immunodominant epitopes. Eur J Immunol 39: 704–16. 33. Kastenmuller W, Gasteiger G, Gronau JH, Baier R, Ljapoci R, et al. (2007) Cross-competition of CD8+ T cells shapes the immunodominance hierarchy during boost vaccination. J Exp Med 204: 2187–98. 53. Cazorla SI, Frank FM, Becker PD, Arnaiz M, Mirkin GA, et al. (2010) Redirection of the immune response to the functional catalytic domain of the cystein proteinase cruzipain improves protective immunity against Trypanosoma cruzi infection. J Infect Dis 202: 136–44. 34. Riedl P, Wieland A, Lamberth K, Buus S, Lemonnier F, et al. (2009) Elimination of immunodominant epitopes from multispecific DNA-based vaccines allows induction of CD8 T cells that have a striking antiviral potential. J Immunol 183: 370–80. J 54. Blanchard N, Gonzalez F, Schaeffer M, Joncker NT, Cheng T, et al. (2008) Immunodominant, protective response to the parasite Toxoplasma gondii requires antigen processing in the endoplasmic reticulum. Nat Immunol 9: 937–44. 35. Immunity to Subdominant/Cryptic Epitopes Yewdell JW (2006) Confronting complexity: real-world immunodominance in antiviral CD8+ T cell responses. Immunity 25: 533–43. 44. Arau´jo AF, de Alencar BC, Vasconcelos JR, Hiyane MI, Marinho CR, et al. (2005) CD8+-T-cell-dependent control of Trypanosoma cruzi infection in a highly susceptible mouse strain after immunization with recombinant proteins based on amastigote surface protein 2. Infect Immun 73: 6017–25. 24. Irvine K, Bennink J (2006) Factors influencing immunodominance hierarchies in TCD8+ -mediated antiviral responses. Expert Rev Clin Immunol 2: 135–47. 25. La Gruta NL, Kedzierska K, Pang K, Webby R, Davenport M, et al. (2006) A virus-specific CD8+ T cell immunodominance hierarchy determined by antigen dose and precursor frequencies. Proc Natl Acad Sci USA 103: 994–9. g p 45. Rodrigues M, Nussenzweig RS, Romero P, Zavala F (1992) The in vivo cytotoxic activity of CD8+ T cell clones correlates with their levels of expression of adhesion molecules. J Exp Med 175: 895–905. p y dose and precursor frequencies. Proc Natl Acad Sci USA 103: 9 26. Kotturi MF, Scott I, Wolfe T, Peters B, Sidney J, et al. (2008) Naive precursor frequencies and MHC binding rather than the degree of epitope diversity shape CD8+ T cell immunodominance. J Immunol 181: 2124–33. 46. Li S, Rodrigues M, Rodriguez D, Rodriguez JR, Esteban M, et al. (1993) Priming with recombinant influenza virus followed by administration of recombinant vaccinia virus induces CD8+ T-cell-mediated protective immunity against malaria. Proc Natl Acad Sci USA 90: 5214–18. 27. Sacha JB, Reynolds MR, Buechler MB, Chung C, Jonas AK, et al. (2008) Differential antigen presentation kinetics of CD8+ T-cell epitopes derived from the same viral protein. J Virol 82: 9293–8. against malaria. Proc Natl Acad Sci USA 90: 5214–18. 47. Rodrigues M, Li S, Murata K, Rodriguez D, Rodriguez JR, et al. (1994) Influenza and vaccinia viruses expressing malaria CD8+ T and B cell epitopes. Comparison of their immunogenicity and capacity to induce protective immunity. J Immunol 153: 4636–48. 28. La Gruta NL, Rothwell WT, Cukalac T, Swan NG, Valkenburg SA, et al. (2010) Primary CTL response magnitude in mice is determined by the extent of naive T cell recruitment and subsequent clonal expansion. J Clin Invest 120: 1885–94. 48. Sedegah M, Jones TR, Kaur M, Hedstrom R, Hobart P, et al. (1998) Boosting with recombinant vaccinia increases immunogenicity and protective efficacy of malaria DNA vaccine. Proc Natl Acad Sci USA 95: 7648–53. 29. Immunity to Subdominant/Cryptic Epitopes 17. Rosenberg CS, Martin DL, Tarleton RL (2010) CD8+ T cells specific for immunodominant trans-sialidase epitopes contribute to control of Trypanosoma cruzi infection but are not required for resistance. J Immunol 185: 560–8. 38. Haolla FA, Claser C, de Alencar BC, Tzelepis F, de Vasconcelos JR, et al. (2009) Strain-specific protective immunity following vaccination against experimental Trypanosoma cruzi infection. Strain-specific protective immunity following vaccination against experimental Trypanosoma cruzi infection. Vaccine 27: 5644–53. 18. Freire-de-Lima L, Alisson-Silva F, Carvalho ST, Takiya CM, Rodrigues MM, et al. (2010) Trypanosoma cruzi subverts host cell sialylation and may compromise antigen-specific CD8+ T cell responses. J Biol Chem 285: 13388–96. 39. Claser C, Espı´ndola NM, Sasso G, Vaz AJ, Boscardin SB, et al. (2007) Immunologically relevant strain polymorphism in the Amastigote Surface Protein 2 of Trypanosoma cruzi. Microbes Infect 9: 1011–9. 19. Oliveira AC, de Alencar BC, Tzelepis F, Klezewsky W, da Silva RN, et al. (2010) Impaired innate immunity in Tlr4(-/-) mice but preserved CD8+ T cell responses against Trypanosoma cruzi in Tlr4-, Tlr2-, Tlr9- or Myd88-deficient mice 29; 6(4): e1000870. 40. Oukka M, Manuguerra JC, Livaditis N, Tourdot S, Riche N, et al. (1996) Protection against lethal viral infection by vaccination with nonimmunodomi- nant peptides. J Immunol 157: 3039–45. 20. Rodrigues MM, de Alencar BC, Claser C, Tzelepis F (2009) Immunodomi- nance: a new hypothesis to explain parasite escape and host/parasite equilibrium leading to the chronic phase of Chagas’ disease? Braz J Med Biol Res 42: 220–3. 41. Friedrich TC, Valentine LE, Yant LJ, Rakasz EG, Piaskowski SM, et al. (2007) Subdominant CD8+ T-cell responses are involved in durable control of AIDS virus replication. J Virol 81: 3465–76. 21. Chen W, Anto´n LC, Bennink JR, Yewdell JW (2000) Dissecting the multifactorial causes of immunodominance in class I-restricted T cell responses to viruses. Immunity 12: 83–93. p 42. Holtappels R, Simon CO, Munks MW, Thomas D, Deegen P, et al. (2008) Subdominant CD8 T-cell epitopes account for protection against cytomegalo- virus independent of immunodomination. J Virol 82: 5781–96. 22. Tenzer S, Wee E, Burgevin A, Stewart-Jones G, Friis L, et al. (2009) Antigen processing influences HIV-specific cytotoxic T lymphocyte immunodominance. Nat Immunol 10: 636–46. 43. Ruckwardt TJ, Luongo C, Malloy AM, Liu J, Chen M, et al. (2010) Responses against a subdominant CD8+ T cell epitope protect against immunopathology caused by a dominant epitope. J Immunol 185: 4673–80. 23. References 11. Lannes-Vieira J, Silverio JC, Pereira IR, Vinagre NF, Carvalho CM, et al. (2009) Chronic Trypanosoma cruzi-elicited cardiomyopathy: from the discovery to the proposal of rational therapeutic interventions targeting cell adhesion molecules and chemokine receptors–how to make a dream come true. Mem Inst Oswaldo Cruz 104(Suppl 1): 226–35. p 4. Boscardin SB, Torrecilhas AC, Manarin R, Revelli S, Rey EG, et al. (2010) Chagas’ disease: an update on immune mechanisms and therapeutic strategies. J Cell Mol Med 14: 1373–84. ( pp ) 12. Tzelepis F, de Alencar BC, Penido ML, Gazzinelli RT, Persechini PM, et al. (2006) Distinct kinetics of effector CD8+ cytotoxic T cells after infection with Trypanosoma cruzi in naive or vaccinated mice. Infect Immun 74: 2477–81. 5. Junqueira C, Caetano B, Bartholomeu DC, Melo MB, Ropert C, et al. (2010) The endless race between Trypanosoma cruzi and host immunity: lessons for and beyond Chagas disease. Expert Rev Mol Med 12: e29. 6. Kierszenbaum F (2005) Where do we stand on the autoimmunity hypothesis of Chagas disease? Trends Parasitol 21: 513–6. 7. Marin-Neto JA, Cunha-Neto E, Maciel BC, Simo˜es MV (2008) Pathogenesis of chronic Chagas heart disease. Circulation 115: 1109–23. PLoS ONE \| www.plosone.org 11 July 2011 \| Volume 6 \| Issue 7 \| e22011 Immunity to Subdominant/Cryptic Epitopes Immunity to Subdominant/Cryptic Epitopes Vasconcelos JR, Hiyane MI, Marinho CR, Claser C, Machado AM, et al. (2004) Protective immunity against trypanosoma cruzi infection in a highly susceptible mouse strain after vaccination with genes encoding the amastigote surface protein-2 and trans-sialidase. Hum Gene Ther 15: 878–86. 55. Frickel EM, Sahoo N, Hopp J, Gubbels MJ, Craver MP, et al. (2008) Parasite stage-specific recognition of endogenous Toxoplasma gondii-derived CD8+ T cell epitopes. J Infect Dis 198: 1625–33. 36. Machado AV, Cardoso JE, Claser C, Rodrigues MM, Gazzinelli RT, et al. (2006) Long-term protective immunity induced against Trypanosoma cruzi infection after vaccination with recombinant adenoviruses encoding amastigote surface protein-2 and trans-sialidase. Hum Gene Ther 17: 898–908. p p J 56. Wilson DC, Grotenbreg GM, Liu K, Zhao Y, Frickel EM, et al. (2010) Differential regulation of effector- and central-memory responses to Toxoplasma gondii Infection by IL-12 revealed by tracking of Tgd057-specific CD8+ T cells. PLoS Pathog 6(3): e1000815. 37. de Alencar BC, Persechini PM, Haolla FA, de Oliveira G, Silverio JC, et al. (2009) Perforin and gamma interferon expression are required for CD4+ and CD8+ T-cell-dependent protective immunity against a human parasite, Trypanosoma cruzi, elicited by heterologous plasmid DNA prime-recombinant adenovirus 5 boost vaccination. Infect Immun 77: 4383–95. 57. MacHugh ND, Connelley T, Graham SP, Pelle R, Formisano P, et al. (2009) CD8+ T-cell responses to Theileria parva are preferentially directed to a single dominant antigen: Implications for parasite strain-specific immunity. Eur J Immunol 39: 2459–69. PLoS ONE \| www.plosone.org July 2011 \| Volume 6 \| Issue 7 \| e22011 12
W4200372202.txt	https://www.frontiersin.org/articles/10.3389/fmicb.2021.760444/pdf	en		Dual RNA-Seq Analysis of the Interaction Between Edible Fungus Morchella sextelata and Its Pathogenic Fungus Paecilomyces penicillatus Uncovers the Candidate Defense and Pathogenic Factors	Frontiers in microbiology	2,021	cc-by	7,518	ORIGINAL RESEARCH published: 02 December 2021 doi: 10.3389/fmicb.2021.760444 Edited by: Anhuai Lu, Peking University, China Reviewed by: José Ascención Martínez-Álvarez, University of Guanajuato, Mexico Chenyang Huang, Institute of Agricultural Resources and Regional Planning, Chinese Academy of Agricultural Sciences (CAAS), China Correspondence: Yang Yu yangyu0221@139.com Weihong Peng pwh424@163.com Specialty section: This article was submitted to Microbiological Chemistry and Geomicrobiology, a section of the journal Frontiers in Microbiology Received: 18 August 2021 Accepted: 05 November 2021 Published: 02 December 2021 Citation: Yu Y, Tan H, Liu T, Liu L, Tang J and Peng W (2021) Dual RNA-Seq Analysis of the Interaction Between Edible Fungus Morchella sextelata and Its Pathogenic Fungus Paecilomyces penicillatus Uncovers the Candidate Defense and Pathogenic Factors. Front. Microbiol. 12:760444. doi: 10.3389/fmicb.2021.760444 Dual RNA-Seq Analysis of the Interaction Between Edible Fungus Morchella sextelata and Its Pathogenic Fungus Paecilomyces penicillatus Uncovers the Candidate Defense and Pathogenic Factors Yang Yu 1,2, Hao Tan 1,2,3, Tianhai Liu 1,2, Lixu Liu 1,2, Jie Tang 1,2 and Weihong Peng 1,2* National-Local Joint Engineering Laboratory of Breeding and Cultivation of Edible and Medicinal Fungi, Institute of Agricultural Resources and Environment, Sichuan Academy of Agricultural Sciences, Chengdu, China, 2 National Observing and Experimental Station of Agricultural Microbiology, Ministry of Agriculture and Rural Affairs, Chengdu, China, 3 School of Bioengineering, Jiangnan University, Wuxi, China 1 Morels (Morchella spp.) are economically important mushrooms cultivated in many countries. However, their production and quality are hindered by white mold disease because of Paecilomyces penicillatus infection. In this study, we aimed to understand the genetic mechanisms of interactions between P. penicillatus and Morchella. M. sextelata, the most prevalent species of Morchella in China, was inoculated with P. penicillatus; then, the expression profiles of both fungi were determined simultaneously at 3 and 6 days post-inoculation (dpi) using a dual RNA-Seq approach. A total of 460 and 313 differentially expressed genes (DEGs) were identified in P. penicillatus and M. sextelata, respectively. The CAZymes of β-glucanases and mannanases, as well as subtilase family, were upregulated in P. penicillatus, which might be involved in the degradation of M. sextelata cell walls. Chitin recognition protein, caffeine-induced death protein, and putative apoptosis-inducing protein were upregulated, while cyclin was downregulated in infected M. sextelata. This indicates that P. penicillatus could trigger programmed cell death in M. sextelata after infection. Laccase-2, tyrosinases, and cytochrome P450s were also upregulated in M. sextelata. The increased expression levels of these genes suggest that M. sextelata could detoxify the P. penicillatus toxins and also form a melanin barrier against P. penicillatus invasion. The potential pathogenic mechanisms of P. penicillatus on M. sextelata and the defense mechanisms of M. sextelata against P. penicillatus were well described. Keywords: Morchella sextelata, Paecilomyces penicillatus, pathogenic factors, response, transcriptomics, CAZymes, laccase, tyrosinase Frontiers in Microbiology \| www.frontiersin.org 1 December 2021 \| Volume 12 \| Article 760444 Yu et al. Morel Interacted With Paecilomyces penicillatus INTRODUCTION Tan et al., 2019; Wang et al., 2020). P. penicillatus was found to harbor many CAZymes, particularly clusters of chitinase genes, which could be related to its pathogenicity (Wang et al., 2020). Cheng et al. (2021) reported the regulatory mechanism of interaction between P. penicillatus and M. importuna. However, the lack of a control group of M. importuna at the same developmental stage complicated the conclusions from that study. To date, the genetic and pathogenic mechanisms that underlie the interaction between P. penicillatus and Morchella remain unclear, which makes it difficult to efficiently breed resistant varieties and develop specific biological or chemical methods to control WMD in Morchella. In this study, M. sextelata, the most prevalent species of Morchella in China, was selected as the research object and artificially inoculated with P. penicillatus to develop WMD in the field. The transcriptional changes of P. penicillatus in different infection stages were analyzed. The differences in transcriptional changes were compared between different growth stages of infected and healthy M. sextelata. We aimed to profile the infection mechanism of P. penicillatus, the response mechanism of M. sextelata to P. penicillatus infection, and the candidate defense genes of M. sextelata that protect against P. penicillatus. This will provide theoretical support for the effective control of WMD and reduce the economic risk caused by WMD in the cultivation of Morchella. Morels (Morchella spp.), members of the Ascomycetes, Pezizomycetes, Pezizales, Morchellaceae, and Morchella, are commercially important edible mushrooms that are consumed throughout the world (Du et al., 2012). After more than 100 years of trials, the commercial cultivation of Morchella in the field was successful (Ower et al., 1986; Dahlstrom et al., 2000; Masaphy, 2010). Since 2012, the industry that cultivates Morchella in China, with Sichuan as a typical example, developed rapidly (Peng et al., 2016). The cultivation area of morels in China has expanded from 200 ha in 2012 to 10,000 ha in 2020 (Zhao, 2021). The considerable economic benefits of morel cultivation have attracted many farmers and have become their main source of income. The Morchella species currently cultivated in China include M. importuna, M. sextelata, and M. septimelata (Liu et al., 2018). Among them, M. sextelata comprises >90% of the total area of Morchella cultivation (Deng et al., 2021). However, the white mold disease (WMD) caused by Paecilomyces penicillatus (Ascomycetes) seriously harms Morchella (He et al., 2017). WMD outbreak spreads quickly, infecting a large area. In severe cases, 60–80% of Morchella are infected in China, resulting in a decline in their production and commodity value (Chen et al., 2017; He et al., 2017). In addition, a high proportion of P. penicillatus in soil microbial community could inhibit the formation of f Morchella fruiting bodies (Tan et al., 2021a). A large number of farmers suffer financial hardship owing to WMD. Therefore, it is extremely urgent to prevent and control WMD. Edible mushrooms are primarily cultivated and managed indoors. Therefore, the prevention and control of edible mushroom diseases rely on environmental controls during cultivation, including the control of temperature, light, water, and air, as well as the use of physical and chemical sterilization methods (Gea et al., 2021). Morchella is artificially cultivated in the field (Liu et al., 2018). Hence, the common strategies to prevent and control diseases of edible mushrooms are not effective for the control of Morchella diseases, such as WMD. Similar to other major crops, breeding resistant varieties are an effective way to control mushroom diseases (Kage et al., 2016). Several studies have addressed the immune response of mushrooms to mycoparasites, although this area of research is in its infancy. The activity of laccase, particularly encoded by lcc2 gene, was found to contribute to the metabolism of toxin production in Trichoderma aggressivum and enhanced the resistance to green mold disease in Agaricus bisporus (Sjaarda et al., 2015). A total of 17 simple sequence repeat markers associated with resistance to Mycogone perniciosa in A. bisporus have been established (Fu et al., 2016). A quantitative trait loci analysis was used to identify the locations, numbers, and effects of genomic regions associated with resistance to Lecanicillium fungicola in A. bisporus, and four traits related to resistance were analyzed (Foulongne-Oriol et al., 2012). Further research and data mining will help design classical breeding and genetic modification schemes, thus, producing resistant varieties. Recently, the genome-wide profiles of several species of Morchella, including M. importuna, M. sextelata, and M. septimelata, as well as that of the pathogen P. penicillatus, were reported (Wingfield et al., 2018; Mei et al., 2019; Frontiers in Microbiology \| www.frontiersin.org MATERIALS AND METHODS Fungal Strains and Experimental Treatment Cultivated strains of M. sextelata were purchased from Jindi Tianlingjian Biotechnology Co., Ltd., in Chengdu, Sichuan, China. P. penicillatus was isolated in our previous study (He et al., 2017) and was deposited in the Culture Collection Center of the Soil and Fertilizer Institute, Sichuan Academy of Agricultural Sciences, Chengdu, China. The strains of M. sextelata used in this study were cultivated on a farm in Xindu, Chengdu, China (30.8°N, 104.2E). The cultivation, growth, and management of morels were done according to standard practices (Tan et al., 2021b). P. penicillatus was cultured on potato dextrose agar medium for 5 d at 25°C before inoculation. When the fruiting bodies of M. sextelata reached 7 cm high, they were inoculated with P. penicillatus mycelia. Needle tip-sized mycelia were picked up with sterilized toothpicks and inoculated on Morchella fruiting bodies. A total of 120 M. sextelata fruiting bodies were used for the field experiment. Half of them were inoculated with P. penicillatus as explained above, while the remaining half were not inoculated. According to the phenotypic changes of the lesion, samples were collected at 3 and 6 days post-inoculation (dpi). One cm2 fruiting body centered on the lesion was collected for each inoculated morel, while healthy tissue from the same site was collected from the non-inoculated morels. The samples were divided into four groups, including non-inoculated groups on 3 and 6 dpi and inoculated groups on 3 and 6 dpi. Each group contained three 2 December 2021 \| Volume 12 \| Article 760444 Yu et al. Morel Interacted With Paecilomyces penicillatus biological replicates, and groups of six fruiting bodies were pulled into one biological replicate. All the samples were stored at −80°C for subsequent transcriptome analysis. TopGO was used for the Gene Ontology (GO) enrichment analysis, and the calculated value of p was then obtained using the Wallenius non-central hypergeometric distribution method to identify significantly enriched GO terms (p < 0.05) from all of the DEGs (Ashburner et al., 2000). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were performed on the DEGs using clusterProfiler. The value of p was calculated using the hypergeometric distribution method (Kanehisa et al., 2004). RNA Extraction and qRT-PCR Total RNA was isolated and purified from healthy and inoculated morels using the TRIzol reagent (Invitrogen, Carlsbad, CA, United States) following the manufacturer’s instructions. The purity, concentration, and integrity of each sample were determined using a NanoDrop spectrophotometer (Thermo Fisher Scientific, Inc., Waltham, MA, United States), a Qubit RNA Kit (Life Technologies, Carlsbad, CA, United States), and a 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, United States), respectively. About 2 μg of total RNA from each sample was reversetranscribed into cDNA using a RevertAid First Strand cDNA Synthesis Kit (THERMO, United Kingdom). The profiles of expression of the genes expression profiles were determined using SYBR Green Premix (TaKaRa Code RR820A; TaKaRa, Shiga, Japan) on a Roche LightCycle480 system (Roche Applied Science, Rotkreuz, Switzerland). All the PCR reactions were conducted using 40 cycles at 98°C for 10 s, 60°C for 10 s, and 72°C for 10 s, in a 20 μl reaction mixture that contained 10 pmol of each primer and 2 μl of cDNA as a template. All the reactions were performed in triplicate, and 5.8S rRNA was used as the internal control for normalization. Supplementary Table S1 lists the primers used for quantitative real-time reverse transcriptase–PCR (qRT-PCR). Data Availability The Illumina Nova-Seq sequencing data from this study were deposited in the NCBI Sequence Read Archive.2 The reference genome of P. penicillatus can be downloaded from https:// www.ncbi.nlm.nih.gov/genome/79861?genome_assembly_id=1473685, and the reference genome of M. sextelata can be downloaded from https://www.ncbi.nlm.nih.gov/genome/86229?genome_ assembly_id=748597. RESULTS Overview of the RNA-Seq Data White spots were observed on the M. sextelata fruiting bodies at 3 days after inoculation with P. penicillatus, indicating that P. penicillatus was successfully germinated and began to infect Morchella. The typical symptom of white mold disease started to appear at 6 dpi (Figure 1). Based on the observed phenotypic change, we sampled tissues at 3 and 6 dpi, respectively. The non-inoculated tissues at 3 and 6 dpi were used as the control groups for the subsequent dual RNA-Seq analysis. An average of 75,728,447 raw reads was generated for each sample with a Q20 and Q30 of 97.81 and 94.11%, respectively (Supplementary Table S2). After quality control, an average of 68,239,578 clean reads was generated. All the clean reads were mapped to a corresponding reference genome. An average of 92.87% clean reads of the non-inoculated samples was mapped to the M. sextelata genome, while none (< 0.05%) was mapped to the P. penicillatus genome. A total of 90.56 and 84.72% clean reads of the inoculated samples were mapped to the genome of M. sextelata at the early and late stages of infection, respectively. However, 2.01 and 6.85% clean reads from the infected sites were mapped to the P. penicillatus genome at the early and late stages of infection, respectively (Supplementary Table S3). cDNA Library Construction and Sequencing cDNA library preparation and sequencing were conducted by the Personal Biotechnology Co., Ltd., in Shanghai, China. All the libraries were sequenced using an Illumina Nova-Seq platform (Illumina, San Diego, CA, United States). The FASTX toolkit was used to filter the raw data, short-fragment reads, and sequencing adapters, while other low-quality reads were filtered to obtain the clean reads. After preprocessing the RNA-Seq data, the reads were mapped to the P. penicillatus and M. sextelata reference genomes using HISAT2 software.1 The RNA reads were classified as M. sextelata or P. penicillatus based on their similarity to the corresponding genome sequence. Fragments per kilobase of transcript per million mapped fragments were used to normalize the abundance of transcripts (Trapnell et al., 2012). Quantitative RT-PCR Validation Differential Expression Analysis We selected 19 genes, including 11 (laccase-2, ABCB1, tyrosinase, primary amine oxidase, pckA, probable acetate kinase, aconitate hydratase, alcohol dehydrogenase 6, indoleamine 2,3-dioxygenase, pyruvate decarboxylase, and sophorolipid transporter) from M. sextelata and eight (chitinase 1, CMB1, GH20, 3-isopropylmalate dehydratase, phosphotransferase, glucosamine-6-phosphate deaminase, leuB, and serine dehydratases) from P. penicillatus, Differential expression analysis was conducted using DESeq2 (Love et al., 2014). The resulting value of p were adjusted using Benjamini and Hochberg’s approach to control the false discovery rate. Genes with an adjusted p < 0.05 and log2 (fold change) > 1 that were detected by DESeq were designated as differentially expressed. http://ccb.jhu.edu/software/hisat2/index.shtml https://www.ncbi.nlm.nih.gov/sra/ 1 Frontiers in Microbiology \| www.frontiersin.org 2 3 December 2021 \| Volume 12 \| Article 760444 Yu et al. Morel Interacted With Paecilomyces penicillatus FIGURE 1 \| Field phenotype of the white mold disease (WMD) in Morchella sextelata. (A,C) represent the non-inoculated morels at 3 and 6 days, respectively. (B,D) represent the morels at 3 and 6 days post-Paecilomyces penicillatus inoculation, respectively. The red arrow shows the WMD symptom. to validate the RNA-Seq data. qRT-PCR assays were conducted to test their patterns of expression at 3 and 6 dpi. As shown in Figure 2, the qPCR assay identified nine upregulated genes in M. sextelata and two were downregulated. Among them, laccase-2 was highly upregulated at both infection stages by 10.9- (3 dpi) and 21.4-fold (6 dpi). In P. penicillatus, three were upregulated, and two were downregulated. The patterns of expression obtained from the qRT-PCR were consistent with the RNA-Seq results, indicating that the RNA-Seq data were well suited to analyze the interaction between P. penicillatus and M. sextelata. including 25 glycosyl hydrolases (GHs), seven glycosyltransferases (GTs), seven auxiliary activities (AAs) enzymes, four carbohydrate-binding modules, and one carbohydrate esterase (CE). Among them, GH 17 (glucan endo-1,3-β-glucosidase), AA6 (1,4-benzoquinone reductase), and GT2 (chitin synthase 1) were the top three genes that were expressed, and they were all upregulated during the late stage of P. penicillatus infection (Supplementary Table S4), indicating their importance in the pathogenicity of P. penicillatus. Out of the 25 GHs, six β-glucanases, three α-mannosidases, and two α-1,6-mannanases were upregulated, while two chitinases (GH18) were downregulated (Figure 3B). This information indicated that the main cell wall-degrading enzymes (CWDEs) were glucanases and mannanases rather than chitinase when P. penicillatus infected M. sextelata. Differentially Expressed Genes in P. penicillatus The transcriptional expression of P. penicillatus on 3 and 6 dpi was analyzed. A total of 460 DEGs in P. penicillatus were identified by comparing the abundance of transcripts on 6 dpi with those on 3 dpi. A total of 336 were upregulated, and 124 were downregulated. Two proteases, including subtilase family protease and serine proteinase, were the most highly expressed among these DEGs (Figure 3A; Supplementary Table S4). The expressions of CAZymes in P. penicillatus, which play an important role in mycoparasitic fungi (Wang et al., 2020), were analyzed. A total of 45 CAZymes were identified from all the DEGs in P. penicillatus (Figure 3B), Frontiers in Microbiology \| www.frontiersin.org Functional Analysis of Differentially Expressed Genes in P. penicillatus GO category enrichment analysis was utilized to elucidate the functional enrichment of the DEGs in P. penicillatus. A total of 73 GO terms were enriched (p < 0.05) in P. penicillatus (Supplementary Table S5). Four “hydrolase activity (GO:0004553, GO:0016813, GO:0016810, and GO:0016798),” three “oxidoreductase activity (GO:0016614, GO:0016616, and GO:0016491),” one “NAD binding (GO:0008172)” in the category 4 December 2021 \| Volume 12 \| Article 760444 Yu et al. Morel Interacted With Paecilomyces penicillatus A B FIGURE 2 \| Validation of the RNA-Seq data by qRT-PCR. (A) 11 DEGs from M. sextelata were selected for validation, and the y-axis shows the fold change expression at 3 and 6 days post-inoculation (dpi) compared with the non-inoculated samples. (B) Eight DEGs from P. penicillatus were selected, and the y-axis shows the fold change expression at 6 dpi compared with 3 dpi. Each data point was obtained from three biological replicates. DEGs, differentially expressed genes; qRT-PCR, quantitative real-time reverse transcriptase–PCR. A C D B FIGURE 3 \| DEGs at 6 days post-inoculation (dpi) compared with 3 dpi in Paecilomyces penicillatus. (A) Volcano plot of the DEGs in P. penicillatus. Differentially expressed CAZymes (B), proteases (C) and cytochrome P450s (D) in P. penicillatus during the infection process. ST, inoculated samples; 3 and 6 dpi, 3 and 6 days post-inoculation; DEGs, differentially expressed genes. Frontiers in Microbiology \| www.frontiersin.org 5 December 2021 \| Volume 12 \| Article 760444 Yu et al. Morel Interacted With Paecilomyces penicillatus of molecular function, one “carbohydrate metabolic process (GO:0005975)” in the biological processes category, and three “membrane (GO:0016020, GO:0016021, and GO:0031224)” in the cellular component category were among the top 20 GO terms (Figure 4). The DEGs in the “carbohydrate metabolic process” were analyzed. Among the 34 DEGs, 16 of them were CAZymes and also belonged to the GO term of “hydrolase activity, hydrolyzing O-glycosyl compounds” (Supplementary Table S5). Eight proteases were identified from “hydrolase activity,” including three metalloprotease proteins, two proteinase T-like proteins, two eukaryotic aspartyl proteases, and one serine proteinase (Figure 3C), which could act as CWDEs when P. penicillatus infected the morels. Ten cytochrome P450 (CYP) superfamily proteins were identified in the GO term of oxidoreductase activity, and eight of them were downregulated (Figure 3D). The KEGG pathway analysis showed that 11 pathways were significantly enriched (p < 0.05) in P. penicillatus, including five “carbohydrate metabolism,” three “amino acid metabolism,” two “lipid metabolism,” and one “nucleotide metabolism.” Among them, “starch and sucrose metabolism (ko00500)” and “amino sugar and nucleotide sugar metabolism (ko00520)” were the two most significantly enriched pathways (Figure 4; Supplementary Table S5). These results indicate that carbohydrate metabolism plays an important role in the pathogenicity of P. penicillatus on M. sextelata, which was similar to the results of GO category enrichment analysis. on 3 and 6 dpi, respectively. Among them, 63 genes were common DEGs in the two infection stages (Figure 5; Supplementary Table S6). Notably, all the 63 common DEGs were upregulated during the two infection stages. Among them, a laccase-2 gene was the most highly upregulated (log2fc = 5.9; Supplementary Table S6). Interestingly, chitin recognition protein, cyclin-dependent protein kinase, caffeine-induced death protein 2, and allergen Asp f 15 precursor involved in immune responses were identified in these common DEGs (Supplementary Table S6). A total of 85 upregulated and five downregulated genes were found to be the specific DEGs on 3 dpi in M. sextelata, while 164 upregulated and 22 downregulated genes were specifically identified on 6 dpi. A cyclin was downregulated on 3 dpi and expressed at normal levels on 6 dpi (Supplementary Table S6). Functional Analysis of Differentially Expressed Genes in M. sextelata GO category enrichment analysis was conducted in more detail, and 77 and 71 GO terms were identified (p < 0.05) on 3 and 6 dpi, respectively (Supplementary Table S7). “Oxidoreductase activity (GO:0016491),” “oxidation-–reduction process (GO:0055114),” “carboxy-lyase activity (GO:0016831),” and several pathways associated with these GO terms were commonly enriched in both infection stages (Figure 6; Supplementary Table S7). DEGs in the oxidation-reduction process were screened, and sets of genes associated with immunity were identified, including five CYPs (benzoate 4-monooxygenase CYP-like protein, NADPH-P450 reductase, CYP52A4, CYP52A4, and similar to isotrichodermin C-15 hydroxylase), four tyrosinases, three flavin oxidoreductases, two multicopper oxidases (laccase-2 and bilirubin oxidase), one copper amine oxidase, and one cytochrome c peroxidase (Figure 7; Supplementary Table S8), were all upregulated after infection with P. penicillatus and could be related to the Differentially Expressed Genes in M. sextelata We also analyzed the transcriptional expression of M. sextelata after infection with P. penicillatus. A total of 313 DEGs were identified in M. sextelata by comparing the abundance of transcripts of the inoculated morels with those of the non-inoculated morels, and 90 and 186 DEGs were identified A B FIGURE 4 \| Functional enrichment analyses for DEGs in P. penicillatus during the infection process. (A) GO enrichment analyses. GO term enrichment values of p are indicated on the x-axis. BP, biological process; MF, molecular function; CC, cellular component. (B) KEGG enrichment analyses. The x-axis represents the rich factor. The size of the dot indicates the number of DEGs involved in the pathway. Color bars on the right represent the value of p of the KEGG pathway. DEGs, differentially expressed genes; GO, gene ontology; KEGG, kyoto encyclopedia of genes and genomes. Frontiers in Microbiology \| www.frontiersin.org 6 December 2021 \| Volume 12 \| Article 760444 Yu et al. Morel Interacted With Paecilomyces penicillatus A C B FIGURE 5 \| DEGs in M. sextelata at 3 and 6 days post-inoculation. Volcano plot (A), Venn diagrams (B) and heat maps (C) of the DEGs modulated by P. penicillatus infection in M. sextelata. Blue indicates downregulated DEGs, while red indicates upregulated DEGs in heat maps. ST and SC represent the inoculated and non-inoculated samples, respectively. DEGs, differentially expressed genes; dpi, days post-inoculation. detoxification of P. penicillatus toxins. In addition, a putative apoptosis-inducing protein was upregulated (Figure 7; Supplementary Table S8). GO terms as serine-type peptidase activity (GO:0008236), hormone activity (GO:0005179), pantothenate metabolic process (GO:0015939), and FMN binding (GO:0010181) were specifically enriched on 6 dpi (Supplementary Table S7), indicating that M. sextelata could distinguish between the early and late stages of infection with P. penicillatus. KEGG pathway analysis identified two and nine significantly enriched pathways on 3 and 6 dpi, respectively. Among them, “tyrosine metabolism (ko00350)” and “ABC transporters (ko02010)” were the common enriched pathways in both infection stages (Supplementary Table S7). The DEGs in these two pathways were screened. Five tyrosinases and two ABC transporter family proteins were found to be upregulated in Frontiers in Microbiology \| www.frontiersin.org the infected morel, confirming their important roles in the response of M. sextelata to P. penicillatus (Figure 7; Supplementary Table S8). “Glycolysis/Gluconeogenesis (ko00010),” “pantothenate and CoA biosynthesis (ko00770),” and several other pathways were specifically enriched on 6 dpi (Supplementary Table S7), which is similar to the results of GO category enrichment analysis. DISCUSSION Owing to the lack of resistant varieties and the prevalence of P. penicillatus in many regions, WMD poses a serious threat to the cultivation of morels in China. This study identified a number of genes involved in the regulation of the pathogenicity of P. penicillatus to M. sextelata and the response of M. sextelata 7 December 2021 \| Volume 12 \| Article 760444 Yu et al. Morel Interacted With Paecilomyces penicillatus FIGURE 6 \| Functional enrichment analyses for DEGs in M. sextelata at 3 and 6 days post-inoculation. GO enrichment analyses, GO term enrichment values of p are indicated on the x-axis. BP, biological process; MF, molecular function; CC, cellular component. KEGG enrichment analyses, x-axis represents the rich factor; the dot size indicates the number of DEGs involved in the pathway; Color bars on the right represent the values of p of the KEGG pathway. 3 and 6 dpi represent 3 and 6 days post-inoculation. DEGs, differentially expressed genes; GO, gene ontology; KEGG, kyoto encyclopedia of genes and genomes. to infection with P. penicillatus. We defined how P. penicillatus successfully parasitizes M. sextelata and how M. sextelata encodes its immune response to invasion by P. penicillatus. Mycoparasitic fungi usually produce various fungal CWDEs to lyse the host cell wall, including glucanases, chitinase, and proteases (Gruber and Seidl-Seiboth, 2012). A previous study found that P. penicillatus encodes approximately 300 CAZymes, which could be involved in the degradation of fungal cell walls (Wang et al., 2020). P. penicillatus activated six β-glucanases, three α-mannosidases, and two α-1,6-mannanases when it infected M. sextelata (Figure 3B), indicating that glucanases and mannanases were the primary CAZymes of P. penicillatus that are involved in the degradation of M. sextelata cell walls. Only two chitinases (GH18) were identified in P. penicillatus, even though they were the most annotated CAZyme family in the P. penicillatus genome (Wang et al., 2020). They were all unexpectedly downregulated (Figure 3B). Previous studies have shown that mycoparasitic species of Trichoderma activated chitinase when interacting with A. bisporus (Guthrie and Castle, 2006; Ihrmark et al., 2010), in contrast to the findings of this study. However, the chitinases of T. gamsii were all downregulated when interacting with Fusarium graminearum (Zapparata et al., 2021), which was Frontiers in Microbiology \| www.frontiersin.org consistent with the data of this study. This suggests that P. penicillatus could change the expression of chitinases based on demands of specific situations when infecting other Ascomycetes fungi. In this study, chitin synthase 1 was upregulated in P. penicillatus. One possible explanation for this upregulation is that these chitinases were involved in the cell wall remodeling during P. penicillatus growth. Their downregulation could result in strengthened cell walls during interaction. Another possible explanation is that when in contact with M. sextelata, P. penicillatus suppresses its secretion of fungal chitinases to avoid recognition and the induction of defense reactions in M. sextelata. Several proteases, including subtilase family protease, metalloprotease, and serine proteinase, were activated during the infection process of P. penicillatus (Figure 3C). These genes were reported to be involved in the regulation of cell wall degradation in plants and microorganisms (Feng et al., 2014; Schaller et al., 2018). In particular, the subtilase family proteases are good candidates for this function since they were reported to be induced during mycoparasitism before and during contact with the host in different Trichoderma species (Suárez et al., 2007; Seidl et al., 2009). This indicated that these CWDEs play an important role in the parasitism of P. penicillatus on morels. 8 December 2021 \| Volume 12 \| Article 760444 Yu et al. Morel Interacted With Paecilomyces penicillatus FIGURE 7 \| Heat map of DEGs that responded to P. penicillatus infection in M. sextelata. PCD, programmed cell death; ST and SC represent the inoculated and non-inoculated samples, respectively; dpi, days post-inoculation; blue indicates downregulated DEGs, while yellow indicates upregulated DEGs. KEGG, kyoto encyclopedia of genes and genomes. M. sextelata activated a series of mechanisms to respond once it sensed the invasion of P. penicillatus. A chitin recognition protein and an allergen Asp f 15 precursor in M. sextelata were upregulated during both infection stages (Figure 7; Supplementary Table S8), which could activate the downstream immune response (Sanchez-Vallet et al., 2015). M. sextelata upregulated a caffeine-induced death protein 2-domaincontaining protein and a putative apoptosis-inducing protein, as well as downregulating a cyclin (Figure 7; Supplementary Table S8); these genes are widely reported to promote immunity by initiating programmed cell death (Saiki et al., 2011; Qi and Zhang, 2019). We hypothesize that M. sextelata activates an apoptotic response to escape the proliferation of the pathogen when P. penicillatus invades. A change in the reduction-oxidation status is one of the earliest responses detected when cells are attacked (Frederickson Matika and Loake, 2014). The reduction-oxidation process plays a key role in both the host and pathogen during the interaction of P. penicillatus and M. sextelata, while the regulatory mechanisms differ. The cytochrome P450 family genes in M. sextelata and P. penicillatus showed an opposite trend of expression. Six of the seven cytochrome P450 in P. penicillatus were downregulated, while all five in M. sextelata were upregulated (Figures 3D, 7). A previous study reported that ent-kaurene-derived diterpenoids act as a virulence factor interacting with the bacterium Xanthomonas oryzae pv. oryzicola Frontiers in Microbiology \| www.frontiersin.org and rice (Lu et al., 2015) and inhibited the growth of Staphylococcus aureus and Bacillus subtilis (Fatope et al., 2010). Two ent-kaurene oxidase-like proteins that are cytochrome P450s were downregulated in P. penicillatus, which could lead to increased synthesis of diterpenoid cytotoxins in P. penicillatus (Wang et al., 2012). Once the diterpenoid cytotoxins were secreted into M. sextelata cells, the pathogenicity of P. penicillatus to M. sextelata might be enhanced. Interestingly, one isotrichodermin C-15 hydroxylase gene was downregulated in P. penicillatus, while another was upregulated in M. sextelata. Previous study reported that blocking an isotrichodermin C-15 hydroxylase in Fusarium sporotrichioides resulted in the accumulation of trichothecenes, a virulence factor of Fusarium. spp. (McCormick and Hohn, 1997; Ward et al., 2002). This study also observed increased expression of CYPs superfamily genes in M. sextelata after infection with P. penicillatus, including benzoate 4-monooxygenase CYP-like protein and NADPH-P450 reductase. Benzoate 4-monooxygenase of ascomycete Cochliobolus lunatus was reported to be involved in detoxification of benzoic acid (BA), a key intermediate in metabolism of aromatic compounds in fungi, as well as the detoxification of phenolic compounds (Podobnik et al., 2008; Berne et al., 2012). NADPH-P450 reductase was reported to plays a central role in chemical detoxification and insecticide resistance in insects and fungi (Lian et al., 2011; Xu et al., 2021). This information suggests that these CYPs in M. sextelata might participated 9 December 2021 \| Volume 12 \| Article 760444 Yu et al. Morel Interacted With Paecilomyces penicillatus in the detoxification of toxic metabolites produced by P. penicillatus. We also identified two ABCB1 proteins that were upregulated in M. sextelata following infection with P. penicillatus (Figure 7). We hypothesize that M. sextelata could transport harmful substances secreted by P. penicillatus out of its cells by activating the expression of ABCB1. This has been confirmed by previous studies that ABCB1, a member of the ABC transporter family, plays an important physiological role in protecting the tissues from xenobiotics and endogenous metabolites (Roy et al., 2021). In fungi, tyrosinases are generally associated with the formation and stability of spores, defense and virulence, and melanin production (Halaouli et al., 2006). Once the cells have been damaged, melanin produced by the reaction of tyrosinase with its substrate forms a melanin barrier to protect against the invasion of pathogenic bacteria or their toxins (Janusz et al., 2020). In this study, tyrosinase metabolism was the most enriched KEGG pathway in the infected M. sextelata (Figure 6). Five tyrosinase genes in M. sextelata were upregulated after infection by P. penicillatus (Figure 7). This indicated their importance in M. sextelata against infection with P. penicillatus. Fungal laccase, a multicopper oxidase, is usually induced by a variety of phenolic compounds (Piscitelli et al., 2011) and was reported to have a wide range of functions, including defense against stressful conditions (Lakshmanan and Sadasivan, 2016; Chakraborty et al., 2020). One 1,4-benzoquinone reductase was upregulated in P. penicillatus (Figure 3B; Supplementary Table S4) and could be involved in the downstream synthesis of phenolic compounds. Interestingly, laccase-2 was the most upregulated gene in infected M. sextelata (Figure 7; Supplementary Table S6). A previous study found that the activity of laccase-2 enhanced the resistance of A. bisporus to the T. aggressivum toxin (Sjaarda et al., 2015), which is 3,4-dihydro-8-hydroxy-3-methyl isocoumarin that contains a phenolic hydroxyl and lactone ring (Krupke et al., 2003). This suggests that P. penicillatus could synthesize phenolic toxins and secrete them into M. sextelata, and M. sextelata could detoxify these phenolic toxins by activating laccase expression. However, further experiments are needed to verify this hypothesis. M. sextelata. M. sextelata triggered programmed cell death to prevent the excessive proliferation of P. penicillatus and synthesize cytochrome P450 and laccase to detoxify the P. penicillatus toxins. The melanin barrier formed by tyrosinase was also a possible immune pathway in M. sextelata. Although these possible regulatory pathways proposed in this study still need further verification, they provided a theoretical basis for researchers to breed WMD-resistant varieties and develop prevention and control methods for WMD. CONCLUSION SUPPLEMENTARY MATERIAL During P. penicillatus infection of M. sextelata, cell wall-degrading enzymes, such as glucanases, mannanases, and proteases were probably secreted by P. penicillatus to degrade the cell wall of The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fmicb. 2021.760444/full#supplementary-material DATA AVAILABILITY STATEMENT The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material. AUTHOR CONTRIBUTIONS YY: conceptualization, funding acquisition, validation, investigation, data curation, visualization, writing – original draft preparation, and writing – review and editing. HT: funding acquisition, validation, and investigation. TL: investigation and data curation. LL: resources and investigation. JT: resources. WP: funding acquisition and supervision. All authors contributed to manuscript revision, read, and approved the submitted version. FUNDING This research was funded by National Science Foundation of China (NSFC31901119), Science and Technology Project of Sichuan Province (2021YFYZ0026), SAAS-International Cooperation Project (2021ZSSFGH04), Edible Fungus Innovation Team of Sichuan Province (SCCXTD-2021-7), and Special Fund for Talent Introduction and Training of SAAS (510000-01-114852). Chakraborty, T., Toth, R., Nosanchuk, J. D., and Gacser, A. (2020). Multicopper oxidases in Saccharomyces cerevisiae and human pathogenic fungi. J. Fungi 6:56. doi: 10.3390/jof6020056 Chen, C., Li, Q., Huang, W., Wang, J., Fu, R., Luo, X., et al. (2017). Effects of Morchella white mold disease on soil fungal community structure. Microbiology 44, 2652–2659. doi: 10.13344/j.microbiol.china.170030 Cheng, C. A., Rf, A., Jian, W. A., Xl, A., Xc, A., Qiang, L. B., et al. (2021). Genome sequence and transcriptome profiles of pathogenic fungus Paecilomyces penicillatus reveal its interactions with edible fungus Morchella importuna. Computational and structural. Biotechnol. J. 19, 2607–2617. doi: 10.1016/j.csbj.2021.04.065 REFERENCES Ashburner, M., Ball, C. A., Blake, J. A., Botstein, D., and Cherry, J. M. (2000). Gene ontology: tool for the unification of biology. Gene Ontol. Cons. Nature Gene. 25, 25–29. doi: 10.1038/75556 Berne, S., Podobnik, B., Zupanec, N., Novak, M., Krasevec, N., Turk, S., et al. (2012). Virtual screening yields inhibitors of novel antifungal drug target, benzoate 4-monooxygenase. J. Chem. Inf. Model. 52, 3053–3063. doi: 10.1021/ ci3004418 Frontiers in Microbiology \| www.frontiersin.org 10 December 2021 \| Volume 12 \| Article 760444 Yu et al. Morel Interacted With Paecilomyces penicillatus Dahlstrom, J. L., Smith, J. E., and Weber, N. S. (2000). Mycorrhiza-like interaction by Morchella with species of the Pinaceae in pure culture synthesis. Mycorrhiza 9, 279–285. doi: 10.1007/PL00009992 Deng, K., Lan, X., Fang, Q., Li, M., Xie, G., and Xie, L. (2021). Untargeted metabolomics reveals alterations in the primary metabolites and potential pathways in the vegetative growth of Morchella sextelata. Front. Mol. Biosci. 8:632341. doi: 10.3389/fmolb.2021.632341 Du, X. H., Zhao, Q., O'Donnell, K., Rooney, A. P., and Yang, Z. L. (2012). Multigene molecular phylogenetics reveals true morels (Morchella) are especially species-rich in China. Fungal Genet. Biol. 49, 455–469. doi: 10.1016/j. fgb.2012.03.006 Fatope, M. O., Varma, G. B., Alzri, N. M., Marwah, R. G., and Nair, R. S. (2010). Ent-Kaurene diterpenoids from Blepharispermum hirtum. Chem. Biodivers. 7, 1862–1870. doi: 10.1002/cbdv.200900389 Feng, T., Nyffenegger, C., Hojrup, P., Vidal-Melgosa, S., Yan, K. P., Fangel, J. U., et al. (2014). Characterization of an extensin-modifying metalloprotease: N-terminal processing and substrate cleavage pattern of Pectobacterium carotovorum Prt1. Appl. Microbiol. Biotechnol. 98, 10077–10089. doi: 10.1007/ s00253-014-5877-2 Foulongne-Oriol, M., Rodier, A., and Savoie, J. M. (2012). Relationship between yield components and partial resistance to Lecanicillium fungicola in the button mushroom, Agaricus bisporus, assessed by quantitative trait locus mapping. Appl. Environ. Microbiol. 78, 2435–2442. doi: 10.1128/AEM.07554-11 Frederickson Matika, D. E., and Loake, G. J. (2014). Redox regulation in plant immune function. Antioxid. Redox Signal. 21, 1373–1388. doi: 10.1089/ ars.2013.5679 Fu, Y., Wang, X., Li, D., Liu, Y., Song, B., Zhang, C., et al. (2016). Identification of resistance to wet bubble disease and genetic diversity in wild and cultivated strains of Agaricus bisporus. Int. J. Mol. Sci. 17:1568. doi: 10.3390/ ijms17101568 Gea, F. J., Navarro, M. J., Santos, M., Dianez, F., and Carrasco, J. (2021). Control of fungal diseases in mushroom crops while dealing with fungicide resistance. Rev. Micro. 9:585. doi: 10.3390/microorganisms9030585 Gruber, S., and Seidl-Seiboth, V. (2012). Self versus non-self: fungal cell wall degradation in Trichoderma. Microbiology 158, 26–34. doi: 10.1099/ mic.0.052613-0 Guthrie, J. L., and Castle, A. J. (2006). Chitinase production during interaction of Trichoderma aggressivum and Agaricus bisporus. Can. J. Microbiol. 52, 961–967. doi: 10.1139/w06-054 Halaouli, S., Asther, M., Sigoillot, J. C., Hamdi, M., and Lomascolo, A. (2006). Fungal tyrosinases: new prospects in molecular characteristics, bioengineering and biotechnological applications. J. Appl. Microbiol. 100, 219–232. doi: 10.1111/j.1365-2672.2006.02866.x He, X. L., Peng, W. H., Miao, R. Y., Tang, J., Chen, Y., Liu, L. X., et al. (2017). White mold on cultivated morels caused by Paecilomyces penicillatus. FEMS Microbiol. Lett. 364:364. doi: 10.1093/femsle/fnx037 Ihrmark, K., Asmail, N., Ubhayasekera, W., Melin, P., Stenlid, J., and Karlsson, M. (2010). Comparative molecular evolution of trichoderma chitinases in response to mycoparasitic interactions. Evol. Bioinformatics Online 6, 1–26. doi: 10.4137/ ebo.s4198 Janusz, G., Pawlik, A., Swiderska-Burek, U., Polak, J., Sulej, J., Jarosz-Wilkolazka, A., et al. (2020). Laccase properties, physiological functions, and evolution. Int. J. Mol. Sci. 21:966. doi: 10.3390/ijms21030966 Kage, U., Kumar, A., Dhokane, D., Karre, S., and Kushalappa, A. C. (2016). Functional molecular markers for crop improvement. Crit. Rev. Biotechnol. 36, 917–930. doi: 10.3109/07388551.2015.1062743 Kanehisa, M., Goto, S., Kawashima, S., Okuno, Y., and Hattori, M. (2004). The KEGG resource for deciphering the genome. Nucleic Acids Res. 32, D277–D280. doi: 10.1093/nar/gkh063 Krupke, O. A., Castle, A. J., and Rinker, D. L. (2003). The north American mushroom competitor, Trichoderma aggressivum F. aggressivum, produces antifungal compounds in mushroom compost that inhibit mycelial growth of the commercial mushroom Agaricus bisporus. Mycol. Res. 107, 1467–1475. doi: 10.1017/s0953756203008621 Lakshmanan, D., and Sadasivan, C. (2016). Trichoderma viride laccase plays a crucial role in defense mechanism against antagonistic organisms. Front. Microbiol. 7:741. doi: 10.3389/fmicb.2016.00741 Lian, L. Y., Widdowson, P., McLaughlin, L. A., and Paine, M. J. (2011). Biochemical comparison of Anopheles gambiae and human NADPH P450 Frontiers in Microbiology \| www.frontiersin.org reductases reveals different 2'-5'-ADP and FMN binding traits. PLoS One 6:e20574. doi: 10.1371/journal.pone.0020574 Liu, Q., Ma, H., Zhang, Y., and Dong, C. (2018). Artificial cultivation of true morels: current state, issues and perspectives. Crit. Rev. Biotechnol. 38, 259–271. doi: 10.1080/07388551.2017.1333082 Love, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550. doi: 10.1186/s13059-014-0550-8 Lu, X., Hershey, D. M., Wang, L., Bogdanove, A. J., and Peters, R. J. (2015). An ent-kaurene-derived diterpenoid virulence factor from Xanthomonas oryzae pv. Oryzicola. New Phytol. 206, 295–302. doi: 10.1111/nph.13187 Masaphy, S. (2010). Biotechnology of morel mushrooms: successful fruiting body formation and development in a soilless system. Biotechnol. Lett. 32, 1523–1527. doi: 10.1007/s10529-010-0328-3 McCormick, S. P., and Hohn, T. M. (1997). Accumulation of Trichothecenes in liquid cultures of a Fusarium sporotrichioides mutant lacking a functional Trichothecene C-15 hydroxylase. Appl. Environ. Microbiol. 63, 1685–1688. doi: 10.1128/aem.63.5.1685-1688.1997 Mei, H., Qingshan, W., Baiyintala, and Wuhanqimuge (2019). The wholegenome sequence analysis of Morchella sextelata. Sci. Rep. 9:15376. doi: 10.1038/s41598-019-51831-4 Ower, R. D., Mills, G. L., and Malachowski, J. A. (1986). Cultivation of morchella. United States patent US4,866,878. 19 Sep 1989. Peng, W., Tang, J., He, X., Chen, Y., and Tan, H. (2016). Status analysis of morel artificial cultivation in Sichuan. Edible Med. Mushrooms 24, 145–150. Piscitelli, A., Giardina, P., Lettera, V., Pezzella, C., Sannia, G., and Faraco, V. (2011). Induction and transcriptional regulation of laccases in fungi. Curr. Genomics 12, 104–112. doi: 10.2174/138920211795564331 Podobnik, B., Stojan, J., Lah, L., Krasevec, N., Seliskar, M., Rizner, T. L., et al. (2008). CYP53A15 of Cochliobolus lunatus, a target for natural antifungal compounds. J. Med. Chem. 51, 3480–3486. doi: 10.1021/jm800030e Qi, F., and Zhang, F. (2019). Cell cycle regulation in the plant response to stress. Front. Plant Sci. 10:1765. doi: 10.3389/fpls.2019.01765 Roy, S., Breakspear, A., Cousins, D., Torres-Jerez, I., Jackson, K. J., Kumar, A., et al. (2021). Three common symbiotic ABC-B transporters in Medicago truncatula are regulated by a NIN-independent branch of the symbiosis signalling pathway. Mol. Plant-Microbe Interact. doi: 10.1094/ MPMI-02-21-0036-R [Epub ahead of print] Saiki, S., Sasazawa, Y., Imamichi, Y., Kawajiri, S., Fujimaki, T., Tanida, I., et al. (2011). Caffeine induces apoptosis by enhancement of autophagy via PI3K/ Akt/mTOR/p70S6K inhibition. Autophagy 7, 176–187. doi: 10.4161/auto.7.2.14074 Sanchez-Vallet, A., Mesters, J. R., and Thomma, B. P. (2015). The battle for chitin recognition in plant-microbe interactions. FEMS Microbiol. Rev. 39, 171–183. doi: 10.1093/femsre/fuu003 Schaller, A., Stintzi, A., Rivas, S., Serrano, I., Chichkova, N. V., Vartapetian, A. B., et al. (2018). From structure to function - a family portrait of plant subtilases. New Phytol. 218, 901–915. doi: 10.1111/nph.14582 Seidl, V., Song, L., Lindquist, E., Gruber, S., Koptchinskiy, A., Zeilinger, S., et al. (2009). Transcriptomic response of the mycoparasitic fungus Trichoderma atroviride to the presence of a fungal prey. BMC Genomics 10:567. doi: 10.1186/1471-2164-10-567 Sjaarda, C. P., Abubaker, K. S., and Castle, A. J. (2015). Induction of lcc2 expression and activity by Agaricus bisporus provides defence against Trichoderma aggressivum toxic extracts. Microb. Biotechnol. 8, 918–929. doi: 10.1111/1751-7915.12277 Suárez, M., Vizcaíno, J., Llobell, A., and Monte, E. (2007). Characterization of genes encoding novel peptidases in the biocontrol fungus Trichoderma harzianum CECT 2413 using the TrichoEST functional genomics approach. Curr. Genet. 51, 331–342. doi: 10.1007/s00294-007-0130-5 Tan, H., Kohler, A., Miao, R., Liu, T., Zhang, Q., Zhang, B., et al. (2019). Multi-omic analyses of exogenous nutrient bag decomposition by the black morel Morchella importuna reveal sustained carbon acquisition and transferring. Environ. Microbiol. 21, 3909–3926. doi: 10.1111/1462-2920.14741 Tan, H., Liu, T., Yu, Y., Tang, J., Jiang, L., Martin, F. M., et al. (2021a). Morel production related to soil microbial diversity and evenness. Microbiol. Spectr. 9:e00229-21. doi: 10.1128/Spectrum.00229-21 Tan, H., Yu, Y., Tang, J., Liu, T., Miao, R., Huang, Z., et al. (2021b). Build your own mushroom soil: Microbiota succession and nutritional accumulation 11 December 2021 \| Volume 12 \| Article 760444 Yu et al. Morel Interacted With Paecilomyces penicillatus in semi-synthetic substratum drive the fructification of a soil-saprotrophic morel. Front. Microbiol. 12:656656. doi: 10.3389/fmicb.2021.656656 Trapnell, C. R. A., Goff, L., Pertea, G., Kim, D., and Kelley, D. R. (2012). Differential gene and transcript expression analysis of RNA-seq experiments with TopHat and cufflinks. Nat. Protoc. 7, 562–578. doi: 10.1038/ nprot.2012.016 Wang, Q., Hillwig, M. L., Wu, Y., and Peters, R. J. (2012). CYP701A8: a rice ent-kaurene oxidase paralog diverted to more specialized diterpenoid metabolism. Plant Physiol. 158, 1418–1425. doi: 10.1104/pp.111.187518 Wang, X., Peng, J., Sun, L., Bonito, G., Guo, Y., Li, Y., et al. (2020). Genome sequencing of Paecilomyces Penicillatus provides insights into its phylogenetic placement and Mycoparasitism mechanisms on Morel mushrooms. Pathogens 9:834. doi: 10.3390/pathogens9100834 Ward, T. J., Bielawski, J. P., Kistler, H. C., Sullivan, E., and O'Donnell, K. (2002). Ancestral polymorphism and adaptive evolution in the trichothecene mycotoxin gene cluster of phytopathogenic Fusarium. Proc. Natl. Acad. Sci. 99, 9278–9283. doi: 10.1073/pnas.142307199 Wingfield, B. D., Bills, G. F., Dong, Y., Huang, W., Nel, W. J., Swalarsk-Parry, B. S., et al. (2018). IMA genome-F 9: draft genome sequence of Annulohypoxylon stygium, aspergillus mulundensis, Berkeleyomyces basicola (syn. Thielaviopsis basicola), Ceratocystis smalleyi, two Cercospora beticola strains, Coleophoma cylindrospora, Fusarium fracticaudum, Phialophora cf. hyalina, and Morchella septimelata. IMA Fungus 9, 199–223. doi: 10.5598/imafungus.2018.09.01.13 Xu, J. W., Liao, C. C., Hung, K. C., Wang, Z. Y., Tung, Y. T., and Wu, J. H. (2021). Proteomics reveals Octyl Gallate as an environmentally friendly wood preservative leading to reactive oxygen species-driven metabolic inflexibility Frontiers in Microbiology \| www.frontiersin.org and growth inhibition in white-rot fungi (Lenzites betulina and Trametes versicolor). J. Fungi. 7:145. doi: 10.3390/jof7020145 Zapparata, A., Baroncelli, R., Brandstrom Durling, M., Kubicek, C. P., Karlsson, M., Vannacci, G., et al. (2021). Fungal cross-talk: an integrated approach to study distance communication. Fungal Genet. Biol. 148:103518. doi: 10.1016/j. fgb.2021.103518 Zhao, Q. (2021). The 6th National Morchella Conference. Available at: https:// www.emushroom.net/news/show-33505.html (Accessed April 28, 2021). Conflict of Interest: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Publisher’s Note: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. Copyright © 2021 Yu, Tan, Liu, Liu, Tang and Peng. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. 12 December 2021 \| Volume 12 \| Article 760444
https://openalex.org/W4362527880	https://figshare.com/articles/journal_contribution/Supplementary_Figure_S8_from_Metformin_Synergizes_with_BCL-XL_BCL-2_Inhibitor_ABT-263_to_Induce_Apoptosis_Specifically_in_p53-Defective_Cancer_Cells/22508524/1/files/39970354.pdf	Malagasy	null	Supplementary Figure S8 from Metformin Synergizes with BCL-XL/BCL-2 Inhibitor ABT-263 to Induce Apoptosis Specifically in p53-Defective Cancer Cells	null	2,023	cc-by	187	A Supplementary Figure S8 C B p53 GAPDH HCT-116 p53+/+ Met ABT - + - + - - + + p-ATM (S1981) GAPDH HCT-116 p53+/+ HCT-116 p53-/- Met ABT - + - + - - + + - + - + - - + + A Supplementary Figure S8 C B p53 GAPDH HCT-116 p53+/+ Met ABT - + - + - - + + p-ATM (S1981) GAPDH HCT-116 p53+/+ HCT-116 p53-/- Met ABT - + - + - - + + - + - + - - + + A Supplementary Figure S8 Supplementary Figure S8 Supplementary Figure S8 A C B p53 GAPDH HCT-116 p53+/+ Met ABT - + - + - - + + p-ATM (S1981) GAPDH HCT-116 p53+/+ HCT-116 p53-/- Met ABT - + - + - - + + - + - + - - + + B p53 GAPDH HCT-116 p53+/+ Met ABT - + - + - - + + C p-ATM (S1981) GAPDH HCT-116 p53+/+ HCT-116 p53-/- Met ABT - + - + - - + + - + - + - - + + C B HCT-116 p53+/+
https://openalex.org/W3109119186	https://www.repository.cam.ac.uk/bitstream/1810/343656/3/fmicb-12-667790.pdf	English	null	Using Whole Genome Sequences to Investigate Adenovirus Outbreaks, Including Five Deaths in a Haematopoietic Stem Cell Transplant Unit	medRxiv (Cold Spring Harbor Laboratory)	2,020	cc-by	8,529	ORIGINAL RESEARCH published: 02 July 2021 doi: 10.3389/fmicb.2021.667790 Citation: Myers CE, Houldcroft CJ, Roy S, Margetts BK, Best T, Venturini C, Guerra-Assunção JA, Williams CA, Williams R, Dunn H, Hartley JC, Rao K, Rolfe KJ and Breuer J (2021) Using Whole Genome Sequences to Investigate Adenovirus Outbreaks in a Hematopoietic Stem Cell Transplant Unit. Front. Microbiol. 12:667790. doi: 10.3389/fmicb.2021.667790 Using Whole Genome Sequences to Investigate Adenovirus Outbreaks in a Hematopoietic Stem Cell Transplant Unit Chloe E. Myers 1, Charlotte J. Houldcroft 2, Sunando Roy 3, Ben K. Margetts 4, Timothy Best 5, Cristina Venturini 3, Jose A. Guerra-Assunção 3, Charlotte A. Williams 3, Rachel Williams 3, Helen Dunn 5, John C. Hartley 5, Kanchan Rao 5, Kathryn J. Rolfe 1 and Judith Breuer 3,5* Chloe E. Myers 1, Charlotte J. Houldcroft 2, Sunando Roy 3, Ben K. Margetts 4, Timothy Best 5, Cristina Venturini 3, Jose A. Guerra-Assunção 3, Charlotte A. Williams 3, Rachel Williams 3, Helen Dunn 5, John C. Hartley 5, Kanchan Rao 5, Kathryn J. Rolfe 1 and Judith Breuer 3,5* 1 Cambridge Clinical Microbiology and Public Health Laboratory, Public Health England, Cambridge, United Kingdom, 2 Department of Medicine, University of Cambridge, Cambridge, United Kingdom, 3 Division of Infection and Immunity, University College London, London, United Kingdom, 4 Division of Infection, Immunity and Inﬂammation, Great Ormond Street Institute of Child Health, University College London, London, United Kingdom, 5 Department of Microbiology, Virology and Infection Prevention and Control, Great Ormond Street Hospital for Children National Health Service Foundation Trust, London, United Kingdom Specialty section: Specialty section: This article was submitted to Virology, a section of the journal Frontiers in Microbiology This article was submitted to Virology, a section of the journal Frontiers in Microbiology Received: 14 February 2021 Accepted: 05 May 2021 Published: 02 July 2021 Keywords: adenovirus, epidemiology, whole genome sequencing, pediatric infectious disease, nosocomial transmission A recent surge in human mastadenovirus (HAdV) cases, including ﬁve deaths, amongst a haematopoietic stem cell transplant population led us to use whole genome sequencing (WGS) to investigate. We compared sequences from 37 patients collected over a 20- month period with sequences from GenBank and our own database of HAdVs. Maximum likelihood trees and pairwise differences were used to evaluate genotypic relationships, paired with the epidemiological data from routine infection prevention and control (IPC) records and hospital activity data. During this time period, two formal outbreaks had been declared by IPC, while WGS detected nine monophyletic clusters, seven were corroborated by epidemiological evidence and by comparison of single-nucleotide polymorphisms. One of the formal outbreaks was conﬁrmed, and the other was not. Of the ﬁve HAdV-associated deaths, three were unlinked and the remaining two considered the source of transmission. Mixed infection was frequent (10%), providing a sentinel source of recombination and superinfection. Immunosuppressed patients harboring a high rate of HAdV positivity require comprehensive surveillance. As a consequence of these ﬁndings, HAdV WGS is being incorporated routinely into clinical practice to inﬂuence IPC policy contemporaneously. Edited by: Qiwei Zhang, Jinan University, China Jinan University, China Reviewed by: David W. Dyer, University of Oklahoma Health Sciences Center, United States Kanako Koyanagi, Hokkaido University, Japan Correspondence: Judith Breuer j.breuer@ucl.ac.uk Reviewed by: David W. Dyer, University of Oklahoma Health Sciences Center, United States Kanako Koyanagi, Hokkaido University, Japan Correspondence: Judith Breuer j.breuer@ucl.ac.uk INTRODUCTION Clinical infections caused by human mastadenoviruses (HAdVs) are associated with signiﬁcant morbidity (10–89%) and mortality (6–70%) in the immunocompromised host (Echavarría, 2008). Risk factors for poor outcome include pediatric patients (who are susceptible to primary infection), unrelated donor stem cell transplants (SCTs), graft-vs.-host disease, T-cell depletion of graft, and certain immunosuppressive drug regimens (Shields et al., 1985; Runde et al., 2001; Chakrabarti et al., 2004). July 2021 \| Volume 12 \| Article 667790 Frontiers in Microbiology \| www.frontiersin.org Adenovirus and Infection Control Myers et al. The burden of HAdV infection is signiﬁcant; within the pediatric oncology population, HAdV has been reported to account for 15% of all diarrhoeal cases (Mhaissen et al., 2017). Amongst pediatric patients undergoing hematopoietic stem cell transplant (HSCT), HAdV viremia and stool shedding were found in 15 and 42% of patients, respectively (Hiwarkar et al., 2013; Kosulin et al., 2018). As non-enveloped viruses, HAdVs can be resistant to standard alcohol cleaning regimens and can survive as clinically infectious particles for up to 4 weeks (Gordon et al., 1993). Nosocomial transmission has been frequently reported in the literature (Russell et al., 2006; Rutala et al., 2006); however, the identiﬁcation of these outbreaks is likely to be under-reported due to the limitations of existing HAdV typing protocols that are performed infrequently and target only small regions of selected genes (Seto et al., 2013). sequencing of isolates from the HSCT unit to determine what proportion was transmitted. Using the WGS data, we documented the genetic relatedness between isolates and described the possible transmission events. These ﬁndings can be used to interrupt HAdV transmission dynamics and should be used to further develop routine IPC policy and ultimately improved patient care. Frontiers in Microbiology \| www.frontiersin.org Context and Ethics Great Ormond Street Hospital (GOSH) is a 350-bed, pediatric tertiary referral center. Due to the immunocompromised status of patients referred here, over 60% of beds are single room isolation facilities. In addition to those patients who are symptomatic, “high-risk” patients—those who are admitted for hematological transplant or congenital immunodeﬁciencies—are screened weekly and on admission for gastrointestinal infection using polymerase chain reaction (PCR). The PCR methods used by the GOSH diagnostic laboratory have been described previously (Houldcroft et al., 2018). Residual diagnostic samples were collected from patients with PCR conﬁrmed HAdV infection. The PCR cycle threshold (CT) values provided a comparable semiquantitative indicator of viral titer. The use of these samples for research was approved by The National Research Ethics Service Committee London—Fulham (reference: 17/LO/1530). The clinical data were extracted from the hospital databases by the GOSH Digital Research Environment (DRE) team and linked to an anonymized patient number. Advances in whole genome sequencing (WGS) have provided valuable insights into the molecular epidemiology of a number of key hospital pathogens (Brodrick et al., 2016; Eyre et al., 2017; Brown et al., 2019; Roy et al., 2019). This has been well-illustrated recently in the context of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), where the application in real time has allowed prompt feedback supporting epidemiological links and the utility of the existing IPC policies (Meredith et al., 2020). y g p Speciﬁcally, within our population, a tertiary pediatric referral center in which 30% of patients are immunocompromised, HAdV is one of the leading causes of viral gastroenteritis, comprising 44% of all infections (Brown et al., 2016). Over the last ﬁnancial year (2019–2020), there were 642 new HAdV detections, from any sample site, 99 of which were viremias (local audit data; all patient groups). Adenoviremia signiﬁcantly decreases the probability of survival in children following HSCT and also increases the duration of inpatient hospital stay with an associated ﬁnancial burden (Faden et al., 2005; Mattner et al., 2008; Hiwarkar et al., 2013; Swartling et al., 2015). Deﬁnitions, Patients, and Samples Deﬁnitions, Patients, and Samples A HSCT unit nosocomial outbreak is suspected when any new detection of HAdV infection is identiﬁed in a child who was negative on admission screening. Further information on the routine management of outbreaks is provided in Supplementary Material. For surveillance reporting, healthcare acquired infection (HCAI) is deﬁned as a positive diagnostic sample ≥48 h post-admission and community-acquired infection (CAI) deﬁned as a positive diagnostic sample within 48 h of admission and no healthcare contact in the preceding 14 days. There is a wide range of adenovirus infections: primary new infections with a second strain, reactivation of a previously known infection, reactivation of a previously unknown quiescent infection, or a mixture of these. Due to these potential overlapping scenarios, the routine epidemiological data can only suggest that a speciﬁc infection may be healthcare, as opposed to community, associated but cannot conﬁrm or refute it. Adenovirus typing to species or serotype level may refute a cross-infection hypothesis, but it does not provide adequate discrimination to conﬁrm cross infection, for which WGS is necessary (Houldcroft et al., 2018). A total of 169 samples from 74 patients were included in this study (Supplementary Table 1). All patients were known to have either a congenital or acquired immunodeﬁciency, and therefore considered high risk. As part of this investigation, 11 outbreak samples (n = 8 patients) were identiﬁed as two clusters by IPC [infection control cluster one (ICC 1) patients: 54, 55, 56, 57, 62, and 68, and infection control cluster two (ICC 2) patients: 40 and 38] and 37 non-outbreak samples (n = 29 patients, including HCAI and CAI infections) were sequenced and analyzed with a local database of HAdV sequences (127 sequences from 37 patients). Extensive eﬀorts are employed to prevent adenovirus infection, modiﬁed to account for local suspected transmission routes. In our unit, we have implemented rigorous IPC policies including environmental screening (Pankhurst et al., 2014; Cloutman-Green et al., 2015), but transmission is still suspected. In our hospital, over a 20-month period, seven HAdV outbreaks have been investigated by the IPC team, two of them were associated with the HSCT unit, and there had been ﬁve adenovirus-associated deaths. SureSelect Bait Design and Sequencing Methods allowing high-throughput HAdV WGS directly from clinical samples have been developed (Depledge et al., 2011; Batty et al., 2013; Houldcroft et al., 2018). These methods SureSelect Bait Design and Sequencing To further understand the routes of transmission and enable further development of the infection control policy, we undertook extensive epidemiological investigation and July 2021 \| Volume 12 \| Article 667790 2 Adenovirus and Infection Control Myers et al. provide a proof of concept that WGS oﬀers the resolution required to conﬁrm nosocomial transmission of HAdV; however, there were technical improvements to be made with species C viruses (85/107 clinical samples) yielding lower quality sequences (Houldcroft et al., 2018). 120-mer baits (version 2) were redesigned, using an in-house Perl script with a tiling factor of 12× (each position in a given genome is covered by 12 unique bait designs) against all whole HAdV sequences (487) in GenBank (accessed on 24 January 2018). The bait design was uploaded to SureDesign, and biotinylated RNA oligonucleotides (baits) were synthesized by Agilent Technologies, Santa Clara, California (Agilent Technologies, 2021). FIGURE 1 \| Incidence of new HAdV cases diagnosed using PCR by the diagnostic laboratory. (A) The total number of HAdV cases identiﬁed by GOSH increased each year but did not demonstrate any seasonality. (B) The proportion of new positives that were documented as HCAI are shown in red and (C) the proportion of new positives identiﬁed from patients admitted to the high-risk HSCT unit are highlighted in black. Each column represents a calendar month. HAdV, human mastadenovirus; PCR, polymerase chain reaction; GOSH, Great Ormond Street Hospital; HCAI, healthcare-acquired infection. Quality control of sample DNA, library preparation using the SureSelectXT Illumina paired-end protocol, and sequencing on an Illumina MiSeq sequencer were performed as described earlier (Houldcroft et al., 2018), except the utilization of the SureSelectXT low input kit. Base calling and sample demultiplexing were performed as standard for the MiSeq platform, generating paired FASTQ ﬁles for each sample. Genome Mapping, Assembly, and Phylogenetic Analysis HAdV, human mastadenovirus; PCR, polymerase chain reaction; GOSH, Great Ormond Street Hospital; HCAI, healthcare-acquired infection. FIGURE 1 \| Incidence of new HAdV cases diagnosed using PCR by the diagnostic laboratory. (A) The total number of HAdV cases identiﬁed by GOSH increased each year but did not demonstrate any seasonality. (B) The proportion of new positives that were documented as HCAI are shown in red and (C) the proportion of new positives identiﬁed from patients admitted to the high-risk HSCT unit are highlighted in black. Each column represents a calendar month. HAdV, human mastadenovirus; PCR, polymerase chain reaction; GOSH, Great Ormond Street Hospital; HCAI, healthcare-acquired infection. Genome Mapping, Assembly, and Phylogenetic Analysis Sequences for all 169 samples were assembled using a reference- based pipeline in CLC Genomics Workbench version 12.0.1 (QIAGEN, Hilden, Germany); the detailed methodology can be found in Supplementary Figure 1. Brieﬂy, all reads were quality trimmed and the adaptor sequences were removed. The trimmed reads were mapped to a reference database (n = 103), where 90% of each read mapped with a minimum of 90% identity, the best reference match was used to assign a genotype to each sample. If mapped reads generated a good match to more than one genotype, suggesting a mixed infection, samples underwent further investigation (Supplementary Figures 1–5 and Supplementary Tables 2–6). Once a sample had been assigned a genotype, a second pipeline was then used to quality trim, re-map to the best reference match with a length and similarity fraction of 0.8, before extracting a consensus sequence. The areas of low coverage (<10-fold) were assigned the ambiguity symbol N. Robust consensus sequences are required for the downstream analysis; therefore, only samples achieving ≥90% genome coverage and ≥100-fold average read depth (quality cut- oﬀ) were included in further analysis. Consensus sequences (GenBank accessions MW686757–MW686857) were aligned, and phylogenies were constructed using CLC Genomics Workbench (version 12.0.1) (Supplementary Material). Pairwise single-nucleotide variant counts were computed using Molecular Evolutions Genetics Analysis (MEGA) software version 6 (Kumar et al., 2016). FIGURE 1 \| Incidence of new HAdV cases diagnosed using PCR by the diagnostic laboratory. (A) The total number of HAdV cases identiﬁed by GOSH increased each year but did not demonstrate any seasonality. (B) The proportion of new positives that were documented as HCAI are shown in red and (C) the proportion of new positives identiﬁed from patients admitted to the high-risk HSCT unit are highlighted in black. Each column represents a calendar month. HAdV, human mastadenovirus; PCR, polymerase chain reaction; GOSH, Great Ormond Street Hospital; HCAI, healthcare-acquired infection. FIGURE 1 \| Incidence of new HAdV cases diagnosed using PCR by the diagnostic laboratory. (A) The total number of HAdV cases identiﬁed by GOSH increased each year but did not demonstrate any seasonality. (B) The proportion of new positives that were documented as HCAI are shown in red and (C) the proportion of new positives identiﬁed from patients admitted to the high-risk HSCT unit are highlighted in black. Each column represents a calendar month. Epidemiological Support of Phylogenetic Clusters Inter-species reference sequences (A12, A18, and A61) were also removed from phylogeny A and from phylogeny E (F40) to aid visualization of patient samples. The shape of each node correlates with the sample type and color, the ward on which the sample was taken. Reference sequences are identiﬁed by their HAdV genotype and GenBank Accession. A bootstrap threshold of 80% is shown. Any clinical samples from at least two patients, with a bootstrap support >90, were considered a cluster. The parsimony informative site for each phylogeny is as follows: (A) 2979/34444, (B) 7054/36263, (C) 3104/36332, (D) 1348/36046, and (E) 4455/34469. For phylogeny D, there are only three sequences, out of which two are identical. In this case, the sites reported are only the variable sites between the two groups. NPA, nasopharyngeal aspirate; BAL, bronchoalveolar lavage; ET aspirate, endotracheal aspirate. FIGURE 2 \| Maximum likelihood phylogenies of adenovirus full genome sequences, organized by species: (A) species A viruses, (B) species B viruses, (C) species C viruses, (D) species E, and (E) species F viruses. Sequences were aligned and maximum likelihood phylogenies generated using CLC Genomics Workbench (version 12.0.1), 500 bootstraps. Clinical samples are labeled according to their anonymized patient number (PtX) and specimen number (_SX). Additional samples from sequentially sampled patients have been collapsed. Inter-species reference sequences (A12, A18, and A61) were also removed from phylogeny A and from phylogeny E (F40) to aid visualization of patient samples. The shape of each node correlates with the sample type and color, the ward on which the sample was taken. Reference sequences are identiﬁed by their HAdV genotype and GenBank Accession. A bootstrap threshold of 80% is shown. Any clinical samples from at least two patients, with a bootstrap support >90, were considered a cluster. The parsimony informative site for each phylogeny is as follows: (A) 2979/34444, (B) 7054/36263, (C) 3104/36332, (D) 1348/36046, and (E) 4455/34469. For phylogeny D, there are only three sequences, out of which two are identical. In this case, the sites reported are only the variable sites between the two groups. NPA, nasopharyngeal aspirate; BAL, bronchoalveolar lavage; ET aspirate, endotracheal aspirate. FIGURE 2 \| Maximum likelihood phylogenies of adenovirus full genome sequences, organized by species: (A) species A viruses, (B) species B viruses, (C) species C viruses, (D) species E, and (E) species F viruses. Frontiers in Microbiology \| www.frontiersin.org Epidemiological Support of Phylogenetic Clusters Where there were suﬃcient HAdV sequences to clearly distinguish local UK variants from non-UK reference sequences (species A, C, and F), monophyletic clusters, deﬁned as groups comprising two or more samples from at least two patients arising from a common ancestral node, with bootstrap support ≥90% were used to identify the putative outbreaks. Where there were insuﬃcient sequences to provide adequate phylogenetic July 2021 \| Volume 12 \| Article 667790 Frontiers in Microbiology \| www.frontiersin.org 3 Myers et al. Adenovirus and Infection Control Myers et al. Adenovirus and Infection Control FIGURE 2 \| Maximum likelihood phylogenies of adenovirus full genome sequences, organized by species: (A) species A viruses, (B) species B viruses, (C) species C viruses, (D) species E, and (E) species F viruses. Sequences were aligned and maximum likelihood phylogenies generated using CLC Genomics Workbench (version 12.0.1), 500 bootstraps. Clinical samples are labeled according to their anonymized patient number (PtX) and specimen number (_SX). Additional samples from sequentially sampled patients have been collapsed. Inter-species reference sequences (A12, A18, and A61) were also removed from phylogeny A and from phylogeny E (F40) to aid visualization of patient samples. The shape of each node correlates with the sample type and color, the ward on which the sample was taken. Reference sequences are identiﬁed by their HAdV genotype and GenBank Accession. A bootstrap threshold of 80% is shown. Any clinical samples from at least two patients, with a bootstrap support >90, were considered a cluster. The parsimony informative site for each phylogeny is as follows: (A) 2979/34444, (B) 7054/36263, (C) 3104/36332, (D) 1348/36046, and (E) 4455/34469. For phylogeny D, there are only three sequences, out of which two are identical. In this case, the sites reported are only the variable sites between the two groups. NPA, nasopharyngeal aspirate; BAL, bronchoalveolar lavage; ET aspirate, endotracheal aspirate. FIGURE 2 \| Maximum likelihood phylogenies of adenovirus full genome sequences, organized by species: (A) species A viruses, (B) species B viruses, (C) species C viruses, (D) species E, and (E) species F viruses. Sequences were aligned and maximum likelihood phylogenies generated using CLC Genomics Workbench (version 12.0.1), 500 bootstraps. Clinical samples are labeled according to their anonymized patient number (PtX) and specimen number (_SX). Additional samples from sequentially sampled patients have been collapsed. Epidemiological Support of Phylogenetic Clusters Sequences were aligned and maximum likelihood phylogenies generated using CLC Genomics Workbench (version 12.0.1), 500 bootstraps. Clinical samples are labeled according to their anonymized patient number (PtX) and specimen number (_SX). Additional samples from sequentially sampled patients have been collapsed. Inter-species reference sequences (A12, A18, and A61) were also removed from phylogeny A and from phylogeny E (F40) to aid visualization of patient samples. The shape of each node correlates with the sample type and color, the ward on which the sample was taken. Reference sequences are identiﬁed by their HAdV genotype and GenBank Accession. A bootstrap threshold of 80% is shown. Any clinical samples from at least two patients, with a bootstrap support >90, were considered a cluster. The parsimony informative site for each phylogeny is as follows: (A) 2979/34444, (B) 7054/36263, (C) 3104/36332, (D) 1348/36046, and (E) 4455/34469. For phylogeny D, there are only three sequences, out of which two are identical. In this case, the sites reported are only the variable sites between the two groups. NPA, nasopharyngeal aspirate; BAL, bronchoalveolar lavage; ET aspirate, endotracheal aspirate. resolution (species B HAdVs), two or more samples within a monophyletic cluster with short branch lengths (<0.002) were used. Timelines for each monophyletic cluster were visualized using the ggplot2 library (Wickham, 2016), incorporating patient admission data and HAdV PCR positivity. Patients within a cluster were deﬁned as epidemiologically supported if they were present on the same ward or unit becoming positive during the incubation period of the virus [median 5.6 days (95% CI 4.8–6.3) based on respiratory disease (Lessler et al., 2009)] and unsupported if they became positive during admissions July 2021 \| Volume 12 \| Article 667790 Frontiers in Microbiology \| www.frontiersin.org 4 Adenovirus and Infection Control Myers et al. Phylogenetic Investigation of Outbreaks and Deaths to completely diﬀerent wards or had no links with any other sequenced patient. To substantiate nosocomial transmission, maximum likelihood phylogenies were constructed (Figure 2). Nine monophyletic clusters were identiﬁed (Figures 2A–E) and summarized in Table 1. One of these clusters (A31 Cluster 4) had been previously identiﬁed phylogenetically (Houldcroft et al., 2018). Statistical Analysis Statistical tests were performed using two-tailed tests at the 5% signiﬁcance level within GraphPad Prism version 8.3.0 for Mac OS, GraphPad Software, San Diego, CA, United States, www.graphpad.com (Supplementary Material). y g y Four of the six patients from ICC 1 (patients 55, 57, 62, and 68) were phylogenetically linked (Figure 2A, A31 Cluster 1). Additionally, two patients (Pt69 and Pt70) documented as having HCAI, for whom no source of infection had previously been identiﬁed were linked phylogenetically to ICC 1. One patient (Pt69) involved in monophyletic A31 Cluster 1 had a mixed HAdV infection. The sequence data obtained from Pt38 were of insuﬃcient quality to conﬁrm or refute transmission with Pt40 of ICC 2 phylogenetically; however, one of these patients, Pt40, was phylogenetically linked to another patient, Pt42, who had a concurrent HAdV-C2 infection. WGS identiﬁed an additional six monophyletic clusters, involving 16 patients who had not previously been identiﬁed by standard IPC follow-up (Figure 2 and Table 1). Frontiers in Microbiology \| www.frontiersin.org RESULTS Burden of Infection and Viral Genotypes Routine reporting of the ﬁrst HAdV PCR-positive cases by the diagnostic laboratory between 2015 and 2019 is summarized in Figure 1. As expected, no seasonality was observed (Brown et al., 2016); however, cases increased each year with a marked rise between 2017 and 2018 (Figure 1A). The proportion of new cases that were attributed to HCAI during this time period rose from 12% (2015–2017) to 23% (2018–2019) (Figure 1B). On average, 18% of new positives are detected from patients admitted to the high-risk HSCT unit (Figure 1C). Out of ﬁve patients who died (patients 52, 53, 59, 60, and 61) as a result of or in association with overwhelming HAdV infection (Supplementary Table 1), two were linked to a monophyletic cluster. Patient 52 was found to have a mixed genotype (C2 and A31) infection that was dominated by a phylogenetically unlinked C2 (Figure 2C) but with a minority subpopulation of A31 that clustered with two other patients (11 and 73, A31 Cluster 2, Figure 2A). Patient 60 had a single C5 infection that clustered with Pt54 (C5 Cluster 3, Figure 2C). The remaining three patients had phylogenetically unlinked single genotype infections. Between August 2017 and April 2019, IPC identiﬁed 11 outbreak samples (from ICC 1 and ICC 2) and 37 non-outbreak samples (including HCAI and CAI cases) from high-risk patients. Sequences were analyzed with 121 previously sequenced samples. A total of 169 clinical samples containing HAdV genotypes A31 (14%), B3 (2%), B11 (1%), C1 (17%), C2 (21%), C5 (18%), C89 (8%), E4 (1%), and F41 (3%) from 74 patients with either localized (e.g., eye, respiratory, and digestive) or disseminated infection were included. Seven of these samples (4%) failed to sequence and 17 (10% of patients) had mixed HAdV infections (Supplementary Table 1 and Supplementary Figure 6). Traditional Epidemiology, Contact Tracing Supported Phylogeny Assignments Previous work has shown that infections can be linked over many years (Houldcroft et al., 2018), e.g., A31 Cluster 4 potentially transmitted over a 5-year period (temporal relationship shown in Figure 3D). Using the new samples sequenced as part of this investigation, we conﬁrmed the same with a putative transmission cluster occurring over a 3-year period (A31 Cluster 2, Figure 3B) and suggested a prolonged transmission also occurring amongst other HAdV species, e.g., B3 Cluster 1 over a 4-year period (Figure 3E). Improved Sequencing Quality p q g y Of 169 samples across all genotypes, 56 (42%) of 132 HAdV genomes passed the quality cut-oﬀusing version 1 baits and 46 (85%) of 54 HAdV genomes passed using version 2 baits (Supplementary Table 1 and Supplementary Figure 7). Both genome coverage and on-target reads were statistically signiﬁcantly improved for species C viruses using version 2 baits (Supplementary Figure 8). This was despite similar species C viral titers in samples between bait groups (Supplementary Figure 9). Average read depth improved but remained signiﬁcantly lower for species C viruses regardless of the baits used (P = 0.0002 version 1 baits vs. P = 0.05 version 2 baits). Wards B and E were associated with all nine monophyletic clusters (Figures 2, 3). As well as sharing clinical teams, these wards are joined by the same corridor and share facilities (dirty utility, kitchen, parents’ room and laundry); for this reason, they are considered one HSCT unit. Five clusters containing ward B and E patients had temporal links with each other: A31 Cluster 1, A31 Cluster 3, C1 Cluster 2, C5 Cluster 3, and F41 Cluster 1 (Figures 3A,C,G–I and Table 1), supporting nosocomial transmission. Out of the seven samples that failed to sequence, four were sequenced using version 1 baits. Three samples failed using version 2 baits, one with a viral load that had previously been successful. There was an insuﬃcient sample for repeat testing. Using an estimated linear regression model, it is predicted that samples with HAdV CT values of ≤34 would generate a ≤100- fold read depth, with 95% certainty (Supplementary Figure 10). Two clusters contained one patient with no temporal links: Pt11 in A31 Cluster 2 and Pt8 within July 2021 \| Volume 12 \| Article 667790 5 Adenovirus and Infection Control Myers et al. TABLE 1 \| Summary of monophyletic clusters identiﬁed by maximum likelihood phylogeny using whole HAdV genome sequences. Improved Sequencing Quality HAdV type, sequence cluster number Sample code ICC number IPC record Ward involved Temporally relateda Diversity within clusterb Conclusion A31 Cluster 1 Pt69_S1_A31 – HCAI not linked to outbreak B Yes 0 Conﬁrmed transmission cluster Pt70_S1 – HCAI not linked to outbreak E Yes Pt68_S1 1 Chronic HAdV–ICC 1 investigated E Yes Pt62_S1 1 HCAI–ICC 1 investigated B Yes Pt57_S1 1 HCAI–ICC 1 investigated B Yes Pt55_S1 1 HCAI–ICC 1 investigated E Yes A31 Cluster 2 Pt11_S1 – Not classiﬁed B No 6 Likely transmission, unconﬁrmed Pt73_S1_A31 – HCAI E Yes 3 Conﬁrmed transmission cluster Pt52_S1_A31* – HCAI E Yes A31 Cluster 3 Pt41_S1 – CAI E Yes 1 Conﬁrmed transmission cluster Pt46_S1 – CAI B Yes A31 Cluster 4 Pt24_S1 – Not classiﬁed E Yes 0–1 Conﬁrmed transmission cluster Pt8_S1 – Not classiﬁed B No Pt18_S1 – Probable HCAI B Yes Pt3_S1 – Not classiﬁed B Yes B3 Cluster 1 Pt27_S1 – Not classiﬁed A No 13 Unlikely transmission cluster Pt73_S1_B3 Marked as long-term carriage from previous admission E No C2 Cluster 1 Pt52_S1_C2* – HCAI E No 6 Likely transmission, unconﬁrmed Pt66_S1_C2 – HCAI E No C1 Cluster 2 Pt42_S1 – CAI E Yes 0 Conﬁrmed transmission cluster Pt40_S1_S2 2 HCAI–ICC 2 investigated E Yes C5 Cluster 3 Pt60_S1* – Not classiﬁed E Yes 0 Conﬁrmed transmission cluster Pt54_S1 1 HCAI–ICC 1 investigated E Yes F41 Cluster 1 Pt71_S1 – HCAI E Yes 0 Conﬁrmed transmission cluster Pt63_S1 – HCAI B Yes Pt65_S1 – CAI B Yes 3 Conﬁrmed transmission cluster HAdV, human mastadenovirus; ICC, Infection Control Cluster; IPC, Infection Prevention and Control; HCAI, healthcare-acquired infection; CAI, community acquired infection. aTemporally related, HAdV PCR positive whilst admitted to same/linked ward. bDiversity within cluster, expressed as the number of pairwise differences/single-nucleotide polymorphisms (SNPs) across the whole genome. Indicates patients who died from or in association with overwhelming HAdV infection. p y , p bDiversity within cluster, expressed as the number of pairwise differences/single-nucleotide polymorphisms (SNPs) across the whole genome. Indicates patients who died from or in association with overwhelming HAdV infection. Epidemiologically linked monophyletic clusters were found to have ≤3 SNPs diﬀerence. Improved Sequencing Quality This corroborated the number of diﬀerences previously deﬁned for within host (≤2 SNPs, Figure 4C) and directly transmitted viruses (Houldcroft et al., 2018), and further supports the nosocomial transmission between clusters A31 Cluster 1, A31 Cluster 3, C1 Cluster 2, C5 Cluster 3, and F41 Cluster 1 (Table 1). Of the two patients (Pt54 and Pt60) involved in C5 Cluster 3, Pt60 died; however, this patient was admitted and HAdV PCR positive several months before Pt54 (Figure 3H), suggesting that Pt60 may have been the source of this nosocomial infection. A31 Cluster 4 (Figures 3B,D). The two remaining clusters (C2 Cluster 1, Figure 3F; B3 Cluster 1, Figure 3E) did not share any temporal links with each other. A31 Cluster 4 (Figures 3B,D). The two remaining clusters (C2 Cluster 1, Figure 3F; B3 Cluster 1, Figure 3E) did not share any temporal links with each other. Conﬁdence in Genomic Links Using Pairwise Distances To quantify phylogenetic relationships, pairwise diﬀerences (single-nucleotide polymorphisms [SNPs] between aligned consensus sequences) were calculated and grouped according to their epidemiological support (Figure 4). July 2021 \| Volume 12 \| Article 667790 Frontiers in Microbiology \| www.frontiersin.org 6 Adenovirus and Infection Control Myers et al. FIGURE 3 \| Temporal relationship between HAdV patients forming monophyletic clusters. Clinical data were extracted from hospital databases by the GOSH Digital Research Environment (DRE) team and linked to an anonymized patient number. The ggplot2 library was used to visualize the data between positive adenovirus PCR results. Sequenced samples are indicated by a vertical black line, HAdV PCR-positive samples by a red cross, HAdV PCR-negative samples by a black cross and ward stays by colored rectangles. (A) A31 Cluster 1, (B) A31 Cluster 2, (C) A31 Cluster 3, (D) A31 Cluster 4, (E) B3 Cluster 1, (F) C2 Cluster 1, (G) C1 Cluster 2, (H) C5 Cluster 3, and (I) F41 Cluster 1. FIGURE 3 \| Temporal relationship between HAdV patients forming monophyletic clusters. Clinical data were extracted from hospital databases by the GOSH Digital Research Environment (DRE) team and linked to an anonymized patient number. The ggplot2 library was used to visualize the data between positive adenovirus PCR results. Sequenced samples are indicated by a vertical black line, HAdV PCR-positive samples by a red cross, HAdV PCR-negative samples by a black cross and ward stays by colored rectangles. (A) A31 Cluster 1, (B) A31 Cluster 2, (C) A31 Cluster 3, (D) A31 Cluster 4, (E) B3 Cluster 1, (F) C2 Cluster 1, (G) C1 Cluster 2, (H) C5 Cluster 3, and (I) F41 Cluster 1. One monophyletic cluster, A31 Cluster 4, despite containing one patient who was not linked temporally to the other patients (Figure 3D), diﬀered by 0–1 SNPs suggesting nosocomial transmission (Table 1). With potentially unsampled patients or environmental intermediates, it is not possible to determine the route of transmission from these patients to the other patients within these clusters. two sequences (14 SNPs were found between unrelated patient samples of the same genotype within species C viruses). Only three B3 infections were identiﬁed during the study period (92 and 93 SNPs separated this cluster from unrelated Pt14). The close clustering of these samples is therefore likely to be a result of the few publicly available UK HAdV B3 sequences. Conﬁdence in Genomic Links Using Pairwise Distances Further sequencing of HAdV-B3 genotypes is therefore required to substantiate the relationship found between these two patients. The remaining monophyletic clusters A31 Cluster 2, B3 Cluster 1, and C2 Cluster 1 included at least one sequence separated by 6–13 SNPs. This range does not overlap with the number of SNPs found between unrelated patients of the same genotype (14–415 SNPs). However, the interpretation of the epidemiological links between cases based on the genomic data alone can be diﬃcult because we do not currently understand the species-speciﬁc substitution rate of HAdV in chronically infected immunosuppressed patients. Frontiers in Microbiology \| www.frontiersin.org DISCUSSION This study demonstrates the major threat to immunosuppressed children that HAdV presents, providing a snapshot of a larger problem as only a minority of viruses were sequenced. With the level of infection present, it is not possible to recognize HAdV outbreaks contemporaneously using the conventional PCR methods. New infections identiﬁed by the diagnostic B3 Cluster 1 is the monophyletic cluster with the least support, without a temporal relationship between patients (admitted and tested 4 years apart) and 13 SNPs between the B3 Cluster 1 is the monophyletic cluster with the least support, without a temporal relationship between patients (admitted and tested 4 years apart) and 13 SNPs between the July 2021 \| Volume 12 \| Article 667790 7 Adenovirus and Infection Control Myers et al. FIGURE 4 \| Pairwise differences equating to single-nucleotide polymorphisms (SNPs) between clinical samples included in this study. (A) Pairwise differences were plotted by the following categories: “intra patient,” differences between samples taken from the same patient; “intra cluster (epidemiologically supported),” differences between samples from different patients within a monophyletic cluster that are temporally linked (HAdV PCR positive whilst admitted to same/linked ward) by the admission data; “intra cluster (not epidemiologically supported),” differences between samples from different patients within a monophyletic cluster that are not linked temporally by the admission data; “inter patient (same genotype),” differences between samples from different patients within the same genotype; “inter species,” differences between samples from different patients within the same species but different genotypes and “intra species,” differences between samples from different patients between species. Median values with 95% CIs are superimposed and plotted in red. (B) Differences between the ﬁrst four categories to aid visualization. (C) Highlighted precise count of pairwise differences found within the ﬁrst four categories. FIGURE 4 \| Pairwise differences equating to single-nucleotide polymorphisms (SNPs) between clinical samples included in this study. Conclusion The utility of the WGS data is entirely dependent on the quality of sequences obtained. Poor genome coverage and low read depth generate tenuous links to other patients and any SNPs identiﬁed are poorly supported. Newly designed baits improved sequencing success (85% using version 2 baits vs. 42% using version 1 baits), especially for species C viruses (Supplementary Figure 8). Species C HAdVs produce the most severe clinical manifestations amongst immunosuppressed patients, particularly those undergoing HSCT (Lion et al., 2010; Qurashi et al., 2011). The clinical utility of WGS technology for IPC purposes has begun to be realized for a number of important pathogens (Brodrick et al., 2016; Eyre et al., 2017; Brown et al., 2019; Roy et al., 2019; Meredith et al., 2020). In this study, we have demonstrated that using a sensitive technique, HCAI and mixed infection remains a signiﬁcant problem despite the application of thorough IPC containment strategies. PCR alone fails to identify HAdV co-infection and transmission events, which can have catastrophic consequences amongst high-risk patients. In order to combat this deﬁcit, HAdV WGS is being implemented into routine diagnostics within this tertiary referral center. The high incidence of mixed infection (10% of patients) identiﬁed in this study highlights the superiority of WGS over PCR. Not only were patients incorrectly linked using standard methods, demonstrated by ICC 1, but HAdV PCR-positive patients would also go unnoticed if they later acquired a second HAdV co-infection. These secondary HAdV co-infections are important not only because of their role in transmission events, but also they could have diﬀerent tissue tropisms, clinical consequences and provide a sentinel source of recombination (Lukashev et al., 2008; Walsh et al., 2009; Hashimoto et al., 2018). DISCUSSION WGS identiﬁed six patient clusters and 15 patient transmission events (patients not within an ICC) which were not identiﬁed using standard IPC investigations and whilst only focusing on high-risk patients from two wards. Rapid sequencing is now possible within 72 h. This has already been shown to impact the IPC management of SARS-CoV-2 in real time (Meredith et al., 2020) and would have impacted on the additional six patient clusters identiﬁed here. known to facilitate prolonged viral shedding (Kosulin et al., 2016). Patient 8 was, however, discharged at least 6 months prior to related patients, and sample positivity by PCR was absent for over 3 years. This suggests ongoing, undetected nosocomial transmission by unsampled intermediates which could include the environment, other patients, visiting relatives or staﬀmembers. Outbreaks amongst vulnerable patients already harboring a high rate of HAdV positivity requires comprehensive surveillance. As a result, we have begun to implement routine HAdV WGS into a standard diagnostic algorithm to improve the clinical care. DATA AVAILABILITY STATEMENT The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material. DISCUSSION (A) Pairwise differences were plotted by the following categories: “intra patient,” differences between samples taken from the same patient; “intra cluster (epidemiologically supported),” differences between samples from different patients within a monophyletic cluster that are temporally linked (HAdV PCR positive whilst admitted to same/linked ward) by the admission data; “intra cluster (not epidemiologically supported),” differences between samples from different patients within a monophyletic cluster that are not linked temporally by the admission data; “inter patient (same genotype),” differences between samples from different patients within the same genotype; “inter species,” differences between samples from different patients within the same species but different genotypes and “intra species,” differences between samples from different patients between species. Median values with 95% CIs are superimposed and plotted in red. (B) Differences between the ﬁrst four categories to aid visualization. (C) Highlighted precise count of pairwise differences found within the ﬁrst four categories. FIGURE 4 \| Pairwise differences equating to single-nucleotide polymorphisms (SNPs) between clinical samples included in this study. (A) Pairwise differences were plotted by the following categories: “intra patient,” differences between samples taken from the same patient; “intra cluster (epidemiologically supported),” differences between samples from different patients within a monophyletic cluster that are temporally linked (HAdV PCR positive whilst admitted to same/linked ward) by the admission data; “intra cluster (not epidemiologically supported),” differences between samples from different patients within a monophyletic cluster that are not linked temporally by the admission data; “inter patient (same genotype),” differences between samples from different patients within the same genotype; “inter species,” differences between samples from different patients within the same species but different genotypes and “intra species,” differences between samples from different patients between species. Median values with 95% CIs are superimposed and plotted in red. (B) Differences between the ﬁrst four categories to aid visualization. (C) Highlighted precise count of pairwise differences found within the ﬁrst four categories. Frontiers in Microbiology \| www.frontiersin.org Frontiers in Microbiology \| www.frontiersin.org July 2021 \| Volume 12 \| Article 667790 8 Adenovirus and Infection Control Myers et al. laboratory by PCR (Figure 1) are also likely to be over- or under- represented; without genomic information, it is impossible to know whether these are genuine new infections or reactivation of clinically quiescent virus (Kosulin et al., 2016; Dhingra et al., 2019). ETHICS STATEMENT This tertiary referral center already has robust IPC precautions in place for high-risk patients; single room isolation with en-suite facilities and environmental screening post-discharge (Cloutman-Green et al., 2015). Despite these precautions, we were still able to conﬁrm that 21 patients (28%) were involved in a nosocomial transmission cluster, and 15 patients (20%) ordinarily would have gone unnoticed or unlinked (Table 1). Unidentiﬁed acquisition/transmission events can have clinically signiﬁcant consequences including prolonged hospital stay, missed treatment opportunities and even death. Sequence data allowed us to investigate ﬁve HAdV-associated deaths as part of the Trust Patient Safety Review process. This is important for all the cases of HCAI where the patient may have come to harm. In this cohort, no death was related to a virus shown to be acquired at GOSH. The two HAdV-associated deaths that were involved in transmission events here (Pt52 and Pt60) appeared to be the index case in their respective clusters (Figures 2A,C, 3B,H). The studies involving human participants were reviewed and approved by The National Research Ethics Service Committee London—Fulham (reference: 17/LO/1530). AUTHOR CONTRIBUTIONS CM, CH, KJR, and JB planned and designed the study. SR designed the sequencing baits. TB, CM, HD, JH, and KR provided the clinical data. CW, RW, and CM sequenced the samples. CM, SR, BM, and JG-A analyzed the data. CM, CH, and JB drafted the manuscript. All authors read, edited, and approved the ﬁnal manuscript. Frontiers in Microbiology \| www.frontiersin.org REFERENCES 33, 1870–1874. doi: 10.1093/molbev/msw054 Cloutman-Green, E., Canales, M., Pankhurst, L., Evenor, T., Malone, D., Klein, N., et al. (2015). Development and implementation of a cleaning standard algorithm to monitor the eﬃciency of terminal cleaning in removing adenovirus within a pediatric hematopoietic stem cell transplantation unit. Am. J. Infect. Control 43, 997–999. doi: 10.1016/j.ajic.2015.05.016 Lessler, J., Reich, N. G., Brookmeyer, R., Perl, T. M., Nelson, K. E., and Cummings, D. A. (2009). Incubation periods of acute respiratory viral infections: a systematic review. Lancet Infect. Dis. 9, 291–300. doi: 10.1016/S1473-3099(09)70069-6 Lion, T., Kosulin, K., Landlinger, C., Rauch, M., Preuner, S., Jugovic, D., et al. (2010). Monitoring of adenovirus load in stool by real-time PCR permits early detection of impending invasive infection in patients after allogeneic stem cell transplantation. Leukemia 24, 706–714. doi: 10.1038/leu.2010.4 Depledge, D. P., Palser, A. L., Watson, S. J., Lai, I. Y. C., Gray, E. R., Grant, P., et al. (2011). Speciﬁc capture and whole-genome sequencing of viruses from clinical samples. PLoS ONE 6:e27805. doi: 10.1371/journal.pone.0027805 Lukashev, A. N., Ivanova, O. E., Eremeeva, T. P., and Iggo, R. D. (2008). Evidence of frequent recombination among human adenoviruses. J. Gen. Virol. 89, 380–388. doi: 10.1099/vir.0.83057-0 Dhingra, A., Hage, E., Ganzenmueller, T., Böttcher, S., Hofmann, J., Hamprecht, K., et al. (2019). Molecular evolution of human adenovirus (HAdV) species C. Sci. Rep. 9:1039. doi: 10.1038/s41598-018-37249-4 Echavarría, M. (2008). Adenoviruses in immunocompromised hosts. Clin. Microbiol. Rev. 21, 704–715. doi: 10.1128/CMR.00 052-07 Mattner, F., Sykora, K. W., Meissner, B., and Heim, A. (2008). An adenovirus type F41 outbreak in a pediatric bone marrow transplant unit: analysis of clinical impact and preventive strategies. Pediatr. Infect. Dis. J. 27, 419–424. doi: 10.1097/INF.0b013e3181658c46 Eyre, D. W., Fawley, W. N., Rajgopal, A., Settle, C., Mortimer, K., Goldenberg, S. D., et al. (2017). Comparison of control of Clostridium diﬃcile infection in six english hospitals using whole-genome sequencing. Clin. Infect. Dis. 65, 433–441. doi: 10.1093/cid/cix338 Meredith, L. W., Hamilton, W. L., Warne, B., Houldcroft, C. J., Hosmillo, M., Jahun, A. S., et al. (2020). Rapid implementation of SARS-CoV- 2 sequencing to investigate cases of health-care associated COVID-19: a prospective genomic surveillance study. Lancet Infect. Dis. 20, 1263–1272. doi: 10.1016/S1473-3099(20)30562-4 Faden, H., Wynn, R. J., Campagna, L., and Ryan, R. M. (2005). Outbreak of adenovirus type 30 in a neonatal intensive care unit. J. Pediatr. 146, 523–527. doi: 10.1016/j.jpeds.2004.11.032 Mhaissen, M. REFERENCES Hashimoto, S., Gonzalez, G., Harada, S., Oosako, H., Hanaoka, N., Hinokuma, R., et al. (2018). Recombinant type human mastadenovirus D85 associated with epidemic keratoconjunctivitis since 2015 in Japan. J. Med. Virol. 90, 881–889. doi: 10.1002/jmv.25041 Agilent Technologies, Inc. (2021). Community Designs (NGS) Infectious Disease Research Designs. Available online at: https://www.agilent.com/en/product/ next-generation-sequencing/community-designs-ngs/infectious-disease- research-designs-536208 Hiwarkar, P., Gaspar, H. B., Gilmour, K., Jagani, M., Chiesa, R., Bennett-Rees, N., et al. (2013). Impact of viral reactivations in the era of pre-emptive antiviral drug therapy following allogeneic haematopoietic SCT in paediatric recipients. Bone Marrow Transplant. 48, 803–808. doi: 10.1038/bmt.2012.221 Batty, E. M., Wong, T. H. N., Trebes, A., Argoud, K., Attar, M., Buck, D., et al. (2013). A modiﬁed RNA-Seq approach for whole genome sequencing of RNA viruses from faecal and blood samples. PLoS ONE. 8:e66129. doi: 10.1371/journal.pone.0066129 Houldcroft, C. J., Roy, S., Morfopoulou, S., Margetts, B. K., Depledge, D. P., Cudini, J., et al. (2018). Use of whole-genome sequencing of adenovirus in immunocompromised pediatric patients to identify nosocomial transmission and mixed-genotype infection. J. Infect. Dis. 218, 1261–1271. doi: 10.1093/infdis/jiy323 Brodrick, H. J., Raven, K. E., Harrison, E. M., Blane, B., Reuter, S., Török, M. E., et al. (2016). Whole-genome sequencing reveals transmission of vancomycin- resistant Enterococcus faecium in a healthcare network. Genome Med. 8:4. doi: 10.1186/s13073-015-0259-7 Kosulin, K., Berkowitsch, B., Matthes, S., Pichler, H., Lawitschka, A., Pötschger, U., et al. (2018). Intestinal adenovirus shedding before allogeneic stem cell transplantation is a risk factor for invasive infection post-transplant. EBioMedicine 28, 114–119. doi: 10.1016/j.ebiom.2017.12.030 Brown, J. R., Roy, S., Shah, D., Williams, C. A., Williams, R., Dunn, H., et al. (2019). Norovirus transmission dynamics in a pediatric hospital using full genome sequences. Clin. Infect. Dis. 68, 222–228. doi: 10.1093/cid/ciy438 Brown, J. R., Shah, D., and Breuer, J. (2016). Viral gastrointestinal infections and norovirus genotypes in a paediatric UK hospital, 2014–2015. J. Clin. Virol. 84, 1–6. doi: 10.1016/j.jcv.2016.08.298 Kosulin, K., Geiger, E., Vécsei, A., Huber, W. D., Rauch, M., Brenner, E., et al. (2016). Persistence and reactivation of human adenoviruses in the gastrointestinal tract. Clin. Microbiol. Infect. 22, 381.e1–381.e8. doi: 10.1016/j.cmi.2015.12.013 Chakrabarti, S., Milligan, D. W., Moss, P. A., and Mautner, V. (2004). Adenovirus infections in stem cell transplant recipients: recent developments in understanding of pathogenesis, diagnosis and management. Leuk Lymphoma 45, 873–885. doi: 10.1080/10428190310001628176 Kumar, S., Stecher, G., and Tamura, K. (2016). MEGA7: molecular evolutionary genetics analysis version 7.0 for bigger datasets. Mol. Biol. Evol. FUNDING This work was supported by the National Institute for Health Research Biomedical Research Centre at Great Ormond Street Hospital for Children NHS Foundation Trust and University College London. Sequencing of 121 samples, making up the pilot study, was funded by an Action Medical Research Grant (grant number GN2424) (CH). Sequencing of subsequent samples was funded by JB who receives funding from the National Institute of Health Research University College London/University College Immediate action needs to be taken to identify the source of HAdV acquisition in these patients in order to understand and halt transmission. Index cases may acquire their HAdV infection outside of hospital, but we have evidence that widely separated samples are linked; Pt8 within A31 Cluster 4, one SNP diﬀerence (Figure 3D). This patient was immunocompromised long term (X-linked lymphoproliferative disease) which is July 2021 \| Volume 12 \| Article 667790 9 Adenovirus and Infection Control Myers et al. for Children Departments of Microbiology, Virology and Infection Prevention and Control to the University College London Infection Bank. We were particularly grateful to John Booth, for his involvement in updating the Digital Research Environment, and Leila Wilson and Colette Smith for reviewing the statistics. London Hospitals NHS Foundation Trust Biomedical Research Centre. The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication. SUPPLEMENTARY MATERIAL We acknowledge the infrastructure support from the University College London Pathogen Genomics Unit, where the sequencing was conducted, and University College London Medical Research We acknowledge the infrastructure support from the University College London Pathogen Genomics Unit, where the sequencing was conducted, and University College London Medical Research Council Centre for Molecular Medical Virology. Clinical samples were provided by Great Ormond Street Hospital The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fmicb. 2021.667790/full#supplementary-material Council Centre for Molecular Medical Virology. Clinical samples were provided by Great Ormond Street Hospital REFERENCES N., Rodriguez, A., Gu, Z., Zhu, H., Tang, L., Sun, Y., et al. (2017). Epidemiology of diarrheal illness in pediatric oncology patients. J. Pediatr. Infect. Dis. Soc. 6, 275–280. doi: 10.1093/jpids/piw050 Gordon, Y. J., Gordon, R. Y., Romanowski, E., and Araullo-Cruz, T. P. (1993). Prolonged recovery of desiccated adenoviral serotypes 5, 8, and 19 from plastic and metal surfaces in vitro. Ophthalmology 100, 1835–1840. doi: 10.1016/S0161-6420(93)31389-8 Pankhurst, L., Cloutman-Green, E., Canales, M., D’Arcy, N., and Hartley, J. C. (2014). Routine monitoring of adenovirus and norovirus within July 2021 \| Volume 12 \| Article 667790 Frontiers in Microbiology \| www.frontiersin.org 10 Myers et al. Adenovirus and Infection Control the health care environment. Am. J. Infect. Control 42, 1229–1232. doi: 10.1016/j.ajic.2014.07.028 Swartling, L., Allard, A., Törlen, J., Ljungman, P., Mattsson, J., and Sparrelid, E. (2015). Prolonged outbreak of adenovirus A31 in allogeneic stem cell transplant recipients. Transpl. Infect. Dis. 17, 785–794. doi: 10.1111/tid. 12443 the health care environment. Am. J. Infect. Control 42, 1229–1232. doi: 10.1016/j.ajic.2014.07.028 Qurashi, Y. M. A. A., Guiver, M., and Cooper, R. J. (2011). Sequence typing of adenovirus from samples from hematological stem cell transplant recipients. J. Med. Virol. 83, 1951–1958. doi: 10.1002/jmv.22204 Walsh, M. P., Chintakuntlawar, A., Robinson, C. M., Madisch, I., Harrach, B., Hudson, N. R., et al. (2009). Evidence of molecular evolution driven by recombination events inﬂuencing tropism in a novel human adenovirus that causes epidemic keratoconjunctivitis. PLoS ONE 4:e5635. doi: 10.1371/journal.pone.000 5635 Roy, S., Hartley, J., Dunn, H., Williams, R., Williams, C. A., and Breuer, J. (2019). Whole-genome sequencing provides data for stratifying infection prevention and control management of nosocomial inﬂuenza A. Clin. Infect. Dis. 69, 1649–1656. doi: 10.1093/cid/ciz020 Runde, V., Ross, S., Trenschel, R., Lagemann, E., Basu, O., Renzing-Köhler, K., et al. (2001). Adenoviral infection after allogeneic stem cell transplantation (SCT): report on 130 patients from a single SCT unit involved in a prospective multi center surveillance study. Bone Marrow Transplant. 28, 51–57. doi: 10.1038/sj.bmt.1703083 Wickham, H. (2016). ggplot2: Elegant Graphics for Data Analysis, 2nd Edn. Cham: Springer. Frontiers in Microbiology \| www.frontiersin.org July 2021 \| Volume 12 \| Article 667790 Conﬂict of Interest: JB has declared that the adenovirus bait designs are used with permission from Agilent. Russell, K. L., Broderick, M. P., Franklin, S. E., Blyn, L. B., Freed, N. E., Moradi, E., et al. (2006). Transmission dynamics and prospective environmental sampling of adenovirus in a military recruit setting. J. Infect. Dis. 194, 877–885. doi: 10.1086/507426 The remaining authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. Rutala, W. A., Peacock, J. E., Gergen, M. F., Sobsey, M. D., and Weber, D. J. (2006). Eﬃcacy of hospital germicides against adenovirus 8, a common cause of epidemic keratoconjunctivitis in health care facilities. Antimicrob. Agents Chemother. 50, 1419–1424. doi: 10.1128/AAC.50.4.1419-1424.2006 Copyright © 2021 Myers, Houldcroft, Roy, Margetts, Best, Venturini, Guerra- Assunção, Williams, Williams, Dunn, Hartley, Rao, Rolfe and Breuer. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Seto, D., Jones, M. S., Dyer, D. W., and Chodosh, J. (2013). Characterizing, typing, and naming human adenovirus type 55 in the era of whole genome data. J. Clin. Virol. 58, 741–742. doi: 10.1016/j.jcv.2013.09.025 Shields, A. F., Hackman, R. C., Fife, K. H., Corey, L., and Meyers, J. D. (1985). Adenovirus infections in patients undergoing bone-marrow transplantation. N. Engl. J. Med. 312, 529–533. doi: 10.1056/NEJM198502283120901 July 2021 \| Volume 12 \| Article 667790 Frontiers in Microbiology \| www.frontiersin.org 11
https://openalex.org/W1972487833	https://europepmc.org/articles/pmc4212660?pdf=render	English	null	Phenotypic and Proteomic Characteristics of Human Dental Pulp Derived Mesenchymal Stem Cells from a Natal, an Exfoliated Deciduous, and an Impacted Third Molar Tooth	Stem cells international	2,014	cc-by	12,578	Hindawi Publishing Corporation Stem Cells International Volume 2014, Article ID 457059, 19 pages http://dx.doi.org/10.1155/2014/457059 Hindawi Publishing Corporation Stem Cells International Volume 2014, Article ID 457059, 19 pages http://dx.doi.org/10.1155/2014/457059 Hindawi Publishing Corporation Stem Cells International Volume 2014, Article ID 457059, 19 pages http://dx.doi.org/10.1155/2014/457059 Hindawi Publishing Corporation Stem Cells International Volume 2014, Article ID 457059, 19 pages http://dx.doi.org/10.1155/2014/457059 Correspondence should be addressed to Murat Kasap; mkasap2008@gmail.com Correspondence should be addressed to Murat Kasap; mkasap2008@gmail.com Received 9 June 2014; Revised 3 September 2014; Accepted 18 September 2014; Published 14 October 2014 Received 9 June 2014; Revised 3 September 2014; Accepted 18 September 2014; Published 14 October 2014 Academic Editor: Pavla Jendelova Copyright © 2014 Gurler Akpinar et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The level of heterogeneity among the isolated stem cells makes them less valuable for clinical use. The purpose of this study was to understand the level of heterogeneity among human dental pulp derived mesenchymal stem cells by using basic cell biology and proteomic approaches. The cells were isolated from a natal (NDPSCs), an exfoliated deciduous (stem cells from human exfoliated deciduous (SHED)), and an impacted third molar (DPSCs) tooth of three different donors. All three stem cells displayed similar features related to morphology, proliferation rates, expression of various cell surface markers, and differentiation potentials into adipocytes, osteocytes, and chondrocytes. Furthermore, using 2DE approach coupled with MALDI-TOF/TOF, we have generated a common 2DE profile for all three stem cells. We found that 62.3 ± 7% of the protein spots were conserved among the three mesenchymal stem cell lines. Sixty-one of these conserved spots were identified by MALDI-TOF/TOF analysis. Classification of the identified proteins based on biological function revealed that structurally important proteins and proteins that are involved in protein folding machinery are predominantly expressed by all three stem cell lines. Some of these proteins may hold importance in understanding specific properties of human dental pulp derived mesenchymal stem cells. Gurler Akpinar,1 Murat Kasap,1,2 Ayca Aksoy,3 Gokhan Duruksu,3 Gulcin Gacar,3 and Erdal Karaoz3,4 1 DEKART Proteomics Laboratory, Kocaeli University Medical School, Umuttepe, 41380 Kocaeli, Turkey 2 Department of Medical Biology, Kocaeli University Medical School, Umuttepe, 41380 Kocaeli, Turkey 3 Department of Stem Cell, Center for Stem Cell and Gene Therapies Research and Practice, Institute of Health Sciences, Umuttepe, 41380 Kocaeli, Turkey 1 DEKART Proteomics Laboratory, Kocaeli University Medical School, Umuttepe, 41380 Kocaeli, Turkey 2 Department of Medical Biology, Kocaeli University Medical School, Umuttepe, 41380 Kocaeli, Turkey 3 Department of Stem Cell, Center for Stem Cell and Gene Therapies Research and Practice, Institute of Health Sciences, Umuttepe, 41380 Kocaeli, Turkey 4Liv Hospital, Center for Regenerative Medicine and Stem Cell Research & Manufacturing (Liv MedCell), Besikta 34340 Istanbul, Turkey Correspondence should be addressed to Murat Kasap; mkasap2008@gmail.com 1. Introduction 0.25% trypsin- EDTA solution was used to detach the cells from the plate for passaging and the detached cells were centrifuged at 2000 ×g for 10 minutes, resuspended in one mL complete medium, counted twice using a Thoma counting chamber, and then plated in 75 cm2 flasks (BD Biosciences, USA) at densities of 1 × 106 cell/flask. Growth medium was replaced every three days over a 10–14-day period. 2.2. Characterization of MSCs by Flow Cytometry. To confirm that MSCs maintain their phenotypic characteristics in vitro, undifferentiated cells were subjected to flow cytometry anal- yses. After passage three, stem cells were harvested and resus- pended in culture medium at a concentration of 106 cells/mL. Flow cytometry was performed by using FACSCalibur (BD Biosciences, USA) according to Kara¨oz et al. (2010) [6]. The data were analyzed with CellQuest software (BD Biosciences, USA). 1. Introduction Although there is a descriptive definition for mesenchy- mal stem cells (MSCs), the extent of heterogeneity within and among MSC lines is overwhelming [4]. This creates a lack of extensive overlap among the studies performed with MSCs. In addition to the genetic background, methods of derivation, growth conditions, the stage of the cell cycle during sample collection, the age and gender of the donor, and the disease status of the donor are the likely factors that contribute to the heterogeneity problem [5]. In general, characterization of MSCs heavily relies on the use of methods such as immunofluorescence microscopy, reverse transcription PCR, and flow cytometry to establish both stem cell identity and function. However, to facilitate stem cell definition Stem cells are undifferentiated cells that can divide, differenti- ate, and self-renew to produce new stem cells in multicellular organisms [1]. They can be used in biomedical research, drug discovery, and toxicity testing, as a model in understanding diseases and more importantly for therapeutic purposes in regenerative medicine [2]. To use stem cells successfully in the aforementioned areas, homogenous populations of stem cells have to be isolated, identified, and characterized. However, given the extent of heterogeneity within and among the stem cell lines, the isolation of homogenous stem cell populations appears to be a challenging task [3]. Stem Cells International 2 Table 1: Immunophenotypic characteristics of NDPSCs, SHED, and hDDP-MSCs. The figures in the table are percentages of CD positive cells. through cellular phenotypic profile, comparative analysis of gene and protein expression studies should be performed. Currently there is no universally accepted and commonly used cellular phenotypic profile for stem cell characterization. Gene expression profiles are preferred due to their relative ease but they vary greatly with the organisms’ state and environment in ways that cannot be easily interpreted. The signature obtained from analysis of the total cell proteome or cell surface proteome (“protein barcodes”) is promising and proteomic approaches can be powerful in characterizing the entire protein profile of stem cell phenotype from different niches. 2. Materials and Methods 2.1. Isolation and Culture of MSCs from Human Dental Pulps (Natal, Deciduous, and Third Molar). Isolation and culture of human dental pulp derived MSCs were performed according to protocols described elsewhere [6]. Briefly, dental pulps of exfoliated deciduous and impacted third molar teeth were collected by cutting around the cement-enamel junction by using sterilized dental fissure burs to reveal the pulp chamber. The recovery of natal dental pulp is different and harder compared to pulp from adult teeth, where teeth were cut around the cementoenamel junction using dental fissure burs to open the pulp chamber and separate the pulp tissue from the crown and root by an excavator. In the recovery of natal teeth pulp, pliers were used to fracture the dental crown into several parts and the dental pulp was uncovered. The pulp tissue of each sample was gently separated from the crown and root by using sterile excavator and digested with colla- genase type I to generate single cell suspensions. MEM-Earle medium containing 15% fetal bovine serum and 100 IU/mL penicillin-100 𝜇g/mL streptomycin was used and the cells were cultured in 25 cm2 plastic tissue culture flasks, incubated at 37∘C in a humidified atmosphere containing 5% CO2. The stem cells were isolated on the basis of their ability to adhere to the culture plates. On the third day, red blood and other nonadherent cells were removed and the medium was replaced to allow further growth. The adherent cells grown to 70% confluency were defined as passage zero (P0) Immunophenotyping characterization of MSCs was per- formed with antibodies against the following human anti- gens: CD3, CD4, CD13, CD14, CD29, CD34, CD44, CD45, CD73, CD90, CD106, CD117, CD146, CD166, HLA-DR, and HLA-ABC (Table 1). The antigen selection was made based on MSC defining criteria set by Dominici et al. (2006) [7]. All antibodies were from BD Biosciences. More than 50% staining was regarded as positive. 2.3. In Vitro Stem Cell Differentiation. To induce adipogenic differentiation 3000 cells/cm2 were seeded onto six-well plates and cultured with Mesencult MSC Basal Medium sup- plemented with 10% adipogenic supplement (StemCell Tech- nologies Inc., Canada) and 1% penicillin/streptomycin for three weeks. The medium was refreshed every two days. Intracellular lipid droplets indicating adipogenic differenti- ation were confirmed by Oil Red O staining.f 2.3. In Vitro Stem Cell Differentiation. 1. Introduction Markers NDPSCs SHED DPSCs CD13 90.87 ± 5.24 89.73 ± 6.76 94.34 ± 8.42 CD14 0.78 ± 0.62 0.78 ± 0.86 0.16 ± 0.07 CD29 99.43 ± 0.28 99.88 ± 0.08 99.44 ± 0.37 CD34 0.93 ± 0.62 0.24 ± 0.08 0.29 ± 0.24 CD44 98.62 ± 1.33 99.80 ± 0.12 99.77 ± 0.26 CD45 0.65 ± 0.42 0.28 ± 0.09 0.12 ± 0.05 CD73 99.69 ± 0.36 99.76 ± 0.36 99.56 ± 0.14 CD90 97.62 ± 3.17 99.73 ± 0.28 99.49 ± 0.32 CD106 1.04 ± 0.70 1.18 ± 0.39 0.14 ± 0.02 CD117 5.68 ± 0.24 1.74 ± 1.72 0.94 ± 0.85 CD146 81.38 ± 7.33 57.91 ± 0.76 54.44 ± 23.89 CD166 99.15 ± 0.97 99.51 ± 0.60 98.59 ± 0.97 HLA-ABC 94.88 ± 5.25 66.89 ± 6.36 83.15 ± 5.65 HLA-DR 0.56 ± 0.39 0.12 ± 0.02 0.17 ± 0.05 To understand the level of heterogeneity among the MSCs, we isolated MSCs from dental pulps of a natal, an exfoliated deciduous, and an impacted third molar tooth of three different donors. The isolated stem cells were then cultured under the same growth conditions and passaged similarly. The cells were compared on the basis of cellular morphology and expression of MSC specific markers and pluripotent transcription factors. In addition, telomerase activity measurements were performed to collect information about age related changes and cellular senescence. Finally, we compared the protein expression profiles of undifferentiated cells by using 2DE gel electrophoresis followed by MALDI- TOF/TOF MS/MS analysis. We identified 61 proteins that were predominantly expressed by all three stem cell lines. We believe that some of these proteins may hold importance in understanding specific properties of human dental pulp derived mesenchymal stem cells. cells. Later passages were labeled accordingly. 0.25% trypsin- EDTA solution was used to detach the cells from the plate for passaging and the detached cells were centrifuged at 2000 ×g for 10 minutes, resuspended in one mL complete medium, counted twice using a Thoma counting chamber, and then plated in 75 cm2 flasks (BD Biosciences, USA) at densities of 1 × 106 cell/flask. Growth medium was replaced every three days over a 10–14-day period. cells. Later passages were labeled accordingly. 2. Materials and Methods To induce adipogenic differentiation 3000 cells/cm2 were seeded onto six-well plates and cultured with Mesencult MSC Basal Medium sup- plemented with 10% adipogenic supplement (StemCell Tech- nologies Inc., Canada) and 1% penicillin/streptomycin for three weeks. The medium was refreshed every two days. Intracellular lipid droplets indicating adipogenic differenti- ation were confirmed by Oil Red O staining.f i For osteogenic differentiation, 3000 cells/cm2 were seeded onto collagen (type I) coated cover-slips in six-well 3 Stem Cells International urea, 2% SDS, 0.375 M Tris Cl pH 8.8, 20% glycerol, and 2% DTT for the first step and the same buffer without DTT but with iodoacetamide (2.5%) for the second step. Following isoelectric focusing, strips were subjected to SDS-PAGE. Precast gels (Criterion TGX Any kD, Bio-Rad, USA) were used with Criterion Dodeca gel running system (Bio-Rad, USA) to minimize gel to gel variation. Gels were stained with Colloidal Coomassie stain (KeraFast, USA) and visualized with VersaDoc4000 MP (Bio-Rad, USA). PDQuest Advance (Bio-Rad, USA) 2DE analysis software was used for com- parison of protein spot profiles. For automated spot detec- tion, parameters used were sensitivity (13.8), spot size scale (3), and minimum peak intensity (258). plates. The differentiation medium was changed twice a week. After four weeks, osteogenic differentiation was assessed by Alizarin Red staining.f To induce chondrogenic differentiation, pelleted micro- mass of 2.5×105 cells was formed by centrifugation at 1300 ×g for 5 minutes and then cultured with chondrogenic medium for two weeks. Medium was changed twice a week. The pellets, fixed with 4% paraformaldehyde and embedded into paraffin, were stained with Haematoxylin-Eosin and Safranin O to assess chondrogenic differentiation. 2.4. Measurement of Telomerase Activity. Telomerase activity was measured by conventional telomeric repeat amplifica- tion protocol (TRAP) using TeloTAGGG PCR-ELISA kit (Roche Diagnostics, Germany) according to the manufac- turer instructions. In brief, after lysis of 2×105 cells, three 𝜇L of supernatant was incubated with premixed buffer for 20 min at 25∘C to show the presumed telomerase activity, and the products were amplified with biotin-labeled primers by PCR. Streptavidin coated ELISA plates were used to quantify the activity by addition of peroxidase specific substrate (3,3󸀠,5,5󸀠- tetramethylbenzidine dihydrochloride; TMB) and reading the optical density at 450 nm. Relative telomerase activity (RTA) was calculated with respect to the control template equivalent to 0.001 mol/𝜇L DNA. 2.7. Image Analysis and Spot Cutting. 2. Materials and Methods The outside edges of all of the images were identically cropped by using automated crop tool of PDQuest Advance software (Bio-Rad, USA). Stain speckles were filtered and the standardized areas of interest from all gels were matched and warped and the quantity of each spot was normalized by the total valid spot intensity. Total spot numbers and volumes within the nor- malized area were determined from the automated analyses. A manual editing tool was used to inspect the determined protein spots detected by the software. Spots that were prone to variation were excluded if they were hard to identify by visual inspection. Every-member matching protein spots were selected among the three mesenchymal stem cells lines. The spots were cut by using automated spot cutting tool, ExQuest spot cutter (Bio-Rad, USA), and disposed into 96- well plates for protein identification. 2.5. Cell Cycle Analysis. Cell cycle was determined with flow cytometry by using BD Cycletest Plus kit (BD Biosciences, USA) following the protocol provided by the manufacturer. 2.6. Protein Extraction and 2DE Gel Electrophoresis. For analysis of proteome profiles, cells were grown in defined media (DMEM, low glucose, pyruvate, Life Tech. Inc., USA). To start cultures, equal number of cells (3 × 105) was seeded into T-75 flasks in triplicate. When 70% confluency was reached the cells were washed with ample amount of ice- cold wash buffer (10 mM Tris-HCl, pH 7.0, 250 mM sucrose) for three times and removed from the plates by scraping. After 10 min centrifugation at 4∘C at 2000 ×g, excess wash buffer was decanted and 250 𝜇L of cell lysis buffer 2DE rehydration buffer (8 M urea, 2 M thiourea, 4% CHAPS, 30 mM Tris pH 8.5, and 1x protease inhibitor cocktail) was added over each cell pellet. To achieve complete lysis, cells were vigorously vortexed for 1 min in lysis buffer and the supernatant containing the soluble protein fraction was obtained by centrifugation at 20.000 ×g for 30 min at 4∘C and stored in lobind tubes (Eppendorf, USA) at −80∘C after being snap-frozen in liquid nitrogen. Protein concentration was determined by using modified Lowry assay with the BSA standard (Bio-Rad, USA). 80 𝜇g of protein was loaded onto immobilized pH gradient strips (IPG) (11 cm, pH 5–8) (Bio- Rad, USA) via passive rehydration. Separation based on isoelectric points was performed by using Protean isoelectric focusing cell using the recommended conditions (Bio-Rad, USA). 2. Materials and Methods Evaluation of proliferation rates by measuring cell viability at various time points indicated that NDPSCs had higher proliferation rate than SHED and DPSCs (Figure 1(b)). carbamidomethyl (Cys) and variable modification being oxi- dation (Met); peptide charge of 1+ and being monoisotopic. Only significant hits, as defined by the MASCOT probability analysis (𝑃< 0.05), were accepted (see Supplementary Material 1 in Supplementary Material available online at http://dx.doi.org/10.1155/2014/457059).i g Classification of proteins was performed by using a freely available classification system PANTHER (http://www.pantherdb.org/) as well as NCBI (http://www .ncbi.nlm.nih.gov/pubmed) and Swiss-Prot/TrEMBL anno- tations (http://www.expasy.org/) [8]. 2.9. Validation of the Selected Proteins by Western Blot Analy- sis. The same protein preparations from the cells used in the 2DE gel experiments were used for Western blotting (WB) to validate the identified proteins by MALDI-TOF/TOF. Protein extracts (30 𝜇g protein for each lane) were mixed with SDS sample buffer (Laemmli solution) and they were subjected to 12% SDS-PAGE gels. The gels were transferred onto nitrocel- lulose membranes (0.45 𝜇m, Bio-Rad) at 30 mA constant cur- rent (TurboBlot, Bio-Rad). To prevent nonspecific bindings, the membranes were blocked for 1 h at room temperature in 5% nonfat dried milk, 1% (v/v) Tween 20 in TBS (50 mM Tris, 150 mM NaCl, adjust pH with HCl to pH 7.6). After blocking, the membranes were incubated overnight at 4∘C with appropriately diluted polyclonal primary antibodies in TBS-T against Galectin-1 (1 : 2000, Santa Cruz, sc-28248), DJ- 1 (1 : 1000, Abcam, ab-4150), HNRNPH1 (1 : 10000, Abcam, ab10374), GAPDH (1 : 2500, Abcam, ab-9485), and mono- clonal UCHL-1 (1 : 5000, Pierce, MAI-46079). GAPDH was used for normalization of each protein sample to ensure equal protein loading. Following three washes with TBS-T, the membranes were incubated with the secondary anti-mouse IgG HRP antibody (Bio-Rad) or anti-rabbit IgG (Bio-Rad) for 1 h at room temperature. Then the proteins were visualized with an ECL Plus Western blotting detection system (GE Healthcare). The bands were quantitated by using Quantity One 1D image analysis software (Bio-Rad). As analyzed by FACS, majority of the stem cells (78%) were in G1 phase of the cell cycle indicating that the cells were growing in size and synthesizing their mRNA and proteins (Figure 2). The cells that were in S and G2 phases were present in lower ratios (12% and 10%, resp.); thus the number of cells undergoing mitotic division was less than the actively growing cells. 2. Materials and Methods A stepwise incremental voltage program was applied to each strip (250 V for 20 min (linear), 4000 V for 2 hr (linear), and 10000 V/hr (rapid)) by using Protean IEF system (Bio-Rad, USA). The strips were then subjected to a two- step equilibration in equilibration buffers containing 6 M 2.8. Identification of Proteins. Protein identification experi- ments were performed at Kocaeli University DEKART pro- teomics laboratory (http://kabiproteomics.kocaeli.edu.tr/) by using ABSCIEX MALDI-TOF/TOF 5800 system. In-gel tryp- tic digestion of the proteins was performed by using an in- gel digestion kit following the recommended protocol by the manufacturer (Pierce, USA). Before deposition onto a MALDI plate, all samples were desalted and concentrated with a 10 𝜇L ZipTipC18 following the recommended protocol (Millipore, USA). Peptides were eluted in a volume of 1 𝜇L using a concentrated solution of 𝛼-CHCA in 50% acetonitrile and 0.1% trifluoroacetic acid in water and spotted onto the MALDI target plate. The TOF spectra were recorded in the positive ion reflector mode with a mass range from 400 to 2000 Da. Each spectrum was the cumulative average of 200 laser shots. The spectra were calibrated with the trypsin autodigestion ion peaks m/z (842.510 and 2211.1046) as internal standards. Ten of the strongest peaks of the TOF spectra per sample were chosen for MS/MS analysis. All of the PMFs were searched in the MASCOT ver- sion 2.5 (Matrix Science) by using a streamline software, ProteinPilot (ABSCIEX, USA), with the following criteria: National Center for Biotechnology Information nonredun- dant (NCBInr); species restriction to H. sapiens; enzyme of trypsin; at least five independent peptides matched; at most one missed cleavage site; MS tolerance set to ±50 ppm and MS/MS tolerance set to ±0.4 Da; fixed modification being 4 Stem Cells International tooth at the age of six and the DPSCs were from a dental pulp of an impacted third molar tooth that belonged to a 27- year-old female. All three stem cell lines were characterized to some extent. Morphological analysis of each stem cell line indicated the presence of cells with large, flattened, or fibroblast-like shape (Figure 1(a)). No noteworthy changes in morphology and growth characteristics were observed throughout the passages. Analysis of the common MSC markers with flow cytometry indicated that CD13, CD29, CD44, CD73, CD90, CD146, CD166, and HLA-ABC were expressed by all three stem cell lines (Table 1). 2. Materials and Methods Age related telomerase shortening due to loss of telom- erase activity is an indicator of stem cell aging [9–11]. We observed a donor-age associated decrease in telomerase activities. The highest telomerase activity was observed with NDPSCs (Figure 3) (considered to be 100% in relative ratio). There was a 30% decrease in telomerase activity of SHED in comparison to NDPSCs. The lowest telomerase activity observed was in DPSCs.f Measurement of differentiation potential is required to assess multipotential characteristics of stem cells. Histo- chemical and immunofluorescence methods were used to demonstrate differentiation potentials of each cell line into adipogenic, osteogenic, and chondrogenic states. During adipogenic differentiation, the lipid droplets enlarged and invaded the entire cytoplasm (Figure 4. (a1), (b1), and (c1)) while during osteogenic differentiation calcium deposits were clearly identified by Alizarin Red staining (Figure 4. (a2), (b2), and (c2)). Chondrogenic differentiation was character- ized by the presence of round cells which resembled hyalin chondrocytes. The presence of GAG was shown by Safranin O staining (Figure 4. (a3), (b3), and (c3)). The same staining procedures were performed with the undifferentiated stem cells and were presented as negative staining in Supplemen- tary Figure 2. 2.10. Statistical Analysis. Flow cytometry, stem cell differenti- ation and telomerase activity measurements were performed on passage three cells. Any data presented showing the results of a set of analysis were carried out on all cell types. The number of experimental replications, both technical and biological was kept at three. A comparative proteomic analysis was performed to identify protein spots that were commonly expressed among the dental pulp derived MSCs. For this purpose stem cells were grown in defined media and protein samples from each MSC type were solubilized and separated by using IPG strips and SDS-PAGE. Following Colloidal Coomassie staining, well resolved and reproducible 2DE gel maps were produced as shown in Figure 5. When it was necessary, statistical analysis of data was carried out by one-way analysis of variance (ANOVA), fol- lowed by Tukey’s test for multiple comparisons to determine the values that were significantly different. Differences were considered statistically significant at 𝑃< 0.05. 3. Results An average of 560 ± 30 well stained spots per analytical gel were detected when the gels were subjected to automated spot detection and analysis. However, the number of well- resolved spots that matched every member was 183 (Master Gel) with an overall mean coefficient of 60.8 percent. The rest of the spots were either not matchable or not expressed MSC lines that were isolated and used throughout the study were from three different donors. The NDPSCs were isolated from dental pulp of a tooth that belonged to a female baby who was born with a tooth. The SHED was isolated from dental pulp of a female juvenile who lost one of her deciduous 5 Stem Cells International NDPSCs SHED DPSCs 100 𝜇m 100 𝜇m 100 𝜇m (a) Days of incubation DPSCs SHED NDPSCs 0 0.5 1 1.5 2 2.5 3 3.5 4 0 5 10 15 20 25 30 Absorbance at 450nm (b) Figure 1: Illustration of MSC morphologies and relevant growth kinetics. (a) Morphology of NDPSCs, SHED, and DPSCs. The cells were at passage three. Images were taken with an inverted microscope (40X). (b) Growth curves for NDPSCs, SHED, and DPSCs over a period of 25 days. The cells were cultured in 96-well plates in triplicate and their growth rates were followed by WST-1 assay, which measures cell viability. DPSCs displayed relatively lower growth rate than SHED and NDPSCs. SHED 100 𝜇m (a) SHED DPSCs 100 𝜇m 100 𝜇m (a) DPSCs 100 𝜇m NDPSCs 100 𝜇m Days of incubation DPSCs SHED NDPSCs 0 0.5 1 1.5 2 2.5 3 3.5 4 0 5 10 15 20 25 30 Absorbance at 450nm Days of incubation 0 0.5 1 1.5 2 2.5 3 3.5 4 0 5 10 15 20 25 30 Absorbance at 450nm (b) Figure 1: Illustration of MSC morphologies and relevant growth kinetics. (a) Morphology of NDPSCs, SHED, and DPSCs. The cells were at passage three. Images were taken with an inverted microscope (40X). (b) Growth curves for NDPSCs, SHED, and DPSCs over a period of 25 days. The cells were cultured in 96-well plates in triplicate and their growth rates were followed by WST-1 assay, which measures cell viability. DPSCs displayed relatively lower growth rate than SHED and NDPSCs. by all three stem cell types. By using PDQuest Advance gel analysis software, changes in spot intensities among these 183 matching spots were compared. 3. Results 6 0 10 20 30 40 50 60 70 80 90 100 G1 G2 S G1 G2 S G1 G2 S NDPSCs SHED DPSCs Cells (%) Figure 2: Cell cycle analysis of NDPSCs, SHED, and DPSCs. Each sample (10,000 cells) was counted in triplicate by a flow cytometer. The values are expressed as percentage mean ± standard deviation. Stem Cells International Stem Cells International 6 6 0 10 20 30 40 50 60 70 80 90 100 G1 G2 S G1 G2 S G1 G2 S NDPSCs SHED DPSCs Cells (%) MSC markers, and in vitro differentiation potentials into adipocytes, osteoblasts, and chondrocytes. Telomerase activ- ities were also measured to monitor the proliferative history of each cell type. Finally, proteome analysis was performed by 2DE gel electrophoresis with MALDI-TOF/TOF MS/MS to create a protein expression pattern to help with under- standing of metabolic pathways that were shared among these MSCs. For proteomic characterization experiments, a better controlled environment was created by using defined media. y gi In this study, we demonstrated that similarly isolated and grown MSCs from human dental pulp displayed similar morphology, antigenic phenotype, and differentiation poten- tial. These findings do not differ from the previous findings reported by the other studies [12, 15–17]. However, in this study, we demonstrated that similarly isolated and grown MSCs from dental pulps of three different individuals carried considerable variation in their proteomes. A recent study reported similar findings among the proteomes of MSCs isolated from dental tissues (follicle, pulp, and papilla tissue) of a single donor [12]. They reported an average protein spot match ratio of 65% among the isolated MSCs. We also found low spot conservation scores even though we took an extreme precaution during cell culture and 2DE gel electrophoresis to prevent experimental variations. The precautions included the following. (I) Stem cells from three different dental pulps were isolated by using the same stem cell isolation protocol. (II) Cells were grown in defined growth media under the same growth conditions. (III) Before each passage, cells were allowed to grow to 70% confluency and equal numbers of cells were used to start new cultures. (IV) To prevent hand to hand variation, the same researchers performed the isolation, growth, and passaging experiments. (V) Protein isolation and 2DE experiments were also performed by a group of the same researchers. 3. Results Spots that were up- or downregulated more than 2-fold were considered to be subjected to regulation (Supplementary Figure 1). We found that 62.3 ± 7% of the protein spots were conserved. Ratios for up- and downregulated protein spots were given in Figure 5. In overall, analysis of spot scattering plots and conservation scores indicated that NDPSCs cell proteome was more similar to the SHED proteome than the DPSCs proteome. To provide evidence to prove that the results of proteomic analysis were reliable, WB analysis for selected proteins was performed (Figure 7). The results agreed with the proteome analysis indicating that the observed changes in protein spots on 2DE gels were reliable and real.i by all three stem cell types. By using PDQuest Advance gel analysis software, changes in spot intensities among these 183 matching spots were compared. Spots that were up- or downregulated more than 2-fold were considered to be subjected to regulation (Supplementary Figure 1). We found that 62.3 ± 7% of the protein spots were conserved. Ratios for up- and downregulated protein spots were given in Figure 5. In overall, analysis of spot scattering plots and conservation scores indicated that NDPSCs cell proteome was more similar to the SHED proteome than the DPSCs proteome. Most of the identified proteins were shown to play roles in cellular architecture. Also, proteins that are part of protein folding machinery were described. Some of these proteins are inducible chaperons that are expressed under stress condi- tions. Transcription, protein biosynthesis, and degradation related proteins that we detected indicated the presence of active cell growth as predicted from cell cycle experiments. Redox metabolism related scavenger proteins that were detected may be an indicator of active cell growth as well. The presence of apoptosis, transcription, protein biosynthesis/ degradation, and energy metabolism related proteins indi- cated the presence of cellular self-renewal and proliferation. To further characterize conserved protein spots, the spots were cut from a preparative gel with an automated spot cut- ting instrument and subjected to in-gel tryptic digestion fol- lowed by MALDI-TOF/TOF analysis. Collectively, 61 protein spots were identified (Table 2). The identified proteins were subjected to analysis based on their molecular function and their involvement in biological processes (Table 3) (Figure 6). 3. Results (VI) To prevent gel to gel variations in 2DE experiments, precast gels were used. They ran and stained under the same conditions by using Criterion gel running and staining systems under appropriate conditions. Figure 2: Cell cycle analysis of NDPSCs, SHED, and DPSCs. Each sample (10,000 cells) was counted in triplicate by a flow cytometer. The values are expressed as percentage mean ± standard deviation. 0 20 40 60 80 100 120 0 0.5 1 1.5 2 3 2.5 NDPSCs SHED DPSCs Cell (%) %RTA/𝜇g total protein Figure 3: Telomerase activities of NDPSCs, SHED, and DPSCs. The activities were measured by conventional telomeric repeat ampli- fication protocol (TRAP) and relative telomerase activities (RTA) were calculated with respect to the control template equivalent to 0.001 mol/𝜇L DNA. Figure 3: Telomerase activities of NDPSCs, SHED, and DPSCs. The activities were measured by conventional telomeric repeat ampli- fication protocol (TRAP) and relative telomerase activities (RTA) were calculated with respect to the control template equivalent to 0.001 mol/𝜇L DNA. g y pp p Sixty-one of the every-member matching protein spots were identified by MALDI-TOF/TOF MS/MS. The identified proteins were subjected to analysis based on their molecular function and their involvement in biological processes. A total of 10 groups can be formed (Figure 6). Classification of the identified proteins based on molecular function revealed that 11 of the proteins can be grouped as structurally impor- tant ones. Among them, actin related protein complex 2/3 regulates tight junctions and function in establishment of branched actin networks [18]. To the best of our knowledge, the existence of this protein has not been described in MSCs. Calponin, a thin filament-associated protein that is implicated in the regulation and modulation of smooth muscle contraction, was also expressed by all our stem cell types. The expression of calponin in human hair follicle derived stem cells and bone marrow derived stem cells was previously reported [19]. Calponin was occasionally used as a marker to characterize smooth muscle cell derived stem cells [20]. Caldesmon is an actin- and myosin-binding protein and cellular component of movement and muscle contraction [21]. There is a strong interaction between caldesmon and calponin [22]. The identification of caldesmon and calponin 4. Discussion In the presented study, we attempted to compare multilineage potentials and proteomic profiles of dental pulp derived MSCs isolated from dental pulps of three donors. Our aim was to minimize sample complexity by keeping sample number at three. One might argue that using more material from more donors might produce much more scientifically relevant data. However, increasing the number of donors makes analysis more complicated as each donor has different metabolic status (age, smoking, obesity, metabolic diseases, ethnicity, gender, and undetermined/unknown conditions). Indeed, our initial trials of proteome comparisons using sam- ples from multiple donors generated more variable data and as the number of donors increased in groups the proteome profiles started to deviate more. Therefore, although there are examples in the literature which used samples from multiple donors [12–14], we preferred to use single donor for each group.h The MSCs were compared on the basis of their morpho- logical properties, proliferation rates, expression of common 7 Stem Cells International 7 Table 2: Protein spots that were identified by MALDI-TOF/TOF (MS/MS) analysis. 4. Discussion SSP Swiss-Prot accession Best protein description Sequence coverage/matches MASCOT probability/expect (p) Theoretical MW (Da) Experimental p𝐼 3001 P 09382 Galectin-1 40%/9 fragment 236/5.10e −20 14706 5.34 3005 P 06703 Protein S100-A6 24%/7 fragment 124/8.10e −09 10173 5.33 4003 P 62988 Ubiquitin 31%/5 fragment 151/1.60e −11 8560 6.56 5002 O 75368 SH3 domain-binding glutamic acid-rich-like protein 22%/5 fragment 181/1.60e −14 12766 5.22 5001 Q 14019 Coactosin-like protein 28%/7 fragment 93/1.00e −05 15935 5.54 4001 O 75347 Tubulin-specific chaperone A 48%/10 fragment 295/6.40e −26 12847 5.25 3003 Q 9NQ39 Putative 40S ribosomal protein S10-like 20%/7 fragment 22/1.10e + 02 20108 10.13 3002 P 63241 Eukaryotic translation initiation factor 5A-1 15%/5 fragment 177/4.00e −14 16821 5.08 3004 O 95881 Thioredoxin domain-containing protein 12 12%/4 fragment 47/0.41 19194 5.24 3101 P 52815 39S ribosomal protein L12, mitochondrial 15%/4 fragment 164/8.10e −13 21335 9.04 4101 O 75947 ATP synthase subunit d, mitochondrial 32%/8 fragment 322/1.30e −28 18480 5.21 4102 P 09936 Ubiquitin carboxyl-terminal hydrolase isozyme L1 22%/8 fragment 212/1.30e −17 24808 5.33 4002 P 07858 Cathepsin B 5%/1 fragment 107/4.00e −07 37797 5.88 3201 P 43487 Ran-specific GTPase-activating protein 28%/10 fragment 167/4.00e −13 23296 5.19 6102 P 32119 Peroxiredoxin-2 34%/15 fragment 536/5.10e −50 21878 5.66 6104 P 30085 UMP-CMP kinase 13%/5 fragment 140/2.00e −10 22208 5.44 6202∗ P 04792 Heat shock protein beta-1 33%/12 fragment 321/1.60e −28 22768 5.98 7101 P 00441 Superoxide dismutase [Cu-Zn] 9%/2 fragment 101/1.60e −06 15926 5.7 6001 O 15511 Actin related protein 2/3 complex subunit 5 12%/3 fragment 78/0.00035 16310 5.47 7001 Q 12912 Lymphoid-restricted membrane protein 1%/2 fragment 25/67 62069 5.62 7204 P 07339 Cathepsin D 20%/17 fragment 503/1.00e −46 44524 6.1 7205 Q 99436 Proteasome subunit beta type 7 14%/9 fragment 150/2.00e −11 29946 7.57 7102 P 09211 Glutathione S-transferase P 16%/5 fragment 215/6.4e −018 23341 5.43 4201A∗ P 07858 Cathepsin B 28%/14 fragment 574/8.10e −54 37797 5.88 4201B O 00299 Chloride intracellular channel protein 1 10%/6 fragment 130/2.00e −09 26906 5.09 8101 P 04792 Heat shock protein beta-1 45%/22 fragment 707/4.00e −67 22768 5.98 8201 P 30048 Thioredoxin-dependent peroxide reductase, mitochondrial 23%/10 fragment 346/5.10e −31 27675 7.67 9102 O 95571 Protein ETHE1, mitochondrial 35%/12 fragment 439/2.60e −40 27855 6.35 9101 Q 99497 Protein DJ-1 33%/10 fragment 239/2.60e −20 19878 6.33 8202 P 30084 Enoyl-CoA hydratase, mitochondrial 25%/10 fragment 183/1.00e −14 31367 8.34 9202∗ P 60174 Triosephosphate isomerase 64%/24 fragment 872/1.30e −83 26653 6.45 9203∗ P 07339 Cathepsin D 16%/14 fragment 233/1.00e −19 44524 6.1 5.34 5.33 6.56 5.22 5.54 5.25 10.13 5.08 5.24 9.04 5.21 5.33 5.88 5.19 5.66 5.44 5.98 5.7 5.47 5.62 6.1 7.57 5.43 5.88 5.09 5.98 7.67 6.35 6.33 8.34 6.45 6.1 8 Stem Cells International Stem Cells Intern Table 2: Continued. 4. Discussion SSP Swiss-Prot accession Best protein description Sequence coverage/matches MASCOT probability/expect (p) Theoretical MW (Da) Experimental p𝐼 8002 P 31949 Protein S100-A11 56%/10 fragment 312/1.30e −27 11733 6.56 9003 P 49773 Histidine triad nucleotide-binding protein 1 16%/4 fragment 139/2.60e −10 13793 6.43 9002 Q 969H8 UPF0556 protein C19orf10 30%/12 fragment 437/4.00e −40 18783 6.2 9201 P 60174 Triosephosphate isomerase 71%/28 fragment 994/8.10e −96 26653 6.45 9215 P 18669 Phosphoglycerate mutase 1 41%/13 fragment 627/4.00e −59 28786 6.67 9213 O 00560 Syntenin-1 24%/13 fragment 257/4.00e −22 32424 7.05 9212 Q 99439 Calponin-2 11%/9 fragment 79/0.00027 33675 6.95 9211 O 14979 Heterogeneous nuclear ribonucleoprotein D-like 15%/13 fragment 361/1.60e −32 46409 9.59 9209 P 04083 Annexin A1 15%/9 fragment 233/1.00e −19 38690 6.57 9221 Q 99729 Heterogeneous nuclear ribonucleoprotein A/B 18%/8 fragment 145/6.40e −11 36202 8.22 9308 Q 14103 Heterogeneous nuclear ribonucleoprotein D0 20%/11 fragment 301/1.60e −26 38410 7.62 9305 P 60709 Actin, cytoplasmic 1 24%/14 fragment 360/2.00e −32 41710 5.29 9402 P 06733 Alpha-enolase 24%/19 fragment 494/8.10e −46 47139 7.01 9401 P 13929 Beta-enolase 14%/8 fragment 165/6.40e −13 46902 7.59 9401 P 09104 Gamma-enolase 11%/6 fragment 159/2.60e −12 47239 4.91 9903 Q 8WXF1 Paraspeckle component 1 17%/14 fragment 236/5.10e −20 58706 6.26 9902 P 04264 Keratin, type II cytoskeletal 1 9%/9 fragment 416/5.10e −38 65999 8.15 9901 Q 05682 Caldesmon 24%/23 fragment 512/1.30e −47 93194 5.63 8901 Q 15942 Zyxin 21%/19 fragment 287/4.00e −25 61238 6.22 8502 P 17661 Desmin 14%/10 fragment 137/4.00e −10 53503 5.21 7401 P 31943 Heterogeneous nuclear ribonucleoprotein H 49%/29 fragment 966/5.10e −93 49198 5.89 9603 P 08670 Vimentin 75%/57 fragment 1290/2.00e −125 53619 5.06 8302 O 14773 Tripeptidyl-peptidase 1 4%/4 fragment 189/2.60e −15 61210 6.01 8301 P 40121 Macrophage-capping protein 23%/18 fragment 465/6.40e −43 38494 5.88 5202 P 52907 F-actin-capping protein subunit alpha-1 44%/11 fragment 210/2.00e −17 32902 5.45 6903 P 11142 Heat shock cognate 71 kDa protein 24%/19 fragment 472/1.3e −043 70854 5.37 5902 P 10809 60 kDa heat shock protein, mitochondrial 27%/18 fragment 475/6.4e −044 61016 5.7 5604 Q 16352 Alpha-internexin 5%/3 fragment 72/0.0012 55357 5.34 4902∗ Q 96AY3 Peptidyl-prolyl cis-trans isomerase FKBP10 5%/5 fragment 73/0.00093 64204 5.36 4901 P 11021 78 kDa glucose-regulated protein 22%/16 fragment 426/5.1e −039 72288 5.07 7003 P 15531 Nucleoside diphosphate kinase A 37%/11 fragment 348/3.20e −31 17138 5.83 9005 Q 01469 Fatty acid-binding protein 27%/5 fragment 246/5.10e −21 15155 6.6 9204 P 25786 Proteasome subunit alpha type 1 25%/9 fragment 166/5.1e −013 29537 6.15 9206∗ P 60174 Triosephosphate isomerase 30%/8 fragment 116/5.1e −008 26653 6.45 9010 P 62937 Peptidyl-prolyl cis-trans isomerase A 17%/5 fragment 150/2.00e −11 18001 7.68 ∗Some of the identified proteins appeared as more than one spot on the gels as indicated by the presence of more than one SSP number. 4. Discussion Table 2: Continued. 4. Discussion SSP Swiss-Prot accession Best protein description Sequence coverage/matches MASCOT probability/expect (p) Theoretical MW (Da) Experimental p𝐼 8002 P 31949 Protein S100-A11 56%/10 fragment 312/1.30e −27 11733 6.56 9003 P 49773 Histidine triad nucleotide-binding protein 1 16%/4 fragment 139/2.60e −10 13793 6.43 9002 Q 969H8 UPF0556 protein C19orf10 30%/12 fragment 437/4.00e −40 18783 6.2 9201 P 60174 Triosephosphate isomerase 71%/28 fragment 994/8.10e −96 26653 6.45 9215 P 18669 Phosphoglycerate mutase 1 41%/13 fragment 627/4.00e −59 28786 6.67 9213 O 00560 Syntenin-1 24%/13 fragment 257/4.00e −22 32424 7.05 9212 Q 99439 Calponin-2 11%/9 fragment 79/0.00027 33675 6.95 9211 O 14979 Heterogeneous nuclear ribonucleoprotein D-like 15%/13 fragment 361/1.60e −32 46409 9.59 9209 P 04083 Annexin A1 15%/9 fragment 233/1.00e −19 38690 6.57 9221 Q 99729 Heterogeneous nuclear ribonucleoprotein A/B 18%/8 fragment 145/6.40e −11 36202 8.22 9308 Q 14103 Heterogeneous nuclear ribonucleoprotein D0 20%/11 fragment 301/1.60e −26 38410 7.62 9305 P 60709 Actin, cytoplasmic 1 24%/14 fragment 360/2.00e −32 41710 5.29 9402 P 06733 Alpha-enolase 24%/19 fragment 494/8.10e −46 47139 7.01 9401 P 13929 Beta-enolase 14%/8 fragment 165/6.40e −13 46902 7.59 9401 P 09104 Gamma-enolase 11%/6 fragment 159/2.60e −12 47239 4.91 9903 Q 8WXF1 Paraspeckle component 1 17%/14 fragment 236/5.10e −20 58706 6.26 9902 P 04264 Keratin, type II cytoskeletal 1 9%/9 fragment 416/5.10e −38 65999 8.15 9901 Q 05682 Caldesmon 24%/23 fragment 512/1.30e −47 93194 5.63 8901 Q 15942 Zyxin 21%/19 fragment 287/4.00e −25 61238 6.22 8502 P 17661 Desmin 14%/10 fragment 137/4.00e −10 53503 5.21 7401 P 31943 Heterogeneous nuclear ribonucleoprotein H 49%/29 fragment 966/5.10e −93 49198 5.89 9603 P 08670 Vimentin 75%/57 fragment 1290/2.00e −125 53619 5.06 8302 O 14773 Tripeptidyl-peptidase 1 4%/4 fragment 189/2.60e −15 61210 6.01 8301 P 40121 Macrophage-capping protein 23%/18 fragment 465/6.40e −43 38494 5.88 5202 P 52907 F-actin-capping protein subunit alpha-1 44%/11 fragment 210/2.00e −17 32902 5.45 6903 P 11142 Heat shock cognate 71 kDa protein 24%/19 fragment 472/1.3e −043 70854 5.37 5902 P 10809 60 kDa heat shock protein, mitochondrial 27%/18 fragment 475/6.4e −044 61016 5.7 5604 Q 16352 Alpha-internexin 5%/3 fragment 72/0.0012 55357 5.34 4902∗ Q 96AY3 Peptidyl-prolyl cis-trans isomerase FKBP10 5%/5 fragment 73/0.00093 64204 5.36 4901 P 11021 78 kDa glucose-regulated protein 22%/16 fragment 426/5.1e −039 72288 5.07 7003 P 15531 Nucleoside diphosphate kinase A 37%/11 fragment 348/3.20e −31 17138 5.83 9005 Q 01469 Fatty acid-binding protein 27%/5 fragment 246/5.10e −21 15155 6.6 9204 P 25786 Proteasome subunit alpha type 1 25%/9 fragment 166/5.1e −013 29537 6.15 9206∗ P 60174 Triosephosphate isomerase 30%/8 fragment 116/5.1e −008 26653 6.45 9010 P 62937 Peptidyl-prolyl cis-trans isomerase A 17%/5 fragment 150/2.00e −11 18001 7.68 ∗Some of the identified proteins appeared as more than one spot on the gels as indicated by the presence of more than one SSP number. 4. Discussion 6.56 6.43 6.2 6.45 6.67 7.05 6.95 9.59 6.57 8.22 7.62 5.29 7.01 7.59 4.91 6.26 8.15 5.63 6.22 5.21 5.89 5.06 6.01 5.88 5.45 5.37 5.7 5.34 5.36 5.07 5.83 6.6 6.15 6.45 7.68 Table 2: Continued. Stem Cells International 9 Table 3: Classification of the proteins that were identified by MALDI-TOF/TOF analysis. The classifications were made based on PANTHER analysis and Swiss-Prot annotations. Protein name Function Biological process involved Apoptosis related Galectin Apoptosis, cell proliferation, and differentiation Apoptosis Cathepsin B Thiol protease which is believed to participate in intracellular degradation and turnover of proteins. Regulation of apoptotic process and proteolysis Cathepsin D Acid protease active in intracellular protein breakdown Regulation of apoptotic process and proteolysis Histidine triad nucleotide-binding protein 1 Hydrolyzes purine nucleotide phosphoramidates Apoptosis Signal transduction SA100A6 May function as calcium sensor and modulator, contributing to cellular calcium signaling Signal transduction Protein S100-A11 Signal transduction Annexin A1 Calcium/phospholipid-binding protein which promotes membrane fusion and is involved in exocytosis Cell surface receptor signaling pathway Protein biosynthesis and degradation Ubiquitin Behaves as a molecular tag for degradation of proteins Ubiquitination Putative 40S ribosomal protein S10-like Ribonucleoprotein Protein biosynthesis Eukaryotic translation initiation factor 5A-1 Ribosome binding Translation elongation factor activity Protein biosynthesis Ubiquitin carboxyl-terminal hydrolase Processing of ubiquitin precursors and of ubiquitinated proteins Ubl conjugation pathway 39S ribosomal protein L12, mitochondrial Ribonucleoprotein Mitochondrial protein biosynthesis Proteasome subunit alpha type 1 The proteasome is a multicatalytic proteinase complex which is characterized by its ability to cleave peptides with Arg, Phe, Tyr, Leu, and Glu adjacent to the leaving group at neutral or slightly basic pH Immunity Proteasome subunit beta type 7 Cleavage of peptide bonds with very broad specificity Protein degradation Transcription related Heterogeneous nuclear ribonucleoprotein A/B Binds single-stranded RNA. Has a high affinity for G-rich and U-rich regions of hnRNA Transcription, transcription regulation Heterogeneous nuclear ribonucleoprotein D0 Binds with high affinity to RNA molecules that contain AU-rich elements (AREs) found within the 3󸀠-UTR of many protooncogenes and cytokine mRNAs Transcription, transcription regulation 10 Stem Cells International Table 3: Continued. Protein name Function Biological process involved Heterogeneous nuclear ribonucleoprotein H This protein is a component of the heterogeneous nuclear ribonucleoprotein (hnRNP) complexes which provide the substrate for the processing events that pre-mRNAs undergo before becoming functional, translatable mRNAs in the cytoplasm. Mediates pre-mRNA alternative splicing regulation Mediates pre-mRNA alternative splicing regulation Heterogeneous nuclear ribonucleoprotein D-like Acts as a transcriptional regulator. 4. Discussion Promotes transcription repression. Transcription, transcription regulation Paraspeckle component 1 Regulates, cooperatively with NONO and SFPQ, androgen receptor-mediated gene transcription activity in Sertoli cell line Transcription, transcription regulation Nucleotide metabolism UMP-CMP kinase Catalyzes the phosphorylation of pyrimidine nucleoside monophosphates at the expense of ATP Nucleotide metabolism Nucleoside diphosphate kinase A Major role in the synthesis of nucleoside triphosphates other than ATP Nucleotide metabolism Redox metabolism related Thioredoxin domain-containing protein 12 Possesses significant protein thiol-disulfide oxidase activity Oxidoreductase Peroxiredoxin-2 Involved in redox regulation of the cell. Reduces peroxides with reducing equivalents provided through the thioredoxin system Hydrogen peroxide catabolic process removal of superoxide radicals Superoxide dismutase [Cu-Zn] Destroys radicals which are normally produced within the cells and which are toxic to biological systems. Antioxidant, oxidoreductase Thioredoxin-dependent peroxide reductase Peroxidase activity, peroxiredoxin activity Involved in redox regulation of the cell Glutathione S-transferase P Conjugation of reduced glutathione to a wide number of exogenous and endogenous substrates Glutathione derivative biosynthetic process Persulfide dioxygenase ETHE1, mitochondrial Plays ab essential role in hydrogen sulfide catabolism in the mitochondrial matrix Hydrogen sulfide metabolic process Protein folding/stress Tubulin-specific chaperone A Involves the early step of the tubulin folding pathway. Chaperone binding UPF0556 protein C19orf10 Involved in unfolded protein response, cellular protein metabolic process Protein DJ-1 Protects cells against oxidative stress and cell death. Oxidative stress Heat shock protein beta-1 Involved in stress resistance and actin organization. Stress response Heat shock cognate 71 kDa protein Acts as a repressor of transcriptional activation Stress response 60 kDa heat shock protein, mitochondrial Implicated in mitochondrial protein import and macromolecular assembly de novo’ protein folding, stress response 11 11 Stem Cells International Table 3: Continued. Protein name Function Biological process involved 78 kDa glucose-regulated protein Plays a role in facilitating the assembly of multimeric protein complexes inside the endoplasmic reticulum ER overload response Peptidyl-prolyl cis-trans isomerase A It catalyzes the cis-trans isomerization of proline imidic peptide bonds in oligopeptides. PPIases accelerate the folding of proteins Tripeptidyl-peptidase 1 Lysosomal serine protease with tripeptidyl-peptidase I activity Involved in unfolded protein response Energy metabolism ATP synthase subunit D ATP synthesis Energy metabolism Phosphoglycerate mutase 1 Interconversion of 3- and 2-phosphoglycerate Glycolysis, gluconeogenesis Alpha-enolase Role in glycolysis, plays a part in various processes such as growth control, hypoxia tolerance, and allergic responses. Glycolysis, plasminogen activation Transcription, transcription regulation Beta-enolase Striated muscle development and regeneration. 4. Discussion Glycolysis Gamma-enolase Has neurotrophic and neuroprotective properties on a broad spectrum of central nervous system (CNS) neurons Glycolysis Triosephosphate isomerase D-Glyceraldehyde 3-phosphate = glycerone phosphate. Energy metabolism Structural proteins Coactosin-like protein Acts as a chaperone for ALOX5 (5LO), influencing both its stability and activity in leukotrienes synthesis. Binds to F-actin in a calcium-independent manner Actin related protein 2/3 complex subunit 5 Component of the Arp2/3 complex which is involved in regulation of actin polymerization Structural constituent of cytoskeleton Syntenin-1 An adapter protein, in adherens junctions may function to couple syndecans to cytoskeletal proteins or signaling components Actin cytoskeleton organization Calponin-2 Thin filament-associated protein that is implicated in the regulation and modulation of smooth muscle contra Actin binding Actin, cytoplasmic 1 Cell motility Cell structure formation Keratin, type II cytoskeletal 1 Regulate the activity of kinases such as PKC and SRC via binding to integrin beta-1 (ITB1) and the receptor of activated protein kinase C Complement activation, lectin pathway, fibrinolysis, and regulation of angiogenesis Caldesmon Actin- and myosin-binding protein Cellular component movement, muscle contraction Zyxin Adhesion plaque protein. Binds alpha-actinin and the CRP prot Cell adhesion 12 Stem Cells International Table 3: Continued. Protein name Function Biological process involved Desmin Class-III intermediate filaments found in muscle cells Cytoskeleton organization, muscle filament sliding Vimentin Vimentins are class-III intermediate filaments found in various nonepithelial cells, especially mesenchymal c Intermediate filament organization Macrophage-capping protein Calcium-sensitive protein which reversibly blocks the barbed ends of actin filaments but does not sever preformed actin filaments. May play an important role in macrophage function. Barbed-end actin filament capping, cell projection assembly F-actin-capping protein subunit alpha-1 F-actin-capping proteins bind in a Ca2+-independent manner to the fast growing ends of actin filaments Actin cytoskeleton organization, actin filament capping, cellular component movement Other proteins SH3 domain-binding glutamic acid-rich-like protein SH3/SH2 adaptor activity SH3-binding Ran-specific GTPase-activating protein Inhibits GTP exchange on Ran Positive regulation of GTPase activity Lymphoid-restricted membrane protein Vesicle fusion Vesicle targeting Chloride intracellular channel protein 1 Can be inserted into membranes and form chloride ion channels. Positive regulation of osteoblast differentiation Enoyl-CoA hydratase Straight-chain enoyl-CoA thioesters from C4 up to at least C16 Fatty acid metabolism Alpha-internexin Class-IV neuronal intermediate filament Differentiation, neurogenesis Fatty acid-binding protein High specificity for fatty acids. Highest affinity for C18 chain length. 4. Discussion Transport 13 13 Stem Cells International Adipogenic Osteogenic Chondrogenic (a1) (a2) (a3) (b1) (c1) (b2) (c2) (b3) (c3) SHED hNDP-MSCs hDP-MSCs Figure 4: In vitro differentiation potential of NDPSCs, SHED, and DPSCs. ((a1), (b1), and (c1)). Adipogenic differentiation is visually marked by accumulation of neutral lipid vacuoles in cultures (Oil Red staining). Actin ((a1) and (b1)) and vimentin (c1) expression were shown in green and nuclei in blue with DAPI. ((a2), (b2), and (c2)) Osteogenic differentiation was indicated by the formation of calcified nodule with Alizarin Red S staining. ((a3), (b3), and (c3)) The analyzed sections were positive for Safranin O staining after chondrogenic differentiation. Osteogenic (a2) Chondrogenic (a3) Adipogenic (a1) hNDP-MSCs hNDP-MSCs (a1) (a3) (b3) (b2) (b1) SHED (b1) (b3) (b2) (c3) (c1) hDP-MSCs (c2) hDP-MSCs (c1) (c2) (c3) Figure 4: In vitro differentiation potential of NDPSCs, SHED, and DPSCs. ((a1), (b1), and (c1)). Adipogenic differentiation is visually marked by accumulation of neutral lipid vacuoles in cultures (Oil Red staining). Actin ((a1) and (b1)) and vimentin (c1) expression were shown in green and nuclei in blue with DAPI. ((a2), (b2), and (c2)) Osteogenic differentiation was indicated by the formation of calcified nodule with Alizarin Red S staining. ((a3), (b3), and (c3)) The analyzed sections were positive for Safranin O staining after chondrogenic differentiation. in our stem cell populations may indicate that certain sub- populations of stem cells in our cultures might be committed to vascular smooth muscle lineage. However, the expression of actin interactin proteins may be due to well known motile nature of MSCs. In addition, expression of caldesmon and annexin A1 is proposed to be a marker for slow growth rate [23]. Because our stem cells were in defined medium their growth rates are expected to be slow and thus they might have expressed these two proteins at a respectable level. Zyxin binds alpha-actinin and the CRP protein and helps cell adhesion [24] and the cells used in this study expressed Zyxin. Zyxin was also determined to be expressed in hMSCs and proposed to be involved in cell to matrix contacts [25]. Previously, coactosin-like protein is shown to be highly expressed in amniotic-fluid derived MSCs [26]. Coactosin- like protein is involved in cytoskeletal organization and movement by binding actin. 4. Discussion Both Coactosin-like protein and actin expression were identified in this study.ii glutathione S-transferase-P is known to function in detoxi- fication of cells from xenobiotics and decrease susceptibility to cancer. In a previous study, glutathione S-transferase-P is used to distinguish multipotent MSCs isolated from bone marrow [27]. Observation of expression of glutathione S- transferase-P in our stem cells may indicate that our cells did have multipotent MSC properties. Increased generation of reactive oxygen species (ROS) may be associated with differentiation process. On that line of thought, Clair et al. (1994) demonstrated the expression of MnSOD on cellular differentiation of fibroblasts [28]. They found that MnSOD greatly enhanced fibroblast differenti- ation into myoblasts. In addition, the protective effect of SOD was demonstrated in bone marrow derived stem cells and in hematopoietic stem cells [29, 30]. We also detected SOD expression in our stem cells. In addition to SOD, three other proteins that are involved in clearance of ROS were detected in this study. These were peroxiredoxin-2, thioredoxin domain-containing protein 12, and thioredoxin- dependent peroxide reductase. Peroxiredoxin reduces per- oxides with reducing equivalents provided through the We have identified five proteins that have roles in redox metabolism and detoxification of the cells. Among them, 14 14 Stem Cells International 100 kDa 50kDa 30kDa 10kDa pH 8 5 pH 8 5 pH 8 5 NDPSCs SHED DPSCs (a) 120 kDa 100 kDa 50kDa 30kDa 10kDa pH 8 5 Master gel (b) Upregulated (%) Downregulated (%) Conserved (%) Upregulated (%) Downregulated (%) Conserved (%) Upregulated (%) Downregulated (%) Conserved (%) NDPSCs SHED DPSCs 72% 18% 10% 57% 24% 19% 58% 18% 24% (c) gure 5: Comparative proteome analysis of NDPSCs, SHED, and DPSCs. (a) Undifferentiated MSCs were subjected to protein isolation a ded onto pH 5 to pH 8 IPG strips for the first dimension and precast SDS-PAGE gels for the second dimension separation and stain h Colloidal Coomassie Blue for 24 hr after 24 hours of fixation. The gels are representative of three gels from each MSC type. (b) Mas image was created by PDQuest Advance software to determine the number of spots that matched every member (183 spots). The numb resent SSP numbers assigned to each spot by the software. These spots were cut from the gels and were subjected to MALDI-TOF/TO alysis. 4. Discussion Protein identification was performed by peptide mass finger printing by MASCOT (c) Pie charts to illustrate proteome conservati os among hNDP, SHED, and hDP-MSCs. Changes in spot intensities among 183 matching spots were compared. Spots that were up- wnregulated more than 2-fold were considered to be subjected to regulation. pH 8 5 SHED (a) pH 8 5 DPSCs 100 kDa 50kDa 30kDa 10kDa pH 8 5 NDPSCs 8 (a) pH Master gel 120 kDa 100 kDa 50kDa 30kDa 10kDa pH 8 5 Master gel (b) Master gel pH 8 5 (b) DPSCs 58% 18% 24% ( ) SHED DPSCs 57% 24% 19% 58% 18% 24% SHED 57% 24% 19% 24% NDPSCs 72% 18% 10% Upregulated (%) Downregulated (%) Conserved (%) (c) Upregulated (%) Downregulated (%) Conserved (%) Upregulated (%) Downregulated (%) Conserved (%) (c) (c) Figure 5: Comparative proteome analysis of NDPSCs, SHED, and DPSCs. (a) Undifferentiated MSCs were subjected to protein isolation and loaded onto pH 5 to pH 8 IPG strips for the first dimension and precast SDS-PAGE gels for the second dimension separation and stained with Colloidal Coomassie Blue for 24 hr after 24 hours of fixation. The gels are representative of three gels from each MSC type. (b) Master gel image was created by PDQuest Advance software to determine the number of spots that matched every member (183 spots). The numbers represent SSP numbers assigned to each spot by the software. These spots were cut from the gels and were subjected to MALDI-TOF/TOF analysis. Protein identification was performed by peptide mass finger printing by MASCOT (c) Pie charts to illustrate proteome conservation ratios among hNDP, SHED, and hDP-MSCs. Changes in spot intensities among 183 matching spots were compared. Spots that were up- or downregulated more than 2-fold were considered to be subjected to regulation. Majority of the proteins that we identified have roles in protein folding machinery. For instance, peptidyl-prolyl cis- trans isomerase (PIN1) accelerates the folding of proteins by catalyzing cis-trans isomerization of proline imidic bonds in oligopeptides [33]. The role of PIN1 in odontogenic and adipogenic differentiation in human dental pulp derived stem cells was studied [34]. It was shown that PIN1 acted as an thioredoxin system. The expression of proteins displaying antioxidant activities by MSCs was previously recognized [31]. In fact MSCs were proposed to be used as a tool to clear out ROS from niches that are under oxidative stress [32]. 4. Discussion Therefore, the presence of PIN1 in our stem cells may be related to stemness properties of human dental pulp derived stem cells.h We identified several proteins that play roles in tran- scription and nucleotide metabolism. Among them, PSPC1 is interesting since the functions of paraspeckles in eukaryotic cells are still not known. The presence of a paraspeckle protein (PSP1 alpha) was reported in human embryonic stem cells although the existence of paraspeckles in human embryonic stem cells was reported to occur only upon differentiation [43]. Our cells expressed a paraspeckle component, PSP1C, and this suggested that paraspeckle components are not only expressed by human embryonic stem cells but also expressed by human mesenchymal stem cells and their expression may not require a differentiation process.i f Twenty percent of the identified proteins displayed bind- ing activities and five of these proteins are involved in calcium or calcium binding metabolism (protein S100-A6, annexin A1, protein S100-A11, peptidyl-prolyl cis-trans isomerase, and gelsolin). The importance of calcium metabolism has long been recognized in stem cells. In fact, calcium hydroxide is shown to increase recruitment, migration, proliferation, and mineralization of the dental pulp derived MSCs [44]. In a recent study, Chen et al. (2013) demonstrated the effect of S100A4, another calcium binding protein, on proliferation, survival, and differentiation of human osteosarcoma cells [45]. In a more recent study, SA100A6 is proposed to be a novel marker for neural stem cells and may be important for generation of astrocytes in the adult hippocampus [46]. Annexin, which interacts with SA100A6, is required for midbody formation and completion of the terminal phase of cytokinesis and was identified in our study as a part of the calcium binding metabolism.h There were four heat shock proteins (60 kDa heat shock protein, heat shock cognate 71 kDa protein, heat shock protein Beta-1, and 78 kDa glucose-regulated protein) that we identified and they were expressed by all three stem cells. Inducible and constitutive HSP proteins confer synergetic resistance against metabolic changes [39]. Protein DJ-1 is another chaperone protein that protects cells against oxida- tive stress and cell death. Its role in Parkinson’s disease is well known but its significance in stem cell research has not been studied [40]. UPF0556 protein C19orf10 (interleukin-25) is another protein that we detected in our gels. This protein is a secreted protein and is proposed to activate signaling pathways that is involved in unfolded protein response [41]. 4. Discussion We believe that all four enzymes that we report here protect cells from antioxidants and may have proliferative effects by eliminating peroxides. 15 Stem Cells International Apoptosis related 7% Signal transduction 5% Protein biosynthesis and degradation 11% Transcription related 8% Nucleotide metabolism 3% Redox metabolism related 10% Protein folding/stress 15% Energy metabolism 10% Structural proteins 20% Other proteins 11% Figure 6: Classification of identified proteins based on their molecular function and their involvement in biological processes. Pie chart representing the distribution of the 61 identified proteins based on their molecular function and biological processes. Assignments were made on the basis of information from PANTHER analysis (http://www.pantherdb.org/) as well as NCBI (http://www.ncbi.nlm.nih.gov/pubmed) and Swiss-Prot/TrEMBL annotations (http://www.expasy.org/). Apoptosis related 7% Other proteins 11% Structural proteins 20% Nucleotide metabolism 3% Redox metabolism related 10% Figure 6: Classification of identified proteins based on their molecular function and their involvement in biological processes. Pie chart representing the distribution of the 61 identified proteins based on their molecular function and biological processes. Assignments were made on the basis of information from PANTHER analysis (http://www.pantherdb.org/) as well as NCBI (http://www.ncbi.nlm.nih.gov/pubmed) and Swiss-Prot/TrEMBL annotations (http://www.expasy.org/). of chaperons or proteins that play roles in protein folding machinery may indicate the efforts of preservation of stem cell integrity.i important modulator of odontogenic and adipogenic differ- entiation of DPSCs. PIN1 was also shown to promote sur- vival, enhance repair, improve differentiation, and antagonize senescence [35]. PIN1 mRNA and protein levels were upreg- ulated in a time-dependent manner during adipogenic differ- entiation in human dental pulp derived stem cells [34]. PIN1 may act as an important modulator in human dental pulp derived stem cells and may have significant implications in regenerative medicine. In a study by Nakamura et al. (2012), depression of PIN1 has shown to suppress neuronal differen- tiation while its overexpression enhanced it [36]. Our group has demonstrated that human dental pulp stem cells display better neural and epithelial stem cell properties [37]. A distinct role for PIN1 has been assigned in induction and maintenance of pluripotency [38]. PIN1 was described to be an indispensable factor for the self-renewal and maintenance of pluripotent stem cells because it can regulate Oct4 and other substrates via activation of phosphorylation cascades. 4. Discussion The last protein that we detected and plays a role in protein folding was tubulin-specific chaperone A which is involved in early steps of the tubulin folding pathway [42]. The presence The presence of four apoptosis related proteins indicated that self-renewal and proliferation of stem cell populations is controlled by induction of apoptosis. Therefore, there is always a balance between the number of actively growing 16 Stem Cells International Relative band intensity levels Relative band intensity levels Relative band intensity levels Relative band intensity levels Relative band intensity levels SHED DPSCs NDPSCs 3000 6000 0 SHED DPSCs NDPSCs 3000 6000 0 SHED DPSCs NDPSCs 3000 6000 0 SHED DPSCs NDPSCs 3000 6000 0 SHED DPSCs NDPSCs 3000 6000 0 SHED DPSCs NDPSCs SHED DPSCs NDPSCs SHED DPSCs NDPSCs SHED DPSCs NDPSCs SSP 3001 SSP 4003 SSP 4102 SSP 9101 67693 PPM SSP 3001 6323 PPM 928 PPM INT ∗area INT ∗area INT ∗area INT ∗area SSP 4003 SSP 4102 7316 PPM SSP 9101 SHED DPSCs NDPSCs Galectin-1 DJ-1 UCHL-1 HNRNPH1 GAPDH SHED DPSCs SHED DPSCs SHED DPSCs SHED DPSCs NDPSCs NDPSCs NDPSCs NDPSCs (a) (b) (c) (d) igure 7: WB validation of the selected proteins identified by MALDI-TOF/TOF analysis. Western blot analysis was used for validatio f proteomic results. Conventional 12% SDS gels were run with whole cell extracts. (a) Three lanes for SHED, DPSCs, and NDP-MSC a hown for each antibody. Galectin-1, DJ-1, UCHL-1, and HNRNPH1 antibodies were used for validation of expressions of the identifie roteins. GAPDH was used for the normalization of each protein sample. (b) Relative protein intensities of bands were measured by Quanti One 1D analysis software (Bio-Rad). Each WB was repeated for three times. (c) Representative images of protein spots selected for W nalysis (protein identities corresponding to SSP numbers can be found in Table 2). (d) Comparative intensity graphs of selected prote pots generated by PDQuest Advance software (Bio-Rad). 5. Conclusions In this study we used basic cell biology and proteomics techniques to uncover the heterogeneity of proteins and their characteristics among three stem cell types. The cells dis- played similar features related to morphology, proliferation rates, expression of various cell surface markers, and differ- entiation potentials. Furthermore, using 2DE with MALDI- TOF/TOF MS/MS approach, we have generated a proteomic profile. We mainly focused on the 61 proteins that were predominantly expressed by all three stem cell types. We are aware that other differentially expressed proteins in each cell type may also be important in understanding specific proteins representing three MSCs from different age groups and those proteins may be a subject of another study. Some of the identified proteins in this study may hold importance in understanding specific properties of human dental pulp derived mesenchymal stem cells. Clarification of the similar- ities and differences among the proteomes of human dental pulp derived mesenchymal stem cells will help to have a better understanding of these cells and this study is a step taken towards that direction. Authors’ Contributions stem cells and apoptotic stem cells. In overall, the dynamic nature of stem cell populations is evident by the presence of apoptosis related proteins as well as proteins that are involved in protein biosynthesis/degradation, energy metabolism, and transcription. The overall design of the study, including planning, and var- ious aspects of interpretation of results were done by Murat Kasap, Gurler Akpinar, and Erdal Karaoz. Murat Kasap and Gurler Akpinar wrote and revised the paper and designed and carried out the proteomics experiments including MALDI- TOF/TOFMS/MS analysis. Ayca Aksoy and Erdal Karaoz did all the cell culture experiments. Gokhan Duruksu performed telomerase assay. Gulcin Gacar performed flow cytometer/immunophenotyping and cell cycle analysis. 4. Discussion Relative band intensity levels SHED DPSCs NDPSCs 3000 6000 0 SHED DPSCs NDPSCs SSP 3001 67693 PPM SSP 3001 INT ∗area SHED DPSCs NDPSCs Galectin-1 Relative band intensity levels SHED DPSCs NDPSCs 3000 6000 0 SHED DPSCs NDPSCs SSP 3001 67693 PPM SSP 3001 INT ∗a s NDPSCs Relative band intensity levels 3 6 Relative band intensity levels SHED DPSCs NDPSCs SSP 9101 INT ∗area 7316 PPM SSP 9101 (c) (d) (c) (d) Relative band intensity levels N SHED DPSCs NDPSCs 3000 6000 0 (b) (a) (b) Figure 7: WB validation of the selected proteins identified by MALDI-TOF/TOF analysis. Western blot analysis was used for validation of proteomic results. Conventional 12% SDS gels were run with whole cell extracts. (a) Three lanes for SHED, DPSCs, and NDP-MSC are shown for each antibody. Galectin-1, DJ-1, UCHL-1, and HNRNPH1 antibodies were used for validation of expressions of the identified proteins. GAPDH was used for the normalization of each protein sample. (b) Relative protein intensities of bands were measured by Quantity One 1D analysis software (Bio-Rad). Each WB was repeated for three times. (c) Representative images of protein spots selected for WB analysis (protein identities corresponding to SSP numbers can be found in Table 2). (d) Comparative intensity graphs of selected protein spots generated by PDQuest Advance software (Bio-Rad). Figure 7: WB validation of the selected proteins identified by MALDI-TOF/TOF analysis. Western blot analysis was used for validation of proteomic results. Conventional 12% SDS gels were run with whole cell extracts. (a) Three lanes for SHED, DPSCs, and NDP-MSC are shown for each antibody. Galectin-1, DJ-1, UCHL-1, and HNRNPH1 antibodies were used for validation of expressions of the identified proteins. GAPDH was used for the normalization of each protein sample. (b) Relative protein intensities of bands were measured by Quantity One 1D analysis software (Bio-Rad). Each WB was repeated for three times. (c) Representative images of protein spots selected for WB analysis (protein identities corresponding to SSP numbers can be found in Table 2). (d) Comparative intensity graphs of selected protein spots generated by PDQuest Advance software (Bio-Rad). 17 Stem Cells International Abbreviations DP: Dental pulp derived hNDP: Natal dental pulp derived SHED: Stem cells from human exfoliated deciduous MSCs: Mesenchymal stem cells MALDI: Matrix assisted laser desorption ionization TOF: Time of flight WB: Western blot EDTA: Ethylenediaminetetraacetic acid TMB: 3,3󸀠,5,5󸀠-Tetramethylbenzidine dihydrochloride RTA: Relative telomerase activity CHAPS: 3-[(3-Cholamidopropyl) dimethylammonio]-1-propanesulfonate DTT: Dithiothreitol IPG: Immobilized pH gradient MS: Mass spectrometry SDS-PAGE: Sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Ethical Approval This study was approved by Ethics Committee of Kocaeli University. Conflict of Interests The authors declare that they have no conflict of interests. DP: Dental pulp derived hNDP: Natal dental pulp derived SHED: Stem cells from human exfoliated deciduous MSCs: Mesenchymal stem cells MALDI: Matrix assisted laser desorption ionization TOF: Time of flight WB: Western blot EDTA: Ethylenediaminetetraacetic acid TMB: 3,3󸀠,5,5󸀠-Tetramethylbenzidine dihydrochloride RTA: Relative telomerase activity CHAPS: 3-[(3-Cholamidopropyl) dimethylammonio]-1-propanesulfonate DTT: Dithiothreitol IPG: Immobilized pH gradient MS: Mass spectrometry SDS-PAGE: Sodium dodecyl sulfate-polyacrylamide gel electrophoresis. [6] E. Kara¨oz, B. N. Do˘gan, A. Aksoy et al., “Isolation and in vitro characterisation of dental pulp stem cells from natal teeth,” Histochemistry and Cell Biology, vol. 133, no. 1, pp. 95–112, 2010. [7] M. Dominici, K. le Blanc, I. Mueller et al., “Minimal crite- ria for defining multipotent mesenchymal stromal cells: the International Society for Cellular Therapy position statement,” Cytotherapy, vol. 8, no. 4, pp. 315–317, 2006. [8] H. Mi, A. Muruganujan, J. T. Casagrande, and P. D. Thomas, “Large-scale gene function analysis with the PANTHER classi- fication system.,” Nature protocols, vol. 8, no. 8, pp. 1551–1566, 2013. [9] I. Flores and M. A. Blasco, “The role of telomeres and telomerase in stem cell aging,” FEBS Letters, vol. 584, no. 17, pp. 3826–3830, 2010. [10] C. Yang, S. Przyborski, M. J. Cooke et al., “A key role for telomerase reverse transcriptase unit in modulating human embryonic stem cell proliferation, cell cycle dynamics, and in vitro differentiation,” Stem Cells, vol. 26, no. 4, pp. 850–863, 2008. [11] E. Hiyama and K. Hiyama, “Telomere and telomerase in stem cells,” British Journal of Cancer, vol. 96, no. 7, pp. 1020–1024, 2007. [12] R. Patil, B. M. Kumar, W.-J. Lee et al., “Multilineage potential and proteomic profiling of human dental stem cells derived from a single donor,” Experimental Cell Research, vol. 320, no. 1, pp. 92–107, 2014. References [1] C. Cookson, “Stem cells: East...and West. Country report: United Kingdom,” Scientific American, vol. 293, no. 1, article A23, 2005. [2] M. Mimeault, R. Hauke, and S. K. Batra, “Stem cells: a revolution in therapeutics—recent advances in stem cell biology and their therapeutic applications in regenerative medicine and cancer therapies,” Clinical Pharmacology and Therapeutics, vol. 82, no. 3, pp. 252–264, 2007. [3] A. D. Ho, W. Wagner, and W. Franke, “Heterogeneity of mesenchymal stromal cell preparations,” Cytotherapy, vol. 10, no. 4, pp. 320–330, 2008. [4] G. C. Parker, M. Anastassova-Kristeva, L. M. Eisenberg et al., “Stem cells: shibboleths of development, part II: toward a functional definition,” Stem Cells and Development, vol. 14, no. 5, pp. 463–469, 2005. [5] W. Wagner, R. E. Feldmann Jr., A. Seckinger et al., “The heterogeneity of human mesenchymal stem cell preparations— evidence from simultaneous analysis of proteomes and tran- scriptomes,” Experimental Hematology, vol. 34, no. 4, pp. 536– 548, 2006. Conflict of Interests The authors declare that they have no conflict of interests. Stem Cells International 18 [28] D. K. St. Clair, T. D. Oberley, K. E. Muse, and W. H. St. Clair, “Expression of manganese superoxide dismutase promotes cellular differentiation,” Free Radical Biology and Medicine, vol. 16, no. 2, pp. 275–282, 1994. [13] E. Eleuterio, O. Trubiani, M. Sulpizio et al., “Proteome of human stem cells from periodontal ligament and dental pulp,” PLoS ONE, vol. 8, no. 8, Article ID e71101, 2013. [14] K. M. Mrozik, P. S. Zilm, C. J. Bagley et al., “Proteomic characterization of mesenchymal stem cell-like populations derived from ovine periodontal ligament, dental pulp, and bone marrow: analysis of differentially expressed proteins,” Stem Cells and Development, vol. 19, no. 10, pp. 1485–1499, 2010. [29] A. Petkau, K. Kelly, W. S. Chelack, S. D. Pleskach, C. Barefoot, and B. E. Meeker, “Radioprotection of bone marrow stem cells by superoxide dismutase,” Biochemical and Biophysical Research Communications, vol. 67, no. 3, pp. 1167–1174, 1975. [15] R. Hass, C. Kasper, S. B¨ohm, and R. Jacobs, “Different popu- lations and sources of human mesenchymal stem cells (MSC): a comparison of adult and neonatal tissue-derived MSC,” Cell Communication and Signaling, vol. 9, article 12, 2011. [30] J. Eastgate, J. Moreb, H. S. Nick, K. Suzuki, N. Taniguchi, and J. R. Zucali, “A role for manganese superoxide dismutase in radioprotection of hematopoietic stem cells by interleukin-1,” Blood, vol. 81, no. 3, pp. 639–646, 1993. [16] L. Ma, B. Sun, L. Hood, and Q. Tian, “Molecular profiling of stem cells,” Clinica Chimica Acta, vol. 378, no. 1-2, pp. 24–32, 2007. [31] G. Kasper, L. Mao, S. Geissler et al., “Insights into mesenchymal stem cell aging: involvement of antioxidant defense and actin cytoskeleton,” Stem Cells, vol. 27, no. 6, pp. 1288–1297, 2009. [17] H. Baharvand, A. Fathi, D. van Hoof, and G. H. Salekdeh, “Concise review: trends in stem cell proteomics,” Stem Cells, vol. 25, no. 8, pp. 1888–1903, 2007. [32] S. Pulavendran, G. Thiyagarajan, and V. Ramakrishnan, “Antioxidant and anti inflammatory action of stem cells in car- diac disease,” Journal of Advanced Laboratory Research in Bio- logy, vol. 1, no. 2, pp. 71–78, 2010. [18] K. Zhou, A. Muroyama, J. Underwood et al., “Actin-related protein2/3 complex regulates tight junctions and terminal dif- ferentiation to promote epidermal barrier formation,” Proceed- ings of the National Academy of Sciences of the United States of America, vol. 110, no. Conflict of Interests 40, pp. E3820–E3829, 2013. [33] K. P. Lu, G. Finn, T. H. Lee, and L. K. Nicholson, “Prolyl cis-trans isomerization as a molecular timer,” Nature Chemical Biology, vol. 3, no. 10, pp. 619–629, 2007. [34] Y. M. Lee, S. Y. Shin, S. S. Jue, I. K. Kwon, E. H. Cho, and E. S. Cho, “The role of PIN1 on odontogenic and adipogenic differentiation in human dental pulp stem cells,” Stem Cells and Development, vol. 23, no. 6, pp. 618–630, 2014. [19] Z. C. Xu, Q. Zhang, and H. Li, “Differentiation of human hair follicle stem cells into endothelial cells induced by vascular endothelial and basic fibroblast growth factors,” Molecular Medicine Reports, vol. 9, no. 1, pp. 204–210, 2014. [35] H. Toko, N. Hariharan, M. H. Konstandin et al., “Differential regulation of cellular senescence and differentiation by prolyl isomerase Pin1 in cardiac progenitor cells,” The Journal of Biological Chemistry, vol. 289, no. 9, pp. 5348–5356, 2014. [20] S. A. Salem, A. N. M. Hwei, A. Saim et al., “Human adipose tissue derived stem cells as a source of smooth muscle cells in the regeneration of muscular layer of urinary bladder wall,” Malaysian Journal of Medical Sciences, vol. 20, no. 4, pp. 80–87, 2013. [36] K. Nakamura, I. Kosugi, D. Y. Lee et al., “Prolyl isomerase pin1 regulates neuronal differentiation via 𝛽-catenin,” Molecular and Cellular Biology, vol. 32, no. 15, pp. 2966–2978, 2012. [21] P. A. J. Huber, C. S. Redwood, N. D. Avent, M. J. A. Tanner, and S. B. Marston, “Identification of functioning regulatory sites and a new myosin binding site in the C-terminal 288 amino acids of caldesmon expressed from a human clone,” Journal of Muscle Research and Cell Motility, vol. 14, no. 4, pp. 385–391, 1993. [37] E. Kara¨oz, P. C. Demircan, ¨O. Saflam, A. Aksoy, F. Kaymaz, and G. Duruksu, “Human dental pulp stem cells demonstrate better neural and epithelial stem cell properties than bone marrow-derived mesenchymal stem cells,” Histochemistry and Cell Biology, vol. 136, no. 4, pp. 455–473, 2011. [22] P. Graceffa, L. P. Adam, and K. G. Morgan, “Strong interaction between caldesmon and calponin,” The Journal of Biological Chemistry, vol. 271, no. 48, pp. 30336–30339, 1996. [38] M. Nishi, H. Akutsu, S. Masui et al., “A distinct role for pin 1 in the induction and maintenance of pluripotency,” The Journal of Biological Chemistry, vol. 286, no. 13, pp. 11593–11603, 2011. [23] M. A. Conflict of Interests Hayat, “Stem cells and cancer stem cells,” in Therapeutic Applications in Disease and Injury, K. Tanimoto, Y.-C. Huang, and K. Tanne, Eds., Springer, Berlin, Germany, 2012. [39] K. Y. Chong, C.-C. Lai, and C.-Y. Su, “Inducible and constitutive HSP70s confer synergistic resistance against metabolic chal- lenges,” Biochemical and Biophysical Research Communications, vol. 430, no. 2, pp. 774–779, 2013. [24] T. Macalma, J. Otte, M. E. Hensler et al., “Molecular characteri- zation of human zyxin,” Journal of Biological Chemistry, vol. 271, no. 49, pp. 31470–31478, 1996. [40] D. W. Miller, R. Ahmad, S. Hague et al., “L166P mutant DJ-1, causative for recessive Parkinson’s disease, is degraded through the ubiquitin-proteasome system,” Journal of Biological Chemistry, vol. 278, no. 38, pp. 36588–36595, 2003. [25] J. K. van Tam, K. Uto, M. Ebara, S. Pagliari, G. Forte, and T. Aoyagi, “Mesenchymal stem cell adhesion but not plasticity is affected by high substrate stiffness,” Science and Technology of Advanced Materials, vol. 13, no. 6, Article ID 064205, 2012. [41] T. Weiler, Q. Du, O. Krokhin et al., “The identification and characterization of a novel protein, c19orf10, in the synovium,” Arthritis Research and Therapy, vol. 9, article R30, 2007. [26] M. G. Roubelakis, K. I. Pappa, V. Bitsika et al., “Molecular and proteomic characterization of human mesenchymal stem cells derived from amniotic fluid: comparison to bone marrow mesenchymal stem cells,” Stem Cells and Development, vol. 16, no. 6, pp. 931–951, 2007. [42] A. Guasch, K. Aloria, R. P´erez, J. Avila, J. C. Zabala, and M. Coll, “Three-dimensional structure of human tubulin chaperone cofactor A,” Journal of Molecular Biology, vol. 318, no. 4, pp. 1139–1149, 2002. [27] C. Conrad, E. Zeindl-Eberhart, S. Moosmann, P. J. Nelson, C. J. Bruns, and R. Huss, “Alkaline phosphatase, glutathione-S- transferase-P, and cofilin-1 distinguish multipotent mesenchy- mal stromal cell lines derived from the bone marrow versus peripheral blood,” Stem Cells and Development, vol. 17, no. 1, pp. 23–27, 2008. [43] L. L. Chen and G. G. Carmichael, “Altered nuclear retention of mRNAs containing inverted repeats in human Embryonic stem cells: functional role of a nuclear noncoding RNA,” Molecular Cell, vol. 35, no. 4, pp. 467–478, 2009. 19 19 Stem Cells International [44] Y.-M. Ji, S. H. Jeon, J.-Y. Park, J.-H. Chung, Y.-H. Choung, and P.-H. Choung, “Dental stem cell therapy with calcium hydrox- ide in dental pulp capping,” Tissue Engineering A, vol. 16, no. 6, pp. 1823–1833, 2010. [45] X. Chen, G. Conflict of Interests Luther, W. Zhang et al., “The E-F hand calcium- binding protein s100a4 regulates the proliferation, survival and differentiation potential of human osteosarcoma cells,” Cellular Physiology and Biochemistry, vol. 32, no. 4, pp. 1083–1096, 2013. [46] J. Yamada and S. Jinno, “S100A6 (calcyclin) is a novel marker of neural stem cells and astrocyte precursors in the subgranular zone of the adult mouse hippocampus,” Hippocampus, vol. 24, no. 1, pp. 89–101, 2014.
https://openalex.org/W2944086253	https://www.e3s-conferences.org/10.1051/e3sconf/20199403011/pdf	English	null	The Potency of the Modernized GNSS Signals for Real-Time Kinematic Positioning	E3S web of conferences	2,019	cc-by	4,663	The Potency of the Modernized GNSS Signals for Real-Time Kinematic Positioning CGCS2000 is a global, three-dimensional, right-handed, orthogonal, geocentric coordinate system which referred to ITRF97 (International Terrestrial Reference Frame 97) at the epoch 2000.0. The Potency of the Modernized GNSS Signals for Real-Time Kinematic Positioning Dhota Pradipta1, Heri Andreas1, Dina A. Sarsito1, and Hendy F. Suhandri2 1Geodesy Research Group, Faculty of Earth Sciences and Technology, Institute of Technology, Bandung, Indonesia 2Coastal, Ocean and Maritime Engineering Research Group, Faculty of Earth Sciences and Technology, Institute of Technology, Bandung, Indonesia Abstract. GNSS positioning has become popular in the past decade as an efficient method of precise and real-time positioning. It is relatively low cost and ease-of-use. Up to now, several parameters were defined to characterize the performance of real-time positioning: availability, precision, accuracy. This article evaluates the performance of signal linear combinations for real-time positioning, both for static as well as the kinematic positioning. This article starts with the investigation of linear combinations (LC) rising from the carrier frequencies of the GNSS systems. Some Linear Combination shows potential benefits in carrier phase integer ambiguity resolution, particularly utilizing the Galileo and Beidou signal phase carrier. For each system, a set of combinations was studied, analyzed, and then selected during the development of GNSS positioning method utilizing the Least-squares Ambiguity Decorrelation Adjustment (LAMBDA). Special signal selection can affect the estimated position and its standard deviation. To further analyze, the results obtained from data processing are compared with respect to baselines and signals. The ambiguity fixing rate is correlated with the baseline length and the method as well as the signals that were used. The analysis of the measurement noise level was first conducted to set a baseline for the real-time GNSS positioning application. According to the results and to assess the data quality and positioning performance of GNSS in respect with GPS (Global Positioning System), an experimental test has established using MGEX data. This research investigates the satellite visibilities, multipath, Signal to Noise Ratio (SNR), and positioning performance. It is shown that in every epoch, at least 8 satellites are visible. The SNR’s are up to 60 dBHz, the code multipath residuals varies within ~1 m, while the phase residuals varies by about ~2 cm. hence the modernized GNSS signals have potencies to improves the RTK positioning. Differs from GPS, BDS adopts the China Geodetic Coordinate System 2000 (CGCS2000) as the coordinate system and BeiDou navigation satellite system time (BDT) as the time system of BDS. GAL using Galileo Terrestrial Reference Frame (GTRF) and Galileo System Time (GST). https://doi.org/10.1051/e3sconf/20199403011 https://doi.org/10.1051/e3sconf/20199403011 E3S Web of Conferences 94, 03011 (2019) ISGNSS 2018 1. Introduction Differs from GPS, BDS adopts the China Geodetic Coordinate System 2000 (CGCS2000) as the coordinate system and BeiDou navigation satellite system time (BDT) as the time system of BDS. GAL using Galileo Terrestrial Reference Frame (GTRF) and Galileo System Time (GST). CGCS2000 is a global, three-dimensional, right-handed, orthogonal, geocentric coordinate system which referred to ITRF97 (International Terrestrial Reference Frame 97) at the epoch 2000.0. Satellite based navigation technology is continuously and rapidly growing in recent time. The Global Navigation Satellite System (GNSS) are Galileo System (GAL) and BeiDou System (BDS), which is developed and operated by European Union and China. Currently, GAL and BDS is entering the third phase of development that is expected to enhance the positioning performance will continue to the final phase on 2020 with global coverage of positioning and navigation system by a constellation of 35 satellites, which are consists of 5 Geostationary Earth Orbit (GEO) satellites, 3 Inclined Geosynchronous Orbit (IGSO) satellites and 27 Medium Earth Orbit (MEO) satellites [1] Table 1.1 Parameters of WGS84, CGS2000, GTRF ellipsoid Similar to GPS, GAL and BDS provides triple-frequency signal. Table I gives the frequencies and wavelengths of the signal from the GAL and the BDS as well as the GPS. © The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (http://creativecommons.org/licenses/by/4.0/). Corresponding author: dhota@fitb.itb.ac.id 2. Data Quality Assessment In order to investigate the advantages of GNSS, the satellite visibilities were analysed. Better satellite visibilities lead a better geometry of satellites constellation, namely Dilution of Precision (DOP), which introduces to a better position accuracy and precision. DOP can be divided into several terms, which are vertical DOP (VDOP), horizontal DOP (HDOP), position DOP (PDOP) and time DOP (TDOP). Those terms can be generalized by using Geometric DOP (GDOP) term. In order to achieve the highest position accuracy and precision (low DOP value), the satellites geometry is not supposed to clustered together in a single quadrant. Fig. 2.2 Code Multipath for GNSS signals (GPS, Glonass, Galileo, and Beidou) 2.1 Multipath This section described the methods that used to assess the data quality of GNSS. The multipath of GNSS can be investigate by applied the Multipath Combination (MPC) algorithm. The MPC can be expressed as [6]: Several studies have been conducted in several areas, particularly in northern hemisphere. The combination of GPS, GAL, and BDS can increase the observation data quality as well as the positioning performance [1]. 𝑀𝑃1 = 𝑃1 − 𝑓12+𝑓22 𝑓12−𝑓22 𝐵1 + 2𝑓22 𝑓12−𝑓22 𝐵2 (1) 𝑀𝑃2 = 𝑃2 − 2𝑓12 𝑓12−𝑓22 𝐵1 + 𝑓12+𝑓22 𝑓12−𝑓22 𝐵2 (2) (1) With a highly growth of infrastructure in Indonesia, the use of GNSS technology is inevitable, especially for the surveying and mapping purposes. Further research over GNSS data quality and positioning performance in Indonesia is needed. To assess the data quality and positioning performance GNSS in respect with GPS (Global Positioning System) over Indonesia, an experimental network had established using MGEX data. This research investigates the satellite visibilities, multipath, Signal to Noise Ratio (SNR), and the positioning performance. (2) where P and L are the GNSS pseudorange and carrier phase range respectively and f is the frequency of the carrier phase. By using these combinations, the first order of ionospheric delay and geometric range from satellite and receiver are eliminated, while noise is assumed to be negligible due to the accuracy of the pseudorange observation. where P and L are the GNSS pseudorange and carrier phase range respectively and f is the frequency of the carrier phase. By using these combinations, the first order of ionospheric delay and geometric range from satellite and receiver are eliminated, while noise is assumed to be negligible due to the accuracy of the pseudorange observation. Fig. 2.2 Code Multipath for GNSS signals (GPS, Glonass, Galileo, and Beidou) Fig. 2.2 Code Multipath for GNSS signals (GPS, Glonass, Galileo, and Beidou) * Corresponding author: dhota@fitb.itb.ac.id E3S Web of Conferences 94, 03011 (2019) ISGNSS 2018 https://doi.org/10.1051/e3sconf/20199403011 E3S Web of Conferences 94, 03011 (2019) Fig. 2.1 shows the number of observed satellites for GPS, GAL, BDS, and combined systems with a cut off angle of 10o. It can be found that the satellite visibilities of GNSS combinations is more stable than the GPS. By using GNSS, at least 10 satellites can be observed for more than 80% of observation time., while the number of combined systems observed satellites is vary from 12 to 22 with an average of 14 satellites. The ellipsoid used in CGCS2000 is slightly different from WGS84 (World Geodetic System 1984) as seen on Table II. Theoretically, the differences induced by those ellipsoids is within 1 millimeter in the equator [1]. BDT adopts seconds without any leap seconds as the international system of unit (SI). The BDT’s start epoch was 00:00:00, 1 January 2006 of Universal Time Coordinated (UTC). In respect with GPS Time (GPT), BDT is 14 s ahead GPT (BDT = GPT + 14 s) [1]. Fig. 2.1 Satellite visibility 2.1 Multipath Fig. 2.1 Satellite visibility Table 1.2 Frequencies and wavelength of GNSS signals (GPS, Galileo, and Beidou) Table 1.2 Frequencies and wavelength of GNSS signals (GPS, Galileo, and Beidou) Fig. 2.1 Satellite visibility Fig. 2.1 Satellite visibility 3. Performance Test The user should be careful when comparing different GNSS receivers, particularly for older models e.g. Trimble that provided the signal quality in “arbitrary manufacture unit” (AMU). AMU units are dependent and need to be converted by a conversion formula because the value can differ by up to 3 dB from the original value [4]. Different generations of GNSS satellites have inherently SNR(dB) = S/N Assuming the GNSS signal strength is S and the noise level is N, the basic formula to measure the GNSS signal strength is S/N. If the carrier waves facing obstructions, S will get affected by attenuation, because many signals have a very wide dynamic range and are expressed using the logarithmic decibel scale, signal and noise may be expressed in decibels (dB). Assuming the system noise (N0) is several magnitudes smaller than the signal strength (S), the normalized signal quality is [3][4]: Because multipath errors are site-specific and particularly affect the code ranges, the use of E5 Galileo provides an advantage, as this signal shows a low multipath/noise residual behavior compared to all other GNSS signals. Moreover, due to its higher signal strength, the E5 Galileo signal is preferable for positioning because the signal is more resilient against outside interference than other GNSS signals and offers advantages to mitigate multipath/noise and ionospheric errors. 𝑆𝑁𝑅 (𝑑𝐵𝐻𝑧) = 𝑆−𝑁= 10. 𝑙𝑜𝑔10(𝑆)(𝑑𝐵) − (𝑁0. 𝐵𝐿)(𝑑𝐵𝐻𝑧) (6) (6) 2.2 Signal to Noise (SNR) Ratio Concerning satellite configuration, site-specific factors as well as atmospheric effects, the quality of GNSS observations may become inconsistent and affecting SNR. In addition, observation weighting plays a dominant 2 E3S Web of Conferences 94, 03011 (2019) E3S Web of Conferences 94, 03011 (2019) ISGNSS 2018 https://doi.org/10.1051/e3sconf/20199403011 different signal strengths, which could cause different SNR values with nothing wrong at all. role when GNSS receivers calculate positions by measuring pseudo-distances to transmitting satellites. A GNSS receiver performance mainly depends on the signal power in the receiver’s tracking loops [2]. Different GNSS receivers with the same antenna tracking the same satellite at the same time may provide different SNR values. These differences could be from band limitation or processing algorithms. In case of independent acquisition and tracking algorithms used by a receiver, the values could be considered to indicate the quality of the received signal when antenna and receiver type, design, and performance are neglected. Hence the SNR depends on the receiver bandwidth, signal acquisition and tracking parameter. There are several methods to measure the GNSS signal strength. However, as the data sent by GNSS are through radio signals, it is a known fact that radio signals cannot maintain their strength for longer distances. The GNSS system employs phase modulation to superimpose data on the radio signals for better reception by the GNSS receiver and the manufacturers employ different algorithms to retrieve the data from the signals for offering the desired data. In order to analyze the behavior of SNR and code multipath/noise, all GNSS signals were processed. By using an epoch by epoch method, each epoch was processed to get SNR and code range residuals and then an averaging method was adapted to get mean values for the figures. All the factors correlating with the elevation angle of the transmitted GNSS signals like the SNR normally grow with increasing satellite elevation angle. SNR is usually expressed in decibels and it refers to the ratio of the signal power and noise power in a given bandwidth. Due to the fact that noise and signal are amplified in the same way, these ratios can be expressed as [3]: Fig. 2.3 SNR mean of GNSS signals (GPS, Galileo, and Beidou) Fig. 2.2 Signal to Noise (SNR) Ratio 2.3 SNR mean of GNSS signals (GPS, Galileo, and Beidou) 𝐶𝑎𝑛𝑡 𝑁𝑎𝑛𝑡≈ 𝑆𝑐𝑜𝑟𝑟 𝑁𝑐𝑜𝑟𝑟≈𝑆 (3) (3) Signal to noise ratio can usually be found in the context of signal baseband of the modulated signal at correlator output (Scorr). The quality of a received GNSS signal is commonly described by carrier to noise ratio of the modulated carrier at the receiving antenna (Cant). The system noise affects the signal quality and the noise and signal are amplified in the same way in the antenna (Nant) and at the correlator output (Ncorr). As the system noise is several magnitudes smaller than Cant and Scorr, therefore the values are normally converted to decibels (dB) to represent a specific bandwidth, thus: Signal to noise ratio can usually be found in the context of signal baseband of the modulated signal at correlator output (Scorr). The quality of a received GNSS signal is commonly described by carrier to noise ratio of the modulated carrier at the receiving antenna (Cant). The system noise affects the signal quality and the noise and signal are amplified in the same way in the antenna (Nant) and at the correlator output (Ncorr). As the system noise is several magnitudes smaller than Cant and Scorr, therefore the values are normally converted to decibels (dB) to represent a specific bandwidth, thus: Fig. 2.3 SNR mean of GNSS signals (GPS, Galileo, and Beidou) In case of the E1 Galileo signal, the SNR and the code multipath/noise residual seems to perform higher than the L1 GPS with up to 1-2 dBHz for signal strength and up to 0.1 m for residual difference, refer to figure 2.2 and 2.3. On the other hand, E5 Galileo performs best in SNR and code multipath/noise residuals at every station tested. When selecting promising signals for future real-time positioning, the combined E5 Galileo signal emerges as one of the alternatives (better than E5a) when using Galileo signals. 𝑆(𝑑𝐵) = 10. 𝑙𝑜𝑔10 (𝑆) and (4) SNR(dB) = S/N (5) (4) (5) (5) (7) GPS solutions are in cm level (~2 cm) 𝐷{𝜙𝑖𝑗} = 𝑡4+𝑛4 (𝑡2−𝑛2)2 𝜎Φ 2 (14) (14) where 𝐷{𝜙𝑖𝑗} denotes the mathematical dispersion. (7) The range of observable 𝜌 appears in the same way as in the original phase observation equation. Moreover, the ionospheric delays are eliminated ( 𝑡2 𝑡2−𝑛2 − 𝑛2 𝑡2−𝑛2 𝑡2 𝑛2) 𝑖𝑖≈ 0, and a combined ambiguity term remains, which does not seem to be integer valued. However, using equation (12) with t and n integer, it is possible to rewrite the ambiguity term to be: ( ) where Δ is the difference between receivers A and B. The superscript -j is the observed satellite. The satellite clock error is eliminated by taking single difference between receivers that observed the same satellite, while the atmospheric biases like tropospheric and ionospheric may be eliminated depending to the length of the baseline. Multipath would be considered as a noise error that could not be eliminated. 𝐸{𝜙𝑖𝑗} = 𝜌+ 𝑡2 𝑡2−𝑛2 𝜆𝑖 (𝑡𝑁𝑖− 𝑛𝑁𝑗) + 𝑒𝑖𝑗= 𝜌+ 𝜆𝑖𝑗 𝑁𝑖𝑗+ 𝑒𝑖𝑗 (13) The remaining receiver clock error is then eliminated by subtracting two single difference observation. Mathematically, a DD is defined as follows: 𝛻𝛥𝐿𝐴,𝐵 𝑗𝑘= 𝛥𝜌𝑗𝑘 𝐴,𝐵+ +𝑀𝐿𝑗𝑘 𝐴,𝐵+ 𝜆𝑁𝑗𝑘 𝐴,𝐵+ 𝜗𝐿𝑗𝑘 𝐴,𝐵 (13) where 𝜆𝑖𝑗 denotes the artificial wavelength and 𝑁𝑖𝑗 the integer ambiguity of the ionosphere free combination and 𝑒𝑖𝑗 denotes multipath and noise values. (8) ( ) where the two satellites are denoted as he superscript-j and k. The atmospheric biases are negligible in Eq.8, however the noise error is multiplied up to two times with respect to SD. where the two satellites are denoted as he superscript-j and k. The atmospheric biases are negligible in Eq.8, however the noise error is multiplied up to two times with respect to SD. A consequence of taking the ionosphere-free linear combination is that the noise of the ionosphere-free observable is increased compared to the noise of the original phase observations. When it is assumed that two original observables are uncorrelated and have same precision 𝜎Φ𝑖= 𝜎Φ𝑗= 𝜎Φ in DD mode, the variance of LC follows: By using MGEX data, we assumed that the precision of GAL and BDS are similar with GPS for all of the position components using DD method in kinematic solutions. The precision of the horizontal and vertical components is in cm level. The ambiguity fixing rate of GPS, BDS, GAL, and combined are more than 90% with the coordinate differences w.r.t. 3.1 Relative Positioning The performance of relative positioning was analyzed by using double difference (DD) positioning using carrier phase range that constructed by differencing two single 3 3 E3S Web of Conferences 94, 03011 (2019) E3S Web of Conferences 94, 03011 (2019) ISGNSS 2018 https://doi.org/10.1051/e3sconf/20199403011 difference (SD) observation. SD can be described as follows [6]: 𝐸{𝜙𝑖𝑗} = 𝜌+ 𝑡2 𝑡2−𝑛2 𝜆𝑖𝑁𝑖− 𝑛2 𝑡2−𝑛2 𝜆𝑗𝑁𝑗− ( 𝑡2 𝑡2−𝑛2 − 𝑛2 𝑡2−𝑛2 𝑡2 𝑛2) 𝑖𝑖+ 𝑒𝑖𝑗 (12) 𝛥𝐿𝐴,𝐵 𝑗 = 𝛥𝜌𝑗 𝐴,𝐵+ 𝑑𝑗 𝑡𝑟𝑜𝑝𝐴,𝐵−𝑑𝑗 𝑖𝑜𝑛𝐴,𝐵+ 𝑐(𝑑𝑡− 𝑑𝑇)𝑗 𝐴,𝐵+ +𝑀𝐿𝑗 𝐴,𝐵+ 𝜆𝑁𝑗 𝐴,𝐵+ 𝜗𝐿𝑗 𝐴,𝐵 ( (12) 3.2 Ionosphere Free Linear Combination Denoting the greatest common divisor as c, we may write for the numerator and the denominator of the wavelength ratio 𝑡= 𝑐· t𝑐 and 𝑛= 𝑐· n𝑐, where c ≥ 1. Several linear combinations could be introduced beside the standard L1/L2 GPS ionosphere-free linear combination [2][9]. Moreover, it is even possible to design useful triple frequency linear combinations from all frequencies in GNSS. As the ionospheric delay is dispersive, the phase observables delay of Φ𝑖 can be related to the delay of Φ𝑗 by the known ratio of wavelengths of the two observables: 3.3. Data Processing and methodology Table 3.1 Data for data processing Observation Data (RINEX) DOY (280, 2017) CEBR-VILL IGS (35 km) GPS+GLO+GAL+BDS+SBAS IGS data precise orbit file IGS (pre) Auxiliary data (updated using data from NGS) antenna information file receiver information file satellite problem file noise model parameter file information coordinate file The processing strategy was designed as follow: 3.3. Data Processing and methodology Table 3.1 Data for data processing 𝑖𝑗= (𝜆𝑗 2 𝜆𝑖 2)𝑖𝑖 ⁄ (9) (9) where the two observables are in order, 𝜆𝑗> 𝜆𝑖, and the ratio can be denoted as: where the two observables are in order, 𝜆𝑗> 𝜆𝑖, and the ratio can be denoted as: 𝜆𝑗 𝜆𝑖= 𝑡 𝑛 , 𝑡> 𝑛 (10) (10) where both t and n are (positive) integers. Using the wavelength ratio, the ionosphere free linear combination 𝜙𝑖𝑗 of two observables is obtained as: The processing strategy was designed as follow:  The following calculations were carried out using data from observation, i.e. 1 hour of observation Rinex data 1 second sampling rate. Calculations were also carried out with Glonass signals (if available) during similar observation 𝐸{𝜙𝑖𝑗} = 𝑡2 𝑡2−𝑛2 𝐸{𝜙𝑖} − 𝑛2 𝑡2−𝑛2 𝐸{𝜙𝑗} (11) (11) Introducing the integer ambiguities 𝑁𝑖 and 𝑁𝑗 (11) can be detailed as: 4 E3S Web of Conferences 94, 03011 (2019) ISGNSS 2018 https://doi.org/10.1051/e3sconf/20199403011 Figure 4.1 and 4.2 shows the estimated kinematic position with GNSS data. To further analyze, the pseudorange and the carrier phase residuals obtained from observation are compared. The figure shows that there is similar accuracy up to 2 cm. According to the figure 4.1 and 4.2, all signals provides similar results position (about 2 cm of position deviation) to L1/L2 GPS solution as reference. However, the Galileo gives a smooth pattern of result. 3.2 Ionosphere Free Linear Combination On the other hand, when combination signals were processed, a combination between GPS+BDS displays a slightly good trend with respect to L1/L2 GPS comparing with other combinations. periods in equal satellite geometry. Integer ambiguities are estimated and resolved on an epoch-by-epoch basis in kinematic positioning mode. The initial scenario assumes all the observations are processed in difference mode. Hence, the atmospheric errors and clock errors are taken into account. To create a scenario with different baseline, the data were sorted out by the stations to arrange the conditions required. As the local time and the location of the test are essentially affecting the results due to variations in the ionosphere and troposphere activity, calculations of the measurement effect on same time window sessions were performed. Fig. 4.1 Kinematic Solution from GNSS sateliltes with baseline up to 35 km (GPS, Glonass, Galileo, and Beidou) Fig. 4.2 Kinematic Solution from GNSS satellite combinations with baseline up to 35 km (GPS, Glonass, Galileo, and Beidou) Fig. 4.1 Kinematic Solution from GNSS sateliltes with baseline up to 35 km (GPS, Glonass, Galileo, and Beidou)  Mask angles of ten degrees were selected. The receiver noise level was neglected, and when required, IGS precise orbit corrections were used for reference. Tropospheric effects are also independent of frequency. And using a standard model correction.  Furthermore, the observations were grouped in short-term observations (less than and equal to 60 minutes) for positioning for static(reference) and kinematic positioning.  Additionally, since most observation data covers satellites with low elevation angles, observation to satellites which are below 10 degrees are removed to minimize the influence of errors in the data processing. Fig. 4.1 Kinematic Solution from GNSS sateliltes with baseline up to 35 km (GPS, Glonass, Galileo, and Beidou)  The distance between base and rover was up to 40 kilometers. ( ) Fig. 4.2 Kinematic Solution from GNSS satellite combinations with baseline up to 35 km (GPS, Glonass, Galileo, and Beidou)  Ionosphere free linear combination and the MLAMBDA method was used to estimate the carrier phase ambiguities [5][8]. 5. Conclusion Fig. 4.3 shows code residuals for all available satellites. By taking a large number of all satellite observations, the code multipath/noise residuals for L5 and E5 are smaller than L1/L2 GPS residuals. On average, the L5 and E5 provides ~1 m deviation. However, by using L5 and E5 signals are allowed more accurate combined code-and-carrier observable to mitigate ionospheric errors because it has the strongest signal strength of the modernized GNSS signals tested. This research has been conducted to study the performance of the GNSS real time positioning technology introducing modernized signals on medium baseline, and by means of a free software application. Most of the investigated signal linear combinations are capable to deliver accurate, precise and consistent position solutions, although they perform differently in terms of noise, ionospheric content and ambiguity resolution ability. This research focuses on single baselines observed over short to medium periods. Fig. 4.4 Phase Residual from GNSS satellite signals (GPS, Glonass, Galileo, and Beidou) Because multipath errors are site-specific the rover environment has to be chosen carefully. In case of evident multipath the E5 Galileo delivers an advantage compared to other signals due to its more multipath resistant codes which also goes in parallel with a higher received signal strength. Other unencrypted signals provide similar SNR- values with a difference of approximately 1-5 dBHz and range noise in zenith direction of up to 0.25 m. On the other hand, the GPS L2 signal in this research shows the lowest SNR value with differences up to 10 dBHz and zenith range noise up to 0.4 m. Fig. 4.4 Phase Residual from GNSS satellite signals (GPS, Glonass, Galileo, and Beidou) Fig. 4.4 shows carrier-phase residual for medium baseline (35 km) from experiment test (MGEX data). Good solutions were possible due to several available signals and satellites. Hence, the utilization of GNSS data as combined signal. GPS, GAL, BDS, and its combination provides similar results that varies up to 5 cm for kinematic positioning. The standard deviation of the position differences using GNSS dual frequency data of several minutes (up to 60 minutes) is very close to the L1/L2 GPS solution. Galileo and Beidou signals can be used as advanced substitute and simultaneously in case of insufficient or improper L2 GPS observations. Need to be taken into account, the results are of course affected by the number of available satellites, satellite geometry, and observation time spans. 4. Results 4.3 Code Residual from GNSS satellite signals (GPS, Glonass, Galileo, and Beidou) These characteristics can open the possibility of performing code-range measurements using modernized GNSS signals at the decimeter level and enable a better mitigation of multipath effects, particularly for signal combination with small code residuals. These characteristics can open the possibility of performing code-range measurements using modernized GNSS signals at the decimeter level and enable a better mitigation of multipath effects, particularly for signal combination with small code residuals. Fig. 4.5 Standard deviation from GNSS satellite combinations for baseline up to 35 km (GPS, Glonass, Galileo, and Beidou) 4. Results Result from data processing shows that a combined signal (GPS+BDS and GPS+GAL) can be improved the resolution time when the GEO satellites is introduced for ambiguity fixing in case of obstruction problems. Based on the results, it is shown that in every epoch, at least ten satellites of GNSS are visible. The SNR for BDS vary within 25-55 dBHz (figure 3) which is less power than other GNSS signals, the code multipath variation of all GNSS signals varies within ~1 m, while the phase residuals variation varies up to ~3 cm. In addition, a GNSS modernized signal has better precision compared with L1/L2 GPS, and the combined used of GPS+BDS and GPS+GAL improves the solutions on positioning, particularly on kinematic positioning. Fig. 4.2 Kinematic Solution from GNSS satellite combinations with baseline up to 35 km (GPS, Glonass, Galileo, and Beidou) 5 5 E3S Web of Conferences 94, 03011 (2019) E3S Web of Conferences 94, 03011 (2019) ISGNSS 2018 https://doi.org/10.1051/e3sconf/20199403011 cm) than the L1/L2 GPS standard deviation of ionosphere- free linear combination. On the other hand, the standard deviation of the L1/L5 GPS-LC and B1/B2 BDS-LC corresponds with the L1/L2 GPS pattern. A short time observation (at least 60 minutes) seems to be sufficient to resolve the ambiguity for kinematic positioning up to 40 km compared with the GNSS ionosphere free linear combination in case of the VILL-CEBR baseline with a length approx. 35 km. With a modernized GNSS system, the E1/E5 Galileo linear combination seems to perform much better than the other frequency combinations. However, when insufficient Galileo observations are not available (less than 5 satellites) the standard deviation becomes worse than L1/L2 GPS. The residual level normally increases at the VILL-CEBR baseline with 35 km, as not all the ambiguities could not be fixed completely by utilized all satellites within observation time span. However, there is a systematic residual pattern in the code and phase residuals which indicates systematic noise that affected the position result (refer to figure 4.3 and 4.4). Fig. 4.3 Code Residual from GNSS satellite signals (GPS, Glonass, Galileo, and Beidou) Fig. 4.5 Standard deviation from GNSS satellite combinations for baseline up to 35 km (GPS, Glonass, Galileo, and Beidou) Fig. 5. Conclusion the standard deviation of the coordinate time series shows interesting features (figure 4.5). The dual carrier phase combination, particularly from Galileo observable performs extremely well. It is mostly better (less than 2 6 https://doi.org/10.1051/e3sconf/20199403011 E3S Web of Conferences 94, 03011 (2019) ISGNSS 2018 E3S Web of Conferences 94, 03011 (2019) Modernized GNSS signals deliver a potency to provide best performance in static and kinematic positioning. Compared to GPS L1/L2 a clear advantage to use Galileo and Beidou linear combination becomes apparent in order to reduce signal noise and multipath significantly with reveals biases up to a few centimeters (~2 cm). These offsets might be caused by un-modeled intersystem biases between GPS, Galileo, and Beidou signals. References [1] Hofmann-Wellenhof, B., Lichtenegger, E. Wasle. (2008). GNSS- Global Navigation Satellite Systems. Springer, Wien. [2] Odijk, D. (2003). Ionosphere-Free Phase Combinations for Modernized GPS. Journal of Surveying Engineering Vol. 129. November [3] Rost, C. and Wanninger, L. (2009). Carrier Phase multipath mitigation based on GNSS signal quality measurements. Journal of Applied Geodesy 3, de Gruyter, pp. 1-8. DOI 10.1515/JAG.2009.009. [4] Amiri-Sikooci, A., Tiberius, C., C., J., M. (2007). Assessing Receiver Noise Using GPS Short Baseline Time Series. GPS Solution, Vol. 11. Pp21-35 DOI 10.1007/s10291-006-0026-8. [5] Davaine, M. (2011). Code Bias and Multipath Estimation with Cascaded Kalman Filter. Thesis. Insitute for Communication and Navigation, Technische Universitaet Muenchen. Munich, November. [6] Xu, G. (2007). GPS Theory, Algorithms and Application. Berlin: Springer. [7] Joseph, A. and Petovello, M. (2010). Measuring GNSS Signal Strength. Inside GNSS. November, pp. 20-25. [8] Teunissen, P. J. G., (1995). The Least-square Ambiguity Decorrelation Adjustment: A Method for Fast GPS Ambiguity Estimation. Journal Geodesy, vol.70. [9] Richert, T., and N. El-Sheimy (2007). Optimal linear combinations of triple frequency carrier phase data from future global navigation satellite systems. GPS Solutions, Vol. 11, No. 1, pp. 11-19. DOI 10.1007/s10291-006-0024-x. 7 7
https://openalex.org/W2535749208	https://boris.unibe.ch/97720/1/pone.0164475.pdf	English	null	Phylogeny of the Viral Hemorrhagic Septicemia Virus in European Aquaculture	PloS one	2,016	cc-by	11,768	RESEARCH ARTICLE Phylogeny of the Viral Hemorrhagic Septicemia Virus in European Aquaculture Michael Cieslak1, Susie S. Mikkelsen2, Helle F. Skall3, Marine Baud4, Nicolas Diserens5, Marc Y. Engelsma6, Olga L. M. Haenen6, Shirin Mousakhani1, Valentina Panzarin7, Thomas Wahli5, Niels J. Olesen2, Heike Schu¨tze1 1 Institute of Infectology (IMED) of the Friedrich-Loeffler-Institut (FLI), Federal Research Institute for Animal Health, Insel Riems, Germany, 2 Section for Virology, National Veterinary Institute, DTU, Frederiksberg, Denmark, 3 Fish Health, Department of Animal Science, Aarhus University, Tjele, Denmark, 4 Ploufragan- Plouzane´ Laboratory (ANSES), Viral Fish Pathology Unit, Universite´ Europe´enne de Bretagne, Technopoˆle Brest-Iroise, Plouzane´, France, 5 Centre for Fish and Wildlife Health (FIWI), Vetsuisse-Faculty, University of Bern, Bern, Switzerland, 6 Central Veterinary Institute of Wageningen UR, NRL for Fish and Shellfish Diseases, Lelystad, the Netherlands, 7 Istituto Zooprofilattico Sperimentaledelle Venezie, OIE Reference Laboratory for Viral Encephalopathy and Retinopathy, Legnaro, Italy a1111 cieslak-michael@gmx.de * cieslak-michael@gmx.de OPEN ACCESS One of the most valuable aquaculture fish in Europe is the rainbow trout, Oncorhynchus mykiss, but the profitability of trout production is threatened by a highly lethal infectious dis- ease, viral hemorrhagic septicemia (VHS), caused by the VHS virus (VHSV). For the past few decades, the subgenogroup Ia of VHSV has been the main cause of VHS outbreaks in European freshwater-farmed rainbow trout. Little is currently known, however, about the phylogenetic radiation of this Ia lineage into subordinate Ia clades and their subsequent geographical spread routes. We investigated this topic using the largest Ia-isolate dataset ever compiled, comprising 651 complete G gene sequences: 209 GenBank Ia isolates and 442 Ia isolates from this study. The sequences come from 11 European countries and cover the period 1971–2015. Based on this dataset, we documented the extensive spread of the Ia population and the strong mixing of Ia isolates, assumed to be the result of the Europe-wide trout trade. For example, the Ia lineage underwent a radiation into nine Ia clades, most of which are difficult to allocate to a specific geographic distribution. Further- more, we found indications for two rapid, large-scale population growth events, and identi- fied three polytomies among the Ia clades, both of which possibly indicate a rapid radiation. However, only about 4% of Ia haplotypes (out of 398) occur in more than one European country. This apparently conflicting finding regarding the Europe-wide spread and mixing of Ia isolates can be explained by the high mutation rate of VHSV. Accordingly, the mean period of occurrence of a single Ia haplotype was less than a full year, and we found a sub- stitution rate of up to 7.813 × 10−4 nucleotides per site per year. Finally, we documented sig- nificant differences between Germany and Denmark regarding their VHS epidemiology, apparently due to those countries’ individual handling of VHS. Citation: Cieslak M, Mikkelsen SS, Skall HF, Baud M, Diserens N, Engelsma MY, et al. (2016) Phylogeny of the Viral Hemorrhagic Septicemia Virus in European Aquaculture. PLoS ONE 11(10): e0164475. doi:10.1371/journal.pone.0164475 Editor: Pedro L. Oliveira, Universidade Federal do Rio de Janeiro, BRAZIL Editor: Pedro L. Oliveira, Universidade Federal do Rio de Janeiro, BRAZIL Copyright: © 2016 Cieslak et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Introduction data collection and analysis, decision to publish, or preparation of the manuscript. The rainbow trout, Oncorhynchus mykiss (Walbaum, 1792), is one of the dominant fish species in European aquaculture in terms of produced weight and value. In 2014, 272,938 metric tons of rainbow trout (from inland waters, brackish waters, and marine waters) [1] worth about 1,000,000 USD were produced in Europe (approximately 4 USD per kilo) [2]. European rain- bow trout production is threatened, however, by viral hemorrhagic septicemia (VHS), which annually leads to substantial economic losses in the trout farming industry. These losses can be a direct consequence of fish mortality or an indirect consequence of disease control measures. In view of its large economic impact, VHS is categorized as a notifiable disease by the OIE (World Organization for Animal Health) [3] and the European Union (Council Directive 2006/88/EC) [4]. The etiologicalagent of VHS is the VHS virus (VHSV), an enveloped single- stranded, negative-sense RNA virus that belongs to the genus Novirhabdovirus in the family Rhabdoviridae [5]. The genome of this virus contains approximately 11,200 nucleotides and encodes six proteins in the following arrangement: 30-N-P-M-G-NV-L-50 (a non-structural protein (NV) and five structural proteins: nucleoprotein (N), phosphoprotein (P), matrix pro- tein (M), glycoprotein (G), and RNA polymerase (L)) [6, 7]. Competing Interests: The authors have declared that no competing interests exist. Competing Interests: The authors have declared that no competing interests exist. Abbreviations: G gene, glycoprotein gene; VHS, viral hemorrhagic septicemia; VHSV, VHS virus. Abbreviations: G gene, glycoprotein gene; VHS, viral hemorrhagic septicemia; VHSV, VHS virus. During acute VHS infection, fish do exhibit external disease signs such as a darkening of the skin, bulging eyes, anemia, bleeding on the skin, gills, eyes, and internal organs, and a bloated abdomen [8]. Mortality can be as high as 100% in fry, but ranges from 5% to 90% in older rain- bow trout [8, 9]. VHSV is transmitted through contact with the urine or reproductive fluids of other infected fish and also through virus-contaminated water or objects [8]. VHS is com- monly a cool- or cold-water disease that is most prevalent at temperatures of 9–12°C [8]. The first records of a rainbow trout exhibiting symptoms similar to those of the current VHS disease are from German trout farms in the early 1930s [10]. Soon after, the disease was reported from several other European countries [9]. Phylogeny of VHSV in European Aquaculture OPEN ACCESS 7892/boris. 7892/boris. Data Availability Statement: All G gene sequences of the 431 isolates are available from the GenBank and Fishpathogens database (accession numbers are within the paper and its Supporting Information files). source: https://doi.org/10.7 source: https://doi.org/10.7 Funding: This work was supported by EMIDA-ERA Net (EU 7th Framework programme), German Federal Office for Agriculture and Food (2811ERA174), Dutch Ministry of Economic Affairs (CVI project no. 1600074-01), and French Agency for Food, Environmental and Occupational Health & Safety. The funders had no role in study design, 1 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0164475 October 19, 2016 PLOS ONE \| DOI:10.1371/journal.pone.0164475 October 19, 2016 Materials and Methods Ethics statement Ethical approval was not required for this study. VHSV samples were obtained from European health servicesand regional laboratories and were isolated from fish on the basis of the COUN- CIL DIRECTIVE 2006/88/EC of the European Union on animal health requirements for aqua- culture animals and products thereof, and on the prevention and control of certain diseases in aquatic animals [4]. Phylogeny of VHSV in European Aquaculture VHSV isolates from European freshwater rainbow trout farms began in 1962, most VHS out- breaks have been found to be caused by subgenogroup Ia isolates [22]. VHSV isolates from European freshwater rainbow trout farms began in 1962, most VHS out- breaks have been found to be caused by subgenogroup Ia isolates [22]. He and colleagues (2014) suggested that the ancestral Ia lineage arose from a pathogenic virus of freshwater-farmed rainbow trout during the 1950s in France [13]. Little is still known, however, about the Ia lineage radiation into Ia clades and their respective spread routes. Cur- rently, there is a hypothesis that the Ia lineage can be divided into two sublineages (Ia-1 and Ia-2) [22]. This hypothesis is based on the observation that most of the isolates from the Ia-1 group were sampled from Danish farms, whereas isolates from the Ia-2 group were largely from other European countries. Given that in RNA viruses (such as the VHSV), the accumulation of mutational changes and biogeographical processes (e.g., an event of spread, vicariance, extinction, or rapid popula- tion growth) take place at similar temporal scales due to the high substitution rate [23], it is assumed that the phylogeny and phylogeographic pattern of VHSV will provide important insights into the epidemiologyof VHS. Accordingly, in this study, we sought to shed new light on the following topics: (1) the phylogenetic radiation of the Ia lineage into subordinate Ia clades; (2) the country-specificdistribution and individual spread route of each Ia clade; and finally, (3) the question if there are indications for a hard polytomy, or a sudden demographic expansion, which would point to a rapid radiation. Due to the particularly high spatiotemporal density of Danish and German Ia isolates in our dataset from approximately the last 20 years, we particularly focused our phylogenetic analysis on those two countries during this period. Our study was carried out within the scope of a multidisciplinary trans-European research project, MOLTRAQ (molecular tracing of viral pathogens in aquaculture), by using the largest dataset of Ia isolates ever compiled. We believe this work will provide an enhanced epidemio- logical understanding of European VHS disease, and our goal is to contribute the epidemiologi- cal knowledge that is a basic prerequisite for developing and implementing efficient VHSV prevention and eradication measures in the future. PLOS ONE \| DOI:10.1371/journal.pone.0164475 October 19, 2016 Introduction However, the viral agent in this disease was not successfully isolated until the early 1960s [11]. Until the mid-1970s, VHS was considered to be a disease specific to trout in Europe, but in the decades that followed, VHSV was also isolated from a variety of farmed and wild fish spe- cies in Europe, North America, Japan, and Korea; VHSV has now been isolated from more than 80 wild and farmed fish species [8]. The chronology of VHSV isolation did not occur in parallel with a simultaneous spreading event of VHSV, however, it resulted mainly from an increase in monitoring efforts, which initially focused mostly on European rainbow trout farms [9]. Europe harbors its own endemic VHSV genogroups in comparison to North America and Asia. Whereas genogroups I (marine and freshwater), II (marine), and III (marine) are endemic to Europe, genogroup IV (marine and freshwater) has almost exclusively been isolated from North American and Asian fish. Currently, there is substantial evidence that these gen- ogroups had already split before the 20th century [12, 13] and therefore long before the virus was discovered in North America (1988) [14] and Asia (1996) [15]. Genogroup I is further divided into six subgenogroups: I(unclassified) (Denmark, freshwater), Ia (predominantly continental Europe), Ib (Northern Europe, marine), Ic (continental Europe), Id (Scandinavia–BalticSea and freshwater), and Ie (Black Sea region); and genogroup IV is divided into three subgenogroups: IVa (North American Pacific Coast), IVb (Great Lakes), and IVc (North American Atlantic Coast) [12, 13, 16, 17]. This phylogenetic classifica- tion has largely been derived from sequences of the G and N genes. In Europe, VHS has a major adverse effect on the production of freshwater-farmed rainbow trout, but disastrous VHS outbreaks have also been recorded in a few cases of farmed turbot (Scophthalmus maximus) and brown trout (Salmo trutta) [18–21]. Since the collection of 2 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0164475 October 19, 2016 RNA extraction Total RNA was extracted from the infected cell culture after two freeze-thaw cycles using the RNeasy mini kit from Qiagen. Then 1.2 mL of the sample was centrifuged at 14,000 rpm for 60 min at 12°C and the resulting pellet was dissolved in 600 μL of RLT buffer. The remaining extraction steps were performed according to the manufacturer’s instructions. Lastly, the RNA was eluted in 30μL of RNase-free water. RT-PCR and G gene sequencing Primers for RT-PCR and sequencing were designed based on the published sequence of VHSV in the GenBank database under the accession number Y18263 (S2 Table). The sequence of the complete G gene was amplified using the primers V2782for and V4664rev (S2 Table). The pre- dicted 1883-bp RT-PCR product (VHSV genome: nucleotides 2782–4664) encompasses the 1524-bp-long G gene, which is localized between nucleotides 2959 and 4482 of the VHSV genome (accession number: Y18263) [25]. RT-PCR was performed using the One Step RT-PCR Kit (Qiagen). The 25-μL reaction mixture consisted of the one-step RT-PCR buffer (Qiagen, including 12.5 mM MgCl2), 40 mM deoxynucleosidetriphosphate (dNTP) mix (10mM of each dNTP), 10 pmol each of the forward and reverse primers, 1.0 μL of the enzyme mix (Qiagen), 20U of RNasin (Promega), and 10–100 ng of total extracted RNA. The following PCR cycling conditions were used: 45 min at 45°C for reverse transcription; 15 min at 95°C for initial activation of DNA polymerase and for inactivation of reverse transcriptase and denatur- ation of the cDNA template; and then 35 cycles of 94°C for 30 s (denaturation), 57°C for 30 s (annealing), and 68°C for 10 min (extension). The reactions were conducted in an automated thermal cycler (MastercyclerGradient, Eppendorf). RT-PCR products were analyzed on 0.7% agarose gels in TAE buffer, and products of the predicted size (approximately 1.9 kb) were eluted by using a QIAquick gel-extraction kit (Qia- gen) according to the manufacturer’s instructions. Next, 50 ng of the eluted RT-PCR product was sequenced directly using specific primers as listed in S2 Table; the sequence of both DNA strands was determined by cycle sequencing using a Big Dye Terminator version 1.1 Cycle Sequencing kit (Applied Biosystems) according to the manufacturer’s instructions. The sequencing product was purified using either Sigma Spin Post-Reaction purification columns (Sigma Aldrich) or Nucleo Spin Columns (Machery Nagel). After denaturation with Hi-Di Formamide, samples were analyzed on an automatic sequencer (ABI 377, Applied Biosystems), and nucleotide sequences were evaluated using the sequencer’s Scanner software version 1.0 (ABI) and Geneious software version 7.1.7 (Biomatters Ltd.) [26]. The full-length G gene sequence was deposited in GenBank and the Fish Pathogens Database [27]. Collection of VHSV isolates We sampled 452 VHSV isolates: 2 from Austria, 1 from the Czech Republic, 152 from Den- mark, 11 from France, 222 from Germany, 21 from Italy, 1 from Kattegat (Scandinavian sea area), 3 from the Netherlands, 37 from Switzerland, and 2 from Turkey. These samples covered the period from 1977–2015 (S1 Table). The viruses were sampled and isolated according to the standardized methods describedin 2001/183/EC [24]. The date of collection,site of collection, and host species of each isolate are summarized in the S1 Table. Viral isolates were propagated at 15°C in the cell lines RTG-2 (CCLV Rie 686) and EPC (CCLV Rie 173). Moreover, 282 G gene sequences of VHSV isolates from GenBank (NCBI: National Center for Biotechnology Information) were also incorporated into our sample collection. For inclusion, the sequence of the complete G gene, the date of collection, and the geographic site of collection all had to be recorded. Finally, our complete VHSV dataset comprises 734 isolates from Austria, Denmark, France, Georgia, Germany, Poland, Slovenia, Switzerland, Turkey, the United Kingdom, waters 3 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0164475 October 19, 2016 Phylogeny of VHSV in European Aquaculture of the Atlantic and Northern Coast of Central Europe, waters of Northern Europe, and the North American Pacific Coast, collected from 1971–2009 (S1 Table). of the Atlantic and Northern Coast of Central Europe, waters of Northern Europe, and the North American Pacific Coast, collected from 1971–2009 (S1 Table). Identification of a polytomy in the phylogeny to indicate a potential hard polytomy A polytomy means that a node splits into more than two nodes. However, in case of a phyloge- netic tree (in our case the bifurcating ML tree), its bifurcation algorithm does not permit the correct phylogenetic resolution of a polytomy. Normally, there are two possible causes for a polytomy. The first possible cause is taxon sampling bias [30, 31] due to a lack of sufficient data or inappropriate analysis of characters. The second possible cause is a hard polytomy [30, 31]. This means that a lineage splits into more than two descendant lineages, also referred to as multifurcation. Such polytomy nodes can occur, for example, when isolates (with a specific haplotype) were simultaneously separated into more than two different environments (e.g., geographic regions). After this, different local scenarios of selection and genetic drift result in the development of individual sublineages (clades) with a common time of origin. Therefore, a hard polytomy can be interpreted as a rapid radiation event. It is widely recognized that polyt- omy relationships are relatively common in intraspecific gene phylogenies [32]. Although an extensive dataset of Ia isolates is available for this study, a sampling bias cannot generally be excluded. Therefore, detected polytomies can be interpreted as a potential indica- tor for a hard polytomy rather than evidence for such. To discover nodes with a tendency toward polytomy we adopted the following procedure. First, we created a phylogenetic Median Joining (MJ) network using the computer program NETWORK version 4.6.1.2 (http://www. fluxus-engineering.com)[33]. This method is well suited to represent the polytomy structure of a node. The program’s default setting of Epsilon (0) was chosen and the transition/transver- sion bias (R) was based on a maximum likelihood estimate obtained using MEGA version 5.2 [29]. Due to the lack of a robustness measure for the MJ network in the NETWORK program, we sought to enhance the reliability of the MJ network topology by examining: (1) if the haplo- types of each ML Ia clade are also grouped in the MJ network; and (2) if they have a phylotem- poral structure. In case of the latter, the expectation would be that older isolates (earlier collection dates) would cluster closer to the ancestral node of the lineage than isolates with more recent collection dates. Secondly, we created a NeighborNet (NN) network [34] using the program SplitsTree version 4.13.1 [35]. Determination of the haplotype The haplotype of a VHSV isolate was determined based on the substitution differences within the complete G gene sequence. Specimens that featured a variation of two possible nucleotides at a single position of the G gene sequence were split into two haplotypes and phylogenetically handled as though they were two isolates. However, if the specimen varied at more than one nucleotide position of the G gene sequence, they were excluded from the dataset. Multiple sequence alignment was performed using Geneious Pro version 7.1.7 (Biomatters Ltd.) [26]. The number of haplotypes was calculated using DnaSP version 5.10.01 [28]. 4 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0164475 October 19, 2016 Phylogeny of VHSV in European Aquaculture Phylogenetic subdivision of the Ia subgenogroup (Ia lineage) into Ia clades Following the principle the lower the bootstrap support value, the less reliable the grouping, we divided the haplotypes of the Ia subgenogroup into subordinate Ia clades on the basis of a Max- imum Likelihood(ML) bootstrap support value of 75%. The bifurcating ML tree, with each node splitting into exactly two descendant branches, was constructed by the computer program MEGA version 5.2 [29]. The best-fit nucleotide substitution model was selected using the Bayesian Information Criterion score with Find Best DNA Model in MEGA version 5.2 [29]. As a result, the general time reversible (GTR) model with gamma rate heterogeneity and invari- ant sites was chosen, and 250 bootstrap replicates were generated to assess the reliability of the clades obtained in the tree. In addition, we further divided those Ia clades with a more complex structure in the MJ network into subsets. This extra division was to better illustrate the phylo- temporal arrangement of their haplotypes, or to group those haplotypes that form a star-like pattern for a separate analysis. PLOS ONE \| DOI:10.1371/journal.pone.0164475 October 19, 2016 The mean period of occurrence in years of a Ia haplotype or clade The mean period of occurrence,in years, of a single Ia haplotype, or an Ia clade, was calculated using the following criteria: 0, occurrencewithin one calendar year; 1, occurrencewithin two calendar years; and so forth. Identification of a polytomy in the phylogeny to indicate a potential hard polytomy This method is also well suited to represent the polyt- omy structure of a node, as it can present alternative parallel branches (splits) in the phyloge- netic topology and even allows bootstrapping. We generated 1,000 bootstrap replicates to assess the reliability of the splits obtained in the splits graph. Finally, for each polytomy node (noticed in the MJ and NN networks), we examined whether such a node was associated with a 5 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0164475 October 19, 2016 Phylogeny of VHSV in European Aquaculture bifurcation conflict in the ML tree. For this purpose, we used a quick principle: the number of internal nodes is less than the number of tips minus 1 [36]. In other words, each polytomy node leads to at least one additional node (depending on the number of descendant lineages, for example: 1, in the case of a trifurcating node; 2, in the case of a quadfurcating node; and so forth) with a low bootstrap support value, which is only present in the ML tree, and not in the MJ nor NN networks. Phylogeographic analysis and nucleotide diversity The phylogeographic pattern of the Ia population was illustrated by the MJ network con- structed by NETWORK version 4.6.1.2 [33]. Nucleotide diversity (π) was calculated using DnaSP version 5.10.01 [28]. Indication of rapid population growth A rapid increase in the size of an Ia clade population (a sudden demographic expansion due to increased population growth) was indicated using the following methods. (1) Visual inspection for a major star-like pattern in the MJ phylogenetic network constructed by NETWORK ver- sion 4.6.1.2 [33]. (2) A coalescent Bayesian skyline analysis [37, 38]. The Bayesian skyline plot was constructed using the program BEAST version 1.8.2 [39] and was visualized with the pro- gram Tracer version 1.5 (from BEAST). A further output of that method was the meanRate (substitution rate) of the sequence used for this calculation. The BEAST input gene files were generated using BEAUti version 1.8.2. Our analyses used the random starting tree for the Mar- kov chain Monte Carlo (MCMC) search, selecting a general time GTR model along with an uncorrelated lognormal relaxed molecular clock. We ran 90 million cycles sampling every 9,000 cycles. Outputs were assessed in TRACER version 1.5 to ensure that values reached sta- tionarity. (3) Statistical tests, including the raggedness index (r), Tajima’s D [40], and Fu’s Fs [41], using DnaSP version 5.10.01 [28]. (4) A mismatch distribution analysis [42, 43], also using DnaSP version 5.10.01 [28]. However, the above mentioned methods may not separate the effects of rapid population growth from a recent selective sweep [44]. Therefore, their results can only be used as an indicator for such a demographic event. Results We successfully sequenced 452 VHSV isolates, and they fell into the previously published phy- logenetic classification of genogroups on the basis of their haplotypes (1 I(unclassified) isolate, 442 Ia isolates, 1 Ib isolate, 6 Ic isolates, and 2 Ie isolate) (S1 Table). The GenBank and Fishpatho- gens database accession numbers of the additional sequences obtained for this study are also listed in S1 Table. In total, the dataset (GenBank samples and samples sequenced for this study) consists of 734 VHSV isolates (4 I(unclassified) isolates, 651Ia isolates, 27 Ib isolates, 11 Ic isolates, 32 Id isolates, 6 Ie isolates, and 1 each of isolates II, III, and IV). Genogroup Ia isolates can be subdivided into 398 haplotypes (S1 Table). Approximately 95% of these 651 Ia isolates were obtained from Oncorhynchus mykiss, 3% from Salmo trutta, 1% from Esox lucius, and 1% from 6 further fish species (isolates from an unknown host species are excluded) (S1 Table). 6 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0164475 October 19, 2016 Phylogeny of VHSV in European Aquaculture Fig 1. Maximum Likelihood trees and NeighborNet networks of the Ia subgenogroup. In diagrams a) and b) the phylogenetic relationship between the Ia clades (clades 1–9), based on the complete G gene sequence, is illustrated as a bifurcating maximum likelihood tree, whereas in diagrams c) and d) it is shown as a NeighborNet network. Diagram a) is based on 650 Ia isolates (S1 Table), diagram b) on 651 Ia isolates (the same isolates as in a) plus the isolate U288000, designated as X). Diagrams c) is based on isolates with the GenBank accession-number AY546571 (9i), LN877188 (9ii), EU708732 (9iii), EU708755 (8iii), EU708748 (8ii), LN877010 (8i), LN876935 (7), AJ233396 (6), LN876803 (5), EU708742 (4), FRG2192 (3), LN876782 (2), and AY546617 (1), diagram d) is based on the same isolates as in c) plus the isolate X. These isolates each represent the Ia clade isolate with the oldest collection date. Isolate AY546576 (VHSV genogroup II) is the outgroup in each diagram. The phylogenetic trees are pictured as cladograms. Numbers above branches represent the bootstrap support values obtained from 250 replicates. In case of the networks, the formation of parallelograms indicates possible alternative split events, and the small gray numbers are bootstrap values for each branch (shown only for values>50%). Black circles marked “1–3” represent nodes that correspond in the networks with a polytomy Gray squares marked “BC1–3” represent Fig 1. Results Maximum Likelihood trees and NeighborNet networks of the Ia subgenogroup. In diagrams a) and b) the phylogenetic relationship between the Ia clades (clades 1–9), based on the complete G gene sequence, is illustrated as a bifurcating maximum likelihood tree, whereas in diagrams c) and d) it is shown as a NeighborNet network. Diagram a) is based on 650 Ia isolates (S1 Table), diagram b) on 651 Ia isolates (the same isolates as in a) plus the isolate U288000, designated as X). Diagrams c) is based on isolates with the GenBank accession-number AY546571 (9i), LN877188 (9ii), EU708732 (9iii), EU708755 (8iii), EU708748 (8ii), LN877010 (8i), LN876935 (7), AJ233396 (6), LN876803 (5), EU708742 (4), FRG2192 (3), LN876782 (2), and AY546617 (1), diagram d) is based on the same isolates as in c) plus the isolate X. These isolates each represent the Ia clade isolate with the oldest collection date. Isolate AY546576 (VHSV genogroup II) is the outgroup in each diagram. The phylogenetic trees are pictured as cladograms. Numbers above branches represent the bootstrap support values obtained from 250 replicates. In case of the networks, the formation of parallelograms indicates possible alternative split events, and the small gray numbers are bootstrap values for each branch (shown only for values>50%). Black circles marked “1–3” represent nodes that correspond in the networks with a polytomy. Gray squares marked “BC1–3” represent nodes that occur only in the trees due to a bifurcation conflict. In the tree diagrams, the bootstrap support value of node Y was increased from 40% to 96% when isolate X was excluded. Small white circles marked “1–3” on a branch indicate the respective connection of this branch to the node 1–3 in the networks. doi:10.1371/journal.pone.0164475.g001 We divided the Ia haplotypes into nine clades (Ia clades 1–9). Their ML bootstrap support values ranged from 78% and 99% (mean value = 94%; standard deviation = 8.62%) (Fig 1a). This division was identical to the Ia-haplotype clade structure in the MJ phylogenetic network (Fig 2). A striking exception was isolate U28800 (hereafter referred to as isolate X) that was the 7 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0164475 October 19, 2016 Phylogeny of VHSV in European Aquaculture Fig 2. Median-joining network based on the complete G gene sequence from 734 VHSV isolates (transition/transversion bias = 5). The country or region of origin is indicated using a color code. The North Sea (labeled by an asterisk) includes also the Barents Sea, English Channel, Kattegat, Norwegian Sea, Rogaland, and Skagerrak. The years represent the range of collection dates for isolates of every European subgenogroup I (I(unclassified), and Ia–e) and Ia clades (clade 1–9). The clades 5, 8, and 9 are additionally subdivided into the subsets 5i–iii, 8i–iii, and 9i–iii. The black circles with the number 1 and 3 represent clear polytomy nodes (node 1 and 3), whereas the black circle with the number 2 represents a intermediate stage between a polytomy and bifurcating node (node 2). The node labeled by the gray circle Y may represents the bifurcating node between the Ia clades, or it may represent the more or less direct ancestor of the polytomy node 3. doi:10 1371/journal pone 0164475 g002 Fig 2. Median-joining network based on the complete G gene sequence from 734 VHSV isolates (transition/transversion bias = 5). The country or region of origin is indicated using a color code. The North Sea (labeled by an asterisk) includes also the Barents Sea, English Channel, Kattegat, Norwegian Sea, Rogaland, and Skagerrak. The years represent the range of collection dates for isolates of every European subgenogroup I (I(unclassified), and Ia–e) and Ia clades (clade 1–9). The clades 5, 8, and 9 are additionally subdivided into the subsets 5i–iii, 8i–iii, and 9i–iii. The black circles with the number 1 and 3 represent clear polytomy nodes (node 1 and 3), whereas the black circle with the number 2 represents a intermediate stage between a polytomy and bifurcating node (node 2). The node labeled by the gray circle Y may represents the bifurcating node between the Ia clades, or it may represent the more or less direct ancestor of the polytomy node 3. Fig 2. Median-joining network based on the complete G gene sequence from 734 VHSV isolates (transition/transversion bias = 5). The country or region of origin is indicated using a color code. The North Sea (labeled by an asterisk) includes also the Barents Sea, English Channel, Kattegat, Norwegian Sea, Rogaland, and Skagerrak. doi:10.1371/journal.pone.0164475.g002 We found many polytomy nodes, as well as several circle structures within individual Ia clades in the MJ network (Fig 2), and many nodes with weak bootstrap support within individ- ual Ia clades in the ML tree (not shown) (Fig 1). Furthermore, we discovered three polytomies (node 1–3) among the Ia clades. In the MJ network, node 1 and node 3 were clear polytomy nodes, whereas node 2 matched a polytomy rather than bifurcation. Node 1 corresponds to the common origin of all Ia isolates and splits into the lineage of clade 1, the lineage of clades 2–5, and the lineage of clades 6–9. Node 2 originated from node 1 and it splits into the lineage of clade 2, the lineage of clades 3 and 4, and the lineage of clade 5; by contrast, node 3 originated from the bifurcating node Y and splits into the lineage of clade 7, the lineage of clade 8, and the lineage of clade 9. Furthermore, the low ML bootstrap support value of the three nodes BC1–3 and the fact that these nodes were only present in the ML tree and not in the MJ nor NN net- works (Figs 1a, 1b and 2), supported the assumption of a polytomy. The nucleotide diversity (π) of the total European Ia population (number of used sequences: 651) was 0.029 π. The nucleotide diversity of an individual Ia clade ranged from 0.005 (clade 7) to 0.009 (clade 9). By comparison, the nucleotide diversity of the Danish Ia population (num- ber of used sequences: 236) was 0.014 π, and the nucleotide diversity of the German Ia popula- tion (number of used sequences: 296) was 0.024 π. The number of haplotypes that were included within an individual Ia clade varied from two (clade 3) to 158 (clade 9). The mean period of occurrenceof a single Ia haplotype was 1.482 years (standard deviation (σ) = 0.55 years) (S1 Table). A striking exception was one haplotype of clade 6 (number of haplotypes: 140) that included a German isolate from 1990 (V01-90he) and a French isolate from 1971 (FR-0771), which was thus re-isolated 19 years after its first iso- lation. This exceptional case has led us to exclude this haplotype from the calculation of the mean period of occurrenceof the Ia haplotypes. The years represent the range of collection dates for isolates of every European subgenogroup I (I(unclassified), and Ia–e) and Ia clades (clade 1–9). The clades 5, 8, and 9 are additionally subdivided into the subsets 5i–iii, 8i–iii, and 9i–iii. The black circles with the number 1 and 3 represent clear polytomy nodes (node 1 and 3), whereas the black circle with the number 2 represents a intermediate stage between a polytomy and bifurcating node (node 2). The node labeled by the gray circle Y may represents the bifurcating node between the Ia clades, or it may represent the more or less direct ancestor of the polytomy node 3. doi:10.1371/journal.pone.0164475.g002 only taxon in the Ia dataset with varying clade-membership when comparing the ML, MJ, and NN Ia phylogeny. In the ML tree, isolate X clustered with clade 6 (Fig 1b), but with a very low bootstrap support value of 19%. In contrast, in the MJ network, the lineage of X originated from the common ancestor of clade 2 (Fig 2), whereas in the case of the NN network, it origi- nated close to the ancestral node of the Ia lineage (node 1), which was approximately between the origins of the lineage of clade 2 and the lineage of clade 6 (Fig 1d). However, when isolate X is excluded from the dataset, the bootstrap support value of node Y increased from 40% to 96% (Fig 1a and 1b). In addition, the number of splits was significantly lower in the NN network of Fig 1c in comparison to Fig 1d. The Ia phylogeny in the MJ network was largely characterized by a phylotemporal structure: older Ia clades (those that included isolates with earlier collection dates) clustered closer to the phylogenetic origin of the Ia lineage (node 1) than did clades with more recent isolates. For example, clades 1, 2, and 6 include haplotypes from the oldest collected Ia isolates (Fig 2 and S1 Table). Moreover, we further divided the Ia clades 5, 8, and 9 into subsets (clade 5 into sub- set 5i–ii, 8 into subset 8i–iii, and 9 into subset 9i–iii) to better illustrate the phylotemporal 8 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0164475 October 19, 2016 Phylogeny of VHSV in European Aquaculture arrangement of their haplotypes, and in the case of clade 5 and 9, to group those haplotypes that form a star-like pattern (5i and 9i). PLOS ONE \| DOI:10.1371/journal.pone.0164475 October 19, 2016 Phylogeny of VHSV in European Aquaculture this clade contained the highest number of Ia haplotypes that occurredin more than one coun- try (Fig 1 and S1 Table). We found indications for two instances of large-scale rapid population growth during the evolution of the Ia lineage. One of these events took place in the 9i population of clade 9, and the other event occurred in the 5i population of clade 5. This finding was based on two major star-like patterns in the MJ network (Fig 2). During these events, the substitution rate was 7.813 × 10−4 nucleotides per site per year for the 5i haplotypes, and 5.651 × 10-4nucleotides per site per year for the 9i haplotypes. These potential demographic expansions were also indicated through a Bayesian skyline plot and a mismatch distribution via a unimodal curve (Figs 3 and 4). On the basis of the Bayesian skyline plot, this rapid increase in the population size of thlle 5i population occurred between approximately 2000 and 2001, and in the case of the 9i popula- tion, it occurredbetween approximately 2001 and 2003. Moreover, we confirmed a low value for the raggedness index and statistically significant negative values in tests for Tajima’s D and Fu’s Fs for both datasets (Table 1). This was also the case when we examined separately either the Danish 5i haplotypes or the German 9i haplotypes (Figs 3 and 4, Table 1). However, when we calculated the mismatch distribution of clade 9i haplotypes exclusively from Switzerland, Italy, or Poland (Fig 3), and also of the haplotypes from 5ii, 8ii, 8iii, and 9ii (graphs not shown), the data only poorly fit the unimodal curve. Instead of a smooth unimodal shape, these curves exhibit a more ragged shape. Accordingly, the raggedness index was significantly increased and a negative non-significant P value was also obtained in tests for Tajima’s D and Fu’s Fs (Table 1). By contrast, the multimodal curve of haplotypes from clade 2 had no similar- ity to a unimodal curve (graph not shown). In case of clades 1, 3, 4, 6, and 7, and 8iii and 9iii, we did not perform a demographic expansion analysis because of the small number of different haplotypes in those clades. PLOS ONE \| DOI:10.1371/journal.pone.0164475 October 19, 2016 In addition, as this haplotype is a member of clade 6, we have excluded this clade from the calculation of the mean period of occurrenceof the Ia clades. The mean period of occurrenceof a Ia clade was 11.4 years (σ = 6.5 years) (Fig 2 and S1 Table). Notably, German Ia haplotypes from 2005 to 2015 fall into four distinct Ia clades (4, 5, 8, and 9), whereas Danish Ia haplotypes from 1999 to 2009 fall into two Ia clades (5 and 9). By comparison, the mean value was 4.7 Ia clades per calendar year (σ = 1.658 per cal- endar year) for the entire dataset, calculated over the last 20 years (1996–2015). On the basis of the phylogeographic pattern of the Ia population (Fig 2), clades 1, 3, 4, and 6 indicate a trend toward Germany and France, whereas clades 2 and 7 suggest a trend toward Germany and Denmark. In case of clades 8 and 9, the phylogeographic pattern was particularly difficult to interpret, as each of them included isolates from an especially large number of European coun- tries (Fig 2 and S1 Table). While the isolates of clade 8 were from Austria, the Czech Republic, France, Germany, Italy, Poland, and Switzerland, the isolates of clade 9 were from Austria, Denmark, France, Germany, Italy, Poland, Slovenia, and Switzerland. However, both clades were dominated by German isolates. By contrast, although clade 5 also includes several Ger- man isolates and an isolate from the United Kingdom, this clade exhibited a clear trend toward Denmark (Fig 2). Furthermore, 4.02% of the Ia haplotypes appeared in more than one European country (nine haplotypes of clade 9, three of clade 8, one of clade 6, and three of clade 5) (Fig 2 and S1 Table). The most widespread Ia haplotype—haplotype number 161, occurringin Germany, France, Poland, and Switzerland—was located very close to the center of clade 9, which was the most widespread Ia clade with haplotypes isolated from seven European countries: Austria, France, Germany, Poland, Italy, Slovenia, and Switzerland (Fig 2 and S1 Table). In addition, 9 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0164475 October 19, 2016 Discussion Since the first Ia isolate was detected in France in 1971, the Ia population has become widely distributed in Europe, especially where rainbow trout is intensively produced in freshwater aquaculture [9]. This extensive spread is clearly confirmed by our phylogenetic results. First, the Ia lineage underwent a considerable phylogenetic radiation, since we were able to divide this subgenogroup into nine Ia clades (Fig 2). Normally, such radiation is based on different local scenarios of selection and genetic drift resulting from the spread events. In addition, iso- late X may even indicate—as a tenth Ia clade—the existence of further clades which remain, however, undetected using this dataset. Secondly, each of the three polytomies (nodes 1–3) (Fig 2) that were found among the Ia clades could be an indicator for a rapid radiation, as they reflect a simultaneous spread from one single place to several places. For example, in the case of even a single delivery of VHS-infected trout, one single Ia haplotype can be spread to various regions of Europe. However, this is only true, if these polytomies are hard polytomies and not the result of a potential taxon sampling bias in our dataset. Third, the Europe-wide distribution of clades 8 and 9, in particular, indicates a strong mixing of Ia isolates through the trout trade. Clade 8 includes isolates from seven different European countries, and clade 9 from eight (Fig 2 and S1 Table). Finally, the indications for the two large-scale rapid population growth events in the Ia population could also point to, at least temporally, a sudden extensive demographic expansion, which is normally the result of the spatial spread of a species [42, 43]. However, a clear reconstruction of the spread routes of ancestral lineages of individual Ia clades was largely impossible. The reason for that was that most Ia clades include several nodes with weak ML bootstrap support (not shown), a result of the low resolving power of mutational differences among their closely related haplotypes, and apparently due to homoplasy (parallel PLOS ONE \| DOI:10.1371/journal.pone.0164475 October 19, 2016 10 / 18 Phylogeny of VHSV in European Aquaculture Fig 3. Mismatch distributions and Bayesian skyline plot of 9i sequences. Discussion The mismatch distribution is based on the complete G gene sequence and is calculated separately for the following: 213 9i sequences from Austria, France, Germany, Italy, Poland, Switzerland, and Slovenia, collected between 1999 and 2015 145 German 9i sequences from 2002 to 2015; 27 Swiss 9i sequences from 1999 to 2012; 13 Italian 9i sequences from 2002 to 2011; and 18 Polish 9i sequences from 2006 to 2009. Table 1 lists the respective 75 O t b 19 2016 11 / 1 Fig 3. Mismatch distributions and Bayesian skyline plot of 9i sequences. The mismatch distribution is based on the complete G gene sequence and is calculated separately for the following: 213 9i sequences from Austria, France, Germany, Italy, Poland, Switzerland, and Slovenia, collected between 1999 and 2015; 145 German 9i sequences from 2002 to 2015; 27 Swiss 9i sequences from 1999 to 2012; 13 Italian 9i sequences from 2002 to 2011; and 18 Polish 9i sequences from 2006 to 2009. Table 1 lists the respective PLOS ONE \| DOI:10.1371/journal.pone.0164475 October 19, 2016 11 / 18 Phylogeny of VHSV in European Aquaculture Phylogeny of VHSV in European Aquaculture value of the raggedness index (r), nucleotide diversity (PI), Tajima’s D, and Fu’s Fs of each dataset. The Bayesian skyline plot shows changes of the 9i population size between 1999 and 2015. The plot was generated using all 213 9i sequences (complete G gene sequence). X axis: time in years, Y axis: population size. The middle solid line is the median estimate, and the area between the blue lines shows the 95% highest probability density (HPD). value of the raggedness index (r), nucleotide diversity (PI), Tajima’s D, and Fu’s Fs of each dataset. The Bayesian skyline plot shows changes of the 9i population size between 1999 and 2015. The plot was generated using all 213 9i sequences (complete G gene sequence). X axis: time in years, Y axis: population size. The middle solid line is the median estimate, and the area between the blue lines shows the 95% highest probability density (HPD). doi:10 1371/journal pone 0164475 g003 doi:10.1371/journal.pone.0164475.g003 or back mutation events) [33]. This poor phylogenetic resolution was also apparent in the MJ network through the many polytomy structures, and several circle structures within the Ia clades (Fig 2). Furthermore, this reconstruction was additionally complicated by an almost complete absence of a country-specificdistribution for most of the Ia clades (Fig 2). Discussion As men- tioned above, it seems that the Ia isolates were permanently and complexly changed due to the Europe-wide mixing of Ia isolates via the trout trade. And finally, the reconstruction of the Fig 4. Mismatch distributions and Bayesian skyline plot of 5i sequences. The mismatch distribution is based on the complete G gene sequence and is calculated separately for the following: 152 5i sequences from Denmark, Germany, and the United Kingdom, collected between 2000 and 2011; and 137 Danish 5i sequences from 2000 to 2009. Table 1 lists the respective values of the raggedness index (r), nucleotide diversity (PI), Tajima’s D, and Fu’s Fs of each dataset. The Bayesian skyline plot shows changes of the 5i population size between 2000 and 2009. The plot was generated using 137 Danish 5i sequences (complete G gene sequence). X axis: time in years, Y axis: population size. The middle solid line is the median estimate, and the area between the blue lines shows the 95% highest probability density (HPD). doi:10 1371/journal pone 0164475 g004 match distributions and Bayesian skyline plot of 5i sequences. The mismatch distribution is based on e G gene sequence and is calculated separately for the following: 152 5i sequences from Denmark Fig 4. Mismatch distributions and Bayesian skyline plot of 5i sequences. The mismatch distribution is based o Fig 4. Mismatch distributions and Bayesian skyline plot of 5i sequences. The mismatch distribution is based on the complete G gene sequence and is calculated separately for the following: 152 5i sequences from Denmark, Germany, and the United Kingdom, collected between 2000 and 2011; and 137 Danish 5i sequences from 2000 to 2009. Table 1 lists the respective values of the raggedness index (r), nucleotide diversity (PI), Tajima’s D, and Fu’s Fs of each dataset. The Bayesian skyline plot shows changes of the 5i population size between 2000 and 2009. The plot was generated using 137 Danish 5i sequences (complete G gene sequence). X axis: time in years, Y axis: population size. The middle solid line is the median estimate, and the area between the blue lines shows the 95% highest probability density (HPD). Fig 4. Mismatch distributions and Bayesian skyline plot of 5i sequences. doi:10.1371/journal.pone.0164475.g004 (1) Haplotypes of the dominant Danish clade (clade 5)—the clade with the largest number of different Danish haplotypes— were almost exclusively detected in Denmark, whereas many haplotypes of clade 9—the clade with the largest number of different German haplotypes—were also found outside Germany in several other European countries (Fig 1 and S2 Table). In addition, we found no isolate of clade 9 i D k th h thi th t id d l d i E ( i i i Table 1. Parameters of the analysis of a rapid population growth based on the complete G gene sequence data. Clade and subset membership, as well as sampling period Number of used sequences (n) Nucleotide diversity (PI) Tajima’s D Fu’s Fs Raggedness index (r) 5i (2000–2011) 152 0.00349 -2.36505** -73.024 0.0051 5i Denmark (2000–2009) 137 0.00333 -2.16059* -63.686 0.0057 9i (1999–2015) 213 0.00583 -2.48568* -149.539 0.0030 9i Germany (2002–2015) 145 0.00584 -2.10903 -72.958 0.0042 9i Italy (2002–2011) 13 0.00407 -1.25092˚ -8.124 0.0155 9i Poland (2006–2009) 18 0.00354 -1.76634˚˚ -4.204 0.0120 9i Switzerland (1999–2012) 27 0.00609 -0.96486˚ -2.551 0.0116 P<0.05 P<0.01 P<0.001 ˚P>0.10 (not signiﬁcant) ˚˚0.10>P>0.05 (not signiﬁcant) doi:10.1371/journal.pone.0164475.t001 Table 1. Parameters of the analysis of a rapid population growth based on the complete G gene sequence data. Clade and subset membership, as well as sampling period Number of used sequences (n) Nucleotide diversity (PI) Tajima’s D Fu’s Fs Raggedness index (r) 5i (2000–2011) 152 0.00349 -2.36505 -73.024 0.0051 5i Denmark (2000–2009) 137 0.00333 -2.16059* -63.686 0.0057 9i (1999–2015) 213 0.00583 -2.48568* -149.539 0.0030 9i Germany (2002–2015) 145 0.00584 -2.10903 -72.958 0.0042 9i Italy (2002–2011) 13 0.00407 -1.25092˚ -8.124 0.0155 9i Poland (2006–2009) 18 0.00354 -1.76634˚˚ -4.204 0.0120 9i Switzerland (1999–2012) 27 0.00609 -0.96486˚ -2.551 0.0116 of a rapid population growth based on the complete G gene sequence data. . Parameters of the analysis of a rapid population growth based on the complete G gene sequence data. spread routes was further complicated, by a sampling bias in our dataset. Although an exten- sive dataset of Ia isolates was used here, spatiotemporal sampling bias may exist in the case of Ia isolates obtained from Austria, the Czech Republic, France, Georgia, Italy, Poland, Slovenia, Switzerland, and the United Kingdom, and from Germany and Denmark before 1996 (S1 Fig). This is possible, as outbreak samples from European trout farms (let alone from farms with latent infections) were collected using different levels of strictness depending on individual national surveillanceof VHS. Therefore, we believe that this study may represent an incom- plete pan-European picture of the country-specificdistribution for most Ia clades. This cir- cumstance has also contributed to our decision to largely avoid the reconstruction of the spread routes for the ancestral lineages of these Ia clades. For example, we consciously decided against speculating on the place of origin of the Ia lineage. This place might be obscured by an extensive Europe-wide spread of the Ia lineage shortly after its origin. On the one hand, this could explain the polytomy structure of node 1, provided that it reflects a hard polytomy(Figs 1c and 2). But in view of the above mentioned potential sampling bias, this polytomy cannot be seen as an argument for a real hard polytomy. On the other hand, the isolates that cluster clos- est to this node 1 (clades 1, 2, and 6) were from different European countries (Denmark, Ger- many, and France) (Fig 2). In addition, the determination of the place of origin is further complicated because our dataset exhibits a spatiotemporal sampling bias of Ia isolates from the period when the first Ia isolates were collected (S1 Fig). Due to the particularly high spatiotemporal density of Danish and German Ia isolates in our dataset from approximately the last 20 years, we particularly focused our phylogenetic analysis on those two countries during this period. Thus, in comparison to the German situation, we observedan instance of rare gene flow of Ia isolates between Denmark and other European countries. This was supported by the following three findings. (1) Haplotypes of the dominant Danish clade (clade 5)—the clade with the largest number of different Danish haplotypes— were almost exclusively detected in Denmark, whereas many haplotypes of clade 9—the clade with the largest number of different German haplotypes—were also found outside Germany in several other European countries (Fig 1 and S2 Table). In addition, we found no isolate of clade 9 in Denmark, even though this was the most widespread clade in Europe (occurringin nine different European countries) (Fig 1 and S1 Table). Discussion The mismatch distribution is based on the complete G gene sequence and is calculated separately for the following: 152 5i sequences from Denmark, Germany, and the United Kingdom, collected between 2000 and 2011; and 137 Danish 5i sequences from 2000 to 2009. Table 1 lists the respective values of the raggedness index (r), nucleotide diversity (PI), Tajima’s D, and Fu’s Fs of each dataset. The Bayesian skyline plot shows changes of the 5i population size between 2000 and 2009. The plot was generated using 137 Danish 5i sequences (complete G gene sequence). X axis: time in years, Y axis: population size. The middle solid line is the median estimate, and the area between the blue lines shows the 95% highest probability density (HPD). 12 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0164475 October 19, 2016 PLOS ONE \| DOI:10.1371/journal.pone.0164475 October 19, 2016 Phylogeny of VHSV in European Aquaculture spread routes was further complicated, by a sampling bias in our dataset. Although an exten- sive dataset of Ia isolates was used here, spatiotemporal sampling bias may exist in the case of Ia isolates obtained from Austria, the Czech Republic, France, Georgia, Italy, Poland, Slovenia, Switzerland, and the United Kingdom, and from Germany and Denmark before 1996 (S1 Fig). This is possible, as outbreak samples from European trout farms (let alone from farms with latent infections) were collected using different levels of strictness depending on individual national surveillanceof VHS. Therefore, we believe that this study may represent an incom- plete pan-European picture of the country-specificdistribution for most Ia clades. This cir- cumstance has also contributed to our decision to largely avoid the reconstruction of the spread routes for the ancestral lineages of these Ia clades. For example, we consciously decided against speculating on the place of origin of the Ia lineage. This place might be obscured by an extensive Europe-wide spread of the Ia lineage shortly after its origin. On the one hand, this could explain the polytomy structure of node 1, provided that it reflects a hard polytomy(Figs 1c and 2). But in view of the above mentioned potential sampling bias, this polytomy cannot be seen as an argument for a real hard polytomy. On the other hand, the isolates that cluster clos- est to this node 1 (clades 1, 2, and 6) were from different European countries (Denmark, Ger- many, and France) (Fig 2). In addition, the determination of the place of origin is further complicated because our dataset exhibits a spatiotemporal sampling bias of Ia isolates from the period when the first Ia isolates were collected (S1 Fig). Due to the particularly high spatiotemporal density of Danish and German Ia isolates in our dataset from approximately the last 20 years, we particularly focused our phylogenetic analysis on those two countries during this period. Thus, in comparison to the German situation, we observedan instance of rare gene flow of Ia isolates between Denmark and other European countries. This was supported by the following three findings. doi:10.1371/journal.pone.0164475.t001 In contrast, the clear dominance of Danish isolates in clade 5 (Fig 2 and S1 Table) strongly indicates a Danish origin of this clade, and therefore it can be assumed that the ancestral lineage of the German and British isolates of 13 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0164475 October 19, 2016 Phylogeny of VHSV in European Aquaculture clade 5 was introduced from Denmark. This conclusion also agrees with the findings of Kahns and his colleagues [22]. (2) Danish isolates fall into only four of the nine Ia clades detected, whereas German isolates fall into all nine (Fig 2 and S1 Table). Accordingly, the nucleotide diversity of the German Ia population (304 used sequences) was almost more than twice as high as the nucleotide diversity of the Danish Ia population (244 used sequences) (not shown). (3) Danish haplotypes that were also detected outside Denmark were only found in Germany, whereas German haplotypes that were detected outside Germany were also found in several other European countries (Austria, Denmark, France, Italy, the Netherlands, Poland, and Swit- zerland) (Fig 2 and S1 Table). This relatively rare cross-border exchange of Ia isolates between Denmark and other Euro- pean countries very likely reflects the strict control of export and import of rainbow trout in Denmark. As Danish fish farms produce fish mainly under tight economic pressure and their production volumes are mostly large, the potential economic impact of a VHS outbreak is sub- stantial. Therefore, since the 1960s, several extensive sanitation and eradication programs have been implemented to control VHS, and since the spring of 2009, no VHS has been detected in any Danish fish farm [22]. In contrast to Danish trout farming conditions, approximately 95% of all German trout farms are very small (annual production volume of <5 metric tons; Survey and Diagnosis 2014, http://www.eurl-fish.eu/Activities/survey_and_diagnosis),and their trout production is intended more for the local market or even for self-sufficiency. The smaller eco- nomic weight of rainbow trout production in Germany could be one reason for this less strict control. Given that the rapid population growth of the 5i population took place exclusively in Den- mark between approximately 2000 and 2001 (Fig 4) (provided that our analysis reflects a demographic expansion and not a selective sweep), it therefore occurredduring the final period of the Danish VHS-eradication program [22]. PLOS ONE \| DOI:10.1371/journal.pone.0164475 October 19, 2016 This was surprising as such an event can point, under certain circumstances, to a biosecurity gap; for example, when dealing with practices of aquaculture or fish trade for farming and restocking purposes. The underlying idea is that care- less handling of VHSV-infected fish may suddenly result in an increased spread rate of the virus, which is normally the basis for a sudden demographic expansion event [45]. However, such an increased spread rate can also be the result of either a sudden non-human-mediated change in the natural environment (e.g., changes in the climate), or a successful evolutionary adaptation towards increased fitness that may have acted within these circulating VHSV infec- tions. Seen in this light, we believe that the causative factor of this Danish demographic expan- sion event may be of particular interest in follow-up studies. In the case of the potential rapid population growth of the 9i population, we think that Ger- many was a crucial place for its emergence, at least from approximately 2001–2003 (Fig 3). Nevertheless, we found 9i isolates in eight further European countries (Fig 1). Accordingly, their participation in the population growth was not excludable, because of the mentioned sampling bias of Ia isolates in our dataset. In particular, the role of Italy, Poland, and Switzer- land was at issue, as their jagged mismatch curves still exhibited a discernible unimodal shape that may indicate their participation (Fig 3). In the case of Switzerland, however, we know that the trout farming industry did not sell life trout abroad during this period;instead, they bought many life trout from surrounding countries. We only exclude Danish participation, as our dataset comprised no Danish 9i isolates, even though our dataset has a high spatiotemporal density of Danish isolates from that period. At first sight, this Europe-wide spread and mixing of Ia isolates seems to be surprisingly inconsistent with another outcome of this study, namely that only about 4% of the Ia haplo- types occurredin more than one European country (S1 Table). However, we think that this can most probably be explained by the observedhigh substitution rate of VHSV [13, 17]. Our 14 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0164475 October 19, 2016 Phylogeny of VHSV in European Aquaculture dataset indicates that the mean period of occurrence of a single Ia haplotype is typically not longer than one calendar year. Furthermore, distinct isolates from the same VHS outbreak occasionally fall into different but very closely related haplotypes (e.g., isolates V43-14bb, V44- 14bb, and V45-14bb) (S1 Table). However, the observation that one haplotype in our dataset (number 140) was re-isolated (as isolate V01-90he) 19 years after its first isolation (as isolate FR-0771) can be best explained by laboratory contamination or a confusion event. Our dataset suggests that Ia clades become extinct after a period, since the mean period of occurrenceof a clade is approximately 11 years (Fig 2). We consider that this cyclic pattern of newly arising formations and the disappearance of clades have been mainly caused by anthro- pogenic influences such as trout farmers and traders, since the host of the Ia population has mainly been the European farmed rainbow trout. Approximately 95% of the Ia isolates of our dataset are from farmed rainbow trout. This close ecologicalinteraction is supported by the fact that Ia isolates have, in contrast to isolates from other subgenogroups of the genogroup I (Ib–Ie) and the genogroups II–IV, a particularly high virulencefor farmed rainbow trout [12, 46]. We believe that careless breeding and trade practices when dealing with VHS-infected fish have unintentionally supported the geographical spread of the virus and the establishment of new clades, while sanitation and eradication measures have led to their extinction. We also largely exclude a natural host’s adaptation toward resistance against the virus. This is because farmed animal stocks frequently suffer from weaker resistance against infectious disease and often have a lower potential for evolutionary resistance adaptation in comparison to their wild ancestors. Furthermore, housing and trading conditions make farmed stocks particularly vul- nerable to the spread of pathogens. The main reasons for this are: (1) the loss of variation at resistance loci because of founder effects that result from domestication and intentional breed- ing events [47–49]; (2) a decreased chance of successfully adapting towards disease resistance due to artificial mating systems and a consequent relaxation of natural selection [50, 51]; (3) the danger of pleiotropic effects (negative side effects) from intentional breeding that can nega- tively influence disease resistance [52]; (4) factory farming practices that can lead to a stress- induced weak immune system due to artificial housing conditions and high population density [50, 53]; and last but not least, (5) a high population density that can additionally increase virus transmissibility [54]. Supporting Information S1 Fig. Chronologyof the geographic collectionof 422 Ia isolates. (DOCX) S1 Fig. Chronologyof the geographic collectionof 422 Ia isolates. (DOCX) S1 Table. Data on the 717 VHSV isolates used in this study. The samples highlighted in white are the extended samples obtained from GenBank. The table includes the NCBI acces- sion number, Fispathogens.eu database (EU-FP) number, name of isolate, date of collection, site of collection, host species (wild fish; feral fish; no asterisk, characterized as a farmed fish), and phylogenetic classification (Ia: haplotype-subgenogroup-clade; I(unclassified), Ib, Ic, Id, and Ie: haplotype-subgenogroup; and II, III, IV: genogroup). Superscripts in parentheses mark corrections with respect to the original protocol in GenBank of the respective accession number: 1) corrections to the name of host species; 2) to the year of isolation; 3) to the name of isolate. S2 Table. Primer sequences. (DOCX) 15 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0164475 October 19, 2016 Phylogeny of VHSV in European Aquaculture Writing – review& editing: MC SSM HFS OLMH VP TW HS. Writing – review& editing: MC SSM HFS OLMH VP TW HS. Acknowledgments Special thanks to Peter-Joachim Enzmann for his substantial work regarding the establishment of the German VHSV isolate database. Furthermore, the skillful technical assistance of René Grenkowitz, Günter Strebelow, Kersten Biebl and Troels Secher Rundqvistis is highly appreci- ated. We also would like to thank all colleagues from the German health servicesand regional laboratories for providing samples and background information, as well as Henrik Korsholm, the Danish Food and Veterinary Administration and Morten Sichlau Bruun, DTU Vet for information and encouragement. Special thanks also to Eva Schellenbeckfor editing this manuscript. Conceptualization: HS. Data curation: MC SSM HFS HS. Formal analysis: MC. Funding acquisition: HS. Methodology:MC. Project administration: HS. Resources: SSM HFS MB ND MYE OLMH SM VP TW NJO HS. Visualization: MC. Writing – original draft: MC. Writing – review& editing: MC SSM HFS OLMH VP TW HS. Funding acquisition: HS. Methodology:MC. PLOS ONE \| DOI:10.1371/journal.pone.0164475 October 19, 2016 Resources: SSM HFS MB ND MYE OLMH SM VP TW NJO HS. Writing – original draft: MC. References 1. FEAP. European Aquaculture Production Report 2005–2014. 2015. Available: http://www.feap.info/ default.asp?SHORTCUT=582 2. IFB-Potsdam. Jahresbericht zur Deutschen Binnenfischerei und Binnenaquakultur 2014. 2015. Avail- able: http://www.ifb-potsdam.de/de-de/ver%C3%B6ffentlichungen/downloads.aspx 3. OIE. Diseases listed by the OIE. In: Aquatic Animal Health Code 2015 chapter 10.10. Available: http:// www.oie.int/index.php?id=171&L=0&htmfile=chapitre_diseases_listed.htm 4. European Union. Council Directive 2006/88/EC of 24 October 2006 on animal health requirements for aquaculture animals and products thereof, and on the prevention and control of certain diseases in aquatic animals. OJEC. 2006; L 328: 14–56. 5. Walker PJ, Benmansour A, Dietzgen R, Fang R-X, Jackson AO, Kurath G, et al. Family Rhabdoviridae. In: Virus Taxonomy. Classification and Nomenclature of Viruses. Van Regenmortel MHV, Fauquet CM, Bishop DHL, Carstens EB, Estes MK, Lemon SM, et al. editors. Academic Press, San Diego, CA. 2000; 563–583. 6. Schu¨tze H, Mundt E, Mettenleiter TC. Complete genomic sequence of viral hemorrhagic septicemia virus, a fish rhabdovirus. Virus Genes. 1999; 19(1): 59–65. PMID: 10499451 7. Betts AM, Stone DM. Nucleotide sequence analysis of the entire coding regions of virulent and aviru- lent strains of viral haemorrhagic septicemia virus. Virus Genes. 2000; 20(3): 259–62. PMID: 10949954 8. OIE. Viral haemorrhagic septicaemia. In: Manual of Diagnostic Tests for Aquatic Animals 2015 chapter 2.3.9. Available:http://www.oie.int/international-standard-setting/aquatic-manual/access-online/ 9. Skall HF, Olesen NJ, Mellergaard S. Viral haemorrhagic septicaemia virus in marine fish and its impli- cations for fish farming—a review. J Fish Dis. 2005; 28(9): 509–29. doi: 10.1111/j.1365-2761.2005. 00654.x PMID: 16266325 16 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0164475 October 19, 2016 Phylogeny of VHSV in European Aquaculture 10. Scha¨perclaus W. Die Scha¨digungen der deutschen Fischerei durch Fischparasiten und Fischkrankhei- ten. Allg Fischztg. 1938; 41: 256–59, 267–70. 11. Jensen MH. Research on the virus of Egtved disease. Ann N Y Acad Sci. 1965; 126: 422–426. PMID: 4957057 12. Einer-Jensen K, Ahrens P, Forsberg R, Lorenzen N. Evolution of the fish rhabdovirus viral haemorrha- gic septicaemia virus. J Gen Virol. 2004; 85: 1167–1179. doi: 10.1099/vir.0.79820-0 PMID: 15105533 13. He M, Yan XC, Liang Y, Sun XW, Teng CB. Evolution of the viral hemorrhagic septicemia virus: diver- gence, selection and origin. Mol Phylogenet Evol. 2014; 77: 34–40. doi: 10.1016/j.ympev.2014.04.002 PMID: 24727199 14. Winton JR, Batts WN, Nishizawa T, Stehr CM. Characterization of the first North American isolates of viral hemorrhagic septicemia virus. Newsl Am Fish Soc Fish Health Sec.1989; 17: 2–3. 15. Isshiki T, Nishizawa T, Kobayashi T, Nagano T. Miyazaki T. References An outbreak of VHSV (viral hemorrhagic septicemia virus) infection in farmed Japanese flounder Paralichthysolivaceus in Japan. Dis Aquat Organ. 2001; 47(2): 87–99. doi: 10.3354/dao047087 PMID: 11775799 16. Snow M, Bain N, Black J, Taupin V, Cunningham CO, King JA, et al. Genetic population structure of marine viral haemorrhagic septicaemia virus (VHSV). Dis Aquat Org. 2004; 61(1–2): 11–21. doi: 10. 3354/dao061011 PMID: 15584406 17. Pierce LR, Stepien CA. Evolution and biogeography of an emerging quasispecies: diversity patterns of the fish viral hemorrhagic septicaemia virus (VHSv). Mol Phylogenet Evol. 2012; 63(2): 327–41. doi: 10.1016/j.ympev.2011.12.024 PMID: 22266219 18. Ross K, McCarthy U, Huntly PJ, Wood BP, Stuart D, Rough EI, et al. An outbreak of viral haemorrhagic septicaemia (VHS) in turbot (Scophthalmus maximus) in Scotland. Bull Eur Assoc Fish Pathol 1994; 14: 213–214. 19. Schlotfeldt HJ, Ahne W, Jørgensen PEV, Glende W. Occurrence of viral haemorrhagic septicaemia in turbot (Scophthalmus maximus)–a natural outbreak. Bull Eur Assoc Fish Pathol.1991; 11: 105–107. 20. deKinkelin P, Le Berre M. Isolement d’un Rhabdovirus pathogene de la Truite Fario (Salmo trutta, L. 1766). C R Acad Sci Paris.1977; 284D: 101104. 21. Schlotfeldt HJ, Ahne W. Epizootics in brown trout (Salmo trutta fario) caused by VHSV-F1. J Appl Ichthyol.1988; 4(3): 147–48. 22. Kahns S, Skall HF, Kaas RS, Korsholm H, Bang Jensen B, Jonstrup SP, et al. European freshwater VHS–virus genotype Ia isolates divide into two distinct subpopulations. Dis Aquat Organ. 2012; 99: 23–35. doi: 10.3354/dao02444 PMID: 22585300 23. Volz EM, Koelle K, Bedford T. Viral phylodynamics. PLoS Comput Biol. 2013; 9(3): e1002947. doi: 10. 1371/journal.pcbi.1002947 PMID: 23555203 24. European Community (2001) Commission Decision 2001/183/EC–Laying down the sampling plans and diagnostic methods for the detection and confirmation of certain fish diseases and repealing Deci- sion 92/532/EEC, Brussels (OJ L 67, 9.3.2001, p. 65). Available: http://eur-lex.europa.eu/legal- content/en/ALL/?uri=CELEX:32001D0183 25. Schu¨tze H, Enzmann PJ, Mundt E, Mettenleiter TC. Identification of the nonvirion (NV) protein of fish rhabdoviruses viral haemorrhagic septicaemia virus and infectious haematopoietic necrosis virus. J Gen Virol. 1996; 77: 1259–63. doi: 10.1099/0022-1317-77-6-1259 PMID: 8683214 26. Kearse M, Moir R, Wilson A, Stones-Havas S, Cheung M, Sturrock S, et al. Geneious basic: an inte- grated and extendable desktop software platform for the organization and analysis of sequence data. Bioinformatics. 2012; 28(12): 1647–9. doi: 10.1093/bioinformatics/bts199 PMID: 22543367 27. Jonstrup SP, Gray T, Kahns S, Skall HF, Snow M, Olesen NJ. FishPathogens.eu/vhsv: a user-friendly viral haemorrhagic septicaemia virus isolate and sequence database. PLOS ONE \| DOI:10.1371/journal.pone.0164475 October 19, 2016 References J Fish Dis. 2009; 32(11): 925– 29. doi: 10.1111/j.1365-2761.2009.01073.x PMID: 19538460 28. Librado P, Rozas J. DnaSP v5: A software for comprehensive analysis of DNA polymorphism data. Bioinformatics. 2009; 25(11): 1451–2. doi: 10.1093/bioinformatics/btp187 PMID: 19346325 29. Tamura K, Peterson D, Peterson N, Stecher G, Nei M, Kumar S. MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and Maximum Parsimony meth- ods. Mol Biol Evol. 2011; 28(10): 2731–9. doi: 10.1093/molbev/msr121 PMID: 21546353 30. Kuhn TS, Mooers AO, Thomas GH. A simple polytomy resolver for dated phylogenies. Methods Ecol Evol. 2011; 2: 1–10. 31. Lin GN, Zhang C, Xu D. Polytomy identification in microbial phylogenetic reconstruction. BMC Syst Biol. 2011; 5(Suppl 3): S2. doi: 10.1186/1752-0509-5-S3-S2 PMID: 22784621 32. Posada D, and Crandall K. Intraspecific gene genealogies: trees grafting into networks. Trends Ecol Evol 2001; 16: 37–45. PMID: 11146143 17 / 18 PLOS ONE \| DOI:10.1371/journal.pone.0164475 October 19, 2016 Phylogeny of VHSV in European Aquaculture 33. Bandelt HJ, Forster P, Rohl A. Median-joining networks for inferring intraspecific phylogenies. Mol Biol Evol. 1999; 16(1): 37–48. doi: 10.1093/oxfordjournals.molbev.a026036 PMID: 10331250 34. Bryant D and Moulton V. NeighborNet: an agglomerative algorithm for the construction of planar phylo- genetic networks. Mol Biol Evol 2004; 21: 255–265. doi: 10.1093/molbev/msh018 PMID: 14660700 35. Huson DH and Bryant D. Application of Phylogenetic Networks in Evolutionary Studies. Mol Biol Evol 2006; 23(2): 254–267. doi: 10.1093/molbev/msj030 PMID: 16221896 36. Swenson NG. Functional and phylogenetic ecology in R. Springer, New York; 2014. 37. Drummond AJ, Nocholls GK, Rodrigo AG, Solomon W. Estimating mutation parameters, population history and genealogy simultaneously from temporally spaced sequence data. Genetics. 2002; 161: 1307–1320. PMID: 12136032 38. Drummond AJ, Rambaut A, Shapiro B, Pybus O: Bayesian coalescent inference of past population dynamics from molecular sequences. Mol Biol Evol. 2005; 22: 1185–1192. doi: 10.1093/molbev/ msi103 PMID: 15703244 39. Drummond AJ, Rambaut A. BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evol Biol. 2007; 7: 214. doi: 10.1186/1471-2148-7-214 PMID: 17996036 40. Tajima F. Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genet- ics. 1989; 123(3): 585–95. PMID: 2513255 41. Fu YX. Statistical tests of neutrality of mutations against population growth, hitchhiking and back- ground selection. Genetics. 1997; 147(2): 915–25. PMID: 9335623 42. Harpending HC, Batzer MA, Gurven M, Jorde LB, Rogers AR, Sherry ST. Genetic traces of ancient demography. PNAS. 1998; 95(4): 1961–1967. PMID: 9465125 43. Schneider S and Excoffier L. PLOS ONE \| DOI:10.1371/journal.pone.0164475 October 19, 2016 References Estimation of past demographic parameters from the distribution of pair- wise differences when the mutation rates vary among sites: Application to human mitochondrial dna. Genetics. 1999; 152(3): 1079–1089. PMID: 10388826 44. Hahn MW, Rausher MD, Cunningham CW. Distinguishing between selection and population expan- sion in an experimental lineage of bacteriophage T7. Genetics 2002; 161: 11–20. PMID: 12019219 45. Real LA, Biek R. Spatial dynamics and genetics of infectious diseases on heterogeneous landscapes. J. R. Soc. Interface. 2007: 4: 935–948. doi: 10.1098/rsif.2007.1041 PMID: 17490941 46. Skall HF, Slierendrecht WJ, King JA, Olesen NJ. Experimental infection of rainbow trout Oncor- hynchus mykiss with viral haemorrhagic septicaemia virus isolates from European marine and farmed fishes. Dis Aquat Organ. 2004; 58: 99–110. doi: 10.3354/dao058099 PMID: 15109131 47. Spielman D, Brook BW, Briscoe DA, Frankham R. Does inbreeding and loss of genetic diversity reduce disease resistance? Conserv Genet. 2004; 5: 439–448. 48. Wang L, Shi X, Su Y, Meng Z, Lin H. Loss of genetic diversity in the cultured stocks of the large yellow croaker, Larimichthyscrocea, revealed by microsatellites. Int J Mol Sci. 2012; 13: 5584–97. doi: 10. 3390/ijms13055584 PMID: 22754317 49. Dunham RA. Transgenic fish resistant to infectious diseases, their risk and prevention of escape into the environment and future candidate genes for disease transgene manipulation. Comp ImmunolMi- crobiol Infect Dis. 2009; 32: 139–61. 50. Frankham R. Stress and adaptation in conservation genetics. J Evol Biol. 2005; 18: 750–755 doi: 10. 1111/j.1420-9101.2005.00885.x PMID: 16033545 51. Grant W. Understanding the adaptive consequences of hatchery–wild interactions in Alaska salmon. Environ Biol Fish. 2012; 94: 325–342. doi: 10.1007/s10641-011-9929-5 52. Gjedrem T. Selection and Breeding Programs in Aquaculture. Dordrecht, The Netherlands: Springer; 2005. 53. Tort L. Stress and immune modulation in fish. Dev CompImmunol. 2011; 35: 1366–75. 54. Lafferty KD, Harvell CD, Conrad JM, Friedman CS, Kent ML, Kuris AM, Powell EN, Rondeau D, Sak- sida SM. Infectious Diseases Affect Marine Fisheries and Aquaculture Economics. Annu Rev Mar Sci. 2015; 7: 471–96. 18 / 18
https://openalex.org/W3025868205	https://europepmc.org/articles/pmc7243933?pdf=render	English	null	Should Root Plasticity Be a Crop Breeding Target?	Frontiers in plant science	2,020	cc-by	15,242	Edited by: Idupulapati Madhusudana Rao, International Center for Tropical Agriculture (CIAT), Colombia Idupulapati Madhusudana Rao, International Center for Tropical Agriculture (CIAT), Colombia Reviewed by: Hillel Fromm, Tel Aviv University, Israel Philip Benfey, Duke University, United States Correspondence: Jonathan P. Lynch jpl4@psu.edu Reviewed by: Hillel Fromm, Tel Aviv University, Israel Philip Benfey, Should Root Plasticity Be a Crop Breeding Target? Hannah M. Schneider and Jonathan P. Lynch Department of Plant Science, The Pennsylvania State University, University Park, PA, United States Hannah M. Schneider and Jonathan P. Lynch* Root phenotypic plasticity has been proposed as a target for the development of more productive crops in variable environments. However, the plasticity of root anatomical and architectural responses to environmental cues is highly complex, and the consequences of these responses for plant ﬁtness are poorly understood. We propose that root phenotypic plasticity may be beneﬁcial in natural or low-input systems in which the availability of soil resources is spatiotemporally dynamic. Crop ancestors and landraces were selected with multiple stresses, competition, signiﬁcant root loss and heterogenous resource distribution which favored plasticity in response to resource availability. However, in high-input agroecosystems, the value of phenotypic plasticity is unclear, since human management has removed many of these constraints to root function. Further research is needed to understand the ﬁtness landscape of plastic responses including understanding the value of plasticity in different environments, environmental signals that induce plastic responses, and the genetic architecture of plasticity before it is widely adopted in breeding programs. Phenotypic plasticity has many potential ecological, and physiological beneﬁts, but its costs and adaptive value in high-input agricultural systems is poorly understood and merits further research. REVIEW published: 15 May 2020 doi: 10.3389/fpls.2020.00546 published: 15 May 2020 doi: 10.3389/fpls.2020.00546 Keywords: anatomy, architecture, breeding, crop, ideotype, plasticity, root Specialty section: Specialty section: This article was submitted to Plant Abiotic Stress, a section of the journal Frontiers in Plant Science Received: 11 February 2020 Accepted: 09 April 2020 Published: 15 May 2020 Correspondence: Correspondence: Jonathan P. Lynch jpl4@psu.edu Unpredictable growth environments, decreasing freshwater availability, altered precipitation patterns, ongoing soil degradation, and the rising cost of nitrogen and phosphorus fertilizer demand the development of crop varieties that are resilient to abiotic stress (Tebaldi and Lobell, 2008; Brisson et al., 2010; Woods et al., 2010; Sandhu et al., 2016; Lynch, 2019). Root phenotypic plasticity is a widespread and important phenomenon for the optimized capture of edaphic resources. An array of biotic and abiotic constraints limit plant productivity, and phenotypic plasticity is an important phenomenon to enable plants to adapt to spatiotemporal changes in their environment. In this article we consider the beneﬁts and tradeoﬀs of root phenotypic plasticity in the development of more productive annual agricultural crops. Many studies of phenotypic plasticity measure the plastic response of allometric traits (length, volume, or biomass), which display plasticity, but may not be adaptive, as they merely reﬂect growth itself. Many ecological studies of phenotypic plasticity focus on comparisons of distinct species, which is not as relevant to crop improvement as comparisons of genotypes within a species. We will not attempt to provide a comprehensive review of a large and disparate literature, much of which only has tangential relevance to annual crops, but instead focus on opportunities and costs of plasticity for root anatomical and architectural phenotypes in agroecosystems. INTRODUCTION Duke University, United States FIGURE 1 \| Schematic diagram of plastic responses. In (A) the phene value does not change across environments, however, phene expression varies between genotypes. In (B) the phene value changes across environments but the reaction norm runs parallel because the response to the environment is the same for both genotypes. In (C) one genotype does not exhibit plasticity for a speciﬁc phene, while another genotype demonstrates signiﬁcant environmental plasticity. In (D) the reaction norms cross because there is a strong plastic phenotypic response to different environments for both genotypes. The classic paradigm is that a phenotype (P) is the product of genetics or intrinsic developmental processes (G), environment (E), and the interaction between genetics and the environment (G × E) (Sultan, 2000). Phenotypic plasticity is the ability of an organism to alter its phenotype in response to the environment and may involve changes in physiology, morphology, anatomy, development, or resource allocation (Figure 1; Sultan, 2000). Plasticity is not a characteristic of an organism as a whole, but rather is a characteristic of a given phene (“phene” is to “phenotype” as “gene” is to “genotype”) (Lynch, 2011; Pieruschka and Poorter, 2012; York et al., 2013) in response to a given environment. A phene state is the outcome of complex synergistic developmental systems, inﬂuenced by many genes and gene products, as well as the environment (Miklos and Rubin, 1996; Trewavas and Malho, 1997). Plastic responses can aﬀect the ﬁtness of a genotype and be a response to physical, chemical, and biological processes or resource limitations (Weiner, 2004). The phenotypic spectrum, or an array of possible phenotypes a single genotype can display in a single environment, illustrates that many factors inﬂuence the expression of a phenotype. For example, the eﬀects of roots of neighboring plants and priority eﬀects determined by germination time may have large eﬀects on the expression of a phenotype in a single environment (Xie et al., 2019). Phenotypic plasticity may include components of genotype by environment interaction, adaptation, and acclimation. Biologists have long been aware of plasticity (which is one reason that many experiments are performed in controlled environmental conditions), and for much of the past century phenotypic plasticity has been regarded as “noise” and was thought to obstruct the true or native phenotype of an organism. Frontiers in Plant Science \| www.frontiersin.org Citation: Schneider HM and Lynch JP (2020) Should Root Plasticity Be a Crop Breeding Target? Front. Plant Sci. 11:546. doi: 10.3389/fpls.2020.00546 Schneider HM and Lynch JP (2020) Should Root Plasticity Be a Crop Breeding Target? Front. Plant Sci. 11:546. doi: 10.3389/fpls.2020.00546 May 2020 \| Volume 11 \| Article 546 1 Frontiers in Plant Science \| www.frontiersin.org Plasticity as a Breeding Target Schneider and Lynch FIGURE 1 \| Schematic diagram of plastic responses. In (A) the phene value does not change across environments, however, phene expression varies between genotypes. In (B) the phene value changes across environments but the reaction norm runs parallel because the response to the environment is the same for both genotypes. In (C) one genotype does not exhibit plasticity for a speciﬁc phene, while another genotype demonstrates signiﬁcant environmental plasticity. In (D) the ROOT PHENES ARE IMPORTANT FOR RESOURCE CAPTURE providing opportunity for plastic responses to evolve. Phenotypic plasticity has utility in enabling a genotype to produce better adapted phenotypes and phenotype-environment combinations across more environments than would be otherwise be possible. However, if no tradeoﬀs or constraints existed, organisms should be able to exhibit perfect or inﬁnite plasticity by expressing the more adaptive phene or combinations of phenes in every environment with no cost. Costly, but maladaptive or neutral phenotypic responses are expected to go extinct (Dewitt et al., 1998) and we would only expect costly forms of plasticity to persist if they have ﬁtness value. Root phenes have important roles in soil resource capture, especially in environments with suboptimal water and nutrient availability. Root anatomical and architectural phenes determine the temporal and spatial distribution of root foraging in speciﬁc soil domains and hence the capture of mobile and immobile resources (Lynch, 1995, 2013, 2019; Hirel et al., 2007; Lynch and Brown, 2012; Lynch and Wojciechowski, 2015). Mobile soil resources, including nitrate and water, are generally more available in deeper soil domains over time due to crop uptake, evaporation, and leaching throughout the growth season. In contrast, immobile soil nutrients, including phosphorus and potassium, are more available in the topsoil (Lynch and Brown, 2001; Lynch and Wojciechowski, 2015). Plants that are able to acquire edaphic resources at reduced metabolic cost will have increased productivity and performance by permitting greater resource allocation to growth, continued soil resource acquisition, and reproduction (Lynch, 2013, 2015, 2018, 2019). For example, root growth angle inﬂuences root depth, and therefore plant performance in nutrient and water stress conditions (Bonser et al., 1996; Uga et al., 2011; Trachsel et al., 2013; York et al., 2013; Dathe et al., 2016) since steep growth angles enable deeper rooting and the capture of mobile nutrients in deep soil domains (Trachsel et al., 2013; Dathe et al., 2016) while shallow growth angles are more beneﬁcial for the capture of immobile resources in the topsoil (Bonser et al., 1996; Lynch and Brown, 2001; Ho et al., 2005; Zhu et al., 2005c). A plastic response does not imply an adaptive response, although many types of plasticity have important adaptive eﬀects. Adaptive plasticity (positively associated with ﬁtness) and apparent plasticity [lacking adaptive value (e.g., speciﬁc types of allometry or stress responses); Correa et al., 2019] are both types of plasticity. ROOT PHENES ARE IMPORTANT FOR RESOURCE CAPTURE Maladaptive plasticity can occur when a plastic response that was adaptive in an evolutionary context is counterproductive in a novel environment. This is especially relevant for crop breeding, since many agroecosystems, especially high-input agroecosystems, diﬀer sharply from ancestral selection environments, as discussed below. By deﬁnition, allometric responses to the environment may be considered plastic responses, however, they are often just a function of alterations in plant size (or development), and they may not necessarily be adaptive. For example, maize plants with greater biomass had increased stele cross-sectional area and number of metaxylem vessels, which is not necessarily an adaptive response (Yang et al., 2019). However, changes in allometric partitioning (e.g., changes in root to shoot partitioning) may be adaptive by refocusing plant resources to address resource shortfalls (Bloom et al., 1985). In order to interpret diﬀerences in biomass allocation, it is necessary to distinguish these sources of variation. It is diﬃcult to distinguish apparent plasticity from plasticity that may be adaptive. Root anatomical phenes improve plant growth and performance in edaphic stress by reducing the nutrient and carbon costs of tissue construction and maintenance (Lynch, 2013, 2015, 2018, 2019). Root cortical aerenchyma are air-ﬁlled lacunae that result from programmed cell death in root cortical cells (Drew et al., 2000). Air-ﬁlled lacunae replace living cortical parenchyma, thereby reducing root segment respiration and nutrient demand (Saengwilai et al., 2014a; Chimungu et al., 2015; Galindo-Castañeda et al., 2018). The reduction in tissue maintenance costs associated with the formation of root cortical aerenchyma enable roots to explore deeper soil domains and improve the capture of water and nitrogen, and thereby improve plant growth and yield in environments with low water and nitrogen availability (Zhu et al., 2010a; Jaramillo et al., 2013; Saengwilai et al., 2014a; Lynch, 2015; Chimungu et al., 2015). Similar to root cortical aerenchyma, a reduction in the number of cortical cell ﬁles or an increase in cortical cell size also results in a reduction in tissue maintenance and/or construction costs which enables deeper rooting and improved plant growth in drought environments (Chimungu et al., 2014a,b). In temperate small grains, root cortical senescence enables greater exploration of deeper soil domains and greater plant growth in edaphic stress due to reduced cortical burden (Schneider et al., 2017a,b; Schneider and Lynch, 2018). In a paper entitled “The problem of environment and selection,” Falconer argued that environmental eﬀects were a major problem in breeding programs since they interfered with the artiﬁcial selection of a trait (Falconer, 1952). However, it is now understood that plasticity is genetically controlled, heritable, and important for the evolution of the species (Bradshaw, 2006). Phenotypic plasticity is now recognized as a signiﬁcant source of phenotypic variation and diversity and is an important aspect of how organisms develop, function, and evolve (Sultan, 2000). Phenotypic plasticity may be adaptive, maladaptive, or neutral in regard to ﬁtness. In the heterogenous matrix of soil, many phenes and combinations of phene states may have utility for resource capture and display a wide range of variation, May 2020 \| Volume 11 \| Article 546 2 Schneider and Lynch Plasticity as a Breeding Target Frontiers in Plant Science \| www.frontiersin.org ROOT PHENES ARE IMPORTANT FOR RESOURCE CAPTURE In rice, plasticity in lateral root length and density (Kano et al., 2011; Kano-Nakata et al., 2013), root length density, and total root length (Kano-Nakata et al., 2011; Tran et al., 2014) correlated with greater shoot biomass, water uptake, and photosynthesis in drought. The number of nodal roots in rice (Suralta et al., 2010) and maize (Gao and Lynch, 2016), lateral branching density and length in maize (Zhan et al., 2015), and deep rooting in wheat (Ehdaie et al., 2012; Wasson et al., 2012), millet (Rostamza et al., 2013), rice (Hazman and Brown, 2018), and maize (Nakamoto, 1993) also have displayed plastic responses to water deﬁcit. A plastic response of lateral root proliferation was induced in barley in response to patches of nitrogen (Figure 5; Drew et al., 1975) and in maize in response to phosphorus patches (Yano and Kume, 2005). Hydropatterning is a plastic response involving the development of lateral branches, root hairs, and aerenchyma toward available water (Bao et al., 2014). Maize genotypes with plastic root hairs that became longer under low phosphorus had better performance under low phosphorus availability than genotypes with constitutively long root hairs (Figure 6; Zhu et al., 2010b). Root anatomical and architectural phenes express a wide range of plastic responses in a wide range of environments. However, it is unclear which plastic responses are adaptive and how phenes interact to create adaptive responses to edaphic stress. In the ﬁeld, plants may be exposed to successive or multiple, simultaneous stresses. For example, in conditions of terminal drought, seeds are planted in moist soil but the soil progressively dries from the surface due to drainage, evaporation, and plant water uptake, resulting in relatively greater water availability in deeper soil strata and progressively harder topsoils in most agroecosystems (Lynch, 2013; Lynch et al., 2014). Root tissue construction and maintenance demand signiﬁcant resources, and in bean cumulative tissue maintenance demands may exceed root tissue construction costs after 1 week of growth (Nielsen et al., 1994, 2001). The investment of those carbon and nutrient resources in tissue construction and maintenance early in plant growth limits the opportunity for the construction of additional roots in diﬀerent soil domains as resource availability changes. ROOT PHENES ARE IMPORTANT FOR RESOURCE CAPTURE In common bean, reduced secondary growth resulted in reduced speciﬁc root respiration and subsequently greater shoot mass and root length in phosphorus-stress conditions (Strock et al., 2018). Plastic By deﬁnition, edaphic stress reduces plant growth, which is a plastic response but is not necessarily adaptive. For example, reduced grain yield or total root biomass under drought is not an adaptive response, but is a plastic response to stress (Ehdaie et al., 2012) and diﬀerent growing environments and/or diﬀerent genotypes may display diﬀerent rates or types of developmental retardation in response to the same stress. In contrast, the plastic response of genotypes during stress recovery may be adaptive. Phenotypic plasticity encompasses a wide range of environmental responses. Here we focus on understanding the ﬁtness landscape (i.e., how phenes aﬀect crop performance in an array of environments and phene combinations) of root anatomical and architectural phenotypes in agroecosystems. We discuss the beneﬁts and trade-oﬀs to plasticity and the utility of root plasticity in monocots and dicots, acid soils, high and low input environments, and polycultures. We also review the genetic architecture and potential breeding strategies of root phene plasticity. Additionally, we highlight future research directions for root plasticity to enable a comprehensive understanding of the ﬁtness landscape and integration into breeding programs. May 2020 \| Volume 11 \| Article 546 3 Plasticity as a Breeding Target Schneider and Lynch responses of root phenes may have large implications in the capture of edaphic resources. which reduced the metabolic cost of accessing water in deep soil domains (Prince et al., 2017). In drought and low phosphorus environments, increased plasticity of root architecture traits correlated with high yield stability in rice (Sandhu et al., 2016). In water stress, plasticity in root length and root cortical aerenchyma formation has been observed in rice and was associated with greater shoot biomass and yield (Niones et al., 2012, 2013). In water stress in wheat and rice, xylem vessel diameter and number and stele diameter were highly plastic (Kadam et al., 2017). Greater phenotypic plasticity in wheat root anatomical traits may be associated with greater stress tolerance compared to rice (Kadam et al., 2017). In common bean, plasticity in secondary root growth inﬂuenced root depth and shoot growth in low phosphorus environments (Figure 4; Strock et al., 2018). ROOT PHENES ARE IMPORTANT FOR RESOURCE CAPTURE For example, if roots proliferate early in the growth season in the moist topsoil, this limits the opportunity for the construction of roots in deeper soil domains where resources are likely to be located later in the growth season. In addition, early root proliferation in topsoil may not have utility in hard, dry soils later in the season. Root deployment therefore implies opportunity costs, especially during multiple successive or simultaneous stresses. Root architectural and anatomical phenes have important roles in the capture of soil resources in speciﬁc environments, for example sustained nitrogen or phosphorus stress (Lynch, 2013, 2018, 2019), however, root phene states can be functionally maladaptive in ﬂuctuating environments or environments with multiple simultaneous stresses (Ho et al., 2005; Poot and Lambers, 2008). For example, shallow growth angles can improve topsoil foraging and improve the capture of phosphorus, but may be functionally maladaptive for the capture of deep resources like water (Ho et al., 2005). In common bean, shallow growth angle and greater number of basal root whorls and hypocotyl-borne roots increase total root length in the topsoil resulting in greater phosphorus acquisition (Rangarajan et al., 2018). However, as the number of axial roots and/or basal root whorl number increase, the resulting carbon limitation leads to a reduced root depth and therefore trade-oﬀs for the capture of deep resources, such as nitrogen (Rangarajan et al., 2018). In monocots, in which axial roots emerge from shoot nodes, shallow roots lack the ability to forage for deep resources, while deep rooting permits the capture of deep resources like nitrogen and water while also being capable of capturing shallow resources, thus creating asymmetric phenotypic trade-oﬀs for the capture of deep and shallow resources (Lynch, 2013). No single phene state is optimal across a range of environments and management practices (Dathe et al., 2016; Tardieu, 2018; Rangarajan et al., 2018). POTENTIAL BENEFITS AND TRADEOFFS OF PHENOTYPIC PLASTICITY There are many examples of adaptive plasticity of root phenes, including the increased development of root cortical aerenchyma, fewer lateral root branches in water deﬁcit, or deeper distribution of lateral root branches, and it has been proposed that phenotypic plasticity may be the future of crop breeding since it would enable the development of more eﬃcient crops that could adapt to changing environments (Giﬀord et al., 2013; Hazman and Brown, 2018; Lobet et al., 2019). Adaptive plasticity may promote establishment and persistence in novel environments and allows genotypes to have broader tolerance and greater ﬁtness across environments. It has been proposed that understanding the genetic and mechanistic basis of root phenotypic plasticity will enable the rapid development of more productive crop varieties that will be robust and stable in future climates (Topp, 2016). MANY ROOT PHENES ARE PLASTIC Plasticity has been observed for a number of root anatomical and architectural phenes (Figures 2, 3). In soybean grown under drought, metaxylem vessel number increased, thereby improving root hydraulic conductivity, while reducing total cortical area May 2020 \| Volume 11 \| Article 546 Frontiers in Plant Science \| www.frontiersin.org 4 Schneider and Lynch Plasticity as a Breeding Target FIGURE 2 \| Adaptive root phene plasticity to optimize soil resource capture in edaphic stress. A number of root phenes have been demonstrated to have an adaptive plastic response to edaphic stress. In phosphorus stress, plants with many nodal roots with a steep angle, many short lateral branches, root exudation, root cortical aerenchyma formation, and long, dense root hairs are adaptive or proposed adaptive responses for stress tolerance. In nitrogen and water stress, few crown roots with a steep angle, few long lateral branches, root cortical aerenchyma formation, and long root hairs are adaptive or proposed adaptive responses for stress tolerance. FIGURE 2 \| Adaptive root phene plasticity to optimize soil resource capture in edaphic stress. A number of root phenes have been demonstrated to have an adaptive plastic response to edaphic stress. In phosphorus stress, plants with many nodal roots with a steep angle, many short lateral branches, root exudation, root cortical aerenchyma formation, and long, dense root hairs are adaptive or proposed adaptive responses for stress tolerance. In nitrogen and water stress, few crown roots with a steep angle, few long lateral branches, root cortical aerenchyma formation, and long root hairs are adaptive or proposed adaptive responses for stress tolerance. the regulatory loci producing the plastic response to modify expression of other genes. With little known about the molecular mechanisms and genetic control of the plastic response, linkage and pleiotropic eﬀects could severely limit the productivity of plastic crop varieties. The adaptation of taxa to sudden environmental changes, like those caused by human disturbance, could also be an advantage of plasticity since these changes generally occur at too rapid of a pace for an evolutionary response, or the development of new crop cultivars through breeding. In speciﬁc environmental scenarios, plasticity may limit plant productivity. For example, if environmental information is not reliable, plastic organisms can produce maladapted phenotypes when environmental cues are incorrectly interpreted, or when correct signals are interpreted about the initial environment, but the environment ﬂuctuates or is highly variable. Frontiers in Plant Science \| www.frontiersin.org MANY ROOT PHENES ARE PLASTIC In many cases, especially with developmental or morphological plasticity, the development of tissues takes time and often there is a lag time between environmental cues and the development of tissues expressing the plastic response. For example, nitrogen, phosphorus, and water are all growth regulators but have diﬀerent mobilities in soil. Nitrogen and water are mobile and can move faster through the soil proﬁle than plants are able to respond by constructing new tissues or modifying established tissues. Even in the case of phosphorus, an immobile soil resource, changes in phosphate uptake kinetics contribute more to increased phosphorus acquisition than root proliferation in heterogeneous soil environments (Jackson et al., 1990; Caldwell et al., 1992). It also has been suggested that genotypes with ﬁxed development (i.e., non-plastic phenes) may be able to express However, “perfect” plasticity is unattainable due to an inability to consistently produce the optimum phenotype, ﬂuctuating environmental signals, and/or because phenotypic plasticity comes at a cost (León, 1993; Via and Lande, 2006). A cost of plasticity is when a plastic organism exhibits less ﬁtness while producing the same phene state as a ﬁxed organism. Costs of plasticity have been identiﬁed in a variety of systems (Relyea, 2002; Merilä et al., 2004). Maintenance cost of phenotypic plasticity may be incurred if facultative development requires the maintenance or construction of sensory and regulatory machinery that ﬁxed development does not require. Genetic costs of plasticity also exist. Phenotypic plasticity may manifest because structural genes or their products are directly aﬀected by the external environment (i.e., allelic sensitivity) or because regulatory genes are aﬀected by the environment which in turn aﬀect the expression of structural genes (Via et al., 1995). However, genetic linkage may cause genes associated with plasticity to be linked with genes conferring reduced ﬁtness, plasticity genes may have negative pleiotropic eﬀects on phenes other than the plastic phene, or epistasis may cause May 2020 \| Volume 11 \| Article 546 5 Schneider and Lynch Plasticity as a Breeding Target FIGURE 3 \| Gentoypes vary in their plastic response to environment, nitrogen stress, and drought. Architectural and anatomical images are presented from a single genotype in response to different environments and edaphic stress conditions. Phenotypic plasticity is shown for root architecture, root anatomy, and lateral branching length and density. Scale bar represents 2 cm (root crown and lateral branch) and 1 mm (anatomy). MANY ROOT PHENES ARE PLASTIC FIGURE 3 \| Gentoypes vary in their plastic response to environment, nitrogen stress, and drought. Architectural and anatomical images are presented from a single genotype in response to different environments and edaphic stress conditions. Phenotypic plasticity is shown for root architecture, root anatomy, and lateral branching length and density. Scale bar represents 2 cm (root crown and lateral branch) and 1 mm (anatomy). FIGURE 4 \| Secondary root growth in common bean (Phaseolus vulgaris) is plastic in response to phosphorus availability. Comparison of basal root anatomy under high P and P stress in greenhouse conditions at 46 DAP. A11 cross-sections are at the same scale. Modiﬁed and reproduced with permission from Strock et al. (2018). FIGURE 4 \| Secondary root growth in common bean (Phaseolus vulgaris) is plastic in response to phosphorus availability. Comparison of basal root anatomy under high P and P stress in greenhouse conditions at 46 DAP. A11 cross-sections are at the same scale. Modiﬁed and reproduced with permission from Strock et al. (2018). FIGURE 5 \| Lateral root proliferation of barley in response to a nutrient patch. (A) A plant supplied with a uniform treatment of nitrate has a uniform lateral branching density and length along the axial root. (B) A plant supplied with nitrate through a banded treatment displays lateral root proliferation in the banded region. Modiﬁed and reproduced with permission from Drew (1975). FIGURE 4 \| Secondary root growth in common bean (Phaseolus vulgaris) is plastic in response to phosphorus availability. Comparison of basal root anatomy under high P and P stress in greenhouse conditions at 46 DAP. A11 cross-sections are at the same scale. Modiﬁed and reproduced with permission from Strock et al. (2018). FIGURE 5 \| Lateral root proliferation of barley in response to a nutrient patch. (A) A plant supplied with a uniform treatment of nitrate has a uniform lateral branching density and length along the axial root. (B) A plant supplied with nitrate through a banded treatment displays lateral root proliferation in the banded region. Modiﬁed and reproduced with permission from Drew (1975). more extreme phene states than plastic genotypes since there may be a trade-oﬀbetween the developmental range that can be expressed across habitats and the magnitude of expression within an environment (Wilson and Yoshimura, 1994; DeWitt, 1998). MANY ROOT PHENES ARE PLASTIC more extreme phene states than plastic genotypes since there may be a trade-oﬀbetween the developmental range that can be expressed across habitats and the magnitude of expression within an environment (Wilson and Yoshimura, 1994; DeWitt, 1998). Costly, but maladaptive or neutral phenotypic responses are expected to go extinct (Dewitt et al., 1998) and we would only expect costly forms of plasticity to persist if they have ﬁtness May 2020 \| Volume 11 \| Article 546 Frontiers in Plant Science \| www.frontiersin.org 6 Plasticity as a Breeding Target Schneider and Lynch FIGURE 6 \| Root hair length is plastic in response to 10W phosphorus environments. Root hairs become longer in low phosphorus environments and are associated with greater shoot biomass. Root hair image courtesy of Anica Massas. biotic and abiotic stress inﬂuencing root function may not have utility in these high-input environments (Lynch, 2018). Root phenotypes that explore deep soil domains, whether plastic or not, enhance the capture of deep resources like water and nitrogen in most agricultural systems, despite the fact that water and nitrogen availability are sometimes greater in surface soils of high-input systems (Manschadi et al., 2006; Gowda et al., 2011; Henry et al., 2011). If a population is exposed to a novel environment and becomes successful, but becomes restricted to that environment, alleles that contributed to plastic responses in the new environment should trend toward ﬁxation in the absence of gene ﬂow from other populations and therefore their ability to confer plasticity is also reduced (Mitchell-Olds et al., 2007; Anderson et al., 2017; Gage et al., 2017). The utility of phenotypic plasticity in successful and highly productive modern crop varieties in heavily managed high-input environments is limited and not required for the survival or migration of the species. Frontiers in Plant Science \| www.frontiersin.org PLASTICITY IN THE CONTEXT OF TEMPORAL RESOURCE DURATION The expression of plant phenes as a result of plasticity may be of variable duration and plastic responses may be long- or short-term. Short-term plasticity, is also referred to as physiological plasticity, allows plants to adjust to temporally variable aspects of the environment such as water or nitrogen availability. For example, the expression of aquaporins or nitrate transporters ﬂuctuates as a short-term response to water or nitrogen availability (Feng et al., 2011; Zargar et al., 2017). In contrast, changes due to morphological or developmental plasticity may be of longer duration (Sultan, 2000). For example, the size and number of cortical cells or initial root angle is established near the growing root apex, and potential for change in mature tissues is limited. Phenotypic plasticity that is established early in development, such as root growth angle, may be beneﬁcial in conditions of sustained edaphic stress (e.g., low phosphorus availability), but may be maladaptive in stresses that ﬂuctuate on shorter time scales (e.g., drought, low nitrogen availability) by creating sustained responses to ephemeral conditions (Lynch, 2013). In addition, the timing of development itself, and its response to the environment, may be plastic. Developmental plasticity may be limited to early growth stages, or its timing may vary in diﬀerent genotypes or species (Pigliucci and Schlichting, 1995). For example, the development of root cortical senescence has the greatest utility in edaphic stress conditions when development occurs relatively early in plant growth, however, genotypic contrasts exist for the rate and timing of its development in root cortical tissues (Schneider et al., 2017b). FIGURE 6 \| Root hair length is plastic in response to 10W phosphorus environments. Root hairs become longer in low phosphorus environments and are associated with greater shoot biomass. Root hair image courtesy of Anica Massas. value in some seasons or environments. We speculate that plasticity was a useful mechanism for crop ancestors to grow and develop in novel environments and thrive in unmanaged, unfertilized, and non-irrigated natural ecosystems. In low-input systems, plasticity may be advantageous by exploiting resource patches with increased lateral root proliferation which may confer a competitive advantage (Lynch, 2018). However, in modern agricultural environments with high-inputs, plasticity may come at a greater cost than a beneﬁt. PLASTICITY IN THE CONTEXT OF TEMPORAL RESOURCE DURATION Indirect evidence for this is the observation that during selection of modern temperate maize breeding, regions of the genome contributing to G × E variance and plasticity were not directly or indirectly selected to increase plant productivity and yield stability (Gage et al., 2017). Short duration plasticity, or physiological plasticity, in variable environments may be advantageous in speciﬁc environments, however, plasticity may be maladaptive in high-input environments with intensive fertilization and greater nutrient availability. In high-input environments, constraints for soil resource acquisition and plant growth in stress are mitigated and strategies that evolved in environments with value in some seasons or environments. We speculate that plasticity was a useful mechanism for crop ancestors to grow and develop in novel environments and thrive in unmanaged, unfertilized, and non-irrigated natural ecosystems. In low-input systems, plasticity may be advantageous by exploiting resource patches with increased lateral root proliferation which may confer a competitive advantage (Lynch, 2018). However, in modern agricultural environments with high-inputs, plasticity may come at a greater cost than a beneﬁt. Indirect evidence for this is the observation that during selection of modern temperate maize breeding, regions of the genome contributing to G × E variance and plasticity were not directly or indirectly selected to increase plant productivity and yield stability (Gage et al., 2017). value in some seasons or environments. We speculate that plasticity was a useful mechanism for crop ancestors to grow and develop in novel environments and thrive in unmanaged, unfertilized, and non-irrigated natural ecosystems. In low-input systems, plasticity may be advantageous by exploiting resource patches with increased lateral root proliferation which may confer a competitive advantage (Lynch, 2018). However, in modern agricultural environments with high-inputs, plasticity may come at a greater cost than a beneﬁt. Indirect evidence for this is the observation that during selection of modern temperate maize breeding, regions of the genome contributing to G × E variance and plasticity were not directly or indirectly selected to increase plant productivity and yield stability (Gage et al., 2017). Short duration plasticity, or physiological plasticity, in variable environments may be advantageous in speciﬁc environments, however, plasticity may be maladaptive in high-input environments with intensive fertilization and greater nutrient availability. UTILITY OF ROOT PLASTICITY VARIES BETWEEN DICOTS AND MONOCOTS Monocots and dicots have diﬀerent foraging strategies for edaphic resources. Throughout the growth season, monocots continually produce new roots from stem nodes, and tillers. In contrast, new roots of dicots are predominately lateral roots arising from older root axes. Dicots do have younger hypocotyl- borne roots that emerge throughout the growth season, however, they normally do not comprise a large portion of the root system, which usually consists of relatively few axial roots of larger diameter with a highly developed lateral root system having multiple orders of lateral branching. Monocots may have superior topsoil foraging, as new ﬂushes of roots are continuously pushed down through shallow soils, whereas in dicots many new roots form in deeper soil domains (Lynch, 2013). In addition, in tillering monocot species, an optimum number of tillers should exist to enhance capture of edaphic resources as the number of tillers is directly related to the number of adventitious roots (Hecht et al., 2016). Reduced crown root number improves plant growth with low nitrogen (Saengwilai et al., 2014b) and drought (Gao and Lynch, 2016) by reducing inter- and intra- plant competition for internal and external resources, thereby increasing root depth and acquisition of deep soil resources. However, greater crown root number improves plant growth in low phosphorus soil by reducing axial root elongation and improving topsoil foraging (Sun et al., 2018). We speculate that the number of tillers (and therefore the number of adventitious roots originating from tillers), and its plastic response to plant density and stress, is important for edaphic stress tolerance in monocot species. Acid subsoils (generally deﬁned as having a pH < 5) present several challenges to root growth and resource acquisition including aluminum (Al) toxicity, deﬁciency of phosphorus (P), calcium (Ca), magnesium (Mg) and potassium (K), and possibly manganese (Mn) toxicity. In acid soils, the solubility of Al increases and injury to root apices occurs, therefore reducing root growth, soil exploration, and subsequent resource acquisition. Commonly, acidic soils are located in humid environments with weathered soils, and acidity increases with soil depth. Plasticity of root phenes that increase topsoil foraging would be beneﬁcial by improving the capture of resources that have greater availability in the topsoil, including P, Ca, Mg, and K (Lynch, 2019), while also avoiding subsoils with greater acidity and Al toxicity (Lynch and Wojciechowski, 2015). PLASTICITY IN THE CONTEXT OF TEMPORAL RESOURCE DURATION In high-input environments, constraints for soil resource acquisition and plant growth in stress are mitigated and strategies that evolved in environments with Short duration plasticity, or physiological plasticity, in variable environments may be advantageous in speciﬁc environments, however, plasticity may be maladaptive in high-input environments with intensive fertilization and greater nutrient availability. In high-input environments, constraints for soil resource acquisition and plant growth in stress are mitigated and strategies that evolved in environments with In response to heterogeneous soil conditions, root plasticity can also vary spatially. Lateral root branches have been documented in some species and genotypes to proliferate in response to localized patches of nutrient availability (Figure 5; Drew, 1975; Zhu and Lynch, 2004). Lateral root proliferation in response to nutrient patches has been proposed as a beneﬁcial May 2020 \| Volume 11 \| Article 546 7 Plasticity as a Breeding Target Schneider and Lynch acquisition. RCS reduces the carbon and nutrient costs of soil exploration by destroying living cortical tissue, thereby reducing carbon and nutrient costs of maintaining a living cortex. The development of RCS may be plastic as limited phosphorus and nitrogen availability accelerate the development of RCS (Schneider et al., 2017a,b). After the development of RCS in monocots or secondary growth in dicots, assimilates that would have been partitioned to the root for maintenance of the cortex may be used for the growth of shoots or new roots, which can increase soil exploration. Monocots and dicots have diﬀerent foraging and resource acquisition strategies and therefore may have diﬀerent adaptive plastic responses for soil resource capture. strategy for enhanced nitrogen acquisition (Mi et al., 2010), however, if mobile resources move faster through the soil proﬁle than roots can proliferate, this response may be maladaptive. In some species, plasticity of lateral root branching in response to local nutrient patches may enhance nutrient resource capture in environments with sustained nutrient sources or in conditions of interspeciﬁc competition (Robinson et al., 1999). However, this can be detrimental when proliferation in response to local nutrients diverts resources from other soil domains with greater resource availability, particularly deeper soil domains in leaching precipitation regimes later in the growing season (Lynch, 2013, 2018). Frontiers in Plant Science \| www.frontiersin.org PLASTICITY IN THE CONTEXT OF POLYCULTURES In many low-input agroecosystems, which traditionally consist of polycultures and generally experience greater weed competition, interplant competition with other species has important implications in plant performance. For example, the maize/bean/squash polyculture used in small-scale subsistence farming has a yield advantage over the average yield of the respective monocultures (Mt. Pleasant and Burt, 2010). Maize, bean, and squash have contrasting root architectures (Postma and Lynch, 2012; Zhang et al., 2014) and diﬀerences in root architecture and vertical root distribution result in diﬀerences in spatial niches and allows polycultures to be productive when plants are competing for soil resources (Zhang et al., 2014). In these polyculture systems, species co-optimize, and spatial niches allow a yield advantage by reducing competition for edaphic resources. In polyculture or multiline systems, highly plastic root architectural phenes could disrupt complementary spatial niche foraging strategies (Zhang et al., 2014). If these species had highly plastic root architectural phenotypes, this would create more competition for the same soil resources, which would be detrimental. For example, if roots of all species proliferate in response to localized patches of nutrient availability, this creates greater inter-plant and species competition. In this scenario, phenotypic plasticity may not be adaptive, as complementary spatial niches are needed for the success of all species or the population. UTILITY OF ROOT PLASTICITY VARIES BETWEEN DICOTS AND MONOCOTS The tradeoﬀ of reduced access to deep soil water would probably be less important in humid environments because of greater water availability in shallower soil domains. Topsoil foraging can be improved through a shallower axial root growth angle (Bonser et al., 1996; Liao et al., 2001), greater production of axial roots (Walk et al., 2006; Miguel et al., 2013; Rangarajan et al., 2018; Sun et al., 2018), denser lateral roots (Postma et al., 2014; Jia et al., 2018), and greater root hair length and density (Zhu et al., 2010b; Miguel et al., 2015). Reduced root metabolic cost improves growth in soils with low phosphorus availability. In maize, the formation of root cortical aerenchyma reduces root respiration and the phosphorus cost of maintaining root tissue therefore improving plant growth in low phosphorus (Postma and Lynch, 2011; Galindo-Castañeda et al., 2018). In bean, phosphorus stress inhibits secondary growth of roots which reduces root costs and improves phosphorus capture and plant growth in low phosphorus soils (Strock et al., 2018). Plastic root phenes that improve topsoil foraging may be beneﬁcial for improved capture of phosphorus in acidic soils. There are important diﬀerences between the anatomy of monocot and dicot roots. Roots of dicot species radially expand through secondary growth, which has important implications for edaphic stress tolerance. Phosphorus stress reduces secondary growth in Phaseolus vulgaris in a genotype-dependent manner, and genotypes with greater reduction of secondary growth had reduced metabolic costs, increased root length, improved phosphorus capture, and increased shoot biomass in low phosphorus soil (Figure 4; Strock et al., 2018). In monocots, temperate small grain species develop root cortical senescence (RCS), a type of programmed cell death. Simulation studies suggest that RCS may be an adaptive trait for water and nutrient Plasticity in carboxylate exudation may also be an important mechanism for phosphorus uptake in acidic soils. Carboxylate exudation into the rhizosphere solubilizes phosphorus from metal complexes (Ryan et al., 2012). Carboxylates also can precipitate toxic levels of aluminum in the soil (Lambers et al., 2003). ROOT PLASTICITY IN THE CONTEXT OF HIGH AND LOW INPUT ENVIRONMENTS It has been proposed that wild crop ancestors and landraces produce more roots than directly needed for the capture of edaphic resources to compensate for root loss from biotic stress, edaphic stress, and competition for soil resources with neighboring plants (Lynch, 2018). We speculate that plasticity was a useful mechanism for crop ancestors in natural ecosystems. Short duration plasticity, or physiological plasticity, in variable environments may be advantageous in speciﬁc environments, however, plasticity may be maladaptive in high-input environments with intensive fertilization, greater nutrient availability, and reduced biotic stress. In high-input agroecosystems, parsimonious, non-plastic root phenotypes including e.g., fewer axial roots, reduced density and length of lateral roots, reduced cortical cell ﬁle number, and reduction of cortical parenchyma through formation of aerenchyma and senescence may be beneﬁcial by permitting deeper rooting and the capture of deep resources like water and nitrogen (Lynch, 2018). Plastic responses to increase topsoil foraging in response to shallow localization of water and N early in the growth season may optimize resource capture in natural systems or low-input agroecosystems, characterized by intense belowground competition from neighboring plants. However, in high-input monocultures, where immobile resources like P and K are likely to be non-limiting, non-plastic phenotypes would be advantageous since eventually water and N would be localized at depth regardless of early season patterns, and resources lost to neighboring plants would still contribute to stand-level ﬁtness (i.e., yield) in high density monocultures (Lynch, 2018). We propose that in low-input systems, highly plastic root phenotypes with a variable number of axial and lateral roots, variable UTILITY OF ROOT PLASTICITY VARIES BETWEEN DICOTS AND MONOCOTS Exudation of carboxylates in plant roots is a common phenomenon in many plants including rice (Kirk et al., 1999), May 2020 \| Volume 11 \| Article 546 8 Plasticity as a Breeding Target Schneider and Lynch root growth angle, variable length and density of root hairs, variable formation of root cortical aerenchyma and cortical cell ﬁles, would be beneﬁcial for the capture of heterogeneous soil resources in environments with signiﬁcant root loss due to biotic factors. However, in high-input systems, a sparser root system with fewer axial roots may be more beneﬁcial, since the negative eﬀects of biotic stress is diminished (Lynch, 2018). wheat (Ryan et al., 1995), and lupin (Gardner et al., 1983). The exudation of carboxylates including citrate and malate into the rhizosphere can incur a large carbon cost (Lambers et al., 2013). Plasticity in the spatiotemporal control of carboxylate exudation, i.e., exudation triggered by aluminum toxicity and phosphorus stress may permit a reduction in the metabolic burden of the root. Low Ca availability is a major challenge to root growth in acid subsoils (Foy et al., 1969). Diﬀerences in cell wall composition may inﬂuence tissue Ca requirements and plants with reduced internal Ca requirement therefore may be more productive in acid soils (Lynch and Wojciechowski, 2015). Cortical cell size, ﬁle number, and aerenchyma all inﬂuence the amount of cell wall material per root volume and therefore aﬀect tissue Ca requirement. Genotypes with reduced pectin content, which has a reduced demand for Ca, may also reduce the Ca requirement of the root (Marschner, 1995). We propose that plasticity of phenes that reduce tissue Ca requirements, like increased cortical cell size, reduced ﬁle number, reduced pectin content, and increased aerenchyma formation may be beneﬁcial in acid soils. Crops with a reduced Ca tissue requirement may be able to continue to explore acidic subsoils, despite reduced Ca availability and Al toxicity. Frontiers in Plant Science \| www.frontiersin.org PROGENY MAY BE PRIMED FOR A PLASTIC RESPONSE Plants cannot only respond to environmental signals by adjusting their own phenotypes, but also can inﬂuence the phenotypes of their oﬀspring, through changes in the quantity and quality of seed production and the structure and quality of the seed coat and fruit tissues (Sultan, 2000). The phenotype of oﬀspring can be inﬂuenced by the parental environment. For example, plants can respond to speciﬁc environments by changing the structure of thickness of the seed coat while maintaining the quantity and quality of the embryo and endosperm tissues (Sultan, 1996; Lacey et al., 1997). Genotypes may vary in the extent to which seedling and mature root phenes are aﬀected by parental stress. For example, progeny of some common bean genotypes from drought-stressed parents developed fewer and shorter basal roots with smaller diameters (Lorts et al., 2019). Progeny from some genotypes from phosphorus stress parents May 2020 \| Volume 11 \| Article 546 9 Plasticity as a Breeding Target Schneider and Lynch 2017). It is important to note that maladaptive plasticity in a speciﬁc environment may be adaptive in diﬀerent environments, including future climates. Plasticity that is not currently adaptive can provide sources of variation that may be important for phenotypic evolution or variation for breeding (Lande, 2009). developed fewer shoot-borne roots and had a greater basal root whorl number (Lorts et al., 2019). Progeny of nutrient-deprived plants increase allocation to root biomass compared to progeny of plants with ample nutrients (Wulﬀand Bazzaz, 1992). Oﬀspring of light-deprived plants reduce root elongation relative to shoot growth compared to progeny of plants grown in high light (Sultan, 2000). In addition, epigenetic processes, including DNA methylation and histone modiﬁcation, may alter gene expression and therefore may be important drivers in phenotypic plasticity (Chinnusamy and Zhu, 2009; Nicotra et al., 2010). These plastic changes may enable oﬀspring to maintain critical aspects of plant growth and function, even if the initial seedling biomass is reduced by parental stress. The genetic architecture of plasticity is highly complex and quantitative. Many genes with small eﬀects control plastic responses and distinct genes control plastic responses of diﬀerent root phenes and in response to diﬀerent stresses and environments (Schneider et al., 2020a,b). This can pose a challenge for breeding programs that use conventional tools like single-trait breeding strategies and marker assisted selection, as hundreds of genes would need to be stacked for the development of desirable root ideotypes for speciﬁc environments. GENETIC ARCHITECTURE OF PLASTICITY AND BREEDING STRATEGIES In addition, genes controlling root anatomical and architectural phenes and their plastic responses are probably highly pleiotropic. For example, multiple root anatomical and architectural phenes are regulated by ethylene (Takahashi et al., 2015; Schneider et al., 2018). Ethylene signaling induces root cortical aerenchyma and RCS formation via programmed cell death (Evans, 2003; Schneider et al., 2018) and presumably common signaling pathways (e.g., ethylene) control expression of other root phenes under a range of edaphic stresses [i.e., lateral root formation (Negi et al., 2008)]. For example, the upregulation of an ethylene-related gene may be intended to increase aerenchyma formation for adaptation in drought environments, however, increased ethylene production may also have unintended eﬀects such as reduced axial root elongation which may be maladaptive in these environments. We must fully understand the genetic architecture of phene plasticity as well as the function of phenes and phene aggregates in order to develop adaptive crop cultivars for speciﬁc environment. Some genetic loci have been associated with root phenes including root stele and xylem vessel diameter in rice (Uga et al., 2008, 2010), xylem vessel phenes in wheat (Sharma et al., 2010), root cortical aerenchyma in Zea species (Mano et al., 2006, 2007), areas of cross section, stele, cortex, aerenchyma, and cortical cells, root cortical aerenchyma, cortical cell ﬁle number, and length, number, and diameter of nodal roots in maize (Burton et al., 2014a,b). However, genes associated with phene expression are distinct from those associated with plasticity for that expression. Genes associated with plasticity have been identiﬁed for root hair length (Zhu et al., 2005a) and lateral root branching and length (Zhu et al., 2005b) in low phosphorus availability in maize, root length density and root dry weight (Sandhu et al., 2016) in rice in response to drought, lateral root branching in rice in response to ﬂuctuating moisture levels (Niones et al., 2015), and wheat and rice root anatomical phenes in response to drought (Kadam et al., 2017). In maize, genes associated with plasticity in response to water deﬁcit and diﬀerent environments are distinct for cortical phenes, root angle, and lateral branching phenes (Schneider et al., 2020a,b; Table 1). Understanding the genetic architecture of plasticity could provide useful breeding targets for crop improvement in speciﬁc environments and improve our understanding of phenotypic plasticity. Plasticity is heritable, and this enables selection for or against plasticity in manmade populations (Pigliucci, 2005). PROGENY MAY BE PRIMED FOR A PLASTIC RESPONSE However, modern breeding methods, like genomic selection enable the selection of multiple loci. Frontiers in Plant Science \| www.frontiersin.org GENETIC ARCHITECTURE OF PLASTICITY AND BREEDING STRATEGIES Historically, breeding programs have focused on selecting crop varieties based on uniformity and yield stability in speciﬁc environments and management practices, and plasticity has often considered to be a breeding obstacle (Basford and Cooper, 1998; Cooper et al., 1999). Large and complex genotype by environment interactions complicate the design and implementation of breeding strategies (Cooper et al., 1999) and breeders often select for a low genotype by environment contribution to enable genotypes to perform predictably in speciﬁc environments. Crop breeding has made huge advancements in the development of productive varieties that are stable across diverse conditions and recent studies have suggested that plasticity was not directly or indirectly selected for in the development of modern crop varieties (Gage et al., In breeding programs with capabilities to use genomic selection, selection should include phenes and integrated phenotypes (and their plastic responses), not just selection for yield. Selection for individual phenes has merits compared with brute-force yield selection for edaphic stress (Lynch, 2019). In training sets for genomic selection, consideration must be given to wild germplasm and landraces, since elite germplasm has been developed through selection in high-input environments and often against plasticity. Landraces and wild germplasm presumably express more phenotypic plasticity than uniform, stable elite crop germplasm and could provide unique sources of phenotypic variation. Phenotypic selection for plasticity may also be a viable strategy for breeding programs, however, selection must occur in speciﬁc targeted environments or under speciﬁc edaphic stresses. A genotype that displays adaptive plastic responses to water stress may not express an adaptive (or any) plastic responses to other edaphic stresses such as limited nutrient availability. The phenotyping of plasticity should be evaluated for individual phenes, as plasticity in a variety of phenes and phene combinations can result in similar yield or measures of plant performance. May 2020 \| Volume 11 \| Article 546 10 Plasticity as a Breeding Target Schneider and Lynch TABLE 1 \| Summary of identiﬁed genetic loci associated with root plasticity and architecture. GENETIC ARCHITECTURE OF PLASTICITY AND BREEDING STRATEGIES Species Root trait Response References Soybean Root length, Number of adventitious roots, Number of root tips Waterlogging Ye et al., 2018 Rice Root diameter, Stele diameter, Cortical diameter, Metaxylem vessel number and diameter, Root length, Speciﬁc root length, Root volume, Root surface area Drought Kadam et al., 2017 Rice Lateral root branching Drought Niones et al., 2015 Rice Root length density and root dry weight Drought Sandhu et al., 2016 Arabidopsis Root volume, Weight, Deep root weight Drought Li et al., 2017 Arabidopsis Root length and dry weight Drought El-Soda et al., 2015 Maize Lateral root branching and length Low Phosphorus Zhu et al., 2005b Maize Root hair length Low Phosphorus Zhu et al., 2005a Maize Root cortical aerenchyma, Cortical cell size and ﬁle number, Metaxylem vessel area, Cortical area, Stele area, Root cross-sectional area Drought Schneider et al., 2020a Maize Root angle, Lateral root branching length and density, Distance to the ﬁrst lateral branch Drought Schneider et al., 2020b Several genes have been identiﬁed in many different species for a number of anatomical and architectural phenes. The adaptive value of plasticity in breeding programs is limited by distinct genetically controlled plasticity responses to diﬀerent environmental conditions. Breeders may need to target a speciﬁc plastic response of a speciﬁc phene or set of phenes to a speciﬁc abiotic or biotic stress or environment, rather than just breed for a variety that highly expresses phenotypic plasticity. Genotypes that have a plastic response to water deﬁcit are not the same set of genotypes with a plastic response to diﬀerent environments (i.e., G × E) (Schneider et al., 2020a,b). Breeding eﬀorts to develop varieties that are plastic to a wide range of environments and stresses, may be maladaptive in environments with multiple stresses or stresses that ﬂuctuate on short time scales or that vary throughout the growth season. The development of new crop varieties can take decades, and the utility of phene states in the current target environment may change in future environments and climates. Since each plastic response to an environmental cue has distinct genetics, use of plasticity as a selection criterion is challenging for breeders who must target each plastic response to a speciﬁc environment or stress. studies have demonstrated that phenotypic plasticity is phene- speciﬁc, not necessarily genotype-speciﬁc (Schneider et al., 2020a,b) so it is important to measure individual phenes as opposed to phene aggregates. GENETIC ARCHITECTURE OF PLASTICITY AND BREEDING STRATEGIES When plasticity of a phene aggregate, or combinations of multiple elemental phenes, is measured, it may reﬂect a plastic response of one or multiple phenes. In addition, when phene aggregates are measured, phenotypic plasticity may be masked by diﬀerent responses of elemental phenes. For example, the diameter of the root may not exhibit plasticity, but the stele size, cortical cell ﬁle number, or size of cortical cells may have changed their phenotype. Many combinations of elemental phenes have the potential to produce the same expression of combinations of phenes. Contrary to earlier neo-Darwinian views of plasticity as trivial “noise,” plasticity is now considered to be an important source of phenotypic variation. Root systems consist of multiple phenes, each under distinct genetic control, that interact with each other and the environment to determine ﬁtness. The ﬁtness landscape of root phenes and their plastic responses that vary among genotypes, species, and environment is poorly understood. Plants are not equipped with unlimited phenotypic plasticity, which suggests that there are constraints to its expression (Schlichting, 1986). Frontiers in Plant Science \| www.frontiersin.org FUTURE RESEARCH DIRECTIONS In the context of plant roots, there are many combinations of phenes that aﬀect ﬁtness of a plant in a speciﬁc environment. In order to interpret the adaptive value, utility of phenotypic plasticity, and consider plasticity in breeding programs we must ﬁrst understand the ﬁtness landscape of individual phenes and phene combinations. (e.g., multiple, simultaneous dynamic stresses). For example, understanding plastic responses to multiple constraints is important. Very few studies have tested plastic responses to multiple simultaneous abiotic and biotic stresses. Short-term or dynamic plasticity is an important but poorly understood component of plasticity that includes the rate of phenotypic response or patterns of development. Plasticity of short duration may be important in maintaining ﬁtness, particularly in ﬂuctuating environments. Dynamic plasticity is challenging to measure, as it requires phenes to be measured over time in many individuals in diﬀerent environments. Common phenotyping tools require destructive harvests at ﬁxed times or at ﬁxed growth stages and are slow and costly. However, this is critical to understanding plasticity, as the determination of whether plasticity is adaptive or maladaptive depends strongly on its temporal expression. The extent of variation in expression of plasticity still remains to be explored in many root phenes. Phenotypic plasticity may be an important source of genetic variation to be exploited for the development of crop varieties for future environments. However, breeding for genotypes with plastic responses will be complicated by their complex genetic architecture, genetic and metabolic costs of plasticity, and potential maladaptive responses in many environments. To understand patterns of plasticity, we need to better understand and monitor local environments and changes in the environment. Subtle changes in the environment, such as localized nutrient patches, may induce a phenotypic response and if the environment is not carefully monitored, it makes interpretation of the plastic responses challenging (Schneider et al., 2020a). Field environments are often heterogeneous and diﬃcult to monitor and replicate. In silico approaches enable the evaluation of many environment and phenotype combinations including those that do not exist in nature (Dunbabin et al., 2013). The use of modern in silico approaches will be necessary to understand the complex interactions of the root ﬁtness landscape that are not possible empirically. We propose that some of the main ideas discussed here regarding root phenotypic plasticity are applicable to shoot phenotypic plasticity. In high-input environments, shoot architecture and anatomy is optimized for enhanced plant performance. FUTURE RESEARCH DIRECTIONS Several recent studies have focused on the utility of speciﬁc phenes in edaphic stress (Trachsel et al., 2013; Chimungu et al., 2014a,b; Saengwilai et al., 2014a,b; Schneider et al., 2017a; Strock et al., 2018), however, the utility of many other root phenes in edaphic stress remains to be explored. In addition, recent studies have explored interactions between root phenes which may be synergistic or antagonistic in nature (Miguel et al., 2015; Rangarajan et al., 2018). For example, in dicots tradeoﬀs exist between shallow and deep soil foraging (Ho et al., 2005). Recent studies suggest that plasticity is phene- speciﬁc and a single genotype may produce an adaptive plastic response for one phene and maladaptive plastic response for a diﬀerent phene on the same plant (Schneider et al., 2020a). Should root plasticity be a breeding target? The answer is complex. The ﬁtness landscape of root phenotypic plasticity is dependent on speciﬁc agroecologies and management practices, and the genetic control of plasticity is in general highly quantitative and is dependent on many loci having small eﬀects. To better understand and interpret plasticity, ﬁrst we need a comprehensive understanding of the utility of individual phenes. Numerous studies evaluate plasticity of speciﬁc root length, root biomass, or yield. However, speciﬁc root length may depend on the expression of many individual phenes including the formation of root cortical aerenchyma, cortical cell ﬁle number, and stele area. Previous Frontiers in Plant Science \| www.frontiersin.org May 2020 \| Volume 11 \| Article 546 11 Plasticity as a Breeding Target Schneider and Lynch Presumably, genes controlling adaptive phenotypic plasticity would have to be stacked in breeding programs to create a suite of adaptive synergistic phenes. Understanding the utility of root phenes and their interactions will have important implications in understanding adaptive or maladaptive plasticity under speciﬁc edaphic stresses. In many cases, more detailed and reﬁned phenotyping methods are needed to be able to characterize and phenotype phene states, rather than phene aggregates. In many knockout collections, the annotation of “no visible phenotype” is common and is partly due to the lack of capacity for the plant science community to analyze subtle and complex elemental phenes. Field phenotyping is a bottleneck in crop breeding programs and high-throughput, industrial-scale phenotyping often does not allow for the identiﬁcation and understanding of subtle, complex phene states. FUTURE RESEARCH DIRECTIONS Similar to root plasticity, we speculate that in high- input environments, shoot plasticity may not be advantageous, since human management has removed many constraints to shoot function. However, shoot phenotypic plasticity ideotypes, beneﬁts, and trade-oﬀs in many ways are not equivalent to root plasticity as soil resources are spatially and temporally dynamic and much more complex than above-ground environments. Like root phenes, we must fully understand the ﬁtness landscape of shoot phenotypic plasticity before its integration into breeding programs. p p y Growth diﬀerences between controlled and ﬁeld environments are often overlooked. Planting density, light, temperature, and other growing conditions have large eﬀects on plant growth and are often dramatically diﬀerent in the ﬁeld compared to controlled environments and phenotypic correlations between lab and ﬁeld data are often poor (Poorter et al., 2016). Controlled environments and growing systems do not represent the heterogeneous matrix of the soil and therefore are diﬃcult to use to discover true plant responses. There is a clear need to employ abiotic conditions that are overall more similar to those which the plants experience in the ﬁeld (e.g., more natural soils, appropriate planting densities, light intensity). Many previous studies on phenotypic plasticity have focused on environmental responses in straightforward traits including biomass and root-shoot ratios (Bradshaw and Hardwick, 1989) and numerous studies have observed plastic or genotype by environment responses of below- and above-ground plant phenes (Robinson et al., 1999; Gage et al., 2017; Rabbi et al., 2017). Now that we have a basic understanding of plasticity, we can move to understanding more complex and subtle aspects of phenotypic plasticity. Single-factor experiments have been important in understanding plastic responses, however, more realistic environmental complexity is needed in studies To harness the power and knowledge of genomic information and agricultural application of plasticity, we need to be able to comprehensively link genetic information to “real world” phenotypes in “real world” environments. We need to measure the adaptive signiﬁcance of patterns in plasticity and understand the complex pathways that lead from environmental cues to a plastic response. The ﬁtness landscape of plasticity is highly complex, yet poorly understood and merits further research to understand the utility of plasticity for edaphic stress tolerance. The study of phenotypic plasticity involves many disciplines including ecology, physiology, development morphology, genetics, in silico biology and evolution and oﬀers many research opportunities to understand links among these areas. Frontiers in Plant Science \| www.frontiersin.org REFERENCES 37, 137–155. doi: 10.1111/j.1095- 8312.1989.tb02099.x Brisson, N., Gate, P., Gouache, D., Charmet, G., Oury, F. X., and Huard, F. (2010). Why are wheat yields stagnating in Europe? A comprehensive data analysis for France. F. Crop. Res. 119, 201–212. doi: 10.1016/j.fcr.2010.07.012 Evans, D. E. (2003). Aerenchyma formation. New Phytol. 161, 35–49. Falconer, D. S. (1952). The problem of environment. Am. Nat. 86, 293–298 Burton, A. L., Johnson, J., Foerster, J., Hanlon, M. T., Kaeppler, S. M., Lynch, J. P., et al. (2014a). QTL mapping and phenotypic variation of root anatomical traits in maize (Zea mays L.). Theor. Appl. Genet. 128, 93–106. doi: 10.1007/s00122- 014-2414-8 Feng, H., Yan, M., Fan, X., Li, B., Shen, Q., Miller, A. J., et al. (2011). Spatial expression and regulation of rice high-aﬃnity nitrate transporters by nitrogen and carbon status. J. Exp. Bot. 62, 2319–2332. doi: 10.1093/jxb/erq403 Foy, C. D., Fleming, A. L., and Armiger, W. H. (1969). Aluminum tolerance of soybean varieties in relation to calcium nutrition. Agron. J. 61, 505. doi: 10.2134/ agronj1969.00021962006100040007x Burton, A. L., Johnson, J. M., Foerster, J. M., Hirsch, C. N., Buell, C. R., Kaeppler, S. M., et al. (2014b). QTL mapping and phenotypic variation for root architectural traits in maize (Zea mays L.). Theor. Appl. Genet. 127, 2293–2311. doi: 10.1007/s00122-014-2353-4 Gage, J. L., Jarquin, D., Romay, C., Lorenz, A., Buckler, E. S., Kaeppler, S., et al. (2017). The eﬀect of artiﬁcial selection on phenotypic plasticity in maize. Nat. Commun. 8, 1–11. doi: 10.1038/s41467-017-01450-2 Caldwell, M. M., Dudley, L. M., and Lilieholm, B. (1992). Soil solution phosphate, root uptake kinetics and nutrient acquisition: implications for a patchy soil environment. Oecologia 89, 305–309. doi: 10.1007/BF00317406 Galindo-Castañeda, T., Brown, K. M., and Lynch, J. P. (2018). Reduced root cortical burden improves growth and grain yield under low phosphorus availability in maize. Plant Cell Environ. 7, 1579–1592. doi: 10.1111/pce.13197 Chimungu, J., Brown, K., and Lynch, J. (2014a). Large root cortical cell size improves drought tolerance in maize. Plant Physiol. 166, 2166–2178. doi: 10. 1104/pp.114.250449 Gao, Y., and Lynch, J. P. (2016). Reduced crown root number improves water acquisition under water deﬁcit stress in maize (Zea mays L.). J. Exp. Bot. 67, 4545–4587. doi: 10.1093/jxb/erw243 Chimungu, J. G., Brown, K. M., and Lynch, J. P. (2014b). Reduced root cortical cell ﬁle number improves drought tolerance in maize. Plant Physiol. 166, 1943–1955. doi: 10.1104/pp.114.249037 Gardner, W. K., Barber, D. A., and Parbery, D. G. (1983). REFERENCES Dewitt, T. J., Sih, A., and Wilson, D. S. (1998). Costs and limits of phenotypic plasticity. Tree 13, 77–81. Anderson, J., Lee, C.-R., Rushworth, C., Colautti, R., and Mitchell-Olds, T. (2017). Genetic trade-oﬀs and conditional neutrality contribute to local adaptation. PLoS ONE 32:e0178059. doi: 10.1371/journal.pone.0178059 Drew, M. (1975). Comparison of the eﬀects of a localized supply of phosphate, nitrate, ammonium and potassium on the growth of the seminal root system, and the shoot, in barley. New Phytol. 75, 479–490. Drew, M. C., He, C. J., and Morgan, P. W. (2000). Programmed cell death and aerenchyma formation in roots. Trends Plant Sci. 5, 123–127. doi: 10.1016/ S1360-1385(00)01570-3 Bao, Y., Aggarwal, P., Robbins, N. E., Sturrock, C. J., Thompson, M. C., Tan, H. Q., et al. (2014). Plant roots use a patterning mechanism to position lateral root branches toward available water. Proc. Natl. Acad. Sci. U.S.A. 111, 9319–9324. doi: 10.1073/pnas.1400966111 Drew, M. C., Saker, L. R., and Ashley, T. W. (1975). Nutrient supply and the growth of the seminal root system in barley. J. Exp. Bot. 24, 1189–1202. doi: 10.1093/jxb/24.6.1189 Basford, K. E., and Cooper, M. (1998). Genotype x environment interactions and some considerations of their implications for wheat breeding in Australia. Aust. J. Agric. Res. 49, 153–174. Dunbabin, V. M., Postma, J. A., Schnepf, A., Pagès, L., Javaux, M., Wu, L., et al. (2013). Modelling root-soil interactions using three-dimensional models of root growth, architecture and function. Plant Soil 372, 93–124. doi: 10.1007/s11104- 013-1769-y Bloom, A. J., Chapin, F. S., and Mooney, H. A. (1985). Resource limitation in plants-an economic analogy. Annu. Rev. Ecol. Syst. 16, 363–392. growth, architecture and function. Plant Soil 372, 93–124. doi: 10.1007/s11104- 013-1769-y Bonser, A., Lynch, J., and Snapp, S. (1996). Eﬀect of phosphorus deﬁciency on growth angle of basal roots in Phaseolus vulgaris. New Phytol. 132, 281–288. Ehdaie, B., Layne, A. P., and Waines, J. G. (2012). Root system plasticity to drought inﬂuences grain yield in bread wheat. Euphytica 186, 219–232. doi: 10.1007/ s10681-011-0585-9 Bradshaw, A. (2006). Unraveling phenotypic plasticity – Why should we bother? New Phytol. 170, 644–648. El-Soda, M., Kruijer, W., Malosetti, M., Koornneef, M., and Aarts, M. G. M. (2015). Quantitative trait loci and candidate genes underlying genotype by environment interaction in the response of Arabidopsis thaliana to drought. Plant Cell Environ. 38, 585–599. doi: 10.1111/pce.12418 Bradshaw, A., and Hardwick, K. (1989). Evolution and stress—genotypic and phenotypic components. Biol. J. Linn. Soc. REFERENCES The acquisition of phosphorus by Lupinus albus L. – III. The probable mechanism by which phosphorus movement in the soil/root interface is enhanced. Plant Soil 70, 107–124. doi: 10.1007/BF02374754 Chimungu, J. G., Maliro, M. F. A., Nalivata, P. C., Kanyama-Phiri, G., Brown, K. M., and Lynch, J. P. (2015). Utility of root cortical aerenchyma under water limited conditions in tropical maize (Zea mays L.). F. Crop. Res. 171, 86–98. doi: 10.1016/j.fcr.2014.10.009 Giﬀord, M. L., Banta, J. A., Katari, M. S., Hulsmans, J., Chen, L., Ristova, D., et al. (2013). Plasticity regulators modulate speciﬁc root traits in discrete nitrogen environments. PLoS Genet. 9:e1003760. doi: 10.1371/journal.pgen.1003760 Chinnusamy, V., and Zhu, J. K. (2009). Epigenetic regulation of stress responses in plants. Curr. Opin. Plant Biol. 12, 133–139. doi: 10.1016/j.pbi.2008. 12.006 Gowda, V. R. P., Henry, A., Yamauchi, A., Shashidhar, H. E., and Serraj, R. (2011). Root biology and genetic improvement for drought avoidance in rice. F. Crop. Res. 122, 1–13. doi: 10.1016/j.fcr.2011.03.001 Cooper, M., Rajatasereekul, S., Somrith, B., Sriwisut, S., Immark, S., Boonwite, C., et al. (1999). Rainfed lowland rice breeding strategies for Northeast Thailand II. Comparison of intrastation and interstation selection. F. Crop. Res. 64, 153–176. doi: 10.1016/S0378-4290(99)00057-X Hazman, M., and Brown, K. M. (2018). Progressive drought alters architectural and anatomical traits of rice roots. Rice 11:62. doi: 10.1186/s12284-018-0252-z Hecht, V. L., Temperton, V. M., Nagel, K. A., Rascher, U., and Postma, J. A. (2016). Sowing density: a neglected factor fundamentally aﬀecting root distribution and biomass allocation of ﬁeld grown spring barley (Hordeum Vulgare L.). Front. Plant Sci. 7:944. doi: 10.3389/fpls.2016.00944 Correa, J., Postma, J. A., Watt, M., and Wojciechowski, T. (2019). Root system architectural plasticity and soil compaction: a review. J. Exp. Bot. 70, 6019–6034. doi: 10.1093/jxb/erz383 Dathe, A., Postma, J., Postma-Blaauw, M., and Lynch, J. (2016). Impact of axial root growth angles on nitrogen acquisition in maize depends on environmental conditions. Ann. Bot. 118, 401–414. doi: 10.1093/aob/mcw112 Henry, A., Gowda, V. R. P., Torres, R. O., Mcnally, K. L., and Serraj, R. (2011). Variation in root system architecture and drought response in rice (Oryza sativa): phenotyping of the OryzaSNP panel in rainfed lowland ﬁelds. F. Crop. Res. 120, 205–214. doi: 10.1016/j.fcr.2010.10.003 conditions. Ann. Bot. 118, 401–414. doi: 10.1093/aob/mcw112 DeWitt, T. (1998). Costs and limits of phenotypic plasticity: tests with predator- induced morphology and life history in a freshwater snail. J. Evol. Biol. 11, 465–480. AUTHOR CONTRIBUTIONS JL conceived the idea. JL and HS contributed to the writing of this manuscript. May 2020 \| Volume 11 \| Article 546 12 Plasticity as a Breeding Target Schneider and Lynch We thank Kathleen Brown, Meredith Hanlon, and Christopher Strock for helpful comments. We thank Kathleen Brown, Meredith Hanlon, and Christopher Strock for helpful comments. We thank Kathleen Brown, Meredith Hanlon, and Christopher Strock for helpful comments. FUNDING This work was supported by USDOE ARPA-E Award Number DE-AR0000821. REFERENCES Opportunities and challenges in the subsoil: pathways to deeper rooted crops. J. Exp. Bot. 66, 2199–2210. doi: 10.1093/jxb/eru508 Kano, M., Inukai, Y., Kitano, H., and Yamauchi, A. (2011). Root plasticity as the key root trait for adaptation to various intensities of drought stress in rice. Plant Soil 342, 117–128. doi: 10.1007/s11104-010-0675-9 Mano, Y., Omori, F., Takamizo, T., Kindiger, B., Bird, R. M. K., and Loaisiga, C. H. (2006). Variation for root aerenchyma formation in ﬂooded and non-ﬂooded maize and teosinte seedlings. Plant Soil 281, 269–279. doi: 10.1007/s11104-005- 4268-y Kano-Nakata, M., Gowda, V. R. P., Henry, A., Serraj, R., Inukai, Y., Fujita, D., et al. (2013). Functional roles of the plasticity of root system development in biomass production and water uptake under rainfed lowland conditions. F. Crop. Res. 144, 288–296. doi: 10.1016/j.fcr.2013.01.024 Mano, Y., Omori, F., Takamizo, T., Kindiger, B., Bird, R. M., Loaisiga, C. H., and Takahashi, H. (2007). QTL mapping of root aerenchyma formation in seedlings of a maize × rare teosinte “Zea nicaraguensis” cross. Plant Soil 295, 103–113. doi: 10.1007/s11104-007-9266-9 Kano-Nakata, M., Inukai, Y., Wade, L. J., Siopongco, J. D., and Yamauchi, A. (2011). Root development, water uptake, and shoot dry matter production under water deﬁcit conditions in two CSSLs of rice: functional roles of root plasticity. Plant Prod. Sci. 14, 307–317. doi: 10.1626/pps.14.307 Manschadi, A. M., Christopher, J., deVoil, P., and Hammer, G. L. (2006). The role of root architectural traits in adaptation of wheat to water-limited environments. Funct. Plant Biol. 33, 823. doi: 10.1071/FP06055 Kirk, G. J. D., Santos, E. E., and Santos, M. B. (1999). Phosphate solubilization by organic anion excretion from rice growing in aerobic soil: rates of excretion and decomposition, eﬀects on rhizosphere pH and eﬀects on phosphate solubility and uptake. New Phytol. 142, 185–200. doi: 10.1046/j.1469-8137.1999.00400.x Marschner, H. (1995). Mineral Nutrition of Higher Plants, 2nd Edn. San Diego, CA: Academic. Merilä, J., Laurila, A., and Lindgren, B. (2004). Variation in the degree and costs of adaptive phenotypic plasticity among Rana temporaria populations. J. Evol. Biol. 17, 1132–1140. doi: 10.1111/j.1420-9101.2004. 00744.x Lacey, E. P., Smith, S., and Case, A. L. (1997). Parental eﬀects on seed mass: seed coat but not embryo/endosperm eﬀects. Am. J. Bot. 84, 1617–1620. doi: 10.2307/2446624 Lambers, H., Clements, J. C., and Nelson, M. N. (2013). How a phosphorus- acquisition strategy based on carboxylate exudation powers the success and agronomic potential of lupines (Lupinus. Fabaceae). Am. J. Bot. 100, 263–288. REFERENCES doi: 10.1046/j.1420-9101.1998.11040465.x Hirel, B., Le Gouis, J., Ney, B., and Gallais, A. (2007). The challenge of improving nitrogen use eﬃciency in crop plants: towards a more central role for genetic May 2020 \| Volume 11 \| Article 546 Frontiers in Plant Science \| www.frontiersin.org 13 Schneider and Lynch Plasticity as a Breeding Target variability and quantitative genetics within integrated approaches. J. Exp. Bot. 58, 2369–2387. doi: 10.1093/jxb/erm097 Lynch, J. P. (2011). Root phenes for enhanced soil exploration and phosphorus acquisition: tools for future crops. Plant Physiol. 156, 1041–1049. doi: 10.1104/ pp.111.175414 variability and quantitative genetics within integrated approaches. J. Exp. Bot. 58, 2369–2387. doi: 10.1093/jxb/erm097 Ho, M. D., Rosas, J. C., Brown, K. M., and Lynch, J. P. (2005). Root architectural tradeoﬀs for water and phosphorus acquisition. Funct. Plant Biol. 32, 737–748. doi: 10.1071/FP05043 Lynch, J. P. (2015). Root phenes that reduce the metabolic costs of soil exploration: opportunities for 21st century agriculture. Plant Cell Environ. 38, 1775–1784. doi: 10.1111/pce.12451 Jackson, R. B., Manwaring, J. H., and Caldwell, M. M. (1990). Rapid physiological adjustment of roots to localized soil enrichment. Nature 344, 58–60. doi: 10. 1038/344058a0 Lynch, J. P. (2018). Rightsizing root phenotypes for drought resistance. J. Exp. Bot. 69, 3279–3292. Jaramillo, R. E., Nord, E. A., Chimungu, J. G., Brown, K. M., and Lynch, J. P. (2013). Root cortical burden inﬂuences drought tolerance in maize. Ann. Bot. 112, 429–437. doi: 10.1093/aob/mct069 Lynch, J. P. (2019). Root phenotypes for improved nutrient capture: an underexploited opportunity for global agriculture. New Phytol. 223, 548–564. doi: 10.1111/nph.15738 Jia, X., Liu, P., and Lynch, J. P. (2018). Greater lateral root branching density in maize improves phosphorus acquisition from low phosphorus soil. J. Exp. Bot. 69, 4961–4790. doi: 10.1093/jxb/ery252 Lynch, J. P., and Brown, K. M. (2012). New roots for agriculture: exploiting the root phenome. Philos. Trans. R. Soc. Ser. B 367, 1598–1604. doi: 10.1098/rstb.2011. 0243 Lynch, J. P., Chimungu, J. G., and Brown, K. M. (2014). Root anatomical phenes associated with water acquisition from drying soil: targets for crop improvement. J. Exp. Bot. 65, 6155–6166. doi: 10.1093/jxb/eru162 Kadam, N. N., Tamilselvan, A., Lawas, L. M. F. M. F., Quinones, C., Bahuguna, R. N., Thomson, M. J., et al. (2017). Genetic control of plasticity in root morphology and anatomy of rice in response to water-deﬁcit. Plant Physiol. 174, 2302–2315. doi: 10.1104/pp.17.00500 Lynch, J. P., and Wojciechowski, T. (2015). REFERENCES doi: 10.3732/ajb.1200474 Mi, G. H., Chen, F. J., Wu, Q. P., Lai, N. W., Yuan, L. X., and Zhang, F. S. (2010). Ideotype root architecture for eﬃcient nitrogen acquisition by maize in intensive cropping systems. Sci. China Life Sci. 53, 1369–1373. doi: 10.1007/ s11427-010-4097-y j Lambers, H., Juniper, D., Cawthray, G. R., Veneklaas, E. J., and Martínez-Ferri, E. (2003). The pattern of carboxylate exudation in Banksia grandis (Proteaceae) is Lambers, H., Juniper, D., Cawthray, G. R., Veneklaas, E. J., and Martínez-Ferri, E. (2003). The pattern of carboxylate exudation in Banksia grandis (Proteaceae) is aﬀected by the form of phosphate added to the soil. Plant Soil 238, 111–122. Miguel, M., Widrig, A., Vieira, R., Brown, K., and Lynch, J. (2013). Basal root whorl number: a modulator of phosphorus acquisition in common bean (Phaseolus vulgaris). Ann. Bot. 112, 973–982. doi: 10.1093/aob/mct164 Lande, R. (2009). Adaptation to an extraordinary environment by evolution of phenotypic plasticity and genetic assimilation. J. Evol. Biol. 22, 1435–1446. doi: 10.1111/j.1420-9101.2009.01754.x Miguel, M. A., Postma, J. A., and Lynch, J. (2015). Phene synergism between root hair length and basal root growth angle for phosphorus acquisition. Plant Physiol. 167, 1430–1439. doi: 10.1104/pp.15.00145 León, J. A. (1993). “Plasticity in ﬂuctuating environments,” in Adaptation in Stochastic Environments, eds J. Yoshimura and C. W. Clark (Berlin: Springer). Miklos, G. L. G., and Rubin, G. M. (1996). The role of the genome project in determining gene function: insights from model organisms. Cell 86, 521–529. doi: 10.1016/S0092-8674(00)80126-9 Li, X., Guo, Z., Lv, Y., Cen, X., Ding, X., Wu, H., et al. (2017). Genetic control of the root system in rice under normal and drought stress conditions by genome-wide association study. PLoS Genet. 13:1–24. doi: 10.1371/journal. pgen.1006889 Mitchell-Olds, T., Willis, J. H., and Goldstein, D. B. (2007). Which evolutionary processes inﬂuence natural genetic variation for phenotypic traits? Nat. Rev. Genet. 8, 845–856. doi: 10.1038/nrg2207 Liao, H., Rubio, G., Yan, X., Cao, A., Brown, K. M., and Lynch, J. P. (2001). Eﬀect of phosphorus availability on basal root shallowness in common bean. Plant Soil 232, 69–79. Mt. Pleasant, J., and Burt, R. F. (2010). Estimating productivity of traditional iroquoian cropping systems from ﬁeld experiments and historical literature. J. Ethnobiol. 30, 52–79. doi: 10.2993/0278-0771-30.1.52 Lobet, G., Paez-Garcia, A., Schneider, H., Junker, A., Atkinson, J. A., and Tracy, S. (2019). Demystifying roots: a need for clariﬁcation and extended concepts in root phenotyping. Plant Sci. 282, 11–13. REFERENCES A., Dathe, A., and Lynch, J. (2014). The optimal lateral root branching density for maize depends on nitrogen and phosphorus availability. Plant Physiol. 166, 590–602. doi: 10.1104/pp.113. 233916 Sharma, S., Demason, D. A., Ehdaie, B., Lukaszewski, A. J., and Waines, J. G. (2010). Dosage eﬀect of the short arm of chromosome 1 of rye on root morphology and anatomy in bread wheat. J. Exp. Bot. 61, 2623–2633. doi: 10.1093/jxb/erq097 Postma, J. A., and Lynch, J. P. (2011). Theoretical evidence for the functional beneﬁt of root cortical aerenchyma in soils with low phosphorus availability. Ann. Bot. 107, 829–841. doi: 10.1093/aob/mcq199 Strock, C. F., Morrow de la Riva, L., and Lynch, J. (2018). Reduction in root secondary growth as a strategy for phosphorus acquisition. Plant Physiol. 176, 691–703. doi: 10.1104/pp.17.01583 Sultan, S. (1996). Phenotypic plasticity for oﬀspring traits in Polygonum Persicaria. Ecol. Monogr. 77, 1791–1807. Postma, J. A., and Lynch, J. P. (2012). Complementarity in root architecture for nutrient uptake in ancient maize/bean and maize/bean/squash polycultures. Ann. Bot. 110, 521–534. doi: 10.1093/aob/mcs082 Sultan, S. E. (2000). Phenotypic plasticity for plant development, function and life history. Trends Plant Sci. 5, 537–542. doi: 10.1016/S1360-1385(00)01797-0 Prince, S. J., Murphy, M., Mutava, R. N., Durnell, L. A., Valliyodan, B., Grover Shannon, J., et al. (2017). Root xylem plasticity to improve water use and yield in water-stressed soybean. J. Exp. Bot. 66, 2027–2036. doi: 10.1093/jxb/erw472 Sun, B., Gao, Y., and Lynch, J. (2018). Large crown root number improves topsoil Sun, B., Gao, Y., and Lynch, J. (2018). Large crown root number improves topsoil foraging and phosphorus acquisition. Plant Physiol. 177, 90–104. doi: 10.1104/ pp.18.00234 foraging and phosphorus acquisition. Plant Physiol. 177, 90–104. doi: 10.1104/ pp.18.00234 Rabbi, S. M. F., Guppy, C. N., Tighe, M. K., Flavel, R. J., and Young, I. M. (2017). Root architectural responses of wheat cultivars to localised phosphorus application are phenotypically similar. J. Plant Nutr. Soil Sci. 180, 169–177. doi: 10.1002/jpln.201600503 Suralta, R. R., Inukai, Y., and Yamauchi, A. (2010). Dry matter production in relation to root plastic development, oxygen transport, and water uptake of rice under transient soil moisture stresses. Plant Soil 332, 87–104. doi: 10.1007/ s11104-009-0275-8 Rangarajan, H., Postma, J. A., and Lynch, J. (2018). Co-optimisation of axial root phenotypes for nitrogen and phosphorus acquisition in common bean. Ann. Bot. 122, 485–499. doi: 10.1093/aob/mcy092 Takahashi, H., Yamauchi, T., Rajhi, I., Nishizawa, N. K., and Nakazono, M. (2015). REFERENCES doi: 10.1016/j.plantsci.2018.09.015 Lobet, G., Paez-Garcia, A., Schneider, H., Junker, A., Atkinson, J. A., and Tracy, S. (2019). Demystifying roots: a need for clariﬁcation and extended concepts in Nakamoto, T. (1993). Eﬀect of soil water content on the gravitropic behavior of nodal roots in maize. Plant Soil 152, 261–267. doi: 10.1007/BF00029096 Negi, S., Ivanchenko, M. G., and Muday, G. K. (2008). Ethylene regulates lateral root formation and auxin transport in Arabidopsis thaliana. Plant J. 55, 175– 187. doi: 10.1111/j.1365-313X.2008.03495.x Lorts, C., Lynch, J. P., and Brown, K. M. (2019). Parental eﬀects and provisioning under drought and low phosphorus stress in common bean. Food Energy Secur. 9:e192. doi: 10.1002/fes3.192 Lynch, J. (2013). Steep, cheap and deep: an ideotype to optimize water and N acquisition by maize root systems. Ann. Bot. 112, 347–357. doi: 10.1093/aob/ mcs293 Nicotra, A. B., Atkin, O. K., Bonser, S. P., Davidson, A. M., Finnegan, E. J., Mathesius, U., et al. (2010). Plant phenotypic plasticity in a changing climate. Trends Plant Sci. 15, 684–692. doi: 10.1016/j.tplants.2010.09.008 Lynch, J., and Brown, K. (2001). Topsoil foraging – An architectural adaptation of plants to low phosphorus availability. Plant Soil 237, 225–237. Nielsen, K. L., Eshel, A., and Lynch, J. P. (2001). The eﬀect of phosphorus availability on the carbon economy of contrasting common bean (Phaseolus vulgaris L.) genotypes. J. Exp. Bot. 52, 329–339. doi: 10.1093/jxb/52. 355.329 Lynch, J. P. (1995). Root architecture and plant productivity. Plant Physiol. 109, 7–13. May 2020 \| Volume 11 \| Article 546 Frontiers in Plant Science \| www.frontiersin.org 14 Schneider and Lynch Plasticity as a Breeding Target Nielsen, K. L., Lynch, J. P., Jablokow, A. G., and Curtis, P. S. (1994). Carbon cost of root systems: an architectural approach. Plant Soil 165, 161–169. doi: 10.1007/BF00009972 Saengwilai, P., Tian, X., and Lynch, J. P. (2014b). Low crown root number enhances nitrogen acquisition from low-nitrogen soils in maize. Plant Physiol. 166, 581–589. doi: 10.1104/pp.113.232603 Niones, J. M., Inukai, Y., Suralta, R. R., and Yamauchi, A. (2015). QTL associated with lateral root plasticity in response to soil moisture ﬂuctuation stress in rice. Plant Soil 391, 63–75. doi: 10.1007/s11104-015-2404-x Sandhu, N., Raman, K. A., Torres, R. O., Audebert, A., Dardou, A., Kumar, A., et al. (2016). Rice root architectural plasticity traits and genetic regions for adaptability to variable cultivation and stress conditions. Plant Physiol. 171, 2562–2576. doi: 10.1104/pp.16.00705 Niones, J. M., Suralta, R. R., Inukai, Y., and Yamauchi, A. (2012). REFERENCES Field evaluation on functional roles of root plastic responses on dry matter production and grain yield of rice under cycles of transient soil moisture stresses using chromosome segment substitution lines. Plant Soil 359, 107–120. doi: 10.1007/s11104-012- 1178-7 Schlichting, C. D. (1986). The evolution of phenotypic plasticity in plants. Annu. Rev. Ecol. Syst. 17, 667–693. Schneider, H., Klein, S., Hanlon, M., Brown, K., Kaeppler, S., and Lynch, J. (2020a). Genetic control of root anatomical plasticity in maize. Plant Genome e20003. doi: 10.1002/tpg2.20003 Niones, J. M., Suralta, R. R., Inukai, Y., and Yamauchi, A. (2013). Roles of root aerenchyma development and its associated QTL in dry matter production under transient moisture stress in rice. Plant Prod. Sci. 16, 205–216. doi: 10. 1626/pps.16.205 Schneider, H., Klein, S., Hanlon, M., Nord, E., Kaeppler, S., Brown, K., et al. (2020b). Genetic control of root architectural plasticity in maize. J. Exp. Bot. eraa084. doi: 10.1093/jxb/eraa084 Pieruschka, R., and Poorter, H. (2012). Phenotyping plants: genes, phenes and machines. Funct. Plant Biol. 39, 813–820. doi: 10.1071/FPv39n11-IN Schneider, H. M., and Lynch, J. P. (2018). Functional implications of root cortical senescence for soil resource capture. Plant Soil 423, 13–26. doi: 10.1007/s11104- 017-3533-1 Pigliucci, M. (2005). Evolution of phenotypic plasticity: where are we going now? Trends Ecol. Evol. 20, 481–486. doi: 10.1016/j.tree.2005.06.001 Schneider, H. M., Postma, J. A., Wojciechowski, T., Kuppe, C., and Lynch, J. (2017a). Root cortical senescence improves barley growth under suboptimal availability of nitrogen, phosphorus, and potassium. Plany Physiol. 174, 2333– 2347. Pigliucci, M., and Schlichting, C. (1995). Ontogenetic reaction norms in Lobelia siphilitica (Lobeliaceae): response to shading. Ecology 76, 2134–2144. Poorter, H., Fiorani, F., Pieruschka, R., Putten, W. H., Van Der, Kleyer, M., et al. (2016). Pampered inside, pestered outside? Diﬀerences and similarities between plants growing in controlled conditions and in the ﬁeld. New Phytol. 122, 838–855. Schneider, H. M., Wojciechowski, T., Postma, J. A., Brown, K. M., Lücke, A., Zeisler, V., et al. (2017b). Root cortical senescence decreases root respiration, nutrient content and radial water and nutrient transport in barley. Plant Cell Environ. 40, 1392–1408. doi: 10.1111/pce.12933 Poot, P., and Lambers, H. (2008). Shallow-soil endemics: adaptive advantages and constraints of a specialized root-system morphology. New Phytol. 178, 371–381. doi: 10.1111/j.1469-8137.2007.02370.x Schneider, H. M., Wojciechowski, T., Postma, J. A., Brown, K. M., and Lynch, J. P. (2018). Ethylene modulates root cortical senescence in barley. Ann. Bot. 28, 95–105. doi: 10.1093/aob/mcy059 Postma, J. REFERENCES Transcript proﬁles in cortical cells of maize primary root during ethylene- induced lysigenous aerenchyma formation under aerobic conditions. Ann. Bot. 115, 879–894. doi: 10.1093/aob/mcv018 Relyea, R. (2002). Costs of phenotypic plasticity. Am. Nat. 159, 272–282. doi: 10.2307/3079078 Robinson, D., Hodge, A., Gri, B. S., and Fitter, A. H. (1999). Plant root proliferation in nitrogen-rich patches confers competitive advantage. Proc. R. Soc. Lond. 266, 431–435. Tardieu, F. (2018). Any trait or trait-related allele can confer drought tolerance: just design the right drought scenario. J. Exp. Bot. 63, 25–31. doi: 10.1093/jxb/err269 Tebaldi, C., and Lobell, D. B. (2008). Towards probabilistic projections of climate change impacts on global crop yields. Geophys. Res. Lett. 41, 1–14. doi: 10.1029/ 2008GL033423 Rostamza, M., Richards, R. A., and Watt, M. (2013). Response of millet and sorghum to a varying water supply around the primary and nodal roots. Ann. Bot. 112, 439–446. doi: 10.1093/aob/mct099 Topp, C. N. (2016). Hope in change: the role of root plasticity in crop yield stability. Plant Physiol. 172, 5–6. doi: 10.1104/pp.16.01257 Ryan, M. H., Tibbett, M., Edmonds-Tibbett, T., Suriyagoda, L. D. B., Lambers, H., Cawthray, G. R., et al. (2012). Carbon trading for phosphorus gain: the balance between rhizosphere carboxylates and arbuscular mycorrhizal symbiosis in plant phosphorus acquisition. Plant Cell Environ. 35, 2170–2180. doi: 10.1111/ j.1365-3040.2012.02547.x Trachsel, S., Kaeppler, S. M., Brown, K. M., and Lynch, J. P. (2013). Maize root growth angles become steeper under low N conditions. F. Crop. Res. 140, 18–31. doi: 10.1016/j.fcr.2012.09.010 Tran, T. T., Kano-Nakata, M., Suralta, R. R., Menge, D., Mitsuya, S., Inukai, Y., et al. (2014). Root plasticity and its functional roles were triggered by water deﬁcit but not by the resulting changes in the forms of soil N in rice. Plant Soil 386, 65–76. doi: 10.1007/s11104-014-2240-4 Ryan, P. R., Delhaize, E., and Randall, P. J. (1995). Characterisation of Al- stimulated eﬄux of malate from the apices of Al-tolerant wheat roots. Planta 196, 103–110. doi: 10.1007/BF00193223 Trewavas, A. J., and Malho, R. (1997). Signal perception and transduction: the origin of the phenotype. Plant Cell 9, 1181–1195. doi: 10.1105/tpc.9.7.1181 Saengwilai, P., Nord, E. A., Chimungu, J. G., Brown, K. M., and Lynch, J. P. (2014a). Root cortical aerenchyma enhances nitrogen acquisition from low- nitrogen soils in maize. Plant Physiol. 166, 726–735. doi: 10.1104/pp.114. 241711 Uga, Y., Okuno, K. and Yano, M. (2008). QTLs underlying natural variation in stele and xylem structures of rice root. 58, 7–14. REFERENCES Plant Soil 270, 299–310. doi: 10.1007/s11104-004- 1697-y Weiner, J. (2004). Allocation, plasticity, and allometry in plants. Perspect. Plant Ecol. Evol. Syst. 6, 207–215. Wilson, D. S., and Yoshimura, J. (1994). On the coexistence of specialists and generalists. Am. Nat. 144, 692–707. Zhu, J., Kaeppler, S. M., and Lynch, J. P. (2005b). Mapping of QTLs for lateral root branching and length in maize (Zea mays L.) under diﬀerential phosphorus supply. Theor. Appl. Genet. 111, 688–695. doi: 10.1007/s00122-005- 2051-3 Woods, J., Williams, A., Hughes, J. K., Black, M., and Murphy, R. (2010). Energy and the food system. Philos. Trans. R. Soc. B Biol. Sci. 365, 2991–3006. doi: 10.1098/rstb.2010.0172 Wulﬀ, R., and Bazzaz, F. (1992). Eﬀect of the parental nutrient regime on growth of the progeny in Abutilon theophrasti (Malvaceae). Am. J. Bot. 79, 1102–1107. Zhu, J., Kaeppler, S. M., and Lynch, J. P. (2005c). Topsoil foraging and phosphorus acquisition eﬃciency in maize (Zea mays). Funct. Plant Biol. 32, 749–762. doi: 10.1071/FP05005 Xie, L., Burridge, J., Klepp, N., Chutoe, C., Miller, J., Saengwilai, P., et al. (2019). “The phenotypic spectrum: identifying whole root architecture types in genotypes of common bean (Phaseolus vulgaris L.),” in Poster at CROPS 2019, Huntsville, AL. Zhu, J., and Lynch, J. P. (2004). The contribution of lateral rooting to phosphorus acquisition eﬃciency in maize (Zea mays) seedlings. Funct. Plant Biol. 31, 949–958. doi: 10.1071/FP 04046 Yang, J. T., Schneider, H. M., Brown, K. M., and Lynch, J. P. (2019). Genotypic variation and nitrogen stress eﬀects on root anatomy in maize are node speciﬁc. J. Exp. Bot. 70, 5311–5325. doi: 10.1093/jxb/erz293 Conﬂict of Interest: The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. Yano, K., and Kume, T. (2005). Root morphological plasticity for heterogeneous phosphorus supply in Zea mays L. Plant Prod. Sci. 8, 427–432. doi: 10.1626/pps. 8.427 Copyright © 2020 Schneider and Lynch. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. REFERENCES doi: 10.1270/jsbbs.58.7 May 2020 \| Volume 11 \| Article 546 Frontiers in Plant Science \| www.frontiersin.org 15 Plasticity as a Breeding Target Schneider and Lynch Uga, Y., Okuno, K. and Yano, M. (2010). Fine mapping of Sta1, a quantitative trait locus determining stele transversal area, on rice chromosome 9. Mol. Breeding 26, 533–538. doi: 10.1007/s11032-010-9450-0 Zargar, S. M., Nagar, P., Deshmukh, R., Nazir, M., Wani, A. A., Masoodi, K. Z., et al. (2017). Aquaporins as potential drought tolerance inducing proteins: toward instigating stress tolerance. J. Proteomics. 169, 233–238. doi: 10.1016/j.jprot. 2017.04.010 Uga, Y., Okuno, K., and Yano, M. (2011). Dro1, a major QTL involved in deep rooting of rice under upland ﬁeld conditions. J. Exp. Bot. 62, 2485–2494. doi: 10.1093/jxb/erq429 Zhan, A., Schneider, H., and Lynch, J. P. (2015). Reduced lateral root branching density improves drought tolerance in maize. Plant Physiol. 168, 1603–1615. doi: 10.1104/pp.15.00187 Via, S., Gomulkiewicz, R., De Jong, G., Scheiner, S., Schlichting, C., and Van Tienderen, P. (1995). Adaptive phenotypic plasticity: consensus and controversy. Tree 10, 212–217. doi: 10.1016/S0169-5347(00)89061-8 Zhang, C., Postma, J. A., York, L. M., and Lynch, J. P. (2014). Root foraging elicits niche complementarity-dependent yield advantage in the ancient “three sisters” (maize/bean/squash) polyculture. Ann. Bot. 114, 1719–1733. doi: 10.1093/aob/ mcu191 Via, S., and Lande, R. (2006). Genotype-environment interaction and the evolution of phenotypic plasticity. Evolution (N. Y). 39:505. doi: 10.2307/2408649 Via, S., and Lande, R. (2006). Genotype-environment interaction and the evolution of phenotypic plasticity. Evolution (N. Y). 39:505. doi: 10.2307/2408649 Zhu, J., Brown, K. M., and Lynch, J. P. (2010a). Root cortical aerenchyma improves the drought tolerance of maize (Zea mays L.). Plant. Cell Environ. 33, 740–749. doi: 10.1111/j.1365-3040.2009.02099.x Walk, T. C., Jaramillo, R., and Lynch, J. P. (2006). Architectural tradeoﬀs between adventitious and basal roots for phosphorus acquisition. Plant Soil 279, 347– 366. doi: 10.1007/s11104-005-0389-6 Wasson, A. P., Richards, R. A., Chatrath, R., Misra, S. C., Prasad, S. V. S., Rebetzke, G. J., et al. (2012). Traits and selection strategies to improve root systems and water uptake in water-limited wheat crops. J. Exp. Bot. 63, 3485–3498. doi: 10.1093/jxb/ers111 Zhu, J., Zhang, C., and Lynch, J. P. (2010b). The utility of phenotypic plasticity of root hair length for phosphorus acquisition. Funct. Plant Biol. 37, 313–322. doi: 10.1071/FP09197 Zhu, J., Kaeppler, S. M., and Lynch, J. P. (2005a). Mapping of QTL controlling root hair length in maize (Zea mays L.) under phosphorus deﬁciency. Frontiers in Plant Science \| www.frontiersin.org REFERENCES Ye, H., Song, L., Chen, H., Valliyodan, B., Cheng, P., Ali, L., et al. (2018). A major natural genetic variation associated with root system architecture and plasticity improves waterlogging tolerance and yield in soybean. Plant Cell Environ. 41, 2169–2182. doi: 10.1111/pce.13190 York, L. M., Nord, E. A., and Lynch, J. P. (2013). Integration of root phenes for soil resource acquisition. Front. Plant Sci. 4:355. doi: 10.3389/fpls.2013.00355 May 2020 \| Volume 11 \| Article 546 Frontiers in Plant Science \| www.frontiersin.org 16
https://openalex.org/W2980100050	https://europepmc.org/articles/pmc6836256?pdf=render	English	null	Starch as a Green Binder for the Formulation of Conducting Glue in Supercapacitors	Polymers	2,019	cc-by	10,047	Received: 18 September 2019; Accepted: 9 October 2019; Published: 11 October 2019 Abstract: This work describes the use of commercially available starch as a binder for the preparation of conductive glue and electrode materials. It is demonstrated that starch can be successfully implemented as a binder in energy storage systems with non-aqueous electrolytes. These devices are characterized by a stable cycle life (for 50,000 cycles) at a nominal voltage of 2.5 V. Moreover, the use of starch-based conductive glue improves the electrochemical performance, especially reducing the internal resistance of the device. Starch-bound electrodes display lower equivalent distributed resistance (EDR) values than electrodes using carboxymethylcellulose (CMC) as the binder. This is due to the noticeably lower pore clogging by starch. An electric double-layer capacitor (EDLC) in organic electrolyte (1 mol L−1 TEABF4 in ACN) at a nominal voltage of 2.5 V can reach a speciﬁc power and energy of 100 kW kg−1 and 12 Wh kg−1, respectively. This study shows that starch-based conductive glues and electrode materials can be incorporated in EDLC systems. Keywords: binder; charge propagation; eco-friendly; energy storage polymers polymers Paweł Je˙zowski 1,* and Przemysław Łukasz Kowalczewski 2 1 Institute of Chemistry and Technical Electrochemistry, Poznan University of Technology, Berdychowo 4, 60-965 Pozna´n, Poland 2 Institute of Food Technology of Plant Origin, Pozna´n University of Life Sciences, 31 Wojska Polskiego St., 60-624 Pozna´n, Poland; przemyslaw.kowalczewski@up.poznan.pl * Correspondence: pawel.jezowski@put.poznan.pl 1 Institute of Chemistry and Technical Electrochemistry, Poznan University of Technology, Berdychowo 4, 60-965 Pozna´n, Poland 2 Institute of Food Technology of Plant Origin, Pozna´n University of Life Sciences, 31 Wojska Polskiego St., 60-624 Pozna´n, Poland; przemyslaw.kowalczewski@up.poznan.pl * Correspondence: pawel.jezowski@put.poznan.pl Starch as a Green Binder for the Formulation of Conducting Glue in Supercapacitors Paweł Je˙zowski 1,* and Przemysław Łukasz Kowalczewski 2 Keywords: binder; charge propagation; eco-friendly; energy storage 1. Introduction Electrochemical energy storage devices are composed of two electrodes separated by a porous membrane and soaked in electrolyte. In particular, for electric double-layer capacitors (EDLCs), the energy is stored in an electric double layer (EDL) formed at the electrode/electrolyte interface. Cations and anions in electrolytic solution are attracted to the oppositely charged electrode surfaces, which creates an electric potential diﬀerence (voltage). The higher the nominal voltage is (the diﬀerence in measured potential between the positive and negative electrode during cell operation), the higher the energy stored (1) and power density (2), according to Equations [1–3]: E = C/2·U2 (1) P = U2/(4·ERS) (2) (1) P = U2/(4·ERS) (2) (2) where C is the capacitance (F), U is the nominal voltage (V), and ERS stands for the equivalent series, which represents all of the factors contributing to the overall value of resistance (Ω). where C is the capacitance (F), U is the nominal voltage (V), and ERS stands for the equivalent series, which represents all of the factors contributing to the overall value of resistance (Ω). To reach high values of energy and power, it is necessary to reach a high nominal voltage and lower the resistance of the cell. Nonetheless, the capacitance part of the equation should not be omitted. The capacitance value increases the overall energy of the constructed device and strongly depends on the electrode materials used and ﬁnal cell construction, according to the equation provided by Helmholtz [4] (3): , 11, 1648; doi:10.3390/polym11101648 www.mdpi.com/journal/polymers Polymers 2019, 11, 1648; doi:10.3390/polym11101648 www.mdpi.com/journal/polymers www.mdpi.com/journal/polymers Polymers 2019, 11, 1648 2 of 12 C = (ε·S)/d (3) (3) C = (ε·S)/d where ε is the electric permittivity (F m−1), S is the accessible surface area [m2], d is the thickness of the EDL (m), and the accessible surface area S is strictly related to the porosity of the electrode material. Equation (3) was developed later by Stern [5], who took into account that the diﬀusion coeﬃcient at the electrolyte/electrode interphase caused by the agglomeration of diﬀerently charged species (ions) near the electrode surface. Calculation of capacitance value for supercapacitors according to the equation (3) has informative value. The theoretical values of electric permittivity found in the literature may diﬀer from the practical values. The accessible surface for ions is not the same as the value of speciﬁc surface area (SSA) calculated from gas adsorption isotherms [6]. 1. Introduction Moreover, the penetration of pores plays important role especially in the organic electrolytes due to the sieving eﬀect [7]. Additional research and studies were done in this topic to analyze how the capacitance can be theoretically calculated and some mathematical models were proposed in the recent years by Nagy et al. [8] and Huang et al. [9], where the properties of the electrode/electrolyte interphase as well as the material porosity were taken into account. However, for the practical reasons the formula (4) is used to calculate the capacitance [3]: C = Q/U = (I·t)/U (4) (4) where Q is the charge accumulated (C), I is the current used during charging (A), t is the time of the charging process (s), and U is voltage (V). Purely physical energy storage mechanisms are very fast and allow high values of power close to 10 kW kg−1 to be achieved [10,11]. To further improve such power values, research has synthesized and used diﬀerent materials such as activated carbons, carbon cloths, carbon ﬁbers, and other active materials due to their large surface area. However, electrode materials are composed not only of an active material but also a material for increasing the conductivity and a substance binding all components together. Recent reports have described the fabrication of electrode materials with desired properties, such as certain pore sizes and distributions thereof [12–18], or with the use of specially designed electrolytes [19–25]. Another approach is to increase the energy by so-called internal hybridization, where a cell is composed of two diﬀerent electrodes, where one electrode exhibits battery-like behaviour, while the second electrode accumulates charge in an EDL [26–31]. To date, however, the impact of the binder has been omitted in the majority of these studies. There are only a few reports concerning the use of novel binders that can be categorized as “green” materials, which have a low impact on the natural environment. Some examples include cellulose [32], carboxymethylcellulose (CMC) [33,34], chitin/chitosan [35,36] and casein [37], just to mention a few. Nevertheless, there is a possibility of preparing so-called binder free electrodes; however, in most cases, for the preparation of electrodes, it is necessary to use sophisticated methods, conditions, or equipment [38,39]. In light of recent studies regarding starch-based binders, which present good cohesive properties, we have decided to pursue the interesting topic of “green” components for electrode preparation. 2. Materials and Methods The conducting glue was prepared by mixing 1:1:1 (mass ratio) deionized water (solvent), starch and soot (Super C65, Imerys, France). The mixture was stirred with a mechanical stirrer at 500 RPM at an elevated temperature of 60 ◦C. Once gelation of the mixture was observed, the glue was spread on the surface of a current collector. To evenly distribute the conductive glue, stainless steel foil (S316L, thickness 10 µm, Materialis Elements, Wrocław, Poland) was placed on the perforated table of an automatic ﬁlm applicator with a doctor blade attachment. The current collector sheet covered with conductive glue was transferred to a fume hood, and the excess solvent was evaporated by using a heating plate at 100 ◦C for 30 min. In the absence of the adhesive glue, extremely low charge propagation and high internal resistance were observed (Figure S1 in Supplementary Materials). The current collector sheet prepared for the CMC-based electrode material was covered with a commercially available conductive glue (ElectrodagTM PF-407C, Acheson, Düsseldorf, Germany). The casting and drying method was identical to that described above. The electrode material was prepared according to the following composition: 80 wt. % activated carbon (Kuraray YP 80F, Tokyo, Japan, with characteristic Vmicro < 2 nm = 0.652 and SSA = 2093 m2 g−1), 10 wt. % carbon black (Super C65, Imerys, France) and 10 wt. % starch or carboxymethyl cellulose (referred to as CMC). Distilled water was added in ratios of 1 mL to 1 g of the dry electrode material mixture to decrease the viscosity until reaching a ﬂuid-like consistency. The mixture was stirred at 20 ◦C using a vacuum homogenizer until a homogeneous suspension (slurry) was obtained. In the next step, the slurry was spread on the surface of the stainless steel foil covered, with conducting glue being used as the current collector. The gap between the blade and surface of the current collector was adjusted to 100 µm. The sheet of the current collector covered with the conductive glue and electrode material was transferred to a fume hood, and the excess solvent was evaporated using a heating plate at 100 ◦C for 1 h. Finally, to remove water from the pores of the activated carbon material, the electrode material sheet was dried under vacuum in a dryer (VACUCELL MM44, MMM Medcenter Einrichtungen GmbH, München, Germany) overnight at 100 ◦C. 1. Introduction In one study by Passerini [40], it was emphasized that for starch-based electrodes, the “EIS measurements suggest that the coating process needs to be optimized”, which gives space for improvement in this ﬁeld. One reasonable approach would be to try and modify the contact surface between the current collector and electrode material [41–43]. Nevertheless, the etching of current collectors is diﬃcult, and it is necessary to provide strict and harsh conditions to achieve suitable results. Varzi et al. [40] used an etched current collector however, we suggest a completely diﬀerent approach: an electrode material and conductive glue based on starch, one casted on the other. Starch is a naturally abundant polymer produced by many plants as a source of energy. It is worth noting that starch is the second most abundant biomaterial in nature. On the one hand, starch is environmentally friendly due to its renewable and biodegradable nature; on the other hand, it is inexpensive and widely available [44–46]. Therefore, starch has been used for years not only 3 of 12 Polymers 2019, 11, 1648 as a component of adhesives used for gluing paper but also as a binder in many industries [47]. Nevertheless, the binding capacity of starch in its native form is not strong enough [48]; thus, additives that increase its adhesive properties, such as polyacrylamides, polyvinyl alcohol, or polyvinyl acetate, are often used in the corresponding glue formulations [49–52]. The industrial use of native starch as a binder also limits the diﬃculty of obtaining an adhesive paste containing a suﬃciently high amount of starch (on a dry basis) to create a continuous weld between the bound surfaces [53,54]. In addition, the gelling process, i.e., heating the starch slurry to 60–70 ◦C signiﬁcantly increases the viscosity [55,56], making it diﬃcult to pump the adhesive properly in automated production lines and spread it evenly over the surface to be bound. To reduce the viscosity, controlled depolymerization is performed, and appropriately selected auxiliary agents that modify the adhesive properties of the obtained glue are added [57,58]. The abovementioned reasons were the driving force for this work, which seeks to establish a preparation method for a conducting glue based on starch for energy storage devices, which could improve the charge propagation and power performance of electrochemical devices and compete with conductive glues available on the market. 2. Materials and Methods The dried material was subjected to calendaring to further improve cohesion and adhesion. The ﬁnal thickness of the electrode material was ca. 70 µm. Electrodes in the form of discs with a diameter of 10 mm were cut out using an EL-CUT precision disc cutter (EL-CELL GmbH, Hamburg, Germany), and the measured mass loading for the electrodes was ca. 2 mg cm−1. The scheme of experimental procedure is presented in Figure 1. 4 of 12 4 of 12 Polymers 2019, 11, 1648 Polymers 2019 11 x FO Figure 1 Scheme of the experimental procedure Figure 1. Scheme of the experimental procedure. Figure 1 Scheme of the experimental procedure Figure 1. Scheme of the experimental procedure. g p p The prepared electrodes were analyzed by scanning electron microscopy (SEM) (Quanta 250 FEG, Thermo Fisher Scientific, Waltham, MA,USA) at an accelerating voltage of 5 kV and a probe f 100 A The prepared electrodes were analyzed by scanning electron microscopy (SEM) (Quanta 250 FEG, Thermo Fisher Scientiﬁc, Waltham, MA, USA) at an accelerating voltage of 5 kV and a probe current of 100 pA. current of 100 pA. A portion of the electrode material was used for the estimation of the specific surface area (SSA) and pore size distribution (PSD) according to the Brunauer–Emmett–Teller (BET) equation. SSA was estimated using the accelerated surface area and porosimetry system ASAP 2020 (Micromeritics Instrument Co., Norcross, GA, USA). Degassing of the investigated samples was performed for 24 h 50 °C d G d i i d 196 °C i h N h d b A portion of the electrode material was used for the estimation of the speciﬁc surface area (SSA) and pore size distribution (PSD) according to the Brunauer–Emmett–Teller (BET) equation. SSA was estimated using the accelerated surface area and porosimetry system ASAP 2020 (Micromeritics Instrument Co., Norcross, GA, USA). Degassing of the investigated samples was performed for 24 h at 50 ◦C under vacuum. Gas adsorption was carried out at −196 ◦C with N2 as the adsorbate. at 50 °C under vacuum. Gas adsorption was carried out at −196 °C with N2 as the adsorbate. Swagelok cells were assembled in a glove box (UNILab 1200/780, M. Braun Inertgas-Systeme GmbH, Garching, Germany) with an oxygen and water level below 1 ppm to determine the capacitance value. Electrodes with the same diameter and similar masses were placed in a Swagelok cell. 2. Materials and Methods Both electrodes were separated by a porous glass fiber separator (GF/A, Whatman, Darmstadt, Germany, thickness 270 μm), and then electrolyte, one molar solution of tetraethylammonium tetrafluoroborate dissolved in acetonitrile (1 mol L−1 TEBF4 in ACN, 250 μL), was introduced into the system. A VMP3 potentiostat/galvanostat (Bio-Logic, France) was used for electrochemical testing. The techniques included galvanostatic cycling with potential limitation (GCPL) with a current density up to 50 A g−1 (the value was based on the total weight of the electrode material) in the voltage range from 0 to 2.5 V and electrochemical impedance spectroscopy (EIS) in the frequency range from 100 kH 10 H i h li d f 5 V Swagelok cells were assembled in a glove box (UNILab 1200/780, M. Braun Inertgas-Systeme GmbH, Garching, Germany) with an oxygen and water level below 1 ppm to determine the capacitance value. Electrodes with the same diameter and similar masses were placed in a Swagelok cell. Both electrodes were separated by a porous glass ﬁber separator (GF/A, Whatman, Darmstadt, Germany, thickness 270 µm), and then electrolyte, one molar solution of tetraethylammonium tetraﬂuoroborate dissolved in acetonitrile (1 mol L−1 TEBF4 in ACN, 250 µL), was introduced into the system. A VMP3 potentiostat/galvanostat (Bio-Logic, France) was used for electrochemical testing. The techniques included galvanostatic cycling with potential limitation (GCPL) with a current density up to 50 A g−1 (the value was based on the total weight of the electrode material) in the voltage range from 0 to 2.5 V and electrochemical impedance spectroscopy (EIS) in the frequency range from 100 kHz to 10 mHz with an amplitude of 5 mV. 100 kHz to 10 mHz with an amplitude of 5 mV. A VMP3 potentiostat/galvanostat (Bio-Logic Science Instruments, Seyssinet-Pariset, France) was used for electrochemical testing. GCPL was performed with a current density of C/20 and C/2 in the potential range from 10 mV to 1.5 V vs. Li0/Li+ (where C corresponds to the theoretical capacity of graphite i e 372 mAh g−1) A VMP3 potentiostat/galvanostat (Bio-Logic Science Instruments, Seyssinet-Pariset, France) was used for electrochemical testing. GCPL was performed with a current density of C/20 and C/2 in the potential range from 10 mV to 1.5 V vs. Li0/Li+ (where C corresponds to the theoretical capacity of graphite, i.e., 372 mAh g−1). 3 R lt d Di i 3. Results and Discussion 3. Results and Discussion The gas adsorption analysis data presented in Figure 2 shows that the highest SSA was observed for the starch-based electrode material (1781 m2 g−1, solid green line) compared to that of polytetrafluoroethylene (PTFE) (1651 m2 g−1, dashed orange line), and the lowest value was detected for the CMC (1463 m2 g−1, dotted red line) bound electrode material. Based on the PSD data (Figure 2b), it seems that starch does not block micropores to the same extent as the PTFE or CMC binders. The starch-based electrode shows a cumulative micropore volume of Vmicro < 2 nm = 0.688, while that for the PTFE-bound electrode material is smaller, with a total micropore volume of Vmicro < 2 nm = 0.505; the value is even smaller for the electrode material with CMC (Vmicro < 2 nm = 0.453). The blockage of pores in AC materials by CMC has been reported by Richner et al supporting the The gas adsorption analysis data presented in Figure 2 shows that the highest SSA was observed for the starch-based electrode material (1781 m2 g−1, solid green line) compared to that of polytetraﬂuoroethylene (PTFE) (1651 m2 g−1, dashed orange line), and the lowest value was detected for the CMC (1463 m2 g−1, dotted red line) bound electrode material. Based on the PSD data (Figure 2b), it seems that starch does not block micropores to the same extent as the PTFE or CMC binders. The starch-based electrode shows a cumulative micropore volume of Vmicro < 2 nm = 0.688, while that for the PTFE-bound electrode material is smaller, with a total micropore volume of Vmicro < 2 nm = 0.505; the value is even smaller for the electrode material with CMC (Vmicro < 2 nm = 0.453). The blockage of pores in AC materials by CMC has been reported by Richner et al., supporting the ﬁndings 5 of 12 Polymers 2019, 11, 1648 of this study [59]. Pristine AC powder (Kuraray YP80F) (dashed-dotted black line) has been added to Figure 2 as a reference. Polymers 2019, 11, x FOR PEER REVIEW 5 of 12 olymers 2019, 11, x FOR PEER REVIEW 5 of 12 Figure 2 as a reference. Polymers 2019, 11, x FOR PEER REVIEW 5 of Figure 2. 3 R lt d Di i 3. Results and Discussion nd (b) pore size distribution (PSD) curves for pristine activated carbon (Kuraray YP 80F) (dashed- dotted black line) and electrodes with different binders: starch (solid green line), PTFE (dashed range line) and CMC (dotted red line). Composition of electrode material 80 wt. % activated carbon YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % binder. h E d h h l f h l d l h h Figure 2. Gas adsorption data: (a) adsorption/desorption Brunauer-Emmett-Teller (BET) isotherms and (b) pore size distribution (PSD) curves for pristine activated carbon (Kuraray YP 80F) (dashed- dotted black line) and electrodes with different binders: starch (solid green line), PTFE (dashed orange line) and CMC (dotted red line). Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % binder. Figure 2. Gas adsorption data: (a) adsorption/desorption Brunauer-Emmett-Teller (BET) isotherms and (b) pore size distribution (PSD) curves for pristine activated carbon (Kuraray YP 80F) (dashed-dotted black line) and electrodes with diﬀerent binders: starch (solid green line), PTFE (dashed orange line) and CMC (dotted red line). Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % binder. nd (b) pore size distribution (PSD) curves for pristine activated carbon (Kuraray YP 80F) (dashed- otted black line) and electrodes with different binders: starch (solid green line), PTFE (dashed range line) and CMC (dotted red line). Composition of electrode material 80 wt. % activated carbon YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % binder. The SEM images demonstrate the morphologies of the electrode materials with starch (Figures 3a,b), PTFE (Figures 3c,d) and CMC (Figures 3e,f) as binders. In the case of starch and CMC, the morphology of electrode materials is very similar. For PTFE we observed a “spider-web” connections, where threads of PTFE link and bind material particles. In all of the cases, the obtained electrodes were homogenous and the soot particles were well dispersed among the activated carbon The SEM images demonstrate the morphologies of the electrode materials with starch (Figure 3a,b), PTFE (Figure 3c,d) and CMC (Figure 3e,f) as binders. In the case of starch and CMC, the morphology of electrode materials is very similar. For PTFE we observed a “spider-web” connections, where threads of PTFE link and bind material particles. 3 R lt d Di i 3. Results and Discussion Gas adsorption data: (a) adsorption/desorption Brunauer-Emmett-Teller (BET) isotherms and (b) pore size distribution (PSD) curves for pristine activated carbon (Kuraray YP 80F) (dashed- dotted black line) and electrodes with different binders: starch (solid green line), PTFE (dashed orange line) and CMC (dotted red line). Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % binder. Figure 2. Gas adsorption data: (a) adsorption/desorption Brunauer-Emmett-Teller (BET) isotherms and (b) pore size distribution (PSD) curves for pristine activated carbon (Kuraray YP 80F) (dashed-dotted black line) and electrodes with diﬀerent binders: starch (solid green line), PTFE (dashed orange line) and CMC (dotted red line). Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % binder. Figure 2. Gas adsorption data: (a) adsorption/desorption Brunauer-Emmett-Teller (BET) isotherms and (b) pore size distribution (PSD) curves for pristine activated carbon (Kuraray YP 80F) (dashed- dotted black line) and electrodes with different binders: starch (solid green line), PTFE (dashed orange line) and CMC (dotted red line). Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % binder. The SEM images demonstrate the morphologies of the electrode materials with starch (Figures Figure 2 Gas adsorption data: (a) adsorption/desorpt ion Brunauer-Emmett-Teller (BET) isotherms tt T ll (BET) i Figure 2. Gas adsorption data: (a) adsorption/desorption Brunauer-Emmett-Teller (BET) isotherms and (b) pore size distribution (PSD) curves for pristine activated carbon (Kuraray YP 80F) (dashed- dotted black line) and electrodes with different binders: starch (solid green line), PTFE (dashed orange line) and CMC (dotted red line). Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % binder. Figure 2. Gas adsorption data: (a) adsorption/desorption Brunauer-Emmett-Teller (BET) isotherms and (b) pore size distribution (PSD) curves for pristine activated carbon (Kuraray YP 80F) (dashed-dotted black line) and electrodes with diﬀerent binders: starch (solid green line), PTFE (dashed orange line) and CMC (dotted red line). Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % binder. 3 R lt d Di i 3. Results and Discussion For the higher current ensities for example 50 A g−1 the observable ohmic drop (IR drop) is related with the internal esistance of the cell which increases when the value of passing current increases [60–62]. Moreover, n Figure S3 (in Supplementary Materials), we present additional data obtained for starch-bound ectrodes operating in aqueous electrolyte (1 mol L−1 Na2SO4). However, this subject still needs dditional work regarding the use of different types of starch and their modifications, which could revent the dissolution of starch in water-based electrolytes for EDLCs during electrochemical Figure 4 presents charging/discharging voltage proﬁles for EDLCs with starch-bound electrode materials coated on the surface of starch-based conductive glue. The triangular shape of these proﬁles and coulombic eﬃciency near 100% indicate no parasitic reactions in the presented system. For the lower current densities (0.5 A g−1), the lower eﬃciency at the initial cycles can be caused by the functional groups on surface of activated carbon. However, after several cycles it reaches the columbic eﬃciency of nearly 100% (Figure S2 in Supplementary Materials). For the higher current densities for example 50 A g−1 the observable ohmic drop (IR drop) is related with the internal resistance of the cell which increases when the value of passing current increases [60–62]. Moreover, in Figure S3 (in Supplementary Materials), we present additional data obtained for starch-bound electrodes operating in aqueous electrolyte (1 mol L−1 Na2SO4). However, this subject still needs additional work regarding the use of diﬀerent types of starch and their modiﬁcations, which could prevent the dissolution of starch in water-based electrolytes for EDLCs during electrochemical cycling. ycling. To establish the long-term stability of the electrode materials with starch binders, cells were continuously charged and discharged at a current of 25 A g−1 in a voltage range from 0.0 to 2.5 V. The table capacitance value (green solid line) during 50,000 cycles (capacitance retention of 95% compared to the initial value before cycling) and high coulombic eﬃciency of 98%, as shown in Figure 5, suggest that starch can be used as a binder for electrode materials and as a conductive glue in EDLCs with long lifespans. g p Nyquist plots (Figure 6) show that the EDLC constructed with conductive glue and electrodes incorporating the starch binder has an equivalent distributed resistance (EDR) = 0.9, which is lower than for CMC-bound electrodes, where the EDR value is 2.2 Ω. 3 R lt d Di i 3. Results and Discussion In all of the cases, the obtained electrodes were homogenous and the soot particles were well dispersed among the activated carbon particles. The SEM images demonstrate the morphologies of the electrode materials with starch (Figures a,b), PTFE (Figures 3c,d) and CMC (Figures 3e,f) as binders. In the case of starch and CMC, the morphology of electrode materials is very similar. For PTFE we observed a “spider-web” onnections, where threads of PTFE link and bind material particles. In all of the cases, the obtained lectrodes were homogenous and the soot particles were well dispersed among the activated carbon articles. articles. Figure 3. Cont. Figure 3. Cont. 6 of 12 Polymers 2019, 11, 1648 Figure 3. SEM images of electrodes with different binders: starch (a,b), PTFE (c,d) and CMC (e,f), at different magnifications (scale added to the images). Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % binder. Figure 3. SEM images of electrodes with diﬀerent binders: starch (a,b), PTFE (c,d) and CMC (e,f), at diﬀerent magniﬁcations (scale added to the images). Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % binder. gure 3. SEM images of electrodes with different binders: starch (a,b), PTFE (c,d) and CMC (e,f), at fferent magnifications (scale added to the images). Composition of electrode material 80 wt. % tivated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % binder. Figure 3. SEM images of electrodes with diﬀerent binders: starch (a,b), PTFE (c,d) and CMC (e,f), at diﬀerent magniﬁcations (scale added to the images). Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % binder. Figure 4 presents charging/discharging voltage profiles for EDLCs with starch-bound electrode materials coated on the surface of starch-based conductive glue. The triangular shape of these profiles nd coulombic efficiency near 100% indicate no parasitic reactions in the presented system. For the ower current densities (0.5 A g−1), the lower efficiency at the initial cycles can be caused by the unctional groups on surface of activated carbon. However, after several cycles it reaches the olumbic efficiency of nearly 100% (Figure S2 in Supplementary Materials). 3 R lt d Di i 3. Results and Discussion T ble capacitance value (green solid line) during 50,000 cycles (capacitance retention of 95% compar the initial value before cycling) and high coulombic efficiency of 98%, as shown in Figure 5, sugge at starch can be used as a binder for electrode materials and as a conductive glue in EDLCs wi ng lifespans. Figure 5. Cyclability of EDLC cells with starch binder in the voltage range from 0.0 to 2.5 V at a current density of 25 A g−1 and 1 mol L−1 TEABF4 in ACN used as the electrolyte. Capacitance values and columbic efficiency are represented as a solid green line and dotted orange line, respectively. Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) d 10 t % t h Figure 5. Cyclability of EDLC cells with starch binder in the voltage range from 0.0 to 2.5 V at a current density of 25 A g−1 and 1 mol L−1 TEABF4 in ACN used as the electrolyte. Capacitance values and columbic eﬃciency are represented as a solid green line and dotted orange line, respectively. Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % starch. ng. Figure 4. Charge/discharge proﬁles in the voltage range 0.0 to 2.5 V for EDLCs operating at diﬀerent current densities: (a) 0.5 A g−1, (b) 2.5 A g−1, (c) 10 A g−1 and (d) 50 A g−1, with 1 mol L−1 TEABF4 in ACN as the electrolyte. Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % starch. lymers 2019, 11, x FOR PEER REVIEW 7 of 1 Figure 4. Charge/discharge profiles in the voltage range 0.0 to 2.5 V for EDLCs operating at different current densities: (a) 0.5 A g−1, (b) 2.5 A g−1, (c) 10 A g−1 and (d) 50 A g−1, with 1 mol L−1 TEABF4 in ACN as the electrolyte. Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % starch. To establish the long-term stability of the electrode materials with starch binders, cells wer ontinuously charged and discharged at a current of 25 A g−1 in a voltage range from 0.0 to 2.5 V. 3 R lt d Di i 3. Results and Discussion Th ble capacitance value (green solid line) during 50,000 cycles (capacitance retention of 95% compare the initial value before cycling) and high coulombic efficiency of 98%, as shown in Figure 5, sugge hat starch can be used as a binder for electrode materials and as a conductive glue in EDLCs wit ng lifespans. ng. ymers 2019, 11, x FOR PEER REVIEW Figure 4. Charge/discharge profiles in the voltage current densities: (a) 0.5 A g−1, (b) 2.5 A g−1, (c) 1 ACN as the electrolyte. Composition of electrode % carbon black (Super C65) and 10 wt. % starch. To establish the long-term stability of the e ntinuously charged and discharged at a curren 7 of ange 0.0 to 2.5 V for EDLCs operating at different A g−1 and (d) 50 A g−1, with 1 mol L−1 TEABF4 in material 80 wt. % activated carbon (YP 80F) 10 wt. ctrode materials with starch binders, cells we f 25 A g−1 in a voltage range from 0.0 to 2.5 V. T ange fr g discha Figure 4. Charge/discharge proﬁles in the voltage range 0.0 to 2.5 V for EDLCs operating at diﬀerent current densities: (a) 0.5 A g−1, (b) 2.5 A g−1, (c) 10 A g−1 and (d) 50 A g−1, with 1 mol L−1 TEABF4 in ACN as the electrolyte. Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % starch. e capacitance value (green solid line) during 50,000 cycles (capacitance retention of 95% compa e initial value before cycling) and high coulombic efficiency of 98%, as shown in Figure 5, sugg starch can be used as a binder for electrode materials and as a conductive glue in EDLCs w lifespans. Figure 5. Cyclability of EDLC cells with starch binder in the voltage range from 0.0 to 2.5 V at a current density of 25 A g−1 and 1 mol L−1 TEABF4 in ACN used as the electrolyte. Capacitance values and columbic efficiency are represented as a solid green line and dotted orange line, respectively. Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % starch. Figure 5. 3 R lt d Di i 3. Results and Discussion The two EDLC systems operating at 2.5 V with starch or CMC as the binder are compared in a Ragone plot (Figure 7). The starch-based EDLC has a higher speciﬁc energy of 20 Wh kg−1 than the EDLC with CMC-bound electrodes (17 Wh kg−1). At high power values (>10 kW kg−1), the energy density of the cell with commercially available conductive glue and CMC binder starts to decrease, most likely because of the higher internal resistance; thus, the speciﬁc energy is 4 Wh kg−1 at a speciﬁc power of 100 kW kg−1. In the case of the EDLC utilizing starch-bound electrodes and conductive glue, it is possible to reach a high speciﬁc energy of 12 Wh kg−1 at a speciﬁc power of 100 kW kg−1. Obtained results were compared with other data found in the literature about symmetric electrochemical capacitors with carbon electrodes. All of the cells operate in organic electrolyte with TEABF4 where salt concentration is 1 mol L−1 [63–66]. It shows that the use starch conductive glue allows us to reach high values of power without a noticeable fade of energy, especially when reaching the value of speciﬁc power 10 kW kg−1. 7 of 12 uld ical Polymers 2019, 11, 1648 dditional work regard revent the dissolutio g. Figure 4. Charge/discharge proﬁles in the voltage range 0.0 to 2.5 V for EDLCs operating at diﬀerent current densities: (a) 0.5 A g−1, (b) 2.5 A g−1, (c) 10 A g−1 and (d) 50 A g−1, with 1 mol L−1 TEABF4 in ACN as the electrolyte. Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % starch. ymers 2019, 11, x FOR PEER REVIEW 7 of Figure 4. Charge/discharge profiles in the voltage range 0.0 to 2.5 V for EDLCs operating at different current densities: (a) 0.5 A g−1, (b) 2.5 A g−1, (c) 10 A g−1 and (d) 50 A g−1, with 1 mol L−1 TEABF4 in ACN as the electrolyte. Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % starch. To establish the long-term stability of the electrode materials with starch binders, cells we ntinuously charged and discharged at a current of 25 A g−1 in a voltage range from 0.0 to 2.5 V. 3 R lt d Di i 3. Results and Discussion Comparative Nyquist plots for EDLCs with starch as an electrode material binder before (solid blue line) and after cycling life studies (dashed green line) and with CMC as a binder (dotted red line). Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % binder. 1 mol L−1 TEABF4 in ACN was used as the electrolyte. Figure 6. Comparative Nyquist plots for EDLCs with starch as an electrode material binder before (solid blue line) and after cycling life studies (dashed green line) and with CMC as a binder (dotted red line). Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % binder. 1 mol L−1 TEABF4 in ACN was used as the electrolyte. Figure 6. Comparative Nyquist plots for EDLCs with starch as an electrode material binder before (solid blue line) and after cycling life studies (dashed green line) and with CMC as a binder (dotted red line). Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % binder. 1 mol L−1 TEABF4 in ACN was used as the electrolyte. Figure 6. Comparative Nyquist plots for EDLCs with starch as an electrode material binder before (solid blue line) and after cycling life studies (dashed green line) and with CMC as a binder (dotted red line). Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % binder. 1 mol L−1 TEABF4 in ACN was used as the electrolyte. Figure 7. Comparative Ragone plot for EDLC cells in which starch (solid green line) or CMC (dashed red line) was used as a binder. Cells were tested up to 2.5 V in 1 mol L−1 TEABF4 in ACN. Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % binder. The reference curves were made according to the data from: [63] (dot-dashed light blue line), [64] (double grey line), [65] (dotted light pink line), [66] (long dashed light orange line and dot-dot-dashed light purple line). Figure 7. Comparative Ragone plot for EDLC cells in which starch (solid green line) or CMC (dashed red line) was used as a binder. 3 R lt d Di i 3. Results and Discussion Cyclability of EDLC cells with starch binder in the voltage range from 0.0 to 2.5 V at a current density of 25 A g−1 and 1 mol L−1 TEABF4 in ACN used as the electrolyte. Capacitance values and columbic eﬃciency are represented as a solid green line and dotted orange line, respectively. Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % starch. Figure 5. Cyclability of EDLC cells with starch binder in the voltage range from 0.0 to 2.5 V at a current density of 25 A g−1 and 1 mol L−1 TEABF4 in ACN used as the electrolyte. Capacitance values and columbic efficiency are represented as a solid green line and dotted orange line, respectively. Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % starch. Figure 5. Cyclability of EDLC cells with starch binder in the voltage range from 0.0 to 2.5 V at a current density of 25 A g−1 and 1 mol L−1 TEABF4 in ACN used as the electrolyte. Capacitance values and columbic eﬃciency are represented as a solid green line and dotted orange line, respectively. Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % starch. 8 of 12 8 of 12 8 of 12 Polymers 2019, 11, 1648 Polymers 2019, 11, x FOR Polymers 2019, 11, x FOR Figure 6. Comparative Nyquist plots for EDLCs with starch as an electrode material binder before (solid blue line) and after cycling life studies (dashed green line) and with CMC as a binder (dotted red line). Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % binder. 1 mol L−1 TEABF4 in ACN was used as the electrolyte. Figure 6. Comparative Nyquist plots for EDLCs with starch as an electrode material binder before (solid blue line) and after cycling life studies (dashed green line) and with CMC as a binder (dotted red line). Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % binder. 1 mol L−1 TEABF4 in ACN was used as the electrolyte. Figure 6. 3 R lt d Di i 3. Results and Discussion Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % binder. The reference curves were made according to the data from: [63] (dot-dashed light blue line), [64] (double grey line), [65] (dotted light pink line), [66] (long dashed light orange line and dot-dot-dashed light purple line). 3 R lt d Di i 3. Results and Discussion Cells were tested up to 2.5 V in 1 mol L−1 TEABF4 in ACN. Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % binder. The reference curves were made according to the data from: [63] (dot-dashed light blue line), [64] (double grey line), [65] (dotted light pink line), [66] (long dashed light orange line and dot-dot-dashed light purple line). Figure 7. Comparative Ragone plot for EDLC cells in which starch (solid green line) or CMC (dashed red line) was used as a binder. Cells were tested up to 2.5 V in 1 mol L−1 TEABF4 in ACN. Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % binder. The reference curves were made according to the data from: [63] (dot-dashed light blue line), [64] (double grey line), [65] (dotted light pink line), [66] (long dashed light orange line and dot-dot-dashed light purple line). Figure 7. Comparative Ragone plot for EDLC cells in which starch (solid green line) or CMC (dashed red line) was used as a binder. Cells were tested up to 2.5 V in 1 mol L−1 TEABF4 in ACN. Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % binder. The reference curves were made according to the data from: [63] (dot-dashed light blue line), [64] (double grey line), [65] (dotted light pink line), [66] (long dashed light orange line and dot dot dashed light purple line) Figure 7. Comparative Ragone plot for EDLC cells in which starch (solid green line) or CMC (dashed red line) was used as a binder. Cells were tested up to 2.5 V in 1 mol L−1 TEABF4 in ACN. Composition of electrode material 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % binder. The reference curves were made according to the data from: [63] (dot-dashed light blue line), [64] (double grey line), [65] (dotted light pink line), [66] (long dashed light orange line and dot-dot-dashed light purple line). Figure 7. Comparative Ragone plot for EDLC cells in which starch (solid green line) or CMC (dashed red line) was used as a binder. Cells were tested up to 2.5 V in 1 mol L−1 TEABF4 in ACN. 4 C l i 4. Conclusions 4. Conclusions 4. Conclusions Starch can be considered as a “green” binder material for the fabrication of conducting glue and electrodes for energy storage devices. The manufacturing process does not involve any toxic or harmful solvents. The electrochemical characteristics of electrode materials incorporating starch as a binder and a layer of starch-based conductive glue provide better results than the commonly used Starch can be considered as a “green” binder material for the fabrication of conducting glue and electrodes for energy storage devices. The manufacturing process does not involve any toxic or harmful solvents. The electrochemical characteristics of electrode materials incorporating starch as a binder and a layer of starch-based conductive glue provide better results than the commonly used Starch can be considered as a “green” binder material for the fabrication of conducting glue and electrodes for energy storage devices. The manufacturing process does not involve any toxic or harmful solvents. The electrochemical characteristics of electrode materials incorporating starch as a binder and a layer of starch-based conductive glue provide better results than the commonly used “green” 9 of 12 Polymers 2019, 11, 1648 binder, CMC. The use of relatively high current loads, such as 50 A g−1, is possible for cells where the electrode material is ﬁxed to the current collector with a starch-based conducting glue, and such cells exhibit better charge propagation, excellent capacitance retention and relatively small resistance. Moreover, preliminary studies have shown that starch binder can be used in aqueous electrolytes, which will be investigated further in the future. Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4360/11/10/1648/s1, Figure S1: Comparative Nyquist plots for EDLC with starch as a binder in conductive glue for improved adhesion of electrode material to current collector (green solid line) and without adhesive conductive layer (red dotted line). Composition of electrodes 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % starch. Figure S2: Cyclability of EDLC cells with starch binder in the voltage range from 0.0 V to 2.5 V at a current density of 0.5 A g−1 and 1 mol L−1 TEABF4 in ACN used as the electrolyte. Capacitance values and columbic eﬃciency are represented as a solid green line and dotted orange line, respectively. Composition of electrodes 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % starch. 4 C l i 4. Conclusions 4. Conclusions Figure S3: Electrode potential proﬁles and cell voltage of EDLC operating with aqueous electrolyte (1 mol L−1 Na2SO4) at current density 1.0 A g−1. Composition of electrodes 80 wt. % activated carbon (YP 80F) 10 wt. % carbon black (Super C65) and 10 wt. % starch. Author Contributions: Conceptualization, P.J.; Investigation, P.J. and P.Ł.K.; Methodology, P.J.; Writing—original draft, P.J. and P.Ł.K. Author Contributions: Conceptualization, P.J.; Investigation, P.J. and P.Ł.K.; Methodology, P.J.; Writing—original draft, P.J. and P.Ł.K. Funding: The European Commission (EC), in the form of the European Research Council (ERC), is acknowledged for funding provided within the ERC-StG-2017 project (grant agreement No. 759603) under the European Union’s Horizon 2020 research and innovation programme (PI: Krzysztof Fic) and Polish Ministry of Science and Higher Education 03/31/SBAD/0383. Acknowledgments: Authors would like to express gratitude towards Krzysztof Fic for the scientiﬁc and merit discussion, as well as François Béguin and El˙zbieta Fr ˛ackowiak for their input in the scientiﬁc development of the Power Sources Group. Conﬂicts of Interest: On behalf of all authors, the corresponding author states that there is no conﬂict of interest. Conﬂicts of Interest: On behalf of all authors, the corresponding author states that there is no conﬂict of interest. Conﬂicts of Interest: On behalf of all authors, the corresponding author states that there is no conﬂict of interest. References 1. Kötz, R.; Carlen, M. Principles and applications of electrochemical capacitors. Electrochim. Acta 2000, 45, 2483–2498. [CrossRef] 1. Kötz, R.; Carlen, M. Principles and applications of electrochemical capacitors. Electrochim. Acta 2000, 45, 2483–2498. [CrossRef] . Conway, B.E. Electrochemical Supercapacitors; Springer: Boston, MA, USA, 1999; ISBN 978-1-4757-3060-9 3. Lu, M. Supercapacitors: Materials, Systems, and Applications; Béguin, F., Fr ˛ackowiak, E., Eds.; Wiley-VCH: Weinheim, Germany, 2013; ISBN 9783527328833. 3. Lu, M. Supercapacitors: Materials, Systems, and Applications; Béguin, F., Fr ˛ackowiak, E., Eds.; Wiley-VCH: Weinheim, Germany, 2013; ISBN 9783527328833. 4. Helmholtz, H. Ueber einige Gesetze der Vertheilung elektrischer Ströme in körperlichen Leitern mit Anwendung auf die thierisch-elektrischen Versuche. Annalen der Physik 1853, 165, 211–233. [CrossRef] 4. Helmholtz, H. Ueber einige Gesetze der Vertheilung elektrischer Ströme in körperlichen Leitern mit Anwendung auf die thierisch-elektrischen Versuche. Annalen der Physik 1853, 165, 211–233. [CrossRef] 5. Stern, O. Zur Theorie Der Elektrolytischen Doppelschicht. Zeitschrift für Elektrochemie und Angewandte Physikalische Chemie 1924, 30, 508–516. [CrossRef] 5. Stern, O. Zur Theorie Der Elektrolytischen Doppelschicht. Zeitschrift für Elektrochemie und Angewandte Physikalische Chemie 1924, 30, 508–516. [CrossRef] 6. Koresh, J. Double Layer Capacitance and Charging Rate of Ultramicroporous Carbon Electrodes. J. Electrochem. Soc. 1977, 124, 1379. [CrossRef] 6. Koresh, J. Double Layer Capacitance and Charging Rate of Ultramicroporous Carbon Electrodes. J. Electrochem. Soc. 1977, 124, 1379. [CrossRef] 7. Salitra, G.; Soﬀer, A.; Eliad, L.; Cohen, Y.; Aurbach, D. Carbon Electrodes for Double-Layer Capacitors I. Relations Between Ion and Pore Dimensions. J. Electrochem. Soc. 2000, 147, 2486. [CrossRef] 8. Nagy, T.; Henderson, D.; Boda, D. Simulation of an Electrical Double Layer Model with a Low Dielectric Layer between the Electrode and the Electrolyte. J. Phys. Chem. B 2011, 115, 11409–11419. [CrossRef] 9. Huang, J.; Sumpter, B.G.; Meunier, V. Theoretical Model for Nanoporous Carbon Supercapacitors. Angew. Chem. Int. Ed. 2008, 47, 520–524. [CrossRef] [PubMed] 0. Gogotsi, Y.; Simon, P. True Performance Metrics in Electrochemical Energy Storage. Science 2011, 334, 917– [CrossRef] [PubMed] 11. Brousse, T.; Bélanger, D.; Long, J.W. To Be or Not To Be Pseudocapacitive? J. Electrochem. Soc. 2015, 162, A5185–A5189. [CrossRef] 12. Hulicova-Jurcakova, D.; Seredych, M.; Lu, G.Q.; Bandosz, T.J. Combined Eﬀect of Nitrogen- and Oxygen-Containing Functional Groups of Microporous Activated Carbon on its Electrochemical Performance in Supercapacitors. Adv. Funct. Mater. 2009, 19, 438–447. [CrossRef] 10 of 12 10 of 12 Polymers 2019, 11, 1648 13. Raymundo-Piñero, E.; Kierzek, K.; Machnikowski, J.; Béguin, F. References Relationship between the nanoporous texture of activated carbons and their capacitance properties in diﬀerent electrolytes. Carbon 2006, 44, 2498–2507. [CrossRef] 14. Frackowiak, E.; Lota, G.; Machnikowski, J.; Vix-Guterl, C.; Béguin, F. Optimisation of supercapacitors using carbons with controlled nanotexture and nitrogen content. Electrochim. Acta 2006, 51, 2209–2214. [CrossRef] 15. Stoeckli, F.; Centeno, T.A. On the determination of surface areas in activated carbons. Carbon 2005, 43, 1184–1190. [CrossRef] 4. Frackowiak, E.; Lota, G.; Machnikowski, J.; Vix-Guterl, C.; Béguin, F. Optimisation of supercapacitors u carbons with controlled nanotexture and nitrogen content. Electrochim. Acta 2006, 51, 2209–2214. [Cross 15. Stoeckli, F.; Centeno, T.A. On the determination of surface areas in activated carbons. Carbon 2005, 43, 1184–1190. [CrossRef] 16. Hulicova, D.; Kodama, M.; Hatori, H. Electrochemical Performance of Nitrogen-Enriched Carbons in Aqueous and Non-Aqueous Supercapacitors. Chem. Mater. 2006, 18, 2318–2326. [CrossRef] 17. Paraknowitsch, J.P.; Thomas, A. Doping carbons beyond nitrogen: An overview of advanced heteroatom doped carbons with boron, sulphur and phosphorus for energy applications. Energy Environ. Sci. 2013, 6, 2839. [CrossRef] 18. Béguin, F.; Presser, V.; Balducci, A.; Frackowiak, E. Carbons and Electrolytes for Advanced Supercapacitors. Adv. Mater. 2014, 26, 2219–2251. [CrossRef] [PubMed] 9. Chmiola, J. Anomalous Increase in Carbon Capacitance at Pore Sizes Less Than 1 Nanometer. Science 2 313, 1760–1763. [CrossRef] [PubMed] 19. Chmiola, J. Anomalous Increase in Carbon Capacitance at Pore Sizes Less Than 1 Nanometer. Science 2006, 313, 1760–1763. [CrossRef] [PubMed] 20. Gao, Q.; Demarconnay, L.; Raymundo-Piñero, E.; Béguin, F. Exploring the large voltage range of carbon/carbon supercapacitors in aqueous lithium sulfate electrolyte Energy Environ Sci 2012 5 9611 [CrossRef] 20. Gao, Q.; Demarconnay, L.; Raymundo-Piñero, E.; Béguin, F. Exploring the large voltage range of carbon supercapacitors in aqueous lithium sulfate electrolyte. Energy Environ. Sci. 2012, 5, 9611. [CrossRef 21. Fic, K.; Lota, G.; Meller, M.; Frackowiak, E. Novel insight into neutral medium as electrolyte for high-voltage supercapacitors. Energy Environ. Sci. 2012, 5, 5842–5850. [CrossRef] 22. Abbas, Q.; Babuchowska, P.; Fr ˛ackowiak, E.; Béguin, F. Sustainable AC/AC hybrid electrochemical capacitors in aqueous electrolyte approaching the performance of organic systems. J. Power Sources 2016, 326, 652–659. [CrossRef] 23. Moreno-Fernández, G.; Schütter, C.; Rojo, J.M.; Passerini, S.; Balducci, A.; Centeno, T.A. On the interaction of carbon electrodes and non conventional electrolytes in high-voltage electrochemical capacitors. J. Solid State Electrochem. 2018, 22, 717–725. [CrossRef] 24. References Aradilla, D.; Gao, F.; Lewes-Malandrakis, G.; Müller-Sebert, W.; Gentile, P.; Boniface, M.; Aldakov, D.; Iliev, B.; Schubert, T.J.S.; Nebel, C.E.; et al. Designing 3D Multihierarchical Heteronanostructures for High-Performance On-Chip Hybrid Supercapacitors: Poly(3,4-(ethylenedioxy)thiophene)-Coated Diamond/Silicon Nanowire Electrodes in an Aprotic Ionic Liquid. ACS Appl. Mater. Interfaces 2016, 8, 18069–18077. [CrossRef] [PubMed] 25. Dai, Z.; Peng, C.; Chae, J.H.; Ng, K.C.; Chen, G.Z. Cell voltage versus electrode potential range in aqueous supercapacitors. Sci. Rep. 2015, 5, 9854. [CrossRef] [PubMed] 26. Je˙zowski, P.; Fic, K.; Crosnier, O.; Brousse, T.; Béguin, F. Use of sacriﬁcial lithium nickel oxide for loading graphitic anode in Li-ion capacitors. Electrochim. Acta 2016, 206, 440–445. [CrossRef] 27. Du Pasquier, A.; Plitz, I.; Menocal, S.; Amatucci, G. A comparative study of Li-ion battery, supercapacitor and nonaqueous asymmetric hybrid devices for automotive applications. J. Power Sources 2003, 115, 171–178. [CrossRef] 28. Je˙zowski, P.; Fic, K.; Crosnier, O.; Brousse, T.; Béguin, F. Lithium rhenium(VII) oxide as a novel material for graphite pre-lithiation in high performance lithium-ion capacitors. J. Mater. Chem. A 2016, 4, 12609–12615. [CrossRef] 29. Naoi, K.; Ishimoto, S.; Isobe, Y.; Aoyagi, S. High-rate nano-crystalline Li4Ti5O12 attached on nano-ﬁbers for hybrid supercapacitors. J. Power Sources 2010, 195, 6250–6254. [CrossRef] 30. Je˙zowski, P.; Crosnier, O.; Deunf, E.; Poizot, P.; Béguin, F.; Brousse, T. Safe and recyclable lithium-ion capacitors using sacriﬁcial organic lithium salt. Nat. Mater. 2018, 17, 167–173. [CrossRef] 31. Aida, T.; Murayama, I.; Yamada, K.; Morita, M. Improvement in Cycle Performance of a High-Voltage Hybrid Electrochemical Capacitor. Electrochem. Solid-State Lett. 2007, 10, A93. [CrossRef] 32. Böckenfeld, N.; Jeong, S.S.; Winter, M.; Passerini, S.; Balducci, A. Natural, cheap and environmentally friendly binder for supercapacitors. J. Power Sources 2013, 221, 14–20. [CrossRef] 33. Krause, A.; Balducci, A. High voltage electrochemical double layer capacitor containing mixtures of ionic liquids and organic carbonate as electrolytes. Electrochem. Commun. 2011, 13, 814–817. [CrossRef] 34. Krause, A.; Kossyrev, P.; Oljaca, M.; Passerini, S.; Winter, M.; Balducci, A. Electrochemical double layer capacitor and lithium-ion capacitor based on carbon black. J. Power Sources 2011, 196, 8836–8842. [CrossRef] 11 of 12 Polymers 2019, 11, 1648 11 of 12 35. Kolodziej, A.; Fic, K.; Frackowiak, E. Towards sustainable power sources: Chitin-bound carbon electrodes for electrochemical capacitors. J. Mater. Chem. A 2015, 3, 22923–22930. [CrossRef] 36. Choudhury, N.A.; Northrop, P.W.C.; Crothers, A.C.; Jain, S.; Subramanian, V.R. Chitosan hydrogel-based electrode binder and electrolyte membrane for EDLCs: Experimental studies and model validation. J. Appl. Electrochem. References 2012, 42, 935–943. [CrossRef] 37. Varzi, A.; Raccichini, R.; Marinaro, M.; Wohlfahrt-Mehrens, M.; Passerini, S. Probing the characteristics of casein as green binder for non-aqueous electrochemical double layer capacitors’ electrodes. J. Power Sources 2016, 326, 672–679. [CrossRef] 38. Gao, H.; Lian, K. Proton-conducting polymer electrolytes and their applications in solid supercapacitors: A review. RSC Adv. 2014, 4, 33091–33113. [CrossRef] 39. Zhang, Y.; Sun, X.; Pan, L.; Li, H.; Sun, Z.; Sun, C.; Tay, B.K. Carbon nanotube–ZnO nanocomposite electrodes for supercapacitors. Solid State Ion. 2009, 180, 1525–1528. [CrossRef] 39. Zhang, Y.; Sun, X.; Pan, L.; Li, H.; Sun, Z.; Sun, C.; Tay, B.K. Carbon nan for supercapacitors. Solid State Ion. 2009, 180, 1525–1528. [CrossRef] 40. Varzi, A.; Passerini, S. Enabling high areal capacitance in electrochemical double layer capacitors by means of the environmentally friendly starch binder. J. Power Sources 2015, 300, 216–222. [CrossRef] 1. Portet, C.; Taberna, P.; Simon, P.; Laberty-Robert, C. Modiﬁcation of Al current collector surface by sol deposit for carbon–carbon supercapacitor applications. Electrochim. Acta 2004, 49, 905–912. [CrossRef] 41. Portet, C.; Taberna, P.; Simon, P.; Laberty Robert, C. Modiﬁcation of Al current collector surface by sol gel deposit for carbon–carbon supercapacitor applications. Electrochim. Acta 2004, 49, 905–912. [CrossRef] 42 Taberna PL ; Portet C ; Simon P Electrode surface treatment and electrochemical impedance spectroscopy 42. Taberna, P.L.; Portet, C.; Simon, P. Electrode surface treatment and electrochemical impedan study on carbon/carbon supercapacitors. Appl. Phys. A 2006, 82, 639–646. [CrossRef] 43. Je˙zowski, P.; Nowicki, M.; Grzeszkowiak, M.; Czajka, R.; Béguin, F. Chemical etching of stainless steel 301 for improving performance of electrochemical capacitors in aqueous electrolyte. J. Power Sources 2015, 279, 555–562. [CrossRef] 44. Gross, R.A.; Kalra, B. Biodegradable Polymers for the Environment. Science 2002, 297, 803–807. [CrossRef] [PubMed] 45. Lu, D.R.; Xiao, C.M.; Xu, S.J. Starch-based completely biodegradable polymer materials. Express Polym. Lett. 2009, 3, 366–375. [CrossRef] 46. Makowska, A.; Szwengiel, A.; Kubiak, P.; Tomaszewska-Gras, J. Characteristics and structure of starch isolated from triticale. Starch Stärke 2014, 66, 895–902. [CrossRef] 7. Kennedy, H.M. Starch- and Dextrin-Based Adhesives. In Adhesives from Renewable Resources; ACS Publica Washington, DC, USA, 1989; pp. 326–336. 48. Imam, S.H.; Gordon, S.H.; Mao, L.; Chen, L. Environmentally friendly wood adhesive from a renewable plant polymer: Characteristics and optimization. Polym. Degrad. Stab. 2001, 73, 529–533. [CrossRef] 49. Wang, Z.; Li, Z.; Gu, Z.; Hong, Y.; Cheng, L. Preparation, characterization and properties of starch-based wood adhesive. Carbohydr. Polym. References 2012, 88, 699–706. [CrossRef] 50. Zhang, Y.; Ding, L.; Gu, J.; Tan, H.; Zhu, L. Preparation and properties of a starch-based wood adhesive with high bonding strength and water resistance. Carbohydr. Polym. 2015, 115, 32–37. [CrossRef] [PubMed] 51. Le Thanh-Blicharz, J.; Lubiewski, Z.; Voelkel, E.; Lewandowicz, G. Evaluation of rheological properties of commercial native starches. Zywnosc Nauka Technologia Jakosc/Food Sci. Technol. Qual. 2011, 18, 53–65. [CrossRef] 52. Makowska, A.; Kubiak, P.; Bialas, W.; Lewandowicz, G. Eﬀect of urea, sodium nitrate and ethylene glycol addition on the rheological properties of corn starch pastes. Polimery 2015, 60, 343–350. [CrossRef] 52. Makowska, A.; Kubiak, P.; Bialas, W.; Lewandowicz, G. Eﬀect of urea, sodium nitrate and ethylene glycol addition on the rheological properties of corn starch pastes. Polimery 2015, 60, 343–350. [CrossRef] 53. Kilbride, P.; Rull, M.V.; Townsend, A.; Wilson, H.; Morris, J. Shear-thickening ﬂuids in biologically relevant 3. Kilbride, P.; Rull, M.V.; Townsend, A.; Wilson, H.; Morris, J. Shear-thickening ﬂuids in biologically relev agents. Biorheology 2019, 1–12. [CrossRef] 54. Ma, Y.; Wu, S.; Tong, J.; Zhang, X.; Peng, J.; Liu, X. Rheological Properties of Corn Starch Dispersions in Pregelatinized Starch Solution. In Proceedings of the 2018 IEEE International Conference on Manipulation, Manufacturing and Measurement on the Nanoscale (3M-NANO), Hangzhou, China, 13–17 August 2018; pp. 146–150. 55. Ai, Y.; Jane, J. Gelatinization and rheological properties of starch. Starch Stärke 2015, 67, 213–224. [CrossRef] 55. Ai, Y.; Jane, J. Gelatinization and rheological properties of starch. Starch Stärke 2015, 67, 213–224. [CrossRef] 56. Baranowska, H.M.; Sikora, M.; Krystyjan, M.; Tomasik, P. Analysis of the formation of starch—Hydrocolloid 56. Baranowska, H.M.; Sikora, M.; Krystyjan, M.; Tomasik, P. Analysis of the formation of starch—Hydrocolloid binary gels and their structure based on the relaxation times of the water molecules. Polimery 2011, 56, 478–483. [CrossRef] 57. Thirumdas, R.; Kadam, D.; Annapure, U.S. Cold Plasma: An Alternative Technology for the Starch Modiﬁcation. Food Biophys. 2017, 12, 129–139. [CrossRef] 12 of 12 Polymers 2019, 11, 1648 58. Wang, Y.-J.; Truong, V.-D.; Wang, L. Structures and rheological properties of corn starch as aﬀected by acid hydrolysis. Carbohydr. Polym. 2003, 52, 327–333. [CrossRef] 59. Richner, R.; Müller, S.; Bärtschi, M.; Kötz, R.; Wokaun, A. Physically and Chemically Bonded Carbonaceous Material for Double-Layer Capacitor Applications. J. New Mater. Electrochem. Syst. 2002, 5, 297–304. 60. Maletin, Y.; Strelko, V.; Stryzhakova, N.; Zelinsky, S.; Rozhenko, A.B.; Gromadsky, D.; Volkov, V.; Tychina, S.; Gozhenko, O.; Drobny, D. © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). References Carbon Based Electrochemical Double Layer Capacitors of Low Internal Resistance. Energy Environ. Res. 2013, 3. [CrossRef] 61. Maletin, Y.; Stryzhakova, N.; Zelinsky, S.; Chernukhin, S.; Tretyakov, D.; Tychina, S.; Drobny, D. Electrochemical Double Layer Capacitors and Hybrid Devices for Green Energy Applications. Green 2014, 4. [CrossRef] 62. Yurii Maletin, N.S.; Sergii Zelinskyi, S.C.; Dmytro Tretyakov, H.M.; Natalia, D.; Dmytro, D. New Approach to Ultracapacitor Technology: What it Can Oﬀer to Electriﬁed Vehicles. J. Energy Power Eng. 2015, 9. [CrossRef] 62. Yurii Maletin, N.S.; Sergii Zelinskyi, S.C.; Dmytro Tretyakov, H.M.; Natalia, D.; Dmytro, D. New Approach to Ultracapacitor Technology: What it Can Oﬀer to Electriﬁed Vehicles. J. Energy Power Eng. 2015, 9. [CrossRef] 63. Sevilla, M.; Fuertes, A.B. Direct Synthesis of Highly Porous Interconnected Carbon Nanosheets and Their Application as High-Performance Supercapacitors ACS Nano 2014 8 5069–5078 [CrossRef] 63. Sevilla, M.; Fuertes, A.B. Direct Synthesis of Highly Porous Interconnected Carbon Nanosheets and Their Application as High-Performance Supercapacitors. ACS Nano 2014, 8, 5069–5078. [CrossRef] 64. Brandt, A.; Balducci, A. Theoretical and practical energy limitations of organic and ionic liquid-based electrolytes for high voltage electrochemical double layer capacitors. J. Power Sources 2014, 250, 343–351. [CrossRef] 65. Krüner, B.; Odenwald, C.; Quade, A.; Kickelbick, G.; Presser, V. Inﬂuence of Nitrogen-Doping for Carbide-Derived Carbons on the Supercapacitor Performance in an Organic Electrolyte and an Ionic Liquid. Batter. Supercaps 2018, 1, 135–148. [CrossRef] 66. Krüner, B.; Lee, J.; Jäckel, N.; Tolosa, A.; Presser, V. Sub-micrometer Novolac-Derived Carbon Beads for High Performance Supercapacitors and Redox Electrolyte Energy Storage. ACS Appl. Mater. Interfaces 2016, 8, 9104–9115. [CrossRef] [PubMed] © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
https://openalex.org/W2741583580	https://acp.copernicus.org/articles/14/6177/2014/acp-14-6177-2014.pdf	English	null	TransCom N&lt;sub&gt;2&lt;/sub&gt;O model inter-comparison, Part II: Atmospheric inversion estimates of N&lt;sub&gt;2&lt;/sub&gt;O emissions	null	2,014	cc-by	15,146	Sopron, Hungary Swiss Federal Laboratories for Materials Science and Technology (Empa), Dübendorf, Switzerland 17Institute for Marine and Atmospheric Research Utrecht, University of Utrecht, Utrecht, the Netherlands nnish Meteorological Institute, Helsinki, Finland 19Umweltbundesamt, Messstelle Schauinsland, Kirchzarten, Germany 20 20Centre for Isotope Research, University of Groningen, Groningen, the Netherlands 21 21School of GeoSciences, The University of Edinburgh, Edinburgh, UK now at: School of Environmental Sciences University of East Anglia Norw 21School of GeoSciences, The University of Edinburgh, Edinburgh, UK now at: School of Environmental Sciences, University of East Anglia, Norwich, UK Correspondence to: R. L. Thompson (rona.thompson@nilu.no) Correspondence to: R. L. Thompson (rona.thompson@nilu.no) Received: 15 November 2013 – Published in Atmos. Chem. Phys. Discuss.: 27 February 2014 Revised: 9 May 2014 – Accepted: 16 May 2014 – Published: 23 June 2014 Received: 15 November 2013 – Published in Atmos. Chem. Phys. Discuss.: 27 February 2014 Revised: 9 May 2014 – Accepted: 16 May 2014 – Published: 23 June 2014 Abstract. This study examines N2O emission estimates from ﬁve different atmospheric inversion frameworks based on chemistry transport models (CTMs). The ﬁve frameworks differ in the choice of CTM, meteorological data, prior un- certainties and inversion method but use the same prior emis- sions and observation data set. The posterior modelled atmo- spheric N2O mole fractions are compared to observations to assess the performance of the inversions and to help diag- nose problems in the modelled transport. Additionally, the mean emissions for 2006 to 2008 are compared in terms of the spatial distribution and seasonality. Overall, there is a good agreement among the inversions for the mean global total emission, which ranges from 16.1 to 18.7 TgN yr−1 and is consistent with previous estimates. Ocean emissions Atmos. Chem. Phys., 14, 6177–6194, 2014 www.atmos-chem-phys.net/14/6177/2014/ doi:10.5194/acp-14-6177-2014 © Author(s) 2014. CC Attribution 3.0 License. TransCom N2O model inter-comparison – Part 2: Atmospheric inversion estimates of N2O emissions R. L. Thompson1,2, K. Ishijima3, E. Saikawa4,5, M. Corazza6, U. Karstens7, P. K. Patra3, P. Bergamaschi6, F. Chevallier2, E. Dlugokencky8, R. G. Prinn4, R. F. Weiss9, S. O’Doherty10, P. J. Fraser11, L. P. Steele11, P. B. Krummel11, A. Vermeulen12, Y. Tohjima13, A. Jordan7, L. Haszpra14,15, M. Steinbacher16, S. Van der Laan17,, T. Aalto18, F. Meinhardt19, M. E. Popa7,20, J. Moncrieff21, and P. Bousquet2 1Norwegian Institute for Air Research, Kjeller, Norway 2Laboratoire des Sciences du Climat et l’Environnement, Gif sur Yvette, France 3Research Institute for Global Change, JAMSTEC, Yokohama, Japan 4Center for Global Change Science, MIT, Cambridge, MA, USA 5Emory University, Atlanta, GA, USA 6Institute for Environment and Sustainability, JRC, Ispra, Italy 7Max Planck Institute for Biogeochemistry, Jena, Germany 8NOAA Earth System Research Laboratory, Global Monitoring Division, Boulder, CO, USA 9Scripps Institution of Oceanography, La Jolla, CA, USA 10Atmospheric Chemistry Research Group, School of Chemistry, University of Bristol, Bristol, UK 11Centre for Australian Weather and Climate Research, CSIRO, Marine and Atmospheric Research, Aspendale, Victoria, Australia 12Energy Research Centre of the Netherlands (ECN), Petten, the Netherlands 13National Institute for Environmental Studies, Tsukuba, Japan 14Hungarian Meteorological Service, Budapest, Hungary 15Geodetic and Geophysical Institute, Research Centre for Astronomy and Earth Sciences, Hungarian Academy of Sciences, Sopron, Hungary 16Swiss Federal Laboratories for Materials Science and Technology (Empa), Dübendorf, Switzerland 17Institute for Marine and Atmospheric Research Utrecht, University of Utrecht, Utrecht, the Netherlands 18Finnish Meteorological Institute, Helsinki, Finland 19Umweltbundesamt, Messstelle Schauinsland, Kirchzarten, Germany 20Centre for Isotope Research, University of Groningen, Groningen, the Netherlands 21School of GeoSciences, The University of Edinburgh, Edinburgh, UK now at: School of Environmental Sciences, University of East Anglia, Norwich, UK Atmos. Chem. Phys., 14, 6177–6194, 2014 www.atmos-chem-phys.net/14/6177/2014/ doi:10.5194/acp-14-6177-2014 © Author(s) 2014. CC Attribution 3.0 License. TransCom N2O model inter-comparison – Part 2: Atmospheric inversion estimates of N2O emissions 2Laboratoire des Sciences du Climat et l’Environnement, Gif sur Yvette, France 2Laboratoire des Sciences du Climat et l’Environnement, Gif sur Yvette, France boratoire des Sciences du Climat et l’Environneme 3Research Institute for Global Change, JAMSTEC, Yokohama, Japan 3Research Institute for Global Change, JAMSTEC, Yokohama, Japan 4Center for Global Change Science, MIT, Cambridge, MA, USA 4Center for Global Change Science, MIT, Cambridge, MA, USA 5Emory University, Atlanta, GA, USA 6Institute for Environment and Sustainability, JRC, Ispra, Italy 6Institute for Environment and Sustainability, JRC, Ispra, Italy 7Max Planck Institute for Biogeochemistry, Jena, Germany 7Max Planck Institute for Biogeochemistry, Jena, Germany 7Max Planck Institute for Biogeochemistry, Jena, Germany g y y 8NOAA Earth System Research Laboratory, Global Monitoring Division 8NOAA Earth System Research Laboratory, Global Monitoring Division, Boulder, CO, USA 10Atmospheric Chemistry Research Group, School of Chemistry, University of Bristol, Bristol, UK 11 11Centre for Australian Weather and Climate Research, CSIRO, Marine and Atmospheric Research, Aspendale, Victoria, Australia 12Energy Research Centre of the Netherlands (ECN), Petten, the Netherlands 13National Institute for Environmental Studies, Tsukuba, Japan 14Hungarian Meteorological Service, Budapest, Hungary 15 15Geodetic and Geophysical Institute, Research Centre for Astronomy and Earth Sciences, Hungarian Academy of Sciences, Sopron, Hungary R. L. Thompson et al.: TransCom N2O model inter-comparison – Part 2 6178 represent between 31 and 38 % of the global total com- pared to widely varying previous estimates of 24 to 38 %. Emissions from the northern mid- to high latitudes are likely to be more important, with a consistent shift in emissions from the tropics and subtropics to the mid- to high latitudes in the Northern Hemisphere; the emission ratio for 0–30◦N to 30–90◦N ranges from 1.5 to 1.9 compared with 2.9 to 3.0 in previous estimates. The largest discrepancies across inver- sions are seen for the regions of South and East Asia and for tropical and South America owing to the poor observa- tional constraint for these areas and to considerable differ- ences in the modelled transport, especially inter-hemispheric exchange rates and tropical convective mixing. Estimates of the seasonal cycle in N2O emissions are also sensitive to er- rors in modelled stratosphere-to-troposphere transport in the tropics and southern extratropics. Overall, the results show a convergence in the global and regional emissions compared to previous independent studies. there is a natural nitrogen turnover leading to N2O emissions but these may be enhanced by the input of reactive nitrogen from fertilizers and other anthropogenic sources by atmo- spheric transport, erosion and leaching, leading to so-called indirect anthropogenic emissions (Galloway et al., 2003). p g ( y ) An alternative and complementary approach to upscaling small-scale ﬂuxes and processes to estimate regional and global N2O budgets, is to use a top-down approach. Atmo- spheric inversion is one such top-down approach and uses observations of N2O mole fractions with a model of at- mospheric transport and chemistry in a statistically rigor- ous way to constrain surface ﬂuxes. This approach has been used previously for estimating N2O emissions on regional (Corazza et al., 2011; Thompson et al., 2011b) and global scales (Hirsch et al., 2006; Huang et al., 2008; Kort et al., 2011; Prinn et al., 1990; Saikawa et al., 2014; Thompson et al., 2014a). One major advantage of the atmospheric inver- sion approach is that it provides a constraint on the total N2O emission since the atmosphere integrates the ﬂuxes and re- quires that the change in atmospheric N2O abundance be bal- anced by the sum of its sources and sinks. 1 Introduction Nitrous oxide (N2O) currently has the third largest contribu- tion to net radiative forcing after CO2 and CH4, and currently has radiative forcing of 0.17 Wm−2 (Myhre et al., 2013). Fur- thermore, N2O plays an important role in stratospheric ozone loss and currently the ozone-depleting-potential weighted emissions of N2O are thought to be the highest of any ozone- depleting substance (Ravishankara et al., 2009). The atmo- spheric mole fraction of N2O has increased signiﬁcantly since the mid-20th century largely as a result of agricultural activities and, in particular, the use of nitrogen fertilizers (Park et al., 2012). Currently, agricultural emissions from fer- tilizer use and manure management (4.3–5.8 TgN yr−1) and emissions from natural soils (6–7 TgN yr−1) account for 60– 70 % of global N2O emissions (Syakila and Kroeze, 2011; Zaehle et al., 2011). The remaining 30–40 % of emissions is from oceans (4.5 TgN yr−1) (Duce et al., 2008) and, to a smaller extent, from fuel combustion, industry (Olivier et al., 2005) and biomass burning (together 1.7 TgN yr−1) (van der Werf et al., 2010). Part 1 of the TransCom N2O experiment examined the im- portance of atmospheric transport and surface ﬂuxes on tro- pospheric N2O mole fractions and, speciﬁcally, looked at the inﬂuence of transport model errors on N2O mole fractions on seasonal to annual timescales (Thompson et al., 2014b). In this paper (Part 2), we present N2O emission estimates from ﬁve inversion frameworks based on ﬁve different at- mospheric chemistry transport models (CTMs), all of which also participated in Part 1. In this context, the objectives of this paper are to: N2O is dominantly produced by microbial processes in soils, sediments and water bodies – speciﬁcally, by nitriﬁ- cation and denitriﬁcation. Although a lot is already known about these processes from laboratory studies under con- trolled conditions and in situ chamber ﬂux measurements, upscaling to emissions on national or regional scales is ham- pered by the strongly variable nature of soil ﬂuxes. N2O pro- duction in soils is dependent on a multitude of environmen- tal factors such as soil moisture and temperature, soil type, among others, which interact in complex ways and are difﬁ- cult to predict. In agricultural soils, the type of nitrogen fertil- izer and the timing of its application are also important con- siderations for estimating N2O emissions. In natural soils, – compare the posterior emissions (i.e. R. L. Thompson et al.: TransCom N2O model inter-comparison – Part 2 In general terms, upscaling approaches provide a detailed picture of the pro- cesses and source types while top-down approaches provide an integrated picture of the regional and long-term emissions and a check on the total budget. However, atmospheric in- versions also have sources of error. The estimated ﬂuxes are sensitive to errors in the modelled transport and, to varying degrees, the chemistry, as these are non-random errors that are extremely difﬁcult to estimate and account for in an inver- sion framework. Particularly for N2O, errors in stratosphere– troposphere exchange (STE) represent an important source of model error since there is a strong N2O mole fraction gra- dient across the tropopause owing to the loss of N2O through photolysis and reaction with O(1D) in the stratosphere (see Part 1, Thompson et al., 2014b). – present regional emissions estimates and their uncer- tainties – present regional emissions estimates and their uncer- tainties by the CTM in each inversion framework. Depending on the inversion framework, H is either a matrix or a non-linear op- erator. The frameworks differ in how the minimum of the cost function (Eq. 1) is sought. Approaches for ﬁnding the x that minimizes this equation fall into one of the follow- ing categories: (1) variational methods, such as those used in weather forecasting (Courtier et al., 1994) and (2) analytical methods (Tarantola, 2005). Variational methods ﬁnd the opti- mal x using an iterative descent algorithm, usually requiring calculation of the gradient of J at each iteration, and do not require H to be a matrix operator (Chevallier et al., 2005). Analytical methods require that the transport operator H is linear and deﬁned (i.e. H) and the optimal (posterior) x is found by solving Eq. (2) or (3) directly (for a derivation of these equations refer to Tarantola, 2005): This paper is divided into four main sections. In Sect. 2, we outline the inversion frameworks and CTMs, as well as the prior ﬂux estimates and atmospheric observations used in this study. Section 3.1 presents a validation of the inversion re- sults by comparing the mole fractions simulated using the posterior ﬂuxes with observations, while Sect. 3.2 analyses the spatial and temporal distribution of the posterior ﬂuxes. In Sect. 3.3, we compare these estimates with those of pre- vious studies and conclude with a discussion of the major challenges for estimating N2O emissions from atmospheric inversions. 2 Methods x = xb + (HTR−1H + B−1)−1HTR−1(y −Hxb) (2) x = xb + BHT HBHT + R −1 (y −Hxb) (3) (2) (3) (3) 2.1 Inversion frameworks (3) Five different inversion frameworks participated in Part 2 of this experiment. In this paper, we refer to each of the frame- works according to the CTM used followed by “-I” to indi- cate that this is the inversion framework. Although the frame- works may be used with a different CTM, in this study the naming is unambiguous as a different CTM was used with each one (see Table 1). All frameworks use the Bayesian inversion method to ﬁnd the optimal surface ﬂuxes, that is, the ﬂuxes that provide the best ﬁt to the atmospheric obser- vations, y, while being guided by the prior ﬂux estimates, xb, and their uncertainties (for details about the Bayesian method refer to Tarantola, 2005). Based on Bayesian theory, and Gaussian-error hypotheses, the optimal ﬂuxes are those that minimize the cost function (analytical methods can also be used in the case that H is non-linear if it is still differentiable and that the linear H can be deﬁned over a small range of x, in which case, an outer loop is also required to better approximate H). For in- versions falling into the ﬁrst category, an adjoint model of the atmospheric chemistry and transport is used to calculate the gradient and to ﬁnd the minimum; the TM5-I, TM3-I, and LMDZ4-I frameworks fall into this category (see Ta- ble 2). For inversions in the second category, the chemistry- transport operator, H, represents the sensitivity of the ob- servations to the ﬂuxes in each of a given number of pre- deﬁned regions. Each column of H can be found by run- ning the CTM, perturbing the ﬂuxes in a given region, and determining the resulting change atmospheric mole fraction for all observations. The emissions sensitivity is thus the ra- tio of the change in mole fraction to the change in ﬂux; the ACTMt42l67-I and MOZART4-I frameworks fall into this category (see Table 2). (analytical methods can also be used in the case that H is non-linear if it is still differentiable and that the linear H can be deﬁned over a small range of x, in which case, an outer loop is also required to better approximate H). 1 Introduction resulting from the inversions) in a standardized way – analyse the posterior emissions in terms of spatial dis- tribution, seasonal variability, and to identify robust fea- tures common to all inversions – identify regions where there are discrepancies between inversions and investigate their cause www.atmos-chem-phys.net/14/6177/2014/ Atmos. Chem. Phys., 14, 6177–6194, 2014 R. L. Thompson et al.: TransCom N2O model inter-comparison – Part 2 Table 1. Overview of the CTMs used in the inversions. Note that the horizontal resolution is given as longitude by latitude Model Institute Resolution Top boundary Meteorology horizontal vertical pressure (hPa) MOZART4 MIT 2.5◦× 1.88◦ 56 σ 1 2 MERRA ACTMt42l67 JAMSTEC 2.8◦× 2.8◦ 67 σ 0.01 JRA25 TM3 MPI-BGC 5.0◦× 3.75◦ 26 η2 1 ERA-Interim TM5 JRC 6.0◦× 4.0◦ 25 η 0.5 ERA-Interim LMDZ4 LSCE 3.75◦× 2.5◦ 19 η 4 ERA-Interim 1 σ refers to the sigma terrain-following vertical coordinate system. 1 η refers to the eta coordinate system that smoothly transitions from the sigma coordinate near the surface to a pressure coordinate in the stratosphere. 6179 R. L. Thompson et al.: TransCom N2O model inter-comparison – Part 2 w of the CTMs used in the inversions. Note that the horizontal resolution is given as longitude by latitude. 2.1 Inversion frameworks For in- versions falling into the ﬁrst category, an adjoint model of the atmospheric chemistry and transport is used to calculate the gradient and to ﬁnd the minimum; the TM5-I, TM3-I, and LMDZ4-I frameworks fall into this category (see Ta- ble 2). For inversions in the second category, the chemistry- transport operator, H, represents the sensitivity of the ob- servations to the ﬂuxes in each of a given number of pre- deﬁned regions. Each column of H can be found by run- ning the CTM, perturbing the ﬂuxes in a given region, and determining the resulting change atmospheric mole fraction for all observations. The emissions sensitivity is thus the ra- tio of the change in mole fraction to the change in ﬂux; the ACTMt42l67-I and MOZART4-I frameworks fall into this category (see Table 2). J (x) = (x −xb)T B−1 (x −xb) (1) + (H (x) −y)T R−1 (H (x) −y), (1) where the prior ﬂux uncertainties are described by the error covariance matrix, B, the observation uncertainties are de- scribed by the error covariance matrix, R, and H is an op- erator of the atmospheric transport and chemistry as deﬁned Atmos. Chem. Phys., 14, 6177–6194, 2014 www.atmos-chem-phys.net/14/6177/2014/ 2.2.1 Stratospheric N2O loss As in Part 1, all participants were requested to use the same atmospheric observations, prior ﬂux estimates and ap- proximate magnitude of the stratospheric sink. Thereby, the sources of differences between inversion results are limited to the choice of CTM and meteorological data, the inver- sion method, and uncertainties assigned to the prior ﬂuxes and the observations. Furthermore, since the CTMs used in each of the inversion frameworks are the same as those used in Part 1, the analysis of the transport model performance can be directly applied in this study when considering differ- ences between posterior ﬂuxes. All inversions were run for the period 2005–2009 but only output from 2006 onwards was analysed as 2005 was used as a spin-up year. A spin- up period is required to minimize the inﬂuence of the initial conditions on the posterior emissions. Here, we chose 1 year for the spin-up as all models started with their best initial conditions estimates established after previous longer inte- grations of the CTMs. Also, when presenting mean emission results, the years 2006–2008 are used, as the end of 2009 is not as well constrained in the inversions (to constrain the end of 2009, observations at the beginning of 2010 would need Loss of N2O in the stratosphere through photolysis (circa 90 % of the loss; Minschwaner et al., 1993) and reaction with O(1D) (circa 10 %) was calculated in each model in every grid cell and time step. Although the exact photolysis and ox- idation rates varied between models (according to the CTM used to calculate the photolysis rate and O(1D) concentra- tion) these were scaled such that the global annual total loss of N2O was approximately 12.5 TgN, consistent with esti- mates of the atmospheric abundance and the lifetime of N2O, which is estimated to be between 124 and 130 years (Prather et al., 2012; Volk et al., 1997). 2.2 Experiment protocol 2.2 Experiment protocol Table 3. Prior ﬂux model overview (totals shown for 2005). Table 3. Prior ﬂux model overview (totals shown for 2005). Category Data set Resolution Total (TgN yr−1) Terrestrial biosphere ORCHIDEE O-CN monthly 10.83 Ocean PISCES monthly 4.28 Waste water EDGAR-4.1 annual 0.21 Solid waste EDGAR-4.1 annual 0.004 Solvents EDGAR-4.1 annual 0.05 Fuel production EDGAR-4.1 annual 0.003 Ground transport EDGAR-4.1 annual 0.18 Industry combustion EDGAR-4.1 annual 0.41 Residential and other combustion EDGAR-4.1 annual 0.18 Shipping EDGAR-4.1 annual 0.002 Other sources EDGAR-4.1 annual 0.0005 Biomass burning GFED-2 monthly 0.71 Total monthly 16.84 to be included and these were not all available at the time of preparing this study). to be included and these were not all available at the time of preparing this study). www.atmos-chem-phys.net/14/6177/2014/ 6180 R. L. Thompson et al.: TransCom N2O model inter-comparison – Part 2 Table 2. Overview of the inversion frameworks. The matrix B is the prior ﬂux error covariance matrix (see Eq. 1). Model Resolution Inversion method Scale length in B (km) Spatial Temporal Land Ocean MOZART4-I1 13 regions monthly Analytical none none ACTMt42l67-I2 22 regions6 monthly Analytical none none TM3-I3 5.0◦× 3.75◦ monthly Variational 500 500 TM5-I4 6.0◦× 4.0◦ monthly Variational 200 200 LMDZ4-I5 3.75◦× 2.5◦ monthly Variational 500 1000 1 Saikawa et al. (2013), 2 based on Rayner et al. (1999), 3 Rödenbeck (2005), 4 Corazza et al. (2011), 5 Thompson et al. (2011a), 6 TransCom-3 regions. Table 3. Prior ﬂux model overview (totals shown for 2005). Category Data set Resolution Total (TgN yr−1) Terrestrial biosphere ORCHIDEE O-CN monthly 10.83 Ocean PISCES monthly 4.28 Waste water EDGAR-4.1 annual 0.21 Solid waste EDGAR-4.1 annual 0.004 Solvents EDGAR-4.1 annual 0.05 Fuel production EDGAR-4.1 annual 0.003 Ground transport EDGAR-4.1 annual 0.18 Industry combustion EDGAR-4.1 annual 0.41 Residential and other combustion EDGAR-4.1 annual 0.18 Shipping EDGAR-4.1 annual 0.002 Other sources EDGAR-4.1 annual 0.0005 Biomass burning GFED-2 monthly 0.71 Total monthly 16.84 2.2 Experiment protocol As in Part 1, all participants were requested to use the same atmospheric observations, prior ﬂux estimates and ap- proximate magnitude of the stratospheric sink. Thereby, the sources of differences between inversion results are limited to be included and these were not all available at the time of preparing this study). 2.2.1 Stratospheric N2O loss Loss of N2O in the stratosphere through photolysis (circa 90 % of the loss; Minschwaner et al., 1993) and reaction with O(1D) ( i 10 %) l l t d i h d l i 6180 R. L. Thompson et al.: TransCom N2O model inter-comparison – Part 2 Table 2. Overview of the inversion frameworks. The matrix B is the prior ﬂux error covariance matrix (see Eq. 1). Model Resolution Inversion method Scale length in B (km) Spatial Temporal Land Ocean MOZART4-I1 13 regions monthly Analytical none none ACTMt42l67-I2 22 regions6 monthly Analytical none none TM3-I3 5.0◦× 3.75◦ monthly Variational 500 500 TM5-I4 6.0◦× 4.0◦ monthly Variational 200 200 LMDZ4-I5 3.75◦× 2.5◦ monthly Variational 500 1000 1 Saikawa et al. (2013), 2 based on Rayner et al. (1999), 3 Rödenbeck (2005), 4 Corazza et al. (2011), 5 Thompson et al. (2011a), 6 TransCom-3 regions. Table 2. Overview of the inversion frameworks. www.atmos-chem-phys.net/14/6177/2014/ The matrix B is the prior ﬂux error covariance matrix (see Eq. 1). 2. Overview of the inversion frameworks. The matrix B is the prior ﬂux error covariance matrix (see Eq. 1). Table 3. Prior ﬂux model overview (totals shown for 2005). 2.3 Atmospheric observations Atmospheric observations of N2O mole fractions (nmol mol−1 equivalently parts per billion, abbreviated as ppb) were pooled from two global networks, NOAA CCGG (Car- bon Cycle and Greenhouse Gases) and AGAGE (Advanced Global Atmospheric Gases Experiment), as well as from a number of smaller regional networks and independent sta- tions (see Fig. 1 and Table 4). From the NOAA CCGG net- work, 42 sites were included. Approximately weekly dis- crete air samples are taken at these sites, which are subse- quently analysed for N2O using GC-ECD (Gas Chromatog- raphy Electron Capture Detector). These data are reported on the NOAA-2006A calibration scale (Hall et al., 2007) and have a reproducibility of 0.4 ppb based on the mean differ- ence of ﬂask pairs. The AGAGE network consists of ﬁve in situ GC-ECD instruments. These data are reported on the SIO-1998 scale and have a reproducibility of approximately 0.1 ppb (Prinn et al., 2000). The MPI-BGC (Max Planck In- stitute for Biogeochemistry) network consists of three sites for discrete air samples and two sites with in situ GC-ECD instruments. These data are also reported on the NOAA- 2006A scale and have a reproducibility of about 0.3 ppb. In addition, data from nine independently run stations with in situ GC-ECD instruments were included (see Table 4). Similarly, each inversion framework has a different esti- mation method for the representation uncertainty, that is, the uncertainty due to the modelled transport and the tempo- ral representation uncertainty. The transport uncertainties are calculated as follows: TM5-I and LMDZ4-I use the method of Bergamaschi et al., 2010; TM3-I uses pre-determined values for the uncertainty at marine (1.6 ppb), mountain (2.4 ppb), continental (4.8 ppb), and coastal (2.4 ppb) loca- tions; MOZART4-I uses the gradient of the monthly mean mixing ratio between the grid cell where the observation site is located and the eight surrounding grid cells; and ACTMt42l67 uses the 3-D gradient (using the four horizon- tal and two vertical grid cells) surrounding the observation site. 2.2.3 Uncertainty estimates The prior ﬂux uncertainties were determined following the method usually used for the respective frameworks. For TM5-I, the uncertainties were calculated for each grid cell as 100 % of the annual mean prior value. For LMDZ4-I, the uncertainties were chosen for each grid cell as 100 % of the maximum of the eight surrounding grid cells plus the one of interest. These were used to form the variances and the square root of total of the prior error covariance matrix was scaled to be equal to 2 TgN yr−1. For TM3-I the uncertainties of the prior ﬂux were calculated as 100 % of the ﬂux per grid cell and month. Lastly, for ACTMt42l67-I and MOZART4- I, the regional uncertainty was chosen to be 100 and 50 %, respectively, of the regional emission. 2.2.4 Degrees of freedom The number of degrees of freedom in the inversion is an important factor for determining how closely the poste- rior ﬂuxes resemble the prior ones. For MOZART4-I and ACTMt42l67-I, which solve the inversion using coarse re- gions, the number of degrees of freedom is substantially re- duced, representing a strong constraint on the inversion as only the mean ﬂux in each region is optimized and the ﬂux pattern within each region remains as described a priori. On the other hand, solving for ﬁne regions i.e. at the resolution of the transport model, as in TM5-I, TM3-I and LMDZ4- I, beneﬁts from additional regularization constraints, such as spatial correlations of the prior ﬂux errors (used in the deﬁni- tion of B). For TM5-I the spatial correlation length (200 km) means that the grid cells are only weakly correlated to one another resulting in a weak constraint, whereas in LMDZ4-I, longer scale lengths are used (500 km for land and 1000 km for ocean) resulting in a stronger constraint (see Table 2). R. L. Thompson et al.: TransCom N2O model inter-comparison – Part 2 R. L. Thompson et al.: TransCom N2O model inter-comparison – Part 2 (CRU-NCEP, Climate Research Unit–National Centre for Environmental Prediction), N-fertilizer application, and at- mospheric N-deposition data and provides inter-annually varying estimates at monthly and 3.75◦× 2.5◦(longitude by latitude) resolution as described in Zaehle et al. (2011). For the ocean ﬂux, we used the Pelagic Interaction Scheme for Carbon and Ecosystem Studies (PISCES) ocean biogeo- chemistry model (Dutreuil et al., 2009), which provides inter- annually varying ﬂuxes at monthly and 1.0◦× 1.0◦resolu- tion. For waste, fuel combustion and industrial emissions, we used EDGAR-4.1 (Emission Database for Greenhouse gas and Atmospheric Research, available at: http://edgar.jrc.ec. europa.eu/index.php), which are estimated for the reference year 2005 and were provided annually at 1.0◦× 1.0◦reso- lution. Biomass burning estimates from GFED-2.1 (Global Fire Emissions Database) (van der Werf et al., 2010) were used, which were provided monthly and at 1.0◦× 1.0◦res- olution. In total, the global emission for 2005 to 2009 was 16.8, 16.3, 16.8, 16.2 and 16.4 TgN yr−1, respectively. ACTMt42l67 also assimilates the monthly mean mixing ra- tio and uses the standard error of the monthly mixing ratio multiplied by a factor of 2 to account for the underestimation of the variability in the model; and TM3-I assimilates the weekly mean mixing ratio but does not include an estimate of the temporal representation uncertainty. ACTMt42l67 also assimilates the monthly mean mixing ra- tio and uses the standard error of the monthly mixing ratio multiplied by a factor of 2 to account for the underestimation of the variability in the model; and TM3-I assimilates the weekly mean mixing ratio but does not include an estimate of the temporal representation uncertainty. Atmos. Chem. Phys., 14, 6177–6194, 2014 2.2.2 Prior ﬂuxes The prior N2O ﬂux was comprised from estimates of the dif- ferent sources, that is, from soils (including both natural and agricultural soils), ocean, biomass burning, waste, fuel com- bustion and industry (see Table 3). For soil ﬂuxes, we used the terrestrial biosphere model, O-CN (Zaehle and Friend, 2010), which is driven by reconstructed observed climate Atmos. Chem. Phys., 14, 6177–6194, 2014 www.atmos-chem-phys.net/14/6177/2014/ 6181 2.3 Atmospheric observations The temporal representation uncertainties are calculated as follows: TM5-I and LMDZ4-I assimilate afternoon/night- time mean mixing ratios for low-altitude/mountain sites, respectively, and use the standard deviation of the after- noon/night mean mixing ratio at each site; MOZART4-I assimilates the monthly mean mixing ratio and uses the standard error of the monthly mixing ratio at each site; These stations do not all use the same calibration scale and, thus, offsets exist between the measurements. Further- more, even in the case where the measurements are reported www.atmos-chem-phys.net/14/6177/2014/ www.atmos-chem-phys.net/14/6177/2014/ Atmos. Chem. Phys., 14, 6177–6194, 2014 R. L. Thompson et al.: TransCom N2O model inter-comparison – Part 2 Atmospheric observation sites using in the inversions. (F = Flask, C = Continuous). Altitude is speciﬁed as metres above sea level ID Station Operator Type Latitude Longitude Altitude (m a.s.l.) ALT Alert, Canada NOAA F 82.5◦N 62.5◦W 210 ASC Ascension Isl., UK NOAA F 7.9◦S 14.4◦W 54 ASK Assekrem, Algeria NOAA F 23.2◦N 5.4◦E 2728 AZR Azores, Portugal NOAA F 38.8◦N 27.4◦W 40 BAL Baltic Sea, Poland NOAA F 55.4◦N 17.2◦E 7 BIK Bialystok, Poland MPI-BGC C 55.3◦N 22.8◦E 460 BKT Bukit Kototabang, Indonesia NOAA F 0.2◦S 100.3◦E 865 BME St. Davis Head, Bermuda, UK NOAA F 32.4◦N 64.7◦W 30 BMW Tudor Hill, Bermuda, UK NOAA F 32.3◦N 64.9◦W 30 BRW Barrow, Alaska NOAA F 71.3◦N 156.6◦W 11 BSC Black Sea, Romania NOAA F 44.2◦N 28.7◦E 3 CBA Cold Bay, Alaska NOAA F 55.2◦N 162.7◦W 21 CBW Cabauw, Netherlands ECN C 52.0◦N 4.9◦E 118 CGO Cape Grim, Tasmania AGAGE C 40.7◦S 144.7◦E 164 CHR Christmas Isl. www.atmos-chem-phys.net/14/6177/2014/ Atmos. Chem. Phys., 14, 6177–6194, 2014 www.atmos-chem-phys.net/14/6177/2014/ NOAA F 1.7◦N 157.2◦W 3 COI Cape Ochi-ishi, Japan NIES C 43.2◦N 145.5◦E 45 CRZ Crozet Isl., France NOAA F 46.45◦S 51.9◦E 120 CVR Calhau, Cape Verde MPI-BGC F 16.9◦N 24.9◦W 10 EIC Easter Island, Chile NOAA F 27.2◦S 109.5◦W 50 GMI Mariana Isl., Guam NOAA F 13.4◦N 144.8◦E 2 HAT Hateruma, Japan NIES C 24.1◦N 123.8◦E 10 HBA Halley Stn., Antarctica NOAA F 75.6◦S 26.5◦W 30 HUN Hegyhatsal, Hungary ELTE C 46.9◦N 16.7◦E 344 ICE Vestmannaeyjar, Iceland NOAA F 63.3◦N 20.3◦W 118 IZO Tenerife, Spain NOAA F 28.3◦N 16.5◦W 2360 JFJ Jungfraujoch, Switzerland EMPA C 46.6◦N 8.0◦E 3580 KEY Key Biscayne, Florida NOAA F 25.7◦N 80.2◦W 3 KUM Cape Kumukahi NOAA F 19.5◦N 154.8◦W 3 KZD Sary Tauku, Kazakhstan NOAA F 44.1◦N 76.8◦E 601 LEF Park Falls, Wisconsin NOAA F 45.9◦N 90.3◦W 868 LLN Lulin, Taiwan NOAA F 23.5◦N 120.9◦E 2867 LUT Lutjewad, Netherlands RUG-CIO C 53.4◦N 6.4◦E 60 MHD Mace Head, Ireland AGAGE C 53.3◦N 9.9◦W 25 MLO Mauna Loa, Hawaii NOAA F 19.5◦N 155.6◦W 3397 NWR Niwot Ridge NOAA F 40.0◦N 105.5◦W 3526 NMB Gobabeb, Namibia NOAA F 23.6◦S 15.0◦E 456 OXK Ochsenkopf, Germany MPI-BGC C 50.1◦N 11.8◦E 1185 PAL Pallas, Finland FMI C 68.0◦N 24.1◦W 560 PSA Palmer Stn, Antarctica NOAA F 64.9◦S 64.0◦W 10 PTA Point Arena, California NOAA F 39.0◦N 123.7◦W 55 RPB Ragged Point, Barbados AGAGE C 13.2◦N 59.4◦W 45 SEY Mahé, Seychelles NOAA F 4.7◦S 55.2◦E 3 SHM Shemya Isl., Alaska NOAA F 52.7◦N 174.1◦E 40 SIS Shetland Isl., UK MPI-BGC F 59.9◦N 1.3◦W 46 SSL Schauinsland UBA C 47.9◦N 7.9◦E 1205 SMO Tutuila, American Samoa AGAGE C 14.3◦S 170.6◦W 42 SPO South Pole, Antarctica NOAA F 89.98◦S 24.8◦W 2810 STM ocean stn. M, Norway NOAA F 66.0◦N 2.0◦E 7 SUM Summit, Greenland NOAA F 72.6◦N 38.5◦W 3238 SYO Syowa Stn., Antarctica NOAA F 69.0◦S 39.6◦E 11 TAP Tae-ahn Peninsula, Taiwan NOAA F 36.7◦N 126.1◦E 20 TDF Tierra del Fuego, Argentina NOAA F 54.9◦S 68.5◦W 20 THD Trinidad Head, California AGAGE C 41.1◦N 124.2◦W 107 TTA Grifﬁn, UK UEDIN C 56.6◦N 3.0◦W 535 R. L. Thompson et al.: TransCom N2O model inter-comparison – Part 2 R. L. Thompson et al.: TransCom N2O model inter-comparison – Part 2 6182 R. L. Thompson et al.: TransCom N2O model inter-comparison – Part 2 Table 4. Atmospheric observation sites using in the inversions. (F = Flask, C = Continuous). www.atmos-chem-phys.net/14/6177/2014/ Altitude is speciﬁed as metres above sea level (m a.s.l.). 6182 Table 4. Atmospheric observation sites using in the inversions. (F = Flask, C = Continuous). Altitude is speciﬁed as metres above sea level (m a.s.l.). Atmos. Chem. Phys., 14, 6177–6194, 2014 R. L. Thompson et al.: TransCom N2O model inter-comparison – Part 2 6183 Table 4. Continued. ID Station Operator Type Latitude Longitude Altitude (m a.s.l.) UUM Ulaan-Uul, Mongolia NOAA F 44.5◦N 111.1◦E 914 WIS Negev Desert, Israel NOAA F 31.1◦N 34.9◦E 400 WKT Moody, Texas NOAA F 31.3◦N 97.3◦W 708 WLG Mt. Waliguan, China NOAA F 36.3◦N 100.9◦E 3810 ZEP Ny-Ålesund NOAA F 78.9◦N 11.88◦E 475 ZOT Zotto, Russia MPI-BGC F 60.8◦N 89.4◦E 415 Table 5. A priori and a posteriori calibration offsets (ppb) relative to the NOAA2006A scale. Note that only LMDZ4-I and TM5-I included the opimization of calibration offsets and only TM5 calculated these annually (the range over all years is given in brackets for TM5-I). R. L. Thompson et al.: TransCom N2O model inter-comparison – Part 2 6183 Table 5. A priori and a posteriori calibration offsets (ppb) relative to the NOAA2006A scale. Note that only LMDZ4-I and TM5-I included the opimization of calibration offsets and only TM5 calculated these annually (the range over all years is given in brackets for TM5-I). ID Prior TM5-I LMDZ4-I ID Prior TM5-I LMDZ4-I BIK 0.06 0.22 (0.00–0.47) 0.13 PAL 0.50 0.00 (0.0–0.0) 0.32 CBW 0.27 0.52 (0.25–0.76) 0.84 MHD 0.25 0.08 (0.25–0.76) 0.05 HUN 1.08 0.45 (0.24–0.59) 0.44 THD −0.30 0.04 (-0.01–0.07) 0.28 LUT −3.0 −1.2 (−2.0–0.0) −2.0 RPB 0.00 −0.11 (−0.21–0.0) 0.07 OXK 0.39 0.77 (0.0–1.28) 1.13 SMO 0.20 0.24 (0.14–0.37) 0.36 TTA 0.00 0.56 (0.0–1.03) 0.65 CGO 0.20 0.08 (0.0–0.13) 0.00 JFJ 0.00 −0.47 (−0.69—0.34) −0.60 NIES −0.60 0.00 (0.0–0.0) −0.41 SSL 0.00 0.30 (0.07–0.50) 0.17 MPI-BGC 0.00 0.38 (0.19–0.54) 0.47 Figure 1. Map of surface sites for atmospheric N2O observations. tive to the NOAA-2006A scale (see Table 5). Since the inter- calibration data were not complete for all times and all sites, the offsets were included into the optimization problem in in- version frameworks with this capacity (i.e. in LMDZ4-I and TM5-I, and only TM5-I resolves the offsets temporally us- ing annual resolution). In this case, the best estimates of the offsets were used as prior values. In the case that they could not be optimized (i.e. 3.1 Validation with atmospheric observations Figure 1. Map of surface sites for atmospheric N2O observations. www.atmos-chem-phys.net/14/6177/2014/ in MOZART4-I, ACMTt42l67-I, and TM3-I) the given values were used to correct the observa- tions prior to the inversion. 3.1.1 Meridional gradients Meridional gradients are some of the most commonly used observational parameters to assess CTMs, as they provide a constraint on features such as inter-hemispheric transport and latitudinal ﬂux distributions (Gloor et al., 2007; Patra et al., 2011). Figure 2 shows the observed annual mean meridional mole fractions (2006 to 2009) compared with simulations on the same scale, there still may be offsets owing to sys- tematic errors. These offsets can introduce signiﬁcant errors in the optimized ﬂuxes if they are not accounted for prior to, or in, the inversion. For this reason, calibration offsets were estimated using inter-calibration data for each of the in situ stations, and for the three MPI-BGC ﬂask sites together, rela- 3.1.2 Seasonal cycles In Part 1 of the inter-comparison, considerable attention was paid to the seasonal cycle of N2O as this is sensitive to STE, the height of the PBL, inter-hemispheric mixing, and season- ality in the ﬂuxes. Figure 4 shows the annual mean (2006 to 2009) seasonal cycles from the posterior model simula- tions and observations at six key background sites. In the NH mid- to high latitudes, i.e. at MHD and BRW, the phase and amplitude are reasonably well captured by ACTMt42l67-I, TM5-I and LMDZ4-I with a minimum occurring in August, whereas MOZART4-I and TM3-I simulate a too early mini- mum at both sites by up to 2.5 months, as was also the case for all CTMs a priori. However, all ﬁve CTMs participating in Part 2 were able to capture the correct phase when using an alternative prior ﬂux estimate with no terrestrial biosphere seasonal cycle (see Part 1, Thompson et al., 2014b) suggest- ing that the reason for the too early minimum was not re- lated to transport problems but rather to the seasonality in the ﬂuxes. This also seems to be the case for MOZART4-I and TM3-I, which have the smallest shift in the seasonal cy- cle relative to the prior ﬂuxes (this is discussed in more detail in Sect. 3.2.3). At MLO, all CTMs simulate a too early mini- mum as was also the case using the a priori emissions. How- ever, with the a posteriori emissions, the amplitude is closer to that observed. The timing of the minimum, in April, in the models is consistent with the expected maximum inﬂuence of stratospheric air in the troposphere owing to the down- ward branch of the Brewer–Dobson circulation, which has a maximum in December to February in the NH. However, the fact that the observed minimum occurs later may suggest that Gradients in the pressure-weighted column mean N2O were also compared against observations from HIPPO (Hi- aper Pole-to-Pole Observations, http://hippo.ucar.edu) cam- paigns in January and November 2009 (Fig. 3). In con- trast to the surface, the simulations all underestimate the to- tal column inter-hemispheric gradient in January by about 1 ppb (circa 50 %). In November, the inter-hemispheric gra- dient is smaller and is matched more closely by the mod- els; however, there is an overall offset of about 1 ppb (ex- cept MOZART4-I where its 1.5 ppb offset compensates). www.atmos-chem-phys.net/14/6177/2014/ Atmos. Chem. Phys., 14, 6177–6194, 2014 R. L. Thompson et al.: TransCom N2O model inter-comparison – Part 2 6184 Figure 2. Comparison of the annual mean meridional N2O mole fraction (ppb) from the posterior simulations with that from surface observations (average 2006–2009). The grey shaded area shows the range of values for all models using the prior ﬂuxes. (Legend: obser- vations, black; MOZART4-I, orange; ACTMt42l67-I, green; TM5- I, blue; TM3-I, red; LMDZ4-I, magenta). from the HIPPO campaigns, ﬁnding that the simulations us- ing the optimized ﬂuxes underestimated the column mean N2O but if the total column was assimilated, larger tropi- cal ﬂuxes were obtained and the total column matched the observations. Kort et al. (2011) reasoned that this was be- cause the surface network failed to detect high N2O sig- nals in the tropics and that these were lofted to higher al- titudes with strong tropical convection. If the ﬂuxes in this study were underestimated for this reason, then this would result in a too low growth rate of N2O in the troposphere. However, all models capture the observed growth rate within 0.17 ppb yr−1 (20 %) and most within 0.1 ppb yr−1 (10 %) (Fig. S3). The simulated upper troposphere values of N2O in January may also be underestimated due to model transport errors such as too strong STE as was suggested in Part 1 of the inter-comparison (Thompson et al., 2014b), which would be much more apparent in the mole fractions above the plan- etary boundary layer (PBL) and is consistent with what we ﬁnd in the comparisons up to 2000 m versus up to 10 000 m. If this were the case, and if no bias correction were applied to account for the transport error, then assimilating observations in the upper troposphere may lead to a systematic overesti- mate of the emissions. Figure 2. Comparison of the annual mean meridional N2O mole fraction (ppb) from the posterior simulations with that from surface observations (average 2006–2009). The grey shaded area shows the range of values for all models using the prior ﬂuxes. (Legend: obser- vations, black; MOZART4-I, orange; ACTMt42l67-I, green; TM5- I, blue; TM3-I, red; LMDZ4-I, magenta). by each CTM integrated with the corresponding posterior ﬂuxes. For both the observations and the simulations, the gradients were calculated from detrended and deseasonal- ized N2O mole fractions at background surface sites. www.atmos-chem-phys.net/14/6177/2014/ For each model, a very good agreement was found with the gra- dient derived from surface observations (correlation coefﬁ- cient R2 ≥0.9 for each model). In MOZART4-I, the mean mole fraction is approximately 1.5 ppb higher, which is most likely due to too high mole fractions in the initial conditions (see also Fig. S1 in the Supplement), but it still captures the gradient reasonably well. 3.1.2 Seasonal cycles For a f h i bb i i T bl 4 (L d b i bl k MOZART4 I the inﬂuence of stratosphere to troposphere transport (STT) is overestimated in the models and/or that the seasonality is still not correct in the ﬂuxes at the latitude of MLO. 2 4 6 8 10 −0.3 −0.2 −0.1 0.0 0.1 0.2 0.3 N2O [ppb] BRW 2 4 6 8 10 −0.3 −0.2 −0.1 0.0 0.1 0.2 0.3 MHD 2 4 6 8 10 −0.3 −0.2 −0.1 0.0 0.1 0.2 0.3 MLO 2 4 6 8 10 −0.3 −0.2 −0.1 0.0 0.1 0.2 0.3 N2O [ppb] SMO 2 4 6 8 10 −0.3 −0.2 −0.1 0.0 0.1 0.2 0.3 CGO 2 4 6 8 10 −0.3 −0.2 −0.1 0.0 0.1 0.2 0.3 SPO Months Months Months Figure 4. Comparison of the mean (2006–2008) observed and sim- ulated seasonal cycles (using the posterior ﬂuxes) in N2O mole frac- tion (ppb) at selected key sites. The grey shading indicates the range of uncertainty (1σ standard deviation) in the observations. For a de- scription of the site abbreviations see Table 4. (Legend: observa- tions, black; MOZART4-I, orange; ACTMt42l67-I, green; TM5-I, blue; TM3-I, red; LMDZ4-I, magenta). 2 4 6 8 10 −0.3 −0.2 −0.1 0.0 0.1 0.2 0.3 N2O [ppb] BRW 2 4 6 8 10 −0.3 −0.2 −0.1 0.0 0.1 0.2 0.3 MHD 2 4 6 8 10 −0.3 −0.2 −0.1 0.0 0.1 0.2 0.3 MLO For the Southern Hemisphere sites, SMO and CGO, all models agree well with the observed seasonal cycles except MOZART4-I at SMO and TM3-I at CGO. At SPO, however, all models underestimate the amplitude and MOZART4-I and TM3-I are also out of phase. It has been shown with N2O isotope measurements that the seasonality at CGO is determined by the combined inﬂuences of STT and ocean ﬂuxes leading to the observed minimum in April (Park et al., 2012). With the a priori ﬂuxes, both TM3-I and LMDZ4-I had the phase of the seasonal cycle at CGO out by nearly 6 months indicating a problem with STT in the Southern Hemisphere (see Thompson et al., 2014b). A similar error in MOZART4-I was observed at SPO as well. 3.1.2 Seasonal cycles The offset in November may be in part due to a calibration dif- ference between HIPPO and the NOAA data, which were used in the inversion, as comparisons of the HIPPO data be- tween 0 and 2000 m around 19◦N and 14◦S with the NOAA data at Mauna Loa (19.5◦N, 155.6◦W) and Samoa (14.3◦S, 170.6◦E), respectively, show an offset of about 0.5 ppb. The underestimate of the gradient in January may be due to the models underestimating N2O mole fractions in the upper tro- posphere as the agreement with the observed column is much better up to 2000 m, within a few tenths of a ppb (except north of about 50◦N) (Fig. S2 in the Supplement). Kort et al. (2011) obtained a similar result when they assimilated only surface data (i.e. within 250–750 m above sea level) www.atmos-chem-phys.net/14/6177/2014/ Atmos. Chem. Phys., 14, 6177–6194, 2014 ompson et al.: TransCom N2O model inter-comparison – Part 2 R. L. Thompson et al.: TransCom N2O model inter-comparison – Part 2 6185 Figure 3. Comparison of model simulations (using the posterior ﬂuxes) with observations of N2O mole fraction (ppb) from surface sites (top row) and pressure-weighted column averages (up to 10 000 m) from HIPPO aircraft proﬁles (bottom row) for January (left) and November (right) 2009. (Legend: observations, black; MOZART4-I, orange; ACTMt42l67-I, green; TM5-I, blue; TM3-I, red; LMDZ4-I, magenta). omparison of the mean (2006 – 2008) observed and simulated seasonal cycles osterior fluxes) in N2O mole fraction (ppb) at selected key sites. The grey shaded he range of uncertainty (1σ standard deviation) in the observations. For a of the site abbreviations see Table 4 (Legend: observations black; MOZART4 I Figure 3. Comparison of model simulations (using the posterior ﬂuxes) with observations of N2O mole fraction (ppb) from surface sites (top row) and pressure-weighted column averages (up to 10 000 m) from HIPPO aircraft proﬁles (bottom row) for January (left) and November (right) 2009. (Legend: observations, black; MOZART4-I, orange; ACTMt42l67-I, green; TM5-I, blue; TM3-I, red; LMDZ4-I, magenta). omparison of the mean (2006 – 2008) observed and simulated seasonal cycles sterior fluxes) in N2O mole fraction (ppb) at selected key sites. The grey shaded e range of uncertainty (1σ standard deviation) in the observations. 3.1.2 Seasonal cycles However, a posteriori, LMDZ4-I has a much-improved ﬁt to the phase at CGO and SPO, which was achieved by increasing the ampli- tude of the ﬂux seasonality in the Southern Ocean, whereas TM3-I and MOZART4-I make nearly no adjustment (this is discussed further in Sect. 3.2.3). 2 4 6 8 10 −0.3 −0.2 −0.1 0.0 0.1 0.2 0.3 N2O [ppb] SMO 2 4 6 8 10 −0.3 −0.2 −0.1 0.0 0.1 0.2 0.3 CGO 2 4 6 8 10 −0.3 −0.2 −0.1 0.0 0.1 0.2 0.3 SPO Months Months Months Figure 4. Comparison of the mean (2006–2008) observed and sim- ulated seasonal cycles (using the posterior ﬂuxes) in N2O mole frac- tion (ppb) at selected key sites. The grey shading indicates the range of uncertainty (1σ standard deviation) in the observations. For a de- scription of the site abbreviations see Table 4. (Legend: observa- tions, black; MOZART4-I, orange; ACTMt42l67-I, green; TM5-I, blue; TM3-I, red; LMDZ4-I, magenta). 3.2 Comparison of posterior emissions In this section, we present a comparison of the posterior emission estimates. All posterior emissions were compared after they were interpolated from the corresponding model grid to 1◦× 1◦resolution. Atmos. Chem. Phys., 14, 6177–6194, 2014 3.2.1 Global means Table 6 shows the global total emission a priori and the global total emission and sink a posteriori calculated by each in- version framework. On the basis of the posterior emissions, the inversions can be grouped into two categories: (1) those with low global totals, i.e. MOZART4-I, ACTMt42l67-I and TM3-I and (2) those with high global totals, i.e. TM5-I and LMDZ4-I, where low and high are deﬁned relative to the prior. In the case of MOZART4-I, the low global total (the lowest of all inversions) results from the overestimate of N2O mole fractions in the initial conditions, which leads to the emissions being underestimated and a too low atmospheric N2O growth rate (see Fig. S3 in the Supplement). For this reason, the MOZART4-I estimates are not included further in the ﬂux totals. However, in general, the low/high catego- rization also corresponds to how the observations were as- similated in the inversion; the ﬁrst category inversions assim- ilate monthly (MOZART4-I and ACTMt42l67-I) or weekly (TM3-I) means, while those in the second category (TM5-I and LMDZ4-I) use the afternoon means for sites within the PBL and night-time means for mountain sites. This feature of the category 2 inversions means that they are also sensi- tive to the synoptic variability of the observations, while in the category 1 inversions this signal is smoothed out. Fur- thermore, the category 1 inversions may overestimate the monthly/weekly mean N2O mole fraction a priori as com- pared to the NOAA CCGG ﬂask measurements, since the ﬂask samples are generally collected during meteorolog- ical conditions corresponding to background air, whereas in the model, it is the monthly/weekly mean of all data. MOZART4-1 and ACTMt42l67-1, which have the lowest global total estimates, also differ from the other inversions in that they solve for emissions in large regions as opposed to solving the emissions at the resolution of the transport model. All inversion frameworks had very similar global total sinks, within less than 1 TgN yr−1 of each other for each year, thus differences in the calculated loss rate is not a reason for the differences in global total emissions. To better examine the differences between the a posteriori emissions, we compare the annual mean zonally integrated emissions plotted against latitude and the accumulated emis- sions from south to north (Fig. 8). www.atmos-chem-phys.net/14/6177/2014/ R. L. Thompson et al.: TransCom N2O model inter-comparison – Part 2 6186 Table 6. Overview of the prior and posterior global annual total source (upper panel) and sink (lower panel) (both in TgN yr−1). Table 6. Overview of the prior and posterior global annual total source (upper panel) and sink (lower panel) (both in TgN yr−1). e prior and posterior global annual total source (upper panel) and sink (lower panel) (both in TgN yr−1). Table 6. Overview of the prior and posterior global annual total source (upper panel) and sink (lower panel) p p g ( pp p ) ( p ) ( Year Prior MOZART4-I ACTMt42l67-I TM5-I TM3-I LMDZ4-I 2006 16.3 14.1 16.0 16.9 15.1 17.6 2007 16.8 15.6 16.7 16.9 16.6 19.1 2008 16.2 15.7 16.5 17.2 16.4 19.4 2009 16.4 14.4 15.5 15.4 15.6 18.8 2006 – 12.8 11.9 12.2 12.4 12.7 2007 – 12.6 12.4 12.6 12.4 12.7 2008 – 12.6 12.5 12.4 12.5 12.5 2009 – 12.7 13.0 12.4 12.6 12.7 3.2.1 Global means highest emissions were found in the sub cal regions of South America Africa a highest emissions were found in the subtropical and tropi- cal regions of South America, Africa and Asia, in Europe and the eastern states of the USA. However, the inversions differ in the relative importance of emissions in each of these sub-continental regions. Figure 6 shows the annual mean ﬂux increments made by each inversion, i.e. the posterior minus prior annual mean ﬂux. There are a number of features in the increments that are common to all inversions: (1) lower (relative to the prior) emissions in temperate land regions in the SH, (2) higher emissions in central Europe, (3) higher emissions in central Africa and (4) no signiﬁcant change in northern Eurasia and Canada. On the other hand, the inver- sions differ signiﬁcantly in the direction and/or magnitude of the ﬂux increments for the USA (eastern states), South and East Asia, and tropical South America. This information is summarized in Fig. 7, which shows the median absolute deviation (MAD) of the annual mean emissions from all ﬁve inversions. Regions with highest MAD correspond to regions with the greatest discrepancy among the inversions. highest emissions were found in the subtropical and tropi- cal regions of South America, Africa and Asia, in Europe and the eastern states of the USA. www.atmos-chem-phys.net/14/6177/2014/ However, the inversions differ in the relative importance of emissions in each of these sub-continental regions. Figure 6 shows the annual mean ﬂux increments made by each inversion, i.e. the posterior minus prior annual mean ﬂux. There are a number of features in the increments that are common to all inversions: (1) lower (relative to the prior) emissions in temperate land regions in the SH, (2) higher emissions in central Europe, (3) higher emissions in central Africa and (4) no signiﬁcant change in northern Eurasia and Canada. On the other hand, the inver- sions differ signiﬁcantly in the direction and/or magnitude of the ﬂux increments for the USA (eastern states), South and East Asia, and tropical South America. This information is summarized in Fig. 7, which shows the median absolute deviation (MAD) of the annual mean emissions from all ﬁve inversions. Regions with highest MAD correspond to regions with the greatest discrepancy among the inversions. 3.2.1 Global means By plotting the emissions in this way, differences in the latitudinal distribution of the emissions are more apparent and may be assessed in terms of different features of the CTMs used in the inversions, such as the rate of inter-hemispheric and vertical mixing. Mov- ing from south to north, one can see that all inversions esti- mate lower emissions compared to the prior in the Southern Hemisphere; it is only north of the Equator that some of the inversions have a higher accumulated emission. TM5-I has the highest emission estimate for the Southern Hemisphere tropics and is also the most southern crossing point with the prior accumulated emissions. This is likely related to the fact that TM5-I has a long inter-hemispheric exchange time (1.7 years compared to the observed 1.4 years in 2006, based on SF6 mole fractions at BRW, MLO, CGO and SPO; Patra et al., 2011), which would mean that in order to match the ob- served atmospheric N2O mole fraction in the Southern Hemi- sphere tropics, higher emissions in this region are required. It can be expected that TM5-I would also for this reason Overall, the global distribution of N2O emissions was sim- ilar in all inversions and close to that a priori (Fig. 5). The www.atmos-chem-phys.net/14/6177/2014/ Atmos. Chem. Phys., 14, 6177–6194, 2014 6187 R. L. Thompson et al.: TransCom N2O model inter-comparison – Part 2 Figure 5. Maps of annual mean posterior and prior N2O ﬂux (gN m−2 yr−1) for 2006–2008. Figure 5. Maps of annual mean posterior and prior N2O ﬂux (gN m− yr−1) for 2006–2008 Figure 5. Maps of annual mean posterior and prior N2O ﬂux (gN m−2 yr−1) for 2006–2008. Table 7. Annual mean (2006–2008) regional N2O emission estimates (TgN yr−1). Values for which the inversions differ on the direction of the change with respect to the prior are shown in parentheses (MAD = median absolute deviation). Table 7. Annual mean (2006–2008) regional N2O emission estimates (TgN yr−1). Values for which the inversions differ on the direction of the change with respect to the prior are shown in parentheses (MAD = median absolute deviation). Atmos. Chem. Phys., 14, 6177–6194, 2014 3.2.2 Regional means estimate lower emissions in the Northern Hemisphere trop- ics and subtropics. However, the accumulated emissions still exceed those of e.g. LMDZ until circa 30◦N. The reason for this cannot be determined from these results alone but it may be at least in part also owing to transport errors in LMDZ. At circa 30◦N, LMDZ4-I surpasses both the prior and TM5- I accumulated emissions owing to very large emission es- timates in the Northern Hemisphere subtropics. LMDZ4-I (in the present 19-layer conﬁguration) has a relatively short inter-hemispheric exchange time, 1.2 years in 2006 (Patra et al., 2011), and has been found to have a very diffusive PBL in the Northern Hemisphere midlatitudes (Geels et al., 2007). These features likely lead to too high emissions in the north- ern subtropics and midlatitudes. North of circa 50◦N, the zonally integrated emission differs very little among the in- versions and the prior, however, the accumulated total emis- sion at 90◦N differs owing to the aforementioned disparities. estimate lower emissions in the Northern Hemisphere trop- ics and subtropics. However, the accumulated emissions still exceed those of e.g. LMDZ until circa 30◦N. The reason for this cannot be determined from these results alone but it may be at least in part also owing to transport errors in LMDZ. At circa 30◦N, LMDZ4-I surpasses both the prior and TM5- I accumulated emissions owing to very large emission es- timates in the Northern Hemisphere subtropics. LMDZ4-I (in the present 19-layer conﬁguration) has a relatively short inter-hemispheric exchange time, 1.2 years in 2006 (Patra et al., 2011), and has been found to have a very diffusive PBL in the Northern Hemisphere midlatitudes (Geels et al., 2007). These features likely lead to too high emissions in the north- ern subtropics and midlatitudes. North of circa 50◦N, the zonally integrated emission differs very little among the in- versions and the prior, however, the accumulated total emis- sion at 90◦N differs owing to the aforementioned disparities. Figure 9 shows the annual mean total emissions for seven sub-continental and three ocean regions from each of the in- versions and the prior, the corresponding range, median, and MAD of the emissions, as well as the uncertainty calculated from a single inversion model (LMDZ-I), are given in Ta- ble 7. 3.2.1 Global means Region Prior Posterior LMDZ4-I Range Median MAD uncertainty Ocean 90–30◦S 1.49 0.92–1.34 1.08 0.20 0.39 Ocean 30◦S–30◦N 3.30 3.25–3.69 (3.66) 0.03 0.61 Ocean 30–90◦N 0.95 1.13–1.29 1.20 0.08 0.32 S + Tr America 2.55 1.99–2.62 (2.33) 0.27 1.13 N America 1.00 0.65–1.29 (0.74) 0.11 0.28 Africa 3.07 3.23–3.40 3.36 0.04 0.70 Europe 0.80 0.84–1.20 1.04 0.20 0.19 N Asia 0.40 0.31–0.67 (0.40) 0.09 0.42 S Asia 2.91 2.56–3.81 (2.85) 0.28 0.77 Australasia 0.39 0.27–0.36 0.31 0.01 0.23 www.atmos-chem-phys.net/14/6177/2014/ Atmos. Chem. Phys., 14, 6177–6194, 2014 www.atmos-chem-phys.net/14/6177/2014/ 6188 R. L. Thompson et al.: TransCom N2O model inter-comparison – Part 2 Figure 6. Maps of annual mean ﬂux increments for 2006–2008 (gN m−2 yr−1). Negative values (blue) indicate posterior ﬂuxes that are lower than the prior ﬂuxes and vice versa for positive ones (yellow-red). Figure 6. Maps of annual mean ﬂux increments for 2006–2008 (gN m−2 yr−1). Negative values (blue) indicate posterior ﬂuxes that are lower than the prior ﬂuxes and vice versa for positive ones (yellow-red). 3.2.2 Regional means The calculated uncertainties per region are larger than the corresponding MAD values, indicating that the spread of posterior emissions is smaller than the uncertainty calculated for a single inversion. For only three out of the seven land re- gions is there a signiﬁcant change in emissions with respect to the prior. Here, we deﬁne signiﬁcant to mean that: (1) all inversions agree on the direction of the change and (2) the prior value is outside the range of the posterior median and plus or minus MAD. These regions are Africa, Europe and Australasia. For Australasia, the contribution to the global www.atmos-chem-phys.net/14/6177/2014/ Atmos. Chem. Phys., 14, 6177–6194, 2014 ompson et al.: TransCom N2O model inter-comparison – Part 2 R. L. Thompson et al.: TransCom N2O model inter-comparison – Part 2 6189 Figure 9. Annual mean (2006–2008) regional emission estimates (TgN yr−1) for the seven land regions (ﬁrst two rows) and three ocean regions (last row). The colours refer to the different inversion frameworks as indicated in the legend and the dashed line is the median of the posterior emissions. Figure 7. Map of median absolute deviation (MAD) of annual mean ﬂuxes (gN m−2 yr−1). The colour scale from white to red shows indicates increasing MAD values. Figure 7. Map of median absolute deviation (MAD) of annual mean ﬂuxes (gN m−2 yr−1). The colour scale from white to red shows indicates increasing MAD values. Figure 9. Annual mean (2006–2008) regional emission estimates (TgN yr−1) for the seven land regions (ﬁrst two rows) and three ocean regions (last row). The colours refer to the different inversion Figure 7. Map of median absolute deviation (MAD) of annual mean ﬂuxes (gN m−2 yr−1). The colour scale from white to red shows indicates increasing MAD values. Figure 9. Annual mean (2006–2008) regional emission estimates (TgN yr−1) for the seven land regions (ﬁrst two rows) and three ocean regions (last row). The colours refer to the different inversion frameworks as indicated in the legend and the dashed line is the median of the posterior emissions. Figure 8. Zonally integrated annual mean (2006–2008) ﬂuxes (top) and accumulated from south to north (bottom). (Legend: prior, grey; MOZART4-I, orange; ACTMt42l67-I, green; TM5-I, blue; TM3-I, red; LMDZ4-I, magenta). to the global total emission. 3.2.2 Regional means Africa was also found to have higher emissions relative to the prior (by 10 %, equivalent to 0.29 TgN yr−1) and contributes 20 % to the global total emis- sion. Of the regions where the change was not considered sig- niﬁcant, North America as well as South and Tropical Amer- ica still satisﬁed the second criterion. For North America, all inversions except LMDZ4-I estimated lower emissions (by 26 %, equivalent to 0.26 TgN yr−1), bringing its contribution to the global total to 4 %, and for South and Tropical Amer- ica, all inversions except TM5-I estimate lower emissions (by 9 %, equivalent to 0.22 TgN yr−1) bringing its contribution to the global total to 14 %. For South Asia and North Asia, however, the inversions differed signiﬁcantly both in the di- rection of change as well as in the magnitude. While the total emission from North Asia is small (2 % of the global), that from South Asia is very important (approximately 17 % of the global). There are several reasons why the inversions differ so sub- stantially for South Asia. First, this region is not well covered by the observation network. Emissions from this region are only constrained by the two in situ sites, HAT and COI, and by the discrete sampling sites, BKT, GMI, LLN, and TAP. Second, since the prior ﬂux uncertainties are calculated pro- portionally to the prior ﬂux, the prior uncertainty for this re- gion is large allowing the inversions considerable freedom to adjust the ﬂuxes here. Lastly, differences in the modelled transport, such as the tropical convection, monsoon ﬂow, and shifts in the North Paciﬁc storm track, which are important in determining outﬂow from the Asian continent (Stohl et al., 2002), may also contribute to the disparity among emis- sion estimates for South Asia. Stohl et al. (2002) showed that tracers emitted in Asia south of 30◦N, particularly in India, are readily transported toward the Intertropical Convergence Figure 8. Zonally integrated annual mean (2006–2008) ﬂuxes (top) and accumulated from south to north (bottom). (Legend: prior, grey; MOZART4-I, orange; ACTMt42l67-I, green; TM5-I, blue; TM3-I, red; LMDZ4-I, magenta). total (2 %, median posterior value) and the absolute change relative to the prior (0.08 TgN yr−1) are very small, and thus this region is not discussed further. 3.2.2 Regional means Europe was found to have 30 % (0.24 TgN yr−1) higher emissions than estimated a pri- ori and contributes on average 6 % (median posterior value) www.atmos-chem-phys.net/14/6177/2014/ Mean (2006–2008) seasonal cycle in N2O ﬂux (TgN yr−1) for each of the seven sub-continental regions (ﬁrst two rows) and three ocean regions (last row). Zone and thus could be one reason why LMDZ4-I, with a fast inter-hemispheric mixing rate, predicts the highest emis- sions for South Asia. Similar reasoning also applies to the large discrepancy for South and Tropical America. South and Tropical America is very poorly covered by the observation network (see Fig. 1) and the prior ﬂux uncertainty for this re- gion is very large. The posterior emission estimates for this region are also likely to be sensitive to features of the mod- elled transport, in particular, convective transport. 39 For the region of South Asia, there is some indica- tion in the posterior ﬂuxes of a double maximum, i.e. in ACTMt42l67-I, TM5-I and LMDZ4-I occurring in April and September. This approximately corresponds to the start and end of the Asian monsoon season, which lasts from April to September, while the period of lowest ﬂuxes, from Octo- ber to March, corresponds to the cool–dry season. This is in accordance with what has been found from in situ ﬂux mea- surements in subtropical Southern China, which experiences annual monsoons – that is, that WFPS, soil NO− 3 and NH+ 4 content, and N2O ﬂuxes were signiﬁcantly higher in the hot– humid season than in the cool–dry season (Lin et al., 2010). However, the peak in spring may also partially be an artefact needed to compensate for the too low simulated spring atmo- spheric mole fraction as compared to the observations owing to a too strong inﬂuence of STT. Unlike for the land regions, there is reasonably good agreement among the inversions for the ocean regions. All ocean regions satisfy the second criterion (i.e. the prior value is outside the range of the posterior median and MAD), and only the region 30◦S–30◦N does not also satisfy the ﬁrst criterion (i.e. that all inversions agree on the direction of the change). The emissions for the Southern Ocean (90–30◦S) were found to be smaller than estimated a priori, contribut- ing 6 % (median posterior value) to the global total, while for the tropical (30◦S–30◦N) and northern (30–90◦N) ocean regions, the emissions were found to be larger, contributing 22 % and 7 % to the global total, respectively. www.atmos-chem-phys.net/14/6177/2014/ Atmos. Chem. Phys., 14, 6177–6194, 2014 R. L. Thompson et al.: TransCom N2O model inter-comparison – Part 2 6190 tal region 0 1 2 3 4 5 N2O [TgN y−1] 2 4 6 8 10 S + Tr America 2 4 6 8 10 N America 2 4 6 8 10 Africa 2 4 6 8 10 Europe 0 1 2 3 4 5 N2O [TgN y−1] 2 4 6 8 10 N Asia 2 4 6 8 10 S Asia 2 4 6 8 10 Australasia PRIOR MOZART4 ACTM42L67 TM5 TM3 LMDZ4 −1 0 1 2 3 4 5 N2O [TgN y−1] 2 4 6 8 10 2 4 6 8 10 90 − 30S 2 4 6 8 10 2 4 6 8 10 30S − 30N 2 4 6 8 10 2 4 6 8 10 30 − 90N PRIOR MOZART4 ACTM42L67 TM5 TM3 LMDZ4 Figure 10. Mean (2006–2008) seasonal cycle in N2O ﬂux (TgN yr−1) for each of the seven sub-continental regions (ﬁrst two rows) and three ocean regions (last row). prior. ACTMt42l67-I and LMDZ4-I both estimate an ear- lier and broader maximum, between April and June, while MOZART4-I, TM5-I and TM3-I predict a broader maximum between June and July. In Part 1 of the inter-comparison, it was shown in the CTM integrations using ﬂuxes with a late summer maximum worsened the ﬁt to the atmospheric ob- servations compared to using ﬂuxes with no seasonal cycle. The result for the Northern Hemisphere temperate regions in this study conﬁrms the hypothesis in Part 1, that elevated emissions begin earlier in spring and continue until autumn without a peak in late summer. This is in line with what is ex- pected based on the dependence of N2O ﬂuxes on soil mois- ture (measured by water-ﬁlled pore space, WFPS), soil tem- perature and the availability of nitrogen substrates, particu- larly NO− 3 and NH+ 4 in soils (Butterbach-Bahl et al., 2013, and references therein). N2O ﬂux is maximized with WFPS of between approximately 70–90 % and has positive corre- lation with soil temperature (Smith et al., 1998). Therefore, low soil N2O ﬂux is expected throughout winter and higher N2O is expected in summer so long as there is sufﬁcient soil moisture and nitrogen substrate. N-fertilization usually occurs in spring and mid-summer providing sufﬁcient nitro- gen substrate, but drier soils in late summer may limit N2O ﬂuxes. Figure 10. www.atmos-chem-phys.net/14/6177/2014/ For the Southern Hemisphere regions of South and Trop- ical America and Africa, there is very little seasonality in the prior ﬂuxes. However, all of the inversions estimate a March–April minimum for South and Tropical America and similarly (except LMDZ4-I) for Africa. For South and Trop- ical America, the March–April minimum is not easy to ex- plain in terms of soil N2O ﬂuxes. In fact, from the few existing regional measurements of N2O ﬂuxes in tropical South America only a small seasonal cycle has been observed with elevated ﬂuxes during the wet season from March–May (D’Amelio et al., 2009). Therefore, it is likely that the mini- mum in the optimized ﬂuxes is due to transport errors since the timing of the atmospheric N2O minimum in April, de- termined to a large extent by STT, is not captured by the models. Thus to match observations, the inversions estimate 3.2.3 Seasonal variability At the semi-hemisphere scale, we ﬁnd a few important differences between our me- dian estimates and previous ones: for the region 90◦S to 30◦S we estimate higher emissions (7 % of the global to- tal), for the region 0 to 30◦N we estimate lower emissions (41 %), and for 30–90◦N slightly higher emissions (23 %). Comparing the ratio for emissions in the regions 0–30◦N and 30–90◦N, all our inversions give a lower value (from 1.5 to 1.9) compared with 3.0 (Hirsch et al., 2006) and 2.9 (Huang et al., 2008) for the periods 1998–2005 and 2001–2005, re- spectively. Since our estimates are for a later period (2006– 2008), this difference may reﬂect real changes in emissions. It is known that emissions have been increasing in Asia, par- ticularly, in China, over the past decade, which has also in- creased the overall emission in the region 30–90◦N, while no signiﬁcant trends have been found in other regions (Thomp- son et al., 2014a). The increase in China has primarily been driven by an increase in N-fertilizer usage and to a lesser ex- tent by an increase in industrial and combustion sources of N2O (Thompson et al., 2014a). Figure 11. Comparison of the total emissions for each semi- hemisphere region, the ocean, land and globally from this study with previous estimates. The vertical extents of the red boxes in- dicate the range and the horizontal lines in the interior indicate the median of inversion estimates from this study. The points indicate the values from previous studies: Hirsch et al. (2006), open circles; Huang et al. (2008), triangles; AR4, diamonds; Syakila et al. (2011), solid circles; Zaehle et al. (2011), squares. The error bars indicate the 1σ uncertainty. lower N2O emissions at this time. It is possible that the im- pact of this transport error on the optimized ﬂuxes would not be so strong if there were better observational constraints for South America. The same also applies for Africa where the minimum in March–April cannot be explained in terms of variability in soil ﬂuxes as this time corresponds to the wet season when the highest N2O emissions are expected. For the ocean regions, the phase and amplitude of the sea- sonal cycles a posteriori differ little from those a priori. In the Southern Ocean, the minimum in April and maximum in September–October is consistent with the independent esti- mate of Nevison et al. 3.2.3 Seasonal variability (2005) and is largely driven by the up- welling and subsequent venting of subsurface water, which is enriched in N2O. In LMDZ, however, the amplitude of the seasonal cycle is signiﬁcantly larger, especially owing to the lower minimum in April, which is most likely also an artefact of the modelled transport. In the tropical and northern ocean regions, the seasonal cycle is much smaller in amplitude but is also likely driven by seasonal changes in upwelling. 3.2.3 Seasonal variability The mean seasonal cycle for each of the seven land and three ocean regions was calculated by averaging the total monthly emissions over the period 2006 to 2008 and is shown in Fig. 10. For the Northern Hemisphere temperate land regions, Europe, North America and North Asia, the prior ﬂux seasonal cycle predicts a late summer maximum, i.e. between July and August. However, all inversions esti- mate smaller emissions in July and/or August relative to the www.atmos-chem-phys.net/14/6177/2014/ Atmos. Chem. Phys., 14, 6177–6194, 2014 R. L. Thompson et al.: TransCom N2O model inter-comparison – Part 2 6191 Figure 11. Comparison of the total emissions for each semi- hemisphere region, the ocean, land and globally from this study with previous estimates. The vertical extents of the red boxes in- dicate the range and the horizontal lines in the interior indicate the median of inversion estimates from this study. The points indicate the values from previous studies: Hirsch et al. (2006), open circles; Huang et al. (2008), triangles; AR4, diamonds; Syakila et al. (2011), solid circles; Zaehle et al. (2011), squares. The error bars indicate the 1σ uncertainty. pares the global, land and ocean total emissions, as well as the emission distribution by semi-hemisphere, where avail- able. Although the exact period of each study varies, they all include estimates of the global N2O budget in the 2000s. At the global scale, all estimates agree within the range of uncer- tainties (no uncertainty estimate was provided by Syakila and Kroeze, 2011). Progress, however, has been made in reduc- ing the level of uncertainty from 4.5 TgN yr−1in the IPCC AR4 (Intergovernmental Panel on Climate Change Fourth Assessment Report, 2007) to 0.7 TgN yr−1in this study (1σ, 68 % probability assuming Gaussian error distribution) with the complete range of inversions from 16.1 to 18.7 TgN yr−1 for the mean 2006–2008. Previous studies differ in the appor- tionment between land and ocean emissions, with ocean esti- mates varying from 24 % to 38 % of the global total, whereas we found fairly good agreement among the inversions par- ticipating in this study with ocean estimates varying between 31 % and 38 % of the global total. 3.3 Comparison with other estimates In this study we have compared the N2O emission estimates of ﬁve inversion frameworks and analysed these in terms of their spatial distribution and seasonal variability. In gen- eral, there is a high level of agreement among the ﬁve in- versions participating in this study despite the differences in inversion approach, atmospheric transport model and mete- orological data used. This gives us conﬁdence that there has been substantial progress made in terms of uncertainty re- duction. Moreover, we have identiﬁed emission patterns that are robust – that is, common to all inversion frameworks – as well as those that depend strongly on the modelled transport and/or inversion set-up. The salient results are summarized as follows: To put this study into context with previous work, we com- pare our results to independent N2O emission estimates. We have chosen ﬁve studies, including two atmospheric inver- sions (Hirsch et al., 2006; Huang et al., 2008) and three in- ventory and model-based estimates (Denman et al., 2007; Syakila and Kroeze, 2011; Zaehle et al., 2011), which are global in coverage and include estimates of N2O emissions from all sources and are thus appropriate for this comparison. (The study of Zaehle et al. (2011) is not completely indepen- dent as it uses the same terrestrial biosphere model, O-CN, for the estimate of N2O soil emissions as was used in this study’s prior emissions; however, the O-CN simulations used different climate forcing and N-deposition.) Figure 11 com- Atmos. Chem. Phys., 14, 6177–6194, 2014 www.atmos-chem-phys.net/14/6177/2014/ Atmos. Chem. Phys., 14, 6177–6194, 2014 www.atmos-chem-phys.net/14/6177/2014/ 6192 R. L. Thompson et al.: TransCom N2O model inter-comparison – Part 2 In general, the global N2O budget, the total emissions and their spatial distribution, are close to what has been found from previous studies. One notable difference in our inver- sion estimates compared to previous ones though, is the shift in the distribution in the Northern Hemisphere, with lower emissions in the tropics and subtropics and higher emissions in temperate latitudes. Moreover, our inversions show a con- vergence of estimates both at the global and sub-continental scale. This good agreement is most likely due to the expan- sion of the atmospheric observation network. However, con- siderable uncertainties remain, especially in the less well- constrained regions of South Asia and South and Tropical America. These regions also appear to be very sensitive to uncertainties in the modelled atmospheric transport and are regions that should be targeted for new observation sites. Also sensitive to atmospheric transport, is the seasonal ﬂux variability. Although this appears to be robust in the northern extratropics, for the tropics and southern extratropics, this is strongly dependent on having adequate representation of the timing of the maximum in stratosphere to troposphere trans- port (STT) and in vertical mixing, which is still not the case in most CTMs. However, inter-annual variations in ﬂuxes are likely to be more robust as the year-to-year variations in STT are not as great as the seasonal variations. Improve- ments in the accuracy of N2O emission estimates from atmo- spheric inversions, and a move towards emissions monitor- ing in the context of the Kyoto Protocol, would require im- provements to modelling of atmospheric transport, in partic- ular STT, which has a strong inﬂuence on tropospheric N2O mole fractions. www.atmos-chem-phys.net/14/6177/2014/ – the mean global annual N2O emission ranges between 16.1 and 18.7 with a median and median absolute devi- ation (MAD) of 16.7 and 0.7 TgN yr−1, respectively, for the years 2006 to 2008 – the mean global annual N2O emission ranges between 16.1 and 18.7 with a median and median absolute devi- ation (MAD) of 16.7 and 0.7 TgN yr−1, respectively, for the years 2006 to 2008 – ocean emissions were found to contribute between 31 and 38 % and land emissions between 62 % and 69 % to the global total – the apportionment of emissions to each semi- hemisphere was fairly close among inversions, with 7 ± 1 % to 90–30◦S, 28 ± 2 % to 30◦S–0◦, 41 ± 1 % to 0–30◦N and 23 ± 1 % to 30–90◦N (median and MAD as a percentage of the global total), thus making the Northern Hemisphere tropics and subtropics the most important latitudinal range for N2O emissions globally – all inversions estimated lower emissions for the lati- tudes 90–30◦S relative to the prior; however, the me- dian estimate (7 % of the global total) was still higher than that found in previous studies (0 to 4 %) – the ratio of emissions in 0–30◦N to 30–90◦N is smaller in all inversions (range of 1.52 to 1.91 and median of 1.9) compared to previous studies (2.9 and 3.0), repre- senting a change in the percentage of the global total of −16 % for 0–30◦N and of +3 % for 30–90◦N – all inversions estimated higher emissions in Europe and Africa relative to the prior, con- tributing 6 % (1.04 ± 0.20 TgN yr−1) and 20 % (3.36 ± 0.04 TgN yr−1) (median and MAD values) respectively to the global total compared with 5 % (0.80 TgN yr−1) and 18 % (3.07 TgN yr−1) a priori The Supplement related to this article is available online at doi:10.5194/acp-14-6177-2014-supplement. – all inversions (except LMDZ4) estimate lower emissions in North America, contributing 4 % (0.74 ± 0.11 TgN yr−1) (median values) to the global total compared to 6 % (1.00 TgN yr−1) a priori Acknowledgements. We would like to thank S. Zaehle, L. Bopp, and G. van der Werf for providing their N2O emissions estimates. We also thank E. Kort and S. Wofsy for the use of the HIPPO data. Additionally, we would like to acknowledge everyone who contributes to the ongoing measurement of N2O in all networks without which we would not have been able to make this inter- comparison study. – all inversions (except TM5) estimate lower emissions in South and Tropical America, contributing 14 % (2.33 ± 0.27 TgN yr−1) to the global total compared to 15 % (2.55 TgN yr−1) a priori – the largest uncertainties were found in the estimates for South and Tropical America and South Asia owing to uncertainties in the modelled atmospheric transport and to the poor observational constraint for these regions Edited by: W. Lahoz Edited by: W. Lahoz – differences in the meridional distribution of emissions among the inversions were also found to depend on the inter-hemispheric mixing rate of the CTMs – assimilating monthly mean observations from ﬂask sampling networks most likely leads to an underesti- mate of the emissions www.atmos-chem-phys.net/14/6177/2014/ References Tignor, M., and Miller, H. L., Cambridge University Press. Cam- bridge, UK and New York, NY, USA, 2013. Tignor, M., and Miller, H. L., Cambridge University Press. Cam- bridge, UK and New York, NY, USA, 2013. Butterbach-Bahl, K., Baggs, E. M., Dannenmann, M., Kiese, R., and Zechmeister-Boltenstern, S.: Nitrous oxide emis- sions from soils: how well do we understand the pro- cesses and their controls?, Phil. Trans. Roy. Soc., 368, 1621, doi:10.1098/rstb.2013.0122, 2013. Galloway, J. N., Aber, J., Erisman, J. W., Seitzinger, S. P., Howarth, R. W., Cowling, E. B., and Cosby, B. J.: The Nitrogen Cascade, Bioscience, 53, 341–356, 2003. Geels, C., Gloor, M., Ciais, P., Bousquet, P., Peylin, P., Vermeulen, A. T., Dargaville, R., Aalto, T., Brandt, J., Christensen, J. H., Frohn, L. M., Haszpra, L., Karstens, U., Rödenbeck, C., Ra- monet, M., Carboni, G., and Santaguida, R.: Comparing at- mospheric transport models for future regional inversions over Europe – Part 1: mapping the atmospheric CO2 signals, At- mos. Chem. Phys., 7, 3461–3479, doi:10.5194/acp-7-3461-2007, 2007. Butterbach-Bahl, K. and Dannenmann, M.: Denitriﬁcation and as- sociated soil N2O emissions due to agricultural activities in a changing climate, Curr. Op. Environ. Sustain., 3, 389–395, 2011. Chevallier, F., Fisher, M., Peylin, P., Serrar, S., Bousquet, P., Bréon, F.-M., Chédin, A., and Ciais, P.: Inferring CO2 sources and sinks from satellite observations: Method and application to TOVS data, J. Geophys. Res., 110, D24309, doi:10.1029/2005JD006390, 2005. Gloor, M., Dlugokencky, E., Brenninkmeijer, C., Horowitz, L., Hurst, D. F., Dutton, G., Crevoisier, C., Machida, T., and Tans, P.: Three-dimensional SF6 data and tropospheric trans- port simulations: Signals, modeling accuracy, and implica- tions for inverse modeling, J. Geophys. Res., 112, D15112, doi:10.1029/2006jd007973, 2007. Corazza, M., Bergamaschi, P., Vermeulen, A. T., Aalto, T., Haszpra, L., Meinhardt, F., O’Doherty, S., Thompson, R., Moncrieff, J., Popa, E., Steinbacher, M., Jordan, A., Dlugokencky, E., Brühl, C., Krol, M. and Dentener, F.: Inverse modelling of European N2O emissions: assimilating observations from different net- works, Atmos. Chem. Phys., 11, 2381–2398, doi:10.5194/acp- 11-2381-2011, 2011. Hall, B. D., Dutton, G. S., and Elkins, J. W.: The NOAA nitrous oxide standard scale for atmospheric observations. J. Geophys. Res., 112, D09305, doi:10.1029/2006JD007954, 2007. Courtier, P., Thépaut, J. N., and Hollingsworth, A.: A strategy for operational implementation of 4D-Var, using an incremental ap- proach, Q. J. Roy. Meteor. Soc., 120, 1367–1387, 1994. Hirsch, A. I., Michalak, A. M., Bruhwiler, L. M., Peters, W., Dlugo- kencky, E. References J., and Tans, P. P.: Inverse modeling estimates of the global nitrous oxide surface ﬂux from 1998–2001, Global Bio- geochem. Cy., 20, GB1008, doi:10.1029/2004GB002443, 2006. D’Amelio, M. T. S., Gatti, L. V., Miller, J. B., and Tans, P.: Re- gional N2O ﬂuxes in Amazonia derived from aircraft vertical proﬁles, Atmos. Chem. Phys., 9, 8785–8797, doi:10.5194/acp- 9-8785-2009, 2009. Huang, J., Golombek, A., Prinn, R., Weiss, R., Fraser, P., Sim- monds, P., Dlugokencky, E. J., Hall, B., Elkins, J., Steele, P., Langenfelds, R., Krummel, P., Dutton, P. and Porter, L.: Es- timation of regional emissions of nitrous oxide from 1997 to 2005 using multinetwork measurements, a chemical transport model, and an inverse method, J. Geophys. Res., 113, D17313, doi:10.1029/2007JD009381, 2008. Denman, K. L., Brasseur, G. P., Chidthaisong, A., Ciais, P., Cox, P. M., Dickinson, R. E., Hauglustaine, D., Heinze, C., Holland, E., Jacob, D., Lohmann, U., Ramachandran, S., da Silva Dias, P. L., Wofsy, S. C. and Zhang, X.: Couplings Between Changes in the Climate System and Biogeochemistry, Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, edited by: Solomon, S. D., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K. B., Tignor, M. and Miller, H. L., Cambridge University Press, Cambridge, 499–587, 2007. Kort, E. A., Patra, P. K., Ishijima, K., Daube, B. C., JimÈnez, R., Elkins, J., Hurst, D., Moore, F. L., Sweeney, C. and Wofsy, S. C.: Tropospheric distribution and variability of N2O: Evidence for strong tropical emissions, Geophys. Res. Lett., 38, L15806, doi:10.1029/2011GL047612, 2011. Lin, S., Iqbal, J., Hu, R., and Feng, M.: N2O emissions from differ- ent land uses in mid-subtropical China, Agricult. Ecosyst. Envi- ron., 136, 40–48, doi:10.1016/j.agee.2009.11.005, 2010 Duce, R. A., LaRoche, J., Altieri, K., Arrigo, K. R., Baker, A. R., Capone, D. G., Cornell, S., Dentener, F., Galloway, J., Ganeshram, R. S., Geider, R. J., Jickells, T., Kuypers, M. M., Langlois, R., Liss, P. S., Liu, S. M., Middelburg, J. J., Moore, C. M., Nickovic, S., Oschlies, A., Pedersen, T., Prospero, J., Schlitzer, R., Seitzinger, S., Sorensen, L. L., Uematsu, M., Ulloa, O., Voss, M., Ward, B., and Zamora, L.: Impacts of Atmospheric Anthropogenic Nitrogen on the Open Ocean, Science, 320, 893– 897, 2008. Minschwaner, K., Salawitch, R. J., and McElroy, M. R. L. Thompson et al.: TransCom N2O model inter-comparison – Part 2 6193 www.atmos-chem-phys.net/14/6177/2014/ www.atmos-chem-phys.net/14/6177/2014/ Atmos. Chem. Phys., 14, 6177–6194, 2014 R. L. Thompson et al.: TransCom N2O model inter-comparison – Part 2 6194 Patra, P. K., Houweling, S., Krol, M., Bousquet, P., Belikov, D., Bergmann, D., Bian, H., Cameron-Smith, P., Chipperﬁeld, M. P., Corbin, K., Fortems-Cheiney, A., Fraser, A., Gloor, E., Hess, P., Ito, A., Kawa, S. R., Law, R. M., Loh, Z., Maksyutov, S., Meng, L., Palmer, P. I., Prinn, R. G., Rigby, M., Saito, R., and Wilson, C.: TransCom model simulations of CH4 and related species: linking transport, surface ﬂux and chemical loss with CH4 vari- ability in the troposphere and lower stratosphere, Atmos. Chem. Phys., 11, 12813–12837, doi:10.5194/acp-11-12813-2011, 2011. Syakila, A. and Kroeze, C.: The global nitrous oxide budget revis- ited, Greenhouse Gas Measurement and Management, 1, 17–26, 2011. Tarantola, A.: Inverse problem theory and methods for model pa- rameter estimation, Society for Industrial and Applied Mathe- matics, Philadelphia, 358 pp., 2005. Thompson, R. L., Bousquet, P., Chevallier, F., Rayner, P. J. and Ciais, P.: Impact of the atmospheric sink and vertical mixing on nitrous oxide ﬂuxes estimated using inversion methods. J. Geo- phys. Res., 116, D17307, doi:10.1029/2011JD015815, 2011a. Prather, M. J., Holmes, C. D., and Hsu, J.: Reactive greenhouse gas scenarios: Systematic exploration of uncertainties and the role of atmospheric chemistry, Geophys. Res. Lett., 39, L09803, doi:10.1029/2012GL051440, 2012. Thompson, R. L., Gerbig, C., and Rödenbeck, C.: A Bayesian inver- sion estimate of N2O emissions for western and central Europe and the assessment of aggregation errors, Atmos. Chem. Phys., 11, 3443–3458, doi:10.5194/acp-11-3443-2011, 2011b. Prinn, R. G., Cunnold, D., Rasmussen, R., Simmonds, P., Alyea, F., Crawford, A., Fraser, P., and Rosen, R.: Atmospheric emis- sions and trends of nitrous oxide deduced from 10 years of ALE- GAGE data, J. Geophys. Res., 95, 18369–18385, 1990. Thompson, R. L., Chevallier, F., Crotwell, A. M., Dutton, G., Langenfelds, R. L., Prinn, R. G., Weiss, R. F., Tohjima, Y., Nakazawa, T., Krummel, P. B., Steele, L. P., Fraser, P., O’Doherty, S., Ishijima, K., and Aoki, S.: Nitrous oxide emis- sions 1999 to 2009 from a global atmospheric inversion, At- mos. Chem. Phys., 14, 1801–1817, doi:10.5194/acp-14-1801- 2014, 2014a. Prinn, R. G., Weiss, R. F., Fraser, P. J., Simmonds, P. G., Cun- nold, D. M., Alyea, F. N., O’Doherty, S., Salameh, P., Miller, B. R., Huang, J., Wang, R. H. J., Hartley, D. E., Harth, C., Steele, L. P., Sturrock, G., Midgley, P. M., and McCulloch, A.: A his- tory of chemically and radiatively important gases in air de- duced from ALE/GAGE/AGAGE, J. Geophys. R. L. Thompson et al.: TransCom N2O model inter-comparison – Part 2 Res., 105, 17751– 17792, 2000. Thompson, R. L., Patra, P. K., Ishijima, K., Saikawa, E., Corazza, M., Karstens, U., Wilson, C., Bergamaschi, P., Dlugokencky, E., Sweeney, C., Prinn, R. G., Weiss, R. F., O’Doherty, S., Fraser, P. J., Steele, L. P., Krummel, P. B., Saunois, M., Chipperﬁeld, M., and Bousquet, P.: TransCom N2O model inter-comparison – Part 1: Assessing the inﬂuence of transport and surface ﬂuxes on tropospheric N2O variability, Atmos. Chem. Phys., 14, 4349– 4368, doi:10.5194/acp-14-4349-2014, 2014b. Ravishankara, A. R., Daniel, J. S., and Portmann, R. W.: Nitrous Oxide (N2O): The Dominant Ozone-Depleting Substance Emit- ted in the 21st Century, Science, 326, 123–125, 2009. Rayner, P. J., Enting, I. G., Francey, R. J., and Langenfelds, R.: Reconstructing the recent carbon cycle from atmospheric CO2, δ13C and O2/N2 observations, Tellus B, 51, 213–232, 1999. van der Werf, G. R., Randerson, J. T., Giglio, L., Collatz, G. J., Mu, M., Kasibhatla, P. S., Morton, D. C., DeFries, R. S., Jin, Y., and van Leeuwen, T. T.: Global ﬁre emissions and the contribution of deforestation, savanna, forest, agricultural, and peat ﬁres (1997– 2009), Atmos. Chem. Phys., 10, 11707–11735, doi:10.5194/acp- 10-11707-2010, 2010. Rödenbeck, C.: Estimating CO2 sources and sinks from atmo- spheric mixing ratio measurements using a global inversion of atmospheric transport. Technical Report. Jena, Max Planck In- stitute for Biogeochemistry, 6, 37 pp., 2005. Saikawa, E., Prinn, R. G., Dlugokencky, E., Ishijima, K., Dutton, G. S., Hall, B. D., Langenfelds, R., Tohjima, Y., Machida, T., Manizza, M., Rigby, M., O’Doherty, S., Patra, P. K., Harth, C. M., Weiss, R. F., Krummel, P. B., van der Schoot, M., Fraser, P. J., Steele, L. P., Aoki, S., Nakazawa, T., and Elkins, J. W.: Global and regional emissions estimates for N2O, Atmos. Chem. Phys., 14, 4617–4641, doi:10.5194/acp-14-4617-2014, 2014. Volk, C. M., Elkins, J. W., Fahey, D. W., Dutton, G. S., Gilligan, J. M., Loewenstein, M., Podolske, J. R., Chan, K. R., and Gun- son, M. R.: Evaluation of source gas lifetimes from stratospheric observations, J. Geophys. Res., 102, 25543–25564, 1997. Zaehle, S. and Friend, A. D.: Carbon and nitrogen cycle dynamics in the O-CN land surface model: 1. Model description, site-scale evaluation, and sensitivity to parameter estimates. Global Bio- geochem. Cy., 24, GB1005, doi:10.1029/2009Gb003521, 2010. Smith, K. A., Thomson, P. E., Clayton, H., McTaggart, I. References B.: Absorption of Solar Radiation by O2: Implications for O3 and Lifetimes of N2O, CFCl3, and CF2Cl2, J. Geophys. Res., 98, 10543–10561, 1993. Nevison, C. D., Keeling, R. F., Weiss, R. F., Popp, B. N., Jin, X., Fraser, P. J., Porter, L. W., and Hess, P. G.: Southern Ocean ven- tilation inferred from seasonal cycles of atmospheric N2O and O2/N2 at Cape Grim, Tasmania, Tellus B, 57, 218–229, 2005. Dutreuil, S., Bopp, L., and Tagliabue, A.: Impact of enhanced vertical mixing on marine biogeochemistry: lessons for geo- engineering and natural variability, Biogeosciences, 6, 901–912, doi:10.5194/bg-6-901-2009, 2009. Olivier, J. G. J., Van Aardenne, J. A., Dentener, F., Pagliari, V., Ganzeveld, L. N., and Peters, J. A. H. W.: Recent trends in global greenhouse gas emissions: regional trends 1970–2000 and spa- tial distribution of key sources in 2000, Environ. Sci., 2, 81–99, doi:10.1080/15693430500400345, 2005. Myhre, G., Shindell, D., Bréon, F.-M., Collins, W., Fuglesvedt, J., Huang, J., Koch, D., Lamarque, J.-F., Lee, D., Mendoza, B., Nakajima, T., Robock, A., Stephens, G., Takemura, T., and Zhang, H.: Anthropogenic and Natural Radiative Forcing, in: Climate Change 2013: The Physical Science Basis Contribution of Working Group I to the Fifth Assessment Report of the In- tergovernmental Panel on Climate Change, edited by: Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K. B., Park, S., Croteau, P., Boering, K. A., Etheridge, D. M., Ferretti, D., Fraser, P. J., Kim, K. R., Krummel, P. B., Langenfelds, R. L., van Ommen, T. D., Steele, L. P., and Trudinger, C. M.: Trends and seasonal cycles in the isotopic composition of nitrous oxide since 1940, Nature Geosci., 5, 261–265, 2012. Atmos. Chem. Phys., 14, 6177–6194, 2014 www.atmos-chem-phys.net/14/6177/2014/ R. L. Thompson et al.: TransCom N2O model inter-comparison – Part 2 P., and Co- nen, F.: Effects of temperature, water content and nitrogen fertil- isation on emissions of nitrous oxide by soils, Atmos. Environ., 32, 3301–3309, 1998. Zaehle, S., Ciais, P., Friend, A. D., and Prieur, V.: Carbon beneﬁts of anthropogenic reactive nitrogen offset by nitrous oxide emis- sions, Nature Geosci, 4, 601–605, 2011. Stohl, A., Eckhardt, S., Forster, C., James, P., and Spichtinger, N.: On the pathways and timescales of intercontinental air pollution transport, J. Geophys. Res.-Atmos., 107, 4684, doi:10.1029/2001jd001396, 2002. Atmos. Chem. Phys., 14, 6177–6194, 2014 www.atmos-chem-phys.net/14/6177/2014/
https://openalex.org/W2991335887	https://www.frontiersin.org/articles/10.3389/fnins.2019.01236/pdf	English	null	Association Between BDNF Val66Met Polymorphism and Optic Neuritis Damage in Neuromyelitis Optica Spectrum Disorder	Frontiers in neuroscience	2,019	cc-by	5,344	Keywords: BDNF, Val66Met, optic nerve damage, NMOSD, optic neuritis Association Between BDNF Val66Met Polymorphism and Optic Neuritis Damage in Neuromyelitis Optica Spectrum Disorder Ting Shen1,2, Vivek Gupta1, Con Yiannikas3, Alexander Klistorner1,2, Stuart L. Graham1,2 and Yuyi You1,2 1 Faculty of Medicine and Health Sciences, Macquarie University, Sydney, NSW, Australia, 2 Save Sight Institute, The University of Sydney, Sydney, NSW, Australia, 3 Australia Department of Neurology, Royal North Shore Hospital, Sydney, NSW, Australia Neuromyelitis optica spectrum disorder (NMOSD) is an autoimmune inﬂammatory disease of the central nervous system (CNS). The purpose of the study was to examine the association between the brain-derived neurotrophic factor (BDNF) Val66Met genotype and structural and functional optic nerve damage in the eyes of NMOSD patients. A total of 17 NMOSD subjects (34 eyes) were included in the study and were divided into subgroups based on optic neuritis (ON) history and BDNF Val66Met polymorphisms. The mean (range) age was 47.8 (23–78) years, and the mean (SD) disease duration was 7.4 (2–39) years. All participants had undergone optical coherence tomography (OCT) scans for global retinal nerve ﬁber layer (gRNFL) and ganglion cell- inner plexiform layer (GCIPL) thickness and multifocal visual evoked potential (mfVEP) test for amplitude and latency. BDNF Val66Met polymorphisms were genotyped in all participants. OCT and mfVEP changes were compared between two genotype groups (Met carriers vs. Val homozygotes) by using the generalised estimating equation (GEE) models. The BDNF Val66Met polymorphism was signiﬁcantly associated with more severe nerve ﬁber layer damage and axonal loss in ON eyes of NMOSD subjects. Met carriers had more signiﬁcantly reduced GCIPL (P = 0.002) and gRNFL (P < 0.001) thickness as well as more delayed mfVEP latency (P = 0.008) in ON eyes. No association was found between Val66Met variants and non-ON (NON)-eye of the participants. These ﬁndings suggest that the BDNF Val66Met polymorphism may be associated with optic nerve damage caused by acute ON attacks in NMOSD patients. BRIEF RESEARCH REPORT published: 19 November 2019 doi: 10.3389/fnins.2019.01236 Correspondence: Ting Shen Correspondence: Ting Shen tingshen313@outlook.com Specialty section: This article was submitted to Neurodegeneration, a section of the journal Frontiers in Neuroscience Received: 25 June 2019 Accepted: 01 November 2019 Published: 19 November 2019 Edited by: Christine Nguyen, The University of Melbourne, Australia Reviewed by: Aline Silva Miranda, Federal University of Minas Gerais, Brazil Friedemann Paul, Charité University Medicine Berlin, Germany Correspondence: Ting Shen tingshen313@outlook.com INTRODUCTION Shen T, Gupta V, Yiannikas C, Klistorner A, Graham SL and You Y (2019) Association Between BDNF Val66Met Polymorphism and Optic Neuritis Damage in Neuromyelitis Optica Spectrum Disorder. Front. Neurosci. 13:1236. doi: 10.3389/fnins.2019.01236 Neuromyelitis Optica (NMO) is an idiopathic, autoimmune and inﬂammatory disease of the central nervous system (CNS) that occurs in individuals of all ethnicities. NMO has long been considered as a variant of multiple sclerosis (MS) until the detection of the serum antibodies to the astrocytic water channel aquaporin-4 (AQP4) (Lennon et al., 2004). AQP4 antibodies are detectable in 60– 90% of the patients with NMO, but not in the serum of MS patients (Jarius and Wildemann, 2010). November 2019 \| Volume 13 \| Article 1236 Frontiers in Neuroscience \| www.frontiersin.org 1 BDNF Val66Met Polymorphism in NMOSD Shen et al. diagnostic criteria (Wingerchuk et al., 2015). Exclusion criteria are: patients tested as MOG-positive, retinal, optic nerve, or other neurologic diseases aﬀecting the visual system, including any other types of optic neuropathy, optic atrophy with no known cause. Patients tested within 12 months of the acute ON were also excluded from the study to eliminate the decline of amplitude and prolongation of latency in mfVEP in a post- acute stage of ON (Klistorner et al., 2010). Two participants with visual acuities worse than hand movement were not able to complete mfVEP testing and were excluded from mfVEP analysis. Details on the recruitment, composition and selective attrition are described in detail elsewhere, blood samples from all but two participants with MOG antibodies in the previous study were obtained (Klistorner et al., 2010). The ethnic background of the participants is mainly Caucasian, except for one Asian subject. All participants underwent OCT scans for RNFL and ganglion cell-inner plexiform layer (GCIPL) thickness, multifocal VEP testing for amplitude and latency, and MRI scans for whole-brain (WB) lesion volume (LV) as previously reported. This ﬁnding together with the foveal thinning reported in NMO patients without optic neuritis (ON) history (Oertel et al., 2017; Yamamura and Nakashima, 2017), supports the concept of NMO to be a predominantly serum AQP4 IgG antibody- mediated astrocytopathy [NMO spectrum disorder (NMOSD)] that is distinct from MS (Wingerchuk et al., 2015). Recently, Müller glial dysfunction has been reported in the eyes of seropositive AQP4-IgG NMOSD patients support a subclinical retinal astrocytopathy in the disease (You et al., 2019). INTRODUCTION It has been reported that subtype of reactive astrocytes are neurotoxic, which may explain the severe axonal loss observed in NMOSD (Liddelow et al., 2017). MOG-antibody-positive NMOSD has been recently recognized as a possible new inﬂammatory disorder of the CNS (Jurynczyk et al., 2017), patients with positive MOG- IgG were thus excluded in this study. A higher frequency of familial NMO cases was observed in a previous study, suggested complex genetic susceptibility in NMOSD (Matiello et al., 2010). Though several genes and genetic variants have been evaluated as contributors to NMOSD, the major genes that confer signiﬁcant susceptibility are still unknown (Kim et al., 2010; Wang et al., 2011; Wang et al., 2012; Yoshimura et al., 2013; Estrada et al., 2018). Brain-derived neurotrophic factor (BDNF) has unique roles in neuronal growth, diﬀerentiation, distribution and survival (Poo, 2001; Gupta et al., 2014). The genetic variation that is the focus of this study has a single nucleotide substitution from G to A (Val66Met, NCBI database dbSNP rs6265) in the pro-region of the BDNF gene. The BDNF Val66Met polymorphism has been associated with various neurodegenerative and psychiatric diseases (Notaras et al., 2015; Shen et al., 2018b). In a recent study, carriage of Met alleles was associated with lower rates of optic nerve damage and was reported to be reducing the long-term open- angle glaucoma (OAG) progression in female patients (Shen et al., 2019). However, it remains to be elucidated whether Val66Met is associated with neuroaxonal damage in NMOSD. Optical coherence tomography (OCT) has been used to examine rates of axonal damage through retinal nerve ﬁbre layer (RNFL) and macular thickness in NMO and MS which could help distinguish the patterns of optic nerve damage in the two diseases (Naismith et al., 2009; Ratchford et al., 2009). Multifocal visual evoked potentials (mfVEPs) provide functional assessments of axonal loss (amplitude) and demyelination (latency) in the visual pathway (Klistorner et al., 2010). NMOSD patients were suggested to have a more severe axonal loss in ON eyes by more signiﬁcantly reduced mfVEP amplitude compared to MS (Shen et al., 2018a). The aim of the present study was to clarify whether the carriage of Met allele is associated with axonal loss and demyelination in ON and non-ON (NON) eyes of NMOSD participants by using both structural (OCT) and functional (mfVEP) analysis. Statistical Analysis Statistical analysis was performed using SPSS software version 22 (SPSS, Inc., Chicago, IL, United States). Data from both ON eyes and NON eyes were included for the comparison between two genotype groups using the generalised estimating equation model (GEE). The gender was included as a cofactor, and the age, disease duration and number of ON episodes of all patients were included as covariates in the GEE analysis. The correlation between OCT and mfVEP parameters were also examined using the GEE method with adjustment of intrasubject factors. A P-value of less than 0.05 was considered statistically INTRODUCTION The study was conducted in accordance with the Declaration of Helsinki and approved by the Human Research Ethics Committee of the University of Sydney (Sydney, NSW, Australia). Written informed consent was obtained from all study participants. Genotyping The genomic DNA was isolated from peripheral blood with a commercially available DNA extraction kit (Qiagen, Hilden, Germany). Quantiﬁcation of isolated DNA was carried out with a spectrophotometer (Thermo Scientiﬁc, Rockford, IL, United States). The G →A nucleotide substitution, identifying the of BDNF Val66Met polymorphism, was assayed by polymerase chain reaction (PCR, Eppendorf, Hamburg, Germany). The primers used in this study were as follows: forward 5′ ACTCTGGAGAGCGTGAATGG 3′ and reverse 5′ TCCAGGGTGATGCTCAGTAGT 3′. The ampliﬁcation conditions for PCR were initiated at 95◦C for 5 min, followed by 30 cycles comprising of denaturation at 94◦C for 1 min, annealing at 55◦C for 30 s and extension at 72◦C for 1 min, with a ﬁnal extension step of 5 min at 72◦C. The carriage of BDNF Val66Met polymorphism was determined by direct sequencing (both directions, Australian Genome Research Facility, Sydney, NSW, Australia) of genomic DNA. A detailed description of the genotyping procedure is provided elsewhere (Shen et al., 2018b). Frontiers in Neuroscience \| www.frontiersin.org DISCUSSION The genotype frequencies of Val66Met among the NMOSD cohort were 9 (52.9%) for genotype GG (Val/Val), 7 (41.2%) for GA (Val/Met), and 1 (5.9%) for AA (Met/Met). The allele frequency of the A allele (Met) was 26.5% in the study subjects. This study suggests that BDNF Val66Met polymorphism may be associated with more severe ON attacks in NMOSD patients. In cross-sectional analysis, the Met carriers were found to have signiﬁcantly more reduced GCIPL and gRNFL thickness, more diminished mfVEP amplitude (borderline signiﬁcance) and signiﬁcantly more delayed mfVEP latency than Val homozygotes in ON eyes. The genotypic eﬀect of Val66Met on NMOSD remains positive even with adjustment for a range of cofactors and covariates, including sex, age, disease duration, and ON episodes. In contrast, no association was found between BDNF polymorphisms and changes in NON-eyes of patients. Correlation analysis suggested positive associations between structural and functional measurements in the NMOSD cohort. To our knowledge, this is the ﬁrst study to suggest that the presence of Met alleles may convey higher rates of axonal damage in ON in NMOSD. The frequency of the BDNF Val66Met polymorphism varies among diﬀerent ethnicity. In our cohort, the Met allele frequency is 26.5%, which is close to the frequencies previous reported in the United States (28.0%) (Parsian et al., 2004) and Italian cohorts (29.7%) (Ventriglia et al., 2002). OCT is an objective, simple and non-invasive method that allows y y j The OCT and mfVEP measures were compared between two genotype groups and are summarized in Table 2. In ON eyes, there were signiﬁcant diﬀerences between the BDNF Val66Met polymorphism genotypes in averaged GCIPL thickness (P = 0.002) and global retinal nerve ﬁber layer (gRNFL) thickness (P < 0.001). In parallel with the OCT changes, carriage of Met allele was also associated with signiﬁcantly more prolonged latency (P = 0.008) in ON eyes. The reduction of mfVEP amplitude showed the borderline diﬀerence between Met and Val genotypes (P = 0.07) in ON eyes. In NON-eyes, no signiﬁcant diﬀerence was observed in any of the parameters mentioned above (Table 2). As there are possibilities for the AQP4 seronegative patients to be aﬀected by another disease (Jarius et al., 2012), a sub-analysis was performed to validate the diﬀerences in ON eyes and excluded the potential inﬂuence of AQP4 serology. METHODS Neuromyelitis optica spectrum disorder (N = 17) patients were recruited from four tertiary neuro-ophthalmology or neurology clinics in Sydney. The NMOSD diagnosis was made per the 2015 November 2019 \| Volume 13 \| Article 1236 Frontiers in Neuroscience \| www.frontiersin.org 2 BDNF Val66Met Polymorphism in NMOSD Shen et al. signiﬁcant, and a P-value between 0.05 and 0.1 was considered borderline signiﬁcant. and P < 0.001, respectively). There was also a borderline signiﬁcant diﬀerence between Val/Val and Met carriers in mfVEP latency (P = 0.1). The correlation between GCIPL and gRNFL (P < 0.001) was showed in Figure 1A as data visualization. In addition, correlation analysis demonstrated strong positive correlations in NMOSD participants with increased GCIPL (P < 0.001) and gRNFL (P = 0.001) thickness associated with higher mfVEP amplitude (Figure 1B). There was a signiﬁcant correlation between GCIPL and mfVEP latency (P = 0.03). A borderline signiﬁcant correlation was also observed between gRNFL and mfVEP latency (P = 0.06). RESULTS Demographic and clinical features of the participants included in the study are presented in Table 1. Twenty eyes of 17 patients with NMOSD had ON history and was included in the following GEE analysis as ON eyes. Fourteen eyes of the patients had no previous ON attacks and were included as NON-eyes in the statistical analysis. The mean age of the patients was 47.8 years (range, 23–78). No signiﬁcant diﬀerence was found in sex ratio, AQP4 status, age, disease duration, BCVA or T2 WB LV between two genotype groups (Table 1). AQP4-IgG = astrocyte water channel aquaporin-4 immunoglobulin G; BCVA = best-corrected visual acuity; WB = whole brain; LV = lesion volume. Statistics: achi-square test; bgeneralized estimating equation; cMann–Whitney U test. Neuromyelitis optica spectrum disorder treatment: immunosuppressants (n = 13, including azathioprine, mycophenolate, rituximab, and prednisolone), plasma exchange (n = 3), and not receiving treatment (n = 3). No parameter showed signiﬁcant differences between two genotype groups. DISCUSSION In AQP4 seropositive patients (16 eyes), signiﬁcant diﬀerences were observed between genotype groups in GCIPL, gRNFL and mfVEP latency (P < 0.001, P = 0.04, TABLE 1 \| Demographic and clinical characteristics of participants. Total (N = 17) Val/Val (N = 9) Met carriers (N = 8) P-values Eyes, N (%) 34 18 (52.9) 16 (47.1) n.a. ON eyes, N (%)a 20 (58.8) 10 (55.6) 10 (62.5) 0.68 NON-eyes, N (%)a 14 (41.2) 8 (44.4) 6 (37.5) ON history, N (%) 13 (76.5) 6 (66.7) 7 (87.5) 0.31 Unilateral ON history, N (%)a 6 (35.3) 2 (22.2) 4 (50.0) 0.39 Bilateral ON history, N (%)a 7 (41.2) 4 (44.4) 3 (37.5) Female, N (%)a 11 (64.7) 6 (66.7) 5 (62.5) 0.86 AQP4-IgG seropositive, N (%)a 13 (76.5) 6 (66.7) 7 (87.5) 0.31 Age, years, mean (range)b 47.8 (23–78) 42.8 (23–64) 53.4 (38–78) 0.14 Disease duration, years, mean (SD)b 7.4 (10.1) 10.9 (13.2) 3.5 (1.9) 0.1 BCVA, LogMAR, mean (SD)b 0.4 (0.8) 0.3 (0.8) 0.4 (0.9) 0.63 T2 WB LV, cm3, mean (SD)c 860.2 (1858.0) 926.0 (1737.7) 786.3 (2104.2) 0.91 AQP4-IgG = astrocyte water channel aquaporin-4 immunoglobulin G; BCVA = best-corrected visual acuity; WB = whole brain; LV = lesion volume. Statistics: achi-square test; bgeneralized estimating equation; cMann–Whitney U test. Neuromyelitis optica spectrum disorder treatment: immunosuppressants (n = 13, including azathioprine, mycophenolate, rituximab, and prednisolone), plasma exchange (n = 3), and not receiving treatment (n = 3). No parameter showed signiﬁcant differences between two genotype groups. AQP4-IgG = astrocyte water channel aquaporin-4 immunoglobulin G; BCVA = best-corrected visual acuity; WB = whole brain; LV = lesion volume. Statistics: achi-square test; bgeneralized estimating equation; cMann–Whitney U test. Neuromyelitis optica spectrum disorder treatment: immunosuppressants (n = 13, including azathioprine, mycophenolate, rituximab, and prednisolone), plasma exchange (n = 3), and not receiving treatment (n = 3). No parameter showed signiﬁcant differences between two genotype groups. November 2019 \| Volume 13 \| Article 1236 Frontiers in Neuroscience \| www.frontiersin.org 3 BDNF Val66Met Polymorphism in NMOSD Shen et al. TABLE 2 \| Association between BDNF Val66Met polymorphism and optic nerve damage in NMOSD. DISCUSSION Parameters Val/Val Met carriers Difference P-values Value 95% CI ON eyes OCT GCIPL 62.19 59.43 2.76 0.97 to 4.55 0.002∗ gRNFL 82.23 58.84 23.39 20.54 to 26.25 <0.001∗ mfVEP Amp 137.64 68.54 69.1 −4.67 to 142.86 0.07 Lat 152.13 163.39 −11.26 −19.61 to−2.91 0.008∗ NON-eyes OCT GCIPL 78.81 75.44 3.37 −9.13 to 15.88 0.6 gRNFL 98.98 95.2 3.78 −22.72 to 30.28 0.78 mfVEP Amp 171.45 184.71 −13.26 −46.57 to 20.06 0.44 Lat 148.91 149.87 −0.96 −7.32 to 5.40 0.77 ON = optic neuritis; OCT = optic coherence tomography; GCIPL = ganglion cell-inner plexiform layer; gRNFL = global retinal nerve ﬁber layer; mfVEP = multifocal visual evoked potential; NON = non-optic neuritis, n.s. = not signiﬁcant. GEE models were applied assuming additive effect for the minor allele variant (Met carriers), adjusted for sex, age, disease duration, and ON episodes. Statistical signiﬁcance was deﬁned as P < 0.05 an asterisk (∗). TABLE 2 \| Association between BDNF Val66Met polymorphism and optic nerve damage in NMOSD. ON = optic neuritis; OCT = optic coherence tomography; GCIPL = ganglion cell-inner plexiform layer; gRNFL = global retinal nerve ﬁber layer; mfVEP = multifocal visual evoked potential; NON = non-optic neuritis, n.s. = not signiﬁcant. GEE models were applied assuming additive effect for the minor allele variant (Met carriers), adjusted for sex, age, disease duration, and ON episodes. Statistical signiﬁcance was deﬁned as P < 0.05 an asterisk (∗). FIGURE 1 \| Data visualization showing correlation analysis of (A) ganglion cell-inner plexiform layer (GCIPL) and global retinal nerve ﬁber layer (gRNFL) and (B) gRNFL and multifocal visual evoked potential (mfVEP) in neuromyelitis optica spectrum disorders (NMOSD). anglion cell-inner plexiform layer (GCIPL) and global retinal nerve ﬁber layer (gRNFL) and (B) FIGURE 1 \| Data visualization showing correlation analysis of (A) ganglion cell-inner plexiform layer (GCIPL) and global retinal nerve ﬁber layer (gRNFL) and (B) gRNFL and multifocal visual evoked potential (mfVEP) in neuromyelitis optica spectrum disorders (NMOSD). nerve axons (Weibel et al., 1995; Fawcett et al., 1998). As the carriage of BDNF Val66Met polymorphism has been associated with reduced activity-dependent secretion of BDNF (Egan et al., 2003), the more severe axonal loss in ON eyes may be explained by less availability of BDNF within the retina and along the higher visual pathway. in vivo measurement of the thickness of retinal nerve ﬁbers. Frontiers in Neuroscience \| www.frontiersin.org DISCUSSION It has been reported that the actual RNFL thickness measured from histologic sections would be 4–12% larger than the corresponding OCT measurements of nerve ﬁbers (Toth et al., 1997). Previous studies demonstrated that ON in NMO is associated with thinner RNFL, GCIPL, and mfVEP amplitude compared to MS (Naismith et al., 2009; Shen et al., 2018a). The mfVEP amplitude and latency were suggested to be highly correlated with RNFL thickness (Parisi et al., 1999; Trip et al., 2005; Klistorner et al., 2008). The results of the present study demonstrated a potential negative eﬀect BDNF Val66Met polymorphism has on ON damage in NMOSD, which contrasts with the previous glaucoma study where the Met alleles exhibited protective eﬀects on OAG progression in a gender-speciﬁc way (Shen et al., 2019). The opposite eﬀect may suggest diﬀerent mechanisms in the axonal loss in glaucoma and NMOSD. The precursor form of BDNF (pro-BDNF) was suggested to increase cell apoptosis, while in contrast, mature BDNF (mBDNF) enhances cell survival (Lu et al., 2005). In neurons, pro-BDNF is generally converted to mBDNF by proteases such as tissue plasminogen activator (tPA), and cerebral tPA expression was reported to decrease with age in an animal model (Cacquevel et al., 2007). Thus, higher levels of BDNF secretion in Val/Val may not increase neuronal function unless there are enough cleavage molecules to convert BDNF from the precursor form to mature form. The mean age was approximately 70 for the participants in the glaucoma The age distribution was suggested to inﬂuence the AQP4 autoimmunity where females aged 65 years above were more likely to be seropositive than other age groups, and the detection rate of AQP4-IgG rise exponentially in females after 50 years old (Quek et al., 2012). A recent study has also established that the gray matter atrophy and RNFL thinning were correlated with longer disease duration in NMOSD (von Glehn et al., 2014). There were no signiﬁcant diﬀerences between two genotype groups in gender, age or disease duration in this study, and though all these parameters have been included and adjusted for as either cofactors or covariate, the results remained signiﬁcant in ON eye analysis. It has been reported that BDNF is transported anterogradely and retrogradely, and BDNF secretion has a selective eﬀect on the survival and maintenance of lesioned optic November 2019 \| Volume 13 \| Article 1236 4 BDNF Val66Met Polymorphism in NMOSD Shen et al. DISCUSSION study, and only 47 in our NMOSD cohort, the signiﬁcant age diﬀerence between the cohorts may contribute to diﬀerentiated tPA levels and opposite eﬀects of Val66Met on RGC survival. In addition, it is well established that in glaucoma, elevated intraocular pressure is the main cause of RGC apoptosis and optic nerve head degeneration via several mechanisms (Kuehn et al., 2005), while in NMOSD, inﬂammation of the optic nerve and spinal cord were usually dominating the disease and causing axonal damage. It has been reported that BDNF has protective eﬀects against ischemic insult by modulating local inﬂammation in the brain in an animal model (Jiang et al., 2011). The Met allele may interact diﬀerently with acute optic nerve damage in inﬂammation and progressive glaucomatous optic neuropathy; however, future studies are warranted to investigate further the cellular mechanisms of the contrast eﬀects BDNF Val66Met have in two diseases. in NMOSD patients. The diﬀerence was not found in NON- eyes of participants indicating the eﬀect of BDNF Val66Met on RGC axonal loss in NMOSD is speciﬁcally related to acute inﬂammatory ON attacks. ETHICS STATEMENT The studies involving human participants were reviewed and approved by Human Research Ethics Committee of the University of Sydney (Sydney, NSW, Australia). The patients/participants provided their written informed consent to participate in this study. There are several potential limitations to the current study. First, the sample size of NMOSD is relatively small due to the rarity of the disease in the Caucasian population. However, the genotype distribution was favorable with both eyes included in the analysis, which signiﬁcantly increased the sample size and met the minimum requirement for desired statistical power. The P-values were especially small for both gRNFL and GCIPL thickness (Table 1), which were likely representing true diﬀerences between the two genotype groups. Second, the investigation was in the absence of a healthy control group. This study is also limited by the cross-sectional study design, future longitudinal study with larger sample size is necessary to validate the reproducibility of the results, especially a cohort with only AQP4 seropositive subjects to exclude the possibility of seronegative patients inﬂuenced by another disease. In addition, future research could also focus on the BDNF Val66Met polymorphism’s eﬀect on ON in other neuroinﬂammatory diseases such as MS. FUNDING This study was supported by the National Health and Medical Research Council Australia, The Ophthalmic Research Institute of Australia (ORIA), National Multiple Sclerosis Society (NMSS), and Sydney Medical School Foundation. DATA AVAILABILITY STATEMENT The raw data supporting the conclusion of this manuscript will be made available from the corresponding author, without undue reservation, to any qualiﬁed researcher. AUTHOR CONTRIBUTIONS TS: acquisition, analysis and interpretation of data, and drafting and revising the work. VG: revising the work and acquisition of data for the work. CY: acquisition of data and design of the work. AK and YY: design of the work, revising the work, and provided approval for publication of the content. SG: revising the work and provided approval for publication of the content. CONCLUSION Our study demonstrated signiﬁcant associations between carriage of BDNF Met allele and more severe axonal damage in ON eyes REFERENCES Gupta, V., You, Y., Li, J., Gupta, V., Golzan, M., Klistorner, A., et al. (2014). BDNF impairment is associated with age-related changes in the inner retina and exacerbates experimental glaucoma. Biochim. Biophys. Acta 1842, 1567–1578. doi: 10.1016/j.bbadis.2014.05.026 Cacquevel, M., Launay, S., Castel, H., Benchenane, K., Cheenne, S., Buee, L., et al. (2007). Ageing and amyloid-beta peptide deposition contribute to an impaired brain tissue plasminogen activator activity by diﬀerent mechanisms. Neurobiol. Dis. 27, 164–173. doi: 10.1016/j.nbd.2007.04.004 Jarius, S., Ruprecht, K., Wildemann, B., Kuempfel, T., Ringelstein, M., Geis, C., et al. (2012). Contrasting disease patterns in seropositive and seronegative neuromyelitis optica: a multicentre study of 175 patients. J Neuroinﬂammation 9:14. doi: 10.1186/1742-2094-9-14 Egan, M. F., Kojima, M., Callicott, J. H., Goldberg, T. E., Kolachana, B. S., Bertolino, A., et al. (2003). The BDNF val66met polymorphism aﬀects activity-dependent secretion of BDNF and human memory and hippocampal function. Cell 112, 257–269. doi: 10.1016/s0092-8674(03)00035-7 Jarius, S., and Wildemann, B. (2010). AQP4 antibodies in neuromyelitis optica: diagnostic and pathogenetic relevance. Nat. Rev. Neurol. 6, 383–392. doi: 10. 1038/nrneurol.2010.72 Estrada, K., Whelan, C. W., Zhao, F., Bronson, P., Handsaker, R. E., Sun, C., et al. (2018). A whole-genome sequence study identiﬁes genetic risk factors for neuromyelitis optica. Nat. Commun. 9:1929. doi: 10.1038/s41467-018-04 332-3 Jiang, Y., Wei, N., Lu, T., Zhu, J., Xu, G., and Liu, X. (2011). Intranasal brain-derived neurotrophic factor protects brain from ischemic insult via modulating local inﬂammation in rats. Neuroscience 172, 398–405. doi: 10.1016/ j.neuroscience.2010.10.054 Fawcett, J. P., Bamji, S. X., Causing, C. G., Aloyz, R., Ase, A. R., Reader, T. A., et al. (1998). Functional evidence that BDNF is an anterograde neuronal trophic factor in the CNS. J. Neurosci. 18, 2808–2821. doi: 10.1523/jneurosci.18-08- 02808.1998 Jurynczyk, M., Messina, S., Woodhall, M. R., Raza, N., Everett, R., Roca-Fernandez, A., et al. (2017). Clinical presentation and prognosis in MOG-antibody disease: a UK study. Brain 140, 3128–3138. doi: 10.1093/brain/awx276 November 2019 \| Volume 13 \| Article 1236 5 Frontiers in Neuroscience \| www.frontiersin.org BDNF Val66Met Polymorphism in NMOSD Shen et al. sclerosis and neuromyelitis optica spectrum disorder. Ophthalmology 126, 445–453. doi: 10.1016/j.ophtha.2018.06.022 Kim, H. J., Park, H. Y., Kim, E., Lee, K. S., Kim, K. K., Choi, B. O., et al. (2010). Common CYP7A1 promoter polymorphism associated with risk of neuromyelitis optica. Neurobiol. Dis. 37, 349–355. doi: 10.1016/j.nbd.2009. 10.013 Shen, T., You, Y. Y., Joseph, C., Mirzaei, M., Klistorner, A., Graham, S. L., et al. (2018b). REFERENCES BDNF polymorphism: a review of its diagnostic and clinical relevance in neurodegenerative disorders. Aging Dis. 9, 523–536. doi: 10.14336/AD.2017. 0717 Klistorner, A., Arvind, H., Garrick, R., Graham, S. L., Paine, M., and Yiannikas, C. (2010). Interrelationship of optical coherence tomography and multifocal visual-evoked potentials after optic neuritis. Invest. Ophthalmol. Vis. Sci. 51, 2770–2777. doi: 10.1167/iovs.09-4577 Toth, C. A., Narayan, D. G., Boppart, S. A., Hee, M. R., Fujimoto, J. G., Birngruber, R., et al. (1997). A comparison of retinal morphology viewed by optical coherence tomography and by light microscopy. Arch. Ophthalmol. 115, 1425– 1428. Klistorner, A., Arvind, H., Nguyen, T., Garrick, R., Paine, M., Graham, S., et al. (2008). Axonal loss and myelin in early ON loss in postacute optic neuritis. Ann. Neurol. 64, 325–331. doi: 10.1002/ana.21474 Trip, S. A., Schlottmann, P. G., Jones, S. J., Altmann, D. R., Garway-Heath, D. F., Thompson, A. J., et al. (2005). Retinal nerve ﬁber layer axonal loss and visual dysfunction in optic neuritis. Ann. Neurol. 58, 383–391. doi: 10.1002/ana.20575 Kuehn, M. H., Fingert, J. H., and Kwon, Y. H. (2005). Retinal ganglion cell death in glaucoma: mechanisms and neuroprotective strategies. Ophthalmol. Clin. North Am. 18, 383–395. doi: 10.1016/j.ohc.2005.04.002 Ventriglia, M., Bocchio Chiavetto, L., Benussi, L., Binetti, G., Zanetti, O., Riva, M. A., et al. (2002). Association between the BDNF 196 A/G polymorphism and sporadic Alzheimer’s disease. Mol. Psychiatry 7, 136–137. doi: 10.1038/sj. mp.4000952 Lennon, V. A., Wingerchuk, D. M., Kryzer, T. J., Pittock, S. J., Lucchinetti, C. F., Fujihara, K., et al. (2004). A serum autoantibody marker of neuromyelitis optica: distinction from multiple sclerosis. Lancet 364, 2106–2112. doi: 10. 1016/s0140-6736(04)17551-x von Glehn, F., Jarius, S., Lira, R. P. C., Ferreira, M. C. A., von Glehn, F. H. R., Castro, S. M. C. E., et al. (2014). Structural brain abnormalities are related to retinal nerve ﬁber layer thinning and disease duration in neuromyelitis optica spectrum disorders. Mult. Scler. J. 20, 1189–1197. doi: 10.1177/ 1352458513519838 Liddelow, S. A., Guttenplan, K. A., Clarke, L. E., Bennett, F. C., Bohlen, C. J., Schirmer, L., et al. (2017). Neurotoxic reactive astrocytes are induced by activated microglia. Nature 541, 481–487. doi: 10.1038/nature21029 Lu, B., Pang, P. T., and Woo, N. H. (2005). The yin and yang of neurotrophin action. Nat. Rev. Neurosci. 6, 603–614. doi: 10.1038/nrn1726 Wang, H. H., Dai, Y. Q., Qiu, W., Zhong, X. N., Wu, A. M., Wang, Y. G., et al. (2011). REFERENCES HLA-DPB1∗0501 is associated with susceptibility to anti- aquaporin-4 antibodies positive neuromyelitis optica in southern han chinese. J. Neuroimmunol. 233, 181–184. doi: 10.1016/j.jneuroim.2010.11.004 Matiello, M., Kim, H. J., Kim, W., Brum, D. G., Barreira, A. A., Kingsbury, D. J., et al. (2010). Familial neuromyelitis optica. Neurology 75, 310–315. doi: 10.1212/ WNL.0b013e3181ea9f15 Naismith, R. T., Tutlam, N. T., Xu, J., Klawiter, E. C., Shepherd, J., Trinkaus, K., et al. (2009). Optical coherence tomography diﬀers in neuromyelitis optica compared with multiple sclerosis. Neurology 72, 1077–1082. doi: 10.1212/01. wnl.0000345042.53843.d5 Wang, H. H., Zhong, X. N., Wang, K., Qiu, W., Li, J., Dai, Y. Q., et al. (2012). Interleukin 17 gene polymorphism is associated with anti-aquaporin 4 antibody-positive neuromyelitis optica in the southern han chinese - a case control study. J. Neurol. Sci. 314, 26–28. doi: 10.1016/j.jns.2011.11.005 Notaras, M., Hill, R., and van den Buuse, M. (2015). The BDNF gene Val66Met polymorphism as a modiﬁer of psychiatric disorder susceptibility: progress and controversy. Mol. Psychiatry 20, 916–930. doi: 10.1038/mp.2015.27 Weibel, D., Kreutzberg, G. W., and Schwab, M. E. (1995). Brain-derived neurotrophic factor (BDNF) prevents lesion-induced axonal die-back in young rat optic nerve. Brain Res. 679, 249–254. doi: 10.1016/0006-8993(95)00238-l Wingerchuk, D. M., Banwell, B., Bennett, J. L., Cabre, P., Carroll, W., Chitnis, T., et al. (2015). International consensus diagnostic criteria for neuromyelitis optica spectrum disorders. Neurology 85, 177–189. Oertel, F. C., Kuchling, J., Zimmermann, H., Chien, C., Schmidt, F., Knier, B., et al. (2017). Microstructural visual system changes in AQP4-antibody-seropositive NMOSD. Neurol. Neuroimmunol. Neuroinﬂamm. 4:e334. doi: 10.1212/NXI. 0000000000000334 Yamamura, T., and Nakashima, I. (2017). Foveal thinning in neuromyelitis optica: a sign of retinal astrocytopathy? Neurol. Neuroimmunol. Neuroinﬂamm. 4:e347. doi: 10.1212/nxi.0000000000000347 Parisi, V., Manni, G., Spadaro, M., Colacino, G., Restuccia, R., Marchi, S., et al. (1999). Correlation between morphological and functional retinal impairment in multiple sclerosis patients. Invest. Ophth. Vis. Sci. 40, 2520–2527. Yoshimura, S., Isobe, N., Matsushita, T., Yonekawa, T., Masaki, K., Sato, S., et al. (2013). Distinct genetic and infectious proﬁles in japanese neuromyelitis optica patients according to anti-aquaporin 4 antibody status. J. Neurol. Neurosurg. Psychiatry 84, 29–34. doi: 10.1136/jnnp-2012-302925 Parsian, A., Sinha, R., Racette, B., Zhao, J. H., and Perlmutter, J. S. (2004). Association of a variation in the promoter region of the brain-derived neurotrophic factor gene with familial Parkinson’s disease. Parkinsonism Relat. Disord. 10, 213–219. doi: 10.1016/j.parkreldis.2003.12.003 You, Y., Zhu, L., Zhang, T., Shen, T., Fontes, A., Yiannikas, C., et al. (2019). Frontiers in Neuroscience \| www.frontiersin.org REFERENCES Evidence of muller glial dysfunction in patients with aquaporin- 4 immunoglobulin g-positive neuromyelitis optica spectrum disorder. Ophthalmology 126, 801–810. doi: 10.1016/j.ophtha.2019.01.016 Poo, M. M. (2001). Neurotrophins as synaptic modulators. Nat. Rev. Neurosci. 2, 24–32. doi: 10.1038/35049004 Quek, A. M. L., McKeon, A., Lennon, V. A., Mandrekar, J. N., Iorio, R., Jiao, Y. J., Quek, A. M. L., McKeon, A., Lennon, V. A., Mandrekar, J. N., Iorio, R., Jiao, Y. J., et al. (2012). Eﬀects of age and sex on aquaporin-4 autoimmunity. Arch. Neurol. 69, 1039–1043. et al. (2012). Eﬀects of age and sex on aquaporin-4 autoimmunity. Arch. Neurol. 69, 1039–1043. Conﬂict of Interest: The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. Ratchford, J. N., Quigg, M. E., Conger, A., Frohman, T., Frohman, E., Balcer, L. J., et al. (2009). Optical coherence tomography helps diﬀerentiate neuromyelitis optica and MS optic neuropathies. Neurology 73, 302–308. doi: 10.1212/WNL. 0b013e3181af78b8 Copyright © 2019 Shen, Gupta, Yiannikas, Klistorner, Graham and You. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Shen, T., Gupta, V. K., Klistorner, A., Chitranshi, N., Graham, S. L., and You, Y. (2019). Sex-speciﬁc eﬀect of BDNF val66met genotypes on the progression of open-angle glaucoma. Invest. Ophthalmol. Vis. Sci. 60, 1069–1075. doi: 10.1167/ iovs.18-26364 Shen, T., You, Y., Arunachalam, S., Fontes, A., Liu, S., Gupta, V., et al. (2018a). Diﬀering structural and functional patterns of optic nerve damage in multiple November 2019 \| Volume 13 \| Article 1236 Frontiers in Neuroscience \| www.frontiersin.org 6
https://openalex.org/W2113306639	https://iris.unica.it/bitstream/11584/231374/2/pgen.adiponectinlevel_2012.pdf	English	null	Novel Loci for Adiponectin Levels and Their Influence on Type 2 Diabetes and Metabolic Traits: A Multi-Ethnic Meta-Analysis of 45,891 Individuals	PLOS genetics	2,012	cc-by	25,337	Novel Loci for Adiponectin Levels and Their Influence on Type 2 Diabetes and Metabolic Traits: A Multi-Ethnic Meta-Analysis of 45,891 Individuals Winkler10, Bruce M. Psaty37,38, Russell P. Tracy39, Jennifer Brody40, Ida Chen41, Jorma Viikari42, Mika Ka¨ho¨ nen43, Peter P. Pramstaller44,45,46, David M. Evans4, Beate St. Pourcain47, Naveed Sattar48, Andrew R. Wood6, Stefania Bandinelli49, Olga D. Carlson50, Josephine M. Egan50, Stefan Bo¨ hringer24, Diana van Heemst51, Lyudmyla Kedenko52, Kati Kristiansson53, Marja-Liisa Nuotio53, Britt-Marie Loo54, Tamara Harris55, Melissa Garcia55, Alka Kanaya56, Margot Haun20, Norman Klopp57, H.-Erich Wichmann57,58,59, Panos Deloukas16, Efi Katsareli60, David J. Couper61, Bruce B. Duncan62,63, Margreet Kloppenburg64, Linda S. Adair65, Judith B. Borja66, DIAGRAM+ Consortium{, MAGIC Consortium{, GLGC Investigators{, MuTHER Consortium, James G. Wilson67, Solomon Musani68, Xiuqing Guo69, Toby Johnson34,70,71, Robert Semple72, Tanya M. Teslovich13, Matthew A. Allison73, Susan Redline74, Sarah G. Buxbaum75, Karen L. Mohlke25, Ingrid Meulenbelt19, Christie M. Ballantyne76, George V. Dedoussis60, Frank B. Hu22, Yongmei Liu21, Bernhard Paulweber52, Timothy D. Spector15, P. Eline Slagboom77, Luigi Ferrucci14, Antti Jula54, Markus Perola53, Olli Raitakari78, Jose C. Florez30,79,80,81, Veikko Salomaa82", Johan G. Eriksson83,84,85,86,87", Timothy M. Frayling6", Andrew A. Hicks44", Terho Lehtima¨ki12", George Davey Smith4", David S. Siscovick88", Florian Kronenberg20", Cornelia van Duijn17,18", Ruth J. F. Loos8", Dawn M. Waterworth7", James B. Meigs3,30", Josee Dupuis31,89", J. Brent Richards15,90"* Florian Kronenberg20", Cornelia van Duijn17,18", Ruth J. F. Loos8", Dawn M. Waterworth7", James B. Meigs3,30", Josee Dupuis31,89", J. Novel Loci for Adiponectin Levels and Their Influence on Type 2 Diabetes and Metabolic Traits: A Multi-Ethnic Meta-Analysis of 45,891 Individuals Marie-France Hivert2,3., Nicholas Timpson4., John R. B. Perry5,6., Xin Yuan7., Zari Dastani1., Marie-France Hivert2,3., Nicholas Timpson4., John R. B. Perry5,6., Xin Yuan7., Robert A. Scott8., Peter Henneman9., Iris M. Heid10., Jorge R. Kizer11., Leo-Pekka Lyytika¨inen12., Christian Fuchsberger13., Toshiko Tanaka14, Andrew P. Morris5, Kerrin Small15,16, Aaron Isaacs17,18, Marian Beekman19, Stefan Coassin20, Kurt Lohman21, Lu Qi22, Stavroula Kanoni16, James S. Pankow23, Hae-Won Uh24, Ying Wu25, Aurelian Bidulescu26, Laura J. Rasmussen-Torvik27, Celia M. T. Greenwood28, Martin Ladouceur29, Jonna Grimsby3,30, Alisa K. Manning31, Ching-Ti Liu31, Jaspal Kooner32, Vincent E. Mooser7, Peter Vollenweider33, Karen A. Kapur34, John Chambers35, Nicholas J. Wareham8, Claudia Langenberg8, Rune Frants9, Ko Willems-vanDijk9, Ben A. Oostra18,36, Sara M. Willems17, Claudia Lamina20, Thomas W. Winkler10, Bruce M. Psaty37,38, Russell P. Tracy39, Jennifer Brody40, Ida Chen41, Jorma Viikari42, Mika Ka¨ho¨ nen43, Peter P. Pramstaller44,45,46, David M. Evans4, Beate St. Pourcain47, Naveed Sattar48, Andrew R. Wood6, Stefania Bandinelli49, Olga D. Carlson50, Josephine M. Egan50, Stefan Bo¨ hringer24, Diana van Heemst51, Lyudmyla Kedenko52, Kati Kristiansson53, Marja-Liisa Nuotio53, Britt-Marie Loo54, Tamara Harris55, Melissa Garcia55, Alka Kanaya56, Margot Haun20, Norman Klopp57, H.-Erich Wichmann57,58,59, Panos Deloukas16, Efi Katsareli60, David J. Couper61, Bruce B. Duncan62,63, Margreet Kloppenburg64, Linda S. Adair65, Judith B. Borja66, DIAGRAM+ Consortium{, MAGIC Consortium{, GLGC Investigators{, MuTHER Consortium, James G. Wilson67, Solomon Musani68, Xiuqing Guo69, Toby Johnson34,70,71, Robert Semple72, Tanya M. Teslovich13, Matthew A. Allison73, Susan Redline74, Sarah G. Buxbaum75, Karen L. Mohlke25, Ingrid Meulenbelt19, Christie M. Ballantyne76, George V. Dedoussis60, Frank B. Hu22, Yongmei Liu21, Bernhard Paulweber52, Timothy D. Spector15, P. Eline Slagboom77, Luigi Ferrucci14, Antti Jula54, Markus Perola53, Olli Raitakari78, Jose C. Florez30,79,80,81, Veikko Salomaa82", Johan G. Eriksson83,84,85,86,87", Timothy M. Frayling6", Andrew A. Hicks44", Terho Lehtima¨ki12", George Davey Smith4", David S. Siscovick88", Florian Kronenberg20", Cornelia van Duijn17,18", Ruth J. F. Loos8", Dawn M. Waterworth7", James B. Meigs3,30", Josee Dupuis31,89", J. Brent Richards15,90"* Zari Dastani1., Marie-France Hivert2,3., Nicholas Timpson4., John R. B. Perry5,6., Xin Yuan7., Robert A. Scott8., Peter Henneman9., Iris M. Heid10., Jorge R. Kizer11., Leo-Pekka Lyytika¨inen12., Christian Fuchsberger13., Toshiko Tanaka14, Andrew P. Morris5, Kerrin Small15,16, Aaron Isaacs17,18, Marian Beekman19, Stefan Coassin20, Kurt Lohman21, Lu Qi22, Stavroula Kanoni16, James S. Pankow23, Hae-Won Uh24, Ying Wu25, Aurelian Bidulescu26, Laura J. Rasmussen-Torvik27, Celia M. T. Greenwood28, Martin Ladouceur29, Jonna Grimsby3,30, Alisa K. Manning31, Ching-Ti Liu31, Jaspal Kooner32, Vincent E. Mooser7, Peter Vollenweider33, Karen A. Kapur34, John Chambers35, Nicholas J. Wareham8, Claudia Langenberg8, Rune Frants9, Ko Willems-vanDijk9, Ben A. Oostra18,36, Sara M. Willems17, Claudia Lamina20, Thomas W. PLoS Genetics \| www.plosgenetics.org Adiponectin: Multiethnic Genome-Wide Meta-Analysis Adiponectin: Multiethnic Genome-Wide Meta-Analysis Adiponectin: Multiethnic Genome-Wide Meta-Analysis Washington, Seattle, Washington, United States of America, 38 Group Health Research Institute, Group Health Cooperative, Seattle, Washington, United States of America, 39 Departments of Pathology and Biochemistry, University of Vermont, Burlington, Vermont, United States of America, 40 Cardiovascular Health Research Unit, University of Washington, Seattle, Washington, United States of America, 41 Medical Genetics Research Institute, Cedars Sinai Medical Center, Los Angeles, California, United States of America, 42 Department of Medicine, University of Turku and Turku University Hospital, Turku, Finland, 43 Department of Clinical Physiology, University of Tampere and Tampere University Hospital, Tampere, Finland, 44 Center for Biomedicine, European Academy Bozen/Bolzano (EURAC) (Affiliated Institute of the University of Lu¨beck, Lu¨beck, Germany), Bolzano, Italy, 45 Department of Neurology, General Central Hospital, Bolzano, Italy, 46 Department of Neurology, University of Lu¨beck, Lu¨beck, Germany, 47 School of Social and community medicine, University of Bristol, Bristol, United Kingdom, 48 British Heart Foundation Glasgow Cardiovascular Research Centre, University of Glasgow, Glasgow, United Kingdom, 49 Geriatric Unit, Azienda Sanitaria Firenze (ASF), Florence, Italy, 50 Laboratory of Clinical Investigation, National Institute of Aging, Baltimore, Maryland, United States of America, 51 Gerontology and Geriatrics, Leiden University Medical Center, Leiden, The Netherlands, 52 First Department of Internal Medicine, St. Adiponectin: Multiethnic Genome-Wide Meta-Analysis Johann Spital, Paracelsus Private Medical University Salzburg, Salzburg, Austria, 53 Public Health Genomics Unit, Department of Chronic Disease Prevention, National Institute for Health and Welfare, and Institute for Molecular Medicine Finland FIMM, University of Helsinki, Helsinki, Finland, 54 Population Studies Unit, Department of Chronic Disease Prevention, National Institute for Health and Welfare, Turku, Finland, 55 Intramural Research Program, Laboratory of Epidemiology, Demography, and Biometry, National Institute on Aging, National Institutes of Health, Bethesda, Maryland, United States of America, 56 Division of General Internal Medicine, Women’s Health Clinical Research Center, University of California San Francisco, San Francisco, California, United States of America, 57 Institute of Epidemiology, Helmholtz Zentrum Mu¨nchen, German Research Center for Environmental Health, Munich, Germany, 58 Institute of Medical Informatics, Biometry and Epidemiology, Ludwig- Maximilians-Universita¨t, Munich, Germany, 59 Klinikum Großhadern, Munich, Germany, 60 Harokopio University, Athens, Greece, 61 Collaborative Studies Coordinating Center, Department of Biostatistics, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States of America, 62 School of Medicine, Federal University of Rio Grande do Sul, Porto Alegre, Brazil, 63 Department of Epidemiology, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States of America, 64 Department of Rheumatology and Department of Clinical Epidemiology, Leiden, The Netherlands, 65 Department of Nutrition, University of North Carolina, Chapel Hill, North Carolina, United States of America, 66 Office of Population Studies Foundation, University of San Carlos, Cebu City, Philippines, 67 Department of Physiology and Biophysics, University of Mississippi Medical Center, Jackson, Mississippi, United States of America, 68 Department of Medicine, University of Mississippi Medical Center, Jackson, Mississippi, United States of America, 69 Medical Genetics Institute, Cedars-Sinai Medical Center, Los Angeles, California, United States of America, 70 University Institute of Social and Preventative Medicine, Centre Hospitalier Universitaire Vaudois (CHUV) and University of Lausanne, Lausanne, Switzerland, 71 Swiss Institute of Bioinformatics, Lausanne, Switzerland, 72 Metabolic Research Laboratories, Institute of Metabolic Science, University of Cambridge, Addenbrooke’s Hospital, Cambridge, United Kingdom, 73 Department of Family and Preventive Medicine, University of California San Diego, La Jolla, California, United States of America, 74 Brigham and Women’s Hospital, Boston, Massachusetts, United States of America, 75 Jackson Heart Study Coordinating Center, Jackson State University, Jackson, Mississippi, United States of America, 76 Baylor College of Medicine and Methodist DeBakey Heart and Vascular Center, Houston, Texas, United States of America, 77 Molecular Epidemiology, Leiden University Medical Center, Leiden, The Netherlands, 78 Research Centre of Applied and Preventive Cardiovascular Medicine, University of Turku and the Department of Clinical Physiology, Turku University Hospital, Turku, Finland, 79 Program in Medical and Population Genetics, Broad Institute, Cambridge, Massachusetts, United States of America, 80 Center for Human Genetic Research, Massachusetts General Hospital, Boston, Massachusetts, United States of America, 81 Diabetes Research Center, Diabetes Unit, Massachusetts General Hospital, Boston, Massachusetts, United States of America, 82 Chronic Disease Epidemiology and Prevention Unit, Department of Chronic Disease Prevention, National Institute for Health and Welfare, Helsinki, Finland, 83 Diabetes Prevention Unit, Department of Chronic Disease Prevention, National Institute for Health and Welfare, Helsinki, Finland, 84 Unit of General Practice, Helsinki University Central Hospital, Helsinki, Finland, 85 Folkhalsan Research Centre, Helsinki, Finland, 86 Vaasa Central Hospital, Vaasa, Finland, 87 Department of General Practice and Primary Health Care, University of Helsinki, Helsinki, Finland, 88 University of Washington, Seattle, Washington, United States of America, 89 National Heart, Lung, and Blood Institute’s Framingham Heart Study, Framingham, Massachusetts, United States of America, 90 Departments of Medicine, Human Genetics, Epidemiology, and Biostatistics, Lady Davis Institute, Jewish General Hospital, McGill University, Montreal, Canada Copyright: 2012 Dastani et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Novel Loci for Adiponectin Levels and Their Influence on Type 2 Diabetes and Metabolic Traits: A Multi-Ethnic Meta-Analysis of 45,891 Individuals Brent Richards15,90"* 1 Department of Epidemiology, Biostatistics, and Occupational Health, Jewish General Hospital, Lady Davis Institute, McGill University, Montreal, Canada, 2 Department of Medicine, Universite´ de Sherbrooke, Sherbrooke, Canada, 3 General Medicine Division, Massachusetts General Hospital, Boston, Massachusetts, United States of America, 4 MRC CAiTE Centre and School of Social and Community Medicine, University of Bristol, Bristol, United Kingdom, 5 Wellcome Trust Centre for Human Genetics, University of Oxford, Oxford, United Kingdom, 6 Genetics of Complex Traits, Peninsula Medical School, University of Exeter, Exeter, United Kingdom, 7 Genetics, GlaxoSmithKline, King of Prussia, Pennsylvania, United States of America, 8 MRC Epidemiology Unit, Institute of Metabolic Science, Addenbrooke’s Hospital, Cambridge, United Kingdom, 9 Department of Human Genetics, Leiden University Medical Center, Leiden, The Netherlands, 10 Department of Epidemiology and Preventive Medicine, Regensburg University Medical Center, Regensburg, Germany, 11 Departments of Medicine and Public Health, Weill Cornell Medical College, New York, New York, United States of America, 12 Department of Clinical Chemistry, University of Tampere and Tampere University Hospital, Tampere, Finland, 13 Center for Statistical Genetics, Department of Biostatistics, University of Michigan, Ann Arbor, Michigan, United States of America, 14 Clinical Research Branch, National Institute on Aging, Baltimore, Maryland, United States of America, 15 Department of Twin Research and Genetic Epidemiology, King’s College London, London, United Kingdom, 16 Wellcome Trust Sanger Institute, Wellcome Trust Genome Campus, Hinxton, United Kingdom, 17 Genetic Epidemiology Unit, Department of Epidemiology, Erasmus Medical Center, Rotterdam, The Netherlands, 18 Centre for Medical Systems Biology, Leiden, The Netherlands, 19 Section of Molecular Epidemiology, Leiden University Medical Center and The Netherlands Genomics Initiative, The Netherlands Consortium for Healthy Aging, Leiden, The Netherlands, 20 Division of Genetic Epidemiology, Innsbruck Medical University, Innsbruck, Austria, 21 Wake Forest University School of Medicine, Winston-Salem, North Carolina, United States of America, 22 Harvard School of Public Health, Boston, Massachusetts, United States of America, 23 Division of Epidemiology and Community Health, University of Minnesota, Minneapolis, Minnesota, United States of America, 24 Medical Statistics and Bioinformatics, Leiden University Medical Center, Leiden, The Netherlands, 25 Department of Genetics, University of North Carolina, Chapel Hill, North Carolina, United States of America, 26 Cardiovascular Research Institute, Morehouse School of Medicine, Atlanta, Georgia, United States of America, 27 Department of Preventive Medicine, Feinberg School of Medicine, Northwestern University, Chicago, Illinois, United States of America, 28 Lady Davis Institute for Medical Research, Department of Oncology, McGill University, Montreal, Canada, 29 Department of Human Genetics McGill University, Montreal, Canada, 30 Department of Medicine, Harvard Medical School, Boston, Massachusetts, United States of America, 31 Department of Biostatistics, Boston University School of Public Health, Boston, Massachusetts, United States of America, 32 Cardiology, Ealing Hospital National Health Service (NHS) Trust, London, United Kingdom, 33 Department of Internal Medicine, University of Lausanne, Lausanne, Switzerland, 34 Department of Medical Genetics, University of Lausanne, Lausanne, Switzerland, 35 Epidemiology and Biostatistics, Imperial College London, London, United Kingdom, 36 Deptartment of Clinical Genetics and Department of Epidemiology, Erasmus Medical Center, Rotterdam, The Netherlands, 37 Cardiovascular Health Research Unit, Departments of Medicine and Epidemiology, University of PLoS Genetics \| www.plosgenetics.org March 2012 \| Volume 8 \| Issue 3 \| e1002607 1 Citation: Dastani Z, Hivert M-F, Timpson N, Perry JRB, Yuan X, et al. (2012) Novel Loci for Adiponectin Levels and Their Influence on Type 2 Diabetes and Metabolic Traits: A Multi-Ethnic Meta-Analysis of 45,891 Individuals. PLoS Genet 8(3): e1002607. doi:10.1371/journal.pgen.1002607 Abstract Salzburg Atherosclerosis Prevention Program in subjects at High Individual Risk (SHAPIR): Part of this work was funded by the ‘‘Genomics of Lipid-associated Disorders’’ (GOLD) of the ‘‘Austrian Genome Research Programme’’ (GEN-AU) to Florian Kronenberg, and grants from the ‘‘Medizinische Forschungsgesellschaft Salzburg’’ and the ‘‘Kamillo Eisner Stiftung’’ (Switzerland) to Bernhard Paulweber. THISEAS: The genotyping of the THISEAS study was funded by the Wellcome Trust. The recruitment was partially supported by the General Secretary of Research and Technology (PENED 03,EPSILON.,DELTA.474). We are grateful to all the volunteers for their time and help, the medical staff of the hospitals and the field investigators, Eirini Theodoraki, Maria Dimitriou and Kathy Stirrups for her assistance in the genotyping. Cebu Longitudinal Health and Nutrition Survey (CLHNS): We thank the Office of Population Studies Foundation research and data collection teams and the study participants who generously provided their time for this study. This work was supported by National Institutes of Health grants DK078150, TW05596, HL085144, RR20649, ES10126, and DK56350. Coordinating Centre: McGill University. This work was supported by grants from the Canadian Foundation for Innovation, the Canadian Institutes of Health Research (CIHR), Fonds de la recherche en sante´ du Que´bec, the Lady Davis Institute, the Ministe`re du De´veloppement e´conomique, de l’Innovation et de l’Exportation du Que´bec and the Jewish General Hospital. JB Richards and Z Dastani are supported by the CIHR. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: DM Waterworth, X Yuan, and VE Mooser are full-time employees of GlaxoSmithKline. P Vollenweider received grant money from GlaxoSmithKline to fund the CoLaus study The other authors declare no competing financial interests Competing Interests: DM Waterworth, X Yuan, and VE Mooser are full-time employees of GlaxoSmithKline. P Vollenweider received grant money from GlaxoSmithKline to fund the CoLaus study. The other authors declare no competing financial interests. * E-mail: brent.richards@mcgill.ca . These authors contributed equally to this work. " These authors also contributed equally to this work. { Memberships of these consortia are provided in the Acknowledgments. clinical parameters including blood glucose, indices of insulin resistance, proatherogenic dyslipidemia, and risk of type 2 diabetes (T2D), stroke and coronary artery disease [1,2,3,4]. Collectively these conditions account for most of the burgeoning pandemic of Abstract Circulating levels of adiponectin, a hormone produced predominantly by adipocytes, are highly heritable and are inversely associated with type 2 diabetes mellitus (T2D) and other metabolic traits. We conducted a meta-analysis of genome-wide association studies in 39,883 individuals of European ancestry to identify genes associated with metabolic disease. We identified 8 novel loci associated with adiponectin levels and confirmed 2 previously reported loci (P = 4.561028– 1.2610243). Using a novel method to combine data across ethnicities (N = 4,232 African Americans, N = 1,776 Asians, and N = 29,347 Europeans), we identified two additional novel loci. Expression analyses of 436 human adipocyte samples revealed that mRNA levels of 18 genes at candidate regions were associated with adiponectin concentrations after accounting for multiple testing (p,361024). We next developed a multi-SNP genotypic risk score to test the association of adiponectin decreasing risk alleles on metabolic traits and diseases using consortia-level meta-analytic data. This risk score was associated with increased risk of T2D (p = 4.361023, n = 22,044), increased triglycerides (p = 2.6610214, n = 93,440), increased waist-to-hip ratio (p = 1.861025, n = 77,167), increased glucose two hours post oral glucose tolerance testing (p = 4.461023, n = 15,234), increased fasting insulin (p = 0.015, n = 48,238), but with lower in HDL- cholesterol concentrations (p = 4.5610213, n = 96,748) and decreased BMI (p = 1.461024, n = 121,335). These findings identify novel genetic determinants of adiponectin levels, which, taken together, influence risk of T2D and markers of insulin resistance. Citation: Dastani Z, Hivert M-F, Timpson N, Perry JRB, Yuan X, et al. (2012) Novel Loci for Adiponectin Levels and Their Influence on Type 2 Diabetes and Metabolic Traits: A Multi-Ethnic Meta-Analysis of 45,891 Individuals. PLoS Genet 8(3): e1002607. doi:10.1371/journal.pgen.1002607 Editor: Peter M. Visscher, The University of Queensland, Australia Received September 30, 2011; Accepted February 3, 2012; Pub PLoS Genetics \| www.plosgenetics.org March 2012 \| Volume 8 \| Issue 3 \| e1002607 PLoS Genetics \| www.plosgenetics.org 2 Adiponectin: Multiethnic Genome-Wide Meta-Analysis Adiponectin: Multiethnic Genome-Wide Meta-Analysis Adiponectin: Multiethnic Genome-Wide Meta-Analysis Funding: Baltimore Longitudinal Study of Aging (BLSA): The BLSA was supported in part by the Intramural Research Program of the NIH, National Institute on Aging. A portion of that support was through an R&D contract with MedStar Research Institute. Erasmus Rucphen Family (ERF). Abstract The ERF study was supported by grants from The Netherlands Organisation for Scientific Research, Erasmus MC and the Centre for Medical Systems Biology (CMSB), and the European Network for Genetic and Genomic Epidemiology (ENGAGE) consortium. Invecchaire in Chianti (InCHIANTI). JRB Perry is a Sir Henry Wellcome Postdoctoral Research Fellow (092447/Z/10/Z). Framingham Heart Study (FHS): This research was conducted in part using data and resources from the Framingham Heart Study of the National Heart Lung and Blood Institute of the National Institutes of Health and Boston University School of Medicine. The analyses reflect intellectual input and resource development from the Framingham Heart Study investigators participating in the SNP Health Association Resource (SHARe) project. This work was partially supported by the National Heart, Lung, and Blood Institute’s Framingham Heart Study (Contract No. N01-HC-25195) and its contract with Affymetrix, Inc for genotyping services (Contract No. N02-HL- 6-4278). A portion of this research utilized the Linux Cluster for Genetic Analysis (LinGA-II) funded by the Robert Dawson Evans Endowment of the Department of Medicine at Boston University School of Medicine and Boston Medical Center. The Collaborative Health Research in the Region of Augsburg (KORA F3): This study was partially funded by the ‘‘Tiroler Wissenschaftsfonds’’ (Project UNI-0407/29) and by the ‘‘Genomics of Lipid-associated Disorders – GOLD’’ of the ‘‘Austrian Genome Research Programme GEN-AU’’ to F Kronenberg. The MONICA/KORA Augsburg cohort study was financed by the Helmholtz Zentrum Mu¨nchen. It was further funded by the NIH subcontract from the Children’s Hospital, Boston, US, (H-E Wichmann and IM Heid, prime grant 1 R01 DK075787-01A1 to JN Hirschhorn) and the German National Genome Research Net NGFN2 and NGFNplus (H-E Wichmann 01GS0823). TwinsUK: Study was funded by the Wellcome Trust, European Commission Framework (FP7/2007–2013), ENGAGE project HEALTH-F4-2007-201413, and the FP5 GenomEUtwin Project (QLG2-CT-2002-01254). It also receives support from the Arthritis Research Campaign, Chronic Disease Research Foundation, the National Institute for Health Research (NIHR) comprehensive Biomedical Research Centre award to Guy’s and St Thomas’ NHS Foundation Trust in partnership with King’s College London, and a Biotechnology and Biological Sciences Research Council project grant (G20234). Cardiovascular Health Study (CHS): The CHS research reported in this article was supported by contract numbers N01-HC-85079 through N01- HC-85086, N01-HC-35129, N01 HC-15103, N01 HC-55222, N01-HC-75150, N01-HC-45133, grant numbers U01 HL080295 and R01 HL087652, HL105756, and HL094555 from the National Heart, Lung, and Blood Institute, with additional contribution from the National Institute of Neurological Disorders and Stroke. Abstract A full list of principal CHS investigators and institutions can be found at http://www.chs-nhlbi.org/pi.htm. DNA handling and genotyping was supported in part by National Center for Research Resources grant M01-RR00425 to the Cedars-Sinai General Clinical Research Center Genotyping core and National Institute of Diabetes and Digestive and Kidney Diseases grant DK063491 to the Southern California Diabetes Endocrinology Research Center, NHLBI R01-HL085251. Helsinki Birth Cohort Study (HBCS): HBCS has been supported by grants from Academy of Finland (project numbers 114382, 126775, 127437, 129255, 129306, 130326, 209072, 210595, 213225, 216374), Finnish Diabetes Research Society, Finnish Foundation for Pediatric Research, Samfundet Folkha¨lsan, Juho Vainio Foundation, Novo Nordisk Foundation, Finska La¨karesa¨llskapet, Pa¨ivikki and Sakari Sohlberg Foundation, Signe and Ane Gyllenberg Foundation, and Yrjo¨ Jahnsson Foundation. DILGOM survey was funded by the Finnish Academy, grant number 118065. Cardiovascular Risk in Young Finns (YFS): The Young Finns Study has been financially supported by the Academy of Finland: grants 126925, 121584, 124282, 129378 (Salve), 117787 (Gendi), and 41071 (Skidi), the Social Insurance Institution of Finland, Kuopio, Tampere and Turku University Hospital Medical Funds, Juho Vainio Foundation, Paavo Nurmi Foundation, Finnish Foundation of Cardiovascular Research (T.L., OT.R), Tampere Tuberculosis Foundation (Te.Le., Mik, Ka¨), the Emil Aaltonen Foundation (T.L.) and Finnish Cultural Foundation. The expert technical assistance in the statistical analyses by Irina Lisinen and Ville Aalto are gratefully acknowledged. Dietary, Lifestyle, and Genetic determinants of Obesity and Metabolic syndrome (DILGOM): K Kristiansson was supported by the Orion-Farmos Research Foundation and the Academy of Finland (grant no. 125973). M Perola and V Salomaa were supported by the Finnish Foundation for Cardiovascular Research, the Sigrid Juse´lius Foundation, and the Academy of Finland (grants 129322, 129494 and 139635). JG Eriksson was supported by the Academy of Finland (grants 126775, 129255, 129907, and 135072). Fenland study: The Fenland Study is funded by the Wellcome Trust and the Medical Research Council, as well as by the Support for Science Funding programme and CamStrad. We are grateful to all the volunteers for their time and help, and to the General Practitioners and practice staff for help with recruitment. We thank the Fenland Study co-ordination team and the Field Epidemiology team of the MRC Epidemiology Unit for recruitment and clinical testing. Multiethnic Study of Atherosclerosis (MESA): The MESA project is conducted and supported by the National Heart, Lung, and Blood Institute (NHLBI) in collaboration with MESA investigators. Abstract Support is provided by grants and contracts N01 HC-95159, N01-HC-95160, N01-HC- 95161, N01-HC-95162, N01-HC-95163, N01-HC-95164, N01-HC-95165, N01-HC-95166, N01-HC-95167, N01-HC-95168, N01-HC-95169, and RR-024156. Funding for CARe genotyping was provided by NHLBI Contract N01-HC-65226. Jackson Heart Study (JHS): The Jackson Heart Study is supported by the National Heart, Lung, and Blood Institute, through contracts with Jackson State University (N01-HC-95170), the University of Mississippi Medical Center (N01-HC-95171), and Tougaloo College (N01- HC-95172). Adiponectin measurements used in the current study were funded by PHS Award UL1 RR025008 from the Clinical and Translational Science Award program, National Institutes of Health, National Center for Research Resources (NCRR). Health ABC: This research was supported by NIA contracts N01AG62101, N01AG62103, and N01AG62106. The genome-wide association study was funded by NIA grant 1R01AG032098-01A1 to Wake Forest University Health Sciences and genotyping services were provided by the Center for Inherited Disease Research (CIDR). CIDR is fully funded through a federal contract from the National Institutes of Health to The Johns Hopkins University, contract number HHSN268200782096C. This research was supported in part by the Intramural Research Program of the NIH, National Institute on Aging. Genetics, Arthrosis , and Progression) study (GARP): This study was supported the Leiden University Medical Centre and the Dutch Arthritis Association. Pfizer, Groton, CT, USA supported the inclusion of the GARP study. The genotypic work was supported by the Netherlands Organization of Scientific Research (MW 904-61-095, 911-03-016, 917 66344 and 911-03-012), Leiden University Medical Centre and the Centre of Medical System Biology and Netherlands Consortium for Healthy Aging both in the framework of the Netherlands Genomics Initiative (NGI). The adiponectin measurements were supported by TI-Pharma. Atherosclerosis Risk in Communities (ARIC): The Atherosclerosis Risk in Communities Study is carried out as a collaborative study supported by National Heart, Lung, and Blood Institute contracts N01-HC-55015, N01-HC-55016, N01-HC-55018, N01-HC-55019, N01-HC-55020, N01-HC-55021, N01-HC-55022, R01HL087641, R01HL59367, R01HL086694, and RC2 HL102419; National Human Genome Research Institute contract U01HG004402; National Institutes of Health contract HHSN268200625226C; and National Institute of Diabetes and Digestive and Kidney Diseases R01DK056918. Infrastructure was partly supported by Grant Number UL1RR025005,a component of the National Institutes of Health and NIH Roadmap for Medical Research. Competing Interests: DM Waterworth, X Yuan, and VE Mooser are full-time employees of GlaxoSmithKline. P Vollenweider received grant money from GlaxoSmithKline to fund the CoLaus study. The other authors declare no competing financial interests. Results of Meta-Analysis of GWAS The meta-analysis was divided into four phases 1) Discovery phase, which involved cohorts providing GWAS results, 2) In- silico replication phase which included additional GWAS cohorts joining our meta-analysis after the completion of the discovery phase, 3) De-novo genotyping in cohorts without GWAS genotyping and 4) Multi-Ethnic meta-analysis applying a novel method for complex trait mapping using different ethnicities. Discovery phase in individuals of white European origin. The meta-analysis of sex-combined data from 16 GWAS (n = 29,347) of individuals of white European descent identified ten loci associated with adiponectin levels at p#5.061028 (Table 1 and Figure 1A and Figure S1, Table S2). These results include the previously described associations with adiponectin at ADIPOQ (rs6810075[T]; ß = 0.06, p-value = 3.60610241), KNG1 (rs2062632[T]; ß = 0.05, p-value = 2.52610219) on 3q27.3, and CDH13 (rs12922394[T; ß = 20.1, p = 3.16610218) on 16q23.3 (Table 1). Furthermore, we identified variants that showed genome- wide significant association in eight novel independent loci including rs9853056 (within the STAB1 gene, rs4282054 (within the NT5DC2 gene), rs13083798 (within the PBRM1 gene), rs1108842 (within the GNL3 gene), rs11235 (within the NEK4 gene), rs2710323 (within the ITIH1 gene), rs3617 (within the ITIH3 gene), and rs2535627 (within 200 Kb of ITIH4 gene) at 3p21.1; rs1597466 (within 1 Mb of TSC22D2 gene) at 3q25.1; rs2980879 (within 1 Mb of TRIB1 gene) at 8q24.13; rs7955516 (within 1.3 Mb PDE3A gene) at 12p12.2; rs601339 (within the GPR109A gene) at 12q24.31; rs6488898 (within the ATP6V0A2 gene), rs7133378 (within the DNAH10 gene), rs7305864 (within the CCDC92 gene), and rs7978610 (within the ZNF664 gene at 12q24.31, which is 1.3 Mb away from GPR109A); rs2925979 (within the CMIP at 16q23.2 gene); and rs731839 (within the PEPD gene) at 19q13.11. (Figure 2A– 2E, Table 1). obesity-related morbidity and mortality that poses a severe and global healthcare challenge [5]. Murine studies suggest that adiponectin plays a mediating role in at least some of these obesity- related complications, and although less clearly established in humans, this suggests that understanding the pathophysiology of adiponectin may uncover novel therapeutic targets in major, highly prevalent human disease.[6,7]. Twins and family studies have revealed moderate to high estimates of heritability (30–70%) for plasma adiponectin levels [8,9,10,11]. However, until recently, few genes associated with adiponectin levels have been identified. Candidate and genome- wide association studies (GWAS) have shown pronounced associations between common polymorphisms in the adiponectin gene (ADIPOQ) and adiponectin levels [12,13,14,15]. Results of Meta-Analysis of GWAS A recent meta-analysis of three GWAS for adiponectin levels identified variants in a novel candidate gene, ARL15, that were associated with adiponectin levels, coronary heart disease (CHD), T2D and other metabolic traits [16]. Furthermore, CDH13 and KNG1 genes were found to be associated with adiponectin levels in two studies involving East Asian populations [17,18]. Although part of the variance explained by the ADIPOQ locus, most of the heritability of adiponectin levels remains unaccounted for. Therefore, we sought to identify novel common variants influencing adiponectin levels and test their association with risk of T2D and related metabolic traits within the framework of a large multi-ethnic consortium of GWAS. In our analysis a common variant (rs601339, MAF = 0.18, allele G) downstream of the GPR109A gene (the putative niacin receptor) was associated with adiponectin (ß = 0.04, p = 7.94610210) and HDL-C (ß = 0.03, p = 5.5961027) in the global lipid analysis. In a coincident candidate gene analysis 11 SNPs were typed in GPR109A/B in CoLaus and LOLIPOP cohorts, containing individuals of European descent. A single nominally significant coding SNP R311C (rs7314976, MAF = 0.14) within the GPR109A gene was taken forward for replication and found to be consistently associated with adiponectin in the three cohorts (CoLaus, Fenland and MRC Ely study, n = 8285, p = 4.661028) and HDL-cholesterol (HDL-C) in four cohorts (CoLaus, Fenland, Ely study and Lolipop, n = 18425, p = 2.961028) (Figure S2A, S2B). However R311C and rs601339 were not in linkage disequilibirium with each other (r2 = 0.04). Therefore the two variants represent two independent signals from the same locus but with similar effects on HDL- cholesterol and adiponectin. We combined genome-wide association results of 35,355 individuals from three different ethnicities (white Europeans (n = 29,347), African American s(n = 4,232) and East Asians (n = 1,776)), applying a novel meta-analytic method to allow for heterogeneity in allelic effects between populations of different ethnic backgrounds. We next examined whether identified genome-wide significant single nucleotide polymorphisms (SNPs) also associated with expression of their nearest gene in human adipocytes, the main source of adiponectin. Since adiponectin has been associated with T2D, insulin resistance and metabolic traits we next investigated whether a multi-SNP genotypic risk, comprising genome-wide significant SNPs for adiponectin levels, also influenced risk of T2D and related traits measured in the DIAbetes Genetics Replication and Meta-analysis (DIAGRAM+) In silico follow-up phase. Author Summary Serum adiponectin levels are highly heritable and are inversely correlated with the risk of type 2 diabetes (T2D), coronary artery disease, stroke, and several metabolic traits. To identify common genetic variants associated with adiponectin levels and risk of T2D and metabolic traits, we conducted a meta-analysis of genome-wide association studies of 45,891 multi-ethnic individuals. In addition to confirming that variants at the ADIPOQ and CDH13 loci influence adiponectin levels, our analyses revealed that 10 new loci also affecting circulating adiponectin levels. We demonstrated that expression levels of several genes in these candidate regions are associated with serum adiponectin levels. Using a powerful novel method to assess the contribution of the identified variants with other traits using summary-level results from large-scale GWAS consortia, we provide evidence that the risk alleles for adiponectin are associated with deleterious changes in T2D risk and metabolic syndrome traits (triglycerides, HDL, post-prandial glucose, insulin, and waist-to-hip ratio), demonstrating that the identified loci, taken together, impact upon metabolic disease. Introduction Adiponectin is a highly abundant adipocyte-derived plasma protein whose levels correlate inversely with a range of important PLoS Genetics \| www.plosgenetics.org March 2012 \| Volume 8 \| Issue 3 \| e1002607 PLoS Genetics \| www.plosgenetics.org 3 Adiponectin: Multiethnic Genome-Wide Meta-Analysis [19], Meta-Analysis of Glucose and Insulin Related Traits Consortium (MAGIC) [20], Genetic Investigation of ANthropo- metric measures Traits (GIANT) [21] , Global Lipids Genetic Consortium (GLGC) [22], and Body Fat GWAS consortia [23]. PLoS Genetics \| www.plosgenetics.org Results of Meta-Analysis of GWAS In the in-silico follow-up phase 468 SNPs demonstrating genome-wide significant (p,561028 ) or suggestive (p,561026) association with adiponectin in the discovery phase were tested for association in additional European cohorts. (Table S3). These SNPs were tested in seven additional GWAS datasets (n = 6,623 from NHS, HPFS, HABC, ERF2, LLS, GARP and ARIC studies) and the combined meta- March 2012 \| Volume 8 \| Issue 3 \| e1002607 4 Adiponectin: Multiethnic Genome-Wide Meta-Analysis Table 1. Lead SNP per Locus for Genome-Wide Significant SNPs Arising from the Sex-Combined Meta-Analysis in European Populations. Table 1. Lead SNP per Locus for Genome-Wide Significant SNPs Arising from the Sex-Combined Meta-Analysis in European Populations. Nearest** Gene Lead SNP{ Region Chr/position{ EA/NEA" EAF"" Beta1 SE P I2 n Beta1 SE P I2 n Discovery Phase Results Joint Analysis Phase* LYPLAL1 rs3001032 1q41 1/217794402 T/C 0.7 20.02 0.005 1.98E-06 0 29,321 20.02 0.004 3.60E-08 0 35,930 GNL3 rs1108842 3p21.1 3/52695120 C/A 0.50 0.03 0.004 3.66E-11 0.33 29,338 0.03 0.004 1.39E-13 0.2 35,962 TSC22D2 rs1597466 3q25.1 3/151538251 T/G 0.1 20.04 0.008 1.88E-08 0 29,319 20.03 0.007 1.62E-06 0.1 35,794 ADIPOQ rs6810075 3q27.3 3/188031259 T/C 0.6 0.06 0.005 3.60E-41 0 29,140 0.06 0.004 1.19E-43 0 35,749 VEGFA rs998584 6q21.1 6/43865874 C/A 0.5 0.03 0.005 5.84E-08 0.3 28,167 0.03 0.005 3.25E-08 0.2 34,108 TRIB1 rs2980879 8q24.13 8/126550657 T/A 0.7 0.03 0.005 1.08E-08 0 24,084 0.03 0.005 7.13E-09 0 30,708 PDE3A rs7955516 12q12.2 12/20389303 C/A 0.4 0.03 0.005 2.43E-08 0.1 29,178 0.02 0.004 4.45E-08 0 38,276 GPR109A rs601339 12q24.31 12/121740696 G/A 0.2 0.04 0.006 3.87E-11 0 29,325 0.03 0.005 7.81E-10 0.3 35,947 DNAH10 rs7133378 12q24.31 12/122975455 G/A 0.7 20.03 0.005 1.29E-09 0 29,223 20.02 0.004 6.21E-07 0.5 35,697 CMIP rs2925979 16q23.2 16/80092291 T/C 0.3 20.04 0.005 1.87E-18 0 29,347 20.04 0.005 1.21E-20 0 35,970 CDH13 rs12922394 16q23.3 16/81229828 T/C 0.1 20.10 0.011 3.16E-18 0.3 24,466 20.08 0.010 1.99E-15 0.4 31,089 PEPD rs731839 19q13.11 19/38590905 G/A 0.35 20.04 0.005 2.20E-13 0.03 29,166 20.03 0.004 7.97E-12 0.4 35,771 All SNPs achieving genome-wide significance in the joint analysis phase are marked in italics. Joint analysis indicates results from the meta-analysis of discovery and follow-up in-silico and de-novo phases. When possible, plausible biological candidate genes have been listed; otherwise, the closest gene is designated. {Lead SNP is the SNP with the lowest p-value for each locus. 1Betas are estimated from models using the natural log transformed adiponectin. "EA: Effect allele, NEA: Non-effect allele. Results of Meta-Analysis of GWAS ""EAF: Effect allele frequency. doi:10.1371/journal.pgen.1002607.t001 Table 1. Lead SNP per Locus for Genome-Wide Significant SNPs Arising from the Sex-Combined Meta-Analysis in European Populations. Locus for Genome-Wide Significant SNPs Arising from the Sex-Combined Meta-Analysis in European (n = 12,662) separately (Figure S2A, S2B, Tables S5 and S6). Although no novel loci reached genome-wide significance in women or men separately, three loci (chromosome 3p, 8 and 12) associated with adiponectin levels in the sex-combined analysis showed evidence of association (p value,561028) in women (Figure S3). Since different assays were used to measure adiponectin levels, we next tested whether stratification by assay rendered similar results and found the results were highly concordant with the combined analysis. GWAS for high molecular weight adiponectin in the CHS study (n = 2,718) identified 2 SNPs in ADIPOQ (rs17300539, p = 3.0610216) and CMIP (rs2927307, p = 2.761028). These two genes are located within the loci identified in our discovery meta-analysis of adiponectin levels. analysis of the discovery and follow-up in-silico GWAS datasets detected additional loci on chromosomes 1q41 near the LYPLAL1 gene (rs3001032, p = 3.661028) and chromosome 6p21.1 near the VEGFA gene (rs998584, p = 5.8610212) that reached genome-wide significance. While we confirmed seven loci that had reached genome-wide significance at the discovery stage (Table 1, Figure 2F and 2G, Table S2), two identified loci (3q25.1 and 12q24.31) did not remain genome-wide significant in the joint analysis of discovery and follow-up results. De novo follow-up phase. Next, in the de-novo genotyping follow-up phase, we genotyped 10 SNPs with suggestive evidence of association (561028,p,561026) from the meta-analysis of the discovery and in-silico follow-up phases in an additional 3,913 individuals. Meta-analyzing the discovery and 2 follow-up stages identified a SNP in ARL15 (rs6450176 [G]; ß = 0.026, p = 5.861028), which was initially described in a previous GWAS for adiponectin levels (Table S3) [16]. Gene Expression Studies Through gene expression studies we sought to address two questions: First, are any of the SNPs that were genome-wide significant for adiponectin levels associated with expression of their nearest transcripts (cis-eQTLs) and second, whether mRNA levels of loci identified through the GWAS for adiponectin levels are associated with circulating adiponectin levels. To address the first question, we examined whether SNPs within 1 Mb of the SNPs achieving genome-wide significance in the discovery stage were associated with the expression levels of nearby genes in human adipocytes from 776 participants of the MuTHER Consortium [25]. We identified 74 SNPs in three eQTLs to be associated with the expression of five genes in adipocytes, using an array-wide level of statistical significance for eQTLs (P,5.161025. See Materials and Methods for details). These genes included: NT5DC2 on chromosome 3; CCDC92, GPR109A, and ZNF664 on chromo- some12; and PEPD on chromosome 19 (Table 3). The cis-eQTL Multi-ethnic meta-analysis. To identify loci influencing adiponectin levels in non-European individuals we performed an additional analysis in 4,232 individuals from an African American population and 1,776 individuals from an East Asian population. In the African American populations, only associations at the ADIPOQ locus reached genome-wide significance, while in the East Asian population there was evidence of association at the ADIPOQ and CDH13 loci (Table S4). Subsequently, we performed a meta-analysis that combined all available GWAS including those of white European origin, African American and East Asian ancestry using novel method MANTRA [24]. This analysis identified two additional loci in or near IRS1 gene on 2q36.3 and at 6q24.1 within a gene desert. (Table 2, Figure 1B). In the African American populations, only associations at the ADIPOQ locus reached genome-wide significance, while in the East Asian population there was evidence of association at the ADIPOQ and CDH13 loci (Table S4). Subsequently, we performed a meta-analysis that combined all available GWAS including those of white European origin, African American and East Asian ancestry using novel method MANTRA [24]. This analysis identified two additional loci in or near IRS1 gene on 2q36.3 and at 6q24.1 within a gene desert. (Table 2, Figure 1B). Secondary GWAS analyses. We next performed meta- analysis of the GWAS data in women (n = 16,685) and men Secondary GWAS analyses. Gene Expression Studies We next performed meta- analysis of the GWAS data in women (n = 16,685) and men PLoS Genetics \| www.plosgenetics.org March 2012 \| Volume 8 \| Issue 3 \| e1002607 5 Adiponectin: Multiethnic Genome-Wide Meta-Analysis Figure 1. Manhattan plots for meta-analyses in the discovery phase. A) Combined sex analysis in European populations, B) Meta-Analysis of Multiple Ethnicities. The Manhattan plots show 2Log10 (p-value) measures for association between single nucleotide polymorphisms (SNPs) and chromosomal position. The SNPs that achieved genome-wide significance are highlighted in green. doi:10.1371/journal.pgen.1002607.g001 Adiponectin: Multiethnic Genome Wide Meta Analysis Figure 1. Manhattan plots for meta-analyses in the discovery phase. A) Combined sex analysis in European populations, B) Meta-Analysis of Multiple Ethnicities. The Manhattan plots show 2Log10 (p-value) measures for association between single nucleotide polymorphisms (SNPs) and chromosomal position. The SNPs that achieved genome-wide significance are highlighted in green. doi:10.1371/journal.pgen.1002607.g001 SNPs often are proxies for the lead SNPs from the GWAS, however, this relationship may also be influenced through mechanisms that are independent from gene expression, such as gene function. number of statistically independent SNPs (p-value threshold of 561024) among the SNPs that were genome-wide significant for adiponectin. These include associations with HDL-C (n= 104 SNPs), triglycerides (TG) (n = 65 SNPs), total cholesterol (TC, n = 12 SNPs), LDL-cholesterol (LDL–C, n = 11 SNPs), and waist- hip ratio (WHR) (n= 65 SNPs) [26]. (However, we note that since sample sizes are different across different consortia power to identify associations is not consistent.) Among these, coding and intronic variants in STAB1 and NT5DC2 genes were associated with WHR and HDL-C, while the variants 1 Mb near TRIB1 were associated with all lipid traits. The coding and intronic variants ariants in the locus on chromosome 12 harboring ZNF664, CCDC92, and DNAH10 showed evidence of association with WHR, HDL-C, and TG. Finally, variants in the PEPD gene were associated with TG. We next identified that mRNA levels of 18 genes arising from six candidate loci were correlated with circulating adiponectin levels (Table 4). Since circulating adiponectin levels may be associated with a surplus of adipocyte transcripts we next tested for enrichment of signal from the candidate loci. Gene Expression Studies There were 133 transcripts in the identified candidate regions, of which 8.2% (11/ 133) were associated with adiponectin levels at an array-wide level of significance (p,261026), while 7.5% of the 24k probes on the entire array exceeded the same p-value threshold, indicating there was therefore no additional enrichment of signal at these candidate loci. We next calculated a multi-SNP genotypic risk score based genome-wide significant SNPs from the discovery phase. This multi-SNP genotypic risk score explained 5% of the variance of natural log-transformed adiponectin levels. We then tested the association of this risk score with risk of T2D and metabolic related traits. The multi-SNP genotypic risk score was associated with increased risk for T2D (ß = 0.3, p = 4.361023), where ß is the average additive effect of adiponectin-decreasing risk alleles on the PLoS Genetics \| www.plosgenetics.org Biological Relevance of the GWAS Loci In the first step toward understanding the biological relevance of the identified regions, we examined the genes harbored by the associated loci using human disease and animal databases. Although some of the genes in these loci do not have a known function, several signify diverse biological functions. On chromosome 1, the lead SNP was located 300 kb from the LYPLAL1, a protein that regulates phospholipids on cellular membranes. Independent efforts have also identified this locus in other metabolic/obesity related traits GWAS: first with WHR (rs2605100; r2 = 0.49 [21] and rs4846567; r2 = 0.55 [27] respectively with the lead adiponectin SNP, rs3001032), and more recently with fasting insulin by a joint meta-analysis including the interaction between SNP and BMI (MF Hivert for the MAGIC investigators, personal communication). In the same report by MAGIC, variants near IRS1 (insulin receptor substrate 1) and PEPD (a protein that hydrolyzes dipeptides and tripeptides) have also been associated with fasting insulin at genome wide significant levels, demonstrating a close link between adiponectin regulation and insulin resistance pathways. Moreover, both IRS1 and PEPD have been associated with T2D (IRS1 in DIAGRAM [28] and PEPD in a Japanese population [29]; p = 9.3610212 and p = 1.461025, respectively). All three homologous genes GPR109A/B/81 located on chromosome 12 are predominantly expressed in adipocytes and mediate antilipolytic effects [45]. Our eQTL results (Table 3) and the correlation between mRNA and adiponectin levels (Table 4) argue strongly for a role of GPR109A at this locus. GPR109A (also known as NIACR1) is a receptor with a high-affinity, concentra- tion-dependent response to nicotinic acid (niacin) [45]. Treatment by niacin increases serum adiponectin levels by up to 94% in obese men with metabolic syndrome in a time- and dose-dependent manner [46]. Functional studies in GPR109A receptor knockout mice demonstrate that niacin increases serum total and HMW adiponectin concentrations and decreases lipolysis following GPR109A receptor activation [47]. Moreover, a recent meta- analysis on cohorts containing extremes of HDL-C provided evidence suggestive of association in GPR109A/B/81 [48]. The lead SNP at 3p21.1 falls within GNL3 that is located in a genomic region containing many genes which could have potential functions in metabolism. Our data provide evidence that adiponectin levels were correlated with human adipocyte mRNA levels of many genes in this region (GLYCTK, SEMA3G, STAB1, PBRM1, SFMBT1; see Table 4). However, this association does not imply a direct influence of these genes on adiponectin level. Biological Relevance of the GWAS Loci Among those genes, STAB1 encodes for stabilin 1, described as an endocytic receptor for advanced glycation end products and may have a function in angiogenesis, lymphocyte homing, cell adhesion, or receptor scavenging for acetylated low-density lipoprotein [30]. Finally, variants in ZNF664 have been associated with CHD, HDL-C and TG levels in a large meta-analysis of over 100,000 individuals of European ancestry [22]. The sex heterogeneity observed in this study is comparable to our finding that the more loci associated with adiponectin at genome wide significance level have been shown in female stratified analysis. Taken together, the loci identified in this large-scale GWAS for adiponectin levels highlight many genes with demonstrated relationships with metabolic disease. T2D and Metabolic Traits doi:10.1371/journal.pgen.1002607.g002 log odds ratio of T2D), increased TG (ß = 0.25, p = 2.6610214), increased WHR adjusted for BMI (ß = 0.18, p = 1.861025), increased post-prandial glucose (ß = 0.25, p = 0.01), increased fasting insulin (ß = 0.05, p = 0.01), homeostatic model assessment- insulin resistance (HOMA-IR) (ß = 0.04, p = 0.047), and with lower HDL-C concentrations (ß = 20.24, p = 4.5610213) and decreased BMI (ß = 20.16, p = 1.461024). (Table 5). is the vascular endothelial growth factor A gene, a known gene in a variety of vascular endothelial cell functions, such as angiogenesis and maintenance of the glomerular endothelium in nephrons [32]. Variants in this gene are also associated with diabetic retinopathy and WHR [27,33]. Moreover, the product of VEGFA interacts with resveratrol, which has been shown to have a beneficial influence in some metabolic traits, including diabetes [34]. Rodent studies show that resveratrol decreases blood glucose, blood insulin, and glycated hemoglobin, as well as increases insulin sensitivity in animals with hyperglycemia (reviewed in [35]). Resveratrol also inhibits TNF-a-induced reductions in adiponectin levels in 3T3-L1 adipocytes [36]. Furthermore, it has been shown that resveratrol modulates adiponectin expression and improves insulin sensitivity, likely through the inhibition of inflammatory- like response in adipocytes [37]. At this locus, VEGFA mRNA levels in adipocytes were the strongest association with adiponectin levels (Table 4). Also likely involved in vascular biology, TRIB1 encodes a G protein-coupled receptor-induced protein interacting with MAP kinases that regulates proliferation and chemotaxis of vascular smooth muscle cells [38]. TRIB1 expression was shown to be elevated in human atherosclerotic arteries [39]. Several variants (rs2954029, rs2954021, rs17321515; all in moderate LD with our lead SNP) in the TRIB1 gene have been associated with HDL-C, LDL-C and CHD risk in European and Asian populations [22,40,41,42,43]. These two loci (TRIB1 and VEGFA) argue for the importance of vascular biology in adiponectin regulation as underlined previously by findings of adiponectin levels associated with variants near CDH13 (a receptor for adiponectin expressed by endothelial smooth muscle) [44]. PLoS Genetics \| www.plosgenetics.org Discussion In this comprehensive multi-ethnic analysis of the genetic influences on adiponectin levels and their impact on metabolic traits and T2D, we have identified 10 novel loci and confirmed the associations of variants in the ADIPOQ and CDH13 loci with adiponectin levels. The adiponectin risk alleles were associated with T2D and related metabolic traits such as BMI, WHR, TG, HDL-C, 2-hour glucose, HOMA-IR and fasting insulin. These findings demonstrate that adiponectin, T2D and metabolic syndrome have a shared allelic architecture. T2D and Metabolic Traits Using data from several large-scale GWAS consortia, some of the significantly associated variants identified here demonstrated associations with T2D and its related traits (Table S7A, S7B, S7C, and S7D). Several individual SNPs showed evidence for association with T2D and various metabolic traits after accounting for the March 2012 \| Volume 8 \| Issue 3 \| e1002607 6 Adiponectin: Multiethnic Genome-Wide Meta-Analysis Figure 2. Regional plots of eight newly discovered genome-wide significant chromosomal regions associated with adiponectin concentrations in European populations. A) chromosome 16q23.2, B) chromosome 19 q13.11 C) Chromosome 3p21.1, D) two loci on chromosome 12q24.31, E) chromosome 8q24.13, F) chromosome 6p21.1, and G) chromosome 1q41. In each panel, purple diamonds indicate the top PLoS Genetics \| www.plosgenetics.org 7 March 2012 \| Volume 8 \| Issue 3 \| e1002607 Figure 2. Regional plots of eight newly discovered genome-wide significant chromosomal regions associated with adiponectin concentrations in European populations. A) chromosome 16q23.2, B) chromosome 19 q13.11 C) Chromosome 3p21.1, D) two loci on chromosome 12q24.31, E) chromosome 8q24.13, F) chromosome 6p21.1, and G) chromosome 1q41. In each panel, purple diamonds indicate the top PLoS Genetics \| www.plosgenetics.org PLoS Genetics \| www.plosgenetics.org March 2012 \| Volume 8 \| Issue 3 \| e1002607 7 Adiponectin: Multiethnic Genome-Wide Meta-Analysis Adiponectin: Multiethnic Genome-Wide Meta-Analysis SNPs, which have the strongest evidence of association. Each circle shows a SNP with a color scale relating the r2 value for that SNP and the top SNP from HapMap CEU. Blue lines indicate estimated recombination rates from HapMap. The bottom panels illustrate the relative position of genes near each locus. Candidate genes are indicated by red ovals. doi:10.1371/journal.pgen.1002607.g002 SNPs, which have the strongest evidence of association. Each circle shows a SNP with a color scale relating the r2 value for that SNP and the top SNP from HapMap CEU. Blue lines indicate estimated recombination rates from HapMap. The bottom panels illustrate the relative position of genes near each locus. Candidate genes are indicated by red ovals. PLoS Genetics \| www.plosgenetics.org Shared Allelic Architecture of Adiponectin Levels and Metabolic Traits Interestingly, several of the genes near lead genome-wide significant SNPs have been implicated in angiogenesis, which might be important for adipose tissue expansion, highlighting the recurring theme of ‘‘adipose tissue expandability’’ in the genetic origins of obesity-related complications [31]. For example, VEGFA Using a multi-SNP genotypic risk score we attempted to understand if the allelic architecture of adiponectin levels was shared with T2D and metabolic traits. This risk score was March 2012 \| Volume 8 \| Issue 3 \| e1002607 8 Adiponectin: Multiethnic Genome-Wide Meta-Analysis Table 2. Genome-Wide Significant SNPs from the Sex-Combined Multi-Ethnic Meta-Analysis. Nearby Gene Lead SNP{ Gene region chr/position{ EA/NEA" EAF"" (CEU/EA/AA) Multi-Ethnic Fixed Effects Meta-analysis Multi-Ethnic Random Effects Meta-analysis MANTRA N Beta (SE) pvalue Q-Value I2 Beta (SE) pvalue BF1 phet{{ LYPLAL1 rs2791553 1q41 1/217742665 G/A 0.6/0.46/0.54 20.02(0.004) 4.91E-07 25.18 0 20.02(0.004) 4.91E-07 6.3 0.06 37,665 IRS1 rs925735 2q36.3 2/226887874 G/C 0.64/0.89/0.74 20.02(0.004) 1.88E-08 22.15 0.01 20.02(0.004) 2.12E-08 8.1 0.06 37,638 GNL3 rs2590838 3p21.1 3/52597126 G/A 0.5 1/0.34/0.54 20.03(0.004) 4.08E-15 28.85 0.06 20.03(0.004) 1.88E-13 14.1 0.05 37,680 ADIPOQ rs6810075 3q27.3 3/188031259 T/C 0.93/1/0.86 0.06(0.004) 1.10E-43 27.44 0.02 0.06(0.004) 2.41E-42 43.6 0.16 31,533 - rs592423 6q24.1 6/139882386 C/A 0.54/0.36/0.41 0.02(0.004) 3.59E-07 15.46 0 0.02(0.004) 3.59E-07 6.5 0.03 37,430 TRIB1 rs2980879 8q24.13 8/126550657 T/A 0.69/0.77/0.67 0.03(0.004) 9.91E-10 21.08 0 0.03(0.004) 9.91E-10 8.2 0.04 32,426 GPR109A rs601339 12q24.31 12/121740696 G/A 0.19/0.39/0.31 0.03(0.005) 3.77E-09 36.11 0.25 0.03(0.006) 4.31E-06 8.3 0.09 37666 CMIP rs2925979 16q23.2 16/80092291 T/C 0.3 0/0.43/0.31 20.04(0.004) 3.12E-21 23.12 0 20.04(0.004) 3.12E-21 19.8 0.31 37,687 CDH13 rs12051272 16q23.3 16/81220789 T/G 0.03/0.33/0.03 20.26(0.017) 4.74E-51 39.17 0.62 20.26(0.032) 1.10E-14 66.0 1.00 24,216 PEPD rs4805885 19q13.11 19/38597963 T/C 0.39/0.64/0.41 20.03(0.004) 1.65E-11 34.94 0.23 20.03(0.005) 2.05E-08 9.9 0.05 37,479 The novel loci identified using Multi-Ethnic Meta-analysis (that were not identified in the European only analysis) are listed in bold. When possible, plausible biological candidate genes have been listed; otherwise, the closest gene is designated. {Lead SNP is the SNP with the lowest p-value for each locus. {Positions are relative to Human Genome NCBI Build 36. 1log10 Bayes factor (BF) from the MANTRA analysis. A log10 BF of 6 and higher was considered as a conservative threshold for genome-wide significance. {{The posterior probability of heterogeneity between studies. "EA: effect allele, NEA: non-effect allele. ""EAF: Frequency of effect allele in CEU, East Asian, and AA, populations respectively. PLoS Genetics \| www.plosgenetics.org March 2012 \| Volume 8 \| Issue 3 \| e1002607 Shared Allelic Architecture of Adiponectin Levels and Metabolic Traits Nonetheless, in aggregate, these results provide strong evidence that the genetic determinants of adiponectin levels are shared with metabolic disease, and in particular, traits related to insulin resistance. Table 4. The Association of mRNA Levels from Genes in Candidate Loci in Human Adipocytes with Circulating Adiponectin Levels. Gene Gene region GeneStart GeneEnd Beta1 Pvalue GLYCTK 3p21.1 52296875 52304311 0.060 1.77E-20 SEMA3G 3p21.1 52442307 52454083 20.018 9.28E-06 STAB1 3p21.1 52504395 52533551 20.039 2.26E-14 PBRM1 3p21.1 52554407 52688779 0.007 2.49E-04 SFMBT1 3p21.1 52913666 53055110 0.010 2.53E-08 DNAJB11 3q27.3 187771160 187786283 20.014 3.31E-07 EIF4A2 3q27.3 187984054 187990379 0.021 1.53E-08 ADIPOQ 3q27.3 188043156 188058944 0.054 1.03E-13 MAD2L1BP 6q21.1 43711554 43716666 0.009 4.09E-04 VEGFA 6q21.1 43845923 43862199 0.012 2.15E-09 ZCCHC8 12q24.31 121523387 121551471 0.011 2.60E-04 GPR109B 12q24.31 121765255 121767392 0.010 3.74E-06 GPR109A 12q24.31 121778105 121781082 0.026 1.80E-11 PITPNM2 12q24.31* 122033979 122160928 20.010 5.09E-06 U1SNRNPBP 12q24.31 122508604 122516894 0.011 1.72E-04 ATP6V0A2 12q24.31 122762817 122812252 20.008 2.86E-04 ZNF664 12q24.31 123023622 123065922 0.010 8.28E-06 SLC7A10 19q13.11 38391409 38408596 0.072 1.66E-14 1Betas are estimated from log transformed and quantile-quantile normalized values. These two loci are independent loci. doi:10.1371/journal.pgen.1002607.t004 Table 4. The Association of mRNA Levels from Genes in Candidate Loci in Human Adipocytes with Circulating Adiponectin Levels. We note that there are several strengths and limitations of this study. Our main findings, identifying genetic determinants of adiponectin levels, are based on the largest meta-analysis to date and include results from three ethnicities. The availability of expression data from human adipose tissue permitted the association of identified SNPs with mRNA levels at candidate genes and, in turn, correlation of these mRNA levels with circulating adiponectin itself. While access to the data from large consortia permitted assessment of the relevance of the identified SNPs to T2D and components of the metabolic syndrome, we note that a subset of the cohorts included in our GWAS were also included in these external consortia. However, we note that even if we assume that all ADIPOGen study participants were included in the external consortia, for cohorts participating in both studies, that the majority of data in these external consortia still arises from study participants not present in ADIPOGen (minimum percent of non-overlapping subects: 86.8%, 85.5%, 86.4% and 82.5% for MAGIC, GLGC, GIANT, and DIAGRAM+ consortia, respec- tively). Shared Allelic Architecture of Adiponectin Levels and Metabolic Traits adiponectin risk alleles at ZNF664 and PEPD are of considerable interest since they impart deleterious changes on aspects of the metabolic syndrome (increased TC, TG, LDL-C and WHR and decreased HDL-C), but also act to decrease BMI and percent fat. Our data do not provide direct evidence as to whether the genetic determinants of adiponectin levels influence these traits through adiponectin itself, or through pleiotropic pathways and therefore do not constitute a Mendelian randomization study. These findings provide a note of caution for Mendelian randomization studies, which may be prone to erroneous conclusions if pleiotropic effects of tested variants are not considered. Nonetheless, in aggregate, these results provide strong evidence that the genetic determinants of adiponectin levels are shared with metabolic disease, and in particular, traits related to insulin resistance. associated with increased risk of T2D and traits associated with insulin resistance and the metabolic syndrome. However, unexpectedly, adiponectin decreasing alleles were associated with a decrease in BMI. In our adiponectin GWAS, BMI was included as a covariate in order to avoid direct identification of obesity SNPs since BMI is strongly related to adiponectin levels [49,50]. Furthermore, this unexpected direction of effect was entirely explained by SNPs at the ZNF664 and PEPD loci; when these loci were removed from the analysis, the association of the genotypic risk score with BMI disappeared (results not shown). Therefore, associated with increased risk of T2D and traits associated with insulin resistance and the metabolic syndrome. However, unexpectedly, adiponectin decreasing alleles were associated with a decrease in BMI. In our adiponectin GWAS, BMI was included as a covariate in order to avoid direct identification of obesity SNPs since BMI is strongly related to adiponectin levels [49,50]. Furthermore, this unexpected direction of effect was entirely explained by SNPs at the ZNF664 and PEPD loci; when these loci were removed from the analysis, the association of the genotypic risk score with BMI disappeared (results not shown). Therefore, Our data do not provide direct evidence as to whether the genetic determinants of adiponectin levels influence these traits through adiponectin itself, or through pleiotropic pathways and therefore do not constitute a Mendelian randomization study. These findings provide a note of caution for Mendelian randomization studies, which may be prone to erroneous conclusions if pleiotropic effects of tested variants are not considered. Shared Allelic Architecture of Adiponectin Levels and Metabolic Traits doi:10.1371/journal.pgen.1002607.t002 osgenetics.org 9 March 2012 \| Volume 8 \| PLoS Genetics \| www.plosgenetics.org March 2012 \| Volume 8 \| Issue 3 \| e1002607 9 associated with increased risk of T2D and insulin resistance and the metabolic s unexpectedly, adiponectin decreasing alleles a decrease in BMI. In our adiponectin GWA as a covariate in order to avoid direct ide SNPs since BMI is strongly related to adipo Furthermore, this unexpected direction o explained by SNPs at the ZNF664 and PEPD were removed from the analysis, the associa risk score with BMI disappeared (results no Table 3. The Association of Lead Genom Gene Lead SNP- Cis-eQTL{ Chr Transcript Start Site NT5DC2 rs13081028 3 52533424 GPR109A rs2454722 12 121778105 CCDC92 rs10773049 12 122986907 ZNF664 rs825453 12 123074711 PEPD rs8182584 19 38569694 {Lead SNP is the SNP with the lowest p-value for ea {{Lead SNP is the SNP with the lowest p-value for e "EA: Effect allele. ""EAF: Frequency of effect allele. 1Betas are estimated expression levels of the genes P value for lead SNP is the SNP in gene expression P value for lead SNP in meta-analysis from discove $r2 LD between lead SNP from expression and lead doi:10.1371/journal.pgen.1002607.t003 Table 4. The Association of mRNA Leve Candidate Loci in Human Adipocytes wit Adiponectin Levels. Gene Gene region GeneStart GeneE GLYCTK 3p21.1 52296875 523043 SEMA3G 3p21.1 52442307 5245408 STAB1 3p21.1 52504395 5253355 PBRM1 3p21.1 52554407 5268877 SFMBT1 3p21.1 52913666 530551 DNAJB11 3q27.3 187771160 1877862 EIF4A2 3q27.3 187984054 1879903 ADIPOQ 3q27.3 188043156 1880589 MAD2L1BP 6q21.1 43711554 4371666 VEGFA 6q21.1 43845923 4386219 ZCCHC8 12q24.31 121523387 1215514 GPR109B 12q24.31 121765255 1217673 GPR109A 12q24.31 121778105 1217810 PITPNM2 12q24.31* 122033979 1221609 U1SNRNPBP 12q24.31 122508604 1225168 ATP6V0A2 12q24.31 122762817 1228122 ZNF664 12q24.31 123023622 1230659 SLC7A10 19q13.11 38391409 3840859 1Betas are estimated from log transformed and qua values. These two loci are independent loci. doi:10.1371/journal.pgen.1002607.t004 Adiponectin: Multiethnic Genome-Wide Meta-Analysis Table 3. The Association of Lead Genome-Wide Significant SNPs for Adiponectin with mRNA Levels of Their Nearest Gene. Table 3. The Association of Lead Genome-Wide Significant SNPs for Adiponectin with mRNA Levels of Their Nearest Gene. Table 3. The Association of Lead Genome-Wide Significant SNPs for Adiponectin with mRNA Le Table 3. The Association of Lead Genome-Wide Significant SNPs for Adiponectin with mRNA Levels of Their Nearest Gene. Shared Allelic Architecture of Adiponectin Levels and Metabolic Traits Gene Lead SNP- Cis-eQTL{ Chr Transcript Start Site Transcript End Site EA" EAF"" Beta (SE)1 P-Exp P-GWAS** lead SNP- GWAS{{ r2$ NT5DC2 rs13081028 3 52533424 52544133 G 0.444 0.14(0.02) 1.32E-19 1.05E-09 rs1108842 0.84 GPR109A rs2454722* 12 121778105 121781082 G 0.166 20.15(0.03) 1.71E-09 3.87E-11 rs601339 1 CCDC92 rs10773049 12 122986907 123023116 T 0.611 0.15(0.02) 8.09E-22 2.67E-08 rs7133378 0.02 ZNF664 rs825453 12 123074711 123065922 T 0.615 20.04(0.01) 4.51E-05 4.03E-08 rs7978610 0.03 PEPD rs8182584 19 38569694 38704639 T 0.364 20.13(0.02) 9.96E-10 6.64E-11 rs731839 1 {Lead SNP is the SNP with the lowest p-value for each gene in gene expression data. {{Lead SNP is the SNP with the lowest p-value for each locus in meta-analysis from discovery phase. "EA Eff t ll l Gene Lead SNP- Cis-eQTL{ Chr Transcript Start Site Transcript End Site EA" EAF"" Beta (SE)1 P-Exp* P-GWAS** lead SNP- GWAS{{ r2$ NT5DC2 rs13081028 3 52533424 52544133 G 0.444 0.14(0.02) 1.32E-19 1.05E-09 rs1108842 0.84 GPR109A rs2454722* 12 121778105 121781082 G 0.166 20.15(0.03) 1.71E-09 3.87E-11 rs601339 1 CCDC92 rs10773049 12 122986907 123023116 T 0.611 0.15(0.02) 8.09E-22 2.67E-08 rs7133378 0.02 ZNF664 rs825453 12 123074711 123065922 T 0.615 20.04(0.01) 4.51E-05 4.03E-08 rs7978610 0.03 PEPD rs8182584 19 38569694 38704639 T 0.364 20.13(0.02) 9.96E-10 6.64E-11 rs731839 1 {Lead SNP is the SNP with the lowest p-value for each gene in gene expression data. {{Lead SNP is the SNP with the lowest p-value for each locus in meta-analysis from discovery phase. "EA: Effect allele. ""EAF: Frequency of effect allele. 1Betas are estimated expression levels of the genes. P value for lead SNP is the SNP in gene expression data. P value for lead SNP in meta-analysis from discovery phase. $r2 LD between lead SNP from expression and lead SNP from meta-analysis. doi:10.1371/journal.pgen.1002607.t003 adiponectin risk alleles at ZNF664 and PEPD are of considerable interest since they impart deleterious changes on aspects of the metabolic syndrome (increased TC, TG, LDL-C and WHR and decreased HDL-C), but also act to decrease BMI and percent fat. adiponectin risk alleles at ZNF664 and PEPD are of considerable interest since they impart deleterious changes on aspects of the metabolic syndrome (increased TC, TG, LDL-C and WHR and decreased HDL-C), but also act to decrease BMI and percent fat. Shared Allelic Architecture of Adiponectin Levels and Metabolic Traits Therefore, since a substantial majority of participants are independent between ADIPOGen and these consortia, it is unlikely that our findings demonstrating a shared allelic architec- ture between adiponectin levels and these traits are spurious. Further, we suggest that locus, 6q24.1, identified only through multi-ethnic meta-analysis using MANTRA and not confirmed through fixed and random effects meta-analysis, be replicated for confirmation of this finding. In conclusion, the data presented in this study provide strong evidence of association for 10 novel loci for adiponectin levels. March 2012 \| Volume 8 \| Issue 3 \| e1002607 10 Adiponectin: Multiethnic Genome-Wide Meta-Analysis Table 5. Results of Association of Multi-SNP Genotypic Risk Score with Diabetes and Related Traits Table 5. Results of Association of Multi-SNP Genotypic Risk Score with Diabetes and Related Traits. Trait N Effect1 (95% CI) P Consortium T2D* 22,044 0.301 (0.09, 0.51) 4.3E-03 DIAGRAM+ BMI (SD units) 121,335 20.162 (20.25, 20.08) 1.4E-04 GIANT WHR* 77,167 0.177 (0.1, 0.26) 1.8E-05 GIANT Percent Fat 34,853 20.052 (20.15, 0.05) 0.31 Body Fat Percent Fasting Glucose (mmol/L) 46,186 0.011 (20.03, 0.05) 0.58 MAGIC Fasting Insulin(pmol/L) 38,238 0.05 (0.01, 0.09) 1.5E-02 MAGIC HomaB 36,466 0.033 (0, 0.07) 5.1E-02 MAGIC Homa IR 37,037 0.042 (0, 0.08) 4.7E-02 MAGIC 2hr Glucose(mmol/L) 15,234 0.245 (0.06, 0.44) 1.1E-02 MAGIC HbA1C (%) 35,908 20.002 (20.04, 0.03) 0.91 MAGIC TG(SD units) 93,440 0.248 (0.18, 0.31) 2.6E-14 GLGC HDL-C (SD units) 96,748 20.243 (20.31, 20.18) 4.5E-13 GLGC LDL-C (SD units) 92,348 0.023 (20.05, 0.09) 0.52 GLGC TC (SD units) 97,021 0.0003 (20.07, 0.07) 0.99 GLGC T2D: Type 2 diabetes, BMI: Body mass Index, WHR: Waist to hip ratio, HbA1C: hemoglobin A1C, TG: Triglyceride, HDL-C: High Density Lipoprotein Cholesterol, LDL-C: Low Density Lipoprotein Cholesterol, TC: Total Cholesterol. 1Effect is mean change in trait or disease per adiponectin-decreasing allele. Waist to hip ratio adjusted for BMI. Significantly associated trait is coded in bold. doi:10.1371/journal.pgen.1002607.t005 SNPs of the first stage also associate with mRNA levels of nearest gene(s) expressed using adipose tissue of 776 European women. We also tested for association between adiponectin levels and mRNA levels of the genes in our candidate loci in adipose tissue of a subgroup of 436 individuals [25]. Third, we calculated a multi- SNP genotypic risk score using genome-wide significant adipo- nectin-lowering alleles and tested the association of this risk score with T2D and related metabolic traits. Figure 3 shows a flow chart detailing the study design. Study Design Our study consisted of three stages. First, in the discovery stage we performed a meta-analysis of the GWAS summary statistics of 16 studies involving 29,347 participants of white European origin to detect SNPs that are associated with adiponectin levels. All signals with p,561026 were followed up in seven additional cohorts (n = 6,623) with GWAS data (in-silico phase) that later joined the consortium and then a subset of SNPs (n = 10) by de-novo genotyping in 3,913 additional participants from three cohorts (n = 39,883 for the combined analysis in Europeans). We also performed a multi-ethnic meta-analysis by combining summary statistics from the 16 studies of individuals of white European discovery cohorts (n = 29,347) with those of five cohort studies that included African Americans subjects (n = 4,232) and one East Asian cohort (n = 1,776) to obtain a total 35,355 individuals for the GWAS meta-analysis involving different ethnicities. After identi- fying variation near two genes of pharmaceutical importance (GPR109A and GPR109B), which encode the putative niacin receptors, we typed additional rare coding and tagging variants in a subset of cohorts. Second, we examined whether the identified Ethical Consideration In total, 45,891 individuals from 26 European and 7 non- European cohorts participated in the different phases of this meta- analysis. Participating cohorts were either population-based (n = 23), family-based (n = 4), or case-control (n = 4) studies. The age of participants ranged from 10 to 95 years. Adiponectin levels were measured using ELISA or RIA methods. More details on the study cohorts and adiponectin measurement are presented in the Text S1 and Table S1. In addition, genotyping of four coding and tagging SNPs in the candidate genes, GRP109A and GPR109B, was undertaken in samples from the Lausanne, Lolipop, MRC Ely, and Fenland cohorts. All participants provided informed written consent. The research protocol of all studies were reviewed and approved by institutional ethics review committees at the involved institutions. Shared Allelic Architecture of Adiponectin Levels and Metabolic Traits Further analyses confirmed that the level of expression of some of these candidate genes in human adipocytes correlated directly with adiponectin levels. A multi-SNP genotypic risk score, and several of the identified variants, directly influence parameters of the metabolic syndrome and, in particular, markers of insulin resistance. These findings identify novel genetic determinants of adiponectin levels, which, taken together, influence risk of T2D and markers of insulin resistance. PLoS Genetics \| www.plosgenetics.org Genotyping and Imputation —In silico follow-up: 468 SNPs with p,561026 from the discovery phase (which includes both genome-wide significant [n = 196, p,561028] and ‘‘suggestive’’ [n =272, 561028,p, 561026] SNPs Table S3) were tested for their association in 6,623 individuals from seven additional cohorts with GWAS data that joined the consortium after the discovery stage had been finalized. —De novo follow-up: We next selected the lead SNP arising from selected loci from the joint analysis of the discovery and in- silico follow-up phase with p-values greater than 561028 but less than 561026 and genotyped 10 SNPs in 3,164 samples from the SAPHIR cohort and an additional subgroup of the KORA cohort. Finally, these same SNPs, or their proxy SNPs (n = 2), were tested for association in the THISEAS cohort (n = 738), which had been genotyped using the Metabochip [55]. Study-level summary statistics from the follow-up phases were meta-analyzed with the data from the discovery phase. Eleven coding and tagging variants in two candidate genes of pharmaceutical importance (GPR109A encoding the niacin receptor and GPR109B) were genotyped in a parallel study in Lausanne, Lolipop, MRC Ely, and Fenland white subjects. Genotyping was performed using a KASPar-On-Demand SNP Genotyping Assay (KBioscience Ltd., Hoddesdon, UK). In Lausanne and Lolipop samples the genotyping assay was carried out on 3.75 ng of genomic DNA in 1 ml 1536-well plate reactions, dispensed with a Meridian, microfluidic dispenser (KBioscience Ltd., Hoddesdon, UK), thermocycled using a Hydrocycler (KBioscience Ltd., Hoddesdon, UK). A Pherastar (BMG GmbH, Germany) was used for end-point detection and Kraken-LIMS (KBioscience Ltd., Hoddesdon, UK) was used for automated allele calling. In MRC Ely and Fenland samples, the genotyping assay was carried out on 10 ng of genomic DNA in 5 ml 384-well plate reactions using a G-Storm GS4 Thermal Cycler (GRI, Rayne, UK). The ABI PRISM 7900HT Sequence Detection System (Applied Biosystems, Warrington, UK) was used for end-point detection and allele calling. y p Multi-ethnic meta-analysis. In order to perform a meta- analysis of GWAS data from cohorts of different ethnic backgrounds, we utilized the novel MANTRA (Meta-ANalysis of Trans-ethnic Association studies) software [24]. This method combines GWAS from different ethnic groups by taking advantage of the expected similarity in allelic effects between the most closely related populations. Fixed-effects meta-analysis assumes the allelic effect to be the same in all populations, and cannot account for heterogeneity between ethnic groups. Genotyping and Imputation Populations are then clustered according to their similarity in terms of relatedness as measured by the mean allele frequency difference at 10,000 independent SNPs, and to their allelic effects at the variant. If all populations are assigned to the same cluster, this is equivalent to a fixed allelic effect across all populations (i.e. no trans-ethnic heterogeneity). The posterior distribution of the allelic effect in each population under the Bayesian partition model is approximated by means of a Monte-Carlo Markov chain algorithm. Evidence in favor of association of the trait with the variant was assessed by means of a Bayes’ factor (BF). A log10 BF of 6 or higher is considered a relatively conservative threshold for genome-wide significance. We also performed meta-analysis by using both random and fixed effects models including all ethnicities. Those loci that achieved both a BF.6 in MANTRA and a P-value less than 561027 in multiethnic analysis are presented in Table 2. Genotyping and Imputation All cohorts were genotyped using commercially available Affymetrix or Illumina genome-wide genotyping arrays. Quality control was performed for each study independently and genotype imputation was carried out using IMPUTE, MACH, BimBam or Beagle with reference to either the Phase II CEU, CEU+YRI, or CHB+JPT+CEU HapMap according to the origin of population. Imputation of East Asian genotypes was undertaken by first masking genotypes of 200 SNPs and then imputing them based on the CEU+CHB+JPT panel from HapMap. This resulted in an allelic concordance rate of ,96.7%. For the African Americans, a combined CEU+YRI reference panel was created. This panel included SNPs segregating in both CEU and YRI, as well as SNPs PLoS Genetics \| www.plosgenetics.org March 2012 \| Volume 8 \| Issue 3 \| e1002607 11 Adiponectin: Multiethnic Genome-Wide Meta-Analysis segregating in one panel and monomorphic and non-missing in the other (2.74 million SNPs). Due to the overlap of African American individuals on the Affymetrix 6.0 and IBC arrays [51], it was possible to analyze imputation performance at SNPs not genotyped on Affymetrix 6.0. For imputation based on Affymetrix data, the use of the CEU+YRI panel resulted in an allelic concordance rate of ,95.6% (calculated as 120.5 [imputed_ dosage– chip_dosage]). This rate is comparable to rates calculated for individuals of African descent imputed with the HapMap 2 YRI individuals. Table S1 summarizes the genotyping methods used for each cohort, genotype-calling algorithms, imputation algorithms and exclusion thresholds. SNP-level quality control metrics were applied prior to meta-analysis for each cohort. These were: call rate $95%, minor allele frequency (MAF)$1%, Hardy- Weinberg equilibrium (HWE) p.1026, and quality measures for imputed SNPs (r2$0.3, or proper info $0.4, for cohorts imputing their data with MACH and IMPUTE, respectively). pean and non-European populations separately, using the same methods as described above. Presence of heterogeneity in the meta-analysis was assessed by the I2 statistic and Q-test [54]. Since cohorts measured adiponectin concentrations using either RIA or ELISA methods, we also performed a GWA meta-analysis stratified by the method of measurement to test whether this contributed to heterogeneity Follow-up phase. The follow-up phase comprised two stages; in-silico follow-up and de-novo follow-up. Statistical Analysis Genome-wide association studies. All cohorts indepen- dently tested for the additive genetic association of common (MAF.1%) genotyped and imputed SNPs with natural log transformed adiponectin levels, while adjusting for age, sex, body mass index (BMI), principal components of population stratification and study site (where appropriate), and for family structure in cohorts with family members [49,50,52]. The analyses were performed for men and women combined, as well as for men and women separately. The Cardiovascular Health Study cohort (CHS) also provided GWA results for high molecular weight (HMW) adiponectin using the same methods as described above. Meta-analysis of GWAS. The meta-analysis was performed by two analysts independently each using different methods; inverse variance-weighted methods using both fixed and random effect models available through either the METAL (http://www. sph.umich.edu/csg/abecasis/metal/) or GWAMA version 2.0.5 (http://www.well.ox.ac.uk/gwama/) software packages [53]. Summary statistics were crosschecked to ensure consistency of results. Prior to the meta-analysis, study-specific summary statistics were corrected using genomic control (lambda range = 0.99–1.25) and the overall meta-analytic results were additionally corrected for genomic control (lambda = 1.06). To examine whether associations with adiponectin were sex-specific, we performed meta-analyses for men and women separately. A p-value threshold of 561028 was considered to be genome-wide significant. Ethnicity-specific meta-analyses were performed for white Euro- Genotyping and Imputation Conversely, random effects meta-analysis assumes that each population has a different underlying allelic effect, however, populations from the same ethnic group would be more homogeneous than those that are more distantly related. To address this challenge we accounted for the expected similarity in allelic effects between the most closely related populations by means of a Bayesian partition model. For each variant, allelic effects and corresponding standard errors are estimated within each population under the assumption of an additive model. Populations are then clustered according to their similarity in terms of relatedness as measured by the mean allele frequency difference at 10,000 independent SNPs, and to their allelic effects at the variant. If all populations are assigned to the same cluster, this is equivalent to a fixed allelic effect across all populations (i.e. no trans-ethnic heterogeneity). The posterior distribution of the allelic effect in each population under the Bayesian partition model is approximated by means of a Monte-Carlo Markov chain algorithm. Evidence in favor of association of the trait with the variant was assessed by means of a Bayes’ factor (BF). A log10 BF of 6 or higher is considered a relatively conservative threshold for genome-wide significance. We also performed meta-analysis by using both random and fixed effects models including all ethnicities. Those loci that achieved both a BF.6 in MANTRA and a P-value less than 561027 in multiethnic analysis are presented in Table 2. Multi-ethnic meta-analysis. In order to perform a meta- analysis of GWAS data from cohorts of different ethnic backgrounds, we utilized the novel MANTRA (Meta-ANalysis of Trans-ethnic Association studies) software [24]. This method combines GWAS from different ethnic groups by taking advantage of the expected similarity in allelic effects between the most closely related populations. Fixed-effects meta-analysis assumes the allelic effect to be the same in all populations, and cannot account for heterogeneity between ethnic groups. Conversely, random effects meta-analysis assumes that each population has a different underlying allelic effect, however, populations from the same ethnic group would be more homogeneous than those that are more distantly related. To address this challenge we accounted for the expected similarity in allelic effects between the most closely related populations by means of a Bayesian partition model. For each variant, allelic effects and corresponding standard errors are estimated within each population under the assumption of an additive model. Association of Genome-Wide Significant SNPs with Gene Expression (Stage 2) In order to identify cis-expression quantitative trait loci (cis- eQTLs) and test whether mRNA levels of candidate genes arising from our GWAS were associated with adiponectin levels, we used PLoS Genetics \| www.plosgenetics.org March 2012 \| Volume 8 \| Issue 3 \| e1002607 12 Adiponectin: Multiethnic Genome-Wide Meta-Analysis Figure 3. Flow chart of study design. doi:10.1371/journal.pgen.1002607.g003 Figure 3. Flow chart of study design. doi:10.1371/journal.pgen.1002607.g003 expression profiles in human adipocytes from the Multiple Tissue Human Expression Resource (MuTHER) Consortium, (856 female twins from the UK) [25]. mRNA expression profiles from subcutaneous fat and genome-wide genotypes were available for 776 individuals and circulating adiponectin levels for 436 of these women. We note that while adiponectin levels were measured at an earlier time point than the fat biopsies, the BMI at time of adipose expression measurement and time of adiponectin measurement was highly correlated (r2 = 0.9). SNPs (which were estimated to be equivalent to 96 independent statistical tests due to linkage disequilibrium [LD]) [26] were tested for their association, we employed a Bonferroni-corrected threshold of a = 0.0005 (where 0.0005 = 0.05/96) to define the threshold of association for any individual SNP association with T2D and related traits. While any individual SNP may demonstrate a relationship with T2D or related traits, it can be more informative to test whether a multi-SNP genotypic risk score is associated with the outcome of interest. In the absence of pleiotropic effects arising from loci other than ADIPOQ, such a multi-SNP genotypic risk score would enable testing of whether adiponectin levels are causally related to risk of T2D or metabolic traits through a Mendelian randomization framework. Since most of the SNPs that we identified to be genome-wide significant for adiponectin levels were not in the ADIPOQ locus, the presence of such pleiotropy precluded a formal Mendelian randomization study. To create a multi-SNP genotypic risk score we implemented a novel method that approximates the average effect of adiponectin decreasing alleles on T2D or related traits. Further, this method allows the use of consortium-level meta-analytic results for a set of SNPs, rather than requiring the re-analysis of individual-level data in each cohort, thereby providing more accurate effects of each allele (due to the larger sample size in the consortium-level meta-analysis). Association of Genome-Wide Significant SNPs with Gene Expression (Stage 2) The weighted sum of the individual SNP coefficients leads not only to an estimate of the average combined allelic effect, but also to an approximate estimate of the explained variance (when scaled by the inverse of the total meta-analysis sample size) from a multivariate regression model containing these SNPs. cis-eQTLs were defined as associations between SNPs and a transcript within 1 Mb of the identified SNP. To correct for multiple testing, we used QVALUE software [56], and estimated that a genome-wide false discovery rate of 1% corresponds to a p- value threshold of 5.0661025 (this conservative threshold accounts for all multiple arising from the use of the array, rather than multiple testing arising from assessing only transcripts in the genome-wide significant regions). To test whether mRNA levels of candidate genes identified in the GWAS meta-analysis are associated with circulating adiponectin levels, we applied a Bonferoni corrected threshold of p,361024 (where 361024 = 0.05/133 and 133 was the number of transcripts tested at the candidate loci). PLoS Genetics \| www.plosgenetics.org Table S1 Cohort characteristics. (XLSX) Table S2 Comparing the Genome-Wide Significant SNPS from fixed effect model with random effect model. SNP with I2 less than 0.5 are listed in bold, EA: Effect Allele, NEA: Non-Effect Allele. (PDF) ^a% Pm i~1 wi^bisi{2 Pm i~1 w2 i s{2 i (PDF) Table S3 Association Results of SNPs achieving p#561026 in the Discovery phase in European Populations (Sex-Combined Analysis). Denotes SNPs typed in the de-novo follow-up phase. (PDF) with a standard error estimate of se(^a)% ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1 Pm i~1 w2 i s{2 i s se(^a)% ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1 Pm i~1 w2 i s{2 i s Table S4 Genome-Wide Significant SNPs (p,561028) Associ- ated with Adiponectin Levels in Non-Europeans Populations. EA: Effect Allele, NEA: Non-Effect Allele, EA-Freq: Frequency of Effect Allele. (PDF) Therefore, under the assumption of uncorrelated SNPs, their joint effect can be estimated in external data by a weighted mean of the individual SNP effects, weighted by the estimates from the discovery data. All these quantities can be obtained from meta- analysis or summary data, so that individual-level data are not required to obtain these results. Table S5 SNPs associated with adiponectin at genome-wide significant levels (p,561028) using the fixed-effect model in women only in European populations (including Discovery and Follow-Up phases). (PDF) To implement this method, we first selected LD-independent adiponectin associated alleles by LD pruning the set of genome- wide significant adiponectin SNPs from the discovery phase with an LD threshold of r2#0.05 in the HapMap CEU population, yielding 20 independent LD blocks from the 196 SNPs in Table S2. (We also applied the method using an LD threshold of r2#0.01 and found no relevant change in results). Since many SNPs from the same independent blocks were associated with adiponectin, we selected the SNP from the LD block that explained the most variance in adiponectin levels. Next, we approximated the effect of the multi-SNP genetic risk score using b and its standard error as derived from the consortium-level meta-analysis in DIAGRAM+, MAGIC, GLGC, GIANT and Body Fat GWAS consortium. Table S6 SNPs associated with adiponectin at genome-wide significant levels (p,561028) using fixed-effect models in men only in Euopean populations. (PDF) Table S7 Association results of nominally significant SNPs with Type 2 Diabetes in the DIAGRAM+ Consortium. EA: Effect Allele, NEA: Non-Effect Allele. B) Association results of nominally significant SNPs with diabetes-related traits in the MAGIC Consortium. Association of Genome-Wide Significant SNPs with T2D and Metabolic Traits (Stage 3) The DIAGRAM+ (effective n = 22,044) [19], MAGIC (n = up to 46,186) [20], GLGC (n = up to 97,021) [22], GIANT (n = up to 121,335) [21], and Body Fat GWAS (n = up to 36,625) consortia provided summary statistics for the association of each SNP that was genome-wide significant in the discovery phase. Since 196 PLoS Genetics \| www.plosgenetics.org March 2012 \| Volume 8 \| Issue 3 \| e1002607 13 Adiponectin: Multiethnic Genome-Wide Meta-Analysis Specifically, suppose m SNPs have shown association in the discovery phase, and effects are denoted wi. However, suppose that the goal of interest is to estimate the joint effect of these SNPs on an outcome of interest, y. Let j index the individuals in the outcome of interest dataset and let 2½ln L(M1){ ln L(M0)%nR2 If the SNPs are uncorrelated, and if the total percentage of variance explained is small, then the change in log likelihood can be approximated by C{ X m i~1 bi{^bi 2 2s2 i where bi now refers to the effect of SNP i in the outcome data, ^bi is the outcome data estimate, and si is the associated standard error estimate. Assuming that this log likelihood difference approximation is maximized with an appropriate value of C, then it can be shown that a can be estimated by: where bi now refers to the effect of SNP i in the outcome data, ^bi is the outcome data estimate, and si is the associated standard error estimate. Assuming that this log likelihood difference approximation is maximized with an appropriate value of C, then it can be shown that a can be estimated by: Table S1 Cohort characteristics. (XLSX) Supporting Information Figure S1 The comparison between two independent meta- analyses performed in different centers for quality control purposes. The 2log10 p-value of all SNPS with MAF$0.01 in the first analysis are plotted against the 2log10 p-value from the second analysis. (TIF) sj~ Xm i~1 wixij Figure S2 The Manhattan plots of sex-stratified meta-analyses in the discovery phase in the European population. The meta- analysis shown in panel a) is stratified for women and that in panel b) is stratified for men. Manhattan plots demonstrate 2Log 10(p- value) measures for association between single nucleotide poly- morphisms (SNPs) and chromosomal position. The SNPs that achieved genome-wide significance are highlighted in green in the plots. The red ovals identify loci found only in women. (TIF) be a risk score based on the discovery data SNPs, and their associated parameter estimates wi. Therefore, the desired goal is to estimate the parameter in the following equation: yj~y0zasjzej in the outcome of interest dataset. The proportion of variance in y explained by the previous equation, (i.e. the R2) attributable to the risk score can be estimated. Standard linear model theory shows that the change in log likelihood is proportional to the R2, Figure S3 Association Results Near Peaks for Sex-specific Analysis of Adiponectin. SNPs in regions near peak associations are shown for a) chromosome 8 female, b) chromosome 8 males, c) chromosome 12 females and d) chromosome 12 males. Purple diamonds indicate the top SNPs, which have the strongest evidence of association in women. Each circle shows a SNP with a color scale proportional to the r2 value for that SNP and the top SNP from HapMap CEU. Blue lines show the estimated recombination rates from HapMap. The bottom panels illustrate the relative position of each gene in the locus. Consortia Authors’ list: DIAGRAM+: 15. Ontario Institute for Cancer Research, 101 College Street, Suite 800, Toronto, Ontario M5G 0A3, Canada Benjamin F Voight1,2,3, Laura J Scott4, Valgerdur Steinthorsdottir5, Andrew P Morris6, Christian Dina7,8, Ryan P Welch9, Eleftheria Zeggini6,10, Cornelia Huth11,12, Yurii S Aulchenko13, Gudmar Thorleifs- son5, Laura J McCulloch14, Teresa Ferreira6, Harald Grallert11,12, Najaf Amin13, Guanming Wu15, Cristen J Willer4, Soumya Raychaudhuri1,2,16, Steve A McCarroll1,17, Claudia Langenberg18, Oliver M Hofmann19, Jose´e Dupuis20,21, Lu Qi22–24, Ayellet V Segre`1,2,17, Mandy van Hoek25, Pau Navarro26, Kristin Ardlie1, Beverley Balkau27,28, Rafn Benediktsson29,30, Amanda J Bennett14, Roza Blagieva31, Eric Boerwinkle32, Lori L Bonnycastle33, Kristina Bengtsson Bostro¨m34, Bert Bravenboer35, Suzan- nah Bumpstead10, Noe¨l P Burtt1, Guillaume Charpentier36, Peter S Chines33, Marilyn Cornelis24, David J Couper37, Gabe Crawford1, Alex SF Doney38,39, Katherine S Elliott6, Amanda L Elliott1,17,40, Michael R Erdos33, Caroline S Fox21,41, Christopher S Franklin42, Martha Ganser4, Christian Gieger11, Niels Grarup43, Todd Green1,2, Simon Griffin18, Christopher J Groves14, Candace Guiducci1, Samy Hadjadj44, Neelam Hassanali14, Christian Herder45, Bo Isomaa46,47, Anne U Jackson4, Paul RV Johnson48, Torben Jørgensen49,50, Wen HL Kao51,52, Norman Klopp11, Augustine Kong5, Peter Kraft22,23, Johanna Kuusisto53, Torsten Lauritzen54, Man Li51, Aloysius Lieverse55, Cecilia M Lindgren6, Valeriya Lyssenko56, Michel Marre57,58, Thomas Meitinger59,60, Kristian Mid- thjell61, Mario A Morken33, Narisu Narisu33, Peter Nilsson56, Katharine R Owen14, Felicity Payne10, John RB Perry62,63, Ann-Kristin Petersen11, Carl Platou61, Christine Proenc¸a7, Inga Prokopenko6,14, Wolfgang Rathmann64, N William Rayner6,14, Neil R Robertson6,14, Ghislain Rocheleau65–67, Michael Roden45,68, Michael J Sampson69, Richa Saxena1,2,40, Beverley M Shields62,63, Peter Shrader3,70, Gunnar Sigurds- son29,30, Thomas Sparsø43, Klaus Strassburger64, Heather M Stringham4, Qi Sun22,23, Amy J Swift33, Barbara Thorand11, Jean Tichet71, Tiinamaija Tuomi46,72, Rob M van Dam24, Timon W van Haeften73, Thijs van Herpt25,55, Jana V van Vliet-Ostaptchouk74, G Bragi Walters5, Michael N Weedon62,63, Cisca Wijmenga75, Jacqueline Witteman13, Richard N Bergman76, Stephane Cauchi7, Francis S Collins77, Anna L Gloyn14, Ulf Gyllensten78, Torben Hansen43,79, Winston A Hide19, Graham A Hitman80, Albert Hofman13, David J Hunter22,23, Kristian Hveem61,81, Markku Laakso53, Karen L Mohlke82, Andrew D Morris38,39, Colin NA Palmer38,39, Peter P Pramstaller83, Igor Rudan42,84,85, Eric Sijbrands25, Lincoln D Stein15, Jaakko Tuomilehto86, Andre Uitterlinden25, Mark Walker87, Nicholas J Wareham18, Richard M Watanabe76,88, Goncalo R Abecasis4, Bernhard O Boehm31, Harry Campbell42, Mark J Daly1,2, Andrew T Hattersley62,63, Frank B Hu22–24, James B Meigs3,70, James S Pankow89, Oluf Pedersen43,90,91, H.-Erich Wichmann11,12,92, Ineˆs Bar- roso10, Jose C Florez1,2,3,93, Timothy M Frayling62,63, Leif Groop56,72, Rob Sladek65–67, Unnur Thorsteinsdottir5,94, James F Wilson42, Thomas Illig11, 16. Table S1 Cohort characteristics. (XLSX) Fasting glucose and 2 h glucose in mmol/L; Insulin in pmol/L, EA: Effect Allele, NEA: Non-Effect Allele. C) Association results of nominally significant SNPs with diabetes- PLoS Genetics \| www.plosgenetics.org March 2012 \| Volume 8 \| Issue 3 \| e1002607 14 Adiponectin: Multiethnic Genome-Wide Meta-Analysis Philippe Froguel7,95, Cornelia M van Duijn13, Kari Stefansson5,94, David Altshuler1,2,3,17,40,93, Michael Boehnke4, Mark I McCarthy6,14,96. Philippe Froguel7,95, Cornelia M van Duijn13, Kari Stefansson5,94, David Altshuler1,2,3,17,40,93, Michael Boehnke4, Mark I McCarthy6,14,96. related traits in the GIANT and Body fat GWAS consortia. The beta expressed in inverse normally transformed BMI units (i.e. interpretable as SD or Z-score), shows the change in BMI per additional effect allele.,Results that are statistically significant, accounting for the number of independent SNPs, are highlighted in bold., EA: Effect Allele, NEA: Non-Effect Allele, EA-Freq: Frequency of Effect Allele. D) Association results of nominally significant SNPs with lipid traits in the GLGC Consortium. For these traits the effect size is in SD units, based on standard error- weighted meta-analysis. Results that are statistically significant, accounting for the number of independent SNPs are highlighted in bold., EA: Effect Allele, NEA: Non-Effect Allele, EA-Freq: Frequency of Effect Allele. (PDF) y 1. Broad Institute of Harvard and Massachusetts Institute of Technology (MIT), Cambridge, Massachusetts 02142, USA 2. Center for Human Genetic Research, Massachusetts General Hospital, 185 Cambridge Street, Boston, Massachusetts 02114, USA 3. Department of Medicine, Harvard Medical School, Boston, Massachusetts 02115, USA 4. Department of Biostatistics, University of Michigan, Ann Arbor, Michigan 48109-2029, USA 5. deCODE Genetics, 101 Reykjavik, Iceland 6. Wellcome Trust Centre for Human Genetics, University of Oxford, Oxford, OX3 7BN, UK 7. CNRS-UMR-8090, Institute of Biology and Lille 2 University, Pasteur Institute, F-59019 Lille, France 8. INSERM UMR915 CNRS ERL3147 F-44007 Nantes, France 9. Bioinformatics Program, University of Michigan, Ann Arbor MI USA 48109 Text S1 Supplemental data include description of study cohorts and funding. (DOCX) Text S1 Supplemental data include description of study cohorts and funding. (DOCX) 10. Wellcome Trust Sanger Institute, Hinxton, CB10 1HH, UK 11. Institute of Epidemiology, Helmholtz Zentrum Muenchen, 85764 Neuherberg, Germany 12. Institute of Medical Informatics, Biometry and Epidemiology, Ludwig-Maximilians-Universita¨t, 81377 Munich, Germany Acknowledgments g y 13. Department of Epidemiology, Erasmus University Medical Center, P O B CA R d h h l d We thank all study participants, volunteers, and study personnel that made this consortium possible. We would also like to thank Ms. Renee Atallah for her efforts with the writing and correction of the manuscript. 13. Department of Epidemiology, Erasmus University P.O. Box 2040, 3000 CA Rotterdam, The Netherlands. 13. Department of Epidemiology, Erasmus University P.O. Box 2040, 3000 CA Rotterdam, The Netherlands. 14. Oxford Centre for Diabetes, Endocrinology and Metabolism, University of Oxford, OX3 7LJ, UK Adiponectin: Multiethnic Genome-Wide Meta-Analysis Faculty of Health Science, University of Copenhagen, 2200 Copenhagen, Denmark 84. Croatian Centre for Global Health, Faculty of Medicine, University of Split, Soltanska 2, 21000 Split, Croatia 84. Croatian Centre for Global Health, F of Split, Soltanska 2, 21000 Split, Croatia 51. Department of Epidemiology, Johns Hopkins University, Baltimore, Maryland 21287, USA 85. Institute for Clinical Medical Research, University Hospital ‘‘Sestre Milosrdnice’’, Vinogradska 29, 10000 Zagreb, Croatia 52. Department of Medicine, and Welch Center for Prevention, Epidemiology, and Clinical Research, Johns Hopkins University, Balti- more, Maryland 21287, USA 86. Department of Chronic Disease Prevention, National Institute for Health and Welfare, Helsinki FIN-00300, Finland, 87. Diabetes Research Group, Institute of Cellular Medicine, Newcastle University, Framlington Place, Newcastle upon Tyne NE2 4HH, UK 53. Department of Medicine, University of Kuopio and Kuopio University Hospital, FIN-70211 Kuopio, Finland y g p y 88. Department of Preventitive Medicine, Keck Medical School, University of Southern California, Los Angeles, CA, 90089-9001, USA 54. Department of General Medical Practice, University of Aarhus, DK- 8000 Aarhus, Denmark 89. Division of Epidemiology and Community Health, University of Minnesota, Minneapolis, Minnesota 55454, USA 55. Department of Internal Medicine, Maxima MC, PO-Box 90052, 5600 PD Eindhoven, The Netherlands p 90. Department of Biomedical Science, Panum, Faculty of Health Science, University of Copenhagen, 2200 Copenhagen, Denmark 56. Department of Clinical Sciences, Diabetes and Endocrinology Research Unit, University Hospital Malmo¨, Lund University, 205 02 Malmo¨, Sweden 91. Faculty of Health Science, University of Aarhus, DK–8000 Aarhus, Denmark 57. Department of Endocrinology, Diabetology and Nutrition, Bichat- Claude Bernard University Hospital, Assistance Publique des Hoˆpitaux de Paris, 75870 Paris Cedex 18, France 92. Klinikum Grosshadern, 81377 Munich, Germany 93. Diabetes Unit, Massachusetts General Hospital, Boston, Massachu- setts 02144, USA 58. INSERM U695, Universite´ Paris 7, 75018 Paris, France 94. Faculty of Medicine, University of Iceland, 101 Reykjavı´k, Iceland 95. Genomic Medicine, Imperial College London, Hammersmith Hospital, W12 0NN, London, UK 59. Institute of Human Genetics, Helmholtz Zentrum Muenchen, 85764 Neuherberg, Germany 60. Institute of Human Genetics, Klinikum rechts der Isar, Technische Universita¨t Mu¨nchen, 81675 Muenchen, Germany p 96. Oxford National Institute for Health Research Biomedical Research Centre, Churchill Hospital, Old Road Headington, Oxford, OX3 7LJ, UK 96. Oxford National Institute for Health Research Biomedical Research Centre, Churchill Hospital, Old Road Headington, Oxford, OX3 7LJ, UK MAGIC Consortium: 61. Consortia Authors’ list: DIAGRAM+: Division of Rheumatology, Immunology and Allergy, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts 02115, USA 17. Department of Molecular Biology, Harvard Medical School, Boston, Massachusetts 02115, USA 18. MRC Epidemiology Unit, Institute of Metabolic Science, Adden- brooke’s Hospital, Cambridge CB2 0QQ, UK 19. Department of Biostatistics, Harvard School of Public Health, Boston, Massachusetts 02115, USA 20. Department of Biostatistics, Boston University School of Public Health, Boston, Massachusetts 02118, USA 21. National Heart, Lung, and Blood Institute’s Framingham Heart Study, Framingham, Massachusetts 01702, USA 21. National Heart, Lung, and Blood Institute’s Study, Framingham, Massachusetts 01702, USA y g 22. Department of Nutrition, Harvard School of Public Health, 665 Huntington Ave, Boston, MA 02115, USA 23. Department of Epidemiology, Harvard School of Public Health, 665 Huntington Ave, Boston, MA 02115, USA 24. Channing Laboratory, Dept. of Medicine, Brigham and Women’s Hospital and Harvard Medical School, 181 Longwood Ave, Boston, MA 02115, USA 25. Department of Internal Medicine, Erasmus University Medical Centre, PO-Box 2040, 3000 CA Rotterdam, The Netherlands 26. MRC Human Genetics Unit, Institute of Genetics and Molecular 26. MRC Human Genetics Unit, Institute of Genetics and Molecular Medicine, Western General Hospital, Edinburgh, EH4 2XU, UK Medicine, Western General Hospital, Edinburgh, EH4 2XU, UK 27. INSERM U780, F-94807 Villejuif. France 28. University Paris-Sud, F-91405 Orsay, France 29. Landspitali University Hospital, 101 Reykjavik, Iceland 30. Icelandic Heart Association, 201 Kopavogur, Iceland 31. Division of Endocrinology, Diabetes and Metabolism, Ulm University, 89081 Ulm, Germany 32. The Human Genetics Center and Institute of Molecular Medicine, University of Texas Health Science Center, Houston, Texas 77030, USA 32. The Human Genetics Center and Institute of Molecular Medicine, U i it f T H lth S i C t H t T 77030 USA 33. National Human Genome Research Institute, National Institute of Health, Bethesda, Maryland 20892, USA y 34. R&D Centre, Skaraborg Primary Care, 541 30 Sko¨vde, Sweden 35. Department of Internal Medicine, Catharina Hospital, PO-Box 1350, 5602 ZA Eindhoven, The Netherlands 36. Endocrinology-Diabetology Unit, Corbeil-Essonnes Hospital, F- 91100 Corbeil-Essonnes, France 37. Department of Biostatistics and Collaborative Studies Coordinating Center, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, 27599, USA 38. Diabetes Research Centre, Biomedical Research Institute, University of Dundee, Ninewells Hospital, Dundee DD1 9SY, UK PLoS Genetics \| www.plosgenetics.org March 2012 \| Volume 8 \| Issue 3 \| e1002607 15 Adiponectin: Multiethnic Genome-Wide Meta-Analysis Adiponectin: Multiethnic Genome-Wide Meta-Analysis 39. Pharmacogenomics Centre, Biomedical Research Institute, Univer- sity of Dundee, Ninewells Hospital, Dundee DD1 9SY, UK 71. Institut interre´gional pour la Sante´ (IRSA), F-37521 La Riche, France 72. Department of Medicine, Helsinki University Hospital, University of Helsinki, FIN-00290 Helsinki, Finland 40. Department of Genetics, Harvard Medical School, Boston, Massachusetts 02115, USA 73. Department of Internal Medicine, University Medical Center Utrecht, 3584 CG Utrecht,The Netherlands 41. Division of Endocrinology, Diabetes, and Hypertension, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts 02115, USA 74. Molecular Genetics, Medical Biology Section, Department of Pathology and Medical Biology, University Medical Center Groningen and University of Groningen, 9700 RB Groningen, The Netherlands 42. Centre for Population Health Sciences, University of Edinburgh, Teviot Place, Edinburgh, EH8 9AG, UK 75. Department of Genetics, University Medical Center Groningen an University of Groningen, 9713 EX Groningen, The Netherlands 75. Department of Genetics, University Medical Center Groningen and University of Groningen, 9713 EX Groningen, The Netherlands 76. Department of Physiology and Biophysics, University of Southern California School of Medicine, Los Angeles, California 90033, USA 43. Hagedorn Research Institute, DK-2820 Gentofte, Denmark 44. Centre Hospitalier Universitaire de Poitiers, Endocrinologie Diabetologie, CIC INSERM 0801, INSERM U927, Universite´ de Poitiers, UFR, Me´decine Pharmacie, 86021 Poitiers Cedex, France y g g 76. Department of Physiology and Biophysics, University of Southern California School of Medicine, Los Angeles, California 90033, USA 7. National Institute of Health, Bethesda, Maryland 20892, U 45. Institute for Clinical Diabetology, German Diabetes Center, Leibniz Center for Diabetes Research at Heinrich Heine University Du¨sseldorf, 40225 Du¨sseldorf, Germany 78. Department of Genetics and Pathology, Rudbeck Laboratory, Uppsala University, S-751 85 Uppsala, Sweden. 46. Folkha¨lsan Research Center, FIN-00014 Helsinki, Finland y , , 80. Centre for Diabetes, Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London E1 2AT, UK 47. Malmska Municipal Health Center and Hospital, 68601 Jakobstad, Finland 81. Department of Medicine, The Hospital of Levanger, N-7600 Levanger, Norway 48. Diabetes Research and Wellness Foundation Human Islet Isolation Facility and Oxford Islet Transplant Programme, University of Oxford, Old Road, Headington, Oxford, OX3 7LJ, UK 82. Department of Genetics, University of North Carolina, Chapel Hill, North Carolina 27599, USA 49. Research Centre for Prevention and Health, Glostrup University Hospital, DK-2600 Glostrup, Denmark 83. Institute of Genetic Medicine, European Academy Bozen/Bolzano (EURAC), Viale Druso 1, 39100 Bolzano, Italy 50. Adiponectin: Multiethnic Genome-Wide Meta-Analysis Nord-Trøndelag Health Study (HUNT) Research Center, Depart- ment of Community Medicine and General Practice, Norwegian University of Science and Technology, NO-7491 Trondheim, Norway MAGIC Consortium: Jose´e Dupuis1,2,178, Claudia Langenberg3,178, Inga Prokopenko4,5,178, Richa Saxena6,7,178, Nicole Soranzo8,9,178, Anne U Jackson10, Eleanor Wheeler11, Nicole LGlazer12, Nabila Bouatia-Naji13, Anna LGloyn4, Cecilia MLindgren4,5, Reedik Ma¨gi4,5, Andrew P Morris5, Joshua Randall5, Toby Johnson14–16, Paul Elliott17,176, Denis Rybin18, Gudmar Thorleifsson19, Valgerdur Steinthorsdottir19, Peter Henneman20, Harald Grallert21, Abbas Dehghan22, Jouke Jan Hottenga23, Christopher SFrank- lin24, Pau Navarro25, Kijoung Song26, Anuj Goel5,27, John R B Perry28, Josephine MEgan29, Taina Lajunen30, Niels Grarup31, Thomas Sparsø31, Alex Doney32, Benjamin F Voight6,7, Heather MStringham10, Man Li33, Stavroula Kanoni34, Peter Shrader35, Christine Cavalcanti-Proenc¸a13, Meena Kumari36, Lu Qi37, Nicholas J Timpson38, Christian Gieger21, Carina Zabena39, Ghislain Rocheleau40,41, Erik Ingelsson42,43, Ping An44, Jeffrey O’Connell45, Jian’an Luan3, Amanda Elliott6,7, Steven A McCar- roll6,7, Felicity Payne11, Rosa Maria Roccasecca11, Franc¸ois Pattou46, Praveen Sethupathy47, Kristin Ardlie48, Yavuz Ariyurek49, Beverley Balkau50, Philip Barter51, John P Beilby52,53, Yoav Ben-Shlomo54, Rafn Benediktsson55,56, Amanda J Bennett4, Sven Bergmann14,16, Murielle Bochud15, Eric Boerwinkle57, Ame´lie Bonnefond13, Lori LBonnycastle47, Knut Borch-Johnsen58,59, Yvonne Bo¨ttcher60, Eric Brunner36, Suzannah J Bumpstead8, Guillaume Charpentier61, Yii-Der Ida Chen62, Peter Chines47, Robert Clarke63, Lachlan J MCoin17, Matthew NCooper64, Marilyn Cornelis37, Gabe Crawford6, Laura Crisponi65, Ian NMDay38, y gy y 62. Genetics of Complex Traits, Institute of Biomedical and Clinical S i P i l M di l S h l U i it f E t M d l R d y gy y 62. Genetics of Complex Traits, Institute of Biomedical and Clinical Science, Peninsula Medical School, University of Exeter, Magdalen Road, Science, Peninsula Medical School, University of Exeter, Magdalen Road, Exeter EX1 2LU, UK 63. Diabetes Genetics, Institute of Biomedical and Clinical Science, Exeter EX1 2LU, UK 63. Diabetes Genetics, Institute of Biomedical and Clinical Science, P i l M di l S h l U i i f E B k R d E Exeter EX1 2LU, UK 63. Diabetes Genetics, Institute of Biomedical and Clinical Science, Peninsula Medical School, University of Exeter, Barrack Road, Exeter EX2 5DW, UK 64. Institute of Biometrics and Epidemiology, German Diabetes Center, Peninsula Medical School, University of Exeter, Barrack Road, Exeter EX2 5DW, UK 64. Institute of Biometrics and Epidemiology, German Diabetes Center, EX2 5DW, UK 64. Institute of Biometrics and Epidemiology, German Diabetes Center, , 64. Adiponectin: Multiethnic Genome-Wide Meta-Analysis Eco J Cde Geus23, Jerome Delplanque13, Christian Dina13, Michael R Erdos47, Annette CFedson64,66, Antje Fischer-Rosinsky67,68, Nita GFor- ouhi3, Caroline SFox2,69, Rune Frants70, Maria Grazia Franzosi71, Pilar Galan72, Mark OGoodarzi62, Ju¨rgen Graessler73, Christopher J Groves4, Scott Grundy74, Rhian Gwilliam8, Ulf Gyllensten75, Samy Hadjadj76, Go¨ran Hallmans77, Naomi Hammond8, Xijing Han10, Anna-Liisa Hartikainen78, Neelam Hassanali4, Caroline Hayward25, Simon CHeath79, Serge Hercberg80, Christian Herder81, Andrew A Hicks82, David R Hillman66,83, Aroon DHingorani36, Albert Hofman22, Jennie Hui52,84, Joe Hung85,86, Bo Isomaa87,88, Paul R V Johnson4,89, Torben Jørgensen90,91, Antti Jula92, Marika Kaakinen93, Jaakko Kaprio94–96, Y Antero Kesa- niemi97, Mika Kivimaki36, Beatrice Knight98, Seppo Koskinen99, Peter Kovacs100, Kirsten Ohm Kyvik101, GMark Lathrop79, Debbie A Lawlor38, Olivier Le Bacquer13, Ce´cile Lecoeur13, Yun Li10, Valeriya Lyssenko102, Robert Mahley103, Massimo Mangino9, Alisa KManning1, Marı´a Teresa Martı´nez-Larrad39, Jarred B McAteer6,104,105, Laura J McCulloch4, Ruth McPherson106, Christa Meisinger21, David Melzer28, David Meyre13, Braxton DMitchell45, Mario A Morken47, Sutapa Mukherjee66,83, Silvia Naitza65, Narisu Narisu47, Matthew J Neville4,107, Ben A Oostra108, Marco Orru`65, Ruth Pakyz45, Colin NA Palmer109, Giuseppe Paolisso110, Cristian Pattaro82, Daniel Pearson47, John F Peden5,27, Nancy LPedersen42, Markus Perola96,111,112, Andreas F H Pfeiffer67,68, Irene Pichler82, Ozren Polasek113, Danielle Posthuma23,114, Simon CPotter8, Anneli Pouta115, Michael A Province44, Bruce MPsaty116,117, Wolfgang Rathmann118, Nigel WRayner4,5, Kenneth Rice119, Samuli Ripatti96,111, Fernando Rivade- neira22,120, Michael Roden81,121, Olov Rolandsson122, Annelli Sand- baek123, Manjinder Sandhu3,124, Serena Sanna65, Avan Aihie Sayer125, Paul Scheet126, Laura J Scott10, Udo Seedorf127, Stephen J Sharp3, Beverley Shields98, Gunnar Sigurðsson55,56, Eric J GSijbrands22,120, Angela Silveira128, Laila Simpson64,66, Andrew Singleton129, Nicholas LSmith130,131, Ulla Sovio17, Amy Swift47, Holly Syddall125, Ann-Christine Syva¨nen132, Toshiko Tanaka133,134, Barbara Thorand21, Jean Tichet135, Anke To¨njes60,136, Tiinamaija Tuomi87,137, Andre´ GUitterlinden22,120, Ko Willems van Dijk70,138, Mandy van Hoek120, Dhiraj Varma8, Sophie Visvikis-Siest139, Veronique Vitart25, Nicole Vogelzangs140, Ge´rard Waeber141, Peter J Wagner96,111, Andrew Walley142, GBragi Walters19, Kim LWard64,66, Hugh Watkins5,27, Michael NWeedon28, Sarah H Wild24, Gonneke Willemsen23, Jaqueline CMWitteman22, John WGYar- nell143, Eleftheria Zeggini5,8, Diana Zelenika79, Bjo¨rn Zethelius43,144, Guangju Zhai9, Jing Hua Zhao3, MCarola Zillikens120, DIAGRAMCon- sortium145, GIANTConsortium145, Global BPgen Consortium145,Ingrid B Borecki44, Ruth J F Loos3, Pierre Meneton80, Patrik KEMagnusson42, David MNathan104,105, Gordon H Williams69,105, Andrew THattersley98, Kaisa Silander96,111, Veikko Salomaa146, George Davey Smith38, Stefan R Bornstein73, Peter Schwarz73, Joachim Spranger67,68, Fredrik Karpe4,107, Alan R Shuldiner45, Cyrus Cooper125, George V Dedoussis34, Manuel Serrano-Rı´os39, Andrew DMorris109, Lars Lind132, Lyle J Palmer64,66,84, Frank B Hu147,148, Paul WFranks149, Shah Ebrahim150, Michael Marmot36, WH Linda Kao33,151,152, James SPankow153, Michael J Sampson154, Johanna Kuusisto155, Markku Laakso155, Torben Han- sen31,156, Oluf Pedersen31,59,157, Peter Paul Pramstaller82,158,159, H Erich Wichmann21,160,161, Thomas Illig21, Igor Rudan24,162,163, Alan F Wright25, Michael Stumvoll60, Harry Campbell24, James F Wilson24, Anders Hamsten on behalf of Procardis Consortium128, Richard NBergman164, Thomas A Buchanan164,165, Francis SCollins47, Karen LMohlke166, Jaakko Tuomilehto94,167, 168, Timo TValle167, David Altshuler6,7,104,105, Jerome I Rotter62, David SSiscovick169, Brenda WJ H Penninx140, Dorret I Boomsma23, Panos Deloukas8, Timothy DSpector8,9, Timothy MFray- ling28, Luigi Ferrucci170, Augustine Kong19, Unnur Thorsteinsdottir19,171, Kari Stefansson19,171, Cornelia Mvan Duijn22, Yurii SAulchenko22, Antonio Cao65, Angelo Scuteri172,177, David Schlessinger47, Manuela Uda65, Aimo Ruokonen173, Marjo-Riitta Jarvelin17,93, 174, Dawn MWater- worth26, Peter Vollenweider141, Leena Peltonen8,48,96,111,112, Vincent Mooser26, Goncalo R Abecasis10, Nicholas J Wareham3, Robert Sladek40,41, Philippe Froguel13,142, Richard MWatanabe164,175, James B Meigs35,105, Leif Groop102, Michael Boehnke10, Mark I McCarthy4,5,107, Jose CFlorez6,7,104,105 & Ineˆs Barroso11 for the MAGIC investigators 1 D f Bi i i B U i i S h l f P bli 4 Oxford Centre for Diabetes, Endocrinology and Metabolism, University of Oxford, Oxford, UK. Adiponectin: Multiethnic Genome-Wide Meta-Analysis Institute of Biometrics and Epidemiology, German Diabetes Center, Leibniz Center for Diabetes Research at Heinrich Heine University Du¨sseldorf, 40225 Du¨sseldorf, Germany 65. Department of Human Genetics, McGill University, Montreal H3H 1P3, Canada 66. Department of Medicine, Faculty of Medicine, McGill University, Montreal, H3A 1A4, Canada 67. McGill University and Genome Quebec Innovation Centre, Montreal, H3A 1A4. Canada 68. Department of Metabolic Diseases, Heinrich Heine University Du¨sseldorf, 40225 Du¨sseldorf, Germany 69. Department of Endocrinology and Diabetes, Norfolk and Norwich University Hospital NHS Trust , Norwich, NR1 7UY, UK. 70. General Medicine Division, Massachusetts General Hospital, Boston, Massachusetts, USA PLoS Genetics \| www.plosgenetics.org March 2012 \| Volume 8 \| Issue 3 \| e1002607 PLoS Genetics \| www.plosgenetics.org 16 Adiponectin: Multiethnic Genome-Wide Meta-Analysis 1 Department of Biostatistics, Boston University School of Public Health, Boston, Massachusetts, USA. Adiponectin: Multiethnic Genome-Wide Meta-Analysis y 5 Wellcome Trust Centre for Human Genetics, University of Oxford, Oxford, UK. 6 Program in Medical and Population Genetics, Broad Institute, Cambridge, Massachusetts, USA. 7 Center for Human Genetic Research, Massachusetts General Hospital, Boston, Massachusetts, USA. p 8 Wellcome Trust Sanger Institute, Hinxton, Cambridge, UK. 9 Twin Research and Genetic Epidemiology Department, King’s College London, St Thomas’ Hospital Campus, London, UK. 10 Center for Statistical Genetics, Department of Biostatistics, University of Michigan School of Public Health, Ann Arbor, Michigan, USA. 11 Metabolic Disease Group, Wellcome Trust Sanger Institute, Hinxton, Cambridge, UK. 12 Cardiovascular Health Research Unit and Department of Medicine, University of Washington, Seattle, Washington, USA. 13 Centre National de la Recherche Scientifique–Unite´ Mixte de Recherche 8090 , Pasteur Institute, Lille 2 –Droit et Sante´ University, Lille, France. 14 Department of Medical Genetics, University of Lausanne, Lausanne, Switzerland. 15 University Institute of Social and Preventative Medicine, Centre Hospitalier Universitaire Vaudois (CHUV) and University of Lausanne, Lausanne, Switzerland. 16 Swiss Institute of Bioinformatics, Lausanne, Switzerland. 17 Department of Epidemiology and Public Health, Imperial College London, Faculty of Medicine, Norfolk Place, London, UK. 18 Boston University Data Coordinating Center, Boston, Massachusetts, USA. 19 deCODE Genetics, Reykjavik, Iceland. 20 Department of Human Genetics, Leiden University Medical Centre, Leiden, The Netherlands. 21 Institute of Epidemiology, Helmholtz Zentrum Muenchen, German Research Center for Environmental Health, Neuherberg, Germany. 22 Department of Epidemiology, Erasmus Medical College, Rotterdam, The Netherlands. 23 Department of Biological Psychology, VU University Amsterdam, Amsterdam, The Netherlands. 24 Centre for Population Health Sciences, University of Edinburgh, Edinburgh, UK. 25 MRC Human Genetics Unit, Institute of Genetics and Molecular Medicine, Edinburgh, UK. 26 Division of Genetics, Research and Development, GlaxoSmithKline, King of Prussia, Pennsylvania, USA. 27 Department of Cardiovascular Medicine, University of Oxford, Oxford, UK. 28 Genetics of Complex Traits, Institute of Biomedical and Clinical Sciences, Peninsula College of Medicine and Dentistry, University of Exeter, Exeter, UK. 29 National Institute of Aging, Baltimore, Maryland, US 30 Unit for Child and Adolescent Health and Welfare, National Institute for Health and Welfare, Biocenter Oulu, University of Oulu, Oulu, Finland. 31 Hagedorn Research Institute, Gentofte, Denmark. 32 Department of Medicine and Therapeutics, Level 7, Ninewells Hospital and Medical School, Dundee, UK. p 33 Department of Epidemiology, Bloomberg School of Public Health, J h H ki U i it B lti M l d USA 33 Department of Epidemiology, Bloomberg School o Johns Hopkins University, Baltimore, Maryland, USA. 2 National Heart, Lung, and Blood Institute’s Framingham Heart Study, Framingham, Massachusetts, USA. Adiponectin: Multiethnic Genome-Wide Meta-Analysis 41 Genome Quebec Innovation Centre, Montreal, Canada. 79 Centre National de Ge´notypage/Institut de ge´nomique/Commis- sariat a` l’e´nergie atomique, Evry Cedex, France. 80 INSERM U872, Faculte´ de Me´decine Paris Descartes, Paris Cedex, France. 42 Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, Sweden. Faculte´ de Me´decine Paris Descartes, Paris Cedex, France. 81 Institute for Clinical Diabetology, German Diabetes Center, Leibniz Center for Diabetes Research at Heinrich Heine University Du¨sseldorf, Du¨sseldorf, Germany. 43 Department of Public Health and Caring Sciences, Uppsala University, Uppsala, Sweden. 44 Division of Statistical Genomics, Department of Genetics, Washing- ton University School of Medicine, St. Louis, Missouri, USA. 82 Institute of Genetic Medicine, European Academy Bozen/Bolzano (EURAC), Viale Druso, Bolzano, Italy, Affiliated Institute of the University Lu¨beck, Lu¨beck, Germany. 45 Division of Endocrinology, Diabetes and Nutrition, University of Maryland School of Medicine, Baltimore, Maryland, USA. 83 Department of Pulmonary Physiology, Sir Charles Gairdner Hospital, Perth, Australia. y y 46 INSERM U 859 , Universite de Lille-Nord de France, Lille, France. 47 Genome Technology Branch, National Human Genome Research Institute, Bethesda, Maryland, USA. 84 Busselton Population Medical Research Foundation, Sir Charles Gairdner Hospital, Perth, Australia. 48 The Broad Institute, Cambridge, Massachusetts, USA. p 85 Heart Institute of Western Australia, Sir Charles Gairdner Hospital, Nedlands West Australia, Australia. 49 Leiden Genome Technology Center, Leiden University Medical Center, Leiden, The Netherlands. Nedlands West Australia, Australia. , 86 School of Medicine and Pharmacology, University of Western Australia, Nedlands West Australia, Australia. 86 School of Medicine and Pharmacology, University of Western 50 INSERM U 780 , Paris Sud University, Villejuif, France. Australia, Nedlands West Australia, Australia. Australia, Nedlands West Australia, Australia. 87 Folkhalsan Research Centre, Helsinki, Finland. 51 The Heart Research Institute, Sydney, New South Wales, Australia. , , 87 Folkhalsan Research Centre, Helsinki, Finland. 52 PathWest Laboratory of Western Australia, Department of Molecular Genetics, J Block, QEII Medical Centre, Nedlands West Australia, Australia. 88 Malmska Municipal Health Care Center and Hospital, Jakobstad, Finland. 89 Nuffield Department of Surgery, University of Oxford, Oxford, UK. 53 School of Surgery and Pathology, University of Western Australia, Nedlands West Australia, Australia. 90 Research Centre for Prevention and Health, Glostrup University Hospital, Glostrup, Denmark. 54 Department of Social Medicine, University of Bristol, Bristol, UK. 54 Department of Social Medicine, University of Bristol, 91 Faculty of Health Science, University of Copenhagen, Copenhagen, Denmark. 55 Landspitali University Hospital, Reykjavik, Iceland. 56 Icelandic Heart Association, Kopavogur, Iceland. Adiponectin: Multiethnic Genome-Wide Meta-Analysis 92 National Institute for Health and Welfare, Unit of Population Studies, Turku, Finland. 57 The Human Genetics Center and Institute of Molecular Medicine University of Texas Health Science Center, Houston, Texas, USA. 93 Institute of Health Sciences and Biocenter Oulu, University of Oulu, Oulu, Finland. 58 Steno Diabetes Center, Gentofte, Denmark. 59 Faculty of Health Science, University of Aarhus, Aarhus, Denmark. 94 Department of Public Health, Faculty of Medicine, University of Helsinki, Helsinki, Finland. 59 Faculty of Health Science, University of Aarhus, Aarhus, Denmark. 60 Department of Medicine, University of Leipzig, Leipzig, Germany. 60 Department of Medicine, University of Leipzig, Leipzig, Germany. 95 National Institute for Health and Welfare, Unit for Child and Adolescent Mental Health, Helsinki, Finland. 61 Endocrinology–Diabetology Unit, Corbeil-Essonnes Hospital, Es- sonnes, France. 96 Institute for Molecular Medicine Finland (FIMM), University of Helsinki, Helsinki, Finland. 62 Medical Genetics Institute, Cedars-Sinai Medical Center, Los Angeles, California, USA. 97 Department of Internal Medicine and Biocenter Oulu, Oulu, Finland. 63 Clinical Trial Service Unit and Epidemiological Studies Unit, University of Oxford, Oxford, UK. 98 Diabetes Genetics, Institute of Biomedical and Clinical Science, Peninsula College of Medicine and Dentistry, University of Exeter, Exeter, UK. 64 Centre for Genetic Epidemiology and Biostatistics, University of Western Australia, Perth, Australia. , , 65 Istituto di Neurogenetica e Neurofarmacologia (INN), Consiglio 99 National Institute for Health and Welfare, Unit of Living Conditions, Health and Wellbeing, Helsinki, Finland. Nazionale delle Ricerche, c/o Cittadella Universitaria di Monserrato, Monserrato, Cagliari, Italy. g 100 Interdisciplinary Centre for Clinical Research, University of Leipzig, Leipzig, Germany. 66 Western Australian Sleep Disorders Research Institute, Queen Elizabeth Medical Centre II, Perth, Australia. p g p g y 101 The Danish Twin Registry, Epidemiology, Institute of Public Health, University of Southern Denmark, Odense, Denmark. 67 Department of Endocrinology, Diabetes and Nutrition, Charite- Universitaetsmedizin Berlin, Berlin, Germany. 102 Department of Clinical Sciences, Diabetes and Endocrinology, L d U i i U i i H i l M l M l S d 102 Department of Clinical Sciences, Diabetes and Endocrinology, Lund University, University Hospital Malmo¨, Malmo¨, Sweden. 102 Department of Clinical Sciences, Diabetes and En 68 Department of Clinical Nutrition, German Institute of Human Nutrition Potsdam-Rehbruecke, Nuthetal, Germany. 103 Gladstone Institute of Cardiovascular Disease, University of California, San Francisco, California, USA. 69 Division of Endocrinology, Diabetes, and Hypertension, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts, USA. Adiponectin: Multiethnic Genome-Wide Meta-Analysis 33 Department of Epidemiology, Bloomberg School Johns Hopkins University, Baltimore, Maryland, USA. 34 Department of Nutrition–Dietetics, Harokopio University, Athens, Greece. 35 General Medicine Division, Massachusetts General Hospital, Boston, Massachusetts, USA. 36 Department of Epidemiology and Public Health, University College London, London, UK. 37 Departments of Nutrition and Epidemiology, Harvard School of Public Health, Boston, Massachusetts, USA. 1 Department of Biostatistics, Boston University School of Public Health, Boston, Massachusetts, USA. 38 MRC Centre for Causal Analyses in Translational Epidemiology, University of Bristol, Bristol, UK. University of Bristol, Bristol, UK. 2 National Heart, Lung, and Blood Institute’s Framingham Heart Study, Framingham, Massachusetts, USA. 39 Fundacio´n para la Investigacio´n Biome´dica del Hospital Clı´nico San Carlos, Madrid, Spain. Carlos, Madrid, Spain. 3 Medical Research Council (MRC), Epidemiology Unit, Institute of Metabolic Science, Addenbrooke’s Hospital, Cambridge, UK. 40 Departments of Medicine and Human Genetics, McGill University, Montreal, Canada. PLoS Genetics \| www.plosgenetics.org March 2012 \| Volume 8 \| Issue 3 \| e1002607 17 Adiponectin: Multiethnic Genome-Wide Meta-Analysis Adiponectin: Multiethnic Genome-Wide Meta-Analysis Adiponectin: Multiethnic Genome-Wide Meta-Analysis y y g g 125 MRC Epidemiology Resource Centre, University of Southampton, y y g g 125 MRC Epidemiology Resource Centre, University of Southampton, Southampton General Hospital, Southampton, UK. 125 MRC Epidemiology Resource Centre, Univers Southampton General Hospital, Southampton, UK. Southampton General Hospital, Southampton, UK. 126 Department of Epidemiology, University of Texas, M.D. Anderson Cancer Center, Houston, Texas, USA. 126 Department of Epidemiology, University of Texas, M.D. Anderson Cancer Center Houston Texas USA 165 Department of Medicine, Division of Endocrinology, Keck School of Medicine, University of Southern California, Los Angeles, California, USA. 127 Leibniz-Institut fu¨r Arterioskleroseforschung an der Universita¨t Mu¨nster, Mu¨nster, Germany. 166 Department of Genetics, University of North Carolina, Chapel Hill, North Carolina, USA. 128 Atherosclerosis Research Unit, Department of Medicine, Karolinska Institutet, Stockholm, Sweden. 167 National Institute for Health and Welfare, Unit of Diabetes Prevention, Helsinki, Finland. 129 Laboratory of Neurogenetics, National Institute on Aging, Bethesda, Maryland, USA. 168 South Ostrobothnia Central Hospital, Seinajoki, Finland. 130 Department of Epidemiology, University of Washington, Seattle, Washington, USA. 169 Departments of Medicine and Epidemiology, University of Washington, Seattle, Washington, USA. 131 Seattle Epidemiologic Research and Information Center, Depart- ment of Veterans Affairs Office of Research and Development, Seattle, Washington, USA. 170 Longitudinal Studies Section, Clinical Research Branch, National Institute on Aging, NIH, Baltimore, Maryland, USA. 171 Faculty of Medicine, University of Iceland, Reykjavı´k, Iceland. 172 Lab of Cardiovascular Sciences, National Institute on Aging 132 Department of Medical Sciences, Uppsala University, Uppsala, Sweden. 171 Faculty of Medicine, University of Iceland, Reykjavık, Ice 172 Lab of Cardiovascular Sciences, National Institute on National Institutes of Health, Baltimore, Maryland, USA. 171 Faculty of Medicine, University of Iceland, Reykjavık, Iceland. 172 Lab of Cardiovascular Sciences, National Institute on Aging N i l I i f H l h B l i M l d USA 72 Lab of Cardiovascular Sciences, National Institute , National Institutes of Health, Baltimore, Maryland, USA. 133 Medstar Research Institute, Baltimore, Maryland, USA. 173 Department of Clinical Sciences/Clinical Chemistry, University of Oulu, University of Oulu, Oulu, Finland. 134 Clinical Research Branch, National Institute on Aging, Baltimore, Maryland, USA. Maryland, USA. 135 Institut interre´gional pour la sante´ (IRSA), La Riche, France. 136 Coordination Centre for Clinical Trials, University of Leipzig, Leipzig, Germany. y 174 National Institute of Health and Welfare, Oulu, Finland. GLGC Consortium: 1 1 GLGC Consortium: Tanya M. Teslovich,1, 118 , Kiran Musunuru,2, 3, 4, 5, 6, 118 , Albert V. Smith,7, 8 , Andrew C. Edmondson,9, 10 , Ioannis M. Stylianou,10 , Masahiro Koseki,11 , James P. Pirruccello,2, 5, 6 , Samuli Ripatti,12, 13 , 4 14 1 15 GLGC Consortium: Tanya M. Teslovich,1, 118 , Kiran Musunuru,2, 3, 4, 5, 6, 118 , Albert V. Smith,7, 8 , Andrew C. Edmondson,9, 10 , Ioannis M. Stylianou,10 , Masahiro Koseki,11 , James P. Pirruccello,2, 5, 6 , Samuli Ripatti,12, 13 , Daniel I. Chasman,4, 14 , Cristen J. Willer,1 , Christopher T. Johansen,15 , Sigrid W. Fouchier,16 , Aaron Isaacs,17 , Gina M. Peloso,18, 19 , Maja Barbalic,20 , Sally L. Ricketts,21 , Joshua C. Bis,22 , Yurii S. Aulchenko,17 , Gudmar Thorleifsson,23 , Mary F. Feitosa,24 , John Chambers,25 , Marju Orho-Melander,26 , Olle Melander,26 , Toby Johnson,27 , Xiaohui Li,28 , Xiuqing Guo,28 , Mingyao Li,9, 10 , Yoon Shin Cho,29 , Min Jin Go,29 , Young Jin Kim,29 , Jong-Young Lee,29 , Taesung Park,30, 31 , Kyunga Kim,32 , Xueling Sim,33 , Rick Twee-Hee Ong,34 , Damien C. Croteau- Chonka,35 , Leslie A. Lange,35 , Joshua D. Smith,36 , Kijoung Song,37 , Jing Hua Zhao,38 , Xin Yuan,37 , Jian’an Luan,38 , Claudia Lamina,39 , Andreas Ziegler,40 , Weihua Zhang,25 , Robert Y. L. Zee,4, 14 , Alan F. Wright,41 , Jacqueline C. M. Witteman,17, 42 , James F. Wilson,43 , Gonneke Willemsen,44 , H.-Erich Wichmann,45 , John B. Whitfield,46 , Dawn M. Waterworth,37 , Nicholas J. Wareham,38 , Ge´rard Waeber,47 , Peter Vollenweider,47 , Benjamin F. Voight,2, 5 , Veronique Vitart,41 , Andre G. Uitterlinden,17, 42, 48 , Manuela Uda,49 , Jaakko Tuomilehto,50 , John R. Thompson,51 , Toshiko Tanaka,52, 53 , Ida Surakka,12, 13 , Heather 141 Department of Internal Medicine, Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerland. 142 Genomic Medicine, Imperial College London, Hammersmith Hospital, London, UK. 143 Epidemiology and Public Health, Queen’s University Belfast, Belfast, UK. 144 Medical Products Agency, Uppsala, Sweden. 145 See Supplementary Note for a full list of autho 146 National Institute for Health and Welfare, Unit of Chronic Disease Epidemiology and Prevention, Helsinki, Finland. 147 Departments of Nutrition and Epidemiology, Harvard School of Public Health, Boston, Massachusetts, USA. 148 Channing Laboratory, Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts, USA. Adiponectin: Multiethnic Genome-Wide Meta-Analysis 115 Department of Obstetrics and Gynaecology, Oulu University Hospital, Oulu, Finland. 152 The Welch Center for Prevention, Epidemiology, and Clinical Research, School of Medicine and Bloomberg School of Public Health, J h H ki U i i B l i M l d USA g Johns Hopkins University, Baltimore, Maryland, USA 116 Departments of Medicine, Epidemiology and Health Services, University of Washington, Seattle, Washington, USA. J p y y 153 Division of Epidemiology and Community Health, School of Public 153 Division of Epidemiology and Community Healt 153 Division of Epidemiology and Community Health, School Health, University of Minnesota, Minneapolis, Minnesota, USA. 153 Division of Epidemiology and Community Health, School o Health, University of Minnesota, Minneapolis, Minnesota, USA. 117 Group Health Research Institute, Group Health Cooperative, Seattle, Washington, USA. y p 154 Department of Endocrinology and Diabetes, Norfolk and Norwich University Hospital National Health Service Trust, Norwich, UK. 118 Institute of Biometrics and Epidemiology, German Diabetes Centre, 155 Department of Medicine, University of Kuopio and Kuopio 155 Department of Medicine, University of Kuopio and Kuopio University Hospital, Kuopio, Finland. Leibniz Centre at Heinrich Heine University Du¨sseldorf, Du¨sseldorf, Germany. 156 Faculty of Health Science, University of Southern Denmark, Odense, Denmark. 119 Department of Biostatistics, University of Washington, Seattle, Washington, USA. 157 Institute of Biomedical Science, Faculty of Health Science, University of Copenhagen, Copenhagen, Denmark. 120 Department of Internal Medicine, Erasmus Medical College, Rotterdam, The Netherlands. 158 Department of Neurology, General Central Hospital, Bolzano, Italy. 121 Department of Metabolic Diseases, Heinrich Heine University Du¨sseldorf, Du¨sseldorf, Germany. y 159 Department of Neurology, University of Lu¨beck, Lu¨beck, Germany. 122 Department of Public Health and Clinical Medicine, Section for Family Medicine, Umea˚ University, Umea˚, Sweden. 160 Institute of Medical Informatics, Biometry and Epidemiology, Ludwig-Maximilians-Universita¨t, Munich, Germany. 160 Institute of Medical Informatics, Biometry Ludwig-Maximilians-Universita¨t, Munich, Germany. 123 School of Public Health, Department of General Practice, University of Aarhus, Aarhus, Denmark. 161 Klinikum Grosshadern, Munich, Germany. 162 School of Medicine, University of Split, 162 School of Medicine, University of Split, Split, Croatia. y 124 Department of Public Health and Primary Care, Strangeways Research Laboratory, University of Cambridge, Cambridge, UK. 163 Gen-Info Ltd., Zagreb, Croatia. 163 Gen-Info Ltd., Zagreb, Croatia. p y , g y Research Laboratory, University of Cambridge, Cambridge, UK. 164 Department of Physiology and Biophysics, Keck School of Medicine, University of Southern California, Los Angeles, California, USA. Adiponectin: Multiethnic Genome-Wide Meta-Analysis 104 Diabetes Research Center, Diabetes Unit, Massachusetts General Hospital, Boston, Massachusetts, USA. 70 Department of Human Genetics, Leiden University Medical Centre, Leiden, The Netherlands. 105 Department of Medicine, Harvard Medical School, Boston, Massachusetts, USA. 71 Department of Cardiovascular Research, Istituto di Ricerche Farmacologiche ‘Mario Negri’, Milan, Italy. 106 Division of Cardiology, University of Ottawa Heart Institute, Ottawa, Ontario, Canada. 72 Institut National de la Sante´ et de la Recherche Me´dicale, Institut National de la Recherche Agronomique, Universite´ Paris 107 Oxford National Institute for Health Research, Biomedical Research Centre, Churchill Hospital, Oxford, UK. 13 , Bobigny Cedex, France. 108 Department of Clinical Genetics, Erasmus Medical College, Rotterdam, The Netherlands. 73 Department of Medicine III, Division Prevention and Care of Diabetes, University of Dresden, Dresden, Germany. 109 Biomedical Research Institute, University of Dundee, Ninewells Hospital and Medical School, Dundee, UK. 74 Center for Human Nutrition, University of Texas Southwestern Medical Center, Dallas, Texas, USA. 110 Department of Geriatric Medicine and Metabolic Disease, Second University of Naples, Naples, Italy. 75 Department of Genetics and Pathology, Rudbeck Laboratory, Uppsala University, Uppsala, Sweden. 111 National Institute for Health and Welfare, Unit of Public Health Genomics, Helsinki, Finland. 76 Centre Hospitalier Universitaire, de Poitiers, Endocrinologie Diabetologie, CIC INSERM 0802, INSERM U927, Universite´ de Poitiers, Unite´ de Formation et de Recherche, Me´decine Pharmacie, Poitiers, France. 112 Department of Medical Genetics, University of Helsinki, Helsinki, Finland. 113 Department of Medical Statistics, Epidemiology and Medical Informatics, Andrija Stampar School of Public Health, Medical School, 77 Department of Public Health and Clinical Medicine, Section for Nutritional Research, Umea˚ University, Umea˚, Sweden. University of Zagreb, Rockefellerova, Zagreb, Croatia. 78 Department of Clinical Sciences, Obstetrics and Gynecology, University of Oulu, University of Oulu, Finland. 114 Department of Clinical Genetics, VU University and Medical Center, Amsterdam, The Netherlands. PLoS Genetics \| www.plosgenetics.org March 2012 \| Volume 8 \| Issue 3 \| e1002607 March 2012 \| Volume 8 \| Issue 3 \| e1002607 18 Adiponectin: Multiethnic Genome-Wide Meta-Analysis Adiponectin: Multiethnic Genome-Wide Meta-Analysis 35 Institut interre´gional pour la sante´ (IRSA), La Riche, Fran 175 Department of Preventive Medicine, Keck School of Medicine, University of Southern California, Los Angeles, California, USA. 136 Coordination Centre for Clinical Trials, University of Leipzig, Leipzig, Germany. 176 MRC–Health Protection Agency Centre for Environment and Health, Imperial College London, London, UK. 137 Department of Medicine, Helsinki University Hospital, University of Helsinki, Helsinki, Finland. 177 UOC Geriatria, Istituto Nazionale Ricovero e cura per Anziani (INRCA) IRCCS, Rome, Italy. 178 These authors contributed equally to this work. Correspondence should be addressed to M.B. (boehnke@umich.edu), M.I.M. (mark.mccarthy@drl.ox.ac.uk), J.C.F. (jcflorez@partners.org) or I.B. (ib1 @sanger.ac.uk). 138 Department of Internal Medicine, Leiden University Medical Centre, Leiden, The Netherlands. 139 Research Unit, Cardiovascular Genetics, Nancy University Henri Poincare´, Nancy, France. 140 EMGO Institute for Health and Care Research, Department of Psychiatry, VU University Medical Center, Amsterdam, The Netherlands. Adiponectin: Multiethnic Genome-Wide Meta-Analysis 18 Department of Biostatistics, Boston University School of Public Health, Boston, Massachusetts 02118, USA. 19 National Heart, Lung and Blood Institute’s Framingham Heart Study, Framingham, Massachusetts 01702, USA. 20 Human Genetics Center, University of Texas Health Science Center at Houston, Houston, Texas 77030, USA. 21 Department of Public Health and Primary Care, Strangeways Research Laboratory, University of Cambridge, Cambridge CB1 8RN, UK. 22 Cardiovascular Health Research Unit and Department of Medicine, University of Washington, Seattle, Washington 98101, USA. 23 deCODE Genetics, 101 Reykjavik, Iceland. 24 Division of Statistical Genomics in the Center for Genome Sciences, Washington University School of Medicine, St Louis, Missouri 63108, USA. 25 Department of Epidemiology and Public Health, Imperial College London, London W2 1PG, UK. 26 Department of Clinical Sciences, Lund University, SE-20502, Malmo¨, Sweden. 27 Clinical Pharmacology and Barts and the London Genome Centre, William Harvey Research Institute, Barts and the London School of Medicine, Queen Mary University of London, London EC1M 6BQ, UK. Yii-der I. Chen,28 , Mark J. Caulfield,27 , Harry Campbell,43 , Noel P. Burtt,5 , Lori L. Bonnycastle,99 , Dorret I. Boomsma,44 , S. Matthijs Boekholdt,100 , Richard N. Bergman,101 , Ineˆs Barroso,55 , Stefania Bandinelli,102 , Christie M. Ballantyne,103 , Themistocles L. Assimes,104 , 28 Medical Genetics Institute, Cedars-Sinai Medical Center, Los Angeles, California 90048, USA. 29 Center for Genome Science, National Institute of Health, Seoul 122- 701, Republic of Korea. 30 Interdisciplinary Program in Bioinformatics, College of Natural Sciences, Seoul National University, Seoul 151-742, Republic of Korea. 31 Department of Statistics, College of Natural Sciences, Seoul National University, Seoul 151-742, Republic of Korea. 32 Department of Statistics, Sookmyung Women’s University, Seoul 140-742, Republic of Korea. 33 Centre for Molecular Epidemiology, National University of Singapore, Singapore, 117597, Republic of Singapore. 34 Genome Institute of Singapore, Singapore 138672, Republic of Singapore. 35 Department of Genetics, University of North Carolina, Chapel Hill, North Carolina 27599, USA. 36 Department of Genome Sciences, University of Washington, Seattle, Washington 98195, USA. 1 Center for Statistical Genetics, Department of Biostatistics, University of Michigan, Ann Arbor, Michigan 48109, USA. 37 Genetics Division, GlaxoSmithKline R&D, King of Prussia, Pennsylvania 19406, USA. 2 Center for Human Genetic Research, Massachusetts General Hospital, Boston, Massachusetts 02114, USA. 38 MRC Epidemiology Unit, Institute of Metabolic Science, Adden- brooke’s Hospital, Cambridge CB2 0QQ, UK. 3 Cardiovascular Research Center, Massachusetts General Hospital, Boston, Massachusetts 02114, USA. Adiponectin: Multiethnic Genome-Wide Meta-Analysis Kees Hovingh,16 , Jouke-Jan Hottenga,44 , Albert Hofman,17, 42 , Andrew A. Hicks,65 , Christian Hengstenberg,88 , Iris M. Heid,45, 89 , Caroline Hayward,41 , Aki S. Havulinna,50, 90 , Nicholas D. Hastie,41 , Tamara B. Harris,80 , Talin Haritunians,28 , Alistair S. Hall,91 , Ulf Gyllensten,85 , Candace Guiducci,5 , Leif C. Groop,26, 92 , Elena Gonzalez,5 , Christian Gieger,45 , Nelson B. Freimer,93 , Luigi Ferrucci,94 , Jeanette Erdmann,95 , Paul Elliott,86, 96 , Kenechi G. Ejebe,5 , Angela Do¨ring,45 , Anna F. Dominiczak,97 , Serkalem Demissie,18, 19 , Panagiotis Deloukas,55 , Eco J. C. de Geus,44 , Ulf de Faire,98 , Gabriel Crawford,5 , Francis S. Collins,99 , Yii-der I. Chen,28 , Mark J. Caulfield,27 , Harry Campbell,43 , Noel P. Burtt,5 , Lori L. Bonnycastle,99 , Dorret I. Boomsma,44 , S. Matthijs Boekholdt,100 , Richard N. Bergman,101 , Ineˆs Barroso,55 , Stefania Bandinelli,102 , Christie M. Ballantyne,103 , Themistocles L. Assimes,104 , Thomas Quertermous,104 , David Altshuler,2, 4, 5 , Mark Seielstad,34 , Tien Y. Wong,105 , E-Shyong Tai,106 , Alan B. Feranil,107 , Christopher W. Kuzawa,108 , Linda S. Adair,109 , Herman A. Taylor Jr,110 , Ingrid B. Borecki,24 , Stacey B. Gabriel,5 , James G. Wilson,110 , Hilma Holm,23 , Unnur Thorsteinsdottir,8, 23 , Vilmundur Gudnason,7, 8 , Ronald M. Krauss,111 , Karen L. Mohlke,35 , Jose M. Ordovas,112, 113 , Patricia B. Munroe,114 , Jaspal S. Kooner,59 , Alan R. Tall,11 , Robert A. Hegele,15 , John J.P. Kastelein,16 , Eric E. Schadt,115 , Jerome I. Rotter,28 , Eric Boerwinkle,20 , David P. Strachan,116 , Vincent Mooser,37 , Kari Stefansson,8, 23 , Muredach P. Reilly,9, 10 , Nilesh J Samani,117 , Heribert Schunkert,95 , L. Adrienne Cupples,18, 19, 118 , Manjinder S. Sandhu,21, 38, 55, 118 , Paul M Ridker,4, 14, 118 , Daniel J. Rader,9, 10, 118 , Cornelia M. van Duijn,17, 42, 118 , Leena Peltonen,119 , Gonc¸alo R. Abecasis,1, 118 , Michael Boehnke1, 118 , & Sekar Kathiresan, 15 Robarts Research Institute, University of Western Ontario, London, Ontario N6A 5K8, Canada. 16 Department of Vascular Medicine, Academic Medical Centre at the University of Amsterdam, 1105 AZ Amsterdam, The Netherlands. niversity of Amsterdam, 1105 AZ Amsterdam, The Netherlan 17 Department of Epidemiology, Erasmus University Medical Center, P.O. Box 2040, 3000 CA Rotterdam, The Netherlands. P.O. Box 2040, 3000 CA Rotterdam, The Netherlands. GLGC Consortium: 1 1 149 Genetic Epidemiology and Clinical Research Group, Department of Public Health and Clinical Medicine, Section for Medicine, Umea˚ University Hospital, Umea˚, Sweden. y p 150 London School of Hygiene and Tropical Medicine, London, UK. 151 Department of Medicine, School of Medicine, Johns Hopkins University, Baltimore, Maryland, USA. PLoS Genetics \| www.plosgenetics.org March 2012 \| Volume 8 \| Issue 3 \| e1002607 19 Adiponectin: Multiethnic Genome-Wide Meta-Analysis Adiponectin: Multiethnic Genome-Wide Meta-Analysis Wong,105 , E-Shyong Tai,106 , Alan B. Feranil,107 , Christopher W. Kuzawa,108 , Linda S. Adair,109 , Herman A. Taylor Jr,110 , Ingrid B. Borecki,24 , Stacey B. Gabriel,5 , James G. Wilson,110 , Hilma Holm,23 , Unnur Thorsteinsdottir,8, 23 , Vilmundur Gudnason,7, 8 , Ronald M. Krauss,111 , Karen L. Mohlke,35 , Jose M. Ordovas,112, 113 , Patricia B. Munroe,114 , Jaspal S. Kooner,59 , Alan R. Tall,11 , Robert A. Hegele,15 , John J.P. Kastelein,16 , Eric E. Schadt,115 , Jerome I. Rotter,28 , Eric Boerwinkle,20 , David P. Strachan,116 , Vincent Mooser,37 , Kari Stefansson,8, 23 , Muredach P. Reilly,9, 10 , Nilesh J Samani,117 , Heribert Schunkert,95 , L. Adrienne Cupples,18, 19, 118 , Manjinder S. Sandhu,21, 38, 55, 118 , Paul M Ridker,4, 14, 118 , Daniel J. Rader,9, 10, 118 , Cornelia M. van Duijn,17, 42, 118 , Leena Peltonen,119 , Gonc¸alo R. Abecasis,1, 118 , Michael Boehnke1, 118 , & Sekar Kathiresan, M. Stringham,1 , Tim D. Spector,54 , Nicole Soranzo,54, 55 , Johannes H. Smit,56 , Juha Sinisalo,57 , Kaisa Silander,12, 13 , Eric J. G. Sijbrands,17, 48 , Angelo Scuteri,58 , James Scott,59 , David Schlessinger,60 , Serena Sanna,49 , Veikko Salomaa,50 , Juha Saharinen,12 , Chiara Sabatti,61 , Aimo Ruokonen,62 , Igor Rudan,43 , Lynda M. Rose,14 , Robert Roberts,63 , Mark Rieder,36 , Bruce M. Psaty,64 , Peter P. Pramstaller,65 , Irene Pichler,65 , Markus Perola,12, 13 , Brenda W. J. H. Penninx,56 , Nancy L. Pedersen,66 , Cristian Pattaro,65 , Alex N. Parker,67 , Guillaume Pare,68 , Ben A. Oostra,69 , Christopher J. O’Donnell,4, 19 , Markku S. Nieminen,57 , Deborah A. Nickerson,36 , Grant W. Montgomery,46 , Thomas Meitinger,70, 71 , Ruth McPherson,63 , Mark I. McCarthy,72, 73, 74 , Wendy McArdle,75 , David Masson,11 , Nicholas G. Martin,46 , Fabio Marroni,76 , Massimo Mangino,54 , Patrik K. E. Magnusson,66 , Gavin Lucas,77 , Robert Luben,21 , Ruth J. F. Loos,38 , Marja-Liisa Lokki,78 , Guillaume Lettre,79 , Claudia Langenberg,38 , Lenore J. Launer,80 , Edward G. Lakatta,60 , Reijo Laaksonen,81 , Kirsten O. Kyvik,82 , Florian Kronenberg,39 , Inke R. Ko¨nig,40 , Kay-Tee Khaw,21 , Jaakko Kaprio,12, 13, 83 , Lee M. Kaplan,84 , A˚ sa Johansson,85 , Marjo-Riitta Jarvelin,86, 87 , A. Cecile J. W. Janssens,17 , Erik Ingelsson,66 , Wilmar Igl,85 , G. Adiponectin: Multiethnic Genome-Wide Meta-Analysis M. Stringham,1 , Tim D. Spector,54 , Nicole Soranzo,54, 55 , Johannes H. Smit,56 , Juha Sinisalo,57 , Kaisa Silander,12, 13 , Eric J. G. Sijbrands,17, 48 , Angelo Scuteri,58 , James Scott,59 , David Schlessinger,60 , Serena Sanna,49 , Veikko Salomaa,50 , Juha Saharinen,12 , Chiara Sabatti,61 , Aimo Ruokonen,62 , Igor Rudan,43 , Lynda M. Rose,14 , Robert Roberts,63 , Mark Rieder,36 , Bruce M. Psaty,64 , Peter P. Pramstaller,65 , Irene Pichler,65 , Markus Perola,12, 13 , Brenda W. J. H. Penninx,56 , Nancy L. Pedersen,66 , Cristian Pattaro,65 , Alex N. Parker,67 , Guillaume Pare,68 , Ben A. Oostra,69 , Christopher J. O’Donnell,4, 19 , Markku S. Nieminen,57 , Deborah A. Nickerson,36 , Grant W. Montgomery,46 , Thomas Meitinger,70, 71 , Ruth McPherson,63 , Mark I. McCarthy,72, 73, 74 , Wendy McArdle,75 , David Masson,11 , Nicholas G. Martin,46 , Fabio Marroni,76 , Massimo Mangino,54 , Patrik K. E. Magnusson,66 , Gavin Lucas,77 , Robert Luben,21 , Ruth J. F. Loos,38 , Marja-Liisa Lokki,78 , Guillaume Lettre,79 , Claudia Langenberg,38 , Lenore J. Launer,80 , Edward G. Lakatta,60 , Reijo Laaksonen,81 , Kirsten O. Kyvik,82 , Florian Kronenberg,39 , Inke R. Ko¨nig,40 , Kay-Tee Khaw,21 , Jaakko Kaprio,12, 13, 83 , Lee M. Kaplan,84 , A˚ sa Johansson,85 , Marjo-Riitta Jarvelin,86, 87 , A. Cecile J. W. Janssens,17 , Erik Ingelsson,66 , Wilmar Igl,85 , G. Kees Hovingh,16 , Jouke-Jan Hottenga,44 , Albert Hofman,17, 42 , Andrew A. Hicks,65 , Christian Hengstenberg,88 , Iris M. Heid,45, 89 , Caroline Hayward,41 , Aki S. Havulinna,50, 90 , Nicholas D. Hastie,41 , Tamara B. Harris,80 , Talin Haritunians,28 , Alistair S. Hall,91 , Ulf Gyllensten,85 , Candace Guiducci,5 , Leif C. Groop,26, 92 , Elena Gonzalez,5 , Christian Gieger,45 , Nelson B. Freimer,93 , Luigi Ferrucci,94 , Jeanette Erdmann,95 , Paul Elliott,86, 96 , Kenechi G. Ejebe,5 , Angela Do¨ring,45 , Anna F. Dominiczak,97 , Serkalem Demissie,18, 19 , Panagiotis Deloukas,55 , Eco J. C. de Geus,44 , Ulf de Faire,98 , Gabriel Crawford,5 , Francis S. Collins,99 , Yii-der I. Chen,28 , Mark J. Caulfield,27 , Harry Campbell,43 , Noel P. Burtt,5 , Lori L. Bonnycastle,99 , Dorret I. Boomsma,44 , S. Matthijs Boekholdt,100 , Richard N. Bergman,101 , Ineˆs Barroso,55 , Stefania Bandinelli,102 , Christie M. Ballantyne,103 , Themistocles L. Assimes,104 , Thomas Quertermous,104 , David Altshuler,2, 4, 5 , Mark Seielstad,34 , Tien Y. Adiponectin: Multiethnic Genome-Wide Meta-Analysis 49 Istituto di Neurogenetica e Neurofarmacologia (INN), Consiglio Nazionale delle Ricerche, c/o Cittadella Universitaria di Monserrato, Monserrato, Cagliari 09042, Italy. 83 Faculty of Medicine, Department of Public Health, University of Helsinki, FI-00014 Helsinki, Finland. 84 Massachusetts General Hospital Weight Center, Boston, Massachu- setts 02114, USA. g y 50 Department of Chronic Disease Prevention, National Institute for Health and Welfare, FI-00271 Helsinki, Finland. 85 Department of Genetics and Pathology, Rudbeck Laboratory, University of Uppsala, SE-75185 Uppsala, Sweden. 51 Department of Health Sciences, University of Leicester, Leicester LE1 6TP, UK. 86 Department of Epidemiology & Biostatistics, Imperial College London, St Mary’s Campus, Norfolk Place, London W2 1PG, UK. 52 Clinical Research Branch, National Institute on Aging, National Institutes of Health, Baltimore, Maryland 21225, USA. 87 Department of Public Health Science and General Practice, University of Oulu, FI-90220 Oulu, Finland. y 53 Medstar Research Institute, Baltimore, Maryland 21218, USA. y 88 Klinik und Poliklinik fu¨r Innere Medizin II, Universita¨t Regensburg, 93053 Regensburg, Germany. 54 Department of Twin Research and Genetic Epidemiology, King’s College London, London SE1 7EH, UK. g 55 Wellcome Trust Sanger Institute, Hinxton, Cambridge CB10 1SA, UK. g g y 89 Department of Epidemiology and Preventive Medicine Regensburg University Medical Center Franz-Josef-Strauss-Allee 11, 93053 Regens- burg, Germany. 56 Department of Psychiatry, EMGO Institute, Neuroscience Campus Amsterdam, VU University Medical Center, 1007 MB Amsterdam, The Netherlands. 90 Department of Biomedical Engineering and Computational Science, Aalto University School of Science and Technology, FI-00076 Aalto, Finland. 57 Division of Cardiology, Department of Medicine, Helsinki University Central Hospital (HUCH), FI-00029 Helsinki, Finland. 91 LIGHT Research Institute, Faculty of Medicine and Health, University of Leeds, Leeds LS2 9JT, UK. 58 Unita Operativa Geriatria, Istituto Nazionale Ricovero e Cura Anziani (INRCA), Istituto Ricovero e Cura a Carattere Scientifico (IRCCS), Via Cassia 1167, 00189 Rome, Italy. 92 Department of Medicine, Helsinki University Hospital, FI-00029 Helsinki, Finland. 59 Hammersmith Hospital, National Heart and Lung Institute, Imperial College London, London W12 0NN, UK. 93 Department of Psychiatry, Center for Neurobehavioral Genetics, The Jane and Terry Semel Institute for Neuroscience and Human Behavior, David Geffen School of Medicine, University of California, Los Angeles, California 90095, USA. 60 Gerontology Research Center, National Institute on Aging, 5600 Nathan Shock Drive, Baltimore, Maryland 21224, USA. 94 Clinical Research Branch, National Institute on Aging, National Institutes of Health, Baltimore, Maryland 21225, USA. 61 Department of Health Research and Policy, Stanford University, Stanford, California 94305, USA. Adiponectin: Multiethnic Genome-Wide Meta-Analysis 62 Department of Clinical Chemistry, University of Oulu, FI-90220 Oulu, Finland. 95 Medizinische Klinik II, Universita¨t zu Lu¨beck, 23538 Lu¨beck, Germany. 96 MRC-HPA Centre for Environment and Health, Imperial College London, London W2 1PG, UK. 63 The John & Jennifer Ruddy Canadian Cardiovascular Genetics Centre, University of Ottawa, Ottawa K1Y 4W7, Canada. 97 BHF Glasgow Cardiovascular Research Centre, University of Glasgow, 126 University Place, Glasgow G12 8TA, UK. y 64 Departments of Medicine, Epidemiology, and Health Services, University of Washington, Group Health Research Institute, Group Health Cooperative, Seattle, Washington 98101, USA. g , y , g , 98 Division of Cardiovascular Epidemiology, Institute of Environmental g y g 98 Division of Cardiovascular Epidemiology, Institute of Environmental Medicine Karolinska Institutet SE-17177 Stockholm Sweden 98 Division of Cardiovascular Epidemiology, Institute of Environ Medicine, Karolinska Institutet, SE-17177 Stockholm, Sweden. p g 65 Institute of Genetic Medicine, European Academy Bozen/Bolzano (EURAC), Viale Druso 1, 39100 Bolzano, Italy – affiliated institute of the University of Lu¨beck, Germany. p gy Medicine, Karolinska Institutet, SE-17177 Stockholm, Sweden. 99 National Human Genome Research Institute, National Institutes of Health, Bethesda, Maryland, 20892, USA. University of Lu¨beck, Germany. 66 Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, SE-17177 Stockholm, Sweden. 66 Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, SE-17177 Stockholm, Sweden. 100 Departments of Vascular Medicine & Cardiology, Academic Medical Centre, 1105 AZ Amsterdam, The Netherlands. 101 Department of Physiology and Biophysics, University of Southern California, Los Angeles, California 90033, USA. 67 Amgen, Thousand Oaks, California 91320, USA. 67 Amgen, Thousand Oaks, California 91320, USA. 68 Genetic and Molecular Epidemiology Laboratory, McMaster University, Hamilton, Ontario L8N3Z5, Canada. 102 Geriatric Unit, Azienda Sanitaria Firenze (ASF), 50125 Florence, Italy. 69 Department of Clinical Genetics, Erasmus University Medical Center, 3000 CA Rotterdam, The Netherlands. 103 Department of Medicine, Baylor College of Medicine, Houston, Texas 77030, USA. 70 Institut fur Humangenetik, Helmholtz Zentrum Munchen, Deutsches Forschungszentrum fur Umwelt und Gesundheit, 85764 Neuherberg, Germany. 104 Department of Medicine, Stanford University School of Medicine, Stanford, California 94305, USA. 71 Institute of Human Genetics, Klinikum rechts der Isar, Technische Universita¨t Mu¨nchen, 81675 Muenchen, Germany. 105 Singapore Eye Research Institute, National University of Singapore, Singapore 168751, Republic of Singapore. g p , p g p 106 Departments of Medicine/Epidemiology and Public Health, Yong 72 Wellcome Trust Centre for Human Genetics, University of Oxford, Roosevelt Drive, Oxford OX3 7BN, UK. Adiponectin: Multiethnic Genome-Wide Meta-Analysis 39 Division of Genetic Epidemiology, Department of Medical Genetics, Molecular and Clinical Pharmacology, Innsbruck Medical University, Schoepfstrasse 41, A-6020 Innsbruck, Austria. 4 Department of Medicine, Harvard Medical School, Boston, Massachusetts 02115, USA. 40 Institut fu¨r Medizinische Biometrie und Statistik, Universita¨t zu Lu¨beck, 23562 Lu¨beck, Germany. 5 Broad Institute, Cambridge, Massachusetts 02142, USA. 6 Johns Hopkins University School of Medicine, Baltimore, Maryland 21287, USA. 41 MRC Human Genetics Unit, Institute of Genetics and Molecular Medicine, Edinburgh EH4 2XU, UK. 7 Icelandic Heart Association, Heart Preventive Clinic and Research Institute, 201 Kopavogur, Iceland. 42 Netherlands Genomics Initiative (NGI)-sponsored Netherlands Consortium for Healthy Aging (NCHA) and Center of Medical Systems Biology (CMSB), 2300 RC Leiden, The Netherlands. 8 University of Iceland, 101 Reykjavik, Iceland. 9 Cardiovascular Institute, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania 19104, USA. 43 Centre for Population Health Sciences, University of Edinburgh, Edinburgh EH8 9AG, UK. 10 Institute for Translational Medicine and Therapeutics, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania, 19104, USA. 44 Department of Biological Psychology, VU University Amsterdam Van der Boechorststraat 1, 1081 BT Amsterdam, The Netherlands. 45 Institute of Epidemiology, Helmholtz Zentrum Munchen – German Research Center for Environmental Health, 85764 Neuherberg, Germany. 11 Division of Molecular Medicine, Department of Medicine, Columbia University, New York, New York 10032, USA. 46 Genetic Epidemiology Unit, Queensland Institute of Medical Research, PO Royal Brisbane Hospital, Queensland 4029, Australia. 12 Institute for Molecular Medicine Finland FIMM, University of Helsinki, FI-00014 Helsinki, Finland. 47 Department of Internal Medicine, Centre Hospitalier Universitaire Vaudois, 1011 Lausanne, Switzerland. 13 National Institute for Health and Welfare, P.O. Box 104, FI-00251 Helsinki, Finland. 14 Division of Preventive Medicine, Brigham and Women’s Hospital, Boston Massachusetts 02215, USA. 48 Department of Internal Medicine, Erasmus University Medical Center, PO Box 2040, 3000 CA Rotterdam, The Netherlands. PLoS Genetics \| www.plosgenetics.org March 2012 \| Volume 8 \| Issue 3 \| e1002607 PLoS Genetics \| www.plosgenetics.org 20 Adiponectin: Multiethnic Genome-Wide Meta-Analysis References 1. Hivert MF, Sullivan LM, Fox CS, Nathan DM, D’Agostino RB, Sr, et al. (2008) Associations of adiponectin, resistin, and tumor necrosis factor-alpha with insulin resistance. J Clin Endocrinol Metab 93: 3165–3172. 17. Jee SH, Sull JW, Lee JE, Shin C, Park J, et al. (2010) Adiponectin concentrations: a genome-wide association study. Am J Hum Genet 87: 545–552. a genome-wide association study. Am J Hum Genet 87: 545–552. 18. Wu Y, Li Y, Lange EM, Croteau-Chonka DC, Kuzawa CW, et al. (2010) Genome-wide association study for adiponectin levels in Filipino women identifies CDH13 and a novel uncommon haplotype at KNG1-ADIPOQ. Hum Mol Genet 19: 4955–4964. 2. Tilg H, Moschen AR (2006) Adipocytokines: mediators linking adipose tissue, inflammation and immunity. Nat Rev Immunol 6: 772–783. 2. Tilg H, Moschen AR (2006) Adipocytokines: mediators linking adipose tissue, inflammation and immunity. Nat Rev Immunol 6: 772–783. 3. Pischon T, Girman CJ, Hotamisligil GS, Rifai N, Hu FB, et al. (2004) Plasma Adiponectin Levels and Risk of Myocardial Infarction in Men. JAMA: The Journal of the American Medical Association 291: 1730–1737. 19. Voight BF, Scott LJ, Steinthorsdottir V, Morris AP, Dina C, et al. (2010) Twelve type 2 diabetes susceptibility loci identified through large-scale association analysis. Nat Genet 42: 579–589. 4. Li S, Shin HJ, Ding EL, van Dam RM (2009) Adiponectin levels and risk of type 2 di b t t ti i d t l i JAMA 302 179 188 4. Li S, Shin HJ, Ding EL, van Dam RM (2009) Adiponectin levels and risk of type 2 diabetes: a systematic review and meta-analysis. JAMA 302: 179–188. 4. Li S, Shin HJ, Ding EL, van Dam RM (2009) Adiponectin levels and risk of type 2 diabetes: a systematic review and meta-analysis. JAMA 302: 179–188. 20. Dupuis J, Langenberg C, Prokopenko I, Saxena R, Soranzo N, et al. (2010) New genetic loci implicated in fasting glucose homeostasis and their impact on type 2 diabetes risk. Nat Genet 42: 105–116. 5. Stumvoll M, Goldstein BJ, van Haeften TW (2005) Type 2 diabetes: principles of pathogenesis and therapy. Lancet 365: 1333–1346. 21. Lindgren CM, Heid IM, Randall JC, Lamina C, Steinthorsdottir V, et al. (2009) Genome-wide association scan meta-analysis identifies three Loci influencing adiposity and fat distribution. PLoS Genet 5: e1000508. doi:10.1371/ journal.pgen.1000508. 6. Nawrocki AR, Rajala MW, Tomas E, Pajvani UB, Saha AK, et al. References (2006) Mice lacking adiponectin show decreased hepatic insulin sensitivity and reduced responsiveness to peroxisome proliferator-activated receptor gamma agonists. J Biol Chem 281: 2654–2660. 22. Teslovich TM, Musunuru K, Smith AV, Edmondson AC, Stylianou IM, et al. (2010) Biological, clinical and population relevance of 95 loci for blood lipids. Nature 466: 707–713. 7. Wang Y, Zhou M, Lam KS, Xu A (2009) Protective roles of adiponectin in obesity-related fatty liver diseases: mechanisms and therapeutic implications. Arq Bras Endocrinol Metabol 53: 201–212. 23. Kilpela¨inen TO, Zillikens MC, Stancˇa´kova´ A, Finucane FM, Ried JS, et al. (2011) Genetic variation near IRS1 associates with reduced adiposity and an impaired metabolic profile. Nature Genetics In press. 8. Comuzzie AG, Funahashi T, Sonnenberg G, Martin LJ, Jacob HJ, et al. (2001) The genetic basis of plasma variation in adiponectin, a global endophenotype for obesity and the metabolic syndrome. Journal of Clinical Endocrinology and Metabolism 86: 4321–4325. 24. Morris AP (2011) Transethnic meta-analysis of genomewide association studies. Genetic epidemiology 35: 809–822. 9. Vasseur F, Helbecque N, Dina C, Lobbens S, Delannoy V, et al. (2002) Single- nucleotide polymorphism haplotypes in the both proximal promoter and exon 3 of the APM1 gene modulate adipocyte-secreted adiponectin hormone levels and contribute to the genetic risk for type 2 diabetes in French Caucasians. Hum Mol Genet 11: 2607–2614. 25. Nica AC, Parts L, Glass D, Nisbet J, Barrett A, et al. (2011) The Architecture of Gene Regulatory Variation across Multiple Human Tissues: The MuTHER Study. PLoS Genet 7: e1002003. doi:10.1371/journal.pgen.1002003. 26. Nyholt DR (2004) A simple correction for multiple testing for single-nucleotide polymorphisms in linkage disequilibrium with each other. Am J Hum Genet 74: 765–769. 10. Cesari M, Narkiewicz K, De Toni R, Aldighieri E, Williams CJ, et al. (2007) Heritability of plasma adiponectin levels and body mass index in twins. J Clin Endocrinol Metab 92: 3082–3088. 27. Heid IM, Jackson AU, Randall JC, Winkler TW, Qi L, et al. (2010) Meta- analysis identifies 13 new loci associated with waist-hip ratio and reveals sexual dimorphism in the genetic basis of fat distribution. Nat Genet 42: 949–960. 11. Liu PH, Jiang YD, Chen WJ, Chang CC, Lee TC, et al. (2008) Genetic and environmental influences on adiponectin, leptin, and BMI among adolescents in Taiwan: a multivariate twin/sibling analysis. Twin Res Hum Genet 11: 495–504. 28. Rung J, Cauchi S, Albrechtsen A, Shen L, Rocheleau G, et al. Adiponectin: Multiethnic Genome-Wide Meta-Analysis Adiponectin: Multiethnic Genome-Wide Meta-Analysis Borja, LS Adair, M Haun. Analyzed the data: Z Dastani, N Timpson, T Tanaka, KA Kapur, R Semple, X Yuan, P Henneman, A Isaacs, J Dupuis, J Grimsby, AK Manning, CT Liu, JRB Perry, AR Wood, F Kronenberg, IM Heid, TW Winkler, C Langenberg, C Fuchsberger, J Brody, L-P Lyytika¨inen, RA Scott, Y Liu, M Garcia, L Qi, FB Hu, M Beekman, I Meulenbelt, H-W Uh, JS Pankow, S Kanoni, A Bidulescu, SG Buxbaum, Y Wu, AP Morris, K Small, TM Teslovich, BA Oostra, AR Wood. Contributed reagents/materials/analysis tools: DM Evans, B St. Pourcain, N Sattar, CMT Greenwood, TD Spector, M Ladouceur, CV Dedoussis, JB Meigs, F Kronenberg, IM Heid, TW Winkler, C Langenberg, AA Hicks, PP Pramstaller, BM Psaty, I Chen, RP Tracy, A Jula, B-M Loo, PE Slagboom, S Bandinelli, I Meulenbelt, H-W Uh, M Kloppenburg, DJ Couper, H-E Wichmann, B Paulweber, GV Dedoussis, P Deloukas, S Redline, T Johnson, BA Oostra, GD Smith. Wrote the paper: Z Dastani, JB Richards, CMT Greenwood, N Timpson, DM Waterworth, VE Mooser, CV Dedoussis, A Isaacs, JB Meigs, JC Florez, J Dupuis, JRB Perry, M-F Hivert, F Kronenberg, IM Heid, T Lehtima¨ki, M Perola, RA Scott, C Langenberg, RJF Loos, T Harris, DJ Couper, CM Ballantyne, BB Duncan, H-E Wichmann, KL Mohlke, AP Morris, K Small, R Semple. Undertook meta-analysis: Z Dastani, J Dupuis. Did multi-ethnic analysis: AP Morris. Did expression analysis: K Small. Wrote the first draft of the manuscript: Z Dastani. 115 Sage Bionetworks, Seattle, Washington 98109, USA. 116 Division of Community Health Sciences, St George’s University of London, London SW17 0RE, UK. 117 Department of Cardiovascular Sciences, University of Leicester, NIHR Biomedical Research Unit in Cardiovascular Disease, Glenfield Hospital, Leicester LE3 9QP, UK. Author Contributions Conceived and designed the experiments: Z Dastani, JB Richards, DM Waterworth, VE Mooser, C van Duijn, A Isaacs, JB Meigs, JC Florez, M-F Hivert, F Kronenberg, IM Heid, PP Pramstaller, JR Kizer, DS Siscovick, BM Psaty, O Raitakari, T Lehtima¨ki, JG Eriksson, M Perola, V Salomaa, RJF Loos, NJ Wareham, T Harris, L Qi, FB Hu, I Meulenbelt, M Kloppenburg, CM Ballantyne, H-E Wichmann, B Paulweber, L Kendenko, JG Wilson, A Bidulescu, S Musani, SG Buxbaum, S Redline, MA Allison, KL Mohlke, AP Morris, K Small. Performed the experiments: L Ferrucci, JM Egan, OD Carlson, P Vollenweider, C Lamina, NJ Wareham, J Chambers, J Kooner, R Frants, K Willems-vanDijk, SM Willems, TM Frayling, S Bo¨hringer, F Kronenberg, AA Hicks, BM Psaty, I Chen, RP Tracy, M Ka¨ho¨nen, J Viikari, K Kristiansson, M-L Nuotio, K Lohman, A Kanaya, PE Slagboom, M Beekman, D van Heemst, I Meulenbelt, M Kloppenburg, CM Ballantyne, N Klopp, S Coassin, E Katsareli, A Bidulescu, MA Allison, LJ Rasmussen-Torvik, X Guo, JB Meulenbelt, M Kloppenburg, CM Ballantyne, N Klopp, S Coassin, E Katsareli, A Bidulescu, MA Allison, LJ Rasmussen-Torvik, X Guo, JB Katsareli, A Bidulescu, MA Allison, LJ Rasmussen-Torvik, X Guo, JB Katsareli, A Bidulescu, MA Allison, LJ Rasmussen-Torvik, X Guo, JB Adiponectin: Multiethnic Genome-Wide Meta-Analysis Loo Lin School of Medicine, National University of Singapore, Singapore 117597, Republic of Singapore. 73 Oxford Centre for Diabetes, Endocrinology and Medicine, University of Oxford, Churchill Hospital, Oxford OX3 7LJ, UK. 107 Office of Population Studies Foundation, University of San Carlos, Cebu City 6000, Philippines. 74 Oxford NIHR Biomedical Research Centre, Churchill Hospital, Oxford OX3 7LJ, UK. 108 Department of Anthropology, Northwestern University, Evanston, Illinois 60208, USA. 75 Avon Longitudinal Study of Parents and Children, University of Bristol, Bristol BS8 2BN, UK. 109 Department of Nutrition, Carolina Population Center, University of 109 Department of Nutrition, Carolina Population Center, University of North Carolina, Chapel Hill, North Carolina 27516, USA. 109 Department of Nutrition, Carolina Population Center, University of North Carolina, Chapel Hill, North Carolina 27516, USA. 76 Institute of Applied Genomics, via Linussio 51, 33100 Udine, Italy. 110 Department of Medicine, University of Mississippi Medical Center, Jackson, Mississippi 39126, USA. 77 Cardiovascular Epidemiology and Genetics, Institut Municipal d’Investigacio Medica, 08003 Barcelona, Spain. 111 Children’s Hospital Oakland Research Institute, Oakland, Califor- nia 94609, USA. 78 Transplantation Laboratory, Haartman Institute, University of Helsinki, FI-00014 Helsinki, Finland. 79 Montreal Heart Institute (Research Center), Universite´ de Montre´al, Montre´al, Que´bec H1T 1C8, Canada. 112 Department of Cardiovascular Epidemiology and Population Genetics, Centro Nacional de Investigaciones Cardiovasculares, 28029 Madrid, Spain. 80 Laboratory of Epidemiology, Demography, and Biometry, National Institute on Aging, National Institutes of Health, Bethesda, Maryland 20892, USA. 113 Nutrition and Genomics Laboratory, Jean Mayer United States Department of Agriculture Human Nutrition Research Center on Aging at Tufts University, Boston, Massachusetts 02111, USA. 81 Science Center, Tampere University Hospital, FI-33521 Tampere, Finland. 114 Clinical Pharmacology and Barts and The London Genome Centre, William Harvey Research Institute, Barts and The London School of Medicine and Dentistry, Queen Mary University of London, Charterhouse Square, London EC1M 6BQ, UK. 82 Institute of Regional Health Research and the Danish Twin Registry, Institute of Public Health, University of Southern Denmark, JBWinsløws Vej 9B, DK-5000, Odense, Denmark. PLoS Genetics \| www.plosgenetics.org March 2012 \| Volume 8 \| Issue 3 \| e1002607 March 2012 \| Volume 8 \| Issue 3 \| e1002607 21 Adiponectin: Multiethnic Genome-Wide Meta-Analysis 34. Hu Y, Sun CY, Huang J, Hong L, Zhang L, et al. (2007) Antimyeloma effects of resveratrol through inhibition of angiogenesis. Chin Med J (Engl) 120: 1672–1677. 45. Wise A, Foord SM, Fraser NJ, Barnes AA, Elshourbagy N, et al. (2003) Molecular identification of high and low affinity receptors for nicotinic acid. J Biol Chem 278: 9869–9874. J 46. Westphal S, Borucki K, Taneva E, Makarova R, Luley C (2006) Adipokines and treatment with niacin. Metabolism 55: 1283–1285. 35. Szkudelski T, Szkudelska K (2011) Anti-diabetic effects of resveratrol. Ann N Y Acad Sci 1215: 34–39. 47. Plaisance EP, Lukasova M, Offermanns S, Zhang Y, Cao G, et al. (2009) Niacin stimulates adiponectin secretion through the GPR109A receptor. Am J Physiol Endocrinol Metab 296: E549–558. 36. Ahn J, Lee H, Kim S, Ha T (2007) Resveratrol inhibits TNF-alpha-induced changes of adipokines in 3T3-L1 adipocytes. Biochem Biophys Res Commun 364: 972–977. 48. Edmondson AC, Braund PS, Stylianou IM, Khera AV, Nelson CP, et al. (2011) Dense Genotyping of Candidate Gene Loci Identifies Variants Associated with High-Density Lipoprotein Cholesterol. Circ Cardiovasc Genet. 37. Kang L, Heng W, Yuan A, Baolin L, Fang H (2010) Resveratrol modulates adipokine expression and improves insulin sensitivity in adipocytes: Relative to inhibition of inflammatory responses. Biochimie 92: 789–796. High-Density Lipoprotein Cholesterol. Circ Cardiovasc Genet. 49. Weyer C, Funahashi T, Tanaka S, Hotta K, Matsuzawa Y, et al. (2001) Hypoadiponectinemia in obesity and type 2 diabetes: close association with insulin resistance and hyperinsulinemia. J Clin Endocrinol Metab 86: 1930–1935. 38. Kiss-Toth E, Bagstaff SM, Sung HY, Jozsa V, Dempsey C, et al. (2004) Human tribbles, a protein family controlling mitogen-activated protein kinase cascades. J Biol Chem 279: 42703–42708. J 39. Sung HY, Guan H, Czibula A, King AR, Eder K, et al. (2007) Human tribbles-1 controls proliferation and chemotaxis of smooth muscle cells via MAPK signaling pathways. J Biol Chem 282: 18379–18387. 50. Lindsay RS, Funahashi T, Hanson RL, Matsuzawa Y, Tanaka S, et al. (2002) Adiponectin and development of type 2 diabetes in the Pima Indian population. The Lancet 360: 57–58. 40. Waterworth DM, Ricketts SL, Song K, Chen L, Zhao JH, et al. (2010) Genetic variants influencing circulating lipid levels and risk of coronary artery disease. Arterioscler Thromb Vasc Biol 30: 2264–2276. 51. Keating BJ, Tischfield S, Murray SS, Bhangale T, Price TS, et al. References (2009) Genetic variant near IRS1 is associated with type 2 diabetes, insulin resistance and hyperinsulinemia. Nat Genet 41: 1110–1115. 12. Menzaghi C, Trischitta V, Doria A (2007) Genetic influences of adiponectin on insulin resistance, type 2 diabetes, and cardiovascular disease. Diabetes 56: 1198–1209. 29. Takeuchi F, Serizawa M, Yamamoto K, Fujisawa T, Nakashima E, et al. (2009) Confirmation of multiple risk Loci and genetic impacts by a genome-wide association study of type 2 diabetes in the Japanese population. Diabetes 58: 1690–1699. 13. Hivert MF, Manning AK, McAteer JB, Florez JC, Dupuis J, et al. (2008) Common variants in the adiponectin gene (ADIPOQ) associated with plasma adiponectin levels, type 2 diabetes, and diabetes-related quantitative traits: the Framingham Offspring Study. Diabetes 57: 3353–3359. 30. Adachi H, Tsujimoto M (2002) FEEL-1, a novel scavenger receptor with in vitro bacteria-binding and angiogenesis-modulating activities. J Biol Chem 277: 34264–34270. Framingham Offspring Study. Diabetes 57: 3353–3359. 14. Ling H, Waterworth DM, Stirnadel HA, Pollin TI, Barter PJ, et al. (2009) Genome-wide Linkage and Association Analyses to Identify Genes Influencing Adiponectin Levels: The GEMS Study. Obesity (Silver Spring). 31. Gray SL, Vidal-Puig AJ (2007) Adipose tissue expandability in the maintenance of metabolic homeostasis. Nutr Rev 65: S7–12. 15. Heid IM, Henneman P, Hicks A, Coassin S, Winkler T, et al. (2010) Clear detection of ADIPOQ locus as the major gene for plasma adiponectin: results of genome-wide association analyses including 4659 European individuals. Atherosclerosis 208: 412–420. 32. Eremina V, Baelde HJ, Quaggin SE (2007) Role of the VEGF–a signaling pathway in the glomerulus: evidence for crosstalk between components of the glomerular filtration barrier. Nephron Physiol 106: p32–37. 16. Richards JB, Waterworth D, O’Rahilly S, Hivert MF, Loos RJ, et al. (2009) A genome-wide association study reveals variants in ARL15 that influence adiponectin levels. PLoS Genet 5: e1000768. doi:10.1371/journal.pgen.1000768. 33. Buraczynska M, Ksiazek P, Baranowicz-Gaszczyk I, Jozwiak L (2007) Association of the VEGF gene polymorphism with diabetic retinopathy in type 2 diabetes patients. Nephrol Dial Transplant 22: 827–832. PLoS Genetics \| www.plosgenetics.org March 2012 \| Volume 8 \| Issue 3 \| e1002607 March 2012 \| Volume 8 \| Issue 3 \| e1002607 22 Adiponectin: Multiethnic Genome-Wide Meta-Analysis Adiponectin: Multiethnic Genome-Wide Meta-Analysis (2008) Concept, design and implementation of a cardiovascular gene-centric 50 k SNP array for large-scale genomic association studies. PLoS ONE 3: e3583. doi:10.1371/journal.pone.0003583. 41. Park MH, Kim N, Lee JY, Park HY (2011) Genetic loci associated with lipid concentrations and cardiovascular risk factors in the Korean population. J Med Genet 48: 10–15. j p 52. Price AL, Patterson NJ, Plenge RM, Weinblatt ME, Shadick NA, et al. (2006) Principal components analysis corrects for stratification in genome-wide association studies. Nat Genet 38: 904–909. 42. Chasman DI, Pare G, Mora S, Hopewell JC, Peloso G, et al. (2009) Forty-three loci associated with plasma lipoprotein size, concentration, and cholesterol content in genome-wide analysis. PLoS Genet 5: e1000730. doi:10.1371/journal.pgen. 1000730. 53. Magi R, Morris AP (2010) GWAMA: software for genome-wide association meta-analysis. BMC Bioinformatics 11: 288. 54. Higgins JP, Thompson SG, Deeks JJ, Altman DG (2003) Measuring inconsistency in meta-analyses. Bmj 327: 557–560. 43. Willer CJ, Sanna S, Jackson AU, Scuteri A, Bonnycastle LL, et al. (2008) Newly identified loci that influence lipid concentrations and risk of coronary artery disease. Nat Genet 40: 161–169. 55. Theodoraki EV, Nikopensius T, Suhorutsenko J, Peppes V, Fili P, et al. (2010) Fibrinogen beta variants confer protection against coronary artery disease in a Greek case-control study. BMC Med Genet 11: 28. 44. Ivanov D, Philippova M, Antropova J, Gubaeva F, Iljinskaya O, et al. (2001) Expression of cell adhesion molecule T-cadherin in the human vasculature. Histochemistry and cell biology 115: 231–242. 56. Storey JD, Tibshirani R (2003) Statistical significance for genomewide studies. Proc Natl Acad Sci U S A 100: 9440–9445. PLoS Genetics \| www.plosgenetics.org March 2012 \| Volume 8 \| Issue 3 \| e1002607 PLoS Genetics \| www.plosgenetics.org 23
https://openalex.org/W3111226947	https://www.researchsquare.com/article/rs-101451/v1.pdf?c=1604637447000	English	null	Accuracy of Different Metabolic Syndrome Definitions for Detecting Increased Carotid Intima-media Thickness in Peruvian Type 2 Diabetes Mellitus Subjects without Cardiovascular Disease.	Research Square (Research Square)	2,020	cc-by	5,219	Accuracy of Different Metabolic Syndrome Definitions for Detecting Increased Carotid Intima- media Thickness in Peruvian Type 2 Diabetes Mellitus Subjects without Cardiovascular Disease. Alberto Teruya-Gibu (  a.teruya.gibu@gmail.com ) Abstract Backrgound: To estimate the accuracy of metabolic syndrome definitions for detecting Carotid Intima- Media thickness (CIMT) in Peruvian type 2 diabetes mellitus (T2DM) subjects without cardiovascular events. Methods: We performed a cross-sectional evaluation of T2DM subjects from the endocrinology service of two reference health centers without stroke or coronary disease. . Bilateral carotid intima-media thickness was measure by B mode ultrasound one only operator. We performed four definitions of Metabolic Syndrome; 1. Cholesterol Education Program Adult Treatment Panel III (ATP III), 2. Harmonized criteria; 3. Gurka's Metabolic syndrome severity score (MSSS) formula. 4. Non-glucose modified MSSS. We calculated the area under the receiver operator curve (AUROC) for detecting increased CIMT (≥ 0.86 mm) between metabolic syndrome definitions. Results: We included 184 subjects with T2DM, 29% were men with a mean age of 61.5 ± 10.5 years old. Median diabetes time was 10.8 years (IQR 4.8 to 19.4), and 26.8% achieved HbA1c goals. Non-glucose modified MSSS was the only definition significantly correlated with elevated CIMT (r = 0.19; p<0.01), and it showed the best accuracy for predicting elevated CIMT (AUC: 0.61, CI95%: 0.52-0.70). Adjusting for age, sex, and HbA1c, each point increase in the non-glucose MSSS Z score, the risk of altered CIMT increases by 59% RP 1.59 (CI 95% 1.09 - 2.35; p=0.017). Conclusions: Non-glucose modified MSSS had a weak accuracy for elevated CIMT and was the best compared to ATP III, harmonized, and original MSSS. Further research on no MS-factors is required to predict better elevated CIMT. Original investigation License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Page 1/16 Page 1/16 1. Background Cardiovascular disease (CVD) is the leading cause of mortality and morbidity in subjects with T2DM worldwide, and physicians call coronary equivalent to that subject with diabetes mellitus without coronary disease due to the high burden of disease. Subclinical cardiovascular disease (SCVD) determination, through Carotid intima-media (CIMT), coronary artery calcium (CAC), and presence of carotid plaque (CP), is associated with overt CVD, and this structural surrogate helps establish preventive strategies in primary care. Although CIMT has a lower predictive ability to detect overt CVD than CAC and CP (1), it might be considered the earliest structural change compared to the other subclinical parameters. Metabolic syndrome (MS) is the aggregation of factors that include three of five components: waist circumference, glucose, HDL-cholesterol, triglycerides, and blood pressure (2). There are multiple definitions, varying by the cut-off points or one factor's predominance over another (2–4). Both T2DM and MS are associated with an increased prevalence of SCVD and overt CVD (5). Furthermore, the reduction or aggregation of factors does not necessarily determine lower o higher cardiometabolic risk. Gurka et al.(7) designed a continuous score for the severity of metabolic syndrome Page 2/16 Page 2/16 Page 2/16 specific to sex and ethnic group. According to factor analysis models, the metabolic syndrome severity score (MSSS) offers the advantage of converting those dichotomous variables (present or absent) into continuous numerical variables. This severity score has shown a good correlation with risk markers such as insulin, adiponectin (8), and predictive value for cardiovascular risk in T2DM subjects (9), and chronic kidney disease (10). specific to sex and ethnic group. According to factor analysis models, the metabolic syndrome severity score (MSSS) offers the advantage of converting those dichotomous variables (present or absent) into continuous numerical variables. This severity score has shown a good correlation with risk markers such as insulin, adiponectin (8), and predictive value for cardiovascular risk in T2DM subjects (9), and chronic kidney disease (10). For these reasons, it is critical to determine whether MSSS can predict the SCVD through the CIMT measurement and to clarify if this method could be used as an early tool for preclinical CVD. The absence of SCVD seems to attenuate the risk of overt CVD in subjects with T2DM and MS (6). 2.1 Study design. The study was conducted as a cross-sectional design. All subjects gave written informed consent, and the study procedure followed the Helsinsky Declaration's agreements and was approved by the institutional ethics committee. 1. Background Thus, we propose evaluating the accuracy of MSSS and other metabolic syndrome definitions to detect increased carotid intima-media thickness in Peruvian Type 2 diabetes mellitus subjects without CVD attending to a healthcare network social security system. 2.2 Patient A total of 184 subjects with Diabetes mellitus were recruited from the outpatient clinic of the Edgardo Rebagliati Martins Hospital (HNERM) and the Diabetes and Hypertension Center (CEDHI) of the Social security Health System between January 2019 and January 2020. We included subjects older than 20 years and diagnosis of DM based on the ADA criteria. Those patients with a history of coronary artery disease, stroke, peripheral revascularization, or major amputation due to peripheral arterial disease were excluded. Likewise, those with DM1, LADA, NYHA III or IV heart failure, Child B or C liver disease, glomerular filtration rate < 60 ml/min, acute or previous disease up to a range of 6 months, surgery & radiotherapy at the cervical level and use of immunosuppressants or corticosteroids. 2.4 Measurement of MICT We measured CIMT by a single trained physician independent of data collection and processing using a Toshiba Nemio XG ultrasound scanner. We used a B-mode carotid ultrasound with a 7.5 MHz linear transducer to assess the CIMT with standard insonation parameters. Patients were examined supine with the neck hyperextended and identified three segments in the anterior and posterior plane: 1 cm distal to the common carotid proximal to the carotid bifurcation, the bifurcation itself, 1 cm proximal to the internal carotid artery. We obtained the maximum average CIMT value from the far wall of both the right and left common carotid arteries and excluded CIMT ≥ 1.5 mm 2.3 Clinical and biochemical parameters 2.3 Clinical and biochemical parameters Independent, trained medical personnel evaluated and collected data patients. Vital functions and anthropometric measurements were taken according to the established protocol. Waist circumference was measured with a tape scale according to the International Diabetes Federation (IDF) recommendations. A questionnaire was filled out directly from the patient regarding pathological history, recommendations. A questionnaire was filled out directly from the patient regarding pathological history, lifestyle, comorbidities, and pharmacological treatment. We obtained fasting blood samples for biochemical analyzes that included glucose, HbA1c, and lipid profile. The process was according to the central laboratory's routine method. Page 3/16 We defined Metabolic syndrome according to ATP III and harmonized criteria and use the cut-of for abdominal perimeter obtained by Aschner in the harmonized MS criteria. We calculated MSSS (7) and non-glucose MSSS, according to Gurka et al (9), as Z-score through a spreadsheet. We defined Metabolic syndrome according to ATP III and harmonized criteria and use the cut-of for abdominal perimeter obtained by Aschner in the harmonized MS criteria. We calculated MSSS (7) and non-glucose MSSS, according to Gurka et al (9), as Z-score through a spreadsheet. 2.5 Statistical analysis We presented categorical data as absolute and relative frequencies and consecutive data as mean and standard deviation or medians with interquartile range according to normal distribution. We described CIMT according to sex and age. Also, we show the correlation between each SM definition with CIMT through Pearson's coefficient and its scatter charts. A value of p < 0.05 was considered significant. We performed area Under the Receiver Operator Curve (AUROC) of each MS definition to predict CIMT ≥ 0.86 mm, and we calculated sensitivity and specificity according to the optimal cut-off value for each MS definitions. We made a bivariate analysis for related and no related MS factors for CIMT and a regression model A adjusted to age and sex, and a regression model B adjusted to age, sex, and HbA1c. We made a bivariate analysis for related and no related MS factors for CIMT and a regression model A adjusted to age and sex, and a regression model B adjusted to age, sex, and HbA1c. 3. Results This study included 184 subjects with T2DM 29% were men with a mean age of 61.5 ± 10.5 years old. Median diabetes time was 10.8 years (IQR 4.8 to 19.4) with a median HbA1c of 8.1% (IQR 6.9 to 9.85), and only 26.8% achieved HbA1c goals. Smoker and sedentary lifestyle prevalences were 18.4% and 73.9%, respectively. Antihypertensive drugs were used by 48.4% for used drugs, statins by 46.7%, and acetylsalicylic acid by 27.7%. Obesity prevalence was 42.3%, with a median waist circumference of 98 cm (IQR 89.7–104). (Table 1) Page 4/16 Tabla 1 Clinical-epidemiological characteristics in subjects with diabetes mellitus without cardiovascular disease at the Edgardo Rebagliati Hospital. 3. Results Total N = 184 Female N=131 Male N=53 Epidemiological Age, mean (SD) 61.5 ± 10.5 60.4 ± 10.6 64.4 ± 9.9 Diabetes time, median (IQR) 10.8 (4.8 to 19.4) 10.5 (4.7 to 17.7) 11.8 (5.5 to 22.2) Smoker, n(%) 34 (18.4) 18 (13.7) 16 (30.2) Sedentary lifestyle, n(%) 136 (73.9) 31 (23.7) 17 (32.1) Drugs Antihipertensive, n(%) 89 (48.4) 60 (45.8) 29 (54.7) Statins, n(%) 86 (46.7) 61 (45.6) 25 (47.2) Acetylsalicilic acid, n(%) 50 (27.7) 30 (22.9) 20 (37.7) Metformin, n(%) 166 (90.2) 116 (88.6) 50 (94.3) Glyburide, n(%) 34 (18.5) 24 (18.3) 10 (18.9) SGLT2 inhibitor, n(%) 1 (0.5) 0 (0.0) 1 (1.9) GLP1 agonist, n(%) 1 (0.5) 1 (0.0) 0 (0.0) Insulin, n(%) 69 (37.5) 50 (38.2) 19 (35.9) Clinical Body mass index (kg/m2), median (IQR) 28.8 (25.7 to 32.2) 28.9 (25.8 to 32.2) 28.6 (25.2 to 32.4) BMI >30 kg/m2, n(%) 72 (42.3) 51 (42.9) 21 (41.2) Waist circumference (cm), median (IQR) 98 (89.75 to 104) 100 (91 to 105) 97 (88 to 104) Blood pressure >130/85, n(%) 83 (45.1) 63 (48.1) 20 (37.7) Glycemia (mg/dl), median (IQR) 135 (106 to 193) 143 (106 to 213) 131 (106 to 153) Triglycerides, median (IQR) 163 (107.5 to 224) 170 (118 to 244) 138 (101 to 196) Total Cholesterol (mg/dl) 178.5 (151 to 189 (161 to 158 (133 to 195) LDL cholesterol (mg/dl), median (IQR) 110 (89 to 139) 116.5 (94 to 142) 101 (75 to 123) Glomerular filtration (ml/min/24 h), median (IQR) 105.3 (87.6 to 116.7) 108.3 (96 to 117) 94 (85 to 106) A1C (%), median (IQR) 8.1 (6.9 to 9.85) 8.2 (6.9 to 10.1) 8 (6.85 to 9.4) A1c < 7% 48 (26.7) 34 (26.6) 14 (26.9) Metabolic syndrome ATP III criteria, n(%) 135 (73.4) 86 (65.7) 49 (92.5) Harmonizing criteria, n(%) 145 (78.8) 96 (73.3) 49 (92.5) MSSS Z- score MSSS, median (IQR) 1.34 (0.58 to 2.71) 1.78 (0.70 to 3.28) 0.84 (0.46 to 1.48) Z-score MSSS ≥ 0 , n(%) 110 (59.8) 119 (90.8) 47 (88.7) Z-score non-glucose MSSS, median (IQR) -1.08 (-1.54 to -0.54) -1.30 (-1.66 to -0.89) -0.43 (-0.91 to -0.19) Z-score non-glucose MSSS ≥ 0 , n(%) 8 (4.3) 1 (0.7) 7 (13.2) CIMT (mm), Median (IQR) 0.8 (0.7 to 0.9) 0.75 (0.7 to 0.85) 0.85 (0.7 to 0.95) ≥ 0.86 mm, n(%) 54 (29.4) 32 (24.4) 22 (41.5) IQR Interquartile range, SD Standard deviation, HDL high-density lipoprotein, LDL low-density lipoprotein, SGLT2 Sodium-glucose cotransporter -2; GLP1 glucagon-like peptide-1; ATP III: Adult treatment Panel III. 3. Results CIMT: carotid medial intima thickness IQR Interquartile range, SD Standard deviation, HDL high-density lipoprotein, LDL low-density lipoprotein, SGLT2 Sodium-glucose cotransporter -2; GLP1 glucagon-like peptide-1; ATP III: Adult treatment Panel III. CIMT: carotid medial intima thickness Metabolic syndrome for ATP III criteria and Harmonized criteria were 73.4% and 78.8%, respectively. The Median Z-score of the MSSS was 1.34 (IQR 0.58 to 2.71), and the median Z-score of the non-glucose MSSS was − 1.08 (IQR − 1.54 to -0.54). The prevalence of MSSS and non-glucose MSSS above Z-score ≥ 0 were 59.8% and 4.3%, respectively. (Table 1 ) Median CIMT was 0.8 mm (0.7–0.9), and the prevalence of elevated CIMT, defined by Rizzo (≥ 0.86 mm), was 29.4%. CIMT tends to increase according to age in both sexes (Fig. 1) Only non-glucose MSSS showed a correlation with CIMT. (r = 0.19, p < 0.01). (Fig. 2). Non-glucose MSSS obtained the best AUROC 0.61 (0.52–0.70) compared with the other definitions. (Fig. 3). MSSS and Harmonized SM criteria included 0.5 in their confidence interval. A Z-score of -1.0 was the best cut-of Page 6/16 Page 6/16 point of non-glucose MSSS to predict elevated CIMT. This value had 57.4% and 60% of sensitivity and specificity, respectively. (Table 2) Table 2 The area under the ROC of the Metabolic Syndrome Severity Score and Classic Definitions of Metabolic Syndrome for carotid medial intima thickness in Subjects with Diabetes Mellitus without Cardiovascular Disease at the Edgardo Rebagliati Martins Hospital. AUC ROC Optimal cut- off Sensitivity Specificity Metabolic síndrome ATPIII MS criterio 0.59 (0.51 to 0.68) 4 44.4% 71.5% Harmonizing MS criteria 0.56 (0.47 to 0.65) 4 53.7% 59.2% Metabolic Syndrome Severity Score MSSS 0.48 (0.39 to 0.57) 0.30 87.0% 20% Non-glucose MSSS 0.61 (0.52 to 0.70) -1.00 57.4% 60% MS ATP III: 3 of 5 criteria according to Adult Treatment Panel III. MS Harmonizing criteria: : 3 of 5 criteria according to Harmonized Consensus. MS ATP III: 3 of 5 criteria according to Adult Treatment Panel III. MS Harmonizing criteria: : 3 of 5 criteria according to Harmonized Consensus. Only systolic pressure was associated with CIMT (PR 1.01 CI95% 1.01 to 1.03; p < 0.000) of the metabolic syndrome factors. Age, male sex, and the use of hypertensive drugs also showed association with CIMT (Table 3). 3. Results Page 7/16 Page 7/16 Table 3 Table 3 Table 3 associated with elevated carotid medial intima thickness in subjects with t cardiovascular disease at the Edgardo Rebagliati Martins Hospital. The prevalence rate of factors associated with elevated carotid medial intima thickness in subjects with diabetes mellitus without cardiovascular disease at the Edgardo Rebagliati Martins Hospital. diabetes mellitus without cardiovascular disease at the Edgardo Rebagliati Martins Hospital. PR CI 95% P Metabolic Syndrome factors Glycemia (mg/dl) 0.99 0.99–1.00 0.52 Waist circunference (cm) 1.00 0.98–1.02 0.56 HDL colesterol (mg/dl) 0.98 0.96–1.01 0.12 Triglycerides (mg/dl) 1.00 0.99–1.00 0.33 Systolic blood pressure (mmHg) 1.01 1.01–1.03 0.000 Other factors Age (years old) 1.06 1.03–1.09 0.000 Male sex 1.69 1.09–2.63 0.018 Smoker 1.12 0.65–1.95 0.66 Diabetes time (years) 1.01 0.99–1.03 0.12 Body mass Index (kg/m2) 0.99 0.95–1.05 0.99 A1c (%) 0.95 0.86–1.06 0.41 Total cholesterol (mg/dl) 1.00 0.99–1.01 0.95 LDL cholesterol 0.99 0.99–1.00 0.86 Hypertension drug 2.13 1.31–3.47 0.002 Statins drug 0.72 0.45–1.15 0.17 PR: Prevalence rate. CI95% Confidence Interval 95. aUnivariate regression using generalized linear model with robust variance and poisson distribution. A value of 0.86 mm is considered as anormal carotide medial intima thickness. aUnivariate regression using generalized linear model with robust variance and poisson distribution. A value of 0.86 mm is considered as anormal carotide medial intima thickness. In the crude analysis, each point increase in the non-glucose MSSS Z score, the risk of altered CIMT increases by 55% PR 1.55 (CI95% 1.05 to 2.28). Adjusting for age, sex, and HbA1c, MSSS also showed association PR 1.59 (CI 95% 1.09 to 2.35; p = 0.017). (Table 4) Page 8/16 Table 4 Multivariate analysis of the severity score of metabolic syndrome without glycemia for the thickness of the medial carotid intima-media in subjects with diabetes mellitus without cardiovascular disease at Hospital Edgardo Rebagliati. Bivariate Multivariate model 1a Multivariate model 2b PR Ci 95% P PR Ci 95% p PR Ci 95% P Non-glucose MSSS (Z score) 1.55 1.05– 2.28 0.01 1.54 1.07– 2.22 0.02 1.59 1.09– 2.35 0.017 Age (years old) 1.06 1.03– 1.09 0.000 1.06 1.04– 1.09 0.000 1.06 1.04– 1.09 0.000 Male 1.69 1.09– 2.63 0.018 0.95 0.57– 1.57 0.84 0.91 0.55– 1.50 0.71 A1c (%) 0.95 0.86– 1.06 0.41 0.96 0.86– 1.07 0.53 PR: Prevalence rate. CI95% Confidence Interval 95%. MSSS: Metabolic síndrome severity score. aMultivariate regression using a generalized linear model with robust variance and Poisson distribution. 3. Results A value of 0.86 mm is considered as abnormal carotid medial intima thickness. aModel 1: Adjusted to age and sex. bModel 2: Adjusted to age, sex, and A1c. Table 4 Table 4 Multivariate analysis of the severity score of metabolic syndrome without glycemia for the thickness of the medial carotid intima-media in subjects with diabetes mellitus without cardiovascular disease at Hospital Edgardo Rebagliati 4. Discussion MSSS is a prediction model developed from the National Health and Nutrition Examination Survey (NHANES) database. It must be convenient to validate outcome studies in different populations such as Latin American subjects to answer whether this prediction model can predict subclinical atherosclerosis as an early marker of CV events. In the present study, only the non-glucose MSSS definition showed the best diagnostic discrimination for elevated CIMT and presented a weak performance. The original MSSS did not show any discrimination in this population. MSSS performed a significant association with future CVD among type 2 DM subjects from the Atherosclerosis Risk in Communities study (ARIC) (9). This cohort study found that MSSS and non- glucose MSSS at baseline were associated with incident CVD HR 1.29 ( 95% CI 1.21–1.39; p < 0.001).and HR 1.42 (95% CI 1.24–1.62; p < 0.001), respectively. This outcome study explains our better predictive ability of non-glucose MSSS to detect elevated CIMT than the original MSSS. We did not find studies related to the association between MSSS and subclinical atherosclerosis as a surrogate marker of cardiovascular risk. Gurka et al. showed that MSSS has an excellent predictive performance of cardiovascular risk if glucose is not considered (9) or even when fasting values is replaced for non-fasting values into this model (11). The independent association of non-glucose- Page 9/16 Page 9/16 modified MSSS with cardiovascular risk might be explained for a possible weakness of glycemic status for detecting CIMT. Other risk factors as age, male sex, high-density-lipoprotein cholesterol, and total cholesterol also influence their development (12). However, glycemic parameters as post-prandial glucose (13) and glycemic variability are associated as independent risk factors for CIMT (14). The oxidative stress related to glucose fluctuations could explain this association. So glycemic control is an essential therapeutic intervention to decrease the subclinical atherosclerosis process (15). DM and MS are associated with atherosclerotic findings as coronary calcium content, abdominal aortic calcification, elevated CIMT, and abnormal ankle-brachial index (ABI). These parameters, alone or combined, with greater cardiovascular risk rate (16). CIMT is one of the earliest surrogates of subclinical cardiovascular disease, and its measurement is easy to use even in-office settings, although its interpretation is operator dependent. Additionally, identifying biomarkers with the ability to predict early structural change, such as CIMT, should be an essential tool for modifying the atherosclerosis process's natural history as a primary prevention strategy for CVD (17–21). 4. Discussion As a weakness, the study stopped due to the SARS CoV 2 pandemic, and the original sample size did not complete, although the trend would not change in light of the results. Another limitation was that cardiovascular disease background was referred for the subjects without any assessment to rule out subclinical coronary disease. We selected the subjects by non-random sampling from a reference hospital and a primary care center and did not extrapolate beyond the study population. Finally, the small number of males and a wide age range could miss estimating the true prevalence of altered CIMT. As a strength, we evaluated the carotid intima-media thickness with the appropriate technique by on a high- resolution ultrasound and a trained operator. 5. Conclusions Our results show that non-glucose modified MSSS had the greatest accuracy for detecting CIMT compared with MS-ATP III, MS-harmonizing criteria, and original MSSS. Further studies are necessary to evaluate ultrasonography's clinical utility as an early intervention to reduce the risk of CIMT progression. Also, to evaluate the role of metabolic syndrome in the development of subclinical atherosclerosis. Ethics approval and consent to participate: This study have been performed in accordance with the Declaration of Helsinki and it had been approved by an institutional ethics committee of Edgardo Rebagliati Martins Hospital. President's committee: Dr. Julio César Alfaro Mantilla Committee address: Hospital Nacional Edgardo Rebagliati Martins. Avenida Edgardo Rebagliati 490, Jesús María, Lima., Perú. This study have been performed in accordance with the Declaration of Helsinki and it had been approved by an institutional ethics committee of Edgardo Rebagliati Martins Hospital. President's committee: Dr. Julio César Alfaro Mantilla Committee address: Hospital Nacional Edgardo Rebagliati Martins. Avenida Edgardo Rebagliati 490, Jesús María, Lima., Perú. Jesús María, Lima., Perú. Page 10/16 Consent for publication: Not applicable for that section. Page 10/16 Phone number: 511-2654901, annex: 3080 email: comitedeeticadhed@essalud.gob.pe. Informed consent to participate in the study wasobtained from participants. The informed consent document was approved by the institutional ethics committee. Authors' contributions: All authors designed the study and collected the data. ATG, DU, and LMT recollected information. MYA performed statistical analyses. ATG and MYA participated in the interpretation of the results. MYA and ATG performed in the writing of the manuscript. All authors approved the final version of the manuscript. Availability of data and materials: The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request. Acknowledgements: Not applicable for that section. Not applicable for that section. Funding: This study was funded by the Institute for Health Technology Assessment and Research - IETSI, EsSalud - Peru, through the 2018 Kaelin Award (Resolution No. 71 and 77-IETSI-ESSALUD-2018). Competing interests: The authors declare to have no compelling interests with this article. References 1. Gepner AD, Young R, Delaney JA, Tattersall MC, Blaha MJ, Post WS, et al. Comparison of Coronary Artery Calcium Presence, Carotid Plaque Presence, and Carotid Intima-Media Thickness for Cardiovascular Disease Prediction in the Multi-Ethnic Study of Atherosclerosis. Circ Cardiovasc Imaging [Internet]. 2015 Jan;8(1). Available from: https://www.ahajournals.org/doi/10.1161/CIRCIMAGING.114.002262 2. Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. Executive Summary of the Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on 2. Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. Executive Summary of the Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Page 11/16 Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). JAMA J Am Med Assoc. 2001 May 16;285(19):2486–97. Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). JAMA J Am Med Assoc. 2001 May 16;285(19):2486–97. 3. Alberti KGMM, Eckel RH, Grundy SM, Zimmet PZ, Cleeman JI, Donato KA, et al. Harmonizing the Metabolic Syndrome: A Joint Interim Statement of the International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; and International Association for the Study of Obesity. Circulation. 2009 Oct 20;120(16):1640–5. 4. Grundy SM, Cleeman JI, Daniels SR, Donato KA, Eckel RH, Franklin BA, et al. Diagnosis and Management of the Metabolic Syndrome: An American Heart Association/National Heart, Lung, and Blood Institute Scientific Statement. Circulation. 2005 Oct 25;112(17):2735–52. 4. Grundy SM, Cleeman JI, Daniels SR, Donato KA, Eckel RH, Franklin BA, et al. Diagnosis and Management of the Metabolic Syndrome: An American Heart Association/National Heart, Lung, and Blood Institute Scientific Statement. Circulation. 2005 Oct 25;112(17):2735–52. 5. Chen K, Lindsey JB, Khera A, Lemos JAD, Ayers CR, Goyal A, et al. Independent associations between metabolic syndrome, diabetes mellitus and atherosclerosis: observations from the Dallas Heart Study. Diab Vasc Dis Res. 2008 Jun 1;5(2):96–101. 5. Chen K, Lindsey JB, Khera A, Lemos JAD, Ayers CR, Goyal A, et al. Independent associations between metabolic syndrome, diabetes mellitus and atherosclerosis: observations from the Dallas Heart Study. Diab Vasc Dis Res. 2008 Jun 1;5(2):96–101. 6. Ingelsson E, Sullivan LM, Murabito JM, Fox CS, Benjamin EJ, Polak JF, et al. Prevalence and prognostic impact of subclinical cardiovascular disease in individuals with the metabolic syndrome and diabetes. Clin Diabetol. 2007;8(8–9):315–29. 6. References Ingelsson E, Sullivan LM, Murabito JM, Fox CS, Benjamin EJ, Polak JF, et al. Prevalence and prognostic impact of subclinical cardiovascular disease in individuals with the metabolic syndrome and diabetes. Clin Diabetol. 2007;8(8–9):315–29. 7. Gurka MJ, Ice CL, Sun SS, DeBoer MD. A confirmatory factor analysis of the metabolic syndrome in adolescents: an examination of sex and racial/ethnic differences. Cardiovasc Diabetol. 2012;11(1):128. 7. Gurka MJ, Ice CL, Sun SS, DeBoer MD. A confirmatory factor analysis of the metabolic syndrome in adolescents: an examination of sex and racial/ethnic differences. Cardiovasc Diabetol. 2012;11(1):128. 8. DeBoer MD, Gurka MJ, Morrison JA, Woo JG. Inter-relationships between the severity of metabolic syndrome, insulin and adiponectin and their relationship to future type 2 diabetes and cardiovascular disease. Int J Obes. 2016 Sep;40(9):1353–9. 9. Gurka MJ, Guo Y, Filipp SL, DeBoer MD. Metabolic syndrome severity is significantly associated with future coronary heart disease in Type 2 diabetes. Cardiovasc Diabetol. 2018 Dec;17(1):17. 9. Gurka MJ, Guo Y, Filipp SL, DeBoer MD. Metabolic syndrome severity is significantly associated with future coronary heart disease in Type 2 diabetes. Cardiovasc Diabetol. 2018 Dec;17(1):17. 10. DeBoer MD, Filipp SL, Musani SK, Sims M, Okusa MD, Gurka M. Metabolic Syndrome Severity and Risk of CKD and Worsened GFR: The Jackson Heart Study. Kidney Blood Press Res. 2018;43(2):555– 67. 10. DeBoer MD, Filipp SL, Musani SK, Sims M, Okusa MD, Gurka M. Metabolic Syndrome Severity and Risk of CKD and Worsened GFR: The Jackson Heart Study. Kidney Blood Press Res. 2018;43(2):555– 67. 11. DeBoer MD, Filipp SL, Gurka MJ. Associations of a metabolic syndrome severity score with coronary heart disease and diabetes in fasting vs. non-fasting individuals. Nutr Metab Cardiovasc Dis. 2020 Jan;30(1):92–8. 12. Temelkova-Kurktschiev T, Koehler C, Schaper F, Fuecker K, Siegert G, Henkel E, et al. Relationship between fasting plasma glucose, atherosclerosis risk factors and carotid intima media thickness in non-diabetic individuals. Diabetologia. 1998 May 27;41(6):706–12. 13. Esposito K, Ciotola M, Carleo D, Schisano B, Sardelli L, Di Tommaso D, et al. Post-Meal Glucose Peaks at Home Associate with Carotid Intima-Media Thickness in Type 2 Diabetes. J Clin Endocrinol Metab. 2008 Apr;93(4):1345–50. 13. Esposito K, Ciotola M, Carleo D, Schisano B, Sardelli L, Di Tommaso D, et al. Post-Meal Glucose Peaks at Home Associate with Carotid Intima-Media Thickness in Type 2 Diabetes. J Clin Endocrinol Metab. 2008 Apr;93(4):1345–50. 14. References Liu M, Ao L, Hu X, Ma J, Bao K, Gu Y, et al. Influence of blood glucose fluctuation, C-peptide level and conventional risk factors on carotid artery intima–media thickness in Chinese Han patients with type 14. Liu M, Ao L, Hu X, Ma J, Bao K, Gu Y, et al. Influence of blood glucose fluctuation, C-peptide level and conventional risk factors on carotid artery intima–media thickness in Chinese Han patients with type Page 12/16 Page 12/16 2 diabetes mellitus. Eur J Med Res. 2019 Dec;24(1):13. 15. Barbieri M, Rizzo MR, Marfella R, Boccardi V, Esposito A, Pansini A, et al. Decreased carotid atherosclerotic process by control of daily acute glucose fluctuations in diabetic patients treated by DPP-IV inhibitors. Atherosclerosis. 2013 Apr;227(2):349–54. 16. Zhao Y, Evans MA, Allison MA, Bertoni AG, Budoff MJ, Criqui MH, et al. Multisite atherosclerosis in subjects with metabolic syndrome and diabetes and relation to cardiovascular events: The Multi- Ethnic Study of Atherosclerosis. Atherosclerosis. 2019 Mar;282:202–9. 17. Matsumoto K, Sera Y, Abe Y, Tominaga T, Yeki Y, Miyake S. Metformin attenuates progression of carotid arterial wall thickness in patients with type 2 diabetes. Diabetes Res Clin Pract. 2004 Jun;64(3):225–8. 18. Katakami N, Yamasaki Y, Hayaishi-Okano R, Ohtoshi K, Kaneto H, Matsuhisa M, et al. Metformin or gliclazide, rather than glibenclamide, attenuate progression of carotid intima-media thickness in subjects with type 2 diabetes. Diabetologia. 2004 Nov;47(11):1906–13. 19. Rizzo M, Rizvi AA, Patti AM, Nikolic D, Giglio RV, Castellino G, et al. Liraglutide improves metabolic parameters and carotid intima-media thickness in diabetic patients with the metabolic syndrome: an 18-month prospective study. Cardiovasc Diabetol. 2016 Dec;15(1):162. 20. Langenfeld MR, Forst T, Hohberg C, Kann P, Lübben G, Konrad T, et al. Pioglitazone Decreases Carotid Intima-Media Thickness Independently of Glycemic Control in Patients With Type 2 Diabetes Mellitus: Results From a Controlled Randomized Study. Circulation. 2005 May 17;111(19):2525–31. 21. Chiavaroli L, Mirrahimi A, Ireland C, Mitchell S, Sahye-Pudaruth S, Coveney J, et al. Cross-sectional associations between dietary intake and carotid intima media thickness in type 2 diabetes: baseline data from a randomised trial. BMJ Open. 2017 Mar;7(3):e015026. 21. Chiavaroli L, Mirrahimi A, Ireland C, Mitchell S, Sahye-Pudaruth S, Coveney J, et al. Cross-sectional associations between dietary intake and carotid intima media thickness in type 2 diabetes: baseline data from a randomised trial. BMJ Open. 2017 Mar;7(3):e015026. Figures Page 13/16 Page 13/16 Figure 1 Carotid intima-media thickness according to sex and age group in subjects with diabetes mellitus without cardiovascular disease at the Edgardo Rebagliati Hospital. A value of 0.86 mm was considered as abnormal carotid medial intima thickness. 54 subjects (29.4% ) presented abnormal CMIT, 24.5% in female and 41.5% in male. Figure 1 Figure 1 Carotid intima-media thickness according to sex and age group in subjects with diabetes mellitus without cardiovascular disease at the Edgardo Rebagliati Hospital. A value of 0.86 mm was considered as abnormal carotid medial intima thickness. 54 subjects (29.4% ) presented abnormal CMIT, 24.5% in female and 41.5% in male. Page 14/16 Page 14/16 Figure 2 The area under the ROC curve of the metabolic syndrome severity score and classic definitions of metabolic syndrome in subjects with diabetes mellitus and no cardiovascular disease at the Edgardo Rebagliati Martins hospital. MSSS: Metabolic síndrome severity score. MS ATP III: 3 of 5 criteria according to Adult Treatment Panel III. MS Harmonizing criteria: : 3 of 5 criteria according to Harmonized Consensus. A value of 0.86 mm was considered as abnormal carotid medial intima thickness. Figure 2 Figure 2 The area under the ROC curve of the metabolic syndrome severity score and classic definitions of metabolic syndrome in subjects with diabetes mellitus and no cardiovascular disease at the Edgardo Rebagliati Martins hospital. MSSS: Metabolic síndrome severity score. MS ATP III: 3 of 5 criteria according to Adult Treatment Panel III. MS Harmonizing criteria: : 3 of 5 criteria according to Harmonized Consensus. A value of 0.86 mm was considered as abnormal carotid medial intima thickness. Page 15/16 Figure 3 Correlation between carotid intima-media thickness with a severity score of metabolic syndrome and classic MS definitions in subjects with diabetes mellitus without cardiovascular disease at Rebagliati h it l A CIMT d i i l MSSS S l ti 0 05 0 46 B CIMT d MSSS ith t Figure 3 Figure 3 Correlation between carotid intima-media thickness with a severity score of metabolic syndrome and classic MS definitions in subjects with diabetes mellitus without cardiovascular disease at Rebagliati hospital. A. CIMT and original MSSS Spearman correlation = -0.05 ; p=0.46. B. CIMT and MSSS without glycemia Spearman correlation = 0.19; p=0.008. . C. CIMT and ATP III SM criterio Spearman correlation =0.18; p=0.01. D. CIMT and Harmonizing MS criterio Spearmna correlation=0.11; p = 0.12 . Correlation between carotid intima-media thickness with a severity score of metabolic syndrome and classic MS definitions in subjects with diabetes mellitus without cardiovascular disease at Rebagliati hospital. A. CIMT and original MSSS Spearman correlation = -0.05 ; p=0.46. B. CIMT and MSSS without glycemia Spearman correlation = 0.19; p=0.008. . C. CIMT and ATP III SM criterio Spearman correlation =0.18; p=0.01. D. CIMT and Harmonizing MS criterio Spearmna correlation=0.11; p = 0.12 . Page 16/16
https://openalex.org/W4221131200	https://portal.findresearcher.sdu.dk/files/201079903/Open_Access_Version.pdf	English	null	Expression of the Nonclassical MHC Class I, Saha-UD in the Transmissible Cancer Devil Facial Tumour Disease (DFTD)	Pathogens	2,022	cc-by	10,416	Download date: 24. Oct. 2024 Citation for pulished version (APA): Hussey, K., Caldwell, A., Kreiss, A., Skjødt, K., Gastaldello, A., Pye, R., Hamede, R., Woods, G. M., & Siddle, H. V. (2022). Expression of the Nonclassical MHC Class I, Saha-UD in the Transmissible Cancer Devil Facial Tumour Disease (DFTD). Pathogens, 11(3), Article 351. https://doi.org/10.3390/pathogens11030351 Citation: Hussey, K.; Caldwell, A.; Kreiss, A.; Skjødt, K.; Gastaldello, A.; Pye, R.; Hamede, R.; Woods, G.M.; Siddle, H.V. Expression of the Nonclassical MHC Class I, Saha-UD in the Transmissible Cancer Devil Facial Tumour Disease (DFTD). Pathogens 2022, 11, 351. https:// doi.org/10.3390/pathogens11030351 Citation: Hussey, K.; Caldwell, A.; Kreiss, A.; Skjødt, K.; Gastaldello, A.; Pye, R.; Hamede, R.; Woods, G.M.; Siddle, H.V. Expression of the Nonclassical MHC Class I, Saha-UD in the Transmissible Cancer Devil Facial Tumour Disease (DFTD). Pathogens 2022, 11, 351. https:// doi.org/10.3390/pathogens11030351 Academic Editor: Mhairi A. Morris Received: 31 January 2022 Accepted: 3 March 2022 Published: 14 March 2022 Keywords: transmissible cancer; Tasmanian devils; MHC class I; nonclassical; immune evasion Published: 14 March 2022 Article Expression of the Nonclassical MHC Class I, Saha-UD in the Transmissible Cancer Devil Facial Tumour Disease (DFTD) Kathryn Hussey 1 , Alison Caldwell 1, Alexandre Kreiss 2, Karsten Skjødt 3, Annalisa Gastaldello 1 , Ruth Pye Rodrigo Hamede 4, Gregory M. Woods 2 and Hannah V. Siddle 1,* 1 School of Biological Sciences, University of Southampton, Southampton SO17 1BJ, UK; k.a.hussey@soton.ac.uk (K.H.); aly.caldwell@googlemail.com (A.C.); annalisa80it@yahoo.it (A.G.) 2 Menzies Institute for Medical Research, University of Tasmania, Hobart, TAS 7000, Australia; alexandre.kreiss@utas.edu.au (A.K.); ruth.pye@utas.edu.au (R.P.); g.m.woods@utas.edu.au (G.M.W.) 3 Department of Cancer and Inﬂammation, University of Southern Denmark, 5230 Odense, Denmark; kskjoedt@health.sdu.dk y ( ) y g g ( ) y ( ) 2 Menzies Institute for Medical Research, University of Tasmania, Hobart, TAS 7000, Australia; alexandre.kreiss@utas.edu.au (A.K.); ruth.pye@utas.edu.au (R.P.); g.m.woods@utas.edu.au (G.M.W.) ( ) py ( ) g ( 3 Department of Cancer and Inﬂammation, University of Southern Denmark, 5230 Odense, Denm kskjoedt@health.sdu.dk j 4 School of Biological Sciences, University of Tasmania, Hobart, TAS 7005, Australia; rodrigo.hamedeross@utas.edu.au 4 School of Biological Sciences, University of Tasmania, Hobart, TAS 7005, Australia; d i h d @ t d 4 School of Biological Sciences, University of Tasmania, Hobart, TAS 7005, Australia; d i h d @ d * Correspondence: h.v.siddle@soton.ac.uk Abstract: Devil facial tumour disease (DFTD) is a transmissible cancer that has circulated in the Tasmanian devil population for >25 years. Like other contagious cancers in dogs and devils, the way DFTD escapes the immune response of its host is a central question to understanding this disease. DFTD has a low major histocompatibility complex class I (MHC-I) expression due to epigenetic modiﬁcations, preventing host immune recognition of mismatched MHC-I molecules by T cells. However, the total MHC-I loss should result in natural killer (NK) cell activation due to the ‘missing self’. Here, we have investigated the expression of the nonclassical MHC-I, Saha- UD as a potential regulatory or suppressive mechanism for DFTD. A monoclonal antibody was generated against the devil Saha-UD that binds recombinant Saha-UD by Western blot, with limited crossreactivity to the classical MHC-I, Saha-UC and nonclassical Saha-UK. Using this antibody, we conﬁrmed the expression of Saha-UD in 13 DFTD tumours by immunohistochemistry (n = 15) and demonstrated that Saha-UD expression is heterogeneous, with 12 tumours showing intratumour heterogeneity. Immunohistochemical staining for the Saha-UD showed distinct patterns of expression when compared with classical MHC-I molecules. The nonclassical Saha-UD expression by DFTD tumours in vivo may be a mechanism for immunosuppression, and further work is ongoing to characterise its ligand on immune cells. pathogens pathogens pathogens University of Southern Denmark Expression of the Nonclassical MHC Class I, Saha-UD in the Transmissible Cancer Devil Facial Tumour Disease (DFTD) Hussey, Kathryn; Caldwell, Alison; Kreiss, Alexandre; Skjødt, Karsten; Gastaldello, Annalisa; Pye, Ruth; Hamede, Rodrigo; Woods, Gregory M.; Siddle, Hannah V. Hussey, Kathryn; Caldwell, Alison; Kreiss, Alexandre; Skjødt, Karsten; Gastaldello, Annalisa; Pye, Ruth; Hamede, Rodrigo; Woods, Gregory M.; Siddle, Hannah V. Hussey, Kathryn; Caldwell, Alison; Kreiss, Alexandre; Skjødt, Karsten; Gastaldello, Annalisa; Pye, Ruth; Hamede, Rodrigo; Woods, Gregory M.; Siddle, Hannah V. Citation for pulished version (APA): Hussey, K., Caldwell, A., Kreiss, A., Skjødt, K., Gastaldello, A., Pye, R., Hamede, R., Woods, G. M., & Siddle, H. V. (2022). Expression of the Nonclassical MHC Class I, Saha-UD in the Transmissible Cancer Devil Facial Tumour Disease (DFTD). Pathogens, 11(3), Article 351. https://doi.org/10.3390/pathogens11030351 Go to publication entry in University of Southern Denmark's Research Portal Terms of use This work is brought to you by the University of Southern Denmark. Unless otherwise specified it has been shared according to the terms for self-archiving. If no other license is stated, these terms apply: • You may download this work for personal use only. • You may not further distribute the material or use it for any profit-making activity or com • You may freely distribute the URL identifying this open access version If you believe that this document breaches copyright please contact us providing details and we Please direct all enquiries to puresupport@bib.sdu.dk Terms of use This work is brought to you by the University of Southern Denmark. Unless otherwise specified it has been shared according to the terms for self-archiving. If no other license is stated, these terms apply: g y y y Unless otherwise specified it has been shared according to the terms for self-archiving. If no other license is stated, these terms apply: • You may download this work for personal use only y y p g • You may freely distribute the URL identifying this open access version If you believe that this document breaches copyright please contact us providing details and we will investigate your claim. Please direct all enquiries to puresupport@bib.sdu.dk If you believe that this document breaches copyright please contact us providing details and we will investigate your claim. Please direct all enquiries to puresupport@bib.sdu.dk pathogens 1. Introduction Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional afﬁl- iations. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional afﬁl- iations. Devil facial tumour disease (DFTD) is a transmissible cancer circulating in the Tas- manian devil (Sarcophilus harrisii), a marsupial species endemic to the Australian island of Tasmania [1]. These tumour cells transmit as an allograft between devils due to biting behaviour during feeding and mating [2–4], forming new tumours on the face and/or neck of the new host. The disease was ﬁrst observed in 1996 [5] and has since spread across most of the island, leading to a population decline [6] and the registration of Tasmanian devils as an endangered species in 2008 [7]. However, recent phylogenetic analysis of DFTD indicates the cancer is becoming endemic, with reduced transmission rates over time [8], along with an increase in tolerance [9–12]. Copyright: © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). Transmitting as an allograft, DFTD should elicit an immune response due to the disparate major histocompatibility complex class I (MHC-I) molecules between the cancer and its host. MHC-I molecules consist of a heavy chain and a β2-microglobulin (β2m) molecule, which bind to intracellular self or nonself peptides. Classical MHC-I molecules https://www.mdpi.com/journal/pathogens Pathogens 2022, 11, 351. https://doi.org/10.3390/pathogens11030351 Pathogens 2022, 11, 351 2 of 17 2 of 17 are expressed on the surface of most cells and present peptides to CD8+ cytotoxic T cells via the T cell receptor [13,14]. An immune response can be activated if the T cell recognises a nonself or mutated self peptide (for example, from a virus or cancer cell, respectively) or a nonself MHC-I heavy chain, such as in a transplant. Though devils have low genetic diversity at the MHC-I loci [15,16], it has been shown that devils are able to reject both MHC-I-matched and mismatched skin grafts [17], yet an immune response to DFTD is rarely observed. Six wild devils have been found to produce serum antibodies against DFTD, four of these associated with regressions [18], but these cases are rare, with most devils appearing ignorant of a DFTD infection. 1. Introduction Interestingly, some devils are able to tolerate infection, with increased survival in females associated with the host SNPs in genes linked to cell-cycle regulation, cell adhesion, and immunity, though no link was found in males [11]. An important mechanism of immune evasion in DFTD is the loss of MHC-I expression through epigenetic modiﬁcations of genes that encode components of the antigen presenta- tion pathway, β2m and TAP (transporter associated with antigen processing) [19]. These genes can be upregulated by treatment with the inﬂammatory cytokine interferon-gamma (IFN-γ) [19] or overexpression of transcriptional coactivator NLRC5 [20]. More recently, it has been found that DFTD expresses polycomb repressive complex 2 (PRC2), which is a conserved mechanism for transcriptional silencing of the antigen presentation pathway in tumours with extremely low MHC-I expression [21]. y p Low levels of classical MHC-I at the cell surface should trigger a natural killer (NK) cell response due to a lack of MHC-I expression, termed ‘missing self’ [22]. While devil NK cell markers are not well deﬁned, making study of these cells difﬁcult, it has been found that devils have functional NK cells that can engage in the cytotoxic killing of the human cancer cell line, K562, which lacks MHC-I on the cell surface; though this response was not seen against DFTD cells [23]. As a result, it is expected that DFTD cells are employing additional mechanisms to evade the host immune system. y There are two broad categories of MHC-I molecules: classical and nonclassical. Classi- cal MHC-I are highly polymorphic, expressed by most cell types, and bind intracellular peptides for presentation to T cells at the cell surface [24,25]. Nonclassical MHC-I genes have low levels of polymorphism and can bind a variety of ligands, including peptides and lipids. These molecules can interact with a broader range of immune cells, including, but not limited to, NK cells [24–26] where they have an established role as inhibitory ligands [26]. In the Tasmanian devil, there are three classical genes: Saha-UA, Saha-UB, and Saha- UC [16]; in addition, there are six nonclassical genes: Saha-UD, Saha-UK [16], Saha-UM, Saha-CD1, Saha-MR1 [27], and Saha-UT [28]. Eleven Saha-UD alleles have been identiﬁed (at the amino acid level), and all have high nucleotide sequence similarity (>97.7%) with putative peptide binding residues not under positive selection [27]. 2. Results 2.1. Anti-Saha-UD Antibody, α-14-37-3, Speciﬁcally Binds Recombinant Nonclassical Saha-UD Protein 2.1. Anti-Saha-UD Antibody, α-14-37-3, Speciﬁcally Binds Recombinant Nonclassical Saha-UD Protein Saha-UD is a devil MHC-I gene that has been classiﬁed as nonclassical due to low polymorphisms between alleles and tissue-speciﬁc expression [16]. However, phyloge- netic analysis groups Saha-UD with the classical MHC-I genes, Saha-UA, Saha-UB, and Saha-UC [16,27,31]; while other nonclassical MHC-I genes, Saha-UK and Saha-UM, form a separate clade, and Saha-CD1 and Saha-MR1 group with orthologues in other species [27]. Sequence identity analysis on the alpha 1–alpha 2 domain amino acid sequence of the devil classical MHC-I (SahaI-01) and nonclassical Saha-UD (SahaI12), Saha-UK, and Saha-UM is shown in Table 1, conﬁrming sequence identity is highest between classical MHC-I and the nonclassical Saha-UD at 78.2%. The sequence identity and phylogenetic analysis indicates that while Saha-UD shares evolutionary history with the devil classical MHC-I, its polymorphism and expression may denote a unique function. Table 1. Percentage sequence identity of the alpha 1–alpha 2 domain amino acid sequence of devil MHC class I genes: classical MHC class I (SahaI-01) and nonclassical Saha-UD (SahaI12), Saha-UK, and Saha-UM. Table 1. Percentage sequence identity of the alpha 1–alpha 2 domain amino acid sequence of devil MHC class I genes: classical MHC class I (SahaI-01) and nonclassical Saha-UD (SahaI12), Saha-UK, and Saha-UM. Table 1. Percentage sequence identity of the alpha 1–alpha 2 domain amino acid sequence of devil MHC class I genes: classical MHC class I (SahaI-01) and nonclassical Saha-UD (SahaI12), Saha-UK, and Saha-UM. MHC Class I Classical (SahaI-01) Saha-UD (SahaI12) Saha-UK Saha-UM Classical (SahaI-01) 78.2% 60.6% 52.1% Saha-UD (SahaI12) 78.2% 58.1% 52.4% Saha-UK 60.6% 58.1% 43.1% Saha-UM 52.1% 52.4% 43.1% To investigate Saha-UD expression further, anti-Saha-UD antibodies were generated by immunising mice using the peptide sequence ‘WIEKMENVDRDYWE’. This sequence comes from the same region of the alpha-1 domain as has been used previously to generate anti-Saha-UA-UB-UC antibodies [32] and has low sequence similarity to the classical MHC-I sequences and nonclassical MHC-I, Saha-UK (Figure 1A). Eight anti-Saha-UD antibody clones were screened by immunohistochemistry (IHC) using formalin-ﬁxed parafﬁn-embedded (FFPE) devil spleen samples (Figure S1). From this data, two antibodies were chosen that gave the strongest staining with minimal background: clone α-UD_14-37-3 (hereafter referred to as UD(3)) and α-UD_14-37-5 (UD(5)). 1. Introduction Amino acid differences between the alleles are low, and few of these residues are predicted to interact with peptides or the T cell receptor [29]. Saha-UD, while it is in a separate phylogenetic clade [16], is closely related to the classical MHC-I genes [27] and binds to β2m [30]; therefore, it is hypothesised that Saha-UD molecules bind peptides for antigen presentation at the cell surface. Saha-UK is monomorphic [27]; therefore, its expression would not be expected to activate an immune response. However, it is possible it does not bind to β2m [30] and is predicted to have a marsupial-speciﬁc function [16]. Due to the expression of Saha-UD in DFTD cell lines, its potential for peptide presentation, and its low polymorphism in the population, we investigated the expression of nonclassical Saha-UD in DFTD primary tumours as a mechanism for host immune evasion. Pathogens 2022, 11, 351 3 of 17 2. Results The recom- binant Saha-UD heavy chain protein was stained with the UD(3) and UD(5) antibodies via Western blot, along with a pan-speciﬁc anti-Saha-UA-UB-UC antibody (α-UA/UB/UC_15- 25-8) and an anti-Saha-UK antibody (α-UK_15-29-1) [32], hereafter referred to as UABC and UK antibodies, respectively, as negative controls (Figure 1B). UD(3) and UD(5) both identiﬁed the Saha-UD recombinant protein, giving a band at ~44 kDa, the size expected for the recombinant protein. UD(5) produced stronger staining but also stained for multiple bands above and below the ~44 kDa band for the Saha-UD, which were not seen on the UD(3) blot. The UABC and UK antibodies both gave a very faint band for the Saha-UD, validating that these antibodies do not have a strong afﬁnity for the nonclassical Saha-UD. 4 of 17 Pathogens 2022, 11, 351 (A) (D) Tasmanian Devil MHC Class I Saha-UD Saha-UA, -UB & -UC WMEKVQDVDPGYWE WMEKVKDVDPGYWE Saha-UD WIEKMENVDRDYWE Saha-UK WLDKSN~~~QNYWE (B) M UD UD(3) UD(5)UABC UK Size (kDa) ~170 ~130 ~95 ~72 ~55 ~43 ~34 ~26 M UC UK UC UK UC UK UC UK UD(3) UD(5) UABC UK Size (kDa) ~170 ~130 ~95 ~72 ~55 ~43 ~34 ~26 ~17 (C) TD74_Lymph Node UD(3) Isotype Control Secondary Control A) Tasmanian Devil MHC Class I Saha-UD Saha-UA, -UB & -UC WMEKVQDVDPGYWE WMEKVKDVDPGYWE Saha-UD WIEKMENVDRDYWE Saha-UK WLDKSN~~~QNYWE (A) (D) Tasmanian Devil MHC Class I Saha-UD Saha-UA, -UB & -UC WMEKVQDVDPGYWE WMEKVKDVDPGYWE Saha-UD WIEKMENVDRDYWE Saha-UK WLDKSN~~~QNYWE (B) M UD UD(3) UD(5)UABC UK Size (kDa) ~170 ~130 ~95 ~72 ~55 ~43 ~34 ~26 M UC UK UC UK UC UK UC UK UD(3) UD(5) UABC UK Size (kDa) ~170 ~130 ~95 ~72 ~55 ~43 ~34 ~26 ~17 (C) TD74_Lymph Node UD(3) Isotype Control Secondary Control Figure 1. Nonclassical MHC class I, Saha-UD antibody generation and validation. (A) Antibodies were generated against the peptide sequence ‘WIEKMENVDRDYWE’ from the alpha-1 domain of Saha-UA, -UB & -UC WMEKVQDVDPGYWE WMEKVKDVDPGYWE (B) M UD UD(3) UD(5)UABC UK Size (kDa) ~170 ~130 ~95 ~72 ~55 ~43 ~34 ~26 M UC UK UC UK UC UK UC UK UD(3) UD(5) UABC UK Size (kDa) ~170 ~130 ~95 ~72 ~55 ~43 ~34 ~26 ~17 (C) (C) (B) (D) TD74_Lymph Node UD(3) Isotype Control Secondary Control Figure 1. Nonclassical MHC class I, Saha-UD antibody generation and validation. (A) Antibodies were generated against the peptide sequence ‘WIEKMENVDRDYWE’ from the alpha-1 domain of TD74_Lymph Node TD74_Lymph Node (D) Isotype Control Secondary Control Figure 1. 2. Results Nonclassical MHC class I, Saha-UD antibody generation and validation. (A) Antibodies were generated against the peptide sequence ‘WIEKMENVDRDYWE’ from the alpha-1 domain of Pathogens 2022, 11, 351 5 of 17 5 of 17 the nonclassical MHC class I, Saha-UD heavy chain protein. This sequence differs from the classical MHC class I (Saha-UA, -UB, and -UC) and nonclassical Saha-UK sequences, shown in the amino acid alignment. A putative model of the Saha-UD molecule is shown, with the antibody epitope in grey. In the grey box, red indicates where the sequences differ. An antibody against the Saha-UA-UB-UC was generated previously, using the sequence ‘WMEKVQDVDPGYWE’ from the alpha-1 domain [32]. (B,C) Two anti-Saha-UD antibodies, clones α-UD_14-37-3 (UD(3)) and α-UD_14-37-5 (UD(5)), were screened by Western blot for protein speciﬁcity. Blots were stained with antibodies against the classical Saha-UA, -UB, and -UC (UABC), α-UA/UB/UC_15-25-8, and nonclassical Saha-UK (UK), α-UK_15- 29-1, as positive controls for their respective proteins and negative controls for other MHC class I. Blots were run with the following: (B) recombinant Saha-UD heavy chain protein; (C) recombinant Saha-UC and Saha-UK heavy chain proteins. (D) Formalin-ﬁxed parafﬁn-embedded Tasmanian devil submandibular lymph node samples were stained by immunohistochemistry with the α-UD_14-37-3 (UD(3)) anti-Saha-UD antibody, along with a secondary antibody only and an IgG2a isotype antibody as controls. Images were taken at 20× magniﬁcation. Positive cells are stained brown; nuclei are stained blue. Black arrows indicate cells with positive staining. Scale bars = 100 µm. Recombinant MHC-I heavy chain proteins, classical Saha-UC, and nonclassical Saha- UK were probed with the anti-Saha-UD antibodies to exclude the possibility of cross- reactivity with the closely related devil MHC-I heavy chain proteins (Figure 1C). As expected, the UABC and UK antibodies stained, respectively, for the Saha-UC (~38 kDa) and Saha-UK (~35 kDa) recombinant proteins, with the UK antibody also binding weakly to the Saha-UC protein. For their respective proteins, there were additional bands at a lower molecular weight. As these were puriﬁed recombinant protein samples, these bands may be due to protein degradation or alternative splicing. This was tested by running the proteins on an SDS-PAGE gel and staining with Coomassie Blue, where the same bands were observed (Figure S2). The UD(3) antibody produced very faint bands for the Saha-UC and Saha-UK recombinant proteins. As devil classical MHC-I molecules have a high sequence similarity [31], the result observed for the Saha-UC is expected to reﬂect all classical MHC-I. 2. Results Therefore, these results indicate the UD(3) antibody does not have a strong binding afﬁnity for either the classical MHC-I or nonclassical Saha-UK and is speciﬁc for the nonclassical Saha-UD. The UD(5) antibody did not produce a band for the Saha-UK protein; however, it did stain for the Saha-UC (~38 kDa). Therefore, it is not a speciﬁc antibody against the nonclassical MHC-I, Saha-UD (Figure 1C). To further validate the speciﬁcity of the UD(3) antibody, serial lymph node samples were stained using a mouse IgG2a isotype control antibody and a secondary-only control (Figure 1D). 2.2. Nonclassical Saha-UD Is Expressed in Primary DFTD Tumours, but Expression Varies and Tumours Show Intratumour Heterogeneity 2.2. Nonclassical Saha-UD Is Expressed in Primary DFTD Tumours, but Expression Varies and Tumours Show Intratumour Heterogeneity To investigate the Saha-UD expression in DFTD, the UD(3) antibody was used to stain 15 DFTD tumour samples by IHC (Figure 2A). Across these samples, collected between May 2006–April 2015, Saha-UD was expressed in 13 samples, but there was varied Saha-UD expression, from very strong staining seen for Crabtree_T2, Lonnavale, and TD505_T2 to negative tumours TD74 and TD388. Thus, while Saha-UD is expressed in many DFTD tumours, it is not expressed in all tumours and is expressed at different levels. Among the tumours stained by IHC, there were two devils for which we had three tumours: Crabtree and TD505. Interestingly, for both animals, while all three tumours expressed Saha-UD, the strength of staining varied between the different tumours (Figure 2A). 6 of 17 Pathogens 2022, 11, 351 (A) (B) Lonnavale TD505_T2 Crabtree_T2 TD505_T1 TD505_T3 Crabtree_T1 Crabtree_T3 Cygnet Franklin_T1 Chauncy Vale_T1b Christine* Grommit* Tiarna* TD74 TD388 Primary Tumours Marker DFTD_1426 DFTD_4906 DFTD_C5065 DFTD_Crabtree T1 DFTD_Crabtree T2 DFTD_Crabtree T3 DFTD_Cygnet DFTD_TD505 T1 DFTD_TD505 T2 DFTD_TD505 T3 DFTD_Lonnavale DFTD_Franklin T1 DFTD_Chauncy Vale Spleen_TD209 Marker No cDNA Cell Lines * * RPL13A UD * * (A) (A) (B) (C) Lonnavale TD505_T2 Crabtree_T2 TD505_T1 TD505_T3 Crabtree_T1 Crabtree_T3 Cygnet Franklin_T1 Chauncy Vale_T1b Christine* Grommit* Tiarna* TD74 TD388 Crabtree_T1 Crabtree_T2 Crabtree_T3 Franklin_T1 Grommit* TD505_T1 TD505_T2 TD505_T3 Primary Tumours Marker DFTD_1426 DFTD_4906 DFTD_C5065 DFTD_Crabtree T1 DFTD_Crabtree T2 DFTD_Crabtree T3 DFTD_Cygnet DFTD_TD505 T1 DFTD_TD505 T2 DFTD_TD505 T3 DFTD_Lonnavale DFTD_Franklin T1 DFTD_Chauncy Vale Spleen_TD209 Marker No cDNA Cell Lines * * RPL13A UD * * Figure 2. Saha-UD expression varies between and within DFTD tumours. (A) Immunoh istry (IHC) staining of formalin-fixed paraffin-embedded (FFPE) DFTD tumour sample anti-Saha-UD antibody. 2. Results Black arrows indicate DFTD tumour cells with positive staini (B) (B) Primary Tumours Marker DFTD_1426 DFTD_4906 DFTD_C5065 DFTD_Crabtree T1 DFTD_Crabtree T2 DFTD_Crabtree T3 DFTD_Cygnet DFTD_TD505 T1 DFTD_TD505 T2 DFTD_TD505 T3 DFTD_Lonnavale DFTD_Franklin T1 DFTD_Chauncy Vale Spleen_TD209 Marker No cDNA Cell Lines * * RPL13A UD * * (C) Crabtree_T1 Crabtree_T2 Crabtree_T3 Franklin_T1 Grommit* TD505_T1 TD505_T2 TD505_T3 Figure 2 Saha UD expression varies between and within DFTD tumours (A) Immunohistoc (C) Figure 2. Saha-UD expression varies between and within DFTD tumours. (A) Immunohistochem- istry (IHC) staining of formalin-fixed paraffin-embedded (FFPE) DFTD tumour samples using an anti-Saha-UD antibody. Black arrows indicate DFTD tumour cells with positive staining for the Pathogens 2022, 11, 351 7 of 17 Saha-UD. Images were taken at 20× magniﬁcation. Positive cells are stained brown; nuclei are stained blue. Asterisks indicate tumours obtained from devils inoculated with cell line DFTD_C5065. Scale bars = 100 µm. (B) RT-PCR on the DFTD cell lines and primary tumours for Saha-UD. RPL13A, encoding a ribosomal protein, was used as a loading control. A no cDNA control is included for each. Asterisks indicate 250 bp on the DNA ladder. (C) IHC images for FFPE DFTD tumour samples showing heterogeneous intratumour staining for the Saha-UD. Black arrows indicate DFTD tumour cells with strong staining for the Saha-UD, and white arrows indicate DFTD tumour cells that are negative or have lower levels of staining for the Saha-UD. Images were taken at 10× magniﬁcation. Positive cells are stained brown; nuclei are stained blue. Asterisks indicate tumours obtained from devils inoculated with cell line DFTD_C5065. Scale bars = 100 µm. To conﬁrm the positive Saha-UD expression observed with the UD(3) antibody by IHC, 10 primary DFTD samples stored in RNAlater and taken from the same tumours stained by IHC, were tested by RT-PCR. Included in this analysis were three DFTD cell lines (DFTD_1426, DFTD_4906, and DFTD_C5065) and devil spleen (Spleen_TD209) as a positive control. Samples were ampliﬁed using primers speciﬁc for Saha-UD and a ribosomal protein, RPL13A, as a control (primers detailed in Table 2). Consistent with previous results, DFTD cell lines were positive for the Saha-UD heavy chain [30,32,33]. All DFTD primary tumours were positive by IHC and RT-PCR, conﬁrming the results obtained with the UD(3) antibody. Tumours negative for the Saha-UD were not tested by RT-PCR as samples stored in RNAlater were not available. 2. Results These results conﬁrm the Saha-UD expression in DFTD tumours from different individuals, but it is difﬁcult to compare expression levels as, based on the IHC results, some contamination from host cells is expected in the biopsies. Table 2. Primer sets. Gene Ampliﬁed Primer Names Primer Sequence (5′ -> 3′) Product Length (bp) Annealing Temperature (◦C) Nonclassical MHC Class I Saha-UD (exon 2–4) Saha138 GCCATATGGGCTCTCACTCCTT GAGGTATTTCGGCACCAC 1044 63 Saha48 ATGCAAGCTTGGCTTTGGCTGT CAGAGAGACATCTGACC RPL13A Saha120 ACAAGACCAAGCGAGGCCAGG 300 60 Saha121 GCCTGGTATTTCCAGCCAACCTCA Nonclassical MHC Class I Saha-UD Dev Men UD (F) ATGGAGAATGTGGACCGGGAC 275 60 Dev Men UD (R) TGAGTTCACTGCCTCATTCACT In addition to the variations in Saha-UD expression between the DFTD tumours, there was heterogeneity within tumours (Figure 2C). Heterogeneity was classed as samples that contained both tumour cells with high-moderate Saha-UD expression and tumour cells that were low-negative. Thirteen out of the 15 tumours showed intratumour heterogene- ity, where strong/moderate Saha-UD-positive tumour cells neighbour tumour cells with weaker staining or no staining at all. For half of the heterogenous tumour samples, this was observed throughout the tumour, for example, Chauncy Vale_T1b (Figure 2A), TD505_T2, and TD505_T3 (Figure 2C). The other half had areas with consistent Saha-UD expression, either positive or negative, in addition to areas with interspersed tumour cells of varying Saha-UD expression, as observed in Crabtree_T1, Crabtree_T2, and Grommit (Figure 2C). Interestingly, there was no clear pattern for areas that were high expressing and areas that were low or negative. Five tumours had areas where there appeared to be a higher expression of the Saha-UD on the edges of a cluster of tumour cells. The only tumours showing consistent staining throughout the tumour were Christine and TD388, the latter being classed as Saha-UD negative. Pathogens 2022, 11, 351 8 of 17 8 of 17 . Distinct Patterns of Staining for Nonclassical Saha-UD versus Classical (Saha-UA, -UB, UC) 2.3. Distinct Patterns of Staining for Nonclassical Saha-UD versus Classical (Saha-UA, -UB, & -UC) Based on our hypothesis that nonclassical Saha-UD is upregulated in DFTD tumours to prevent NK cell activation due to the ‘missing self’, we would expect higher levels of Saha-UD expression when there is lower classical MHC-I expression. To compare the expression of classical MHC-I, Saha-UA, -UB, and -UC with that of nonclassical Saha-UD, serial sections of the 15 DFTD tumours, previously stained using the UD(3) antibody, were stained for classical MHC-I expression using the pan-speciﬁc UABC antibody (α- UA/UB/UC_15-25-8) (Figure 3). 2. Results UD(3) UABC Secondary Control TD74 Tiarna* Crabtree_T3 Crabtree_T1 Chauncy Vale_T1b Crabtree_T2 Figure 3. Patterns of staining observed when comparing nonclassical Saha-UD versus classical MHC class I staining in DFTD. Representative images of staining patterns observed for DFTD tumours when staining for classical MHC class I, Saha-UA, -UB, and -UC, with antibody α-UA/UB/UC_15-25-8 (UABC), and nonclassical MHC class I, Saha-UD, with antibody α-UD_14-37-3 (UD (3)). Images were taken at 20× magnification. Positive cells are stained brown; nuclei are stained blue. Asterisks indicate tumours obtained from devils inoculated with the cell line DFTD_C5065. Scale bars = 100 µm. Figure 3. Patterns of staining observed when comparing nonclassical Saha-UD versus classical MHC class I staining in DFTD. Representative images of staining patterns observed for DFTD tumours when staining for classical MHC class I, Saha-UA, -UB, and -UC, with antibody α-UA/UB/UC_15-25-8 (UABC), and nonclassical MHC class I, Saha-UD, with antibody α-UD_14-37-3 (UD (3)). Images were taken at 20× magnification. Positive cells are stained brown; nuclei are stained blue. Asterisks indicate tumours obtained from devils inoculated with the cell line DFTD_C5065. Scale bars = 100 µm. Pathogens 2022, 11, 351 9 of 17 9 of 17 As with the nonclassical Saha-UD expression, there is variation in classical MHC-I expression between the DFTD tumours, with staining ranging from strong to negative. This contrasts previous ﬁndings that classical MHC-I transcription is low in cell lines and low or not expressed in DFTD primary tumours [19]. There was also intra-tumour heterogeneity for classical MHC-I expression, present in all 13 tumours that stained positively for classical MHC-I. Heterogeneity was observed throughout the tumours. In nine of these tumours, the heterogeneous classical MHC-I staining occurred in a similar pattern to the Saha-UD staining. g The inverse correlation we had expected between the Saha-UD and classical MHC- I expression was not present across all 15 tumours. However, there were two general expression patterns observed, examples of which are shown in Figure 3. The most common pattern, observed in 11 of the tumours, was that staining for the Saha-UD and classical MHC-I mirrored each other (Figure 3; top panel). For example, there is either strong staining for Saha-UD and classical MHC-I or staining for both is weak. 2. Results While the expression of Saha-UD and classical MHC-I cannot be directly compared, as there may be differences in the binding afﬁnity of the antibodies, it is interesting that most of the tumours had similar levels of staining by both antibodies. The two tumours classiﬁed as Saha-UD-negative, TD388 and TD74, were also negative for classical MHC-I. The second pattern observed is that a subset of four tumours had stronger staining for Saha-UD than classical MHC-I (examples in Figure 3; bottom panel). These tumours had strong staining for the Saha-UD, with staining for classical MHC-I weaker or completely negative. Though this pattern more closely ﬁt with our hypothesis, two of the four tumours did have moderate classical MHC-I expression. 3. Discussion Map of Tasmania with sample locations for DFTD tumours. Each dot represents a single tumour; the colour indicates the expression score. Asterisks indicate tumours inoculated during vaccination [35] or transmission trials in captive devils obtained from these locations. Expression scores were assigned based on immunohistochemical staining of DFTD tumour samples. An expression score of 4 indicates strong staining; 0 indicates no staining. Expression score criteria are detailed in Figure S5. Map created using datawrapper.de. Expression Score 4 0 * * (2010) (2014) (2015) (2014) (2015) (2015) (2015) (2014) (2006) (2008) * (2010) Expression Score 4 0 Figure 4. Expression scores for nonclassical Saha-UD by DFTD sample location. Map of Tasmania with sample locations for DFTD tumours. Each dot represents a single tumour; the colour indicates the expression score. Asterisks indicate tumours inoculated during vaccination [35] or transmission trials in captive devils obtained from these locations. Expression scores were assigned based on immunohistochemical staining of DFTD tumour samples. An expression score of 4 indicates strong staining; 0 indicates no staining. Expression score criteria are detailed in Figure S5. Map created using datawrapper.de. As it has been shown that dominant clades in a particular area can change over time [34], the date of the sample collection may be an important factor in the Saha-UD expression. While time in culture will have altered the C5065 cell line since it was collected in 2005, it is interesting that tumour samples from the inoculated devils match the natural tumour from 2006, TD74. Conversely, newer tumours appear to have a broader range of Saha-UD expression. Thus, Saha-UD variability may have increased with time, highlighting the importance of elucidating both spatial and temporal factors in Saha-UD expression, as the two may be interlinked. y The variation in Saha-UD expression may also be due to the individual interactions of tumours with the host immune system. The expression of Saha-UD is lower in Tiarna than in Grommit and Christine (Figure 2A), despite all being inoculated with the cell line C5065, and Grommit and Tiarna both show intratumour heterogeneity. Intratumour heterogeneity, observed in 13 of the 15 DFTD tumours, highlights that Saha-UD expression may be plastic in response to local tissue conditions, such as inﬁltrating immune cells. It has been shown that DFTD cells exhibit mesenchymal plasticity in response to immune activation following immunisation [36], changing to a dedifferentiated, immunosuppressive state. 3. Discussion We have generated and validated a speciﬁc monoclonal antibody against the devil nonclassical Saha-UD (α-UD_14-37-3), which shows little cross-reactivity with other devil recombinant MHC-I proteins (classical Saha-UC and nonclassical Saha-UK). IHC shows that DFTD tumours express Saha-UD, though expression varies between tumours. Saha- UD expression in 10 of the DFTD samples was also conﬁrmed by RT-PCR using Saha-UD speciﬁc primers. Interestingly, DFTD tumours are heterogeneous for Saha-UD expression, (Figure 2C) and, unexpectedly, we observe varied classical MHC-I expression between tumours. The expression of the nonclassical Saha-UD heavy chain has previously been found in three DFTD cell lines [30,32,33], and this study provides the ﬁrst evidence of Saha-UD expression in primary DFTD biopsies. Based on IHC staining, Saha-UD expression is common in DFTD tumours from different individuals. However, this analysis is restricted to 15 tumours over a limited geographic range. As it is possible that Saha-UD expression is changing over time and by location in DFTD, more samples are needed over a greater period to understand broader trends in Saha-UD expression and the evolution of DFTD immune escape. p Saha-UD expression by DFTD tumours may reﬂect the genetic evolution of the cancer as it has spread across Tasmania. The biopsy samples in this study are derived from two broad geographical areas in Tasmania. Devils captured from the Northwest of Tasmania were inoculated with the DFTD cell line C5065, obtained from Eastern Tasmania in 2005, and wild tumours were collected from locations in and around the d’Entrecasteaux peninsula, in the Southeast (See Figure S3 for a map showing sample locations and Table S1). Interestingly, tumours from the d’Entrecasteaux peninsula exhibited higher expression than both tumours to the East of this area and tumours from the inoculated devils (Figure 4). Though sample number is limiting, it is possible the variation in the Saha-UD expression is due to genetic differences between the DFTD tumours in these areas [34]. Recent genotyping of DFTD tumours has found that early in its clonal evolution, DFTD developed into multiple clades, some of which have persisted and become dominant in particular geographic areas [34]. Therefore, Saha-UD expression may be reﬂective of DFTD sublineages. 10 of 17 10 of 17 Pathogens 2022, 11, 351 Expression Score 4 0 * * (2010) (2014) (2015) (2014) (2015) (2015) (2015) (2014) (2006) (2008) * (2010) Figure 4. Expression scores for nonclassical Saha-UD by DFTD sample location. 3. Discussion Though MHC-I expression was not investigated, it could be controlled by similar pathways. However, Saha-UD is located in a different region as compared to the classical MHC-I genes and is missing many upstream regulatory elements present for classical MHC-I, including Pathogens 2022, 11, 351 11 of 17 11 of 17 an interferon stimulated response element [16]. Thus, its response to local inﬂammation is predicted to be limited. an interferon stimulated response element [16]. Thus, its response to local inﬂammation is predicted to be limited. If Saha-UD expression was solely dependent on the local immune environment, pat- terns of heterogeneity within the tumour would be expected. However, this was not seen in most tumours. Some tumours had small areas with slightly higher Saha-UD expression on the outside of clusters of tumour cells, which might be expected if Saha-UD was being upregulated in response to inﬁltrating immune cells, but this was not conclusive or present throughout the tumour. Thus, Saha-UD expression is likely to be associated with a combi- nation of the sublineage of DFTD, the history of host devils encountered by the tumour cells, and the speciﬁc interactions with the current host devil immune system. This is highlighted by varying levels of Saha-UD expression in different DFTD tumours infecting the same devil (see Crabtree and TD505, Figure 2A). These tumours may originate from bite wounds from different devils and/or derive from an infection with different clades of DFTD [34]. These tumours may have been subjected to different selective pressures as they moved through the population prior to infection in the current devil, which may produce differential interactions with the host immune system. This is akin to the evolution observed in other emerging pathogens as they spread through a population: acquiring mutations and genetic changes in response to the pressures of the host environments they encounter, which leads to the development of variants. The DFTD tumours examined are also heterogenous for classical MHC-I expression, which was unexpected given previous ﬁndings that DFTD cells are MHC-I negative [19]. This may be due to plasticity in the classical MHC-I expression, afforded by epigenetic mutations in genes for components of the peptide presentation pathway [21], which can be upregulated by cytokines (for example, IFN-γ). 4.2. Sequence Analysis To calculate the percentage sequence identity for the alpha 1–alpha 2 domain amino acid sequences of devil MHC-I genes (Table 1), the sequences were aligned in CLC Main Workbench 22 and the percentage sequence identity calculated in Bioedit using the sequence identity matrix function. Sequence identity was compared between the classical MHC-I (devil MHC-I reference sequence SahaI-01; NCBI: NP_001267784.1), nonclassical Saha-UD (SahaI12; NCBI: NP_001267783.1), nonclassical Saha-UK (GenBank: AIS75088.1), and nonclassical Saha-UM (GenBank: AIS75089.1). The model of the devil nonclassical Saha-UD (Figure 1A) was generated using an AlphaFold Colab notebook created by DeepMind [42]. Saha-UD is modelled on the alpha 1–alpha 3 domain of the amino acid sequence (SahaI12; NCBI: NP_001267783.1). 4.1. Animals Wild Tasmanian devils were trapped or found dead from road trauma or other causes. Tissue biopsies and ﬁne needle aspirates were collected postmortem or from live devils, which were subsequently released. Samples collected were formalin ﬁxed (10% neutral buffered formalin) and parafﬁn embedded or were stored in RNAlater. p Two tumour samples were from devils, Grommit and Tiarna, who were immunised with an irradiated cell line DFTD_C5065 (protocol detailed in [35]). C5065 (RRID:CVCL_LB79) is a cell line established from a DFTD biopsy in 2005. Despite irradiation of the cell line, tumours grew at the site of inoculation. Christine was inoculated with DFTD_C5065 as part of an unrelated transmission trial. All animal procedures were performed under a standard operating procedure ap- proved by the general manager through the Natural and Cultural Heritage Division, Tasma- nian Government Department of Primary Industries, Parks, Water and the Environment, or under University of Tasmania Animal Ethics Committee Permit A0014976. Information for the devil samples used in this paper are listed in Table S1, and sample collection locations are shown in Figure S3. 4.3. Anti-Saha-UD Antibody Generation Antibodies against the Tasmanian devil MHC-I molecules have previously been de- signed and validated in our lab [32]. These were against the nonclassical Saha-UK, using peptide sequence ‘RITHRTHPDGKVTL’ from the alpha-3 domain, and a pan-speciﬁc anti- body for the classical Saha-UA, -UB, and -UC, against the peptide sequence ‘WMEKVQD- VDPGYWE’ from the alpha-1 domain (Figure 1A). p g For the generation of an antibody against the nonclassical Saha-UD, MHC-I transcripts representing all known devil MHC-I alleles [31] were aligned and translated into protein se- quences using CLC workbench [32]. The alpha 1 domain was manually assessed for regions of low amino acid identity between Saha-UD alleles, classical MHC-I (Saha-UA, -UB, and -UC) and Saha-UK (Figure 2A). Using this sequence, the peptide ‘WIEKMENVDRDYWE’ was synthesised for immunisations against Saha-UD. Mice were immunised subcuta- neously twice with a mixture of GERBU adjuvant and 50 µg of WIEKMENVDRDYWE-C coupled to diphtheria toxoid via the N-terminal cysteine. The ﬁnal boost was performed by intravenous injection, without GERBU, using 25 ug of antigen. Three days later, spleen lymphocytes were fused with the SP2 cell line using PEG as a fusogen. Hybridomas were selected based on reactivity in the ELISA against both N- and C-terminal-coupled pep- tide. Antibodies were subsequently screened against Tasmanian devil spleen samples and veriﬁed against recombinant expressed MHC-I by Western blot, as described below. 3. Discussion This would be a similar progression as observed in the transmissible cancer in dogs, canine transmissible venereal tumour (CTVT), which regulates the expression of MHC-I during an infection of a single host, with MHC-I downregulated during its progressive growth phase and upregulated in response to cytokines during the regressive phase [37,38]. An alternative scenario is that subclonal lineages of DFTD are emerging that express classical MHC-I. MHC-I expression on DFTD cells has been shown to elicit an immune response in devils [39]. Therefore, emergence of MHC-I-positive tumours may contribute to recent ﬁndings of devil immune responses [18] and tolerance to DFTD [11]. Based on our original hypothesis, the loss of classical MHC-I would be expected to correlate with the gain of Saha-UD expression to mitigate the effects of NK cells. Expression levels of Saha-UD and classical MHC-I cannot be directly compared using IHC as antibody afﬁnity may differ. However, all tumours that expressed classical MHC-I also expressed nonclassical Saha-UD. In nine of the 13 tumours that were heterogeneous for classical MHC-I expression, the areas of heterogeneity match between Saha-UD and classical MHC-I. Thus, it appears that the expression of nonclassical Saha-UD and classical MHC-I are linked, potentially due to genes associated with the antigen presentation pathway. Though these results do not support our hypothesis, nonclassical Saha-UD expression may still play a role in immune evasion of DFTD. Given that classical MHC-I expression in DFTD is antigenic [39] and that Saha-UD is expressed when there is classical MHC-I, nonclassical Saha-UD could be creating an immunosuppressive environment to prevent host recognition. In humans, the expression of nonclassical HLA-E, HLA-F, and HLA-G by extravillous trophoblasts induces maternal immunotolerance of foetal cells, preventing recognition of paternal alloantigens via the classical HLA-C molecules [40,41]. We postulate that Saha-UD binds peptides, due to the similarity to classical MHC-I sequences [16,27], and is expressed at the cell surface as it binds β2m [30]. Therefore, so it is possible that Saha-UD is providing an immunosuppressive signal to the immune system. However, to investigate potential roles for Saha-UD in immune evasion, further work is needed to conﬁrm its ligand binding properties and interactions with immune cells. In addition, a wider range of DFTD samples will be tested for Saha-UD expression to investigate whether there are differences over time, by location, or by DFTD clades. Pathogens 2022, 11, 351 12 of 17 12 of 17 4. Materials and Methods 4.1. Animals 4.1. Animals 4.5. SDS-PAGE and Western Blot To test the speciﬁcity of anti-Saha-UD antibodies, the recombinant Saha-UC and Saha- UK protein (described above) was run on a Western blot. 6 µg of Saha-UC or Saha-UK solubilised recombinant protein was made up to a total volume of 25 µL with a loading buffer (500 µL 2X Lamelli sample buffer, 50 µL β-mercaptoethanol, and 450 µL dd-H2O). The Saha-UD samples described above were run alongside the Saha-UC and Saha-UK. Preparations were heated to 95 ◦C for 10 min, and 20 µL were added per well. p p Samples were run on a 12% acrylamide gel (National Diagnostics ProtoGel® 30%) using Laemmli buffers and the Fisherbrand™Vertical Gel Tank. Proteins were transferred to a nitrocellulose blotting membrane (Amersham Protran GE Healthcare Life Sciences, Up- psala, Sweden) in a transfer buffer (25 mM Tris, 190 mM glycine, and 20% (v/v) methanol) using a vertical gel tank (Geneﬂow). The membrane was blocked for 1 hr with 5% milk in TBST (150 mM NaCl, 0.1% (v/v) Tween20, 5% (w/v) milk powder, 50 mM TrisCl (pH 8)); it was then washed in TBST (150 mM NaCl, 0.1% (v/v) Tween20, 50 mM TrisCl (pH 8)), and incubated with a neat hybridoma supernatant at 4 ◦C overnight. Membranes were washed with TBST, incubated with a secondary antibody (IRDye® 680RD Goat anti-Mouse IgG (H + L)) in 5% milk in TBST for 1 hr at room temperature, and washed in TBST. Membranes were visualised using the LI-COR Odyssey® scanner. 4.4. Recombinant MHC Class I Proteins Recombinant devil MHC-I heavy chain proteins were used to determine the speci- ﬁcity of anti-Saha-UD antibodies, speciﬁcally Saha-UC (ensembl:ENSSHAG00000000117; Devil_ref v7.0), Saha-UK (ensembl:ENSSHAG00000002942; Devil_ref v7.0), and Saha-UD Pathogens 2022, 11, 351 13 of 17 13 of 17 (ensembl: ENSSHAG00000010776; Devil_ref v7.0). The MHC-I classical Saha-UC and nonclassical Saha-UK full length recombinant proteins were previously generated in our lab [32]. For this work, we also generated recombinant protein for the nonclassical Saha-UD using the same method. Brieﬂy, exons 2–4 from Saha-UD were ampliﬁed from the full gene sequence (Ensembl identiﬁer: ENSSHAG00000010776.2) using primers Saha-138 and Saha- 48 (Table 2). The subsequent amplicons were cloned into the pET22b+ vector (Novagen) and transformed into competent dh5α E. coli cells. Clones from single colonies were grown in LB with ampicillin for 12 h, the plasmid was isolated using the NucleoSpin Plasmid kit (Macherey-Nagel, Düren, Germany), and the transcripts were sequenced in both directions. pET22b+-SahaUD plasmids with the correct sequence were transformed into Rosetta pLysS cells (Novagen). Bacterial colonies containing the plasmid were cultured to OD600 0.6, and protein expression was induced with 1mM Isopropyl β-D-1-thiogalactopyranoside (IPTG). The IPTG-induced cell suspension was spun down at 13,000 rpm for 1 min and resuspended in a 25 µL denaturing loading buffer (500 µL 2X Lamelli sample buffer, 50 µL β-mercaptoethanol, and 450 µL dd-H2O) for direct analysis by Western blot (see below). 4.6. Immunohistochemistry (IHC) Spleen and DFTD tissue samples were ﬁxed in a 10% (v/v) PBS-buffered formalin solution for 2 to 4 days before being processed and embedded in parafﬁn blocks. The samples were cut using a microtome at a thickness between 4−10 µm and mounted onto coated slides. Sections were deparafﬁnized in xylene and rehydrated through graded alcohol. Anti- gen retrieval was performed by a water bath (95 ◦C) in a citrate buffer solution (10 mM citric acid, 0.05% Tween20, pH 6) for 40 min followed by cooling for 15 min. Endogenous peroxidase was blocked by the incubation of slides with 3% H2O2 (Sigma Aldrich, St. Louis, MI, USA) for 10 min, and the nonspeciﬁc protein binding was blocked with a 10% (v/v) goat serum in PBS for 30 min. Sections were incubated with primary antibodies (listed in Table S2) at 4 ◦C overnight. A peroxidase-coupled secondary antibody (EnVision Peroxidase/DAB+ kit; Dako) was used to detect primary antibody binding; incubating sections with HRP for 30 min and colour developed with DAB chromogen for 5 min. Sec- tions were counterstained with haematoxylin (Vector hematoxylin nuclear counterstain (Gill’s Formula)) for 4 min, differentiated in 2% (v/v) acetic acid, and blued in 0.2% (v/v) ammoniated water. Sections were dehydrated through graded alcohol, transferred to xylene, and cover-slipped (using National Diagnostics Omnimount Histological Mounting Medium). Images were captured using the Nikon Eclipse 400 microscope, Retiga 2000R Pathogens 2022, 11, 351 14 of 17 camera, and Q-capture pro 7 computer software. Scale bars were added to images using ImageJ software. camera, and Q-capture pro 7 computer software. Scale bars were added to images using ImageJ software. 4.7. RT-PCR The NucleoSpin RNA kit (Macherey-Nagel, Düren, Germany) was used to extract RNA from cells lines and TRI Reagent (Sigma-Aldrich) used for tissue samples, both according to the manufacturer’s instructions. 1 µg of RNA was reverse transcribed to cDNA using the RevertAid First Strand cDNA Synthesis kit (Thermo Fisher, Waltham, MA, USA), according to the manufacturer’s instructions. Dev Men UD primers [32] were used to amplify the nonclassical MHC-I, Saha-UD by PCR. RPL13A primers [19], amplifying a ribosomal protein, were used as a control. Primers and annealing temperatures are listed in Table 2. 500 ng of cDNA was ampliﬁed using Phusion High-Fidelity DNA Polymerase (Thermo Fisher), according to the manufacturer’s instructions (see Table 3 for PCR reaction conditions) and appropriate negative controls. PCR products were run on a 1.2% (w/v) agarose gel with GelRed and visualised using the SynGene PXi machine and SynGene software. Table 3. PCR reagents and conditions. PCR Reagents RCR Reaction Conditions Reagent Final Concentration Cycle Element Temperature (◦C) Time (s) Number of Cycles cDNA 500 ng Initial denaturation 98 30 1 DNA polymerase (Thermo Fisher Phusion) 0.5 U Denaturation 98 10 Primers 0.6 µM Annealing Temperatures for primers in Table 2 30 30 dNTPs 200 µM Elongation 72 30 Phusion High Fidelity buffer 1X ddH2O To total volume of 25 µl Final Elongation 72 300 1 Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/pathogens11030351/s1, Figure S1: Screening of nonclassical anti- Saha-UD antibodies by immunohistochemistry using devil spleen samples; Figure S2: Recombinant heavy chain proteins for classical Saha-UC and nonclassical Saha-UK, run on an SDS-PAGE gel stained with Coomassie Blue; Figure S3: Map of sample locations for Devil Facial Tumour Disease (DFTD) samples; Figure S4: Immunohistochemistry staining of Devil Facial Tumour Disease (DFTD) samples for periaxin expression [43,44]; Figure S5: Expression score criteria and mapped DFTD classical MHC class I expression scores; Table S1: Details of Devil Facial Tumour Disease (DFTD) samples used in this paper; Table S2: Primary antibodies. Table 3. PCR reagents and conditions. 4.7. RT-PCR PCR Reagents RCR Reaction Conditions Reagent Final Concentration Cycle Element Temperature (◦C) Time (s) Number of Cycles cDNA 500 ng Initial denaturation 98 30 1 DNA polymerase (Thermo Fisher Phusion) 0.5 U Denaturation 98 10 Primers 0.6 µM Annealing Temperatures for primers in Table 2 30 30 dNTPs 200 µM Elongation 72 30 Phusion High Fidelity buffer 1X ddH2O To total volume of 25 µl Final Elongation 72 300 1 Table 3. PCR reagents and conditions. Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/pathogens11030351/s1, Figure S1: Screening of nonclassical anti- Saha-UD antibodies by immunohistochemistry using devil spleen samples; Figure S2: Recombinant heavy chain proteins for classical Saha-UC and nonclassical Saha-UK, run on an SDS-PAGE gel stained with Coomassie Blue; Figure S3: Map of sample locations for Devil Facial Tumour Disease (DFTD) samples; Figure S4: Immunohistochemistry staining of Devil Facial Tumour Disease (DFTD) samples for periaxin expression [43,44]; Figure S5: Expression score criteria and mapped DFTD classical MHC class I expression scores; Table S1: Details of Devil Facial Tumour Disease (DFTD) samples used in this paper; Table S2: Primary antibodies. Author Contributions: Conceptualisation, K.H., H.V.S. and G.M.W.; Methodology, K.H., A.C., K.S., A.G. and H.V.S.; Validation, K.H. and A.C.; Formal Analysis, K.H. and H.V.S.; Investigation, K.H.; Resources, A.K., K.S., R.P., R.H. and G.M.W.; Writing—Original Draft Preparation, K.H. and H.V.S.; Writing—Review & Editing, K.H., K.S., R.H. and H.V.S.; Visualisation, K.H.; Supervision, H.V.S., R.H. and G.M.W.; Project Administration, H.V.S.; Funding Acquisition, H.V.S. and R.H. All authors have read and agreed to the published version of the manuscript. Funding: This work was funded by an Established Investigator grant from the Morris Animal Foundation to H.V.S. (D18ZO-104), the Australian Research Council (DP180100520) and funds raised by the Save the Tasmanian Devil Appeal; Wildcare Inc. by Safﬁre. K.H. and A.C. were funded by PhD scholarships from The Gerald Kerkut Charitable Trust and the University of Southampton. The Australian Research Council (DE170101116 and LP170101101) funded ﬁeldwork and data collection for the study. Pathogens 2022, 11, 351 15 of 17 15 of 17 Institutional Review Board Statement: All Tasmanian devil procedures were performed under a standard operating procedure approved by the general manager through the Natural and Cultural Heritage Division, Tasmanian Government Department of Primary Industries, Parks, Water and the Environment, or under University of Tasmania Animal Ethics Committee Permit A0014976. References Pearse, A.; Swift, K.A.D. Transmission of devil facial-tumour disease. Nature 2006, 439, 549. [CrossRef] [Pu 1. Pearse, A.; Swift, K.A.D. Transmission of devil facial-tumour disease. Nature 2006, 439, 549. [Cross 2. Loh, R.; Bergfeld, J.; Hayes, D.; O’Hara, A.; Pyecroft, S.; Raidal, S.; Sharpe, R. The Pathology of Devil Facial Tumor Disease (DFTD) in Tasmanian Devils (Sarcophilus harrisii). Vet. Pathol. 2006, 43, 890–895. [CrossRef] [PubMed] 2. Loh, R.; Bergfeld, J.; Hayes, D.; O’Hara, A.; Pyecroft, S.; Raidal, S.; Sharpe, R. The Pathology of Devil Facial Tumor Disease (DFTD) in Tasmanian Devils (Sarcophilus harrisii). Vet. Pathol. 2006, 43, 890–895. [CrossRef] [PubMed] 3. Hamede, R.K.; McCallum, H.; Jones, M. Seasonal, demographic and density-related patterns of contact between Tasmanian devils (Sarcophilus harrisii): Implications for transmission of devil facial tumour disease. Austral. Ecol. 2008, 33, 614–622. [CrossRef] 3. Hamede, R.K.; McCallum, H.; Jones, M. Seasonal, demographic and density-related patterns of contact between Tasmanian devils (Sarcophilus harrisii): Implications for transmission of devil facial tumour disease. Austral. Ecol. 2008, 33, 614–622. [CrossRef] p p 4. Hamede, R.K.; McCallum, H.; Jones, M. Biting injuries and transmission of Tasmanian devil facial tumour disease. J. Anim. Ecol. 2013, 82, 182–190. [CrossRef] 4. Hamede, R.K.; McCallum, H.; Jones, M. Biting injuries and transmission of Tasmanian devil facial tumour disease. J. Anim. Ecol. 2013, 82, 182–190. [CrossRef] 5. Hawkins, C.; Baars, C.; Hesterman, H.; Hocking, G.; Jones, M.; Lazenby, B.; Mann, D.; Mooney, N.; Pemberton, D.; Pyecroft, S.; et al. Emerging disease and population decline of an island endemic, the Tasmanian devil Sarcophilus harrisii. Biol. Conserv. 2006, 131, 307–324. [CrossRef] ] r, M.W.; Brown, W.E.; Hawkins, C.E.; Hocking, G.J.; Hume, F.; Huxtable, S.; Iles, P.; Jones, M.; Lawrence 6. Lazenby, B.T.; Tobler, M.W.; Brown, W.E.; Hawkins, C.E.; Hocking, G.J.; Hume, F.; Huxtable, S.; Iles, P.; Jones, M.; Lawrence, C.; et al. Density trends and demographic signals uncover the long-term impact of transmissible cancer in Tasmanian devils. J. Appl. Ecol. 2018, 55, 1368–1379. [CrossRef] nds and demographic signals uncover the long-term impact of transmissible cancer in Tasmanian devils. J 1368–1379. [CrossRef] pp 7. Hawkins, C.E.; McCallum, H.; Mooney, N.; Jones, M.; Holdsworth, M. Sarcophilus harrisii. The IUCN Red List of Threatened Species 2008: e.T40540A10331066. 2008. Available online: https://www.iucnredlist.org/species/40540/10331066 (accessed on 30 January 2022). ) 8. Patton, A.H.; Lawrance, M.F.; Margres, M.J.; Kozakiewicz, C.P.; Hamede, R.; Ruiz-Aravena, M.; Hamilton, D.G.; Comte, S.; Ricci, L.E.; Taylor, R.L.; et al. 4.7. RT-PCR The animal study protocol was approved by the Institutional Animal Welfare and Ethical Review Body of the University of Southampton (submission ID: 53590; approved: 16 January 2020). Informed Consent Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: The data presented in this study are available in Hussey et al., 2022 and associated Supplementary Material. Data Availability Statement: The data presented in this study are available in Hussey et al., 2022 and associated Supplementary Material. Acknowledgments: The authors thank Jon Ward and Jeanette Norman from the Histochemistry Research Unit at the University of Southampton for parafﬁn embedding tumour samples. Acknowledgments: The authors thank Jon Ward and Jeanette Norman from the Histochemistry Research Unit at the University of Southampton for parafﬁn embedding tumour samples. Conﬂicts of Interest: The authors declare no conﬂict of interest. β2m β2-microglobulin CTVT Canine Transmissible Venereal Tumour DFTD Devil Facial Tumour Disease 1 FFPE Formalin-Fixed Parafﬁn-Embedded HLA Human Leukocyte Antigen IFN-γ Interferon-γ IHC Immunohistochemistry NK Cell Natural Killer Cell MHC-I Major Histocompatibility Complex Class I PBS Phosphate Buffered Saline PCR Polymerase Chain Reaction PRC2 Polycomb Repressive Complex 2 TAP Transporter Associated with Antigen Processing References [CrossRef] [PubMed] [ ] [ ] 19. Siddle, H.V.; Kreiss, A.; Tovar, C.; Yuen, C.K.; Cheng, Y.; Belov, K.; Swift, K.; Pearse, A.-M.; Hamede, R.; Jones, M.; et al. Reversible epigenetic down-regulation of MHC molecules by devil facial tumour disease illustrates immune escape by a contagious cancer. Proc. Natl. Acad. Sci. USA 2013, 110, 5103–5108. [CrossRef] 20. Ong, C.E.B.; Patchett, A.L.; Darby, J.M.; Chen, J.; Liu, G.-S.; Lyons, A.B.; Woods, G.M.; Flies, A.S. NLRC5 regulates expression of MHC-I and provides a target for anti-tumor immunity in transmissible cancers. J. Cancer Res. Clin. Oncol. 2021, 147, 1973–1991. [CrossRef] 21. Burr, M.; Sparbier, C.; Chan, K.L.; Chan, Y.-C.; Kersbergen, A.; Lam, E.Y.; Azidis-Yates, E.; Vassiliadis, D.; Bell, C.; Gilan, O.; et al. An Evolutionarily Conserved Function of Polycomb Silences the MHC Class I Antigen Presentation Pathway and Enables Immune Evasion in Cancer. Cancer Cell 2019, 36, 385–401. [CrossRef] 22. Ljunggren, H.-G.; Kärre, K. In search of the ‘missing self’: MHC molecules and NK cell recognition. Immunol. Today 1990, 11, 237–244. [CrossRef] 23. Brown, G.K.; Kreiss, A.; Lyons, A.B.; Woods, G.M. Natural Killer Cell Mediated Cytotoxic Responses in the Tasmanian Devil. PLoS ONE 2011, 6, e24475. [CrossRef] [PubMed] 24. Braud, V.M.; Allan, D.S.; McMichael, A.J. Functions of nonclassical MHC and non-MHC-encoded class I molecules. Curr. Opin. Immunol. 1999, 11, 100–108. [CrossRef] 25. Janeway, C.A.; Travers, P.; Walport, M.; Shlomchik, M.J. The major histocompatibility complex and its functions. In Immunobiology: The Immune System in Health and Disease, 5th ed.; Garland Science: New York, NY, USA, 2001. 26. Allen, R.L.; Hogan, L. Non-Classical MHC Class I Molecules (MHC-Ib). eLS 2013. [CrossRef] 27. Cheng, Y.; Belov, K. Characterisation of non-classical MHC class I genes in the Tasmanian devil (Sarcophilus harrisii). Immunogenetics 2014, 66, 727–735. [CrossRef] [PubMed] 28. Papenfuss, A.T.; Feng, Z.-P.; Krasnec, K.; Deakin, J.E.; Baker, M.L.; Miller, R.D. Marsupials and monotremes possess a novel family of MHC class I genes that is lost from the eutherian lineage. BMC Genom. 2015, 16, 535. [CrossRef] [PubMed] 29. Bjorkman, P.J.; Parham, P. Structure, function, and diversity of class I major histocompatibility compl Biochem. 1990, 59, 253–288. [CrossRef] am, P. Structure, function, and diversity of class I major histocompatibility complex molecules. Annu. Rev 3–288. [CrossRef] 30. Gastaldello, A.; Ramarathinam, S.H.; Bailey, A.; Owen, R.; Turner, S.; Kontouli, N.; Elliott, T.; Skipp, P.; Purcell, A.W.; Siddle, H.V. References A transmissible cancer shifts from emergence to endemism in Tasmanian devils. Science 2020, 370, eabb9772. [CrossRef] 9. Epstein, B.; Jones, M.; Hamede, R.; Hendricks, S.; McCallum, H.; Murchison, E.P.; Schönfeld, B.; Wiench, C.; Hohenlohe, P.; Storfer, A. Rapid evolutionary response to a transmissible cancer in Tasmanian devils. Nat. Commun. 2016, 7, 1–7. [CrossRef] 10. Hubert, J.-N.; Zerjal, T.; Hospital, F. Cancer- and behavior-related genes are targeted by selection in the Tasmanian devil 10. Hubert, J.-N.; Zerjal, T.; Hospital, F. Cancer- and behavior-related genes are targeted by selec (Sarcophilus harrisii). PLoS ONE 2018, 13, e0201838. [CrossRef] 11. Margres, M.J.; Jones, M.E.; Epstein, B.; Kerlin, D.H.; Comte, S.; Fox, S.; Fraik, A.K.; Hendricks, S.A.; Huxtable, S.; Lachish, S.; et al. Large-effect loci affect survival in Tasmanian devils (Sarcophilus harrisii) infected with a transmissible cancer. Mol. Ecol. 2018, 27, 4189–4199. [CrossRef] Pathogens 2022, 11, 351 16 of 17 12. Ruiz-Aravena, M.; Jones, M.E.; Carver, S.; Estay, S.; Espejo, C.; Storfer, A.; Hamede, R.K. Sex bias in ability to cope with cancer: Tasmanian devils and facial tumour disease. Proc. R. Soc. B Biol. Sci. 2018, 285, 20182239. [CrossRef] [PubMed] 13. Alberts, B.; Johnson, A.; Lewis, J.; Raff, M.; Roberts, K.; Walter, P. T Cells and MHC Proteins. In Molecular Biology of the Cell, 4th ed.; Garland Science: New York, NY, USA, 2002. 14. Neefjes, J.; Jongsma, M.L.M.; Paul, P.; Bakke, O. Towards a systems understanding of MHC class I and MHC class II antigen presentation. Nat. Rev. Immunol. 2011, 11, 823–836. [CrossRef] [PubMed] 15. Siddle, H.V.; Marzec, J.; Cheng, Y.; Jones, M.; Belov, K. MHC gene copy number variation in Tasmanian devils: Implications for the spread of a contagious cancer. Proc. R. Soc. B Biol. Sci. 2010, 277, 2001–2006. [CrossRef] [PubMed] 16. Cheng, Y.; Stuart, A.; Morris, K.; Taylor, R.; Siddle, H.; Deakin, J.; Jones, M.; Amemiya, C.T.; Belov, K. Antigen-presenting genes and genomic copy number variations in the Tasmanian devil MHC. BMC Genom. 2012, 13, 87. [CrossRef] [PubMed] 17. Kreiss, A.; Cheng, Y.; Kimble, F.; Wells, B.; Donovan, S.; Belov, K.; Woods, G.M. Allorecognition in the Tas harrisii), an Endangered Marsupial Species with Limited Genetic Diversity. PLoS ONE 2011, 6, e22402. 18. Pye, R.; Hamede, R.; Siddle, H.V.; Caldwell, A.; Knowles, G.W.; Swift, K.; Kreiss, A.; Jones, M.; Lyons, A.; Woods, G.M. Demonstration of immune responses against devil facial tumour disease in wild Tasmanian devils. Biol. Lett. 2016, 12, 20160553. References The immunopeptidomes of two transmissible cancers and their host have a common, dominant peptide motif. Immunology 2021, 163, 169–184. [CrossRef] Caldwell, A.; Siddle, H.V. The role of MHC genes in contagious cancer: The story of Tasmanian devils. Im 537–545. [CrossRef] , H.V. The role of MHC genes in contagious cancer: The story of Tasmanian devils. Immunogenetics 2017, 69 32. Caldwell, A.; Coleby, R.; Tovar, C.; Stammnitz, M.R.; Kwon, Y.M.; Owen, R.S.; Tringides, M.; Murchison, E.P.; Skjødt, K.; Thomas, G.J.; et al. The newly-arisen Devil facial tumour disease 2 (DFT2) reveals a mechanism for the emergence of a contagious cancer. eLife 2018, 7, e35314. [CrossRef] 33. Pye, R.; Pemberton, D.; Tovar, C.; Tubio, J.; Dun, K.; Fox, S.; Darby, J.; Hayes, D.; Knowles, G.W.; Kreiss, A.; et al. A second transmissible cancer in Tasmanian devils. Proc. Natl. Acad. Sci. USA 2016, 113, 374–379. [CrossRef] 34. Kwon, Y.M.; Gori, K.; Park, N.; Potts, N.; Swift, K.; Wang, J.; Stammnitz, M.R.; Cannell, N.; Baez Evolution and lineage dynamics of a transmissible cancer in Tasmanian devils. PLoS Biol. 2020, 18, e3 34. Kwon, Y.M.; Gori, K.; Park, N.; Potts, N.; Swift, K.; Wang, J.; Stammnitz, M.R.; Cannell, N.; Baez-Ortega, A.; Comte, S.; et al. Evolution and lineage dynamics of a transmissible cancer in Tasmanian devils. PLoS Biol. 2020, 18, e3000926. [CrossRef] [PubMed] 35. Kreiss, A.; Brown, G.; Tovar, C.; Lyons, A.; Woods, G. Evidence for induction of humoral and cytotoxic immune responses against devil facial tumor disease cells in Tasmanian devils (Sarcophilus harrisii) immunized with killed cell preparations. Vaccine 2015, 33, 3016–3025. [CrossRef] [PubMed] Evolution and lineage dynamics of a transmissible cancer in Tasmanian devils. PLoS Biol. 2020, 18, e3000926. [CrossRef] [PubMed] 35. Kreiss, A.; Brown, G.; Tovar, C.; Lyons, A.; Woods, G. Evidence for induction of humoral and cytotoxic immune responses against devil facial tumor disease cells in Tasmanian devils (Sarcophilus harrisii) immunized with killed cell preparations. Vaccine 2015, 33, 3016–3025. [CrossRef] [PubMed] 36. Patchett, A.L.; Tovar, C.; Blackburn, N.B.; Woods, G.M.; Lyons, A.B. Mesenchymal plasticity of devil facial tumour cells during in vivo vaccine and immunotherapy trials Immunol Cell Biol 2021 99 711 723 [CrossRef] [PubMed] 35. Kreiss, A.; Brown, G.; Tovar, C.; Lyons, A.; Woods, G. Evidence for induction of humoral and cytotoxic immune responses against devil facial tumor disease cells in Tasmanian devils (Sarcophilus harrisii) immunized with killed cell preparations. Vaccine 2015, 33, 3016–3025. [CrossRef] [PubMed] 36. 44. Tovar, C.; Obendorf, D.; Murchison, E.P.; Papenfuss, A.T.; Kreiss, A.; Woods, G.M. Tumor-Speciﬁc Diagnostic Marker for Transmissible Facial Tumors of Tasmanian Devils: Immunohistochemistry Studies. Vet. Pathol. 2011, 48, 1195–1203. [CrossRef] [PubMed] 38. Belov, K. The role of the Major Histocompatibility Complex in the spread of contagious cancers. Mamm. Genome 2011, 22, 83–90. [CrossRef] References Patchett, A.L.; Tovar, C.; Blackburn, N.B.; Woods, G.M.; Lyons, A.B. Mesenchymal plasticity of dev in vivo vaccine and immunotherapy trials. Immunol. Cell Biol. 2021, 99, 711–723. [CrossRef] [PubM 37. Murgia, C.; Pritchard, J.K.; Kim, S.Y.; Fassati, A.; Weiss, R.A. Clonal Origin and Evolution of a Transmissible Cancer. Cell 2006, 126, 477–487. [CrossRef] 17 of 17 17 of 17 Pathogens 2022, 11, 351 38. Belov, K. The role of the Major Histocompatibility Complex in the spread of contagious cancers. Mamm. Genome 2011, 22, 83–90. [CrossRef] 39. Tovar, C.; Pye, R.; Kreiss, A.; Cheng, Y.; Brown, G.K.; Darby, J.; Malley, R.C.; Siddle, H.V.T.; Skjødt, K.; Kaufman, J.; et al. Regression of devil facial tumour disease following immunotherapy in immunised Tasmanian devils. Sci. Rep. 2017, 7, srep43827. [CrossRef] [ ] 40. Djurisic, S.; Hviid, T.V.F. HLA Class Ib Molecules and Immune Cells in Pregnancy and Preeclampsia. Front. Immunol. 2014, 5, 652. [CrossRef] [ ] 41. Tersigni, C.; Meli, F.; Neri, C.; Iacoangeli, A.; Franco, R.; Lanzone, A.; Scambia, G.; Di Simone, N. Role of Human Leukocyte Antigens at the Feto-Maternal Interface in Normal and Pathological Pregnancy: An Update. Int. J. Mol. Sci. 2020, 21, 4756. [CrossRef] [ ] 42. Jumper, J.; Evans, R.; Pritzel, A.; Green, T.; Figurnov, M.; Ronneberger, O.; Tunyasuvunakool, K.; Bates, R.; Žídek, A.; Potapenko, A.; et al. Highly accurate protein structure prediction with AlphaFold. Nature 2021, 596, 583–589. [CrossRef] 43. Murchison, E.P.; Tovar, C.; Hsu, A.; Bender, H.S.; Kheradpour, P.; Rebbeck, C.A.; Obendorf, D.; Conlan, C.; Bahlo, M.; Blizzard, C.A.; et al. The Tasmanian Devil Transcriptome Reveals Schwann Cell Origins of a Clonally Transmissible Cancer. Science 2010, 327, 84–87. [CrossRef] 44. Tovar, C.; Obendorf, D.; Murchison, E.P.; Papenfuss, A.T.; Kreiss, A.; Woods, G.M. Tumor-Speciﬁc Diagnostic Marker for Transmissible Facial Tumors of Tasmanian Devils: Immunohistochemistry Studies. Vet. Pathol. 2011, 48, 1195–1203. [CrossRef] [PubMed]
https://openalex.org/W4382130980	https://link.springer.com/content/pdf/10.1557/s43577-023-00530-7.pdf	English	null	Advanced in vivo fluorescence microscopy of neural electronic interface	MRS bulletin	2,023	cc-by	12,056	Advanc microsc Qianru Yang Neuroelectroni treatment of n interfaces. Des challenges of mechanisms fo neural tissue an around the neu developing and neural interfaci microscopy sys well as recent f modalities. Fina Advanc microsc Qianru Yang Neuroelectroni treatment of n interfaces. Des challenges of mechanisms fo neural tissue an around the neu developing and neural interfaci microscopy sys well as recent f modalities. Fina doi:10.1557/s43577-023-00530-7 Qianru Yang, University of Pittsburgh Swanson School of Engineering, Pittsburgh, USA; qiy24@pitt.edu X. Tracy Cui , University of Pittsburgh Swanson School of Engineering, Pittsburgh, USA; xic11@pitt.edu Qianru Yang and X. Tracy Cui* Neuroelectronic devices are essential tools in neuroscience research, diagnosis, and/or treatment of neurological diseases, as well as in neuro-prosthetics and brain–computer interfaces. Despite a long history of application, neuroelectronic devices are still facing challenges of unsatisfactory chronic stability and a lack of understanding of cellular mechanisms for recording and stimulation. To improve the information transfer between the neural tissue and electronic devices, a comprehensive understanding of the biological activities around the neural electrode is critical. In vivo fluorescent microscopy technologies are rapidly developing and have revolutionized our understanding of cellular dynamics in response to neural interfacing materials. Here, we will provide an overview of the in vivo fluorescence microscopy systems and imaging configurations for studying the neural electronic interface, as well as recent findings in biological mechanisms learned using these advanced optical imaging modalities. Finally, we will discuss the current challenges and future directions. and neuronal loss and neurodegeneration.7 These responses are believed to contribute to the low yield and degradation of neural recording. Traditionally, postmortem histochemistry has been the gold standard to evaluate biological responses toward neural implants because of the rich choices of chemi- cal compounds and antibodies that can label many types of biomolecules and cells. However, postmortem histochemistry only provides a snapshot at discrete end points, omitting the dynamic information in between. Additionally, conventional histochemistry requires removal of the implant for tissue slicing and staining, which inevitably damages the original tissue–device interface. Device-capture histology maintains the electrode–tissue interface by slicing the tissue parallel to the probe shank in thick sections (350–450 μm).8 This method allows direct visualization of the probe in situ, but is still a postmortem approach.l Advanced in vivo fluorescence microscopy of neural electronic interface Qianru Yang and X. Tracy Cui* Brain surface electrodes Confocal microscopy reduces the out-of-focus signals by focusing the illumination light to the focal point and placing a pinhole before the detector, enabling optical sectioning of thick tissue.13 However, due to scattering and absorption of light in the biological tissue, the imaging depth is limited. One approach to increase imaging depth is to extend the excita- tion and emission wavelengths to 1000–1700 nm, the second near-infrared window (NIR-II).14 Together with development of long wavelength fluorescent probes and improvement of single-photon detectors, the imaging depth of confocal micros- copy has recently demonstrated up to 1.7 mm below the sur- face of the mouse brain.15,16 Advancements in fabricating transparent electrode arrays have allowed optical imaging directly through a brain surface electrode. After craniotomy, the electrodes can be positioned above or underneath the dura to record the electrocorticogram (ECoG) signals and to deliver electrical stimulation (Fig- ure 1a). The electrodes are usually flexible, sheet-like, and consist of several layers of different materials. Typically, the conductive material is sandwiched between insulation layers, except for the electrode sites and the bonding pads. The insula- tion materials need to be nonconductive, flexible, transparent, and biocompatible. Popular polymers for the insulation sub- strate are poly(dimethylsiloxane) (PDMS), parylene C, par- ylene HT, SU-8, polyimide, liquid crystal polymer, hydrogel,24 etc. Different materials have different properties and different ways of fabrication that could influence their applications.25 The commonly used conductive materials include metals, carbon/graphene,26,27 indium tin oxide (ITO),28 and poly(3,4- ethylenedioxythiophene) (PEDOT)-based conducting poly- mers. Metallic materials, such as platinum (Pt), gold (Au), and iridium/iridium oxide (Ir/IrOx), have a long history of mak- ing neural electrodes due to their high conductivity, chemical inertness, high charge-storage capacity, and good biocompat- ibility. However, metallic materials are not transparent. The Multiphoton laser scan microscopy has been the most widely employed high-resolution in vivo fluorescent imaging method. Multiphoton microscopy utilizes a laser with a dou- bled or tripled long wavelength compared to single-photon fluorescent microscopy. The fluorescence excitation requires two or more photons to arrive at the point simultaneously, which significantly reduces the out-of-focus fluorescence excitation therefore decreases out-of-focal plane noises and is safer for long-term imaging than traditional single- photon microscopy. In addition, the longer laser wavelength can penetrate deeper into the scattering tissue. Two-photon a b c Figure 1. Experiment setup configurations for imaging neural electrode tissue in vivo. (a) Brain surface electrode. (b) Penetrating electrode. Introduction Since the discovery of action potentials, neural electrodes have been vital tools to read out and modulate neural activi- ties. By recording and stimulating with implanted electrodes, researchers can determine the functions and connections of different brain regions.1,2 The clinical value of brain electrodes was found first in treating tremor with deep brain stimula- tion (DBS), followed by successful treatments of a variety of neurological diseases, such as depression, epilepsy, pain, addiction, dementia, and stroke recovery.3 Neural prosthetics based on implanted neural electrodes are evolving at a rapid pace and have shown great potential in restoring functions of the central nervous system (CNS), including motor, visual, auditory, somatosensory sensation, etc.4–6 Despite these prog- resses, the mechanisms of neural prosthetics at the cellular level and the interactions between the implanted device and host tissue remain poorly understood, and the elucidation of such is absolutely important for improving their efficiency, precision, naturalness, and safety. Recent technology advancements in in vivo fluorescence microscopy allow investigations of real-time and long-term changes of cellular activity in living animals. Here, we will discuss the currently available in vivo fluorescence microscopy technologies for studying cellular activities around implanted neural electronics and findings obtained using these imaging technologies. To date, all of the neural interfacing implants are still facing a significant challenge—the foreign body responses of brain tissue, including blood–brain barrier (BBB) disruption, microg- lia activation, macrophage infiltration, astrogliosis, fibrosis, 506 MRS BULLETIN • VOLUME 48 • MAY 2023 • mrs.org/bulletin © The Author(s) 2023 Advanced in vivo fluorescence microscopy of neural electronic interface In vivo fluorescence microscopyl microscopy typically can reach 500-µm deep, and three-photon microscopy about 900-µm deep, and using longer wavelength could further improve the imaging depth to ~1.6 mm.17,18 Despite the high resolution and deep optical penetration, the imaging speed and the field of view of traditional multiphoton microscopy is limited to <10 Hz and <1 mm2.19 Methodology advancements of multiphoton microscopy are heading toward large field of view, kilohertz scan rate, multiplane, multiregion, and miniaturization for imaging in free-behaving animals.20–23 l In vivo fluorescent imaging is a well-established and widely used method to study biological systems. There are three most popular types of in vivo fluorescence imaging methods: wide field microscopy, confocal microscopy, and multipho- ton microscopy. Wide field microscopy illuminates a large field (typically enough to cover the whole cranial window on rodents) and collects emission light by a camera on a fluo- rescence microscope.9 It is relatively low cost and has faster imaging speed compared to confocal and multiphoton micros- copy, but lacks depth specificity because it receives emitted light from both in-focus and out-of-focus planes, making it difficult to visualize fine cellular structures in thick tissues.10 Technological improvements for high-resolution wide field microscopy are growing.11 For instance, structured illumina- tion and reconstruction algorithms can remove out-of-focus signals, and when combined with adaptive optics, have achieved synapse-resolving imaging in vivo.10,12 In vivo imaging configurations for electrode– tissue interface studiesi Currently, there are three major types of setup configurations for visualizing brain tissue and neural electronic device interfaces with optical imaging systems differentiated based on the position of the electrode and optical light path (Figure 1). Brain surface electrodes The photoelectric effect happens when the light hits the electrode material (sites and traces), electrons in the material absorb the energy and are emitted, producing a current that can be captured by the recording equipment.34 The amplitude of the artifact has been shown to be related with the laser power and scan rate.35,36 This artifact is also stronger for metal materials because the electrons in metal have lower binding energy than nonmetal conductors.27,28,36 Increasing transparency can effectively reduce the photoelectrical artifact. When the amplitude of the photoelectric artifact does not saturate the recording ampli- fier, it is possible to remove the recording artifact by post- processing algorithms, such as multiple band-stop filters or subtracting an average artifact template based on laser scan- ning triggers.35,37 In the chronic setup, the cranial window can be sealed with transparent medium and a cover glass on top. The clarity of the window, however, can be difficult to maintain.43 For exam- ple, it was shown that the imageable depth drops from 600 to 300 µm within a week.43 The decreased window clarity and imaging depth resulted from cascades of host tissue responses including bleedings, bone regrowth, fibrous tissue thickening, recruitment of inflammatory cells at meninges and around the electrode, and sometimes infections due to contaminations or incomplete seal.48 Placing the cover glass intimately onto the brain surface is the key to keeping a clear window by physically inhibiting tissue regrowth. Using a cover glass with a small hole and inserting silicon electrodes through the hole afterward, Mols et al. showed two-photon imaging of the electrode down to 500 µm below the pia after two-week implantation and 200 µm below the pia after 10-week implan- tation.49 For flexible electrodes, it could be slightly easier to keep a clear window over time because the cover glass can be placed more seamlessly onto the brain surface.44,45 Addition- ally, ultraflexible electrodes are expected to cast less chronic mechanical stress to the brain compared to stiff electrodes. Compared to intracortical microelectrodes, brain surface electrodes usually are limited to recording local field potentials (LFPs) and multiunit activity. Resolving single unit activities from surface electrodes is difficult because of the distance to neuronal somas. On the other hand, they are considered less invasive and introduce less damage to the underlying brain parenchyma tissue. Brain surface electrodes (c) Microprism combined with microelectrode array (MEA). c c Figure 1. Experiment setup configurations for imaging neural electrode tissue in vivo. (a) Brain surface electrode. (b) Penetrating electrode. (c) Microprism combined with microelectrode array (MEA). MRS BULLETIN • VOLUME 48 • MAY 2023 • mrs.org/bulletin 507 Advanced in vivo fluorescence microscopy of neural electronic interface accessible region for optical imaging is often limited to areas outside of metal parts. To improve the light transmittance of metal parts, several groups have developed different methods to create mesh-shaped metal.29–31 There are also an increasing number of studies of other materials with better optical trans- parency. For example, the carbon-nanotube/PDMS electrode array has a greater than 85% transmittance across light wave- lengths from 400 to 2500 nm.32 However, the electrochemical impedance of most of the transparent materials is high. Surface modification with conductive polymers or nano-sized metal have been used to improve the electrochemical impedance of transparent materials at the cost of decreased transparency.26,27 the soft brain tissue, can also penetrate the brain, but usually need an accessary shuttle (a needle or a metal wire) or some temporary stiffening strategies to facilitate the insertion.44–46 When the electrodes are transparent or sufficiently thin, they could be inserted more vertically without significantly affect- ing the imaging.45 In the acute setup, the cranial window can be left open for at least 6 h.47 The advantage of having an open cranial win- dow is to be able to rinse the blood out or make adjustment of the implant. However, with the skull open, the brain tissue is essentially exposed to pollution and pathogens in the environ- ment that could influence the brain activity. Alternatively, the cranial window can be sealed with a cover glass and transpar- ent filler such as saline, artificial cerebrospinal fluid (ACSF), silicone, agar, etc. Solid filler can also reduce the micromotion of the brain, thus improving the image stability. p p y Multiphoton imaging and concurrent electrophysiological recording/stimulation through transparent ECoG arrays have been demonstrated in animals.33 The animals are either trans- genically labeled with fluorescent proteins or locally injected with virus or dyes to indicate neural activity. One big chal- lenge for concurrent multiphoton imaging and electrophysiol- ogy recording is the photoelectric artifact. Electrodes combined with penetrating optical devices To date, in vivo imaging of surface and penetrating electrodes has been mostly limited to superficial layers (layer 1–3) of the cortex. In order to see deeper brain tissue, researchers have been developing optical devices that can break the imaging depth limit of conventional configurations. Optical fiber, microlens, and microprism are the three commonly used opti- cal devices for deep brain imaging.i Optical fiber can be implanted to any depth into the brain and/or interface with other optical devices, allowing imag- ing in free-behaving animals due to its flexibility.50 The same optic fiber can both deliver excitation light into the brain tissue and collect emission signals. The conventional optical fibers are based on silica, whereas other materials such as silk,51 hydrogel, and synthetic polymers have also been investi- gated.52 There are several categories of optic fiber, including conventional step-index fiber, fiber bundles, tapered optical Brain surface electrodes Both epidural and subdural implants are usually encapsulated by collagenous tissue and inflammatory cells further increasing the distance to cortical neurons.38,39 Softer implants that are more conformal to the brain curvature could reduce mechanical stress and fibrous encapsulation.40 Yet, chronic in vivo imaging studies of brain surface electrode arrays have been rare.41 Studies of neurons and neuron–electrode interaction Studies of neurons and neuron–electrode interaction All brain functions, from locomotion, sensing to emotion, learning, and memory, are carried out by neuronal activities. It is estimated that an electrode can detect neuronal action potential signals within a cylinder of 140 µm radius, which contains ~1000 neurons in layer 2/3 of rat cortex.76–78 Yet in practice, only a few neurons can be reliably identified through electrical recording on each electrode site. This large dis- crepancy is hypothesized to be due to neural damage by the implanted electrode, or presence of silent neurons or signal amplitude being too small.78 Electrophysiology recording alone cannot pinpoint these mechanisms because the record- able neuron for an electrode is not only dependent on the dis- tance between the electrode and neuron soma, but also related to the anatomical structures. For example, action potentials from layer 5 pyramidal neurons can be recorded up to 400 µm from the soma along the apical dendrite, but not detectable at 100 µm in the opposite direction.79 Combining in vivo imag- ing of neural activity and electrophysiology recording has the potential to better uncover the ground truth of extracellular recording. Simultaneous neural calcium imaging and elec- trophysiology recording have been done with surface ECoG arrays and intracortical microelectrodes.27,70 Yet, there has not been a conclusive study correlating electrophysiologi- cal recording with all recordable neural population activity imaging. With continuous technology advancement in fast, volumetric in vivo calcium/voltage imaging, and fabrication of transparent, flexible, and low impedance microelectrode array, this problem is hopefully to be resolved in the near future. Microprisms are tiny glass prisms, with size customizable from 500 µm to several millimeters. The coating (usually alu- minum) on the slant side will reflect light at 90 degrees, chang- ing the focal plane from horizontal to vertical. This vertical view is favorable for high resolution imaging of interlaminar activities.67 Microprisms can be directly glued under the cover glass or combined with GRIN lens or optic fibers.68,69 Flex- ible electrode arrays can be attached to the microprism imag- ing face after bending to comply the 90 degree angle formed between the microprism and the cover glass70 (Figure 1c). The footprint of optical devices is usually much larger than microelectrode arrays, as such tissue damages caused by implanting these optical devices bring about some concerns. Penetrating electrodes To date, the most common imaging setup for studying for- eign body responses and extracellular electrophysiology is to insert a silicon electrode device into the brain parenchyma at an angle between 30 and 45 degrees to the brain surface with electrode sites facing up42,43 (Figure 1b). This allows the visualization of the electrode sites interfacing with the brain. Ultraflexible electrodes, developed for seamless integration in 508 MRS BULLETIN • VOLUME 48 • MAY 2023 • mrs.org/bulletin Advanced in vivo fluorescence microscopy of neural electronic interface fiber, gradient refractive index (GRIN) fiber, etc.53,54 The classic step-index fibers deliver light based on total internal reflection between the outer cladding and the inner core with distinct refractive indexes. Fiber bundles consist of up to ~100,000 densely packed step-index fibers to transform pix- ilated images.55 The GRIN fiber has refractive index declines with radius from the central axis, so that light propagates at the same speed across different locations within the fiber. The tapered optical fiber has a special tapered tip compared to the traditional flat-cleaved tip, which can extend the light illumi- nation and collection region from a restricted volume near the fiber tip to ~2 mm along the taper.56,57 The size of optical fiber is customizable, ranging from ~10 µm to several millimeters in diameter.58 To enable electrophysiology, microwire elec- trodes (tetrodes) and flexible electrode arrays can be adhered to the side wall of the optic fibers.59 Alternatively, multifunc- tional fibers have been fabricated by combining different components and materials for simultaneous electrophysiology and optical simulation/drug delivery.60 However, optical fiber- based photometry has limitations on the spatial resolution as they measure integrated photon count from a large population of cells near the probe tip (within ~500 µm).61 absorption by red blood cells. It has been reported that bleed- ing and vascular remodeling take about two weeks to settle down after the 1-mm-microprism implantation,71 while the microglial responses need four to eight weeks to return to the homeostatic state.72 In addition, because most optical devices are rigid and need to be anchored to the skull, the mechani- cal mismatch between the brain tissue and the optical glasses increases the possibility of postsurgery damage due to micro- motion.73 Soft optical fibers have shown reduced inflammatory responses compared to rigid fibers.59,74,75 The foreign body responses after implanting optical devices with various shapes and materials await further examination. Recent biological findings at neural electrode– tissue interface The research on brain tissue and electronic interface using the advanced imaging modalities can be categorized in two major directions. One direction is to characterize and understand the foreign body responses in order to improve the integration and biocompatibility of neural electrodes. The other direc- tion is focused on investigating the cellular mechanisms of electrophysiology recording and stimulation for better decod- ing and encoding brain activities. Here, we will summarize discoveries from in vivo imaging modalities in both directions in recent years. Microlenses, including miniaturized objective lenses and GRIN lenses, can provide better spatial resolution and sig- nal efficiency than optic fibers. Objective lenses can provide the best resolution and fluorescent efficiency, but the minia- turization is challenging (minimally ~3 mm in diameter).62 GRIN lenses can be miniaturized to 350–1000 µm in diam- eter and can be customized to lengths from several milli- meters to several centimeters to approach deep brain regions in rodents or larger animal models.63 GRIN lenses can be used for wide field, confocal, multiphoton microscopy, or integrated with optic fibers.63–65 Similar to optical fibers, stiff and flexible electrodes can be placed on the side wall of GRIN lenses.66 Advanced in vivo fluorescence microscopy of neural electronic interface The inserted electrode directly ruptures cell membranes and blood vessels, leading to neuronal death near the implanted device as soon as 1 h postimplantation.80 The neuronal density drop was observed most dramatically in the first two weeks.81 postinsert within 150 µm, then goes down and stabilizes within a minute84 (Figure 2a–b). After 6 min, the calcium level in neural somas subsided to baseline, while the neurites in layer 1 showed sustained high calcium over 27 min.84 This calcium elevation is thought to result from the mechanical distortion of neurons and neurites, potentially through mechanoporation,85 opening of voltage-gated calcium channels triggered by mechanically sensitive sodium channels,86 or intracellular calcium release.87 From 1 to 3 h postinsertion, spherical protrusions were observed in the cortical layer 1, suggesting damage to the neurites.84,88 Additionally, the visually evoked calcium activity around the electrode declined gradually and became unresponsive after two weeks of implantation.88 In a longer-term study, dendritic arbors of fluorescently labeled layer 5 pyramidal neurons near a multicontact electrode shank were tracked in Thy1-YFP transgenic mice using two-photon microscopy (TPM), where persistent dendritic atrophy was observed for three months.89 The inserted electrode directly ruptures cell membranes and blood vessels, leading to neuronal death near the implanted device as soon as 1 h postimplantation.80 The neuronal density drop was observed most dramatically in the first two weeks.81 After two weeks, the neural degeneration marker was found mostly in neural processes rather than the cell body, and was increasing over time.80 In addition to neural degeneration, neural activity and excitability could change after electrode implantation.82 These studies are limited to discrete timepoints due to the postmortem nature of histology method. After two weeks, the neural degeneration marker was found mostly in neural processes rather than the cell body, and was increasing over time.80 In addition to neural degeneration, neural activity and excitability could change after electrode implantation.82 These studies are limited to discrete timepoints due to the postmortem nature of histology method. Advanced in vivo fluorescence microscopy of neural electronic interface In vivo multiphoton imaging of neural calcium activity provided an important method to study neuronal health and activity after device implantation in real time, which can be correlated with electrophysiology recording/stimulation.83 Calcium imaging has been extensively used in neuroscience as calcium influx is associated with most neural depolarization electrical activity.19 During implantation of microelectrodes in mice cortex, both neuron soma and neurites exhibit abnormally prolonged high calcium levels.84 Specifically, the neural calcium activity was observed to peak in several seconds Electrical stimulation has been widely used in neurosci- ence research to dissect neural circuitries, and in clinical a b c d Figure 2. In vivo two-photon microscopy of neural calcium responses to probe insertion and electric stimulation with different surface materials. (a) Electrode array implantation causes neuronal calcium elevation in the vicinity of the implant site. (b) Calcium signal intensity within the dotted blue box in panel (a) showing a steep increase and gradual decrease after implantation. Adapted with permission from Reference 84. (c) Calcium responses to a 35 μA stimulation on electrode sites from a single-shank array. The top row shows responses from IrOx sites. The bottom row shows responses from their neighbor PEDOT/carbon nanotube (CNT) (PC) sites, 50 μm away. (d) Mean normalized calcium responses as a function of stimulation amplitude from two materials. A significantly higher response was observed from the PC electrode at all current amplitude tested demonstrating higher efficiency of the PEDOT/CNT electrodes. Adapted with permission from Reference 107. a b a b b c d d c Figure 2. In vivo two-photon microscopy of neural calcium responses to probe insertion and electric stimulation with different surface materials. (a) Electrode array implantation causes neuronal calcium elevation in the vicinity of the implant site. (b) Calcium signal intensity within the dotted blue box in panel (a) showing a steep increase and gradual decrease after implantation. Adapted with permission from Reference 84. (c) Calcium responses to a 35 μA stimulation on electrode sites from a single-shank array. The top row shows responses from IrOx sites. The bottom row shows responses from their neighbor PEDOT/carbon nanotube (CNT) (PC) sites, 50 μm away. (d) Mean normalized calcium responses as a function of stimulation amplitude from two materials. A significantly higher response was observed from the PC electrode at all current amplitude tested demonstrating higher efficiency of the PEDOT/CNT electrodes. Adapted with permission from Reference 107. Studies of neurons and neuron–electrode interaction Optic fibers or lenses smaller than 1 mm in diameter could be implanted without removing brain tissue but would squeeze or push tissue to the side or to the bottom, resulting in mechani- cal damage and chronic compression stress. Devices larger than 1 mm could require removing part of the brain tissue to reduce the compressive damage. During insertion, the opti- cal devices will inevitably damage some blood vessels, and the bleeding will cast a shadow in the images due to light MRS BULLETIN • VOLUME 48 • MAY 2023 • mrs.org/bulletin 509 509 Imaging glial responses to implantsl Advances in two-photon imaging and transgenic mice that express fluorescent proteins in microglia reveal that in resting state, microglial processes dynami- cally survey around while the soma stay immobile.111 These moving processes will make intimate but transient contact with neuronal synapses.112 This surveillance of microglial processes is dependent on K+ channel.113 Once the brain is damaged or injured, surrounding microglia respond rapidly by sending processes toward the injury within several minutes, then depending on the degree of damage, they could become motile and migrate to the injury in several hours, phagocytose cell debris and proliferate in the following days, and remain residing at the injury site for months.42,72,114,115 The direc- tional response of microglial processes is potentially regulated by ATP release from damaged tissue through G-protein coupled purinergic P2Y receptors.116 Anesthesia can influence microglial surveillance and activation.117 Hence, it is important to take careful consideration of the effect of anesthesia while doing live imaging. When possible, imaging in awake animals is recommended for unbiased microglial analysis tional response of microglial processes is potentially regulated by ATP release from damaged tissue through G-protein coupled purinergic P2Y receptors.116 Anesthesia can influence microglial surveillance and activation.117 Hence, it is important to take careful consideration of the effect of anesthesia while doing live imaging. When possible, imaging in awake animals is recommended for unbiased microglial analysis. Microstimulation with high spatial resolution requires advanced electrode materials with high charge injection capacity and stability. As new electrode materials are devel- oped, characterization of the stimulation performance is nec- essary. Although evaluation of stimulating materials has been conventionally done via in vitro and in vivo electrochemical tests, electrophysiology and histology, in vivo imaging can offer much higher temporal and spatial resolution. To com- pare newly developed poly(3,4-ethylenedioxythiophene) (PEDOT)/carbon nanotubes (CNTs) (PC) electrode coating to clinically applied iridium oxide (IrOx), Zheng et al. coated silicon probes with PC and IrOx on electrode sites in alter- nation and evaluated the neuronal calcium activation upon electrical stimulation in head-fixed behaving GCaMP mice for up to 12 weeks.106,107 Surprisingly, PC electrodes activated a larger and brighter GCaMP response than IrOx (Figure 2c–d) at the same current amplitude, showing higher stimulation efficiency from the PC electrodes.107 Abnormal ictal corti- cal responses or cortical spreading depression were observed through the imaging using stimulation parameters commonly considered safe in intracortical stimulation, which could oth- erwise be overlooked in conventional characterization setup. Imaging glial responses to implantsl treatment of neurological diseases and sensory restora- tions.6,90–94 However, what happens between the electri- cal stimulation and behavior outcomes remains largely unknown. In addition, continuous stimulation can lead to neural reorganization and damage.95,96 There are emerging studies utilizing simultaneous in vivo imaging and electri- cal stimulation to study the cellular mechanism of electrical stimulation. Early electrophysiology and modeling works suggest that increasing stimulation current leads to increas- ing radius of neural activation,97,98 but Histed et al. revealed sparsely distributed neural activation even at low currents using in vivo two-photon imaging around the inserted metal wire electrode in layer 2/3.99 It is explained that distant neu- rons are activated via neural fibers passing by the electrode tip. In contrast, ex vivo studies indicated a laterally con- fined activation field of cortical microstimulation mainly through transsynaptic activation.100,101 Using two-photon calcium imaging and iridium microelectrodes on silicon probes inserted into layer 2/3 of mouse cortex, the Kozai group has published a series of works looking at the effects of stimulation parameters and patterns on the spatial and temporal dynamics of neural activity.102–105 Although these studies began to systematically characterize the neuronal response to microstimulation paradigms, the imaging is only limited to superficial brain layers. By combining transparent and flexible microelectrode array with microprism implanta- tion, Yang et al. expanded the imaging depth and volume, and illustrated the depth dependent neuronal responses of stimulation amplitudes and frequencies across all cortical layers in awake mice.70 As the resident inflammatory cells in the brain, microglia respond rapidly to neural electrode implantation and can initi- ate cascades of inflammatory responses. In immunohistochem- istry studies, activated microglia are found in the vicinity of an implanted electrode, usually surrounded by a dense layer of astrocytes.108,109 Microglia and astrocytes together form the “glial scar” around the electrode, which is believed detrimental to the signal transduction between the electrode and the neural tissue.110 Microglial morphology is highly associated with their functional status. In the homeostatic state, microglia have ramified processes protruding almost evenly to all directions. When the brain is damaged, microglia transform to a phagocytic state featured with enlarged cell bodies and shortened processes. Advanced in vivo fluorescence microscopy of neural electronic interface 510 MRS BULLETIN • VOLUME 48 • MAY 2023 • mrs.org/bulletin Advanced in vivo fluorescence microscopy of neural electronic interface Imaging glial responses to implantsl These studies demonstrate the potential and need to evaluate the efficiency and safety of different electrode materials using real-time imaging. g y The role of microglia in neuronal health is complex. On one hand, absence of microglia after brain injury leads to much more severe neuronal loss, showing the critical role of microglia in neuroprotection.118 On the other hand, activated microglia can secrete cytotoxic factors that result in excessive neural degeneration.119 Modulating the extent of microg‑ lia activation to the appropriate level could be the key to improving neuron survival after electrode implantation.120 Coating of neural electrodes with biomimetic protein or antifouling zwitterionic polymer have been shown to effectively reduce inflammatory responses in postmortem histology studies.121–123 Recently, using two-photon imaging, the dynamic interactions between the microglia processes and implant surface with and without coatings have been investigated (Figure 3a–b).47,124 It was revealed that a neuronal adhesion molecule L1 coating inhibited the microglia coverage on the surface of neural probes (Figure 3a).47 Zwitterionic coating also significantly decreases the microglial coverage, and the mechanism is by suppressing the spreading of microglial end-feet, but not reducing the number (Figure 3b).124 Anti-inflammatory drugs, such as melatonin MRS BULLETIN • VOLUME 48 • MAY 2023 • mrs.org/bulletin 51 511 Advanced in vivo fluorescence microscopy of neural electronic interface a b Figure 3. In vivo two-photon microscopy of microglial responses to implanted silicon probes with different coatings. (a) Microglial encapsulation after implantation of silicon probes with and without L1 protein coating. Blue squares indicate the implanted probe. Adapted with permission from Reference 47. (b) Zwitterionic PSBMA coating decreases the size but not the number of microglial end-feet on the surface of implanted silicon probe. Adapted with permission from Reference 124. Scale bar = 20 µm. a a b b Figure 3. In vivo two-photon microscopy of microglial responses to implanted silicon probes with different coatings. (a) Microglial encapsulation after implantation of silicon probes with and without L1 protein coating. Blue squares indicate the implanted probe. Adapted with permission from Reference 47. (b) Zwitterionic PSBMA coating decreases the size but not the number of microglial end-feet on the surface of implanted silicon probe. Adapted with permission from Reference 124. Scale bar = 20 µm. Imaging glial responses to implantsl and dexamethasone, have also been shown to attenuate microglial responses and improve electrophysiology record ings.125–127 than neurons.138 Moreover, the astrocyte activity is induced by noradrenergic and glutamatergic signaling from neurons.138 Oligodendrocytes generate myelin sheaths along axons in the central nervous system (CNS) that can increase the action potential propagation.139 The myelin sheath development is regulated by neural activity through calcium influx in oligodendrocytes.140 In addition to myelination, oligodendrocytes also provide neurotrophic and metabolic support to neurons.141,142 Ablation of oligodendrocytes has been shown to cause neurodegeneration and excessive inflammation.143,144 Pharmaceutical oligodendrocytes depletion impairs electrophysiology recording quality.145 A four-week long two-photon imaging study characterized the oligodendrocyte soma deformation and myelinosome formation around implanted silicon electrodes, showing a progressive myeline degeneration (Figure 5a–b).146 Besides microglia, other types of glial cells in the brain are important for maintaining brain health, including astrocytes, oligodendrocytes, and pericytes. Astrocytes have thin, branchy processes extensively contacting neural synapses and blood vessels.128 Cumulative evidence indicates that astrocyte processes sense synaptic activity and regulate blood vessel diameters accordingly.129–131 The perivascular processes of astrocytes also play a critical role in water homeostasis and blood–brain barrier permeability.132,133 Astrocytes do not produce action potentials like neurons, but they exhibit spontaneous and neural activity correlated intracellular Ca2+ increases.134,135 In pathological conditions, activated microglia induce a subtype of reactive astrocytes—termed as A1 astrocytes—by releasing cytokines Il-1α, TNF, and C1q.136 A1 astrocytes lose most of the normal neuro-supportive functions and become neurotoxic.136 The astrocyte response to microelectrode implantation has been characterized using in vivo two-photon imaging, showing that astroglia scar develops within the first two weeks and remains stable from week 2 to week 10 postimplantation (Figure 4).49,137 Two-photon in vivo calcium imaging in both neuron and astrocytes reveals that astrocytes start to show calcium response to electrical stimulation at longer pulse width and higher current intensity Oligodendrocyte precursor cells (OPCs) or NG2 + cells, the major proliferating cells in the adult CNS, continuously differentiate into myelinating oligodendrocytes in healthy tissue.147 OPCs play an important role in remyelination, as they can proliferate and differentiate into oligodendrocytes and astrocytes after demyelination.148–150 In vivo two- photon imaging illustrates that OPCs have motile filopodia surveying around, which is similar to microglial processes, but without bulbous endings and move more slowly (0.3 ± 512 MRS BULLETIN • VOLUME 48 • MAY 2023 • mrs.org/bulletin Advanced in vivo fluorescence microscopy of neural electronic interface b a Figure 4. Imaging glial responses to implantsl Astrocyte response to penetrating electrode implantation. (a) Astrocyte (green) soma area increases in the first several days postimplantation. Adapted with permission from Reference 137. Blood vessels are labeled using SR101 (red). Scale bar = 50 µm. (b) No significant increase in astrocyte response from week 2 to week 10 postimplantation. Adapted with permission from Reference 49. Scale bar = 100 µm. a a a b Figure 4. Astrocyte response to penetrating electrode implantation. (a) Astrocyte (green) soma area increases in the first several days postimplantation. Adapted with permission from Reference 137. Blood vessels are labeled using SR101 (red). Scale bar = 50 µm. (b) No significant increase in astrocyte response from week 2 to week 10 postimplantation. Adapted with permission from Reference 49. Scale bar = 100 µm. b Figure 4. Astrocyte response to penetrating electrode implantation. (a) Astrocyte (green) soma area increases in the first several days postimplantation. Adapted with permission from Reference 137. Blood vessels are labeled using SR101 (red). Scale bar = 50 µm. (b) No significant increase in astrocyte response from week 2 to week 10 postimplantation. Adapted with permission from Reference 49. Scale bar = 100 µm. 0.01 μm min−1 of OPCs151 versus 1.47 ± 0.1 μm min−1 of microglia111). In addition, most OPCs continuously change their soma locations and processes orientation except for perivascular cells,151 whereas microglia largely maintain their territory over time. After electrode insertion, OPCs also extend processes toward the implant, but at later time (>12 h) compared to microglia (within several minutes) (Figure 5c–d).152 permeability.153 Vascular damages caused by neural electrodes implantation not only disrupt nutrient and metabolic supply, but also leak the cells and molecules that are cytotoxic to brain tissues.154 For example, erythrocyte/hemoglobin entry increases iron levels in the brain, which leads to oxidative stress, excitotoxicity, and neuronal injury.154–157 Cumulative evidence suggests that a higher level of BBB damage is correlated with increased foreign body responses and worse recording performance.158–160 Two-photon microscopy has the capability to illustrate blood vessel anatomical structures at high resolution and to measure blood flow with injected intravascular fluorescence dyes.161,162 Localized bleeding as a result of neural probe insertion has been characterized using two-photon microscopy as early as 2009, while shrinkage of capillary and decrease in perfusion at the vicinity of implants have lately been documented.162 Using OIS-BOLD under wide field microscopy, decreases in blood oxygenation Implantation‑induced vascular changes Neural tissue is highly vascularized. Implanting electronic devices inevitably disrupts the vascular network and damage the BBB. The BBB regulates substance exchanges between the peripheral circulation system and the CNS. Vascular endothelial cells together with astrocytes, pericytes, neurons, and the extracellular matrix, form a “neurovascular unit” that is essential for regulating blood flow and BBB 513 MRS BULLETIN • VOLUME 48 • MAY 2023 • mrs.org/bulletin Advanced in vivo fluorescence microscopy of neural electronic interface a b c d Figure 5. Demyelination and oligodendrocyte precursor cell (OPC) responses to penetrating electrode implantation. (a) Myelinosomes (red arrows) near the implant dynamically change diameter over time. Scale bar = 20 µm. (b) Demyelination occurs near the implant over the four-week implantation. Adapted with permission from Reference 146. Green: OPCs. Red: Blood vessels (SR101). (c) OPC processes extend toward the probe surface (blue rectangle) beginning 12 h post-insertion. Scale bar = 15 μm. (d) OPCs migrate toward the surface of the probe (blue rectangle). Scale bar = 15 μm. Adapted with permission from Reference 152. a d c c d Figure 5. Demyelination and oligodendrocyte precursor cell (OPC) responses to penetrating electrode implantation. (a) Myelinosomes (red arrows) near the implant dynamically change diameter over time. Scale bar = 20 µm. (b) Demyelination occurs near the implant over the four-week implantation. Adapted with permission from Reference 146. Green: OPCs. Red: Blood vessels (SR101). (c) OPC processes extend toward the probe surface (blue rectangle) beginning 12 h post-insertion. Scale bar = 15 μm. (d) OPCs migrate toward the surface of the probe (blue rectangle). Scale bar = 15 μm. Adapted with permission from Reference 152. Characterizing meningeal response The meninges are a thin structure embedding the brain, consist- ing of three layers in general—dura, arachnoid, and pia matter. Cerebrospinal fluid (CSF) is filled in the subarachnoid space between dura and pia. Recently a fourth layer membrane was uncovered that divides the subarachnoid space into two com- partments.163 There are abundant immune cells residing the meninges, including macrophages and dendritic cells (DCs). Macrophages are numerous in dura and pia and usually respond the quickest to inflammatory stimuli.164 Although DCs are fewer in number, they are antigen-presenting cells that can stimulate and recruit T cells.165 The recent discovery of lymphatic vessels in the meninges further highlights the importance of meninges in the immune system.166,167 The meningeal fibrous encapsulations of neural implants have been well documented in postmortem histology studies for both brain surface arrays and intracortical Implantation‑induced vascular changes Our under- standing of meningeal responses are still preliminary and remain to be elucidated with more chronic in vivo imaging studies. Implantation‑induced vascular changes implants.41,168,169 Lymphangiogenesis of meningeal lymphatic vessels and increased CSF influx are more recently reported after implantation of brain surface EEG electrodes.170 Implanting neu- ral probes into the brain without connection to the meninges has been shown to significantly reduce the glial scar formation compared to trans-meninges unanchored implants, suggesting that part of the glial encapsulation originated from meninges.171 implants.41,168,169 Lymphangiogenesis of meningeal lymphatic vessels and increased CSF influx are more recently reported after implantation of brain surface EEG electrodes.170 Implanting neu- ral probes into the brain without connection to the meninges has been shown to significantly reduce the glial scar formation compared to trans-meninges unanchored implants, suggesting that part of the glial encapsulation originated from meninges.171 In vivo two-photon microscopy has been used to evaluate menin- geal inflammatory responses and meningeal collagen remodeling around implanted silicon probes.48 The collagen fiber is visu- alized by second-harmonic generation signal. It is shown that the meningeal inflammatory cell migration along electrodes is most robust during the first day and attenuate by two days post- implantation, while the collagen continues to grow over the next month.48 Meninges is a potentially powerful target region that can relieve fibrous encapsulation of chronic implants. Our under- standing of meningeal responses are still preliminary and remain to be elucidated with more chronic in vivo imaging studies. were observed around the implanted electrode on day 2 and returned to normal levels by day 7.88 Decreases in blood perfusion could contribute to decreased neural activity in the injured tissue area. With chronic two-photon microscopy, it has been shown that angiogenesis and vascular remodeling are most pronounced in the first two weeks after intracortical implantation of flexible electrodes, coincident with a rapid improvement of electrophysiology recording performance.45 Chronic two-photon imaging of both microglia and blood vessels reveals persistent microglia aggregations around ablated capillaries, suggesting that BBB breakage even in a single microvessel could lead to long-term inflammation.72 In vivo two-photon microscopy has been used to evaluate menin- geal inflammatory responses and meningeal collagen remodeling around implanted silicon probes.48 The collagen fiber is visu- alized by second-harmonic generation signal. It is shown that the meningeal inflammatory cell migration along electrodes is most robust during the first day and attenuate by two days post- implantation, while the collagen continues to grow over the next month.48 Meninges is a potentially powerful target region that can relieve fibrous encapsulation of chronic implants. Current limitations and future directions Optical methods for in vivo imaging of neural electrode arrays have been vastly progressing in recent years. Direct two-pho- ton imaging of silicon probes inserted at 30–45 degrees in a cranial window has been a popular setup that greatly improved our understanding of dynamic properties of different types of cells in response to electrode implantation. Although it is a facile model, the optically accessible depth is limited to layer 2/3 of cortex and decreases over chronic implantation. Intro- ducing optic fibers, GRIN lenses, or microprisms could help us 514 MRS BULLETIN • VOLUME 48 • MAY 2023 • mrs.org/bulletin Advanced in vivo fluorescence microscopy of neural electronic interface gain greater imaging depth. While there are emerging studies on combining optical devices with electrode arrays, most of them are at the stage of proof-of-concept demonstration. Sev- eral areas can be further improved for broad applications: (1) The flexible, transparent microelectrodes are more compatible with optical imaging modalities than the commercial silicon probes. However, almost all transparent and flexible arrays are fabricated in individual academic labs. The product quality could vary from batch to batch, and lab to lab. Standardiza- tion and commercialization of such flexible and transparent microelectrodes would encourage more usage and accelerate studies in this field. (2) Implanting optical devices generate foreign body responses, which could be a confounding factor for understanding the cellular dynamics at the brain-electronic interface. Multiple timepoints or in vivo characterizations of tissue responses to optical devices are needed to validate the methods. Furthermore, it will be beneficial to develop meth- ods to improve the biocompatibility for optical devices, such as antifouling coating, anti-inflammatory coating, etc. (3) Insertion of the optical devices can inevitably cause bleed- ing that impedes imaging in the first several weeks. To be able to acquire information during that recovery period, tech- nological advancements are necessary. For example, to have a removable cranial window that allows inserting electrodes after the tissue reaction subsides. (4) Flexible electrode arrays for concurrent optical imaging and for the next generation of brain–machine interface are rapidly developing recently. Flexible electrodes have often shown reduced foreign body responses compared to conventional rigid electrodes. Yet, flex- ible electrodes are usually more difficult to handle and less durable than rigid electrode material wise, and lack quanti- tative long-term biocompatibility studies. References 1. E. Bizzi, Exp. Brain Res. 6, 69 (1968) p ( ) 2. H. Asanuma, S. Stoney Jr., C. Abzug, J. Neurophysiol. 31, 670 (1968) 3. S. Miocinovic, S. Somayajula, S. Chitnis, J.L. Vitek, JAMA Neurol. 70, 163 (2013) 4. W.H. Dobelle, ASAIO J. 46, 3 (2000) 5. A.B. Schwartz, X.T. Cui, D.J. Weber, D.W. Moran, Neuron 52, 205 (2006) 6. G.A. Tabot, J.F. Dammann, J.A. Berg, F.V. Tenore, J.L. Boback, R.J 6. G.A. Tabot, J.F. Dammann, J.A. Berg, F.V. Tenore, J.L. Boback, R.J. Vogelstein, S.J. Bensmaia, Proc. Natl. Acad. Sci. U.S.A. 110, 18279 (2013) Bensmaia, Proc. Natl. Acad. Sci. U.S.A. 110, 18279 (2013) 7. T.D. Kozai, A.S. Jaquins-Gerstl, A.L. Vazquez, A.C. Michael, X.T 7. T.D. Kozai, A.S. Jaquins-Gerstl, A.L. Vazquez, A.C. Michael, X.T. Cui, ACS Chem. Neurosci. 6, 48 (2015) 8. A.J. Woolley, H.A. Desai, K.J. Otto, J. Neural Eng. 10, 026007 (2013) 9. J.V. Cramer, B. Gesierich, S. Roth, M. Dichgans, M. Düring, A. Liesz, Neuroimage 199, 570 (2019) 10. Z. Li, Q. Zhang, S.-W. Chou, Z. Newman, R. Turcotte, R. Natan, Q. Dai, E.Y. Isacoff, N. Ji, Sci. Adv. 6, 3870 (2020) 11. R. Shi, C. Jin, H. Xie, Y. Zhang, X. Li, Q. Dai, L. Kong, Biomed. Opt. Express 10, 6625 (2019) 12. M.A. Neil, R. Juškaitis, T. Wilson, Opt. Lett. 22, 1905 (1997) 13. U. Dirnagl, A. Villringer, K.M. Einhäupl, J. Microsc. 165, 147 (1992) 14. W. Yu, B. Guo, H. Zhang, J. Zhou, X. Yu, L. Zhu, D. Xue, W. Liu, X. Sun, J. Qian, Sci. B ll 64 410 (2019) 14. W. Yu, B. Guo, H. Zhang, J. Zhou, X. Yu, L. Zhu, D. Xue, W. Liu, X. Sun, J. Qian, Sci. Bull. 64, 410 (2019) 15. F. Wang, F. Ren, Z. Ma, L. Qu, R. Gourgues, C. Xu, A. Baghdasaryan, J. Li, I.E. Zadeh, J.W. Los, Nat. Nanotechnol. 17, 653 (2022) 16. F. Xia, M. Gevers, A. Fognini, A.T. Mok, B. Li, N. Akbari, I.E. Zadeh, J. Qin-Dregely, C. Xu, ACS Photonics 8, 2800 (2021) 16. F. Xia, M. Gevers, A. Fognini, A.T. Mok, B. Li, N. Akbari, I.E. Zadeh, J. Qin-Dregely, C X ACS Photonics 8 2800 (2021) 17. D. Kobat, N.G. Horton, C. Xu, J. Biomed. Opt. 16, 106014 (2011) 18. D. Kobat, M.E. Durst, N. Nishimura, A.W. Wong, C.B. Schaffer, C. Xu, Opt. Express 17, 13354 (2009) 19. C. Stosiek, O. Garaschuk, K. Holthoff, A. Konnerth, Proc. Natl. Acad. References Sci. U.S.A. 100, 7319 (2003) 20. A. Cheng, J.T. Gonçalves, P. Golshani, K. Arisaka, C. Portera-Cailliau, Nat. Methods 8, 139 (2011) 21. J.N. Stirman, I.T. Smith, M.W. Kudenov, S.L. Smith, Nat. Biotechnol. 34, 857 (2016) 22. N.J. Sofroniew, D. Flickinger, J. King, K. Svoboda, eLife 5, e14472 (2016) 23. W. Zong, R. Wu, M. Li, Y. Hu, Y. Li, J. Li, H. Rong, H. Wu, Y. Xu, Y. Lu, Nat. Methods 14, 713 (2017) 24. E. Castagnola, X.S. Zheng, X.T. Cui, “Flexible and Soft Materials and Devices for Neural Interface,” in Handbook of Neuroengineering, ed. by N.V. Thakor (Springer, 24. E. Castagnola, X.S. Zheng, X.T. Cui, “Flexible and Soft Materials and Devices for 24. E. Castagnola, X.S. Zheng, X.T. Cui, “Flexible and Soft Materials and Devices for Neural Interface,” in Handbook of Neuroengineering, ed. by N.V. Thakor (Springer, Singapore, 2023), pp. 79–139 25. Z. Fekete, A. Pongrácz, Sens. Actuators B Chem. 243, 1214 (2017) 26. P. Kshirsagar, S. Dickreuter, M. Mierzejewski, C.J. Burkhardt, T. Chassé, M. Fleischer, P.D. Jones, Adv. Mater. Technol. 4, 1800318 (2019) 27. Y. Lu, X. Liu, R. Hattori, C. Ren, X. Zhang, T. Komiyama, D. Kuzum, Adv. Funct. Mater. 28, 1800002 (2018) Á 28. A. Zátonyi, M. Madarász, Á. Szabó, T. Lőrincz, R. Hodován, B. Rózsa, Z. Fekete, J. Neural Eng. 17, 016062 (2020) 29. W. Lee, D. Kim, N. Matsuhisa, M. Nagase, M. Sekino, G.G. Malliaras, T. Yokota, T. 29. W. Lee, D. Kim, N. Matsuhisa, M. Nagase, M. Sekino, G.G. Malliaras, T. Yokota, T. Someya, Proc. Natl. Acad. Sci. U.S.A. 114, 10554 (2017) Someya, Proc. Natl. Acad. Sci. U.S.A. 114, 10554 (2017) 30. C.F. Guo, T. Sun, Q. Liu, Z. Suo, Z. Ren, Nat. Commun. 5, 3121 (2014) Conflict of interest 31. Y. Qiang, P. Artoni, K.J. Seo, S. Culaclii, V. Hogan, X. Zhao, Y. Zhong, X. Han, P.-M. Wang, Y.-K. Lo, Sci. Adv. 4, eaat0626 (2018) On behalf of all authors, the corresponding author states that there is no conflict of interest. 32. J. Zhang, X. Liu, W. Xu, W. Luo, M. Li, F. Chu, L. Xu, A. Cao, J. Guan, S. Tang, Nano Lett. 18, 2903 (2018) 33. N. Driscoll, R.E. Rosch, B.B. Murphy, A. Ashourvan, R. Vishnubhotla, O.O. Dickens, C S C 33. N. Driscoll, R.E. Rosch, B.B. Murphy, A. Ashourvan, R. Vishnubhotla, O.O. Dickens A.C. Johnson, K.A. Davis, B. Litt, D.S. Bassett, Commun. Biol. 4, 136 (2021) A.C. Johnson, K.A. Davis, B. Litt, D.S. Bassett, Commun. Biol. 4, 136 (2021) Current limitations and future directions Therefore, flexible electrode devices need to be further studied and optimized in terms of biocompatibility and durability before broad applica- tions. With the continuous technological improvement in in vivo fluorescent imaging for neural interfacing electronics, we expect to gather more knowledge on the cellular and subcel- lular mechanisms related to electrophysiological recording and stimulation, which would greatly help to determine the best electrode material and design in the coming future. in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. MRS BULLETIN • VOLUME 48 • MAY 2023 • mrs.org/bulletin 515 Open access Limousin, P. Krack, P. Pollak, A. Benazzouz, C. Ardouin, D. Hoffmann, A.-L. Benabid, N. Engl. J. Med. 339, 1105 (1998) , ( ) 94. P. Limousin, P. Krack, P. Pollak, A. Benazzouz, C. Ardouin, D. Hoffmann, A.-L. B bid N E l J M d 339 110 (1998) 51. S.T. Parker, P. Domachuk, J. Amsden, J. Bressner, J.A. Lewis, D.L. Kaplan, F.G. Omenetto, Adv. Mater. 21, 2411 (2009) 95. D. McCreery, V. Pikov, P.R. Troyk, J. Neural Eng. 7, 036005 (2010) 52. R. Nazempour, Q. Zhang, R. Fu, X. Sheng, Materials (Basel) 11, 1283 (2018) g ( ) 96. M.M. Chakravarty, C. Hamani, A. Martinez-Canabal, J. Ellegood, C. Laliberté, J.N. 53. B.A. Flusberg, E.D. Cocker, W. Piyawattanametha, J.C. Jung, E.L. Cheung, M.J. Schnitzer, Nat. Methods 2, 941 (2005) Nobrega, T. Sankar, A.M. Lozano, P.W. Frankland, J.P. Lerch, Neuroimage 125, 422 (2016) 54. M. Murayama, E. Pérez-Garci, T. Nevian, T. Bock, W. Senn, M.E. Larkum, Nature 457, 1137 (2009) ( ) 97. E. Tehovnik, A. Tolias, F. Sultan, W. Slocum, N. Logothetis, J. Neurophysiol. 96, 512 (2006) 55. W. Göbel, J.N. Kerr, A. Nimmerjahn, F. Helmchen, Opt. Lett. 29, 2521 (2004) ) 98. S. Stoney Jr., W. Thompson, H. Asanuma, J. Neurophysiol. 31, 659 (1 F. Pisanello, G. Mandelbaum, M. Pisanello, I.A. Oldenburg, L. Sileo, J.E. M 56. F. Pisanello, G. Mandelbaum, M. Pisanello, I.A. Oldenburg, L. Sileo, J.E. Markowitz, R.E. Peterson, A. Della Patria, T.M. Haynes, M.S. Emara, Nat. Neurosci. 20, 1180 (2017) R.E. Peterson, A. Della Patria, T.M. Haynes, M.S. Emara, Nat. Neurosci. 20, 1180 (2017) 99. M.H. Histed, V. Bonin, R.C. Reid, Neuron 63, 508 (2009) 57. F. Pisano, M. Pisanello, S.J. Lee, J. Lee, E. Maglie, A. Balena, L. Sileo, B. Spagnolo, M. Bianco, M. Hyun, Nat. Methods 16, 1185 (2019) 100. A.T. Hussin, J.A. Boychuk, A.R. Brown, Q.J. Pittman, G.C. Teskey, Brain Stimul. 8, 742 (2015) 57. F. Pisano, M. Pisanello, S.J. Lee, J. Lee, E. Maglie, A. Balena, L. Sileo, B. Spagnolo, M Bianco M Hyun Nat Methods 16 1185 (2019) 101. Y. Tanaka, T. Nomoto, T. Shiki, Y. Sakata, Y. Shimada, Y. Hayashida, T. Yagi, J. Neural Eng. 16, 036007 (2019) 58. R. Nazempour, B. Zhang, Z. Ye, L. Yin, X. Lv, X. Sheng, Adv. Fiber Mater. 2021, 1 (2021) g , ( ) 102. N.J. Michelson, J.R. Eles, A.L. Vazquez, K.A. Ludwig, T.D. Kozai, J. Neurosci. Res. 97, 620 (2019) 59. Y. Open access 34. I. Adawi, Phys. Rev. 134, A788 (1964) 34. I. Adawi, Phys. Rev. 134, A788 (1964) 35. W.L. Shew, T. Bellay, D. Plenz, J. Neurosci. Methods 192, 75 (2010) This article is licensed under a Creative Commons Attribu- tion 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Com- mons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated other- wise in a credit line to the material. If material is not included 36. Y.U. Cho, J.Y. Lee, U.J. Jeong, S.H. Park, S.L. Lim, K.Y. Kim, J.W. Jang, J.H. Park, H.W. Kim, H. Shin, Adv. Funct. Mater. 32, 2105568 (2022) H.W. Kim, H. Shin, Adv. Funct. Mater. 32, 2105568 (2022) 37. J. Riera, T. Ogawa, R. Hatanaka, T. Goto, A. Sumiyoshi, H.E. Kadji, S. Nakauchi, R. Kawashima, J. Biophotonics 3(3), 147 (2010) 38. A.D. Degenhart, J. Eles, R. Dum, J.L. Mischel, I. Smalianchuk, B. Endler, R.C. Ashmore, E.C. Tyler-Kabara, N.G. Hatsopoulos, W. Wang, J. Neural Eng. 13, 046019 (2016) 39. C. Henle, M. Raab, J.G. Cordeiro, S. Doostkam, A. Schulze-Bonhage, T. Stieglitz, J. Rickert, Biomed. Microdevices 13, 59 (2011) 40. F.Z. Fedor, M. Madarász, A. Zátonyi, Á. Szabó, T. Lőrincz, V. Danda, L. Spurgin, C. Manz, B. Rózsa, Z. Fekete, Adv. Mater. Technol. 7, 2100942 (2022) MRS BULLETIN • VOLUME 48 • MAY 2023 • mrs.org/bulletin 515 Advanced in vivo fluorescence microscopy of neural electronic interface Advanced in vivo fluorescence microscopy of neural electronic interface 85. P.R. Williams, B.-N. Marincu, C.D. Sorbara, C.F. Mahler, A.-M. Schumacher, O. 41. A.A. Schendel, M.W. Nonte, C. Vokoun, T.J. Richner, S.K. Brodnick, F. Atry, S. Frye, A.A. Schendel, M.W. Nonte, C. Vokoun, T.J. Richner, S.K. Brodnick, F. Atry 41. A.A. Schendel, M.W. Nonte, C. Vokoun, T.J. Richner, S.K. Brodnick, F. Atry, S. Frye, P. Bostrom, R. Pashaie, S. Thongpang, J. Neural Eng. 11, 046011 (2014) Griesbeck, M. Kerschensteiner, T. Misgeld, Nat. Commun. 5(1), 5683 (2014) P. Bostrom, R. Pashaie, S. Thongpang, J. Neural Eng. 11, 046011 (2014) 86. J.A. Wolf, P.K. Stys, T. Lusardi, D. Meaney, D.H. Smith, J. Neurosci. 21, 1923 (2001) 42. T.D.Y. Kozai, A.L. Vazquez, C.L. Weaver, S.-G. Kim, X.T. Cui, J. Neural Eng. Open access Zhou, C. Gu, J. Liang, B. Zhang, H. Yang, Z. Zhou, M. Li, L. Sun, T.H. Tao, X. Wei, Microsyst. Nanoeng. 8, 118 (2022) , ( ) 03. J.R. Eles, K.C. Stieger, T.D. Kozai, J. Neural Eng. 18, 015001 (2021) 60. A. Canales, X. Jia, U.P. Froriep, R.A. Koppes, C.M. Tringides, J. Selvidge, C. Lu, C. Hou L Wei Y Fink Nat Biotechnol 33 277 (2015) g g ( ) 104. K.C. Stieger, J.R. Eles, K.A. Ludwig, T.D. Kozai, J. Neurosci. Res. 98, 2072 (2020) Hou, L. Wei, Y. Fink, Nat. Biotechnol. 33, 277 (2015) 104. K.C. Stieger, J.R. Eles, K.A. Ludwig, T.D. Kozai, J. Neurosci. Res. 9 105. J.R. Eles, T.D. Kozai, Biomaterials 234, 119767 (2020) 61. G. Cui, S.B. Jun, X. Jin, M.D. Pham, S.S. Vogel, D.M. Lovinger, R.M. Costa, Nature 494, 238 (2013) 06. X.S. Zheng, Q. Yang, A. Vazquez, X.T. Cui, iScience 25, 104539 (202 ( ) 62. A.R. Rouse, A. Kano, J.A. Udovich, S.M. Kroto, A.F. Gmitro, Appl. Opt. 43, 5763 (2004) 107. X.S. Zheng, Q. Yang, A. Vazquez, X.T. Cui, Adv. Nanobiomed. Res. 1(7), 2000092 (2021) ( ) 108. V.S. Polikov, P.A. Tresco, W.M. Reichert, J. Neurosci. Methods 148(1), 1 (2005) 63. J.C. Jung, A.D. Mehta, E. Aksay, R. Stepnoski, M.J. Schnitzer, J. Neurophysiol. 92, 3121 (2004) 109. D. Szarowski, M. Andersen, S. Retterer, A. Spence, M. Isaacson, H.G. Craighead, J. Turner, W. Shain, Brain Res. 983, 23 (2003) 64. M.J. Levene, D.A. Dombeck, K.A. Kasischke, R.P. Molloy, W.W. Webb, J. Neurophysiol. 91, 1908 (2004) ( ) 110. D. McCreery, S. Cogan, S. Kane, V. Pikov, J. Neural Eng. 13, 0360 111. A. Nimmerjahn, F. Kirchhoff, F. Helmchen, Science 308, 1314 (2005) 65. R.P. Barretto, T.H. Ko, J.C. Jung, T.J. Wang, G. Capps, A.C. Waters, Y. Ziv, A. Attardo, L. Recht, M.J. Schnitzer, Nat. Med. 17, 223 (2011) 112. H. Wake, A.J. Moorhouse, S. Jinno, S. Kohsaka, J. Nabekura, J. Neurosci. 29, 3974 (2009) 66. L.F. Cobar, A. Kashef, K. Bose, A. Tashiro, Sci. Rep. 12, 736 (2022) 113. C. Madry, V. Kyrargyri, I.L. Arancibia-Cárcamo, R. Jolivet, S. Kohsaka, R.M. Bryan, D. Attwell, Neuron 97, 299 (2018) 67. M.L. Andermann, N.B. Gilfoy, G.J. Goldey, R.N. Sachdev, M. Wölfel, D.A. McCormick, R.C. Reid, M.J. Levene, Neuron 80, 900 (2013) R.C. Reid, M.J. Levene, Neuron 80, 900 (2013) 68. M. Murayama, E. Pérez-Garci, H.-R. Luscher, M.E. Larkum, J. Neurophysiol. 98, 1791 (2007) 114. S.J. Ahn, J. Open access 9, 066001 (2012) y y ( ) 87. J.A. Staal, T.C. Dickson, R. Gasperini, Y. Liu, L. Foa, J.C. Vickers, J. Neurochem. 112, 1147 (2010) 43. T.D. Kozai, J.R. Eles, A.L. Vazquez, X.T. Cui, J. Neurosci. Methods 258, 46 (2016) 88. N.J. Michelson, A.L. Vazquez, J.R. Eles, J.W. Salatino, E.K. Purcell, J.J. Williams, 88. N.J. Michelson, A.L. Vazquez, J.R. Eles, J.W. Salatino, E.K. Purcell, J.J. Williams, X.T. Cui, T.D. Kozai, J. Neural Eng. 15, 033001 (2018) 44. L. Luan, X. Wei, Z. Zhao, J.J. Siegel, O. Potnis, C.A. Tuppen, S. Lin, S. Kazmi, R.A. Fowler, S. Holloway, Sci. Adv. 3, e1601966 (2017) 88. N.J. Michelson, A.L. Vazquez, J.R. Eles, J.W. Salatino, E.K. Purcell, J.J. Williams, X.T. Cui, T.D. Kozai, J. Neural Eng. 15, 033001 (2018) 45. F. He, Y. Sun, Y. Jin, R. Yin, H. Zhu, H. Rathore, C. Xie, L. Luan, Biomaterials 291, 121905 (2022) 89. C.G. Welle, Y.-R. Gao, M. Ye, A. Lozzi, A. Boretsky, E. Abliz, D.X. Hammer, Biomaterials 238, 119831 (2020) ( ) 90. S.N. Flesher, J.L. Collinger, S.T. Foldes, J.M. Weiss, J.E. Downey, E.C. Tyler-Kabara, ( ) 46. X. Liu, C. Ren, Y. Lu, Y. Liu, J.-H. Kim, S. Leutgeb, T. Komiyama, D. Kuzum, Nat. Neurosci. 24, 886 (2021) S.J. Bensmaia, A.B. Schwartz, M.L. Boninger, R.A. Gaunt, Sci. Transl. Med. 8, 361ra141 (2016) 47. J.R. Eles, A.L. Vazquez, N.R. Snyder, C. Lagenaur, M.C. Murphy, T.D. Kozai, X.T. Cui, Biomaterials 113, 279 (2017) ( ) 91. C.L. Hughes, S.N. Flesher, J.M. Weiss, M. Boninger, J.L. Collinger, R.A. Gaunt, eLife 10, e65128 (2021) 48. J. Eles, A. Vazquez, T. Kozai, X. Cui, Biomaterials 195, 111 (2019) 92. J. Siegfried, B. Lippitz, Neurosurgery 35, 1126 (1994) 92. J. Siegfried, B. Lippitz, Neurosurgery 35, 1126 (1994) 49. K. Mols, S. Musa, B. Nuttin, L. Lagae, V. Bonin, Sci. Rep. 7(1), 15642 (2017) 93. L. D’Alatri, G. Paludetti, M.F. Contarino, S. Galla, M.R. Marchese, A.R. Bentivoglio, J. Voice 22, 365 (2008) 50. H. Lütcke, M. Murayama, T. Hahn, D.J. Margolis, S. Astori, S. Meyer, W. Göbel, Y. Y W T S Kü l F t N l Ci it 4 9 (2010) 50. H. Lütcke, M. Murayama, T. Hahn, D.J. Margolis, S. Astori, S. Meyer, W. Göbel, Y. Yang, W. Tang, S. Kügler, Front. Neural Circuits 4, 9 (2010) J. Voice 22, 365 (2008) Yang, W. Tang, S. Kügler, Front. Neural Circuits 4, 9 (2010) , ( ) 94. P. Open access Anrather, N. Nishimura, C.B. Schaffer, Stroke 4 114. S.J. Ahn, J. Anrather, N. Nishimura, C.B. Schaffer, Stroke 49, 1719 (2018) 115. T.M. Rotterman, F.J. Alvarez, Sci. Rep. 10, 8648 (2020) Kim, E. Chung, H. Yamashita, K.E. Hung, A. Mizoguchi, R. Kucherlap 116. D. Davalos, J. Grutzendler, G. Yang, J.V. Kim, Y. Zuo, S. Jung, D.R. Littman, M.L. Dustin, W.-B. Gan, Nat. Neurosci. 8, 752 (2005) 116. D. Davalos, J. Grutzendler, G. Yang, J.V. Kim, Y. Zuo, S. Jung, g g Dustin, W.-B. Gan, Nat. Neurosci. 8, 752 (2005) Fukumura, R.K. Jain, S.H. Yun, Nat. Methods 7, 303 (2010) Dustin, W.-B. Gan, Nat. Neurosci. 8, 752 (2005) ukumura, R.K. Jain, S.H. Yun, Nat. Methods 7, 303 (2010) 70. Q. Yang, B. Wu, E. Castagnola, M.Y. Pwint, A.L. Vazquez, X.T. Cui, bioRxiv (2022). https://doi.org/10.1101/2022.07.08.499369 117. W. Sun, K. Suzuki, D. Toptunov, S. Stoyanov, M. Yuzaki, L. Khiroug, A. Dityatev, Front. Neurosci. 13, 421 (2019) p g 71. L. Beckmann, X. Zhang, N.A. Nadkarni, Z. Cai, A. Batra, D.P. Sullivan, W.A. Muller, p g 71. L. Beckmann, X. Zhang, N.A. Nadkarni, Z. Cai, A. Batra, D.P. Sullivan, W.A. Muller, C Sun R Kuranov H F Zhang Biomed Opt Express 10 5235 (2019) 118. J. Neumann, M. Gunzer, H.O. Gutzeit, O. Ullrich, K.G. Reymann, K. Dinkel, FASEB J. 20, 714 (2006) g C. Sun, R. Kuranov, H.F. Zhang, Biomed. Opt. Express 10, 5235 (2019) Neurophysiol. 84, 390 (2000) 124. Q. Yang, B. Wu, J.R. Eles, A.L. Vazquez, T.D. Kozai, X.T. Cui, A 124. Q. Yang, B. Wu, J.R. Eles, A.L. Vazquez, T.D. Kozai, X.T. Cui, Adv. Biosyst. 4, 1900287 (2020) 77. C. Holmgren, T. Harkany, B. Svennenfors, Y. Zilberter, J. Physiol. 551, 139 (2003) ( ) 125. D.D. Krahe, K.M. Woeppel, Q. Yang, N. Kushwah, X.T. Cui, Antioxidants (Basel) 11 1628 (2022) 125. D.D. Krahe, K.M. Woeppel, Q. Yang, N. Kushwah, X.T. Cui, Antioxidants (Basel) 11, 1628 (2022) 78. G. Buzsáki, Nat. Neurosci. 7, 446 (2004) 79. G.R. Buzsáki, A. Kandel, J. Neurophysiol. 79, 1587 (1998) , ( ) 126. A. Golabchi, B. Wu, X. Li, D.L. Carlisle, T.D. Kozai, R.M. Friedlander, X.T. Biomaterials 180 225 (2018) 126. A. Golabchi, B. Wu, X. Li, D.L. Carlisle, T.D. Kozai, R.M. Friedlander, X.T. Cui, Biomaterials 180, 225 (2018) 126. A. Golabchi, B. Wu, X. Li, D.L. Carlisle, T.D. Kozai, R.M. Friedland Bi i l ( ) 80. G.C. McConnell, H.D. Rees, A.I. Levey, C.-A. Gutekunst, R.E. Gross, R.V. 80. G.C. McConnell, H.D. Rees, A.I. Levey, C.-A. Gutekuns 126. A. Golabchi, B. Wu, X. Li, D.L. Carlisle, T.D. Kozai, R.M. Friedlander, X.T. Cui, Biomaterials 180, 225 (2018) 80 G C cCo e , ees, e ey, C Gu e u s , Bellamkonda, J. Neural Eng. 6, 056003 (2009) Biomaterials 180, 225 (2018) Bellamkonda, J. Neural Eng. 6, 056003 (2009) 81. M. Ravikumar, S. Sunil, J. Black, D.S. Barkauskas, A.Y. Haung, R.H. Miller, S.M. 127. Y. Zhong, R.V. Bellamkonda, Brain Res. 1148, 15 (2007) 81. M. Ravikumar, S. Sunil, J. Black, D.S. Barkauskas, A g, , , ( ) 128. K. Kacem, P. Lacombe, J. Seylaz, G. Bonvento, Glia 23, 1 (1998) 129 C I d l M N d d N t N i 10 1369 (2007) Selkirk, J.R. Capadona, Biomaterials 35, 8049 (2014) 128. K. Kacem, P. Lacombe, J. Seylaz, G. Bonvento, Glia 23, 1 (1998) Selkirk, J.R. Capadona, Biomaterials 35, 8049 (2014) 82. J.W. Salatino, B.M. Winter, M.H. Drazin, E.K. Purcell, J. Neurophysiol. 118, 194 (2017) 129. C. Iadecola, M. Nedergaard, Nat. Neurosci. 10, 1369 (2007) 130. M.V. Sofroniew, H.V. Vinters, Acta Neuropathol. 119, 7 (2010) 31. R.C. Koehler, R.J. Roman, D.R. Harder, Trends Neurosci. 32, 160 (2 83. H. Dana, T.-W. Chen, A. Hu, B.C. Shields, C. Guo, L.L. Looger, D.S. Kim, K. Svoboda, PLoS ONE 9, e108697 (2014) 132. M. Simard, G. Arcuino, T. Takano, Q.S. J. 20, 714 (2006) 119. R.B. Banati, J. Gehrmann, P. Schubert, G.W. Kreutzberg, Glia 72. Q. Yang, A.L. Vazquez, X.T. Cui, Biomaterials 276, 121060 (2021) 73. Y.-T. Kim, R.W. Hitchcock, M.J. Bridge, P.A. Tresco, Biomaterials 25, 2229 (2004) 120. M. Schwartz, J. Cereb. Blood Flow Metab. 23, 385 (2003) 74. L. Wang, C. Zhong, D. Ke, F. Ye, J. Tu, L. Wang, Y. Lu, Adv. Opt. Mater. 6, 1800427 (2018) 121. A. Golabchi, K.M. Woeppel, X. Li, C.F. Lagenaur, X.T. Cui, Biosens. Bioelectron. 155, 112096 (2020) 155, 112096 (2020) 75. S. Park, H. Yuk, R. Zhao, Y.S. Yim, E.W. Woldeghebriel, J. Kang, A. Canales, Y. Fink, 122. A. Golabchi, B. Wu, B. Cao, C.J. Bettinger, X.T. Cui, Biomaterials 225, 119519 (2019) G.B. Choi, X. Zhao, P. Anikeeva, Nat. Commun. 12(1), 3435 (2021) 123. E. Azemi, C.F. Lagenaur, X.T. Cui, Biomaterials 32, 681 (2011) 76. D.A. Henze, Z. Borhegyi, J. Csicsvari, A. Mamiya, K.D. Harris, G. Buzsaki, J. Neurophysiol. 84, 390 (2000) Neurophysiol. 84, 390 (2000) 40, 2371 (2020) 141. A.S. Saab, I.D. Tzvetavona, A. Trevisiol, S. Baltan, P. Dibaj, K. Kusch, W. Möbius, B Goetze H M Jahn W Huang Neuron 91 119 (2016) 141. A.S. Saab, I.D. Tzvetavona, A. Trevisiol, S. Baltan, P. Dibaj, K. Kusch, W. Möbius, j B. Goetze, H.M. Jahn, W. Huang, Neuron 91, 119 (2016) B. Goetze, H.M. Jahn, W. Huang, Neuron 91, 119 (2016) 171. N.T. Markwardt, J. Stokol, R.L. Rennaker II., J. Neurosci. Methods 214, 119 (2013) ⃞ 142. A.I. Amaral, T.W. Meisingset, M.R. Kotter, U. Sonnewald, Front. Endocrinol. 4, 54 (2013) 143. M. Nakano, Y. Tamura, M. Yamato, S. Kume, A. Eguchi, K. Takata, Y. Watanabe, Y. Kataoka, Sci. Rep. 7, 42041 (2017)i Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. p ( ) 144. I. Griffiths, M. Klugmann, T. Anderson, D. Yool, C. Thomson, M.H. Schwab, A. Schneider, F. Zimmermann, M. McCulloch, N. Nadon, Science 280, 1610 (1998) ( ) M. Wellman, K. Guzman, K.C. Stieger, L.E. Brink, S. Sridhar, M.T. Dubanie 145. S.M. Wellman, K. Guzman, K.C. Stieger, L.E. Brink, S. Sridhar, M.T. Dubaniewicz, L. Li, F. Cambi, T.D. Kozai, Biomaterials 239, 119842 (2020) 145. S.M. Wellman, K. Guzman, K.C. Stieger, L.E. Brink, S. Sridhar, M.T. Dubaniewicz, L. Li, F. Cambi, T.D. Kozai, Biomaterials 239, 119842 (2020) Qianru Yang is a doctoral candidate at the University of Pittsburgh. She earned her BS degree in biology from the University of Science and Technology of China in 2015. Her research has been focused on in vivo two-photon fluorescent microscopy of the neural electrode– tissue interface, including antifouling coating material, acute and chronic inflammatory responses to neural implants, application of microprism implantation in neural tissue engineering, and investigating cellular mechanisms of electrophysiology recording and stimulation. Yang can be reached by email at qiy24@pitt.edu. Qianru Yang is a doctoral candidate at the University of Pittsburgh. She earned her BS degree in biology from the University of Science and Technology of China in 2015. Her research has been focused on in vivo two-photon fluorescent microscopy of the neural electrode– tissue interface, including antifouling coating material, acute and chronic inflammatory responses to neural implants, application of microprism implantation in neural tissue engineering, and investigating cellular mechanisms of electrophysiology recording and stimulation. Yang can be reached by email at qiy24@pitt.edu. 146. K. Chen, S.M. Wellman, Y. Yaxiaer, J.R. Eles, T.D. Neurophysiol. 84, 390 (2000) Liu, M. Nedergaard, J. Neurosci. 23, 9254 (2003) 132. M. Simard, G. Arcuino, T. Takano, Q.S. Liu, M. Nedergaard, J. Neurosci. 23, 9254 (2003) 84. J.R. Eles, A.L. Vazquez, T.D. Kozai, X.T. Cui, Biomaterials 174, 79 (2018) 133. M. Simard, M. Nedergaard, Neuroscience 129, 877 (2004) 516 MRS BULLETIN • VOLUME 48 • MAY 2023 • mrs.org/bulletin Advanced in vivo fluorescence microscopy of neural electronic interfac Advanced in vivo fluorescence microscopy of neural electronic interface 163. K. Møllgård, F.R. Beinlich, P. Kusk, L.M. Miyakoshi, C. Delle, V. Plá, N.L. Hauglund, T. Esmail, M.K. Rasmussen, R.S. Gomolka, Science 379, 84 (2023) 134. A.H. Cornell-Bell, S.M. Finkbeiner, M.S. Cooper, S.J. Smith, Science 247, 470 (1990) , , , , ( ) 164. Y. Lavin, A. Mortha, A. Rahman, M. Merad, Nat. Rev. Immunol. 15, 731 (2015) ( ) 135. H. Hirase, L. Qian, P. Barthó, G. Buzsáki, PLoS Biol. 2, e96 (2004) 165. H.R. Chinnery, M.J. Ruitenberg, P.G. McMenamin, J. Neuropathol. Exp. Neurol. 69(9), 896 (2010) 136. S.A. Liddelow, K.A. Guttenplan, L.E. Clarke, F.C. Bennett, C.J. Bohlen M.L. Bennett, A.E. Münch, W.-S. Chung, T.C. Peterson, Nature 541, 481 (2017) ( ) ( ) 166. A. Louveau, I. Smirnov, T.J. Keyes, J.D. Eccles, S.J. Rouhani, J.D. Peske, N.C. 137. S.P. Savya, F. Li, S. Lam, S.M. Wellman, K.C. Stieger, K. Chen, J.R. Eles, T.D. Kozai, Biomaterials 289, 121784 (2022) 166. A. Louveau, I. Smirnov, T.J. Keyes, J.D. Eccles, S.J. Rouhani, J.D. Peske, N.C. Derecki, D. Castle, J.W. Mandell, K.S. Lee, Nature 523, 337 (2015) , , y , , , , Derecki, D. Castle, J.W. Mandell, K.S. Lee, Nature 523, 337 (2015) 167. J.A. Coles, P.J. Stewart-Hutchinson, E. Myburgh, J.M. Brewer, Methods 127, 53 (2017) 138. Z. Ma, L. Wei, X. Du, S. Hou, F. Chen, Q. Jiao, A. Liu, S. Liu, J. Wang, H. Shen, J. Neural Eng. 18, 046096 (2021) ( ) 168. X. Cui, J. Wiler, M. Dzaman, R.A. Altschuler, D.C. Martin, Biomaterials 24, 777 (2003) g ( ) 139. G.S. Tomassy, D.R. Berger, H.-H. Chen, N. Kasthuri, K.J. Hayworth, A. Vercelli, H.S. Seung, J.W. Lichtman, P. Arlotta, Science 344, 319 (2014) ( ) 169. P.A. Cody, J.R. Eles, C.F. Lagenaur, T.D. Kozai, X.T. Cui, Biomaterials 161, 117 (2018) 140. A.M. Krasnow, M.C. Ford, L.E. Valdivia, S.W. Wilson, D. Attwell, Nat. Neurosci. 21, 24 (2018) ( ) 170. N.L. Hauglund, P. Kusk, B.R. Kornum, M. Nedergaard, J. Neurosci. Neurophysiol. 84, 390 (2000) Kozai, Biomaterials 268, 120526 (2021) 147. M.R. Dawson, A. Polito, J.M. Levine, R. Reynolds, Mol. Cell. Neurosci. 24, 476 (2003) 148. R.B. Tripathi, L.E. Rivers, K.M. Young, F. Jamen, W.D. Richardson, J. Neurosci. 30, 16383 (2010) 149. A.R. Hackett, D.-H. Lee, A. Dawood, M. Rodriguez, L. Funk, P. Tsoulfas, J.K. Lee, Neurobiol. Dis. 89, 10 (2016) 150. J.M. Levine, R. Reynolds, Exp. Neurol. 160, 333 (1999) 151. E.G. Hughes, S.H. Kang, M. Fukaya, D.E. Bergles, Nat. 151. E.G. Hughes, S.H. Kang, M. Fukaya, D.E. Bergles, Nat. Neurosci. 16, 668 (2013) S 152. S.M. Wellman, T.D. Kozai, Biomaterials 164, 121 (2018) 153. B.T. Hawkins, T.P. Davis, Pharmacol. Rev. 57, 173 (2005) Á 154. C. Bennett, F. Mohammed, A. Álvarez-Ciara, M.A. Nguyen, W.D. Dietrich, S.M. Rajguru W J Streit A Prasad Biomaterials 188 144 (2019) 154. C. Bennett, F. Mohammed, A. Álvarez-Ciara, M.A. Nguyen, W.D. Dietrich, S.M. Rajguru, W.J. Streit, A. Prasad, Biomaterials 188, 144 (2019) , , , g y , , Rajguru, W.J. Streit, A. Prasad, Biomaterials 188, 144 (2019) Rajguru, W.J. Streit, A. Prasad, Biomaterials 188, 144 (2019) 155. L. Goldstein, Z.P. Teng, E. Zeserson, M. Patel, R.F. Regan, J. Neurosci. Res. 73, 113 (2003) 156. J. Wang, Prog. Neurobiol. 92, 463 (2010) 157. C.C. Winterbourn, Toxicol. Lett. 82, 969 (1995) X. Tracy Cui is the William Kepler Whiteford Professor of Bioengineering at the University of Pittsburgh. She earned her PhD degree in macromolecular science and engineering from the University of Michigan. She works in the field of neural engineering with special focuses on the neural electrode–tissue interface, neural tissue engineering, drug delivery, and biosensors. She is a senior member of the National Academy of Inventors, Fellow of the Royal Society of Chemistry, Fellow of the American Institute for Medical and Biological Engineering, and a recipient of the Carnegie Science Emerging Female Scientist Award. Cui can be reached by email at xic11@pitt.edu. 158. T. Saxena, L. Karumbaiah, E.A. Gaupp, R. Patkar, K. Patil, M. Betancur, G.B. , , pp, , , , Stanley, R.V. Bellamkonda, Biomaterials 34, 4703 (2013) Stanley, R.V. Bellamkonda, Biomaterials 34, 4703 (2013) 159. N.F. Nolta, M.B. Christensen, P.D. Crane, J.L. Skousen, P.A. Tresco, Biomaterials 53, 753 (2015) , ( ) 160. L. Grand, L. Wittner, S. Herwik, E. Göthelid, P. Ruther, S. Oscarsson, H. Neves, B. Dombovári, R. Csercsa, G. Karmos, J. Neurosci. Methods 189, 216 (2010) 161. A.Y. Shih, J.D. Driscoll, P.J. Drew, N. Neurophysiol. 84, 390 (2000) Nishimura, C.B. Schaffer, D. Kleinfeld, J. Cereb Blood Flow Metab 32 1277 (2012) , ( ) 160. L. Grand, L. Wittner, S. Herwik, E. Göthelid, P. Ruther, S. Oscarsson, H. Neves, B. Dombovári R Csercsa G Karmos J Neurosci Methods 189 216 (2010) 0. L. Grand, L. Wittner, S. Herwik, E. Göthelid, P. Ruther, S. Oscarsson, H. 160. L. Grand, L. Wittner, S. Herwik, E. Göthelid, P. Ruther, S. Oscarsson, H. Neves, B. Dombovári, R. Csercsa, G. Karmos, J. Neurosci. Methods 189, 216 (2010) Dombovári, R. Csercsa, G. Karmos, J. Neurosci. Methods 189, 216 (201 Dombovári, R. Csercsa, G. Karmos, J. Neurosci. Methods 189, 216 (2010) 161. A.Y. Shih, J.D. Driscoll, P.J. Drew, N. Nishimura, C.B. Schaffer, D. Kleinfeld, J. Cereb. Blood Flow Metab. 32, 1277 (2012) 162. T.D.Y. Kozai, T.C. Marzullo, F. Hooi, N.B. Langhals, A. Majewska, E. Brown, D.R. 161. A.Y. Shih, J.D. Driscoll, P.J. Drew, N. Nishimura, C.B. Schaffer, D. Kleinfeld, J. Cereb. Blood Flow Metab. 32, 1277 (2012) Cereb. Blood Flow Metab. 32, 1277 (2012) 162. T.D.Y. Kozai, T.C. Marzullo, F. Hooi, N.B. Langhals, A. Majewska, E. Brown, D.R. Kipke, J. Neural Eng. 7, 046011 (2010) 162. T.D.Y. Kozai, T.C. Marzullo, F. Hooi, N.B. Langhals, A. Majewska, E. Brown, D.R. Ki k J N l E 7 046011 (2010) Kipke, J. Neural Eng. 7, 046011 (2010) MRS BULLETIN • VOLUME 48 • MAY 2023 • mrs.org/bulletin 517
W4223971745.txt	https://jyx.jyu.fi/bitstream/123456789/80645/1/112036-Artikkelin%2520teksti-227548-1-10-20220325.pdf	fi		Pala Ukrainasta – väkisin, mutta suostutellen	Idäntutkimus	2,022	public-domain	7,689	This is a self-archived version of an original article. This version may differ from the original in pagination and typographic details. Author(s): Ahtamo, Valtteri Title: Pala Ukrainasta – väkisin, mutta suostutellen Year: 2022 Version: Published version Copyright: Rights: © 2022 Venäjän ja Itä-Euroopan tutkimuksen seura ry In Copyright Rights url: http://rightsstatements.org/page/InC/1.0/?language=en Please cite the original version: Ahtamo, V. (2022). Pala Ukrainasta – väkisin, mutta suostutellen. Idäntutkimus, 29(1), 3-22. https://doi.org/10.33345/idantutkimus.112036 artikkeli Pala Ukrainasta – väkisin, mutta suostutellen Tämä artikkeli käsittelee vuoden 2014 Krimin valtauksen läpivientiä painopisteenään Venäjän hallinnon tuottaman oikeutuksen tarkastelu. Päättäjien julkilausumista ja valtionmedian representaatiosta koostuvaa venäjänkielistä aineistoa analysoidaan tapahtumaketjuun suhteuttaen diskursiivis-retorisen geopolitiikan luennalla, mikä valottaa oikeutuskampanjan reaktiivista toteutusmuotoa ja kohdeyleisöihin vedonneen suostuttelun perustaa. Kreml sai johdettua Ukrainan sisäisestä konfliktista ensin juridisuuteen ja turvallisuuteen vetoavat interventiodiskurssit, joilla vakuutteli kotiyleisöä ja krimiläisiä maanmiehiä puolelleen osin jo ennen Krimin miehitystä, ja täydensi oikeutussisältöä niemimaan liittämisen lähestyessä Krimin ja Venäjän historiallisen ykseyden sekä Venäjää alistavan tekopyhän lännen diskurssein. Analyysi tuo esille diskurssit koostavien retoristen elementtien yksityiskohtia ja niiden kumulaatiota, millä Venäjän hallinto muodosti Krimin anastusta puoltavan geopoliittisen mielikuvan. Kuvitelmalla oli ilmeinen painoarvo kohdeyleisösidonnaisen kannatuksen varmistamisessa käytännön toimille. Valtteri Ahtamo Kremlin toimintamallissa ensin katsotaan talviolympialaiset ja sitten hyökätään Ukrainaan. Tämä artikkeli valmistuu vuoden 2022 helmikuun lopulla, kun Vladimir Putin on pitänyt sodanjulistuspuheensa ja käynnistänyt historiallisen hyökkäyksen, mikä yllätti röyhkeydessään myös kotiyleisön. Vaikka akuutti tilanne vasta kehittyy, Ukrainan vuoden 2022 invaasio eroaa radikaalisti Krimin vuoden 2014 valtauksesta. Niemimaan haltuunotto oli melko kontingentti operaatio, jonka Kreml toteutti epäsuorilla sotatoimilla käyttämällä hyväksi Ukrainan sisäistä konfliktia ja rakentamalla sekä koti- että krimiläisyleisöön voimakkaasti vetoavan oikeutuskampanjan. Suurinvaasio näyttäytyi ensin samalla kaavalla suoritettavalta Ukrainan itäosien miehitykseltä, mutta kääntyi häikäilemättömäksi täyshyökkäykseksi, jonka tarkoituksena lienee Valko-Venäjän kaltaisen vasallihallinnon muodostus tai jopa valtion IDÄNTUTKIMUS 1/2022 3 kahtiajako. Operaatio on monessa suhteessa Putinin henkilökohtaista ”venäläisen ruletin” peluuta, esimerkiksi kannatuksen saamisen ja siten kansan suosiossa pysymisen suhteen. Vaikka kotiyleisöä oli valmisteltu jonkinlaiseen interventioon, kuten separatistialueiden ”pelastamiseen”, legitimointi on jäänyt ilmeisen riittämättömäksi Ukrainaan kohdistetun iskun laajuuteen nähden. Tämä artikkeli keskittyy tarkastelemaan Krimin anastuksen oikeutusta. Krimin valtausta ja sen taustakontekstia, Ukrainan kriisiä, on tarkasteltu lukuisissa tieteellisissä teksteissä konfliktiin johtaneiden geopoliittisten prosessien kautta. Lännen ja Venäjän kilpailu Neuvostoliiton jälkeisen alueen integroinnista on keskeisin kehys, josta käsin kriisin syntyä on selitetty ja seurauksia arvioitu (esim. Götz 2015; Mankoff 2014; Mearsheimer 2014; Sakwa 2015a), mutta vähemmän huomiota on saanut se kontingentti tapahtumaketju, jossa geopoliittista karttaa muokattiin käytännössä. Jälkikäteen on helppo todeta, että oli vain ajan kysymys, milloin Moskovan eliitti venyttäisi Venäjän institutionaaliset rajat yli Krimin. Eritoten silloin, kun huomioidaan Georgian sodan (2008) jälkeiset argumentit Kremlin aloittamasta (uus)imperialismista, rajarevisionismista tai Neuvostoliiton jälkeisen tilan muuttamisesta sen juonittelulle alttiiksi poikkeustilavyöhykkeeksi (Prozorov 2010). Tapahtumaa enteili myös Venäjän kansallisidentiteetin diskurssi, johon oli jo vuosikymmenen ajan sisältänyt ajatus Krimin ”palauttamisen” mahdollisuudesta (Hopf 2016). Jos oli ”päivänselvää”, että jotain ”operaatio Krimin” kaltaista oli odotettavissa, mitkä tekijät mahdollistivat tapahtuman juuri keväällä 2014? Tämä tutkimusartikkeli osallistuu valtauksen läpivientiä käsittelevään keskusteluun tarkastelemalla geopoliittisesta viitekehyksestä käsin Kremlin rakentaman oikeutusoperaation sisältöjä, niiden ajallisia painotuksia sekä vuorovaikutteisuutta tapahtumiin ja muiden osapuolten toimiin nähden. Tulkintakulmassa Putinin hallinto ammensi Ukrainan sisäisten tapahtumien sarjasta kohdeyleisöihinsä vetoavan, Krimin anastusta puoltavan geopoliittisen mielikuvan, millä oli ilmeinen painoarvo kannatuksen varmistamisessa käytännön toimille. Jos ukrainalaisosapuolten aikaansaamat olosuhteet tarjosivat lähtökohdat juonittelulle, tehokkaan oikeutusoperaation toteutus kuului Kremlin juonittelukeinoihin, joilla se junaili itselleen akuutin mandaatin niemimaan miehittämiseksi ja Venäjän federaatioon liittämiseksi. Kuten jo muinaisintialainen geopolitiikan dynamiikkaa käsitellyt Kautilya kuvaa (1915, 564), sotatoimi alkaa yrityksenä vallata sekä kotimaan että ulkomaiden väestön ajattelua. Hyökkäävän hallinnon on tuotettava suhteellisen uskottavat perusteet tulevalle invaasiolleen (Kuusisto 1998), eritoten nykyaikana, kun kansainväliset ja kansalliset instituutiot, keskinäisriippuvuus sekä potentiaaliset haitat talouden ja diplomatian alueilla kuten myös sotaisan hallinnon mahdollinen suosion lasku muodostavat kynnyksen voimankäytölle. Käytännössä edes supervalta ei voi aloittaa menestyksekästä sotaa ilman sille rakentamiaan ideationaalisia olosuhteita, joiden valossa operaatio voisi näyttäytyä suhteellisen hyväksyttävältä suhteessa omaan lainsäädäntöön, kansainväliseen oikeuteen sekä kansainvälisen yhteisön ja eritoten kotiyleisön silmissä. USA:n hallitus kampanjoi vuoden ennen hyökkäystään Irakiin. George W. Bush käynnisti suostuttelun nimeämällä Irakin ”pahan akseliin” kuuluvaksi terroristitoimijaksi ja siirtyi Husseinin yhä voimakkaampaan identifiointiin brutaalina orjuuttajana ja ”joukkotuhoaseilla” uhkailevana massamurhaajana, joka pakotti Washingtonin suorittamaan maailmanpelastussodan (Bush 29.1.2002; 12.9.2002; 7.10.2002; 28.1.2003; 26.2.2003). Operaatio vastasi länsimaista sodanoikeutusta rakentuen Riikka Kuusiston mukaan (1998, 606–607) uhkaavien vihollisten, jalojen tehtävien ja viimeisen taistelun narratiiveille, jotka kehystetään poikkeustilalla ja jyrkin vastakohdin. 9/11-tapahtumat ja laajasti tuettu Afganistanin sota tarjosivat olosuh4 IDÄNTUTKIMUS 1/2022 teet Irakin invaasion kehystämiselle seuraavana askeleena terrorisminvastaisessa sodassa (Dodds 2007, 14), mutta koska aie tuomittiin laajasti kansainvälisen oikeuden vastaisena, Washington keskittyi riittävän kannatuksen hankintaan. Kyselyjen perusteella 65–75 % amerikkalaisista oli saatu suhtautumaan sotaan myötämielisesti, mutta kannatuslukemat olivat tätäkin korkeampia skenaariossa, jossa oletettiin USA:n saavan operaatiolleen kansainvälisen tuen (Carlisle 2005, 68–72). Riittäväksi katsottu tuki realisoitui maantieteellisesti spesifin USA-Britannia-johtoisen liittouman muodossa, jonka Bush nimesi ”halukkaiden koalitioksi” (ibid.; Cohen 2015, 5). Tapaus havainnollistaa legitimiteetin hankinnan suhteellisuuden. Tiettyjen alueiden väestöt ovat samaistettavissa verrattain tiiviiksi kollektiiviksi, jota yhdistää käsitys oikeutetusta sodasta lainsuojattomaan muukalaiseen kohdistuen. Näin sotaa ei perustella ainoastaan ”järjellä”, vaan prosessissa rakennetaan kokonaisvaltaisempaa käsitystä tilallisesta ja oikeutetusta meistä ja vihollisestamme, jonka suhteen identiteettimme velvoittaa toimimaan. Kyse on toimia puoltavan geopoliittisen mielikuvan tuottamisesta. Kremlin ei tarvinnut käydä sisäpoliittista debattia operaatioon ryhtymisestä, mutta toimien tehokas legitimointi oli välttämätöntä vastareaktiot kestävän kannatuksen varmistukselle. Valtaussodan konteksti lisäsi oikeutusasetelmaan oman haasteensa – kohdealueen väestön sisällyttämisen osaksi meitä. Jo Machiavelli (1997, 6–7) käsitteli alueliitosten realiteetteja esittäen niiden onnistumisen edellyttävän paikallisten luottamuksen saamista, mihin myötävaikuttavat samankaltaisuudet kielen, kulttuurin ja instituutioiden saralla. Myös Kreml nojasi oikeutuksessaan näihin aspekteihin, joskaan ne eivät itsessään olleet anastuksen perusteita. Krimiläistenkin keskuudessa toimille saatu kannatus (O’Loughlin & Toal 2019) edellytti Kremliltä mielikuvan rakentamisen sekä krimiläisiä että kotiyleisöä yhdistävästä vihollisesta, jolta Krimin kattava Venäjä tuli suojata nopeasti. Operaation aikana Vladimir Putinin suosio kasvoi kotiyleisön keskuudessa 60 %:n tuntumasta 80 %:iin (Deutsche Welle 15.3.2019) ja saavutti reilussa vuodessa ennätykselliset 89 % (Nardelli, Rankin & Arnett 23.7.2015). Muutos heijastelee Kremlin onnistumista sekä akuutissa kannatuksen varmistamisessa että valtauksen jälkeisessä retoriikassa, vaikka Putinin asemaa pyrittiinkin horjuttamaan lännen suunnalta. Oikeastaan länsi asettui pakotepolitiikallaan helposti paketoitavaksi Venäjä-vastaiseksi monoliitiksi (Robertshaw 2015), jonka nojalla Kreml sinetöi alueliitoksensa paikallisen legitimiteetin ja integroi kansalaisia aiempaa tiukemmin kannalleen. Artikkelin suhde aiempaan tutkimukseen Vaikka Krimin anastusta on tutkittu laajasti, itse Kremlin oikeutukseen paneutuva kirjallisuus on verrattain ohutta. Tutkimustieto oikeutusretoriikan yksityiskohtaisesta kronologiasta on puutteellista, eikä suostuttelun sisältöä ole peilattu riittävästi varsinaisiin kohderyhmiin eli koti- ja krimiläisyleisöön. John Biersack ja Shannon O’Lear (2014) sekä Thomas Ambrosio (2016) ovat eritelleet oikeutusta temaattisesti käyttäen aineistoa, jonka Kreml tuotti Krimin liittämishetkellä ja sen jälkeen, jolloin kohderyhmien suostuttelu oli käytännössä ohi ja Kreml oli latistanut tapahtumien yhteydessä korostamansa retoriikan temaattiseen yhteenvetoon painotuksia muuttaen. Tutkimuksissa eritelläänkin oikeutusaspekteja symmetrisinä abstraktioina ja huomioiden kokonaisuudesta vain sen, minkä materiaali mahdollistaa. Esimerkiksi Biersack ja O’Lear määrittävät (2014, 254) aineistonsa suhteen kolme validia oikeutusaspektia: ”Venäjä määritti Krimin kuuluneen kaiken aikaa Venäjään, demonisoi Janukovytšin jälkeisen Ukrainan IDÄNTUTKIMUS 1/2022 5 hallinnon toisen maailmansodan kielikuvilla ja väitti krimiläisten käyttäneen itsemääräämisoikeuttaan liittymisessä.” Yhteenveto on ongelmallinen siinä mielessä, ettei Venäjän johto tosiasiassa edes vihjannut tutkijoiden määrittämiin ensimmäiseen ja kolmanteen kohtaan, kun sotatoimet olivat käynnissä. Miehityksen aikana kohdeyleisöjä suostuteltiin eri retoriikalla. Kremlin oikeutusta on käsitelty myös yhden narratiivin tai argumentin kautta. Magdalena B. Leichtova käsittelee (2016) liittämisen jälkeistä historiaan vetoavaa argumentaatiota. Vasile Rotaru ja Miruna Troncotă (2017) paneutuvat kampanjan lopussa esitettyyn ”Kosovo-argumenttiin”, ja Rotaru rakentaa (2019) erillisen näkökulman Kremlin pyrkimyksistä oikeuttaa toimensa lännelle. Mainituissa tutkimuksissa oikeutusta on tarkasteltu vasta Krimin liittämishetkestä, täydentävää argumentaatiota käsitellen tai länsiyleisön näkökulmasta. Kuitenkin Damian Strycharz painottaa (2020) oikeutuksen alkaneen ennen invaasiota, samoin kuin kotiyleisöön tehokkaasti vedonneen narratiivin, maanmiehiin kohdistuvan uhan, keskeisyyttä legitimoinnissa. Hän keskittyy argumentoimaan, että Kreml olisi helmikuun 2014 lopussa korvannut Ukrainan konfliktia kehystävän etäisen ja maltillisen sekaantumattomuus-narratiivinsa oppositiopoliitikkojen Vladimir Žirinovskin ja Gennadi Zjuganovin vaikutuksesta, sillä he olivat aloittaneet valtiojohdon käyttöönottaman radikaalin kielenkäytön aiemmin. Strycharz ei kuitenkaan arvioi Sotšin talviolympialaisten (7.–23.2.2014) oletettavaa vaikutusta maltillisen linjan säilyttämiselle eikä sitä, että vasta helmikuun kynnyksellä eskaloitunut konflikti tarjosi tapahtumillaan aitoa tarttumapintaa radikaalille retoriikalle. Tämä tutkimus täydentää aihetta käsittelevän kirjallisuuden puutteita kolmella tavalla. 1) Analyysi valottaa oikeutuksen kontingenttia ja vuorovaikutteista toteutusmuotoa tarkastelemalla sen kronologiaa suhteessa ukrainalaisosapuolten toimiin sekä miehityksen ja alueliitoksen ajankohtiin nähden. Näkökulmassa Kreml käynnisti akuutin suostuttelukampanjan Ukrainan sisäisen konfliktin kärjistyttyä ja jatkoi sen rakentamista pitkälti ukrainalaisosapuolten toimiin reagoivalla retoriikalla. Valtiojohto tuotti kuukauden mittaisessa kampanjassaan paitsi keskeisimmät anastusperusteet, esitti ne myös kohdeyleisöjensä kannalta vetovoimaisimmassa järjestyksessä. Kronologisen tulkintakulman kautta aiemmissa tutkimuksissa eritellyt oikeutusaspektit paikannetaan kampanjan aikajanalle, minkä kautta hahmottuu niiden rooli kokonaisuudessa. 2) Oikeutuksen sisällöntulkinta tehdään kohdeyleisöön peilaten diskursiivisretorisen geopolitiikan luennalla. Tulkintatavan kautta tulee esille, miten valtionjohto tuotti käsitteitä, argumentteja ja piennarratiiveja sekä näiden yhdistelmiä muodostaakseen koti- ja krimiläisyleisöön vetoavan mielikuvan Krimin kattavasta Venäjästä. Geopoliittisella kuvitelmalla oli ilmeinen painoarvo kannatuksen varmistamisessa, ja se kulminoitui jo ennen Krimin liittämishetkeä. 3) Artikkelin pääargumentti on, että Kreml rakensi oikeutuskuvitelmansa neljän diskurssin kautta: juridinen, turvallistava, Venäjän ja Krimin historiallinen ykseys sekä Venäjää alistava tekopyhä länsi. Kampanjan alkupuolisko painottui kahden ensimmäisen, loppuvaihe kahden jälkimmäisen tuottamiseen, mutta juridinen diskurssi oli mukana koko ajan. Kreml panosti huomattavasti siihen, että sen toimet näyttäytyisivät laillisilta, mikä ei ole aiemmissa tutkimuksissa tullut kunnolla esille. Geopolitiikka diskursiivis-retorisena operaationa Tutkimuksen aineisto koostuu Vladimir Putinin, Dmitri Medvedevin, Sergei Lavrovin ja alemman portaan päättäjien venäjänkielisistä julkilausumista sekä median, eritoten Rossija24-uutiskanavan representaatioista ajalta 20.2.–21.3.2014. Venäjän hallinto suoritti kuukauden aikana Krimin miehityksen, de facto alueliitoksen sekä toimia osin edeltäneen 6 IDÄNTUTKIMUS 1/2022 oikeutusoperaation, joka käynnistyi Ukrainan sisäisen konfliktin kärjistyttyä. Aineiston tarkastelu pohjustetaan eskalaatiota edeltäneillä tapahtumilla, mutta varsinainen analyysi alkaa ajankohdasta, jolloin Kreml sekaantui konfliktiin retorisesti. Koska oikeutusoperaatio oli kumulatiivinen kokonaisuus alkaen ukrainalaisosapuolten toimien tarkoitushakuisista tulkinnoista ja täydentyen jatkotapahtumista johdetulla retoriikalla, myös analyysikehys on kronologinen. Asetelmalla havainnollistuu oikeutuksen vuorovaikutteinen ja progressiivinen luonne; millaisella retoriikalla Kreml reagoi kuhunkin tapahtumaan ja keskeisten toimijoiden kannanottoihin, miten se vaikutti tapahtumien jatkoon ja edisti Krimin anastusta puoltavan geopoliittisen mielikuvan kasaamista. Oikeutuskampanjan sisältöön pureudutaan diskursiivisretorisen geopolitiikan luennalla, jonka seuraavaksi määrittelen. Päättäjien julkilausumat ovat harkittuja tekoja, joissa jokaisella liikkeellä pyritään edistämään tavoitteita (Hart 2004, 80). Asioita merkityksellistetään valikoidulla kielellä tuottaen rajattua todellisuuskäsitystä, jonka yleisön halutaan omaksuvan (ks. Austin 1962, 94–107). Ihmisen kielen rajat ovat hänen maailmansa rajat, kuten Ludwig Wittgenstein (1961, 68) kiteyttää. Geopolitiikan viitekehyksestä tarkasteltuna päättäjät tuottavat puheissaan valikoitua geopoliittista maailmankuvaa, kuten käsityksiä maantieteen, toimijoiden ja väestöjen välisestä suhteesta – tai suhteen laadun muuttumisesta. Yleensä prosessissa uusinnetaan kohdeyleisöjen hyväksymiä käsityksiä, mutta niitä myös muovataan paitsi palvelemaan valittua agendaa, vastaamaan muuttuvaa poliittista todellisuutta. Eritoten alueellisten konfliktien tai poikkeustilanteiden yhteydessä toimijoilla on pelivaroja muokata vallitsevia käsityksiä, joiden hyväksyttämisellä voidaan tähdätä toiminta-alttiuden luomiseen (ks. Perelman 1996, 19) maantieteeseen sidonnaiselle toiminnalle, ääritapauksessa sotimiselle. Irakin invaasiota edelsivät terroristidemonin heijastaminen Lähi-idän kartalle sekä sitä kurittamaan ryhtyvän liittouman muodostus. Merkittävästä ei-sotilaallisesta esimerkistä käy toisen maailmansodan jälkeisen Saksan länsiosien lohkaisu ja lännen kollektiiviin integrointi geopoliittisen mielikuvaoperaation onnistuttua (Jackson 2006). Näin geopoliittisen maailmankuvan rakentaminen tapahtuu integroivien ja hajaannusta kylvävien ideoiden tuottamisen kautta, millä päättäjät edistävät pohjimmiltaan meidän tilan kontrollointia tai muille kuuluvien tilojen haltuunottoa (Ahtamo 2020). Maantieteeseen sidonnaista toimintaa edeltävä järkeily, mobilisointi ja oikeutus ilmenevät Gearóid Ó Tuathailn ja John Agnewin mukaan (1992, 191) valtiojohdon diskurssissa. Kyseessä on kielenkäytön makrotason kategoria käsittäen tietyn ideakokonaisuuden esiintymisen julkilausumien laajassa skaalassa. Kuitenkin diskursiivinen tarkastelu on metodina liian väljä, jos tarkoituksena on selvittää julkilausumien suostuttelupotentiaalia. Diskurssien erittely ei vielä sinänsä selitä, miksi myös vaihtoehtoisille käsityksille altistunut yleisö niitä hyväksyisi, vaikka niillä operointi soveltuukin laajojen vallankäytön prosessien etäältä tarkasteluun. Tarkemmin katsottuna diskurssit koostuvat pienemmistä merkityskokonaisuuksista, joihin nivoutuu sanoman vakuuttavuuteen vaikuttava retorisuuskin. Tutkimusmetodologiana retoriikkaa on kehitetty julkilausumien suostutteluvoiman tarkasteluun jo antiikin Kreikan ajoilta ja se soveltuu hyvin geopolitiikan tutkimukseen, kuten Pekka Korhonen (2003, 10–11) argumentoi (ks. myös Harle & Moisio 2003). Julkilausumien retorisuutta tarkastellaan suhteessa reetorin kohdeyleisön identiteettiin, sillä kuten Chaïm Perelman kuvaa (1996, 16, 28), esitys on sovitettu ”henkisesti kohdatulle” yleisölle. Puhuja luo suhteen kuulijoihin identifioimalla heidät itseensä ja heidän intressinsä omiinsa, jolloin suostuttelu tapahtuu erilaisin samaistuksin ja puhujan toiveet siirtyvät kuuliIDÄNTUTKIMUS 1/2022 7 joiden mielipiteiksi, kuten Kenneth Burke määrittää (1969, 46, 57). Vakuuttamisessa tarjottu käsitys saadaan näyttämään tutulta, mutta suostuttelu koostuu myös kirjosta yleisösidonnaisia tekniikoita. Suostuttelussa erilaiset asiat linkitetään toisiinsa (ibid., 20–21), herätetään kuulijoissa toivottuja tunteita, todistellaan argumentein, joita vastaanottajat voisivat hyväksyä ja asiat esitetään (suur)narratiivein, jolloin aikajanalle sijoitetut tapahtumat, toimijoiden luonnehdinta, yksityiskohdat ja draaman elementit liikuttavat yleisöä tiettyyn reaktioon (Hart 2004, 81–82, 88–91). Suostuttelu on myös ikään kuin neuvottelua: reetori suhteuttaa väitteet yleisön uskomuksiin, oletuksiin, kokemuksiin ja tunteisiin tarkoituksenaan tehdä ne ”järkeenkäyviksi”. Eritoten ajallisuus on retoriikan tutkimisessa keskeistä, sillä suostuttelussa nojataan akuuttien asioiden tärkeyteen ja käytännöllisiin ehdotuksiin, jotka yleisön on helppo tunnistaa järkeviksi. (ibid., 79, 81.) Näin retoriikan analyysi kohdistuu konkreettiselle mikrotasolle eli esiintymiskontekstissaan tuotettuihin lauseisiin ja sanoihin, joilla päättäjät muodostavat palapelinomaisesti geopoliittisen mielikuvan. Diskursiivis-retorinen geopolitiikan luenta yhdistää käytännön toimien legitimoinnin mikro- ja makrotason tarkastelun. Yhtäältä tulkintakulmalla valotetaan kunkin oikeutusoperaation ainutlaatuisuutta, kuten esiintymiskontekstia, suhdetta käytännön toimiin, retoriikan ja sen tuottamistavan yksityiskohtaisuutta sekä oikeutuksessa rakennetun geopoliittisen mielikuvan suostuttelupotentiaalia kohdeyleisöihin nähden. Toisaalta mikrotason analyysin kautta hahmottuvassa oikeutuskokonaisuudessa voi havaita säännönmukaisuutta, kuten retoristen elementtien jakautumisen diskursiivisiin kategorioihin. Näin oikeutuksessa käytetyt kielelliset resurssit voidaan kiteyttää abstrahoituun muotoon, mikä muodostaa vertailupohjan vastaavien operaatioiden tarkastelulle. Tässä tapauksessa metodin kautta lisätään ymmärrystä siitä, millaista kielellistä vallankäyttöä Venäjän hallinto harjoitti erityisesti vallatessaan Krimiä ja mitkä diskursiiviset rakenteet ilmentävät Kremlin oikeutusta yleisesti. Tapahtumat Kremlin retoriikan muutoksen taustalla Kun Viktor Janukovytš hylkäsi 21.11.2013 EU:n kanssa neuvottelemansa assosiaatio- ja vapaakauppasopimuksen, alkoivat protestit Kiovan Maidanilla. ”Euromaidan” laajeni päivissä maan länsi- ja keskiosiin arviolta 300 000 mielenosoittajan voimin. Joulukuun ensimmäisenä noin puoli miljoonaa kapinoitsijaa kerääntyi Kiovaan, Leninin patsas kumottiin ja poliisi käytti voimakeinoja. Eskaloituminen alkoi 16.1.2014 parlamentin säädettyä lakeja, jotka mahdollistivat pitkät vankilatuomiot levottomuuksien järjestäjille. Vastareaktiot kärjistivät väkivallan ja ensimmäisen protestoijan kuolema 22.1. voimisti kapinan vallankumousliikkeeksi. Aseistautuneet protestoijat valtasivat Volynian alueen sekä Lvivin ja Ternopilin kaupunkien hallintorakennuksia ja anastivat asevarastojen kalustoa, jolla varustivat myös Maidanin mielenosoittajia. 15 poliisia ja 77 kapinoitsijaa menehtyi 20.2. mennessä. Seuraavana päivänä Janukovytš ja opposition edustajat Arseni Jatsenjuk, Vitali Klytško ja Oleg Tjagnibok solmivat sopimuksen todistajinaan Saksan, Ranskan ja Puolan ulkoministerit sekä Venäjän edustaja. Janukovytšin tuli allekirjoittaa 48 tunnissa lakiehdotus, joka palauttaisi vuoden 2004 perustuslain. Perustettaisiin presidentin valtaa rajoittava hallitus, mitä seuraisivat presidentinvaalit. Myös konfliktin päättämiskeinot määriteltiin. Kun oppositioedustajat esittelivät sopimusta Maidanilla, ei se kapinallisille kelvannut. Janukovytš pakeni vallasta syöstynä saman päivän aikana. (Sakwa 2015b, 68–73.) Levadan kyselyissä (26.2.2014) venäläisenemmistön suhtautuminen Ukrainan tapahtumiin muuttui joulukuusta helmikuun puoliväliin tultaessa neutraalista närkästykseen, mutta 8 IDÄNTUTKIMUS 1/2022 65 % vastanneista ei suosinut Janukovytšin hallintoa eikä mielenosoittajiakaan. Samaisella aikavälillä Kreml kehysti kriisin Ukrainan sisäiseksi asiaksi ja painotti etäistä ja maltillista suhtautumistaan siihen (Strycharz 2020, 5). Kuitenkin vallankumouksen kynnyksellä tapahtunut kärjistyminen laajensi sisäpoliittisen yhteenoton kansainväliseksi konfliktiksi (Unikaitė-Jakuntavičienė & Matulkaitė 2017, 300–301), sillä Kreml ja länsijohtajat ajautuivat kiistaan osapuolten ja tapahtumien luonteesta. Valkoisen talon tiedote sisälsi selvän tulkinnan syyllisestä: Olemme raivoissamme kuvista, joissa Ukrainan turvallisuusjoukot ampuvat automaattiaseilla kansalaisia. Kehotamme presidentti Janukovytšia vetämään välittömästi turvallisuusjoukkonsa pois Kiovan keskustasta ja kunnioittamaan oikeutta rauhanomaiseen protestointiin, ja kehotamme mielenosoittajia ilmaisemaan asiansa rauhanomaisesti. (Office of the Press Secretary 20.2.2014.) Ulkoministeri John Kerry (20.2.2014) yhtyi kuvaukseen ja kehotti Janukovytšia perustamaan opposition kanssa väliaikaishallinnon tuottaen Washingtonille tyypillistä opposition valtaanpääsyä edistävää retoriikkaa (ks. Tsygankov 2009, 7). Lavrov puolestaan syytti (20.2.2014) länttä osapuolten ja tapahtumien valikoivasta kehystämisestä ja esitti päinvastaisen tilannekuvan. Žirinovskin ja Zjuganovin kuukausia aiemmin aloittamaan kielenkäyttöön yhtyen (Strycharz 2020) ulkoministeri nimesi (20.2.2014) mielenosoittajat radikaaleiksi, ekstremisteiksi ja antisemitisteiksi, joiden vastapainona oli hyveellinen hallinto: Ei autojen tuhopolttoja, ei poliisin päälle viskottuja polttopulloja, ei murhia, joita tehtiin ekstremistien valtaamissa ”Alueiden Puolueen” tiloissa ole kommentoitu […] Vaiti pysytään eri Ukrainan alueiden kuvernööreihin kohdistetusta väkivallasta, faktoista liittyen ”onnistuneisiin” yrityksiin kaapata asevarastot. Vain hallitusosapuolta kehotetaan lopettamaan väkivallan kohdistamista ”ilmeisen rauhanomaisiin mielenosoittajiin” […] Ukrainan hallitus ja presidentti ovat jo useamman kerran osoittaneet hyvää tahtoa ehdottaen kompromissia, sopivat Maidanilla vangittujen vapauttamisesta vastineeksi hallintorakennusten vapauttamisesta. Hallitus on toteuttanut oman osuuden sopimuksesta, oppositio sen sijaan repi sen. […] Oppositio ei voi tai ei halua erottautua ekstremisteistä, mutta meidän läntiset kumppanimme Euroopassa ja Yhdysvalloissa vierittävät koko syyllisyyden Ukrainan viranomaisille eivätkä määrittele ekstremistien toimia asianmukaisesti. […] ukrainalaisten tulee selvittää asiat itse, oman perustuslakinsa puitteissa, ilman pakottavien ehdotuksien hyväksymistä. (Lavrov 20.2.2014.) Kannatuksen saaminen ulkomaisiin konflikteihin sekaantumiselle edellyttää pahan vihollisen ja asiallisten interventioperusteiden artikulointia toteutuen osapuolten nimeämisillä ja toimien juonenomaisella selittämisellä, millä vaikutetaan tapahtumien jatkoon (Kuusisto 1998, 603, 606). Lavrov rakensi käsitystä järjestystä ylläpitävästä hallinnosta, joka oli jäämässä kaaosta aiheuttavan pahan toimijan uhriksi. Pahaa tuottavia narratiivinkatkelmia olivat raakuuksia ja laittomuuksia tekevät ääriryhmät, opposition epäluotettavuus neuvottelukumppanina ja linkki radikaaleihin sekä lännen tarkoitushakuiset tulkinnat ja painostus väliaikaishallinnon perustamiseen. Niiden kautta muodostui hatara mielikuva eri tasoin linkittyvästä muukalaiskollektiivista, joka ikään kuin kontrolloi rasistisia väkivaltaryhmiä. Hyvän narratiivinkatkelmia olivat vastuullinen ja reilu Janukovytšin hallinto, ukrainalaiset IDÄNTUTKIMUS 1/2022 9 kohtalonsa päättäjinä ja juridisten normien ehdottomuus konfliktin päättämisessä. Puheen tyyli, kuten etäisiksi kuvatut ”meidän läntiset kumppanimme”, ilmentää retoriikan kohdistuneen kotiyleisölle. Kaiken kaikkiaan sitaattiin kiteytyy argumentatiivinen pohja, johon Kreml ankkuroi jatkotapahtumista johtamansa retoriikan – Ukrainan valtiollinen tila tulee säilymään järjestyksen hallinnassa tai romahtaa kaaoksen valtaan. Päästyään valtaan uusi hallinto erotti Janukovytšin (BBC 22.2.2014), muttei noudattanut perustuslain virkasyyteartiklan proseduuria (Biersack & O’Lear 2014, 264; Sakwa 2015b, 76). Janukovytš piti toimea laittomana kaappauksena verraten sitä natsien nousuun 1930-luvun Saksassa eikä suostunut eroamaan (BBC 22.2.2014). Kumousta tukeneet länsijohtajat tunnustivat uuden hallinnon, minkä Kreml näki Ukrainan perustuslain vastaisena kaappauksena (Tsygankov 2014, 193). Dmitri Medvedev väitti (24.2.2014), ettei hallintoa ole ja viittasi länsijohtajien lukeneen vääriä perustuslakeja pitäessään aseellisen kapinan tuloksena valtaan nousseita legitiimeinä. Uusi parlamentti, Verhovna Rada, äänesti 23.2. ukrainan asettamisesta ainoaksi viralliseksi kieleksi ja kumosi vuoden 2012 lain, joka puolsi rinnakkaisen asiointikielen käyttöönoton alueilla, joiden väestöstä 10 %:lla on muu äidinkieli kuin ukraina. 27 hallintoalueesta 13 oli virallistanut venäjän. (Sakwa 2015b, 54–55.) Vuosien 2006–2007 tutkimuksissa 52 % kansalaisista piti ukrainaa ja 25,7 % venäjää äidinkielenään. Uudemmissa kyselyissä venäjää työkielenään käyttäviä on 27,5 miljoonaa ja 80 % väestöstä osaa sitä sujuvasti. (Marples 2015, 10–11.) On myös arvioita, että vain 38 % Ukrainan asukkaista käyttää ukrainaa arkielämässä (Heiskanen 6.3.2014). Sotatoimen oikeutus alkaa Kielilaki tarjosi Kremlille merkittävän retorisen potentiaalin. Sen lisäksi että Lavrov (25.2.2014) tarttui Janukovytšin ja opposition välisen sopimuksen laiminlyöntiin, tulkitsi hän kielilain tarkoittavan venäjänkielisten ja muiden vähemmistöjen perusoikeuksien epäämistä. Medvedev ja Lavrov tuottivat nyt juridisia narratiiveja koostuen laittomasta vallankaappauksesta, sitä edeltäneestä sopimusrikkeestä sekä vähemmistöjä syrjivän kielilain asetuksesta. Näin luotiin käsitys Ukrainasta anarkiaan vajonneena maantieteellisenä tilana, jossa ”barbaarit” polkivat ihmisoikeuksia. Ulkoministeri aloitti (ibid.) myös tulevan miehityksen oikeutusretoriikan kulmakiven, turvallistavan narratiivin tuottamisen: ”nyt radikaalit uhkaavat ’vaeltaa’ niille valtion alueille, jotka ilmaisevat erimielisyytensä opposition käyttämiä metodeja kohtaan”. Vihjatessaan venäjänkielisten kohtaaman epäoikeudenmukaisuuden lisäksi fyysisen uhan mahdollisuuteen, ulkoministeri hahmotteli skenaarion, jossa sotatoimi olisi perusteltavissa viattomiin kohdistuvan hyökkäyksen estämisenä. Barry Buzan, Ole Wæver ja Jaap de Wilde kuvaavat (1998, 21, 23–26) turvallistamisliikkeen puhetekona, jossa rakennetaan käsitystä eksistentiaalisesta uhkatilanteesta ja oikeutetaan sen ehkäisevät poikkeustoimet. Turvallistaminen onnistuu, jos yleisö omaksuu uhkakuvan ja normeista vapautetun torjuntapolitiikan riippumatta turvallisuusuhan todellisuudesta (ibid.). Kreml representoi juridisella retoriikallaan järjestystä ylläpitävän valtiovallan hävinneen Ukrainan maaperältä, mikä tarkoitti myös alueen väestön oikeuksien takaajan ja fyysisen suojelijan kuvitteellista katoamista. Turvallistamalla venäjänkieliset ja venäläiset alttiiksi väkivaltaisten ääriryhmien hyökkäykselle, Venäjän hallinto esitti kotiyleisölleen retorisen kysymyksen: kenen puolustusvastuulla nämä maanmiehemme ovat? 10 IDÄNTUTKIMUS 1/2022 Uhkakuva sovitettiin virallisiin toimintakriteereihin. Ulkopoliittisen konseptin mukaan Venäjän hallinnon tulee ”puolustaa ulkomailla asuvien Venäjän kansalaisten ja maanmiesten oikeuksia ja laillisia intressejä” (Venäjän presidentin hallinto 15.7.2008). Sotilasdoktriinissa määritellään asevoimien ja muiden joukkojen rauhanajan tärkeimmiksi tehtäviksi: ”Venäjän federaation kansalaisten puolustaminen Venäjän federaation rajojen ulkopuolella heihin kohdistuvalta aseelliselta hyökkäykseltä” (ibid. 5.2.2010). Ilmeisenä pyrkimyksenä oli myös kannustaa Itä-Ukrainan venäläisiä ja venäjänkielisiä mellakoimaan uutta hallintoaan vastaan osoittaen heille Venäjän hallinnon tukea. Venäjänkielisten enemmistöstä koostuvissa Krimin ja Itä-Ukrainan kaupungeissa (Sakwa 2015b, 25) kielilaista ei pidetty. Sevastopolissa venäjämieliset protestoivat lakia vaatien Krimille oikeutta päättää asioistaan (Pajunen 27.2.2014; Rossija24 25.2.2014a). Simferopolissa Krimin parlamenttirakennuksen edustalla mieltä osoittivat venäjämielisten lisäksi tataarit, jotka vaativat Krimin säilymistä Ukrainan osana. Ryhmät ottivat yhteen. (Rossija24 26.2.2014a; 27.2.2014a.) Presidentin tehtävää hoitanut Oleksandr Turtšynov reagoi protesteihin 28.2. kieltäytyen allekirjoittamasta kielilakia, mutta Krimin miehitys oli silloin jo käynnissä (Sakwa 2015b, 76–77). Aamulla 27.2. niemimaalle ilmestyi vihreäuniformuisia sotilaita, joilla ei ollut tunnusmerkkejä, mutta kehittyneet aseet. ”Hiljaiset gentlemannit” tai ”pienet vihreät miehet” piirittivät Ukrainan sotalaitteistot ja valtasivat Simferopolin lentokentän sekä hallintorakennuksia. Ukrainan varuskunnat antautuivat taisteluitta ja rauhallisesti toimineet sotilaat työntyivät poliisin ja turvallisuuspalvelun tiloihin. Paikallisina vapaaehtoisina esittäytyneet Venäjän erikoisjoukot saivat laajan paikallisväestön tuen. (Biersack & O’Lear 2014, 249; Galeotti 2015, 159–160; Sakwa 2015b, 82.) Krimin parlamentin puheenjohtaja Vladimir Konstantinov kieltäytyi tunnustamasta Ukrainan uutta hallintoa ja vallatussa rakennuksessa äänestettiin vajaalla edustajamäärällä niemimaan pääministeriksi Sergei Aksjonov. Proseduuri oli Ukrainan lain vastainen. Krimin johto vaati niemimaalle autonomiaa. (Biersack & O’Lear 2014, 250; Katchanovski 2015, 86.) Niemimaan sotavoimien, poliisin ja turvallisuuspalvelun johtajaksi julistautunut Aksjonov vetosi lausunnossaan Venäjän presidenttiin rauhan varmistamiseksi Krimillä (Lewis, Traynor & Harding 28.2.2014). Kreml ei kommentoinut tapahtumia, mutta suostuttelua ylläpitivät media ja rivipoliitikot. Rossija24-kanavan uutisointi (25–28.2.2014) sisälsi yksityiskohtia ja symboliikkaa, joilla kotiyleisöä samaistettiin Ukrainan venäläisiin empatiaa herättäen. Kuviin päätyivät Krimin ja Itä-Ukrainan hallintorakennusten salkoihin vedetyt Venäjän liput ja toisen maailmansodan voittoa symboloivat Yrjön nauhat, joita pitivät uutta hallintoaan vastustavat ihmisjoukot huutaen ”Venäjä, Venäjä” -mantraa. Haastatellut pyysivät Venäjän apua ja toivoivat pääsevänsä valtion hallinnon piiriin. Uutissisältö esitettiin rauhallisesti, mutta otsikoitiin dramaattisesti. Vaikka tarvetta interventiolle ei suoranaisesti esitetty, skenaario enteili pahaa. Ruuduilla näytettiin aseettomien amatööriryhmien tienvalvontapisteitä ja kerrottiin hallintorakennusten päätymisestä itsepuolustusryhmien hallintaan. Mainittiin kuitenkin lentokenttien päätyneen joukoille, jotka eivät silminnähden kantaneet aseita ensi kertaa, mutteivät paljastaneet taustaansa. Duuman rivipoliitikot ehdottivat lakia, joka sallisi halukkaille ukrainalaisille Venäjän kansalaisuuden. (Rossija24 25.2.2014a; 25.2.2014b; 26.2.2014a; 26.2.2014b; 26.2.2014c; 27.2.2014a; 27.2.2014b; 27.2.2014c; 27.2.2014d; 28.2.2014.) Kansalaisuuden myöntöpuheet vahvistivat uhkakuvaa vihjaamalla pakolaisaaltoon, mitä media ruokki uutisoimalla krimiläisten saapumista Venäjän konsulaattiin turvapaikkaa anoen IDÄNTUTKIMUS 1/2022 11 (Lenta 27.2.2014a). Oikeudenmukainen Venäjä -puolue ehdotti kansalaisuuden myöntöä nopeuttavaa lakia (Lenta 27.2.2014b). Massamyönnöt olisivat lisänneet nopeasti de jure puolustettavien määrää. Kreml käytti ”passiensa puolustamista” interventioverukkeenaan Georgiaan, mutta Krimin tapauksessa passeihin ei viitattu. Kuten Jeffrey Mankoff huomioi (2014, 64), Kremlin määritelmä venäläisistä oli joustava, mikä maksimoi suojeltavien määrän. Tieto miehityksestä levisi nopeasti. Barack Obama (28.2.2014) tuomitsi sen Ukrainan suvereniteetin loukkauksena ja kansainvälisen oikeuden vastaisena, vihjasi sanktioista ja antoi tukensa Ukrainan hallinnolle. Vaikka Venäjän valtamedia kuvasi Krimin kohteet kaapanneet itsepuolustusryhminä, uutisoitiin (Kommersant 28.2.2014; Lenta 28.2.2014) myös ukrainalaispoliitikkojen väitteet Venäjän erikoisjoukkojen toimista. Kotiyleisöllä oli mahdollisuus päätellä hallintonsa interventio. Turvallistaminen kulminoituu Maaliskuun ensimmäisenä Kreml julkaisi Putinin (1.3.2014) vetoomuksen liittoneuvostolle, jossa hän anoi Ukrainan poikkeustilasta johtuvaan Venäjän kansalaisten ja maanmiesten hengenvaaraan nojaten puoltoa asevoimien käytölle Ukrainassa, kunnes se yhteiskuntapoliittisesti normalisoituu. Mandaatti myönnettiin (Rossija24 1.3.2014a). Kyseessä oli sotilasdoktriinin interventiokriteeriin nojaava asevoimien käyttöönoton näytösproseduuri, jossa Ukrainan kehystys poikkeustilana ja maanmiesten turvallistaminen tehtiin virallisesti korkeimmalta taholta. Uhkaahan ei todellisuudessa ollut, sillä Krim oli jo vallattu. Siirron järkeilyn tekivät alemman portaan päättäjät. Liittoneuvoston puheenjohtaja Valentina Matvijenko (ibid.) perusteli puoltoa mahdollisuudella suojata Krimin venäläisväestöä, jonka osuudeksi hän mainitsi 60 %. Hän myös vertasi tilannetta Yhdysvaltain koneiston käynnistymiseen uhan kohdatessa yksittäisiä amerikkalaisia (ibid.). Kremlin tiedottaja Dmitri Peskov totesi, ettei toimia ole kuitenkaan aloitettu, mutta niitä harkitaan tilanteen kehittyessä uhkaavaksi (Rossija24 1.3.2014b). Rivipoliitikot maalailivat selkeämpiä uhkakuvia, kritisoivat Obaman lausuntoja ja toimivat Putinin vetoomuksen ”kaikukoppina” (Rossija24 1.3.2014c). Media otsikoi Ukrainan armeijan mobilisointiaikeista (Rossija24 1.3.2014d; 1.3.2014e) näyttäen jo päiviä haastatteluja krimiläisten avunpyynnöistä ja toiveista päästä Venäjän hallinnon piiriin. Turvallistaminen edisti kotiyleisön ja uhreina kuvattujen Ukrainan venäläisten samaistamista sekä ahdistuksen, pelon ja raivon nostattamista, mikä palveli suoritetun miehityksen järkeilyä ja virallisten joukkojen käyttöoptiota, jos Ukrainan johto olisi puolustanut Krimiä aseellisesti. Biersack ja O’Lear arvelevat (2014, 250–251) uuden hallinnon ohjeistaneen Krimin armeijajoukkojaan pidättäytymään hyökkäämästä venäläissotilaisiin estäen Georgiassa tapahtuneen skenaarion, jossa Kreml sai interventioverukkeen Georgian iskiessä Etelä-Ossetian separatistialueelle, jossa oli venäläissotilaita. Viimein maaliskuun kolmantena Venäjän johtoporras perusteli invaasiotaan myöntämättä sitä suoranaisesti. Lavrov piti Genevessä YK:n ihmisoikeusneuvoston kokouksessa puheen, jossa turvallisti yhä järeämmin ja hyökkäsi länsijohtajien kritiikkiä vastaan. Aiemman tarinansa toistettuaan (Lavrov 3.3.2014) hän samaisti Ukrainan hallinnon ja ääriryhmät selkeämmin yhdeksi venäläisiä ja venäjänkielisiä tuhoamaan ryhtyväksi pahaksi: Kiovan keskusta ja monet muut Länsi-Ukrainan kaupungit kaapattiin aseellisten nationalistiradikaalien toimesta äärimmäisiä, venäläisvastaisia ja antisemitistisiä iskulauseita käyttäen […] Ukrainan parlamentti päätti rajoittaa vähemmistökieliä edustavien oikeuksia 12 IDÄNTUTKIMUS 1/2022 […] Vaaditaan venäjän kielen käytön rajoittamista ja sen rangaistavaksi asettamista, vastenmielisten poliittisten puolueiden kieltoa, puhdistuksen suorittamista. (Lavrov 3.3.2014.) Koska uutta hallintoa ei voinut nimetä natsijohdoksi, Lavrov (3.3.2014) identifioi sen sellaiseksi toimijoiden (aiemmin määritetyt) ominaisuudet linkittäen. Ääriryhmät, jotka valtasivat ja polttivat hallintorakennuksia, pahoinpitelivät poliiseja, ryöstivät asevarastoja, nöyryyttivät virkamiehiä, vandalisoivat kirkkoja ja valmistautuivat rankaisemaan uuden hallinnon päätösten vastustajia, assosioituivat natseiksi antisemitisminsä, venäläisvastaisuutensa ja nationalistiradikaaliutensa kautta. Ja hallinto, joka jo Janukovytšin retoriikassa kaappasi vallan natsien tavoin, jakoi radikaalien agendan venäläisten ja venäjänkielisten hävittämisestä, mikä loi mielikuvan yhtenäisestä fasistikollektiivista. Kun suostuttelulogiikan uskottavuus riippuu kuulijakunnasta (Hart 2004, 81–82), tiettyjen järkeilytapojen käyttöönotto rajaa kohdeyleisöä. Paholaisen rooliin valittu natsivihollinen osoittaa, ettei Kreml kohdistanut suostutteluaan länsiyleisölle, jolle fasismin torjunta on “terroristeja kurittavaa maailmanpelastusretoriikkaa” vieraampaa. Valinta vetosi venäjänkielisen maailman syväkulttuuriseen identiteettiin. Natsit ovat voitonpäivää juhlistavalle sankarikansalle pahin Neuvostoliiton kukistama demoni. Lavrov (3.3.2014) muistuttaakin fasismin nujertamisen 70-vuosijuhlasta. Hänen lausuntoonsa ”natsit kaappasivat vallan ja hyökkäävät venäläisiä vastaan” kiteytyy ajatus, jolla voittajakansassa lietsottiin vihaa revanssia tarjoten. Puolustautuminen oli velvollisuus ja identiteettiä uusintava motiivi. Sotatoimen oikeutus edellyttää yleisön arvostamien traditioiden ja periaatteiden liittämistä tiettyihin tapahtumiin ja tekoihin (Kuusisto 1998, 606). Lavrov integroi viholliskollektiivin ja meidät ominaisuuksien perusteella, mutta geopoliittisesti ratkaisevaa oli näiden identifiointi myös maantieteellisinä vastinpareina. Tässä maantieteellisessä asetelmassa Krim näyttäytyy meidän etulinjanamme, joka oli säästynyt toistaiseksi, mutta tarvitsi kollektiivin keskuksen, Kremlin varmistusta selviytyäkseen: Tilanteessa, jossa vallitsi jatkuva ultranationalistien väkivallan uhka altistaen venäläiset ja koko venäjänkielisen väestön hengenvaaraan ja jossa heidän lailliset intressinsä olivat uhattuina, muodostivat ihmiset itsepuolustusjoukkoja, joiden on pitänyt jo estää Krimin hallintorakennusten voimakkaat valtausyritykset sekä aseiden ja ammusten tuonti niemimaalle. Tietoon on tullut uusien provokaatioiden valmistelut kohdistuen myös Ukrainan alueella sijaitsevaan Venäjän Mustanmeren laivastoon. Kyseisissä olosuhteissa tämän autonomisen tasavallan laillisesti valitut viranomaiset kääntyivät presidentti Putinin puoleen pyytäen apua rauhan palauttamiseksi Krimillä. (Lavrov 3.3.2014.) Näin Kreml naamioi joukkojensa toiminnan krimiläisten itsepuolustusryhmien teoiksi ja ylläpitäessään uhkakuvaa pidätti option ”ennakoivalle puolustustoimelle”, joka voitiin toimittaa apuna ”autonomisen tasavallan laillisten toimijoiden” (laittomasti pääministeriksi asetetun Aksjonovin) pyynnöstä. Lavrov rakensi juridisia ja humanitaarisia perusteita viralliselle interventiolle, jolle Putin oli jo saanut liittoneuvostolta mandaatin. Ulkoministeri täsmensi (ibid.) vielä tärkeimmän perusteen mahdollisille toimille: ”Kyse on kansalaistemme ja maanmiestemme puolustamisesta, tärkeimmän ihmisoikeuden puolustamisesta – oikeudesta elämään.” Lännen kritiikkiin Lavrov vastasi bumerangin lailla syyttäen (3.3.2014) sanktioilla uhkailijoita Ukrainan kriisin luomisesta geopoliittisilla peleillään ja yllyttäessään ääriryhmiä väkivaltaiseen hallituksen kaatoon. Sanktioiden perusteita hän ei kuitenkaan kyseenalaistanut, IDÄNTUTKIMUS 1/2022 13 mikä on tulkittavissa sotatoimet myöntävänä eleenä, vaan siirsi huomion ”lännen suurempaan syyllisyyteen”. Päivää myöhemmin Putin piti tiedotustilaisuuden, jossa suostutteli kotiyleisön ohella ukrainalaisia. Hän oli ymmärtävinään Maidanin protestoijia ja suhtautui myötätunnolla korruptioon kyllästyneisiin ukrainalaisiin, jotka olivat kärsineet itsenäistymisestään saakka huonoista hallituksista ja alueita hallitsevista oligarkeista sekä ”tottuneet siihen, että yhdet huijarit vaihdetaan toisiin huijareihin ja roistoihin” (Putin 4.3.2014). Kansalla oli hänen mukaansa luonnollinen halu muutokseen, mutta vallanvaihdosta ei saanut tehdä aseellisesti, koska se johti juridiset instituutiot hävittävään kaaokseen: Miksi piti ryhtyä laittomiin perustuslain vastaisiin toimiin ja viedä valtio, vetää valtio siihen kaaokseen, jossa se on tänä päivänä? […] jos jollekin on sallittua laiminlyödä kaikki [lait], niin tämä oikeus kuuluu myös kaikille muille ja kaaos alkaa. (Putin 4.3.2014.) Analogiat toimivat suostuttelussa todellisuuden rakennetta luovina todistelukeinoina, joilla asiat saadaan näyttämään samankaltaisilta (Perelman 1996, 129–135). Putinin analoginen argumentti kuuluu: kun kerran Kiovassa asetettiin uudet standardit suorittamalla laiton ja aseellinen vallankaappaus, niin uudet standardit saivat olla myös krimiläisten käytössä. Väitteelleen relevanssia hän haki argumenttinsa perustaa eli laittoman kaappauksen tosiasiallisuutta täsmentäen: Legitiimi presidentti, puhtaan juridisesti, on kiistattomasti vain Janukovytš. Ukrainan lain mukaan on kolme tapaa, kuinka presidentti menettää asemansa: kuolema, […] hänen henkilökohtainen eroamisensa ja kolmanneksi – virkasyyte. […] Siihen pitäisi osallistua perustuslakituomioistuin, korkein oikeus, itse Rada. Se on monimutkainen, pitkä proseduuri. Tätä proseduuria ei noudatettu. Tästä syystä juridisesti katsoen kyse on kiistattomasta faktasta. (Putin 4.3.2014.) Olkoonkin Janukovytš epäonnistunut presidentti ja huijari kuten edeltävät ja nykyiset hallitsijat, mutta Putin palautti hänelle kuvitteellisen vallan, sillä tarvitsi tätä oikeutuksessa. Kun Lavrov oli perustellut mahdollista asevoimien käyttöä Krimin johdon avunpyynnöllä, laajensi Putin (ibid.) argumenttia laillisen presidentin pyyntöön nojaten: ”meillä on suora vetoomus virassa olevalta ja legitiimiltä, kuten jo sanoin, Ukrainan presidentti Janukovytšilta asevoimien käyttämiseksi Ukrainan kansalaisten elämän, vapauden ja terveyden suojelemiseksi”. Interventio olisi Putinin mukaan (ibid.) legitiimi ja kansainvälisen oikeuden mukainen, joskin hän toisteli, ettei voimankäytölle ollut sillä hetkellä tarvetta ja painotti ettei Venäjä tulisi koskaan kohdistamaan aseitaan itse veljeskansaan. Putin muutti asetelman haalimalla Kremlin suojelun piiriin kaikki Ukrainan kansalaiset, millä viittasi epäsuorasti yhteiseen viholliseen. Tuon toimijan voimankäyttö kohdistuisi presidentin kuvauksessa lapsiin ja naisiin sekä tavallisiin työssäkäyviin, kuten insinööreihin ja tekniseen henkilöstöön, joita nationalistiradikaalit olivat polttaneet hallintorakennusten mukana elävältä Länsi-Ukrainassa (Putin 4.3.2014). Insinöörien mainita oli tärkeä yksityiskohta. Venäjällä ammattiin valmistuu lähes puoli miljoonaa opiskelijaa vuosittain ja titteli on yleinen myös Itä-Ukrainan Luhanskissa ja Donetskissa, joissa olivat ennen kriisiä maan suurimmat teollisuuskeskukset (Ayres 28.9.2014; Myers 17.9.2015). Näin Putin lietsoi pelkoa ja vihaa herättäen itäistä väestöä mellakoimaan Kremlin tuella: 14 IDÄNTUTKIMUS 1/2022 Ja kun me näemme tällaista, niin ymmärrämme, mikä huolestuttaa Ukrainan kansalaisia. […] Ja jos näemme, että tämä kaaos alkaa itäisillä alueilla, jos ihmiset pyytävät apuamme ja virassa oleva legitiimi presidentti on jo virallisesti vedonnut meihin, niin me pidämme oikeuden käyttää kaikkia keinojamme näiden kansalaisten suojelemiseksi. Ja katsomme sen olevan täysin legitiimiä. (Putin 4.3.2014.) Lavrovin tapaan Putin ei demonisoinut uutta hallintoa eksplisiittisesti, eikä väittänyt suoraan, että Kiova mobilisoi armeijaansa Krimille ja Itä-Ukrainaan. He käyttivät metonymiaa, jossa ”itään vaeltavat uusnatsit” assosioitiin hallinnon sotilaiksi. Putinilla tämä esiintyy tarinallisena analogiana Hitlerin valtaannoususta Ernst Röhmin rynnäkköjoukkojen avulla (ibid.). Turvallistamisen kohdistaminen Itä-Ukrainaan viittaa Kremlin toiveikkuuteen Krimin skenaarion mahdollisuudesta. Syventääkseen mielikuvaa Venäjän ja (itäisen) Ukrainan ykseydestä Putin piti puheenvuoron (ibid.) ”kollegansa saappaissa”. Janukovytš ei ollut kuulemma tyrmännyt EU-sopimusta, vaan halusi keskustella sen vaikutuksista, koska ei voinut nostaa energian hintaa eikä repiä taloussuhdetta kolmanneksen viennistä kattavaan Venäjään. Sopimus olisi tuhonnut koneiden viennin, koska länsi ei niitä ostaisi, ja siten kaatanut yhtiöitä työttömyyttä aiheuttaen. Siirto ei olisi palvellut kansallisia intressejä, ”Putinkovytš” totesi. Numeerisella retoriikalla todistellen (Hart 2004, 84) Putin esitti Ukrainan taloudelle elintärkeät Venäjä-intressit, joiden rinnalla Euromaidan näyttäytyi epärelevantilta, Janukovytšin väärinkäsittäneeltä reaktiolta. Venäjän ja Ukrainan integroinnin taustalla piirtyi mielikuva EU:n muukalaisuudesta. Kun toimittaja kysyi olivatko Krimin kohteita vallanneet sotilaat Venäjän joukkoja, Putin (4.3.2014) katsoi lattiaan ja vastasi: ”ne olivat paikallisia itsepuolustusjoukkoja”. Ja kun toimittaja kommentoi univormujen muistuttaneen Venäjän armeijan asua, Putin näytti hermostuneen. Hän väitti, että vastaavia univormuja saa kaupoista ja käänsi huomion Kiovassa ammattimaisesti toimineiden protestoijien varustukseen. Itse miehitykseen hän viittasi epäsuorasti, lieventävillä asianhaaroilla ja väheksymällä sen vakavuutta: ”Ihmiset tulivat, blokkasivat aseistetut yksiköt […] ja sopivat niiden kanssa siitä, että heidän tulisi noudattaa alueella asuvan kansan vaatimuksia ja tahtoa […] kukaan ei ampunut, ei ollut yhtään laukausta” (ibid.) Argumentin mukaan oli välttämätöntä tehdä minimaalinen rike, jolla estettiin dramaattisempi skenaario. ”Tällä tavoin Krimin jännittynyt tilanne, jonka suhteen oli optiona käyttää asevoimia, oli yksinkertaisesti poistettu”, Putin totesi (ibid.). Kun Krim oli jo hallussa, turvallistamista ei tarvinnut jatkaa. Invaasion vakavuuden vähättely ilmeni myös epäsuorassa suhteellistuksessa Washingtonin sotarikoksiin Lähi-Idässä ja Libyassa; Lavrovin tapaan Putin viittasi länteen suurempana rikollisena (ibid.). Seuraavien päivien (7.–10.3.) kyselyissä 62 % vastanneista Venäjän kansalaisista käsitti Ukrainan olevan vailla legitiimiä hallintoa ja anarkian vallassa, 67 % piti kriisin kärjistymisen aiheuttajina äärikansallisia organisaatioita, 58 % kannatti Venäjän joukkojen lähettämistä Krimille ja muualle Ukrainaan, ja 43 % uskoi nationalistien uhkaamien Ukrainan venäläisten pelastuvan vain Venäjän armeijan toimilla (Levada 13.3.2014). Kreml oli saanut kotiyleisön tuen. Lisäargumenttien vuoro Krimin johto julisti vallan anastettuaan itsenäisyyttä, mutta päivien edetessä niemimaan poliitikkojen, aktivistien ja asukkaiden keskuudessa vahvistui toive Venäjän federaatioon IDÄNTUTKIMUS 1/2022 15 liittymisestä (Biersack & O’Lear 2014, 250). Krimin parlamentin ja Sevastopolin hallituksen 16.3. järjestämän kansanäänestyksen tulokseksi ilmoitettiin 97 % kannatus liittymiselle (Katchanovski 2015, 86). Viikkojen takaisessa kyselyssä kannatus oli 41 % (Dyczok 2015, 202). Venäjän tiedustelupalvelun vuotojen perusteella vain kolmannes äänesti, kun virallinen luku oli 85 % (Uehling 2015, 70). Kuitenkin Pew Centerin huhtikuun selvityksessä 91 % vastanneista krimiläisistä piti äänestystä vapaana ja oikeudenmukaisena (Katchanovski 2015, 86). Mielipidetutkimuksia tutkineet John O’Loughlin ja Gerard Toal argumentoivat (2019) liittymisen vastanneen krimiläisenemmistön tahtoa. Alueliitos sai Venäjällä 79 % kannatuksen (Levada 13.3.2014). Kaksi päivää äänestyksestä Putin puhui (18.3.2014) Kremlissä seremoniallisesti liittoneuvoston, duuman ja aluejohdon edessä. Vasta tässä vaiheessa otettiin käyttöön monien (esim. Biersack & O’Lear 2014) ensisijaiseksi assosioima oikeutusnarratiivi: Krim Venäjän historiallisena osana. Useiden Krimin venäläisyyttä korostavien väitteidensä lisäksi Putin (18.3.2014) tuomitsi Nikita Hruštšovin teon lahjoittaa Krim Ukrainalle laittomana ja järjettömänä. Siirto, jossa ”krimiläiset myytiin kuin säkki perunoita”, ja jossa ”ihmiset menivät nukkumaan yhdessä valtiossa herätäkseen toisessa”, ei ollut Neuvostoliitossa merkittävä, mutta sen romahdettua teko konkretisoitui auttamattomasti, hän lisäsi. Keinotekoisen separaation päättyminen huipentui Saksojen yhdistymisen analogiaan. Presidentin mukaan Krim oli ollut ja jää venäläiseksi, ukrainalaiseksi ja (äidinkielelleen virallisen aseman saavien) tataarien niemimaaksi. (ibid.) Fakta Katariina Suuren 1700-luvulla tekemästä Krimin valtauksesta tataareilta (Dyczok 2015, 202) loisti poissaolollaan. Vaikka Putin (4.3.2014) oli kaksi viikkoa aiemmin vihjannut krimiläisten oikeuteen kohtalonsa määrittäjinä, hän vahvisti argumenttia kansanäänestystulokseen nojaten (Putin 18.3.2014). Putin tukeutui: 1) numeeriseen todisteluun viittaamalla äänestyslukuihin sekä venäläisten (1,5 milj.) ja venäjänkielisten (350 000) määrään Krimillä ja selitti luvuilla jopa tataareilta saatua kannatusta, 2) Ukrainan juridiset normit romuttaneeseen vallankaappaukseen, mikä salli vastaavat toimet aluetoimijoille, 3) YK:n tuomioistuimen päätelmään, jossa kansainvälinen oikeus sallii itsenäisyysjulistuksen, ja vertasi tilannetta Kosovon irtautumiseen Serbiasta. Kosovo-argumentissa länsi oli luonut ennakkotapauksen itsenäistymisille, joita keskushallinto ei puoltanut. (ibid.) Analogialla tähdättiin lännen kritiikin vaimentamiseen (ks. Rotaru 2019) ja uusinnettiin käsitystä tekopyhästä lännestä. Näin Putin sivuutti kansanäänestyksen luotettavuuden ja Krimin miehityksen. Loppupuhe sisälsi tiivistelmän aiemmasta oikeutusretoriikasta sekä Ukrainan ja Venäjän integraation laajuuden todistelusta historialliseen siteeseen, ystävyyteen sekä talouden ja väestöjen riippuvuussuhteeseen vedoten (ibid.). Lopulta Putin vei oikeutuksensa geopoliittisen kamppailun tasolle. Tavanomaista jyrkempi USA-kritiikki (tuomitsi imperialismin, kansainvälisen oikeuden rikkeet ja petetyn Naton laajentumattomuuslupauksen) johdatti presidentin ydinteesiin: Venäjän rajoittamispolitiikka, jota tehtiin 1700-, 1800- ja 1900-luvuilla, jatkuu myös tänä päivänä. Meitä yritetään jatkuvasti ajaa johonkin nurkkaan siitä, että meillä on riippumaton positio ja siitä, että puolustamme sitä; siitä, että kutsumme asioita omilla nimillään, emmekä ole tekopyhiä. Mutta kaikella on rajansa. Ja Ukrainan tapauksessa meidän läntiset kumppanimme ylittivät linjan, käyttäytyivät törkeästi, vastuuttomasti ja epäammattimaisesti. (Putin 18.3.2014.) 16 IDÄNTUTKIMUS 1/2022 Tuottaessaan todellisuuden rakennetta luovan mallin (Perelman 1996, 122–128) lännen sortopolitiikan jatkuvuudesta, Putin (18.3.2014) liitti Ukrainan uuden hallinnon osaksi lännen tilallista muukalaista, jolta meidän krimiläisemme estivät Naton laajenemisen ja toivat voimaannuttavan geopoliittisen torjuntavoiton. Hän kiitti myös Kiinaa ja Intiaa puolueettomuudesta muodostaen hataran mielikuvan toimia tukevasta koalitiosta. Varsinaista tukea Kreml ei saanut, mutta Brasilia, Intia ja Kiina vetäytyivät Venäjän toimet tuomitsevasta YK:n yleiskokouksen lauselmasta ja tuomitsivat Venäjän ulossulkuehdotuksen G20-yhteisöstä (Stuenkel 2014). Kolme päivää myöhemmin Putin (21.3.2014) allekirjoitti lait, joilla Krim ja Sevastopol liitettiin Venäjän federaation osaksi. Johtopäätökset Tämä tutkimus on antanut oman panoksensa Krimin valtauksen läpivientiä käsittelevään keskusteluun valaisemalla Kremlin rakentaman oikeutusoperaation reaktiivista toteutusmuotoa ja sen suhdetta käytännön valtaustoimiin, oikeutuksessa käytettyjä kielellisiä resursseja sekä niiden kautta tuotetun sisällön geopoliittista luonnetta. Oikeutusoperaatiota tarkasteltiin akuutisti käynnistettynä suostuttelukampanjana, joka rakennettiin kontingentissa kontekstissa aluksi ukrainalaisosapuolten toimiin reagoivalla retoriikalla ja lopuksi Krimin liittämistä tukevalla lisäargumentaatiolla. Se oli kuukauden mittainen kampanja, jonka kuluessa Venäjän valtiojohto tuotti paitsi keskeisimmät anastusperusteet, esitti ne kohdeyleisöjensä kannalta vetovoimaisimpaan aikaan ja suoritti kampanjan varjolla niemimaan valtauksen. Vaikka taustalla olisikin ollut suunnitelma ja kannatuspotentiaalia ”operaatio Krimin” suorittamiseksi, on ilmeistä, että vasta ukrainalaisosapuolten luoma sisäinen konflikti muodosti Kremlin tarvitsemat olosuhteet sekä akuutin mandaatin junailulle että käytännön toimille. Krimin anastusta puoltava geopoliittinen mielikuva rakennettiin neljän diskurssin kautta: juridinen, turvallistava, Venäjän ja Krimin historiallinen ykseys sekä Venäjää alistava tekopyhä länsi, mitä havainnollistan vielä seuraavassa tiivistelmässä. Kreml aloitti suostuttelunsa 20.2.2014 kotiolympialaisten päättyessä ja Ukrainan konfliktin väkivaltaisuuksien kärjistyessä, josta ammensi argumentatiivisen pohjan tulevalle retoriikalleen. Konfliktin päätöstä juridisten normien puitteissa painottaen ja Janukovytšin hallintoa suosien Kreml kuvasi Maidanin mielenosoittajat väkivaltaisina nationalistiradikaaleina, joista vastuussa oli oppositio ja näille tukensa antanut länsi. Muodostui mielikuva pahasta muukalaiskollektiivista, joka saattaisi syrjäyttää valtiollista järjestystä ylläpitävän Janukovytšin hallinnon ja viedä Ukrainan kaaokseen. Tätä seurannut tapahtumaketju koostuen Janukovytšin ja opposition sopimuksen laiminlyönnistä, opposition vallankaappauksesta ja uuden parlamentin asettamasta kielilaista (21.–23.2.) tarjosi ratkaisevan tarttumapinnan retoriikan voimistamiselle, minkä Kreml käytti hyväksi tuottamalla sodan oikeutuksen elementtejä. Venäjän johto representoi tapahtumat ensin juridisella argumentaatiolla, jolla se loi käsityksen Ukrainasta anarkiaan vajonneena maantieteellisenä tilana, jossa ”barbaarit” polkivat ihmisoikeuksia, ja aloitti sitten turvallistavan retoriikan. Koska geopoliittisessa kuvitelmassa järjestystä ylläpitänyt valtiovalta oli hävinnyt Ukrainan maaperältä, tarkoitti se myös alueen väestön oikeuksien takaajan ja fyysisen suojelijan kuvitteellista katoamista. Raakalaiset veivät ukrainanvenäläisiltä ja venäjänkielisiltä oikeuden käyttää omaa äidinkieltään ja tulisivat vielä suorittamaan ”puhdistuksen”. Turvallistamalla maanmiehet alttiiksi ääriryhmien vainolle, Kreml vihjasi kohdeyleisölleen olevansa IDÄNTUTKIMUS 1/2022 17 vastuussa heidän puolustamisestaan. Kun käsitys alueen väestöstä valtiollisena entiteettinä oli romutettu, kollektiivien maantieteelliset rajat piirrettiin kulttuuri-identiteetin perusteella. Samoihin aikoihin, kun Venäjän erikoisjoukot suorittivat miehityksen, valtionjohto, alemman portaan päättäjät ja valtionmedia naamioivat toiminnan krimiläisten itsepuolustusjoukkojen valmistautumisena läntisestä Ukrainasta tulevaan hyökkäykseen. Skenaario sovitettiin virallisiin interventiokriteereihin ja uhkakuvaa konkretisoitiin teatraalisesti Putinille myönnetyllä mandaatilla armeijan käytölle Ukrainassa. Vastakkainasettelua korostettiin Ukrainan uuden hallinnon epäsuoralla identifioinnilla natsidemonina, jonka armeijan muodostivat väkivaltaiset nationalistiradikaalit. Myös meidät, koti- ja krimiläisyleisö, integroitiin yhtenäiseksi kollektiiviksi syväkulttuurisesti ja yhteisen uhan kautta samaistaen ja tuomalla esille krimiläisten (viranomaisten) avunpyyntöjä ja toiveita Venäjän hallinnon piiriin pääsystä. Ominaisuuksiensa ohella vastapuolet kuvattiin yhä selkeämmin maantieteellisinä vastinpareina. Geopoliittisessa mielikuvassa Krim oli meidän etulinjamme, joka oli toistaiseksi säästynyt, mutta tarvitsi kollektiivimme keskuksen, Kremlin varmistusta selviytyäkseen. Miehitys suoritettiin vain kaksi päivää turvallistamisen aloittamisesta. Turvallistava ja juridinen retoriikka jatkui intensiivisenä noin viikon miehityksestä, kunnes Putin mainitsi krimiläisten itsepuolustusjoukkojen poistaneen uhan. Oletettavasti Kreml yritti piilotella toimiaan odottaessaan kannatuksen kasvua. Tässä suhteessa kannatuksen varmistus laahasi sotatoimien perässä. Seuraavaksi Putin keskittyi ”vaurioiden minimoimiseen” ja tuotti kuvitelmaa Venäjän ja Ukrainan korvaamattomasta integraatiosta. Toisaalta turvallistava retoriikka jatkui vielä itäisen Ukrainan osalta viitaten Kremlin ylläpitämään optioon Krimin kaltaisen skenaarion toteuttamisesta. Krimin kansanäänestyksen jälkeen suostuttelukampanjan painopiste siirtyi Krimiä ja Venäjää integroivien käsitysten tuottamiseen historiallisen ykseyden ja juridisen retoriikan kautta. Mielikuvassa keinotekoinen separaatio oli päättynyt juhlalliseen jälleenyhdistymiseen. Krimin liittämishetkellä Putin palautti lännelle ensisijaisen vastustajan aseman ja vei oikeutuksen geopoliittisen kamppailun kontekstiin. Krimin alueliitos oli ikään kuin Kremlin torjuntavoitto; Venäjää systemaattisesti sortavan lännen tämänkertainen ekspansiohanke oli estetty. Vaikka maanmiesten turvallistaminen ja puolustajan roolissa esiintyminen vetosivat kohderyhmiin jo sinänsä, Kremlille oli erityisen tärkeää, että se sai rakennettua kuvitelman, jossa geopoliittinen operaatio näyttäytyi myös juridisesti oikeutetulta. Helmikuun 2022 suurinvaasio Ukrainaan näyttäytyi ensin Itä-Ukrainan valtausprojektin loppuunsaattamiselta, mille Kreml oli hakenut perusteita pakottamalla lännen akuuttiin vastakkainasetteluun ja kehystämällä Naton laajenemismahdollisuudesta eksistentiaalisen uhan ja sitä kautta ajankohtaisen puolustautumisverukkeen. Turvallistaminen kohdistui myös Ukrainan väitettyyn ”natsihallintoon”. Oikeutuksen mittakaava saattoi vielä riittää kannatuksen varmistamiselle Itä-Ukrainan separatistialueiden ”pelastamisen” suhteen, mutta Putinin järjetön, Kremlin omien intressien vastainen täyshyökkäys Ukrainaan teki legitimoinnista mahdottoman. Kun laskelmoiva Putin lähti pelaamaan ”venäläistä rulettia”, hän lähti peliin patruunoita säästelemättä. Ilmassa leijuu vuoden 1905 kaltainen vallankumous. 18 IDÄNTUTKIMUS 1/2022 Lähteet Ahtamo, Valtteri (2020), Venäjä osana maailmanhistorian geopoliittisia prosesseja: operaatio Krim, geopoliitikko Dostojevski ja joustavien geopolityjen maailma. Maisterintutkielma, Jyväskylän yliopisto. https://jyx.jyu.fi/handle/123456789/73212 Ambrosio, Thomas (2016), The rhetoric of irredentism: The Russian Federation’s perception management campaign and the annexation of Crimea. – Small Wars & Insurgencies 27:3, 467–490. https:// doi.org/10.1080/09592318.2016.1151653 Austin, John (1962), How to do things with words. London: Oxford University Press. Ayres, Sabra (28.9.2014), Amid staggering destruction, eastern Ukraine looks to rebuild. Aljazeera. http://america.aljazeera.com/articles/2014/9/28/east-ukraine-reconstruction.html (Tarkistettu 5.11.2021). BBC (22.2.2014), Ukrainian MPs vote to oust President Yanukovych. https://www.bbc.com/news/ world-europe-26304842 (Tarkistettu 5.11.2021). Biersack, John & O’Lear, Shannon (2014), The geopolitics of Russia’s annexation of Crimea: narratives, identity, silences, and energy. – Eurasian Geography and Economics 55:3, 247–269. https:// doi.org/10.1080/15387216.2014.985241 Burke, Kenneth (1969), A rhetoric of motives. Berkeley: University of California Press. Bush, George W. (29.1.2002), First (Official) Presidential State of the Union Address. https://www. americanrhetoric.com/speeches/stateoftheunion2002.htm (Tarkistettu 5.11.2021). Bush, George W. (12.9.2002). Speech to the United Nations General Assembly. https://www.americanrhetoric.com/speeches/gwbushunitednations2002.htm (Tarkistettu 5.11.2021). Bush, George W. (7.10.2002). Address to the Nation on the Threat of Iraq. https://www.americanrhetoric.com/speeches/gwbushwariniraqcincinnati.htm (Tarkistettu 5.11.2021). Bush, George W. (28.1.2003). Second Presidential State of the Union Address. https://www.americanrhetoric.com/speeches/stateoftheunion2003.html (Tarkistettu 5.11.2021). Bush, George W. (26.2.2003). Address on the Future of Iraq. https://www.americanrhetoric.com/ speeches/gwbushfutureofiraq2003.htm (Tarkistettu 5.11.2021). Buzan, Barry, Wæver, Ole & de Wilde, Jaap (1998), Security. A New Framework for Analysis. London: Lynne Rienner Publishers. Carlisle, Rodney (2005), Iraq War. New York: Facts On File. Cohen, Saul (2015), Geopolitics. The Geography of International Relations. Maryland: Rowman & Littlefield. Deutsche Welle (15.3.2019), Vladimir Putin’s ‘Crimea effect’ ebbs away 5 years on. https://www. dw.com/en/vladimir-putins-crimea-effect-ebbs-away-5-years-on/a-47941002 (Tarkistettu 5.11.2021). Dodds, Klaus (2007), Geopolitics: A Very Short Introduction. New York: Oxford University Press. Dyczok, Marta (2015), The Ukraine Story in Western Media. – Ukraine and Russia: People, Politics, Propaganda and Perspectives. Ed. Agnieszka Pikulicka-Wilczewska & Richard Sakwa. Bristol: E-International Relations, 197–205. Galeotti, Mark (2015), Hybrid War’ and ‘Little Green Men’: How It Works, and How It Doesn’t. – Ukraine and Russia: People, Politics, Propaganda and Perspectives. Ed. Agnieszka PikulickaWilczewska & Richard Sakwa. Bristol: E-International Relations, 156–164. Götz, Elias (2015), It’s geopolitics, stupid: explaining Russia’s Ukraine policy. – Global Affairs 1:1, 3–10. https://doi.org/10.1080/23340460.2015.960184 Harle, Vilho & Moisio, Sami (2003), Muuttuva geopolitiikka. Teoksessa Vilho Harle ja Sami Moisio (toim.), Muuttuva geopolitiikka. Helsinki: Gaudeamus, 7–16. Hart, Roderic (2004), Modern rhetorical criticism. Boston: Pearson/Allyn & Bacon Heiskanen, Heikki (6.3.2014), Asiantuntijat: Itäukrainalaiset eivät haikaile osaksi Venäjää. Yle. https:// yle.fi/uutiset/3-7124145 (Tarkistettu 5.11.2021). Hopf, Ted (2016), ‘Crimea Is Ours’: A Discursive History. – International Relations 30:2, 227–255. https://doi.org/10.1177%2F0047117816645646 Jackson, Patrick (2006), Civilizing the Enemy, German Reconstruction and the Invention of the West. Ann Arbor: University of Michigan Press Katchanovski, Ivan (2015), Crimea: People and Territory Before and After Annexation. – Ukraine and Russia: People, Politics, Propaganda and Perspectives. Ed. Agnieszka Pikulicka-Wilczewska & IDÄNTUTKIMUS 1/2022 19 Richard Sakwa. Bristol: E-International Relations, 80–89. Kautilya (1915), Arthashastra. Bangalore: Government Press. Kerry, John (20.2.2014), Situation in Ukraine. https://2009-2017.state.gov/secretary/remarks/2014/02/221855.htm (Tarkistettu 5.11.2021). Kommersant (28.2.2014), Glava MVD Ukrainy Arsen Avakov obvinjajet Rossiju v vooružennom vtorženii. https://www.kommersant.ru/doc/2418581?query=%D1%83%D0%BA%D1%80%D0 %B0%D0%B8%D0%BD%D0%B0 (Tarkistettu 5.11.2021). Korhonen, Pekka (2003), Aasiaksi nimetty tila. – Kosmopolis : rauhan-, konfliktin- ja maailmanpolitiikan tutkimuksen aikakauslehti 33:1, 9–37. https://urn.fi/URN:NBN:fi:ELE-1022628 Kuusisto, Riikka (1998), Framing the Wars in the Gulf and in Bosnia: The Rhetorical Definitions of the Western Power Leaders in Action. – Journal of Peace Research 35:5, 603–620. https://www. jstor.org/stable/425701 Lavrov, Sergei (20.2.2014), Vystuplenije i otvety na voprosy SMI Ministra inostrannyh del Rossii S.V. Lavrova v hode sovmestnoi press-konferentsii po itogam peregovorov s Ministrom inostrannyh del Iraka H.Zibari, g. Bagdad, 20 fevralja 2014 goda. http://www.mid.ru/press_service/ minister_speeches//asset_publisher/7OvQR5KJWVmR/content/id/74246 (Tarkistettu 5.11.2021). Lavrov, Sergei (25.2.2014), Vystuplenije i otvety na voprosy SMI Ministra inostrannyh del Rossii S.V.Lavrova v hode sovmestnoi press-konferentsii po itogam peregovorov s Ministrom inostrannyh del Ljuksemburga Ž.Asselbornom, Moskva, 25 fevralja 2014 goda. http://www.mid.ru/press_service/minister_speeches//asset_publisher/7OvQR5KJWVmR/content/id/73790 (Tarkistettu 5.11.2021). Lavrov, Sergei (3.3.2014), Vystuplenije Ministra inostrannyh del Rossii S.V.Lavrova v hode segmenta vysokogo urovnja 25-i sessii Soveta OON po pravam čeloveka, Ženeva, 3 marta 2014 goda. http://www.mid.ru/press_service/minister_speeches/-/asset_publisher/7OvQR5KJWVmR/content/ id/72642 (Tarkistettu 5.11.2021). Leichtova, Magdalena B. (2016), Why Crimea was Always Ours: Legitimacy Building in Russia in the Wake of the Crisis in Ukraine and the Annexation of Crimea. – Russian Politics 1:3, 291–315. https://doi.org/10.1163/2451-8921-00103004 Lenta (27.2.2014a), Krymtšane obratilis v genkonsulstvo RF za statusom bežentsev. https://lenta.ru/ news/2014/02/27/ask/ (Tarkistettu 5.11.2021). Lenta (27.2.2014b), Spravorosy predložili ukraintsam uskorennoje polutšenije pasportov RF. https:// lenta.ru/news/2014/02/27/citizenship/ (Tarkistettu 5.11.2021). Lenta (28.2.2014), V zahvate parlamenta Kryma zapodozrili rossijski spetšnaz. https://lenta.ru/ news/2014/02/28/blame/ (Tarkistettu 5.11.2021). Levada (26.02.2014), Rossijane o sobytijah v Ukraine (”Maidan”). https://www.levada.ru/2014/02/26/ rossiyane-o-sobytiyah-v-ukraine-majdan/ (Tarkistettu 5.11.2021). Levada (13.3.2014), Situatsija v Ukraine i v Krymu. https://www.levada.ru/2014/03/13/situatsiya-vukraine-i-v-krymu/ (Tarkistettu 5.11.2021). Lewis, Paul, Traynor, Ian & Harding, Luke (28.2.2014), Crimea crisis: pro-Russian leader appeals to Putin for help. The Guardian. https://www.theguardian.com/world/2014/feb/28/barack-obamavladimir-putin-ukraine-russia (Tarkistettu 5.11.2021). Machiavelli, Niccolo (1997), The Prince. Teoksessa S. M. Cahn (Ed.), Classics of Modern Political Theory. Machiavelli to Mill. New York: Oxford University Press. Mankoff, Jeffrey (2014), Russia’s Latest Land Grab: How Putin Won Crimea and Lost Ukraine. – Foreign Affairs 93:3, 60–68. https://www.jstor.org/stable/24483406 Marples, David (2015), Ethnic and Social Composition of Ukraine’s Regions and Voting Patterns. – Ukraine and Russia: People, Politics, Propaganda and Perspectives. Ed. Agnieszka PikulickaWilczewska & Richard Sakwa. Bristol: E-International Relations, 9–18. Mearsheimer, John (2014), Why the Ukraine Crisis Is the West’s Fault: The Liberal Delusions That Provoked Putin. – Foreign Affairs 93:5, 77–84. https://www.jstor.org/stable/24483306 Medvedev, Dmitri (24.2.2014), Dmitri Medvedev otvetil na voprosy žurnalistov. http://government. ru/news/10719/ (Tarkistettu 5.11.2021). Myers, Joe (17.9.2015), Which country has the most engineering graduates? WEF. https://www. weforum.org/agenda/2015/09/which-country-most-engineering-manufacturing-and-constructiongraduates/ (Tarkistettu 5.11.2021). 20 IDÄNTUTKIMUS 1/2022 Nardelli, Alberto, Rankin, Jennifer & Arnett, George (23.7.2015), Vladimir Putin’s approval rating at record levels. The Guardian. https://www.theguardian.com/world/datablog/2015/jul/23/vladimirputins-approval-rating-at-record-levels (Tarkistettu 5.11.2021). Obama, Barack (28.2.2014), Statement by the President on Ukraine. https://obamawhitehouse.archives. gov/the-press-office/2014/02/28/statement-president-ukraine (Tarkistettu 5.11.2021). Office of the Press Secretary (20.2.2014), Statement by the Press Secretary on Ukraine. https:// obamawhitehouse.archives.gov/the-press-office/2014/02/20/statement-press-secretary-ukraine (Tarkistettu 5.11.2021). O’Loughlin, John & Gerard Toal (2019), The Crimea conundrum: legitimacy and public opinion after annexation. – Eurasian Geography and Economics 60:1, 6–27. https://doi.org/10.1080/1538721 6.2019.1593873 Ó Tuathail, Gearóid & Agnew, John (1992), Geopolitics and Discourse: Practical Geopolitical Reasoning in American Foreign Policy. – Political Geography 11:2, 190–204. https://doi. org/10.1016/0962-6298(92)90048-X Pajunen, Ilpo (27.2.2014), Yle Krimillä: Mielenosoittajat vastustavat kielilakia ja vaativat Krimille oikeutta päättää asioistaan. Yle. https://yle.fi/uutiset/3-7112067 (Tarkistettu 5.11.2021). Perelman, Chaïm (1996), Retoriikan valtakunta. Tampere: Vastapaino. Prozorov, Sergei (2010), Ethos without nomos: the Russian–Georgian War and the post-Soviet state of exception. – Ethics & Global Politics 3:4, 255–275. https://doi.org/10.3402/egp.v3i4.5665 Putin, Vladimir (1.3.2014), Vladimir Putin vnjos obraštšenije v Sovet Federatsii. http://kremlin.ru/ events/president/news/20353 (Tarkistettu 5.11.2021). Putin, Vladimir (4.3.2014), Vladimir Putin otvetil na voprosy žurnalistov o situatsii na Ukraine. http:// kremlin.ru/events/president/news/20366 (Tarkistettu 5.11.2021). Putin, Vladimir (18.3.2014), Obraštšenije Prezidenta Rossijskoi Federatsii. http://kremlin.ru/events/ president/news/20603 (Tarkistettu 5.11.2021). Putin, Vladimir (21.3.2014), Podpisany zakony oprinjatii Kryma i Sevastopolja v sostav Rossii. http:// kremlin.ru/events/president/news/20625 (Tarkistettu 5.11.2021). Robertshaw Sam (2015), Why the EU got the Ukrainian crisis wrong. – Global Affairs 1:3, 335–343. https://doi.org/10.1080/23340460.2015.1063249 Rossija24 (25.2.2014a), Sevastopol ne nameren podtšinjatsja vlastjam iz Kieva. https://www.vesti.ru/ doc.html?id=1321740&cid=9# (Tarkistettu 5.11.2021). Rossija24 (25.2.2014b), Posle Majdana Sevastopol perehodit k prjamoi demokratii. https://www.vesti. ru/doc.html?id=1320270&cid=9# (Tarkistettu 5.11.2021). Rossija24 (26.2.2014a), ”Berkut” vzjal pod svoe krylo Krym. https://www.vesti.ru/doc. html?id=1325916&cid=9# (Tarkistettu 5.11.2021). Rossija24 (26.2.2014b), Harkov: gorožane govorjat ”Net” jevromajdanu. https://www.vesti.ru/doc. html?id=1325918&cid=9# (Tarkistettu 5.11.2021). Rossija24 (26.2.2014c), Gosduma predlagaet pomotš Ukraine rossijskim graždanstvom. https://www. vesti.ru/doc.html?id=1325919&cid=7 (Tarkistettu 5.11.2021). Rossija24 (27.2.2014a), Neizvestnye v uniforme zahvatili zdanija parlamenta i pravitelstva Kryma. https://www.vesti.ru/doc.html?id=1327808&cid=9# (Tarkistettu 5.11.2021). Rossija24 (27.2.2014b), Sevastopol prosit pomoštši. https://www.vesti.ru/doc.html?id=1328808&cid=5# (Tarkistettu 5.11.2021). Rossija24 (27.2.2014c), Vozle Rady gruppa radikalov trebuet prinjat zakon o ljustratsii. https://www. vesti.ru/doc.html?id=1328601&cid=9# (Tarkistettu 5.11.2021). Rossija24 (27.2.2014d), Zahvativšie zdanija v Simferopole otkazalis ot peregovorov. https://www. vesti.ru/doc.html?id=1328102&cid=9 (Tarkistettu 5.11.2021). Rossija24 (28.2.2014), Moskva objasnila peredviženie bronetehniki v Krymu. https://www.vesti.ru/ doc.html?id=1332622&cid=9# (Tarkistettu 5.11.2021). Rossija24 (1.3.2014a), Sovet Federatsii jedinoglasno odobril ispolzovanije rossijskoi armii v Krymu. https://www.vesti.ru/doc.html?id=1336215&cid=5# (Tarkistettu 5.11.2021). Rossija24 (1.3.2014b), Peskov: Putin primet rešenie o vvode vojsk na Ukrainu, ishodja iz razvitija situatsii. https://www.vesti.ru/doc.html?id=1336540&cid=5 (Tarkistettu 5.11.2021). Rossija24 (1.3.2014c), Gosduma RF stšitajet ispolzovanije rossijskih vojsk v Krymu svojevremennym. https://www.vesti.ru/doc.html?id=1336168&cid=7# (Tarkistettu 5.11.2021). IDÄNTUTKIMUS 1/2022 21 Rossija24 (1.3.2014d), Turtšinov mobilizujet ukrainskuju armiju. https://www.vesti.ru/doc. html?id=1336805&cid=9 (Tarkistettu 5.11.2021). Rossija24 (1.3.2014e), Gennadii Zjuganov podderživajet ispolzovanije rossijskoi armii dlja normalizatsii obstanovki na Ukraine. https://www.vesti.ru/doc.html?id=1336167&cid=7 (Tarkistettu 5.11.2021). Rotaru, Vasile (2019), Silencing the contestant. Legitimizing Crimea’s annexation by mimicking the West. – European Security 29:1, 96–118. https://doi.org/10.1080/09662839.2019.1688303 Rotaru, Vasile & Troncotă, Miruna (2017), Continuity and change in instrumentalizing ‘The Precedent’. How Russia uses Kosovo to legitimize the annexation of Crimea. – Southeast European and Black Sea Studies 17:3, 325–345. https://doi.org/10.1080/14683857.2017.1348044 Sakwa, Richard (2015a), The death of Europe? Continental fates after Ukraine. – International Affairs 91:3, 553–579. https://doi.org/10.1111/1468-2346.12281 Sakwa, Richard (2015b), Frontline Ukraine, Crisis in the Borderlands. London: I.B.Tauris & Co Ltd. Strycharz, Damian (2020), Dominant Narratives, External Shocks, and the Russian Annexation of Crimea. – Problems of Post-Communism 1–12. https://doi.org/10.1080/10758216.2020.1813594 Stuenkel, Oliver (2014), Why Brazil has not criticized Russia over Crimea. NOREF. https://www.files. ethz.ch/isn/180529/65655a04cd21b64dbcc9c8a823a8e736.pdf (Tarkistettu 5.11.2021). Tsygankov, Andrei (2009), Russophobia, Anti-Russian Lobby and American Foreign Policy. New York: Palgrave Macmillan. Tsygankov, Andrei (2014), The Strong State in Russia, Development and Crisis. New York: Oxford University Press. Uehling, Greta (2015), Everyday Life after Annexation: The Autonomous Republic of Crimea. – Ukraine and Russia: People, Politics, Propaganda and Perspectives. Ed. Agnieszka PikulickaWilczewska & Richard Sakwa. Bristol: E-International Relations, 69–79. Unikaitė-Jakuntavičienė, Ingrida. & Matulkaitė, Rita (2017), The Ukraine Crisis Reflections in the Speeches of USA and Russian Political Leaders. – Slovak Journal of Political Sciences 17:3–4, 299–342. https://www.vdu.lt/cris/entities/publication/bf78e533-7515-4c71-bb3e-d30a9ee9ec78/ details Venäjän presidentin hallinto (15.7.2008), Kontseptsija vnešnei politiki Rossijskoi Federatsii. http:// kremlin.ru/acts/news/785 (Tarkistettu 5.11.2021). Venäjän presidentin hallinto (5.2.2010), Vojennaja doktrina Rossijskoi Federatsii. http://kremlin.ru/ supplement/461 (Tarkistettu 5.11.2021). Wittgenstein, Ludwig (1961), Tractatus Logico-Philosophicus. London: Routledge & Kegan Paul. 22 IDÄNTUTKIMUS 1/2022
https://openalex.org/W3150878503	http://pure-oai.bham.ac.uk/ws/files/121432276/s12884_021_03707_7.pdf	English	null	Vitamin D status in women with dichorionic twin pregnancies and their neonates: a pilot study in China	BMC pregnancy and childbirth	2,021	cc-by	7,606	Vitamin D status in women with dichorionic twin pregnancies and their neonates Li, Xin; Yu, Jiaxiao ; Wen, Li; Li, Qingshu; Yan, Jianying; Tian, Jing; Tong, Chao; Tong, Qi; Qi, Hongbo; Saffery, Richard; Kilby, Mark; Baker, Philip N. DOI: 10.1186/s12884-021-03707-7 License: Creative Commons: Attribution (CC BY) Document Version Publisher's PDF, also known as Version of record Citation for published version (Harvard): Li, X, Yu, J, Wen, L, Li, Q, Yan, J, Tian, J, Tong, C, Tong, Q, Qi, H, Saffery, R, Kilby, M & Baker, PN 2021, 'Vitamin D status in women with dichorionic twin pregnancies and their neonates: a pilot study in China', BMC pregnancy and childbirth, vol. 21, no. 1, 279. https://doi.org/10.1186/s12884-021-03707-7 Document Version Publisher's PDF, also known as Version of record Citation for published version (Harvard): Li, X, Yu, J, Wen, L, Li, Q, Yan, J, Tian, J, Tong, C, Tong, Q, Qi, H, Saffery, R, Kilby, M & Baker, PN 2021, 'Vitamin D status in women with dichorionic twin pregnancies and their neonates: a pilot study in China', BMC pregnancy and childbirth, vol. 21, no. 1, 279. https://doi.org/10.1186/s12884-021-03707-7 Link to publication on Research at Birmingham portal Download date: 24. Oct. 2024 General rights l li General rights Unless a licence is specified above, all rights (including copyright and moral rights) in this document are retained by the authors and/or the copyright holders. The express permission of the copyright holder must be obtained for any use of this material other than for purposes permitted by law. eely distribute the URL that is used to identify this publication. •Users may freely distribute the URL that is used to identify this publication. Users may freely distribute the URL that is used to identify this publication. •Users may download and/or print one copy of the publication from the University of Birmingham research portal for the purpose of private study or non-commercial research. y p oad and/or print one copy of the publication from the University of Birmingham research portal for the purpose of private mercial research study o o co e c a esea c •User may use extracts from the document in line with the concept of ‘fair dealing’ under the Copyright, Designs and P •Users may not further distribute the material nor use it for the purposes of commercial gain. Where a licence is displayed above, please note the terms and conditions of the licence govern your use of this document. When citing, please reference the published version. Take down policy down policy the University of Birmingham exercises care and attention in making items available there are rare occasions when an it ded in error or has been deemed to be commercially or otherwise sensitive. this is the case for this document, please contact UBIRA@lists.bham.ac.uk providing details and we will remove access ely and investigate. If you believe that this is the case for this document, please contact UBIRA@lists.bham.ac.uk providing details and we the work immediately and investigate. Download date: 24. Oct. 2024 Li et al. BMC Pregnancy and Childbirth (2021) 21:279 https://doi.org/10.1186/s12884-021-03707-7 Open Access Abstract Background: Vitamin D deficiency is a global public health issue in women and children and is associated with adverse impacts on child growth, such as rickets. However, prior studies have mainly focused on measuring vitamin D levels in singleton pregnant women and their offspring, and very limited studies have revealed the prevalence of vitamin D deficiency in twin pregnant women and their offspring. The aim of this study was to investigate vitamin D levels in twin-pregnant women and their neonates. We also explored the correlation of maternal vitamin D levels with neonatal outcomes and infant growth. Methods: A prospective subcohort investigation was carried out among 72 dichorionic, diamniotic twin-pregnant mothers and their twin offspring from the Longitudinal Twin Study. Peripheral blood was collected from the mothers in the third trimester, and cord blood was collected from neonates at birth to identify 25[OH]D levels. Data on the characteristics of the mothers and neonates were collected. Infant growth data and food sensitivities were also collected. Results: The average maternal 25[OH]D level was 31.78 ng/mL, with 19.4% being deficient and 20.8% insufficient, while the average neonatal 25[OH]D level was 15.37 ng/mL, with 99.3% being deficiency or insufficient. A positive correlation was found between maternal and neonatal 25[OH]D levels (beta-value: 0.43, 95% CI: 0.37, 0.49). Results: The average maternal 25[OH]D level was 31.78 ng/mL, with 19.4% being deficient and 20.8% insufficient, while the average neonatal 25[OH]D level was 15.37 ng/mL, with 99.3% being deficiency or insufficient. A positive correlation was found between maternal and neonatal 25[OH]D levels (beta-value: 0.43, 95% CI: 0.37, 0.49). Interestingly, the higher the maternal 25[OH]D level was, the smaller the cotwin birthweight discordance (beta- value: -2.67, 95% CI: −5.11, −0.23). In addition, the infants of mothers with vitamin D deficiency were more likely to be allergic to foods at 6 months than those of mothers with vitamin D sufficiency. Interestingly, the higher the maternal 25[OH]D level was, the smaller the cotwin birthweight discordance (beta- value: -2.67, 95% CI: −5.11, −0.23). In addition, the infants of mothers with vitamin D deficiency were more likely to be allergic to foods at 6 months than those of mothers with vitamin D sufficiency. Conclusions: Twin neonates were at high risk of vitamin D deficiency, although their mothers’ vitamin D deficiency partially improved. Higher maternal vitamin D levels were associated with smaller discordance of cotwin birthweight. Vitamin D status in women with dichorionic twin pregnancies and their neonates: a pilot study in China Xin Li1,2,3†, Jiaxiao Yu1,2,3†, Li Wen1,2,3†, Qingshu Li2,3,4, Jianying Yan5, Jing Tian6, Chao Tong1,2,3,7* , Qi Tong7, Hongbo Qi1,2,3,4, Richard Saffery3,8, Mark D. Kilby9,10 and Philip N. Baker3,11 Xin Li1,2,3†, Jiaxiao Yu1,2,3†, Li Wen1,2,3†, Qingshu Li2,3,4, Jianying Yan5, Jing Tian6, Chao Tong1,2,3,7* , Qi Tong7, Hongbo Qi1,2,3,4, Richard Saffery3,8, Mark D. Kilby9,10 and Philip N. Baker3,11 * Correspondence: chaotongcqmu@163.com; 167570342@qq.com; qihongbo728@163.com * Correspondence: chaotongcqmu@163.com; 167570342@qq.com; qihongbo728@163.com g †Xin Li, Jiaxiao Yu and Li Wen contributed equally to this work. 1Department of Obstetrics, The First Affiliated Hospital of Chongqing Medical University, 1 Youyi Road, Yuzhong District, Chongqing 400016, China 7Chongqing Population and Family Planning Science and Technology Research Institute, 18 Honghuang Rd, Jiangbei District, Chongqing 400020, China Full list of author information is available at the end of the article Abstract Trial registration: Chinese Clinical Trial Registry ChiCTR-OOC-16008203, 1st April 2016. Trial registration: Chinese Clinical Trial Registry ChiCTR-OOC-16008203, 1st April 2016. Keywords: Vitamin D, Twin pregnancies, Neonates, Prospective study, Nutrition © The Author(s). 2021 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Li et al. BMC Pregnancy and Childbirth (2021) 21:279 Li et al. BMC Pregnancy and Childbirth (2021) 21:279 Page 2 of 9 Page 2 of 9 birth cohort study conducted in Chongqing, which aims to determine the relative contributions of genes and the environment to early-onset intermediate phenotypes re- lated to later adult onset disease. Chongqing is situated in southwestern China at a latitude of 29.35° N and characterized by a subtropical monsoon humid climate. This city has insufficient sunshine (1000–1400 h/year), especially in winter and spring. The LoTiS twin cohort study was established in January 2016 and aims to un- ravel the complex interplay between genes and the en- vironment in specifying early life determinants of illness in infancy (Chinese Clinical Trial Registration Number: ChiCTR-OOC-16008203) and was approved by the Eth- ics Committee of the First Affiliated Hospital of Chong- qing Medical University (No. 201530). Written informed consent was obtained from all participants. This sub- group study included dichorionic twin pregnant women confirmed prenatally by ultrasound with a prepregnancy BMI in the normal range (18.5 ~ 23.9) as well as daily multivitamin (vitamin D: 500 IU) supplementation from the first trimester to delivery. Due to the high incidence of adverse outcomes in monochorionic twin pregnancies, they were excluded from this study. Therefore, this study investigated vitamin D status in dichorionic diamniotic twin pregnancies. Background Vitamin D is a potent steroid hormone that is not only es- sential for building and maintaining bones but also plays an important role in the immune, endocrine and cardio- vascular systems [1–4]. Although both animal-derived and plant-derived food sources can provide some vitamin D, the main source of vitamin D in the body depends on ex- posure to sunlight [5, 6]. However, the amount of time that people spend in the sun in modern society is not enough to meet their vitamin D needs [7, 8], especially since excessive ultraviolet radiation causes skin problems. In addition, ethnicity, latitude and body mass index also influence vitamin D status [9–11]. Vitamin D deficiency is a global public health problem, especially among pregnant women and newborns [12, 13]. A deficiency of vitamin D in pregnant women increases the risk of gestational diabetes, gestational hypertension disorder and insufficient gestational weight gain and may affect fetal growth and bone ossification [14, 15]. Patients with vitamin D deficiency have been shown to have a higher risk of calcemia and respiratory distress syndrome and a higher likelihood of developing food sensitivities, asthma, type I diabetes or autism later in life [16–20]. Pre- vious studies have revealed that vitamin D status in the fetus and newborn is largely dependent on maternal vita- min D status; thus, the main risk factor for newborn vita- min D deficiency is maternal vitamin D deficiency [21– 23]. However, the data on vitamin D status among twin- pregnant women and their offspring are very limited. In recent decades, the prevalence of twin pregnancies has increased 1.8-fold according to the National Vital Statistics Report of the US due to the development of assisted reproductive technology and advanced maternal age [24–26]. In comparison with those with singleton pregnancy, women who are pregnant with twins are con- sidered to undergo more complex physiological changes and obviously have a higher risk of adverse obstetric consequences [27, 28]. More attention should be paid to twin pregnancies with respect to nutrition and vitamin supplementation. Thus, it is essential to clarify whether vitamin D deficiency in mothers and newborns worsens among twin pregnancies. In addition, previous studies have shown that obese pregnant women may have abnormal vitamin D intake [29, 30], so we only included pregnant women with normal prepregnancy BMI. Mothers with chronic metabolic diseases and those using immunosup- pressants were excluded. Twin pairs with a birth weight < 1500 g, significant malformations, or genetic disorders were also excluded. All twin offspring received daily usage of vitamin D supplements (400 IU/day) from birth and had a pediatric follow-up visit thereafter at a corrected age of 6 months. Data collection l Maternal sociodemographic data (age, height, weight, education, occupation, parity, mode of conception), lifestyle behaviors before pregnancy (smoking and alco- hol use) and pre-existing conditions were collected by standardized questionnaire in the first follow-up during 11–16 gestational weeks. The standardized question- naire was self-designed for the LoTiS cohort study, and detailed information is presented in Supplemental File 1. Information about vitamin D supplementation and other nutrients was collected using a structured questionnaire during the third trimester. Detailed infor- mation on this questionnaire is presented in Supple- mental File 2. Pregnancy complications and maternal and neonatal outcomes, including gestational age, pre- term premature rupture of membranes (PPROM, < 37 gestational weeks), neonatal sex, birthweight and small- for-gestational age (SGA, defined as a weight below the Therefore, in the present study, we aimed to investi- gate the status of 25[OH]D in mothers and their new- borns in a twin pregnancy and birth cohort from Southwest China. Given the previous findings reported for singleton pregnancy, we also aimed to determine the impact of maternal 25[OH]D deficiency on maternal and neonatal outcomes as well as child growth. Statistical methods All statistical analyses were performed with SPSS version 25.0 (IBM, Armonk, NY, USA). Categorical variables are presented as the count and percentage and were ana- lyzed by the chi-squared or Fisher’s exact test. Continu- ous variables are presented as the means and standard deviation and were analyzed by Student’s t-test, LSD Student’s t test, one-way analysis of variance or the non- parametric test. Linear correlation analysis was used to explore the correlation between 25[OH]D levels in mothers and newborns and the correlation of 25[OH]D levels between cotwins. Multivariable linear regression analysis was used to detect the associations between neonatal 25[OH]D levels and related clinical characteris- tics. All tests were two-tailed, and p < 0.05 was consid- ered statistically significant. Serum 25[OH]D is the best estimator to assess body vita- min D levels in the body [33]. For analysis, we divided mothers into 3 groups based on 25[OH]D levels: 25[OH]D levels < 20 ng/mL indicated deficiency, 25[OH]D levels in the range of 20–30 ng/mL indicated insufficiency, and 25[OH]D levels > 30 ng/mL were considered sufficiently high [34]. The same cutoffs were used for neonates according to the Chin- ese standard [35]. Measurement of serum 25[OH]D and classification criteria Measurement of serum 25[OH]D and classification criteria Peripheral blood samples were collected from the mothers in the third trimester, and cord blood samples were imme- diately collected from newborns after placental delivery by using a coagulation-promoting blood collection tube. The specimen was transported to the Maternal and Fetal Med- ical Laboratory under refrigeration, where they were cen- trifuged to obtain serum. The serum 25[OH]D level was measured by high-performance liquid chromatography- electrospray tandem mass spectrometry (HPLC-MS-MS, Waters, USA), which is the gold standard method [32]. The concentrations of 25[OH]D3 and 25[OH]D2 were measured separately, and the total level of 25[OH]D was the sum of 25[OH]D3 and 25[OH]D2. Infant growth assessment and skin prick test Infant growth was monitored for weight and length at the corrected age of 6 months. The corrected age was defined as the infant’s chronological age minus the dif- ference between term birth (40 weeks) and chronological gestational weeks of delivery. Weight and length were simultaneously measured unclothed by a digital measur- ing bed (Beideneng, Shanghai, China) operated by trained nurses at the Department of Child Health Care of Chongqing Health Center for Women and Children. The indexes of weight, height and BMI were converted to Z-scores for sex and age according to WHO Child Growth Standards software (https://www.who.int/ childgrowth/software/en/). On the same day, a skin prick test (SPT) was per- formed to preliminarily diagnose food allergies. The SPT was performed on the volar surface of the forearm with a lancet through the use of a drop of an allergen extract, and test results were produced in 15 to 20 min. The rou- tine test screened for food types including milk, egg white, egg yolk, peanut, fish, wheat, soybean, citrus, and apple. In the maternal vitamin D deficiency group, over half of the mothers conceived with ART and were compli- cated with gestational diabetes, and 78.6% delivered in the winter or spring; this group had the highest inci- dence of SGA, with the lowest birthweight as well as the highest birthweight discordance (Table 2). Results The selection process for this study population is pre- sented in Fig. 1. A total of 190 dichorionic twin pregnan- cies from LoTiS were initially recruited into this subgroup study. After excluding mother-twin offspring pairs that did not meet the inclusion criteria and did not have blood samples, a total of 72 mother-twin offspring pairs were available for the final analysis. The descriptive data of the study participants are shown in Table 1. The average maternal age at recruitment was 30.46 ± 2.93 years, and the average gestational age was 36.77 ± 1.16 weeks. The mean birth weight of all neonates was 2626.53 ± 329.96 g. In addition, 62.5% of mothers con- ceived with the aid of assisted reproductive technology (ART), and 66.7% of mothers delivered in the summer or autumn. The average maternal 25[OH]D level was 31.78 ± 11.1 ng/mL, with 19.5% of mothers being defi- cient and 20.8% being insufficient. Unexpectedly, the average neonatal 25[OH]D level was 15.37 ± 4.86 ng/mL, with 78.5% of neonates being deficient and 20.8% being insufficient. Study design and participants The present study was embedded in the Longitudinal Twin Study (LoTiS), an ongoing twin pregnancy and Li et al. BMC Pregnancy and Childbirth (2021) 21:279 Page 3 of 9 Page 3 of 9 10th percentile for GA and sex [31]), were collected from medical records. 10th percentile for GA and sex [31]), were collected from medical records. 10th percentile for GA and sex [31]), were collected from medical records. (800 L/Hr), Cone (150 L/Hr), Nebulizer (7.0 The calibra- tion curve showed good linearity over the range of 1–80 ng/mL for 25[OH]D2 and 4–200 ng/mL for 25[OH]D3, with <±15% RSD and RE. The intra-assay and inter-assay coefficients of variation were < 15% in all assays, indicating good repeatability. (800 L/Hr), Cone (150 L/Hr), Nebulizer (7.0 The calibra- tion curve showed good linearity over the range of 1–80 ng/mL for 25[OH]D2 and 4–200 ng/mL for 25[OH]D3, with <±15% RSD and RE. The intra-assay and inter-assay coefficients of variation were < 15% in all assays, indicating good repeatability. HPLC-MS-MS The mass spectrometry conditions were an electrospray ionization source with multiple reaction ion scanning modes. Voltages: capillary (3 kV), cone (20 V), source off- set (50 V), temperature (400 °C). GAS Flow: Desolvation A significant difference in neonatal 25[OH]D levels was found among the maternal vitamin D deficiency, in- sufficiency and sufficiency groups (Table 2). There was a Li et al. BMC Pregnancy and Childbirth (2021) 21:279 Li et al. BMC Pregnancy and Childbirth Page 4 of 9 (2021) 21:279 Fig. 1 The selection process for this study directly proportional correlation between maternal 25[OH]D levels and neonatal 25[OH]D levels (r = 0.90, p < 0.001) (Fig. 2a). In addition, a significantly positive correlation was found between cotwins in terms of neo- natal 25[OH]D level (r = 0.91, p < 0.001) (Fig. 2b). Multi- variable linear regression showed that neonatal 25[OH]D levels were positively associated with maternal 25[OH]D levels (beta-value: 0.968, 95% CI: 0.459,0.531. p < 0.001) and birth season (beta-value: -0.102, 95% CI: −1.013, 0.198. p = 0.004) and were not found to have a correlation with maternal age, BMI, gestational weight gain, gestational age, GDM, GHP, ICP or neonatal birth- weight (Table 3). Table 1 Description of the maternal and neonatal characteristics Variables Total Mothers 72 Age (year) 30.46 ± 2.93 Pre-pregnancy BMI (kg/m2) 21.15 ± 1.32 Mode of conception Natural conception 27 (37.5%) Assisted reproductive technology 45 (62.5%) Gestational weight gain (kg) 17.65 ± 5.89 Gestational age (week) 36.77 ± 1.16 Preterm birth 33 (45.8%) Sampling season Summer/autumn 24 (33.3%) Winter/spring 48 (66.7%) 25[OH]D level (ng/mL) 31.78 ± 11.1 Deficiency (< 20 ng/mL) 14 (19.5%) Insufficiency (20-30 ng/mL) 15 (20.8%) Sufficiency (> 30 ng/mL) 43 (59.7%) Infants 144 Gender (male) 70 (48.6%) Birthweight (g) 2626.53 ± 329.96 SGA 5 (3.5%) 25[OH]D level (ng/mL) 15.37 ± 4.86 Deficiency (< 20 ng/mL) 113 (78.5%) Insufficiency (20-30 ng/mL) 30 (20.8%) Sufficiency (> 30 ng/mL) 1 (0.7%) The neonatal birthweight and the discordance in birth- weight between cotwins were significantly different be- tween the maternal 25[OH]D deficiency and sufficiency groups (Fig. 3a-c). After adjusting for maternal age, BMI, GDM, GHP, ICP, gestational weight gain, gestational age and birth season, the results suggested that there was no correlation between maternal 25[OH]D levels and neo- natal birthweight (Table 4). HPLC-MS-MS Surprisingly, maternal 25[OH]D levels were negatively correlated with the dis- cordance in birthweight between cotwins, with the dis- cordance in birthweight between cotwins decreasing by 2.67% when the maternal 25[OH]D level increased by 1 ng/mL (95% CI: −5.11, −0.23. p = 0.032) (Fig. 3c). At 6 months of corrected age, the z-scores of the weight-for-age, height-for-age, weight-for-height, and BMI-for-age indexes showed no differences among the maternal 25[OH]D deficiency, insufficiency and suffi- ciency groups for all infants taking vitamin D supple- ments each day. However, in the maternal vitamin D deficiency group, the incidence of allergies to foods was highest (Table 5). Li et al. BMC Pregnancy and Childbirth (2021) 21:279 Page 5 of 9 Table 2 Description of the maternal and neonatal characteristics by maternal vitamin D status Variables Deficiency Insufficiency Sufficiency p-value Mothers N = 14 N = 15 N = 43 Age (year) 29.86 ± 3.57 29.87 ± 3.54 30.86 ± 3.52 0.394a Pre-pregnancy BMI (kg/m2) 20.76 ± 1.30 20.50 ± 1.26 21.50 ± 1.62 0.054a Mode of conception 0.868b Natural conception 6 (42.9%) 5 (33.3%) 16 (37.2%) ART 8 (57.1%) 10 (66.7%) 27 (62.8%) Pregnancy weight gain (kg) 17.50 ± 4.97 19.43 ± 4.44 17.07 ± 6.55 0.412a Gestational age (week) 36.62 ± 0.97 36.98 ± 1.08 36.73 ± 1.24 0.473a Preterm birth 7 (50.0%) 6 (40.0%) 20 (46.5%) 0.856b Pregnancy-induced illness GDM 8 (57.1%) 6 (40.0%) 12 (27.9%) 0.133b GHD 1 (7.1%) 1 (6.7%) 2 (4.7%) 0.510c ICP 4 (28.6%) 0 (0%) 8 (18.6%) 0.919c Sampling season 0.538b Summer/autumn 3 (21.4%) 6 (40.0%) 15 (34.9%) Winter/spring 11 (78.6%) 9 (60.0%) 28 (65.1%) Infants N = 28 N = 30 N = 86 Gender (male) 12 (42.9%) 19 (63.3%) 39 (45.3%) 0.188b Birthweight (g) 2453 (365) 2740(274) 2641 (347) 0.001d Birthweight discordance (%) 8.83 (1.21) 6.93 (3.87) 5.72 (4.06) 0.048d SGA 2 (7.1%) 1 (3.3%) 2 (2.3%) 0.481c 25[OH]D level (ng/mL) 7.93 (6.20) 12.89 (6.34) 18.75 (5.46) < 0.001d BMI body mass index, ART assisted reproductive technology, GDM gestational diabetes, GHD gestational hypertension disorder, ICP intrahepatic cholestasis of pregnancy, SGA small for gestational age aAverage and standard deviation. One-way Analysis of Variance bNumber (percentage). Chi-squared Test cNumber (percentage). Fisher Exact Test dMedian (interquartile range). Kruskal-Wallis Test Fig. 2 Correlation between maternal and neonatal 25[OH]D levels and correlation of cotwins’ 25[OH]D levels. assisted reproductive technology, GDM gestational diabetes, GHD gestational hypertension disorder, ICP intrahepatic cholestasis of gestational age HPLC-MS-MS a There was a directly proportional correlation between maternal 25[OH]D levels and neonatal 25[OH]D levels (r = 0.90, p < 0.001). b A significantly positive correlation was found between cotwins in terms of neonatal 25[OH]D levels (r = 0.91, p < 0.001) Fig. 2 Correlation between maternal and neonatal 25[OH]D levels and correlation of cotwins’ 25[OH]D levels. a There was a directly proportional correlation between maternal 25[OH]D levels and neonatal 25[OH]D levels (r = 0.90, p < 0.001). b A significantly positive correlation was found between cotwins in terms of neonatal 25[OH]D levels (r = 0.91, p < 0.001) Li et al. BMC Pregnancy and Childbirth (2021) 21:279 Page 6 of 9 Page 6 of 9 Table 3 Association between pregnant covariates and neonatal 25[OH]D levels Variables Beta 95% CI P-value Maternal 25[OH]D level 0.968 (0.459,0.531) < 0.001 Maternal age −0.019 (−0.347,0.209) 0.622 Maternal pre-pregnancy BMI −0.002 (−0.114,0.109) 0.965 Maternal gestational weight gain 0.015 (−0.056,0.084) 0.684 Gestational age at delivery −0.187 (−0.649,0.276) 0.423 Birth season −0.102 (−1.013,-0.198) 0.004 Neonatal birth weight −0.020 (−0.002,0.001) 0.672 GDM −0.019 (−1.042,0.598) 0.591 GHD 0.029 (−1.056,2.461) 0.427 ICP −0.013 (−0.522,0.399) 0.791 Table 3 Association between pregnant covariates and neonatal 25[OH]D levels three times per week, the maternal vitamin D deficiency rates improved to 36.6 and 31.6%, respectively [37]. The maternal vitamin D deficiency rate in this study was lower than that in the aforementioned studies. This can be explained by the fact that all subjects in this study took vitamin D supplements daily starting in the first tri- mester. Thus, we speculated that a high frequency of vitamin D supplementation during pregnancy is an ef- fective way to reduce the risk of maternal vitamin D deficiency. Previous studies have revealed that the rates of mater- nal and neonatal vitamin D deficiency are comparable in singleton pregnancy [36, 38]. A similar statement was also found in the northern Indian twin pregnancy popu- lation, with a maternal vitamin D deficiency rate of 90% and neonatal vitamin D deficiency rate of 89% [13]. However, the results of our study showed that the neo- natal vitamin D deficiency rate was 4-fold higher than the maternal vitamin D deficiency rate, which was in- consistent with the aforementioned study. This is prob- ably attributed to the distinct rates of maternal vitamin D deficiency in the two studies. HPLC-MS-MS The significant differ- ence in neonatal vitamin D deficiency and maternal vita- min D deficiency incidence in our study was likely due to the maternal vitamin D supply to the two fetuses. In addition, we also found that the maternal vitamin D level was an independent factor that correlated with the neonatal vitamin D level, which was consistent with pre- vious studies [39]. Discussion BMC Pregnancy and Childbirth (2021) 21:279 Page 7 of 9 Page 7 of 9 Table 4 Association between pregnant covariates and neonatal birthweight Variables Beta 95% CI P-value Maternal 25[OH]D level 0.392 (−0.129,0.913) 0.112 Maternal age (years) 1.104 (−0.193,2.400) 0.139 Maternal pre-pregnancy BMI −0.189 (−2.107,1.728) 0.095 Maternal gestational weight gain 0.007 (0.001,0.014) 0.050 Gestational age at delivery 1.768 (−2.079,5.615) 0.018 Birth season −2.489 (−4.544,-0.434) 0.169 GDM 1.006 (−7.268,9.279) 0.365 GHD 5.398 (0.010,10.787) 0.810 ICP −0.226 (−0.549,0.097) 0.845 Vitamin D deficiency in neonates has been shown to lead to a higher risk of food sensitivities later in life. In our study, because 99.3% of neonates were vitamin D de- ficient or insufficient, we compared the status of infants allergic to foods by maternal vitamin D levels. Although no significant difference was found among the maternal vitamin D deficiency, insufficiency and sufficiency groups in terms of allergies to foods, the allergy rate was highest in the maternal vitamin D deficiency group. It is worth exploring whether maternal vitamin D deficiency or neonatal vitamin D deficiency predominantly influ- ences infant allergies to foods. The strength of our study is the specialized study population. We used strict inclusion and exclusion cri- teria to screen the participants. Monochorionic twin pregnant women were not selected due to the higher risks of maternal and fetal complications, higher rate of preterm birth and lower neonatal birthweight. Prepreg- nancy BMI also has an obvious impact on perinatal out- comes, so we only recruited women with prepregnancy BMIs in the normal range. Additionally, the food sensi- tivities of infants at 6 months were followed, represent- ing a relatively complete study design. Due to the low natural incidence of twin pregnancy [24], it is time- consuming to obtain participants. Therefore, ART twin pregnancies were also recruited to enlarge the sample size of this study. This also conferred us the opportunity to investigate whether ART has an impact on the 25[OH]D levels of twin pregnancies, and our data clearly revealed that ART does not disturb maternal and neo- natal vitamin D levels in twin pregnancies. when twin pregnant women took vitamin D supple- ments of 500 IU daily. Therefore, further investigations are needed to establish an appropriate dose that is ef- fective for improving neonatal vitamin D deficiency and safe for maternal metabolism. Consistent with previous studies [40, 41], we found that newborns born in winter may have lower 25[OH]D levels. Discussion In this prospective preliminary study, we reported that 19.5% of mothers had vitamin D deficiency, and their neonates had a remarkably high prevalence of vitamin D deficiency at birth, with a rate of 78.5%. We noticed very poor vitamin D stores, especially in twin neonates, even though all mothers took multivitamins (including vita- min D: 500 IU) daily during pregnancy. Several studies conducted on singleton pregnancy among the Chinese population have investigated mater- nal 25[OH]D levels. Maternal vitamin D deficiency was reported in 79.2% of a multiethnic population without an investigation of prenatal vitamin D supplementation [36]. Approximately 10% of mothers took prenatal vita- min D, and the maternal vitamin D deficiency rate was also high at 74.9%. If the mothers took prenatal vitamin D daily during the last month before delivery or over The clinical practice guidelines of the Endocrinology Society recommend that pregnant women take vitamin D supplements of at least 600 IU daily. However, twin neonates were at a very high risk of vitamin D deficiency Fig. 3 The impact of maternal 25[OH]D levels on neonatal birthweight. a The neonatal birthweight was significantly different between the maternal 25[OH]D status (deficiency group vs insufficiency group, P < 0.001; deficiency group vs sufficiency group, P = 0.011); b The neonatal birthweight discordance(%) was significantly different between the maternal 25[OH]D status (deficiency group vs sufficiency group, P = 0.016); c Maternal 25[OH]D levels were negatively correlated with the discordance in birthweight between cotwins, with the discordance in birthweight between cotwins decreased 2.67% when maternal 25[OH]D level increased by 1 ng/mL (95% CI: −5.11, −0.23. p = 0.032) Fig. 3 The impact of maternal 25[OH]D levels on neonatal birthweight. a The neonatal birthweight was significantly different between the maternal 25[OH]D status (deficiency group vs insufficiency group, P < 0.001; deficiency group vs sufficiency group, P = 0.011); b The neonatal birthweight discordance(%) was significantly different between the maternal 25[OH]D status (deficiency group vs sufficiency group, P = 0.016); c Maternal 25[OH]D levels were negatively correlated with the discordance in birthweight between cotwins, with the discordance in birthweight between cotwins decreased 2.67% when maternal 25[OH]D level increased by 1 ng/mL (95% CI: −5.11, −0.23. p = 0.032) Li et al. Discussion Chongqing, also known as China’s fog capital, has few sunny days in winter. The long winter and lack of sunshine in Chongqing may be the cause of neonatal vitamin D deficiency. There are conflicting reports about the association be- tween maternal vitamin D levels and neonatal birth- weight in singleton pregnancy. Hajianfar et al found a significant inverse association between maternal vitamin D level and the rate of low birthweight neonates [42]; others have reported no relationship [43, 44]. In our study, although the neonatal birthweight in the maternal vitamin D deficiency group was lower than that in the other group, no correlation was found between maternal vitamin D level and neonatal birthweight after adjusting for potential confounders. Interestingly, we found that the higher the maternal vitamin D level was, the smaller the discordance in birthweight between cotwins, thus presenting a negative association. However, further in- vestigations are needed to detect the relative molecular mechanism. This preliminary study contributes new knowledge about the status of maternal and neonatal vitamin D levels in twin pregnancies, but several limitations of this study should be taken into consideration. First, compared with that of similar studies in singleton pregnancy, the sample size in this study was relatively small. Second, pregnant women in our hospital were routinely advised to take multivitamin supplements daily based on the clinical guidelines, particularly twin pregnant women; thus, there were no women who did not take multivitamin supple- ments as controls. Third, it is more appropriate to Table 5 Description of the growth and food sensitivity by maternal 25[OH]D status Variables Deficiency (N = 28) Insufficiency (N = 30) Sufficiency (N = 86) p-value WHZ 0.39 ± 0.80 0.47 ± 0.73 0.35 ± 0.82 0.828a WAZ 0.41 ± 0.88 0.59 ± 0.97 0.34 ± 0.85 0.527a HAZ 0.29 ± 1.10 0.57 ± 1.21 0.42 ± 0.95 0.673a BAZ 0.34 ± 0.66 0.38 ± 0.98 0.36 ± 0.79 0.981a Allergic to one or more food 4 (14.3%) 2 (6.7%) 4 (4.7%) 0.200b WHZ z-score for weight-for-height, WAZ z-score for weight-for-age, HAZ z-score for height-for-age, BAZ z-score for BMI-for-age aAverage and standard deviation. Student t test bNumber (percentage). chi-square test Li et al. BMC Pregnancy and Childbirth (2021) 21:279 Li et al. BMC Pregnancy and Childbirth (2021) 21:279 Li et al. Funding g This study was supported by grants from National Key Research and Development Program of China (2018YFC1002900), National Natural Science Foundation of China (81520108013 and 81771613), Chongqing Science and Technology Commission (cstc2017jcyjBX0045), and NHC Key Laboratory of Birth Defects and Reproductive Health (2018–5). The funding agencies did not have any role in the design of the study, collection, analysis, and interpretation of data, and in writing the manuscript. 6. Courbebaisse M, Souberbielle JC, Baptiste A, Taieb J, Tsatsaris V, Guibourdenche J, Senat MV, Haidar H, Jani J, Guizani M, Jouannic JM, Haguet MC, Winer N, Masson D, Elie C, Benachi A. Vitamin D status during pregnancy and in cord blood in a large prospective French cohort. Clin Nutr. 2019;38(5):2136–44. https://doi.org/10.1016/j.clnu.2018.08.035. 7. Kohler M, Leiber F, Willems H, Merbold L, Liesegang A. Influence of altitude on vitamin D and bone metabolism of lactating sheep and goats. J Anim Sci. 2013;91(11):5259–68. https://doi.org/10.2527/jas.2013-6702. Acknowledgments h h ld l 1. Foger-Samwald U, Dovjak P, Azizi-Semrad U, Kerschan-Schindl K, Pietschmann P. Osteoporosis: pathophysiology and therapeutic options. EXCLI J. 2020;19:1017–37. g The authors would like to thank all the participants, Beijing Health Biotech Co. Ltd., health professionals and researchers who contributed to this cohort study. 2. Scheffer-Rath ME, Boot AM. The many facets of vitamin D in the pediatric population. Pediatr Endocrinol Rev. 2020;17(4):293–301. Discussion BMC Pregnancy and Childbirth (2021) 21:279 Page 8 of 9 Page 8 of 9 measure vitamin D levels in women pregnant with twins in the first trimester, prescribe adequate dose supplements and monitor vitamin D levels during pregnancy. Then, the related variables of vitamin D metabolism were limited, such as dietary vitamin D intake and daily solarization not involved. Finally, singleton pregnant women taking the same multivitamin supplements and their neonates should be included as a control group. A large-scale study includ- ing singleton pregnancy and twin pregnancies conducted in multiple centers is essential to better understand the prevalence of maternal and neonatal vitamin D deficiency in China among twin pregnancy populations. Center, where the staff are responsible for the collection, management, and distribution of data. All data are stored electronically in an anonymous format and are currently only available to LoTiS researchers. The datasets used and/or analyzed during the current study are available from the corresponding author upon request (chaotongcqmu@163.com). Abbreviations NC: Natural conception group; IVF: In vitro fertilization; BMI: Body mass index; CI: Confidence interval; GA: Gestational age; GDM: Gestational diabetes mellitus; ICP: Intrahepatic cholestasis of pregnancy; LoTiS: Longitudinal Twin Study; OR: Odds ratio; PE: Preeclampsia; SGA: Small for gestational age; ART: Assisted reproductive technology; GHD: Gestational hypertension disorder; WHZ: Z-score for weight-for-height; WAZ: Z-score for weight-for- age; HAZ: Z-score for height-for-age; BAZ: Z-score for BMI-for-age Author details 1 1Department of Obstetrics, The First Affiliated Hospital of Chongqing Medical University, 1 Youyi Road, Yuzhong District, Chongqing 400016, China. 2State Key Laboratory of Maternal and Fetal Medicine of Chongqing Municipality, Chongqing Medical University, Chongqing 400016, China. 3International Collaborative Laboratory of Reproduction and Development of Chinese Ministry of Education, Chongqing Medical University, Chongqing 400016, China. 4Department of Pathology, The First Affiliated Hospital of Chongqing Medical University, Chongqing 400016, China. 5Fujian Provincial Maternity and Children’s Hospital, affiliated hospital of Fujian Medical University, Fuzhou 350001, Fujian, China. 6Department of Obstetrics and Gynecology, University-Town Hospital of Chongqing Medical University, Chongqing 401331, China. 7Chongqing Population and Family Planning Science and Technology Research Institute, 18 Honghuang Rd, Jiangbei District, Chongqing 400020, China. 8Cancer, Disease and Developmental Epigenetics, Murdoch Children’s Research Institute, Parkville, Victoria 3052, Australia. 9Centre for Women’s and Newborn Health, Institute of Metabolism and Systems Research, University of Birmingham, Birmingham B15 2TT, UK. 10Fetal Medicine Centre, Birmingham Women’s & Children’s Foundation Trust, Edgbaston, Birmingham B15 2TG, UK. 11College of Life Sciences, University of Leicester, Leicester LE1 7RH, UK. Competing interests In summary, this study suggested that despite twin- pregnant women taking prenatal vitamin D supplements and the mothers’ vitamin D deficiency partially improv- ing, their twin neonates were at high risk of vitamin D deficiency. These findings indicated that obstetricians should pay special attention to the dose of vitamin sup- plements provided to the twin pregnancy population. g The authors declare that they have no competing interests. The authors declare that they have no competing interests. Ethics approval and consent to participate The experimental design and procedures were approved by the Ethics Committee of the First Affiliated Hospital of Chongqing Medical University (No. 201530). All participants gave signed informed consent for participation in research. Authors’ contributions 3. Al Mheid I, Quyyumi AA. Vitamin D and cardiovascular disease: controversy unresolved. J Am Coll Cardiol. 2017;70(1):89–100. https://doi.org/10.1016/j.ja cc.2017.05.031. CT and LW designed the research protocol; XL and JY1 conducted the study; XL and JT analyzed the data; XL drafted the manuscript; WL, CT, QL, JY2, RS, MDK, PB critically revised the manuscript; CT, QT, HQ provided funding resources. All authors have accepted responsibility for the entire content of this submitted manuscript and approved submission. 4. Hussain Gilani SY, Bibi S, Siddiqui A, Ali Shah SR, Akram F, Rehman MU. Obesity and diabetes as determinants of vitamin D deficiency. J Ayub Med Coll Abbottabad. 2019;31(3):432–5. 5. Khalil I, Barma P. Subcontinental atmosphere and inherent immune system may have impact on novel Corona Virus’ 2019 (nCovid-19) prevalence in South East Asia. Mymensingh Med J. 2020;29(2):473–80. Supplementary Information Th l l y The online version contains supplementary material available at https://doi. org/10.1186/s12884-021-03707-7. Additional file 1. Received: 3 November 2020 Accepted: 10 March 2021 Received: 3 November 2020 Accepted: 10 March 2021 Additional file 2. Availability of data and materials Goswami D, Rani R, Saxena A, Arora MS, Batra S, Sreenivas V. Maternal and neonatal vitamin-D status in twin versus singleton pregnancies. J Obstet Gynecol Res. 2016;42(10):1250–7. https://doi.org/10.1111/jog.13060. 34. Traglia M, Windham GC, Pearl M, Poon V, Eyles D, Jones KL, Lyall K, Kharrazi M, Croen LA, Weiss LA. Genetic contributions to maternal and neonatal vitamin D levels. Genetics. 2020;214(4):1091–102. https://doi.org/10.1534/ genetics.119.302792. y 14. Yoon HK. Gestational diabetes mellitus, fetal growth and vitamin D. J Bone Metab. 2017;24(3):155–9. https://doi.org/10.11005/jbm.2017.24.3.155. 15. Vijay A, Sinha S, Sindgikar SP, Shenoy VD. Effect of gestational diabetes on the vitamin D levels in the neonates: a case control study. Turk J Pediatr. 2020;62(3):431–5. https://doi.org/10.24953/turkjped.2020.03.010. 35. Wei F, Wang Z, Wang J, Xu H, Zhou H. Serum vitamin D levels among children aged 0-12 years in the first affiliated Hospital of Harbin Medical University, China. J Public Health (Oxf). 2018;40(4):721–6. https://doi.org/10.1 093/pubmed/fdy055. 16. Curtis EM, Moon RJ, Harvey NC, Cooper C. Maternal vitamin D supplementation during pregnancy. Br Med Bull. 2018;126(1):57–77. https:// doi.org/10.1093/bmb/ldy010. 36. Wang Y, Li H, Zheng M, Wu Y, Zeng T, Fu J, Zeng D. Maternal vitamin D deficiency increases the risk of adverse neonatal outcomes in the Chinese population: a prospective cohort study. PLoS One. 2018;13(4):e0195700. https://doi.org/10.1371/journal.pone.0195700. 17. Perezabad L, Lopez-Abente J, Alonso-Lebrero E, Seoane E, Pion M, Correa- Rocha R. The establishment of cow’s milk protein allergy in infants is related with a deficit of regulatory T cells (Treg) and vitamin D. Pediatr Res. 2017; 81(5):722–30. https://doi.org/10.1038/pr.2017.12. 37. Zhu T, Liu TJ, Ge X, Kong J, Zhang LJ, Zhao Q. High prevalence of maternal vitamin D deficiency in preterm births in northeast China, Shenyang. Int J Clin Exp Pathol. 2015;8(2):1459–65. 18. Kechichian E, Ezzedine K. Vitamin D and the skin: an update for dermatologists. Am J Clin Dermatol. 2018;19(2):223–35. https://doi.org/10.1 007/s40257-017-0323-8. 38. Windham GC, Pearl M, Anderson MC, Poon V, Eyles D, Jones KL, Lyall K, Kharrazi M, Croen LA. Newborn vitamin D levels in relation to autism spectrum disorders and intellectual disability: a case-control study in California. Autism Res. 2019;12(6):989–98. https://doi.org/10.1002/aur.2092. 19. Sablok A, Batra A, Thariani K, Batra A, Bharti R, Aggarwal AR, Kabi BC, Chellani H. Supplementation of vitamin D in pregnancy and its correlation with feto-maternal outcome. Clin Endocrinol. 2015;83(4):536–41. https://doi. org/10.1111/cen.12751. 39. Ariyawatkul K, Lersbuasin P. Prevalence of vitamin D deficiency in cord blood of newborns and the association with maternal vitamin D status. Availability of data and materials 8. Dinlen N, Zenciroglu A, Beken S, Dursun A, Dilli D, Okumus N. Association of vitamin D deficiency with acute lower respiratory tract infections in The LoTiS is being conducted mainly at the First Affiliated Hospital of Chongqing Medical University and Chongqing Women and Children’s Health Page 9 of 9 Page 9 of 9 Page 9 of 9 Li et al. BMC Pregnancy and Childbirth (2021) 21:279 Page 9 of 9 30. Karlsson T, Andersson L, Hussain A, Bosaeus M, Jansson N, Osmancevic A, Hulthén L, Holmäng A, Larsson I. Lower vitamin D status in obese compared with normal-weight women despite higher vitamin D intake in early pregnancy. Clin Nutr. 2015;34(5):892–8. https://doi.org/10.1016/j.clnu.2 014.09.012. newborns. J Matern Fetal Neonatal Med. 2016;29(6):928–32. https://doi.org/1 0.3109/14767058.2015.1023710. newborns. J Matern Fetal Neonatal Med. 2016;29(6):928–32. https://doi.org/1 0.3109/14767058.2015.1023710. newborns. J Matern Fetal Neonatal Med. 2016;29(6):928–32. https://doi.org/1 0.3109/14767058.2015.1023710. 9. Cashman KD, Ritz C, Adebayo FA, Dowling KG, Itkonen ST, Ohman T, et al. Differences in the dietary requirement for vitamin D among Caucasian and east African women at northern latitude. Eur J Nutr. 2019;58(6):2281–91. https://doi.org/10.1007/s00394-018-1775-1. 31. Matsuda N, Taki A, Tsuji A, Nakajima K, Takasawa K, Morioka C, Minosaki Y, Oku K, Kashimada K, Morio T. Perinatal factors affecting growth and development at age 3 years in extremely low birth weight infants born small for gestational age. Clin Pediatr Endocrinol. 2018;27(1):31–8. https:// doi.org/10.1297/cpe.27.31. 10. Petrenya N, Lamberg-Allardt C, Melhus M, Broderstad AR, Brustad M. Vitamin D status in a multiethnic population of northern Norway: the SAMINOR 2 clinical survey. Public Health Nutr. 2020;23(7):1186–200. https://doi.org/10.1 017/S1368980018003816. 32. Mathew EM, Moorkoth S, Rane PD, Lewis L, Rao P. Cost-effective HPLC-UV method for quantification of vitamin D2 and D3 in dried blood spot: a potential adjunct to newborn screening for prophylaxis of intractable pediatric seizures. Chem Pharm Bull (Tokyo). 2019;67(2):88–95. https://doi. org/10.1248/cpb.c18-00542. 11. Elsori DH, Hammoud MS. Vitamin D deficiency in mothers, neonates and children. J Steroid Biochem Mol Biol. 2018;175:195–9. 11. Elsori DH, Hammoud MS. Vitamin D deficiency in mothers, neonates and children. J Steroid Biochem Mol Biol. 2018;175:195–9. 12. Esmeraldo CUP, Martins MEP, Maia ER, Leite JLA, Ramos JLS, Goncalves J Jr, et al. Vitamin D in term newborns: relation with maternal concentrations and birth weight. Ann Nutr Metab. 2019;75(1):39–46. https://doi.org/10.11 59/000502044. 33. Kovacs CS. Vitamin D in pregnancy and lactation: maternal, fetal, and neonatal outcomes from human and animal studies. Am J Clin Nutr. 2008; 88(2):520S–8S. https://doi.org/10.1093/ajcn/88.2.520S. 13. Availability of data and materials 1999;47(18):1–96. Availability of data and materials Eur J Pediatr. 2018;177(10):1541–5. https://doi.org/10.1007/s00431-018-3210-2. 20. Sotunde OF, Laliberte A, Weiler HA. Maternal risk factors and newborn infant vitamin D status: a scoping literature review. Nutr Res. 2019;63:1–20. 40. Al Zarooni AAR, Al Marzouqi FI, Al Darmaki SH, Prinsloo EAM, Nagelkerke N. Prevalence of vitamin D deficiency and associated comorbidities among Abu Dhabi Emirates population. BMC Res Notes. 2019;12(1):503. https://doi. org/10.1186/s13104-019-4536-1. 21. Fernando S, Kumar S, Bakr M, Speicher D, Lea R, Scuffham PA, Johnson NW. Children's untreated decay is positively associated with past caries experience and with current salivary loads of mutans streptococci; negatively with self-reported maternal iron supplements during pregnancy: a multifactorial analysis. J Public Health Dent. 2019;79(2):109–15. https://doi. org/10.1111/jphd.12301. 41. Fan H, Hui L, Yan X, Hou W, Bai E, Wang L, Yu X. Serum 25 hydroxyvitamin D levels and affecting factors among preconception fertile women. BMC Womens Health. 2020;20(1):146. https://doi.org/10.1186/s12905-020-01018-1. 22. Saraf R, Morton SM, Camargo CA Jr, Grant CC. Global summary of maternal and newborn vitamin D status - a systematic review. Matern Child Nutr. 2016;12(4):647–68. https://doi.org/10.1111/mcn.12210. 22. Saraf R, Morton SM, Camargo CA Jr, Grant CC. Global summary of maternal and newborn vitamin D status - a systematic review. Matern Child Nutr. 2016;12(4):647–68. https://doi.org/10.1111/mcn.12210. 42. Yu X, Wang W, Wei Z, Ouyang F, Huang L, Wang X, Zhao Y, Zhang H, Zhang J. Vitamin D status and related factors in newborns in Shanghai, China. Nutrients. 2014;6(12):5600–10. https://doi.org/10.3390/nu6125600. 23. Stagi S, Rigante D. Vitamin D and juvenile systemic lupus erythematosus: lights, shadows and still unresolved issues. Autoimmun Rev. 2018;17(3):290– 300. https://doi.org/10.1016/j.autrev.2018.01.004. 23. Stagi S, Rigante D. Vitamin D and juvenile systemic lupus erythematosus: lights, shadows and still unresolved issues. Autoimmun Rev. 2018;17(3):290– 300. https://doi.org/10.1016/j.autrev.2018.01.004. 43. Shor DB, Barzel J, Tauber E, Amital H. The effects of maternal vitamin D on neonatal growth parameters. Eur J Pediatr. 2015;174(9):1169–74. https://doi. org/10.1007/s00431-015-2517-5. 24. Martin JA, Hamilton BE, Osterman MJK, Driscoll AK. Births: final data for 2018. Natl Vital Stat Rep. 2019;68(13):1–47. 44. Yun C, Chen J, He Y, Mao D, Wang R, Zhang Y, Yang C, Piao J, Yang X. Vitamin D deficiency prevalence and risk factors among pregnant Chinese women. Public Health Nutr. 2017;20(10):1746–54. https://doi.org/10.1017/ S1368980015002980. 25. Martin JA, Hamilton BE, Ventura SJ, Osterman MJ, Wilson EC, Mathews TJ. Births: final data for 2010. Natl Vital Stat Rep. 2012;61(1):1–72. 26. Ventura SJ, Martin JA, Curtin SC, Mathews TJ. Births: final data for 1997. Natl Vital Stat Rep. Publisher’s Note 27. Chang YL, Wang TH, Abufraijeh SM, Chang SD, Chao AS, Hsieh PCC. Preliminary report of altered insulin secretion pattern in monochorionic twin pregnancies complicated with selective intrauterine growth restriction. Taiwan J Obstet Gynecol. 2017;56(1):51–4. https://doi.org/10.1016/j.tjog.201 5.11.004. Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. 28. Mihaela B, Radu V, Anca B, Alexandra M, Ioana C, Simona V. Monochorionic vs dichorionic twins: kanet test vs postnatal neurodevelopment. Maedica (Buchar). 2020;15(1):61–70. 29. Mousa A, Abell SK, Shorakae S, Harrison CL, Naderpoor N, Hiam D, et al. Relationship between vitamin D and gestational diabetes in overweight or obese pregnant women may be mediated by adiponectin. Mol Nutr Food Res. 2017;61(11).
https://openalex.org/W3105198944	https://link.springer.com/content/pdf/10.1007/s11295-021-01506-x.pdf	English	null	Genetic diversity analysis of Chinese plum (<i>Prunus salicina</i> L.) based on whole-genome resequencing	bioRxiv (Cold Spring Harbor Laboratory)	2,020	cc-by	7,421	Keywords Prunus salicina Lindl. . Whole-genome resequencing . Population structure . Genetic introgression Keywords Prunus salicina Lindl. . Whole-genome resequencing . Population structure . Genetic introgression https://doi.org/10.1007/s11295-021-01506-x Tree Genetics & Genomes (2021) 17: 26 https://doi.org/10.1007/s11295-021-01506-x Tree Genetics & Genomes (2021) 17: 26 ORIGINAL ARTICLE Xiao Wei1 & Fei Shen2 & Qiuping Zhang1 & Ning Liu1 & Yuping Zhang1 & Ming Xu1 & Shuo Liu1 & Yujun Zhang1 & Xiaoxue Ma1 Weisheng Liu1 Xiao Wei1 & Fei Shen2 & Qiuping Zhang1 & Ning Liu1 & Yuping Zhang1 & Ming Xu1 & Shuo Li Xiaoxue Ma1 & Weisheng Liu1 Received: 16 November 2020 /Revised: 22 March 2021 /Accepted: 25 March 2021 # The Author(s) 2021 / Published online: 21 April 2021 Abstract Chinese plum (Prunus salicina L.), also known as Japanese plum, is gaining importance because of its extensive genetic diversity and nutritional attributes that are beneficial for human health. Single-nucleotide polymorphisms (SNPs) are the most abundant form of genomic polymorphisms and are widely used in population genetics research. In this study, we constructed high-quality SNPs through whole-genome resequencing of 67 Prunus accessions with a depth of ~20× to evaluate the genome-level diversity and population structure. Phylogenetic analysis, principal component analysis, and population structure profiling indicated that the 67 plum accessions could be classified into four groups corresponding to their origin location, the southern cultivar group (SCG), the northern cultivar group (NCG), the foreign cultivar group (FG), and the mixed cultivar group (MG). Some cultivars from South China clustered with the other three groups. The genetic diversity indices including private allele number, observed heterozygosity, expected heterozygosity, and the nucleotide diversity of the SCG were higher than those of the NCG. Gene flow from the SCG to FG was also detected. Based on the distribution of wild resources, we concluded that the domestication center of origin of the Chinese plum was southwestern China. This study also provided genetic variation features and the population structure of Chinese plum cultivars, laying a foundation for breeders to use diverse germplasm and allelic variants to improve Chinese plum varieties. 1 Liaoning Institute of Pomology, Yingkou 115009, China 2 Beijing Agro-biotechnology Research Center, Beijing Academy of Agriculture and Forestry Sciences, Beijing 100193, China Plant materials Genetic variability is a prerequisite for any plant breeding program. Learning the extent and structure of genetic variation in germplasm collections is a crucial step for the efficient conservation and utilization of biodiversity in cultivated crops. For plum breeders, using diverse plum resources to broaden the genetic base of worldwide plum cultivars is a critical objective (Liu et al. 2007; Urrestarazu et al. 2018). Previous studies have conducted the genetic diversity of plums; for example, Zhang and Zhou (1998) took the lead in collecting plum germplasm resources and evaluating their ge- netic diversity based on morphological traits and isozyme polymorphisms. However, because morphological traits were highly susceptible to environmental factors, the estimates of genetic diversity were not precise. The isozymes had a low degree of polymorphism and hence were not efficient enough for the characterization of germplasm genetic diversity (Khush 2002). Current research applies DNA-based markers in plum genetic diversity analysis: markers include random amplified polymorphic DNA (RAPD) (Liu et al. 2006; Ben Tamarzizt et al. 2015); simple sequence repeats (SSRs) (Pop et al. 2018; Zhang et al. 2018; Acuña et al. 2019); and inter- simple sequence repeats (ISSRs) (Liu et al. 2007; Wu et al. 2019). DNA-based marker assessments showed that Chinese plums can generally be classified into two major groups: the southern cultivar group (SCG) and the northern cultivar group (NCG) (Liu 2005; Wei et al. 2019). A palynology study showed that the P. salicina system in South China was more primitive than that in North China, with the spread of P. salicina taking place from the south to the north (Guo 2006). A cultivar clustering approach, however, is inadequate for the integrated study of genetic diversity of the Chinese plum. Advances in next-generation sequencing (NGS) tech- nologies, involving genotyping by sequencing (GBS), provide a great wealth of information that makes it possible to identify thousands of single nucleotide polymorphisms (SNPs). This information, after adequate filtering, allows to carry out de- tailed genetic diversity studies (Elshire et al. 2011; Salazar et al. 2019; Zhebentyayeva et al. 2019). Whole-genome resequencing can be used to obtain more SNP datasets for Prunus species. In the case of the peach, large-scale SNP data-based diversity analysis has boosted the deciphering of the evolution and domestication of peach germplasm re- sources (Cao et al. 2014; Li et al. 2019). A diverse collection of 67 Prunus spp. accessions, including 65 P. Plant materials salicina accessions and two P. simonii accessions, was selected for the whole-genome resequencing study (Table S1). These trees were accessed at the NGRPA located in Xiongyue county, Liaoning Province, China (40° 18′ N, 122° 16′ E) at a planting density of 3.0 × 4.0 m, trained to the open vase system with three or four main branches, and maintained un- der conventional management and pest control operations. DNA extraction, library preparation, and sequencing Genomic DNA was isolated from young, healthy leaf samples of 67 accessions using a modified cetyl trimethylammonium bromide (CTAB) protocol (Doyle and Doyle 1990). The qual- ity and integrity of the DNA were examined using a NanoDrop® spectrophotometer (ND-1000, Thermo Fisher Scientific Inc., USA), followed by electrophoresis in 1% aga- rose gels. Quantification of the DNA samples was performed using Qubit™(2.0 Fluorometer, Invitrogen, Carlsbad, CA, USA). High-molecular-weight DNA aliquots with 230/260 and 260/280 ratios ranging between 1.8–2.0 and 1.8–2.2, respec- tively, were then sent to BGI (Shenzhen, China) for library construction and sequencing. The insert size of the libraries was 500 bp, and the length of the pair-end reads was 150 bp. All libraries were sequenced using the Illumina HiSeq 2500 platform (Illumina, San Diego, CA, USA). Materials and methods to 47.09%, suggesting that the genetic variability of Chinese plums is widely distributed. Introduction L., 2n = 2x = 16), also known as the Japanese plum, is widely grown for fresh market consumption and the canning industry, and includes both the pure Chinese plum and its hybrids with other diploid plum species, such as Prunus simonii Carr., Prunus cerasifera Ehrh., and Prunus americana Marsh. (Liu et al. 2007). Plum is one of the most important fruit crops globally and, possesses extensive genetic diversity and high economic value (Topp et al. 2012). The largest plum producer is China, with an annual production of 6,801,187 metric tons in 2018, ac- counting for 53.9% of the world’s total (FAOSTAT 2018). Within the genus Prunus, the Chinese plum (Prunus salicina According to historical records of cultivation, the Chinese plum may have originated in the Yangtze River Basin, and there are abundant plum germplasm resources in China (Hartmann and Neumulle 2009). The Chinese plum has a long growing history and extensive geographical distribution, with more than 1000 indigenous plum cultivars in China, derived from Prunus salicina L. (Zhang 1990). Over 700 of these cultivars are currently preserved at the National Germplasm Repository for Plums and Apricots (NGRPA) located in Xiongyue, Liaoning Province, China (Wei et al. 2020). Yu et al. (2011) carried out comprehensive phenotyping of 405 Chinese plum cultivars and their hybrids from the NGRPA, and investigated a total of 32 morphological and agronomic characters. Coefficients of variation (CVs) ranged from 14.85 Xiao Wei and Fei Shen contributed equally to this work. Communicated by W.-W. Guo * Qiuping Zhang lbzhangqiuping2@163.com * Weisheng Liu wsliulaas@163.com 1 Liaoning Institute of Pomology, Yingkou 115009, China 2 Beijing Agro-biotechnology Research Center, Beijing Academy of Agriculture and Forestry Sciences, Beijing 100193, China Xiao Wei and Fei Shen contributed equally to this work. Communicated by W.-W. Guo Xiao Wei and Fei Shen contributed equally to this work. Communicated by W.-W. Guo * Qiuping Zhang lbzhangqiuping2@163.com * Weisheng Liu wsliulaas@163.com 1 Liaoning Institute of Pomology, Yingkou 115009, China 2 Beijing Agro-biotechnology Research Center, Beijing Academy of Agriculture and Forestry Sciences, Beijing 100193, China * Qiuping Zhang lbzhangqiuping2@163.com * Weisheng Liu wsliulaas@163.com 26 Page 2 of 10 Tree Genetics & Genomes (2021) 17: 26 Read mapping, SNP calling, and SNP annotation The qualified paired-end reads of each accession were aligned against the peach reference genome v2.0 (Verde et al. 2013) using BWA v0.7.12-r1039 with the parameters mem -t 4 -k 32 –M (Li and Durbin 2010), and SNPs were identified using SAMtools v1.4 (Li et al. 2009). Low-quality SNPs were fil- tered out by minimum minor allele frequency (mnMAF < 0.01) and missing data per site (MDpS > 10%), and finally converted into a variant call format file (VCF). Gene-based SNP annotation was performed using the ANNOVAR pack- age v2018-04-16 (Wang et al. 2010). Based on the reference genome annotation, SNPs were categorized as occurring in exonic regions (overlapping with a coding exon); intronic re- gions (overlapping with an intron); upstream and downstream regions (within a 1 kb region upstream or downstream from the transcription start site); or intergenic regions. Additionally, heterozygosity was calculated using VCFtools v0.1.14 (Danecek et al. 2011). The present study was the first in which we used SNP markers generated from high depth whole-genome resequencing data (an average depth of ~20×) to elucidate the pattern of genetic diversity, population structure, and do- mestication of a diverse P. salicina collection. An in-depth understanding of such genetic relationships could benefit plum germplasm conservation and utilization and lead to new cultivar improvement. Tree Genetics & Genomes (2021) 17: 26 Page 3 of 10 26 Population structure analysis higher than 20×; a total 462.6 Gb of clean data was retained with an average of 6.9 Gb for each accession after filtering out low-quality reads. The quality of the sequencing data was high with a Q30 > 85%, and the GC content fluctuating slightly around 38.0%. The mapping rate to the peach genome ranged from 85.73 to 88.57%, with an average of 87.49% (Table S1). This result indicated a high level of collinearity between the genomes of Chinese plum and peach. TreeBeST v1.9.2 software was used to calculate the distance matrix (Vilella et al. 2009). RAxML v8.2.12 was used to construct the maximum likelihood (ML) phylogenetic tree, and 1000 bootstrap replicates were used (Stamatakis 2014). The resulting phylogenetic tree was visualized using MEGA v5.05 (Tamura et al. 2011). We called SNPs with unique mapped reads, using SAMtools v1.4 software; a total of 16,600,033 SNPs and 2,107,664 indels were identified across the 67 accessions. After filtering out those of low quality, we obtained 14,549,234 SNPs that would ensure the accuracy and reliability of subsequent genetic diversity and population structure analyses (Fig. 1). Approximately 43.81% of the SNPs were located in intergenic regions, and 10.52% were in coding regions. The non-synonymous to synonymous substitu- tion ratio (dN/dS) for the SNPs in the coding regions was 1.32. Transitions were found in 59.58% (10,315,764/16,600,033), with a transition/transversion ratio (Ts/Tv ratio) of 1.47 (Fig. 1, Table 1). The principal component analysis was performed using PLINK v1.07 software with default parameters (Purcell et al. 2007). We used ADMIXTURE v1.23 to infer population structure (Alexander et al. 2009). To identify the best genetic clusters K, cross-validation error was tested for each K value from 2 to 10. The termination criterion was 1e−6 (stopping when the log- likelihood increased by less than 1e−6 between iterations). Gene flow analysis The TreeMix software v1.13 was used to evaluate the gene flow among different groups with the parameters -se- bootstrap-k 1000 -m, where the number (-m) varied from one to three (Pickrell and Pritchard 2012). Phylogenetic and population structure To explore the relationships among the Chinese plum acces- sions, a neighbor-joining phylogenetic tree of the 67 acces- sions was constructed using all SNPs (Fig. 2a). The phyloge- netic tree classified the accessions into four main groups, which corresponded to their respective origin locations: (1) the southern cultivar group (SCG), comprising plum cultivars from Sichuan, Guizhou, Yunnan, Guangdong, Guangxi, Zhejiang, and Fujian Provinces; (2) the northern cultivar group (NCG), comprising plum cultivars mainly from Hebei, Henan, Shandong, Shaanxi Province, and the southern part of Liaoning Province; (3) the foreign cultivar group (FG), including plum cultivars from the USA and Japan; and (4) the mixed cultivar group (MG), comprising newly bred cultivars from the NGRPA and several cultivars originated SCG and Linkage disequilibrium (LD) analysis Linkage disequilibrium (LD) was calculated using SNPs with MAF greater than 0.05, using PLINK v1.07 software with the following settings: –file –r2 –ld-window 99999 –ld-window- kb 200 –out. The LD decay was calculated based on the squared correlation coefficient (r2) values between the two SNPs and the physical distance between the two SNPs (Purcell et al. 2007). Genetic diversity and differentiation analysis To evaluate the genetic diversity and differentiation, we used a 100-kb sliding window with a step size of 10 kb to calculate the number of private alleles (AP); observed heterozygosity (Ho); expected heterozygosity (He); nucleotide diversity (π); Tajima’s D value; Wright’s F-statistic (FIS); and population differentiation statistics (fixation index, FST) using VCFtools v0.1.14 (Danecek et al. 2011). Table 1 Statistics of variations across different chromosomes Categories Numbers Ratio Raw SNPs 16,600,033 88.73% Indels 2,107,664 11.27% Filtered SNPs 14,549,234 87.65% Intergenic regions SNPs 7,272,007 43.81% Coding regions SNPs 1,745,725 10.52% Transition 10,315,764 59.58% Transversion 6,999,126 40.42% Transition/transversion -- 147.00% Non-synonymous SNPs 993,073 6.83% Synonymous SNPs 752,652 5.17% Table 1 Statistics of variations across different chromosomes Marker development based on whole-genome resequencing The 67 plum accessions representing different geographic and morphological characteristics (Table S1) were sequenced using the Illumina HiSeq 2500 platform with a sequence depth 26 Page 4 of 10 Tree Genetics & Genomes (2021) 17: 26 ratio). c The counts of different types of transitions and transversions. d The counts of genome-wide insertions and deletions Fig. 1 The statistics of markers generated from the whole-genome resequencing of the 67 plum accessions. a The chromosome-scale SNP distribution. b The distribution of transition/transversion ratio (Ts/Tv ratio). c The counts of different types of transitions and transversions. d The counts of genome-wide insertions and deletions ratio). c The counts of different types of transitions and transversions. d The counts of genome-wide insertions and deletions Fig. 1 The statistics of markers generated from the whole-genome resequencing of the 67 plum accessions. a The chromosome-scale SNP distribution. b The distribution of transition/transversion ratio (Ts/Tv FG. Notably, we found that “Saozouli” and “Shuili” cultivars, collected from South China, were clustered with the NCG; “Zaohuangli” and “Jinshali” with the FG; and “Wanshu huanai,” “Huahongli,” and “Abazhou meiguili” with the MG. These SCG samples showed some degree of correlation, suggesting a close relationship between the SCG and others groups. The results showed that the other groups potentially originated from the SCG. FG. Notably, we found that “Saozouli” and “Shuili” cultivars, collected from South China, were clustered with the NCG; “Zaohuangli” and “Jinshali” with the FG; and “Wanshu huanai,” “Huahongli,” and “Abazhou meiguili” with the MG. These SCG samples showed some degree of correlation, suggesting a close relationship between the SCG and others groups. The results showed that the other groups potentially originated from the SCG. in the SCG, which was consistent with the π value analysis (Fig. 3b). For the FG, we found that the population structure was relatively independent, compared with the SCG and others. Thus, we postulated that the existence of intra- species hybrids in the FG group led to a high π value (Fig. 3b). To further confirm the relationship among the cultivars, we performed principal component analysis (PCA) of the 67 plum accessions (Fig. 2c). As shown in the principal component plot for the first two principal components, the NCG exhibited a relatively close relationship with the SCG, which was consistent with the phylogenetic tree and population structure analysis. Marker development based on whole-genome resequencing As shown in Table 1, a large number of rare alleles were lost in the NCG population. Therefore, we deduced that the NCG could be considered to be a more independent subgroup of the SCG and might have been derived from a specific ecotype. A population genetic structure analysis was performed, based on high-quality SNPs. We employed 5-fold cross-vali- dation to infer the number of ancestral populations, K (Fig. 2b and Figure S1). When the K value was 2–4, the NCG exhib- ited a consistent genetic constitution with the SCG, suggesting that the NCG was derived from the SCG and, to some extent, had also undergone environmental/human selection. The pop- ulation structure of the SCG was complex, especially when the K value was higher, suggesting a higher genetic diversity Page 5 of 10 26 Tree Genetics & Genomes (2021) 17: 26 Fig. 2 The population structure of the 67 plum accessions. a Neighboring-joining phylogenetic tree constructed using SNPs at four- fold degenerate site. Each group was color coded. b Bayesian model- based clustering of the 67 plum accessions with the number of ancestry kinship (K) from 2 to 10. c Principle Component Analysis (PCA) of the 67 plum accessions. FG, the foreign cultivar group; MG, the mixed cul- tivar group; SCG, the southern cultivar group; NCG, the northern cultivar group g ns. a four- model- cestry kinship (K) from 2 to 10. c Principle Component Analysis (PCA) of the 67 plum accessions. FG, the foreign cultivar group; MG, the mixed cul- tivar group; SCG, the southern cultivar group; NCG, the northern cultivar group kinship (K) from 2 to 10. c Principle Component Analysis (PCA) of the 67 plum accessions. FG, the foreign cultivar group; MG, the mixed cul- tivar group; SCG, the southern cultivar group; NCG, the northern cultivar group Fig. 2 The population structure of the 67 plum accessions. a Neighboring-joining phylogenetic tree constructed using SNPs at four- fold degenerate site. Each group was color coded. b Bayesian model- based clustering of the 67 plum accessions with the number of ancestry Genetic diversity, differentiation, and inferred evolutionary path The genus Prunus has shown conserved intraspecific and intragenic collinearity in the Rosaceae fam- ily, with the peach being considered a model species for the genus Prunus for multiple types of genetic research (Arús et al. 2012; Carrasco et al. 2018; Marti et al. 2018). Zhang et al. (2020) found that the correlation of eight chromosomes was greater than 0.8, indicating a high level of collinearity between the Japanese plum and peach genomes. A recently released plum genome has been published, and collinearity analysis has shown that the assembled genomes of the plum also exhibit a high level of genome synteny with the peach (Liu et al. 2020). For the first time, we obtained an average of 6.9 Gb of high-quality data for each acquisition based on whole-genome resequencing, where the mapping rate of 67 plum accessions with the peach genome sequence was above 85%. This result indicates that the peach genome sequence is highly effective as the reference genome of Chinese plum, and ecological groups were close to each other, we found no gene flow between them, indicating a relatively independent do- mestication process. Table 2 The statistical values of genetic diversity within different populations Genetic diversity, differentiation, and inferred evolutionary path The genus Prunus has shown conserved intraspecific and intragenic collinearity in the Rosaceae fam- ily, with the peach being considered a model species for the genus Prunus for multiple types of genetic research (Arús et al. 2012; Carrasco et al. 2018; Marti et al. 2018). Zhang et al. (2020) found that the correlation of eight chromosomes was greater than 0.8, indicating a high level of collinearity between the Japanese plum and peach genomes. A recently released plum genome has been published, and collinearity analysis has shown that the assembled genomes of the plum also exhibit a high level of genome synteny with the peach (Liu et al. 2020). For the first time, we obtained an average of 6.9 Gb of high-quality data for each acquisition based on whole-genome resequencing, where the mapping rate of 67 plum accessions with the peach genome sequence was above 85%. This result indicates that the peach genome sequence is highly effective as the reference genome of Chinese plum and Tree Genetics & Genomes (2021) 17: 26 Fig. 3 The genetic diversity of different groups. a The decay of linkage disequilibrium (LD) measured as the squared correla- tion coefficient (r2) by pairwise physical distance. b The nucleo- tide diversity (π) of different groups. c The genetic differentia- tion analysis between groups. The values between pairs indicate population divergence (FST). FG, the foreign cultivar group; MG, the mixed cultivar group; SCG, the southern cultivar group; NCG, the northern cultivar group Fig. 3 The genetic diversity of different groups. a The decay of linkage disequilibrium (LD) measured as the squared correla- tion coefficient (r2) by pairwise physical distance. b The nucleo- tide diversity (π) of different groups. c The genetic differentia- tion analysis between groups. The values between pairs indicate population divergence (FST). FG, the foreign cultivar group; MG, the mixed cultivar group; SCG, the southern cultivar group; NCG, the northern cultivar group Fig. 3 The genetic diversity of different groups. a The decay of linkage disequilibrium (LD) measured as the squared correla- tion coefficient (r2) by pairwise physical distance. b The nucleo- tide diversity (π) of different groups. c The genetic differentia- tion analysis between groups. The values between pairs indicate population divergence (FST). FG, the foreign cultivar group; MG, the mixed cultivar group; SCG, the southern cultivar group; NCG, the northern cultivar group Pinosio et al. 2020). Genetic diversity, differentiation, and inferred evolutionary path indicated that genetic recombination in the FG was difficult because of the narrow genetic origin of the artificial hybrid cultivars in the group. As shown in Table 2, the expected heterozygosity (He) of the Prunus populations varied between 0.234 and 0.304; the ob- served heterozygosity (Ho) of the Prunus populations ranged between 0.328 and 0.429; Wright’s F-statistic (FIS) of the Prunus populations varied between −0.241 and −0.146; the number of private alleles (AP) in the Prunus populations var- ied between 546 and 31,285; and the nucleotide diversity (π) ranged between 0.00358 and 0.00467. The mean nucleotide variation of P. salicina was higher than that of other perennial crops, such as peach (π = 0.0015) (Verde et al. 2013), cassava (π = 0.0026) (Kawuki et al. 2009), and apricot (π = 0.0027) (Li et al. 2020), but was lower than that of date palms (π = 0.0092) (Hazzouri et al. 2015). The Tajima’s D values of the four groups all tested positive (1.002–1.497) and were signif- icantly different from zero. Thus, the null hypothesis of neu- tral evolution was rejected. As shown in Fig. 3a, the LD decay rate was fastest in the SCG and slowest in the FG, which The FST values of the four groups varied between 0.0485 and 0.1303 (Fig. 3). The FST values of the MG-FG, MG-SCG, and SCG-NCG pairs were relatively lower among all the group pairs analyzed, indicating that the genetic differences within populations were higher than those between popula- tions and that there was possible genetic exchange between populations. The FST value between the NCG and FG groups was the highest of all the groups analyzed, possibly because of geographical isolation, low gene flow between populations, and significant genetic differences. We analyzed the gene flow between the four geographic groups (Fig. 4b, c). Allowing one or two migration events (m = 1 or 2), we observed that gene flow occurred between the SCG and FG accessions, a likely reflection of their many shared genomic components due to hybridization in their do- mestication and breeding histories. Considering the geograph- ical locations of the SCG and NCG, although these two Tree Genetics & Genomes (2021) 17: 26 26 Page 6 of 10 nd no gene endent do- ch in plant , display a arkers, and asco et al. large num- tively low- rsity analy- association dvances in Pinosio et al. 2020). Discussion Chinese plums cultivat- ed in southwest China were tested with high genetic diversity using SSR markers (Wei et al. 2019): the high Shannon index of diversity, the high value of average sufficient allele, and the high expected heterozygosity, combined with the fact that there were wild plum populations discovered in Yunnan, Sichuan, and Guizhou Provinces, suggested high genetic diversity (Wei et al. 2020). In the present study, the ge- netic diversity of the SCG was higher than that of the NCG, while some cultivars from the southwest region were distributed in the other three groups. We therefore suggest that southwestern China was a primitive domesti- cation center of the Chinese plum; the genetic diversity of the SCG was lower than that of the FG and MG, which could be explained by the presence of artificially hybrid- ized cultivars in the FG and MG. Fig. 4 The inferred evolutionary path (a) and gene flow analysis between populations as inferred by TreeMix using a model with one (b) and two (c) admixture events. Admixtures are colored according to their weight. FG, the foreign cultivar group; MG, the mixed cultivar group; SCG, the southern cultivar group; NCG, the northern cultivar group g. 4 The inferred evolutionary ath (a) and gene flow analysis etween populations as inferred y TreeMix using a model with ne (b) and two (c) admixture vents. Admixtures are colored ccording to their weight. FG, the oreign cultivar group; MG, the mixed cultivar group; SCG, the outhern cultivar group; NCG, the orthern cultivar group The inferred evolutionary and gene flow analysis n populations as inferred Mix using a model with and two (c) admixture Admixtures are colored ng to their weight. FG, the cultivar group; MG, the ultivar group; SCG, the n cultivar group; NCG, the n cultivar group Fig. 4 The inferred evolutionary path (a) and gene flow analysis between populations as inferred by TreeMix using a model with one (b) and two (c) admixture events. Admixtures are colored according to their weight. FG, the foreign cultivar group; MG, the mixed cultivar group; SCG, the southern cultivar group; NCG, the northern cultivar group It was originally speculated from historical records that the Chinese plum originated in the Yangtze River basin in China (Hartmann and Neumulle 2009). A palynological study showed that the P. salicina in South China is more primitive than that in North China (Guo 2006). Discussion SNPs have become essential as markers for research in plant genetics because they occur in high frequencies, display a lower mutation rate compared to SSR-based markers, and are uniformly distributed across genomes (Carrasco et al. 2018). The NGS technologies allow discovery of large num- bers of SNPs for extensive genetic studies at relatively low- cast. The technologies include within-species diversity analy- sis; linkage map construction; and genome-wide association studies (GWAS), which have led to significant advances in plant genetics and breeding (Cao et al. 2016; Li et al. 2019; Groups Private allele number (Ap) Observed heterozygosity (Ho) Expected heterozygosity (He) Nucleotide diversity (π) Wright’s F- statistic (FIS) Tajima’s D FG 31,285 0.4290 0.3036 0.00467 −0.2411 1.4974 MG 30,112 0.3658 0.2846 0.00433 −0.1560 1.0094 SCG 27,287 0.3365 0.2606 0.00395 −0.1460 1.0022 NCG 546 0.3280 0.2336 0.00358 −0.1829 1.3211 Page 7 of 10 26 Tree Genetics & Genomes (2021) 17: 26 that the SNPs that were used ensure accurate and comprehen- sive population genetic analysis. Based on the high-quality SNPs, the 67 plum accessions in this study were divided into four groups, the SCG, NCG, FG, and MG. The classification into these four groups was sup- ported by the phylogenetic tree, population structure analysis, and PCA results, which partly aligned with the findings of a previous study which used RAPD, ISSR, and SSR markers (Liu et al. 2006; Liu 2005; Wei et al. 2019). We noticed that several cultivars from the southern part of China, present in the SCG, were distributed in the other three groups (Fig. 2a). We speculated that, for the SCG, the genetic background was relatively broader. Additionally, the fruit weight of the south- ern cultivars exhibited the highest genetic variation a previous analysis of the phenotypic variation of 405 plum accessions (Yu et al. 2011). The π, Ho, and He of the SCG, calculated with SNPs in our study, were higher than those of the NCG, corresponding to the higher genetic diversity of the SCG. This finding showed that gene exchange in the SCG was more frequent than in the other cultivar groups. It was originally speculated from historical records that the Chinese plum originated in the Yangtze River basin in China (Hartmann and Neumulle 2009). A palynological study showed that the P. salicina in South China is more primitive than that in North China (Guo 2006). Discussion Chinese plums cultivat- ed in southwest China were tested with high genetic diversity using SSR markers (Wei et al. 2019): the high Shannon index of diversity, the high value of average sufficient allele, and the high expected heterozygosity, combined with the fact that there were wild plum populations discovered in Yunnan, Sichuan, and Guizhou Provinces, suggested high genetic diversity (Wei et al. 2020). In the present study, the ge- netic diversity of the SCG was higher than that of the NCG, while some cultivars from the southwest region were distributed in the other three groups. We therefore suggest that southwestern China was a primitive domesti- cation center of the Chinese plum; the genetic diversity of the SCG was lower than that of the FG and MG, which could be explained by the presence of artificially hybrid- ized cultivars in the FG and MG. that the SNPs that were used ensure accurate and comprehen- sive population genetic analysis. Based on the high-quality SNPs, the 67 plum accessions in this study were divided into four groups, the SCG, NCG, FG, and MG. The classification into these four groups was sup- ported by the phylogenetic tree, population structure analysis, and PCA results, which partly aligned with the findings of a previous study which used RAPD, ISSR, and SSR markers (Liu et al. 2006; Liu 2005; Wei et al. 2019). We noticed that several cultivars from the southern part of China, present in the SCG, were distributed in the other three groups (Fig. 2a). We speculated that, for the SCG, the genetic background was relatively broader. Additionally, the fruit weight of the south- ern cultivars exhibited the highest genetic variation a previous analysis of the phenotypic variation of 405 plum accessions (Yu et al. 2011). The π, Ho, and He of the SCG, calculated with SNPs in our study, were higher than those of the NCG, corresponding to the higher genetic diversity of the SCG. This finding showed that gene exchange in the SCG was more frequent than in the other cultivar groups. 26 Page 8 of 10 26 Page 8 of 10 Tree Genetics & Genomes (2021) 17: 26 The clustering results in Fig. Conflict of interest The authors declare no competing interests. Conflict of interest The authors declare no competing interests. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adap- tation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, pro- vide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. Genetic differentiation between populations is considered moderate when the FST value is higher than 0.05, and highly differentiated when the FST value is higher than 0.15 (Xu 2009). In our study, the NCG-FG pair possessed the highest FST value, followed by the SCG-FG pair. In contrast, the FST value of the MG-FG was significantly lower, as there were five newly selected plum cultivars by the NGRPA in the MG with foreign lineage from the FG. Human intervention, artifi- cially selecting favorable phenotypic traits to enhance produc- tion and improve desirable agronomic traits, can reduce the levels of genetic variability and skew allele frequencies (Tajima 1989). Discussion 2a showed that, for the five newly bred cultivars from the NGRPA in the MG, one of their parents was from the USA (“Blackamber,” “Friar”) or Japan (“Akihime”) and the other parent was native to China, and also that gene flow occurred between the SCG and FG (Fig. 4). Thus, we assumed that the MG cultivars were genetically related. As mentioned in other studies, including a dendrogram analysis (Liu et al. 2006; Liu et al. 2007; Wei et al. 2019), most of the Chinese plumcultivarsfromJapanandtheimprovedChineseplum hybrids from the USA were distributed across the Chinese indigenous plum group. The P. salicina-predominant genetic components of improved Chinese plum hybrids (Faust and Suranyi 1999) the view that plum cultivars, originating in China, were initially intro- duced to Japan, and then exported from Japan to the USA in 1870 by Luther Burbank (Hartmann and Neumulle 2009). References Acuña CV, Rivas JG, Brambilla SM, Cerrillo T, Frusso EA, García MN, Villalba PV, Aguirre NC, Sabio y García JV, Martínez MC, Hopp EH, Marcucci Poltri SN (2019) Characterization of genetic diversity in accessions of Prunus salicina Lindl: keeping fruit flesh color ideotype while adapting to water stressed environments. Agronomy 9:487 Alexander DH, Novembre J, Lange K (2009) Fast model-based estimation of ancestry in unrelated individuals. Genome Res 19:1655–1664 Arús P, Verde I, Sosinski B, Zhebentyayeva T, Abbott AG (2012) The peach genome. Tree Genet Genomes 8:531–547 The present study is the first in which we constructed high- quality SNPs from 67 accessions in the Chinese plum by whole- genome resequencing, with the peach genome as the reference. Our findings show that the domestication center of origin of the Chinese plum was in the southwest region of China. This study, in providing genetic variation features of plum cultivars and an analysis of their genetic diversity and population structure, lays a foundation for breeders to use diverse germplasm and allelic variants to improve Chinese plum varieties. Ben Tamarzizt H, Ben Mustapha S, Baraket G, Abdallah D, Salhi- Hannachi A (2015) Assessment of genetic diversity and relation- ships among wild and cultivated Tunisian plums (Prunus spp.) using random amplified microsatellite polymorphism markers. Genet Mol Res 14:1942–1956 Cao K, Zheng Z, Wang L, Liu X, Zhu G, Fang W, Cheng S, Zeng P, Chen C, Wang X, Xie M, Zhong X, Wang X, Zhao P, Bian C, Zhu Y, Zhang J, Ma G, Chen C, Li Y, Hao F, Li Y, Huang G, Li Y, Li H, Guo J, Xu X, Wang J (2014) Comparative population genomics reveals the domestication history of the peach, Prunus persica, and human influences on perennial fruit crops. Genome Biol 15:1–15 Cao K, Zhou Z, Wang Q, Guo J, Zhao Pei, Zhu Gengrui, Fang Weichao, Chen Changwen, Wang Xinwei, Wang Xiaoli, Tian Zhixi, Wang Lirong (2016) Genome-wide association study of 12 agronomic traits in peach. Nat Commun 7. https://doi.org/10.1038/ ncomms13246 Supplementary Information The online version contains supplementary material available at https://doi.org/10.1007/s11295-021-01506-x. Supplementary Information The online version contains supplementary material available at https://doi.org/10.1007/s11295-021-01506-x. 94(3):284–294 Hartmann W, Neumulle M (2009) Plum breeding. In: Jain SM, Priyadarshan PM (eds) Breeding Plantation Tree Crops: Temperate Species. Springer Science Business Press, Media, pp 161–231 Lindl.). J. Hortic. Sci. Biotechnol. 94(3):284–294 Stamatakis A (2014) RAxML version 8: a tool for phylogenetic analysis and post-analysis of large phylogenies. Bioinformatics 30:1312–1313 Hazzouri KM, Flowers JM, Visser HJ, Khierallah HSM, Rosas U, Pham GM, Meyer RS, Johansen CK, Fresquez ZA, Masmoudi K, Haider N, El Kadri N, Idaghdour Y, Malek JA, Thirkhill D, Markhand GS, Krueger RR, Zaid A, Purugganan MD (2015) Whole genome resequencing of date palms yields insights into diversification of a fruit tree crop. Nat Commun 6:8824 Tajima F (1989) Statistical method for testing the neutral mutation hy- pothesis by DNA polymorphism. Genetics 123:585–595 Tamura K, Peterson D, Peterson N, Stecher G, Nei M, Kumar S (2011) MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Mol Biol Evol 28:2731–2739 Kawuki RS, Ferguson M, Labuschagne M, Herselman L, Kim D (2009) Identification, characterisation and application of single nucleotide polymorphisms for diversity assessment in cassava (Manihot esculenta Crantz). Mol Breed 23:669–684 Topp BL, Russell DM, Neumuller M, Dalbo M, Liu WS (2012) Plum. In: Badenes ML, Byrne DH (eds) Fruit Breeding. Springer Science Business, Media, pp 571–621 Khush G (2002) Molecular genetics-plant breeder’s perspective. In: Jain SM, Brar DS, Ahloowalia BS (eds) Molecular techniques in crop improvement. Kluwer Academic Press, Dordrecht, pp 1–8 Urrestarazu J, Errea P, Miranda C, Santesteban LG, Pina A, Pang XM (2018) Genetic diversity of Spanish Prunus domestica L. germplasm reveals a complex genetic structure underlying. PLoS One 13:1–21 Li H, Durbin R (2010) Fast and accurate long-read alignment with Burrows-Wheeler transform. Supplementary Information The online version contains supplementary material available at https://doi.org/10.1007/s11295-021-01506-x. Bioinformatics 26:589–595 Verde I, Abbott AG, Scalabrin S, Jung S, Shu SQ, Marroni F, Zhebentyayeva T, Dettori MT, Grimwood J, Cattonaro F, Zuccolo A, Rossini L, Jenkins J, Vendramin E, Meisel LA, Decroocq V, Sosinski B, Prochnik S, Mitros T, Policriti A, Cipriani G, Dondini L, Ficklin S, Goodstein DM, Xuan P, Fabbro CD, Aramini V, Copetti D, Gonzalez S, Horner DS, Falchi R, Lucas S, Mica E, Maldonado J, Lazzari B, Bielenberg D, Pirona R, Miculan M, Barakat A, Testolin R, Stella A, Tartarini S, Tonutti P, Arús P, Orellana A, Wells C, Main D, Vizzotto G, Silva H, Salamini F, Schmutz J, Morgante M, Rokhsar DS (2013) The high-quality draft genome of peach (Prunus persica) identifies unique patterns of genetic diversity, domestication and genome evolu- tion. Nat Genet 45(5):487–494 Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, Marth G, Abecasis G, Durbin R (2009) The sequence alignment/map format and SAMtools. Bioinformatics 25(16):2078–2079 Li W, Liu L, Wang Y, Zhang Q, Fan G, Zhang S, Wang Y, Liao K (2020) Genetic diversity, population structure, and relationships of apricot (Prunus) based on restriction site-associated DNA sequencing. Hortic Res 7:69 Bielenberg D, Pirona R, Miculan M, Barakat A, Testolin R, Stella A, Li Y, Cao K, Zhu G, Fang W, Chen C, Wang X, Zhao P, Guo J, Ding T, Guan L, Zhang Q, Guo W, Fei Z, Wang L (2019) Genomic analyses of an extensive collection of wild and cultivated accessions provide new insights into peach breeding history. Genome Biol 20:36 Vilella AJ, Severin J, Ureta-Vidal A, Heng L, Durbin R, Birney E (2009) EnsemblCompara GeneTrees: complete, duplication-aware phylo- genetic trees in vertebrates. Genome Res 19:327–335 Liu C, Feng C, Peng W, Hao J, Wang J, Pan J, He Y (2020) Chromosome-level draft genome of a diploid plum (Prunus salicina). GigaScience 9(12):130 Wang K, Li M, Hakonarson H (2010) ANNOVAR: Functional annota- tion of genetic variants from high-throughput sequencing data. Nucleic Acids Res 38(16):e164 Liu W (2005) Studies on the genetic diversity among plum germplasm resources and the phylogenetic relationships of main plum species. [Doctor’s thesis]. [Beijing]: China Agricultural University Wei X, Zhang Q, Liu N, Zhang Y, Liu S, Zhang Y, Ma X, Liu W (2019) Genetic diversity of the Prunus salicina L. from different Sources and their related species. Sci Agric Sin 52(3):568–578. https://doi. Supplementary Information The online version contains supplementary material available at https://doi.org/10.1007/s11295-021-01506-x. Data archiving statement All of the raw reads of the plum acces- sions generated in this study have been deposited in the public database of National Center of Biotechnology Information under PRJNA659814. Carrasco B, González M, Gebauer M, García-González R, Maldonado J, Silva H (2018) Construction of a highly saturated linkage map in Japanese plum (Prunus salicina L.) using GBS for SNP marker calling. PLoS One 13:1–14 Danecek P, Auton A, Abecasis G, Albers CA, Banks E, DePristo MA, Handsaker RE, Lunter G, Marth GT, Sherry ST, Vean GM, Durbin R (2011) The variant call format and VCFtools. Bioinformatics 27: 2156–2158 Author contribution ZQP and LWS conceived and designed the experi- ments. WX and SF performed the experiments. WX, SF and ZQP ana- lyzed the data and wrote the paper. LN, ZYP, XM, LS, MXX, and ZYJ collected samples and managed the materials. LWS revised the manu- script. All authors have read and approved the final manuscript. Doyle JJ, Doyle JL (1990) Isolation of plant DNA from fresh tissue. Focus 12:13–15 Elshire RJ, Glaubitz JC, Sun Q, Poland JA, Ken K, Buckler ES, Mitchell SE, Orban L (2011) A robust, simple genotyping-by-sequencing (GBS) approach for high diversity species. PLoS One 6:1–10 Funding This work was supported by National Natural Science Foundation of China (31401826), the Program of Conservation and Utilization of Crop Germplasm Resources (2018NWB003) and National Crop Germplasm Resources Platform of China (NICGR2018- 056). Faust M, Suranyi D (1999) Origin and dissemination of plum. Hortic Rev 23:179–231 Page 9 of 10 Tree Genetics & Genomes (2021) 17: 26 26 Food and Agriculture Organization of the United Nations (2018) World production of plums and sloes. http://www.fao.org/faostat/zh/#data/ QC. Accessed 23 April 2020 Food and Agriculture Organization of the United Nations (2018) World production of plums and sloes. http://www.fao.org/faostat/zh/#data/ QC. Accessed 23 April 2020 Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MAR, Bender D, Maller J, Sklar P, de Bakker PIW, Daly MJ, Sham PC (2007) PLINK: a tool set for whole-genome association and population- based linkage analyses. Am J Hum Genet 81:559–575 Guo Z (2006) Collection and utilization of Prunus salicina germplasm resource in South China. [Doctor’s thesis]. [Nanjing]: Nanjing Agricultural University Salazar JA, Pacheco I, Silva C, Zapata P, Shinya P, Ruiz D, Martínez- Gómez P, Infante R (2019) Development and applicability of GBS approach for genomic studies in Japanese plum (Prunus salicina Lindl.). J. Hortic. Sci. Biotechnol. Zhang Q, Wei X, Liu N, Zhang Y, Xu M, Zhang Y, Ma X, Liu W (2020) Construction of an SNP-based high-density genetic map for Japanese plum in a Chinese population using specific length frag- ment sequencing. Tree Genet Genomes 16:18 Zhebentyayeva T, Shankar V, Scorza R, Callahan A, Ravelonandro M, Castro S, DeJong T, Saski CA, Dardick C (2019) Genetic Supplementary Information The online version contains supplementary material available at https://doi.org/10.1007/s11295-021-01506-x. org/10.3864/j.issn.0578-1752 Liu W, Liu D, Feng C, Zhang A, Li SH (2006) Genetic diversity and phylogenetic relationships in plum germplasm resources revealed by RAPD markers. J Hortic Sci Biotechnol 81:242–250 Liu W, Liu D, Zhang A, Feng C, Yang J, Yoon J, Li SH (2007) Genetic diversity and phylogenetic relationships among plum germplasm resources in China assessed with inter-simple sequence repeat markers. J Am Soc Hortic Sci 132(5):619–628 Wei X, Zhang Q, Liu W (2020) Research progress on plum germplasm resources in China. Acta Hortic Sin 47(6):1203–1212 Wu W, Chen F, Yeh K, Chen J (2019) ISSR analysis of genetic diversity and structure of plum varieties cultivated in Southern China. Biology (Basel) 8:1–13 Marti AFI, Saski CA, Manganaris GA, Gasic K, Crisosto CH (2018) Genomic sequencing of Japanese plum (Prunus salicina Lindl.) mutants provides a new model for rosaceae fruit ripening studies. Front. Plant Sci 9:21 Xu G (2009) Plant population genetics. Science Press, Beijing Yu X, Zhang Q, Liu W, Sun M, Liu N, Zhang Y, Xu M (2011) Genetic diversity analysis of morphological and agronomic characters of Chinese plum (Prunus salicina Lindl.) germplasm. J. Plant. Genet. Resour. 12(3):402–440 Pickrell JK, Pritchard JK (2012) Inference of population splits and mix- tures from genome-wide allele frequency data. PLoS Genet 8(11): e1002967 Pinosio S, Marroni F, Zuccolo A, Vitulo N, Mariette S, Sonnante G, Aravanopoulos FA, Ganopoulos I, Palasciano M, Vidotto M, Magris G, Iezzoni A, Vendramin GG, Morgante M (2020) A draft genome of sweet cherry (Prunus avium L.) reveals genome-wide and local effects of domestication. Plant J 103(4):1420–1432 Zhang J (1990) Report on the investigation for national plum and apricot resources. China Fruits 4:29–34 Zhang J, Zhou E (1998) China fruit-plant monographs, plum flora. China Forestry Press, Beijing Pop R, Hârţa M, Szabo K, Zanescu M, Sisea CR, Catana C, Pamfil D (2018) Genetic diversity and population structure of plum acces- sions from a Romanian germplasm collection assessed by simple sequence repeat (SSR) markers. Not Bot Hort Agrobot 46(1):90–96 ZhangQ,WeiX,LiuW,LiuN,ZhangY,XuM,LiuS,ZhangY,MaX,Dong W(2018)Thegeneticrelationshipandstructureofsomenaturalinterspe- cifichybridsinPrunussubgenusPrunophora,basedonnuclearandchlo- roplast simple sequence repeats. Genet Resour Crop Evol 65:625–636 26 Page 10 of 10 Tree Genetics & Genomes (2021) 17: 26 characterization of worldwide Prunus domestica (plum) germplasm using sequence-based genotyping. Hortic Res 6:12 characterization of worldwide Prunus domestica (plum) germplasm using sequence-based genotyping. Hortic Res 6:12 Publisher’s note Springer Nature remains neutral with regard to jurisdic- tional claims in published maps and institutional affiliations.
https://openalex.org/W2588404019	https://repository.library.noaa.gov/view/noaa/24516/noaa_24516_DS1.pdf	English	null	Practice and philosophy of climate model tuning across six US modeling centers	Geoscientific model development	2,017	cc-by	16,192	Correspondence to: Gavin A. Schmidt (gavin.a.schmidt@nasa.gov) Correspondence to: Gavin A. Schmidt (gavin.a.schmidt@nasa.gov) Received: 2 February 2017 – Discussion started: 14 February 2017 Received: 2 February 2017 – Discussion started: 14 February 2017 Received: 2 February 2017 – Discussion started: 14 February 2017 Revised: 17 July 2017 – Accepted: 24 July 2017 – Published: 1 September 2017 Revised: 17 July 2017 – Accepted: 24 July 2017 – Published: 1 September 2 Revised: 17 July 2017 – Accepted: 24 July 2017 – Published: 1 September 2017 Abstract. Model calibration (or “tuning”) is a necessary part of developing and testing coupled ocean–atmosphere climate models regardless of their main scientiﬁc purpose. There is an increasing recognition that this process needs to become more transparent for both users of climate model output and other developers. Knowing how and why climate models are tuned and which targets are used is essential to avoiding pos- sible misattributions of skillful predictions to data accommo- dation and vice versa. This paper describes the approach and practice of model tuning for the six major US climate mod- eling centers. While details differ among groups in terms of scientiﬁc missions, tuning targets, and tunable parameters, there is a core commonality of approaches. However, prac- tices differ signiﬁcantly on some key aspects, in particular, in the use of initialized forecast analyses as a tool, the ex- plicit use of the historical transient record, and the use of the present-day radiative imbalance vs. the implied balance in the preindustrial era as a target. tem is evident in the myriad processes involved (such as the microphysics of cloud nucleation, land surface heterogene- ity, convective plumes, and ocean mesoscale eddies) and in the dynamic views provided by remote sensing. This com- plexity and wide range of scales that need to be incorporated imply that simulations will necessarily include approxima- tions to well-understood physics and empirical formulations for unresolved effects. The simulations are neither a straight- forward encapsulation of some well-known theory, nor are they laboratory experiments probing the real world, though they have features of both (Schmidt and Sherwood, 2014). Despite this, climate and weather simulations have demon- strated useful predictive skill across many emergent diagnos- tics (Reichler and Kim, 2008; Flato et al., 2013; Bosilovich, 2013). Note that we distinguish ﬁelds or statistics in the model that arise from the interactions of multiple physical effects (“emergent properties”) from those that are closely related to single processes or parameterizations. Correspondence to: Gavin A. Schmidt (gavin.a.schmidt@nasa.gov) Since the pioneering work in climate modeling in the mid- 20th century (e.g., Phillips, 1956; Manabe and Bryan, 1969; Hansen et al., 1983), climate models have increased enor- mously in scope and complexity, going from relatively crude discretizations of atmospheric dynamics to, now, far more detailed atmospheres, combined with ocean, sea ice, carbon cycles, and interactive composition in the atmosphere, in- cluding chemistry and multiple aerosol species. As that com- 1 Introduction Simulation has become an essential tool for understanding processes in the Earth system, interpreting observations and for making predictions over short (weather), medium (sea- sonal), and long (climate) terms. The complexity of this sys- 1NASA Goddard Institute for Space Studies, 2880 Broadway, New York, USA 2 3GFDL/NOAA, Princeton University Forrestal Campus, 201 Forrestal Rd., Princeton, New Jersey 4Columbia University, New York, New York, USA y 5National Center for Atmospheric Research (NCAR), Boulder, Colorado, USA 6Global Modeling and Assimilation Ofﬁce, NASA GSFC, Greenbelt, Maryland, USA 7 Practice and philosophy of climate model tuning across six US modeling centers Gavin A. Schmidt1, David Bader2, Leo J. Donner3, Gregory S. Elsaesser1,4, Jean-Christophe Golaz2, Cecile Hannay5, Andrea Molod6, Richard B. Neale5, and Suranjana Saha7 Gavin A. Schmidt1, David Bader2, Leo J. Donner3, Gregory S. Elsaesser1,4, Jean-Christophe Andrea Molod6, Richard B. Neale5, and Suranjana Saha7 G. A. Schmidt et al.: US climate model tuning Others also emerge from the characteriza- tion of model subgrid-scale variations in the parameteriza- tions, such as the “critical relative humidity” for cloud forma- tion (Schmidt et al., 2006), or equivalent mixing rates for tur- bulent transport, and may be loosely approximated from ei- ther observations or higher-resolution models (e.g., Siebesma and Cuijpers, 1995). Thus it has become increasingly clear that a more trans- parent process is necessary. A survey of modeling groups in- volved in CMIP5 (Hourdin et al., 2017, hereafter H17) pro- vides a good background on tuning practices and makes a plea for better coordination of documentation of these issues. This paper is a more detailed follow-up for a subset of cli- mate models associated with laboratories in the US (three of which were surveyed by H17, three of which were not). The six modeling centers that are the focus of this paper have all developed and maintained Earth system models that (at minimum) have a dynamic atmosphere and coupled ocean components and are global in scope. Additional components (such as ice sheets, the carbon cycle, atmospheric chemistry, and aerosols) are also common. While two of the models dis- cussed (NCEP Climate Forecast System (CFS) and GEOS-5 (Goddard Earth Observing System 5) from NASA GMAO) are primarily used for short-term (daily to seasonal) predic- tions, there is sufﬁcient overlap with the models focused on longer-term problems (decadal to multidecadal periods) to warrant describing them all as “climate models” below. jp ) Individual parameterizations for a speciﬁc phenomenon are generally calibrated to process-level data using high- resolution modeling and/or ﬁeld campaigns to provide con- straints. For instance, boundary layer parameterizations might be tuned to well-observed case studies such as in Larcfrom (Pithan et al., 2016) or DICE (http://appconv. metofﬁce.com/dice/dice.html). However, in some cases, even when parameter values are well constrained physically or ex- perimentally, simulations can often be improved by choos- ing values that violate these constraints. For example, Go- laz et al. (2011) found that cooling by interactions between anthropogenic aerosols and clouds in GFDL’s AM3 model depends strongly on the volume-mean radius at which cloud droplets begin to precipitate. By altering CM3 after its con- ﬁguration used for CMIP5 was established as described in Donner et al. (2011), Golaz et al. G. A. Schmidt et al.: US climate model tuning approximations (“parameterizations”) needed to match the phenomenology of unresolved or poorly understood subgrid- scale or excluded processes (Hourdin et al., 2017). The de- gree of approximation and complexity in the empirical pa- rameterizations vary greatly across models and processes, as well as the resolved scales. Many parameterizations em- ploy an underlying paradigm that makes use of well-known or well-observed processes so that the fundamental depen- dence on the atmospheric state is approximated, albeit only at a phenomenological level. plexity has grown, more processes are explicitly included and the parameterizations are pushed to a more detailed (and more fundamental) level, allowing for better constraints on unknown parameters. However, at the same time, the process of model development has become more convoluted and now involves many more components than it did originally. This has led somewhat predictably to an unfortunate reduction in transparency over time. It is worth expanding on why this matters: ﬁrst, model de- velopment involves expert judgments which are inevitably subjective, and with different choices there would be dif- ferences in emergent responses. For instance, in the MPI model, Mauritsen et al. (2012) show that equally valid but distinct tunings can impact model sensitivity. This kind of be- havior should therefore be reported more widely to improve the assessment of the robustness of speciﬁc responses. Sec- ond, models used as part of international assessment projects (such as the Coupled Model Intercomparison Project, Phase 5: CMIP5) are increasingly being weighted or subset in order to reﬁne predictions. If the skill measure that is used to ﬁl- ter or weight models has been tuned for in some cases rather than in others, the subset or weighted average will be biased towards models where the skill measure was tuned over those in which it was not, and that may not correspond to better physics or better predictions (Knutti et al., 2010). Parameters in climate models vary widely in their phys- ical interpretation. Some are well-determined physical val- ues, such as the Coriolis parameter, the acceleration due to gravity, or the Stefan–Boltzmann constant. Some, such as reaction rates for chemical or microphysical processes, may be inferred from laboratory or ﬁeld measurements (with some uncertainty). Some emerge from the construction of parameterizations but do not correspond directly to well- deﬁned physical processes, e.g., “erosion rates” for clouds (Tiedtke, 1993). G. A. Schmidt et al.: US climate model tuning (2013) found that its 20th- century temperature increase could be simulated more real- istically (larger increase) using values for this threshold drop size smaller than observed (Pawlowska and Brenguier, 2003; Suzuki et al., 2013) (see Sect. 4.1 below). Another example is the variation in the effective diffusion constants for mo- mentum, moisture, and temperature, which has been used to decrease large root-mean-square errors in tropical winds in the NCEP model. Published by Copernicus Publications on behalf of the European Geosciences Union. Published by Copernicus Publications on behalf of the European Geosciences Union. 3208 G. A. Schmidt et al.: US climate model tuning which occurs in a multiyear integration with speciﬁed cli- matological sea surface temperatures (SSTs) when emissions (primary aerosols and short-lived gases), long-lived green- house gas (GHG) concentrations (carbon dioxide, nitrous ox- ide, methane, and the halocarbons), and solar irradiance are changed from preindustrial to present-day values. model against a selected set of target observations. The deci- sions on what to tune, and especially what targets to tune for, undoubtedly involve value judgments (Hourdin et al., 2017; Winsberg, 2012; Schmidt and Sherwood, 2014; Inte- mann, 2015). Notably, there is not any obvious consensus in the modeling community as to the extent to which parame- ter choices should be guided by conforming to process-level knowledge as opposed to optimizing emergent behaviors in climate models. At many centers, the philosophy for the most part has been to tune parameters in ways that make physical sense, with the expectation that in the long run that should be the best strategy. Increasing skill in climate models over time does support this approach (Reichler and Kim, 2008). A second class of tuning targets are well-characterized cli- matological observations which might include annual means, average seasonal cycles, or interannual variance. A third po- tential class are observations of transient events (on daily to centennial scales) or trends. Some observational targets have important (and sometimes unrecognized) structural uncer- tainties and therefore any tuning to those targets risks over- ﬁtting the model to imperfect data, potentially reducing skill in “out-of-sample” predictions (those for which the evalu- ation data either did not exist at the time of the prediction or were not used in model development or tuning). This is a particular problem for transient observations such as esti- mates of early 20th-century temperature changes (Thompson et al., 2008; Richardson et al., 2016), pre-1979 sea-ice extent (Meier et al., 2012; Walsh et al., 2017), pre-1990 ocean heat content change (Levitus et al., 2000; Church et al., 2011), or water vapor trends (Dessler and Davis, 2010), which have all been corrected in recent years, as nonclimate artifacts in the raw observations have been found and adjusted for. In con- trast, many climatologies over the satellite era are robust met- rics whose estimates over any ﬁxed period have not changed appreciably as understanding of the observations evolved. G. A. Schmidt et al.: US climate model tuning Additionally, climate simulations depend not only on pa- rameter choices within an established model structure but also on the structural choices made in the parameteriza- tion itself. Examples include experimentation with alternate closures and triggers for the cumulus parameterization at GFDL during the development of GFDL AM3 (Benedict et al., 2013) and evaluation of two candidate parameteri- zations for cloud macrophysics and convection in NCAR CESM2 (Community Earth System model 2) (Bogenschutz et al., 2013; Park, 2014). Theoretically, all such structural choices could be coded to vary with a parameter and so there is no strong theoretical distinction between parameter and structural variations. In practice, however, perturbed physics ensembles (PPEs) do not span as wide a range of structural variations as multimodel ensembles of opportunity (Yoko- hata et al., 2012). These examples suggest that it can be hard to distinguish model tuning from model development (writ large) in practice, since both happen concurrently. For our purposes, we deﬁne tuning as a change occurring within a ﬁxed structural framework that does not involve adding new physics. Models equipped for data assimilation or that are used for operational forecasts have the additional possibility of tuning parameters to improve skill scores in those forecasts on mul- tiple timescales – whether they be 6-hourly, daily, weekly, or even for many months for seasonal forecasts of, for instance, the state of the tropical Paciﬁc. Targets for possible tuning fall into three classes. First there are targets that need to be satisﬁed in order for use- ful numerical experiments to be performed in the ﬁrst place (usually related to the equilibration of model components with long timescales). The most important of these is a re- quirement of near energy balance at the top of the atmo- sphere and surface in an initial state of a coupled model to prevent temperature drifts over time. Strictly speaking this is not tuning to an observed quantity, but rather is a tuning to a situation that was approximately inferred to hold in the “preindustrial” (PI) period. Note that while the concept of a preindustrial period is a little elusive (Hawkins et al., 2017), we refer to conditions around the mid-19th century around 1850. 2 Why is climate model tuning necessary? Climate and weather models consist of three levels of rep- resentation of physical processes: fundamental physics (such as conservation of energy, mass and momentum), approxi- mations to well-known physical theories (the discretization of the Navier–Stokes equations, broadband approximations to line-by-line radiative transfer codes, etc.), and empirical A number of parameters remain that are not strongly con- strained by process-level observations or theory but that nonetheless have large impacts on emergent properties of the simulation. It is these additional degrees of freedom that are used to “tune” or calibrate the emergent properties of the Geosci. Model Dev., 10, 3207–3223, 2017 www.geosci-model-dev.net/10/3207/2017/ www.geosci-model-dev.net/10/3207/2017/ 3209 G. A. Schmidt et al.: US climate model tuning 3210 one ﬁeld are often accompanied by degradation in others, and thus the ﬁnal choice of parameters involves subjective judg- ments about the relative importance of different aspects of the simulations. For example, the Australian contribution to CMIP5 (ACCESS v.1) used a version of the UK Met Ofﬁce atmosphere model with small modiﬁcations to mitigate prob- lems in the tropics and Southern Hemisphere that affect Aus- tralian forecasts, at the expense of performance elsewhere (Bi et al., 2013). There are additionally many obvious biases in model simulations that persist across model generations, in- dicating that these aspects are robustly stubborn to develop- ment changes in the model (including the tuning) (Masson and Knutti, 2011). efﬁcient, but does not necessarily prove robust when those components are coupled. However, coupled models take a very long time to equilibrate, and their quasi-stable states may be too far from the observed climate to be useful. As- suming that models conserve energy appropriately, all con- trol runs will eventually drift to a quasi-steady state with a near-zero energy balance at the TOA and at the surface of the ocean. However, the realism of the ﬁnal state is not guaran- teed and, indeed, given the long time constants in the ocean, might require many thousands of years of integration to get to the wrong answer. Thus a balance must be struck between approaches. Most discussions of tuning deal with explicit calibration of parameters to match a target observation. However, analy- sis of the CMIP3 ensemble (Kiehl, 2007; Knutti, 2008) sug- gested that there may have been some kind of implicit tun- ing related to aerosol forcing and climate sensitivity among a subset of models, with models with higher sensitivity hav- ing a tendency to have higher (more negative) aerosol forc- ing (this situation was less evident in CMIP5; Forster et al., 2013). Both of these correlations, however, seem rather low (CMIP3: 0.24; CMIP5: 0.19) and so do not provide evidence for a general tuning related to forcing and sensitivity. That models with accurate historical simulations must trade off forcing and sensitivity is not necessarily evidence they have been tuned to do so. Since the CMIP3 models’ aerosol forc- ings were not explicitly tuned to enforce the observed his- torical trend in temperature, the mechanisms that might ex- plain this observation are unclear. G. A. Schmidt et al.: US climate model tuning With further data on the current top-of-atmosphere radiative imbalance (Allan et al., 2014; von Schuckmann et al., 2016), this issue will, however, need to be revisited for the latest generation of models. 3 Speciﬁc practices Each of six US modeling centers described below have spe- ciﬁc missions and foci that drive different aspects of their modeling. For instance, NASA GMAO and NCEP have op- erational data assimilation products for short-term weather, longer seasonal forecasts, and reanalyses that form the core of their tasks. NCAR CESM, GFDL, and NASA GISS have more long-term climate change issues at the forefront of their research, but each with different mandates – respectively, to be a community model, to advance NOAA’s mission goal to understand and predict changes in climate, and to help in- terpret and use NASA remote sensing products. The DOE’s Accelerated Climate Modeling for Energy (ACME) project has been tasked with a very speciﬁc role to serve DOE’s en- ergy planning and computational resource needs. For each modeling group, we describe the principal tar- gets and tuning strategies for their atmosphere-only GCM (general circulation model), their coupled ocean–atmosphere GCM, and additional Earth system components as relevant. The speciﬁc models referred to are described in Table 1. We outline the commonalities of approaches and key differences in Sect. 4, and then discuss the implications and ways for- ward in Sect. 5. Model selection can also act as an implicit form of tun- ing, even though this might be seen by others as simple model development. In deciding between two versions of a dynamical core or convection parameterizations, skill in El Niño–Southern Oscillation (ENSO) variability or reductions of ocean drifts may play an important role. Conceivably, a modeling center may decide not to release or use a particular version because it fails to meet certain criteria perceived to be essential, though more generally this will simply spur further development. One candidate criterium would be a realistic simulation of the 20th century; however, the wide spread in 20th-century trends in the CMIP5 ensemble (Forster et al., 2013, Fig. 7) would indicate that this has not been generally applied (though see below for more detailed discussion of the use of historical changes). G. A. Schmidt et al.: US climate model tuning To avoid dealing with the lack of sufﬁcient observa- tional data from the 19th century, some modeling groups (see below) alternatively choose to tune to present-day (PD) conditions, including an energy imbalance at the top of the atmosphere (TOA) as inferred from ocean observations to- day (Loeb et al., 2009). Another tuning target is sometimes referred to as the radiative forcing perturbation (RFP), or the effective radiative forcing, and is the change in net ﬂux We note here a distinction between ﬁelds that are closely monitored during the model development process (many ex- amples are given below) and speciﬁc tuning targets. Moni- tored diagnostics tend to be complex emergent diagnostics that do not depend in any simple way on adjustable param- eters, and thus are difﬁcult (or impractical) to tune for. For example, note that the range of preindustrial global tem- peratures in CMIP5 is [12.0,14.8] ◦C, which is noticeably wider than the uncertainties in that quantity (±0.5 ◦C; Jones et al., 2012). Changes in such a monitored ﬁeld are kept track of, but unless the values stray beyond a nominal acceptable range no action to change the code would be taken. If the values do stray, the principal action is often taken to go into detail and examine what has happened more closely. The limitations of tuning are well known (Mauritsen et al., 2012; Schmidt and Sherwood, 2014; Hourdin et al., 2017). First, it provides remarkably little leverage in improving overall model skill once a reasonable part of parameter space has been identiﬁed – for instance, tuning has been unable to resolve the persistent so-called “double ITCZ” problem (Lin, 2007; Oueslati and Bellon, 2015). Second, improvements in www.geosci-model-dev.net/10/3207/2017/ Geosci. Model Dev., 10, 3207–3223, 2017 3.1 DOE The prototype version of DOE ACME v0 is closely related to the CESM. The initial version ACME v1, currently under de- velopment, incorporates new ocean and sea-ice components (Model for Prediction Across Scales: MPAS) (Ringler et al., 2013) as well as updated atmosphere and land components. ACME v1 is being developed at two horizontal resolutions: a low-resolution conﬁguration, which includes an atmosphere at approximately 1◦and an ocean with varying resolution between 60 and 30 km, and a high-resolution conﬁguration, which is based on a 1/4◦atmosphere and an eddy-permitting ocean resolution between 18 and 6 km. Within climate models, there is always a choice as to whether to tune a speciﬁc component (such as the atmo- sphere, sea ice, land surface, or ocean) with tightly con- strained boundary conditions or to tune the coupled model as a whole. In practice, both approaches are taken, though the relative importance and computation resources available vary across groups. Tuning components is generally fast and Tuning is performed iteratively at the component levels and on the fully coupled system. Most of the component- level tuning takes place in the atmosphere. The atmosphere Geosci. Model Dev., 10, 3207–3223, 2017 www.geosci-model-dev.net/10/3207/2017/ G. A. Schmidt et al.: US climate model tuning Table 1. Climate models discussed in the text. Modeling group Model Reference Department of Energy (DOE) ACME v1 in preparation NOAA Geophysical Fluid CM3 Donner et al. (2011), Grifﬁes et al. (2011) Dynamics Laboratory (GFDL) NASA Goddard Institute for GISS-E2/2.1 Schmidt et al. (2014) Space Studies (GISS) NASA Global Modeling and GEOS5 Rienecker et al. (2008), Molod et al. (2015) Assimilation Ofﬁce (GMAO) National Center for Atmospheric CESM1 Gent et al. (2011), Hurrell et al. (2013) Research (NCAR) NOAA National Center for CFS v1 & v2 Saha et al. (2006, 2010, 2014) Environmental Prediction (NCEP) 3211 G. A. Schmidt et al.: US climate model tuning Table 1. Climate models discussed in the text. is primarily tuned using short simulations (2 to 10 years) with climatological SSTs and sea-ice boundary conditions, either for present-day (circa 2000) or preindustrial condi- tions. The tuning targets a near-zero TOA radiation balance for 1850 by adjusting cloud-related parameters. 3.1 DOE Overall sim- ulation ﬁdelity is another important aspect of the tuning pro- cess, with the goal of minimizing errors in important clima- tological ﬁelds such as sea level pressure, short- and long- wave cloud radiative effects, precipitation, near-surface land temperature, surface wind stress, 300 hPa zonal wind, aerosol optical depth, zonal mean temperature, and relative humidity. The magnitude of the aerosol indirect effects is also evaluated and adjusted if deemed to be inconsistent with the observed historical warming (speciﬁcally if it has a magnitude greater than 1.5 W m−2). Cess climate sensitivity (Cess et al., 1990) is monitored using idealized SST+4K simulations. To date, there have been no situations where the estimated sensitivity was deemed to be unacceptable based on expert judgment. Should such a situation arise, the model would receive ex- tra scrutiny to better understand what may have caused the climate sensitivity to change compared to previous develop- mental versions. The radiative imbalance in the 21st century with observed SST must be positive, with a target range of 0.5 to 1 W m−2. of ocean heat content, meridional heat transport, overturning circulations, and the Nino3.4 index. Longer coupled simula- tions are often performed in pairs of perpetual present-day and preindustrial forcing to monitor the combined impact of anthropogenic forcings and climate sensitivity and to max- imize the odds of successful historical simulations. To that end, parallel coupled simulations, one with perpetual 1850 forcings and one with perpetual 2000 forcings, will be tested to ensure that the 2000 control simulation is indeed warmer than the 1850 control. Abrupt 4×CO2 experiments are also conducted to estimate the equilibrium climate sensitivity. Geosci. Model Dev., 10, 3207–3223, 2017 G. A. Schmidt et al.: US climate model tuning G. A. Schmidt et al.: US climate model tuning 3212 tions ranged to −3 W m−2 (Lohmann and Feichter, 2005), and GFDL’s AM2, modiﬁed to include cloud–aerosol inter- actions, yielded an associated climate forcing of −2.3 W m−2 (Ming et al., 2005). Since climate forcing by greenhouse gases is around 3 W m−2 (IPCC, 2013), the most extreme estimates of climate forcing by cloud–aerosol interactions would not be compatible with observed historical tempera- ture increases. Given the approximate treatments of cloud– aerosol interactions in climate models, the possibility that some parameter combinations or formulations could lead to these extreme estimates could not be ruled out during model development. Indeed, Golaz et al. (2011) show that the magnitude of climate forcing by cloud–aerosol interac- tions depends strongly on the volume-mean drop radius at which cloud droplets begin to precipitate. Golaz et al. (2011) also ﬁnd that assumptions regarding the subgrid distribution of updraft speeds is an important control, though exerted through re-tuning for radiative balance as the distribution of updraft speeds is changed. The effective cloud-droplet ra- dius and cloud droplet number concentration are both cen- tral to climate forcing by cloud–aerosol interactions and vary strongly with aerosol size distribution (Feingold, 2003; Mc- Figgans et al., 2006). Ming et al. (2006), which is used to parameterize aerosol activation in AM3, supports a range of aerosol size distributions. volume-mean radius at which cloud droplets begin to precipi- tate), cloud erosion scales, and ice fall speeds in the Rotstayn (1997) and Tiedtke (1993) cloud microphysics and macro- physics parameterizations were tuned to improve regional patterns of TOA shortwave and longwave ﬂuxes, TOA short- wave and longwave cloud radiative effects, the Earth’s en- ergy imbalance, precipitation, and implied ocean heat trans- ports. The choices of a closure based on convective available potential energy (CAPE) for the Donner (1993) deep cu- mulus parameterization and the relaxation time and CAPE threshold in that closure were primarily motivated by their effects on the precipitation simulation. Tuning vertical diffu- sion of horizontal momentum in the Donner (1993) deep and Bretherton et al. (2004) shallow cumulus parameterizations impacted tropical precipitation and surface wind stresses. Other tunings related to convection include changes in en- trainment (partly to account for changes in vertical resolu- tion), the moisture budget for mesoscale circulations asso- ciated with deep convection, and maximum heights for the mesoscale circulations. G. A. Schmidt et al.: US climate model tuning These tunings improved precipita- tion, shortwave cloud radiative effects, and implied ocean heat transports. Changes in lateral entrainment for shallow convection (Bretherton et al., 2004) also improved these ﬁelds, limiting excessive low cloudiness in particular. The maximum heights of the mesoscale circulations also exerted a strong control on stratospheric water vapor. Between 100 and 10 hPa, zonally averaged water vapor mixing ratios are between 1.5 and 4 mg kg−1, mostly within 0.5 mg kg−1 of HALOE (Halogen Occultation Experiment) and MLS (Mi- crowave Limb Sounder) observations. The TOA RFP (see Sect. 2) was monitored during AM3 development, as was the Cess climate sensitivity. A conﬁg- uration for which the ratio of the RFP to the Cess sensitiv- ity was about 15 % less than the value for AM2 (The GFDL Global Atmospheric Model Development Team, 2004) was selected for AM3. This imposes a bound on RFP which de- pends on AM3 sensitivity and the forcing-to-sensitivity ratio in AM2. Within the limitations of the Cess sensitivity, this ratio was used to compare, without coupling, the changes in global-mean surface temperature that might be expected from CM3 relative to CM2. The basis for using this ra- tio as a tuning target is past experience that CM2 gener- ally simulates historical temperature change reasonably well. Without coupling, this target, while not constraining forc- ing or sensitivity independently, aimed to exploit knowledge from earlier models about their joint behavior, associated with realistic simulation of historical temperature change. The AM3 RFP is 0.99 W m−2, with the aerosol contribution about −1.6 W m−2 (Golaz et al., 2013). The coupled model CM3 was not further constrained with respect to its simula- tion of 20th-century climate change. Although the ratio of RFP to Cess sensitivity for AM3 is only about 15 % less than for AM2, 20th-century temperature increases in CM3 are less than observed, while CM2.1 temperature increases are greater than observed (Donner et al., 2011). The uncou- pled ratio is clearly limited in its ability to fully indicate tem- perature response in the coupled model. Aspects of AM3 related to variability, including station- ary wave patterns, relationships between the Niño-3 index and regional precipitation, relationships between the North- ern Hemisphere annular mode and regional pressure and tem- perature patterns, tropical cyclones, and the tropical wave spectrum, were monitored during AM3 development (Don- ner et al., 2011). 3.2 GFDL In developing the GFDL atmospheric model AM3 and cou- pled model CM3, parameter choices and some structural choices as to how to deploy parameterizations were guided by multiple goals. In addition to choosing parameters within plausible ranges suggested by observations, experiments, theory, or higher-resolution modeling, these goals included simulating thermodynamic and dynamical ﬁelds, as well as TOA regional shortwave and longwave ﬂuxes, as realistically as possible. The global and annual mean net TOA radiative ﬂux in integrations with speciﬁed, present-day (1981–2000) SSTs was tuned to a slight positive imbalance (0.8 W m−2) within observational estimates (Loeb et al., 2009). Particu- lar attention was also given to surface properties important for successfully coupling AM3 to models for sea ice (high- latitude surface energy balance) and ocean (wind stresses and implied ocean heat transports). Many of the changes in parameters from earlier GFDL models or nominal val- ues in literature describing the model parameterizations are summarized in Donner et al. (2011). For example, the mo- mentum source in the Alexander and Dunkerton (1999) pa- rameterization for gravity wave drag was chosen based on the stratospheric circulation it yielded. To facilitate optimiz- ing input parameters to this parameterization, the orographic wave parameterization was limited in the vertical extent of its application. Additionally, the autoconversion threshold (the Most of the tuning is performed using the low-resolution atmosphere. However, cloud parameterizations need to be re- tuned separately for the high-resolution atmosphere. Because of the cost of the high-resolution atmosphere, it is more ef- fective to use short hindcast simulations (Ma et al., 2015) to ﬁrst evaluate the parameter space. Tuning is also performed with the fully coupled system us- ing perpetual preindustrial or present-day forcing. Ocean and sea-ice initial conditions are either from rest (Locarnini et al., 2013; Zweng et al., 2013) or derived from separate CORE ex- periments (Grifﬁes et al., 2009). Simulations vary in length from a decade to over a century. Priority metrics for the cou- pled preindustrial simulations are top-of-atmosphere radia- tion, surface winds, sea-ice extent and thickness (climatol- ogy and seasonal cycle), sea surface temperatures, stability www.geosci-model-dev.net/10/3207/2017/ Geosci. Model Dev., 10, 3207–3223, 2017 www.geosci-model-dev.net/10/3207/2017/ G. A. Schmidt et al.: US climate model tuning the addition of a subgrid-scale explicit “cold pool” feature greatly enhanced variability on MJO timescales and lead to greatly increased forecast skill in initialized 20-day simula- tions (Del Genio et al., 2015). 0.3 W m−2). Tuning in CM3 was concentrated in the atmo- spheric component AM3. Outside of the atmospheric com- ponent, sea ice, land, and vegetation albedo, along with snow masking, were tuned. These tunings improved the Atlantic Meridional Overturning Circulation in preliminary coupled conﬁgurations, prior to ﬁnal tuning of the atmospheric com- ponent and subsequent initiation of the preindustrial coupled control. The resulting albedos were generally more realistic than those used in CM2.1. The change was made possible by CM3’s improved realism in regions with sea ice (Donner et al., 2011). Further ﬁne tuning in the coupled models, for instance for the exact global-mean surface temperature, is effectively pre- cluded by the long spin-up times and limited resources avail- able. No tuning is done for climate sensitivity or for perfor- mance in a simulation with transient forcing or hindcasts. In transient simulations without an explicit aerosol indirect ef- fect, the aerosol indirect effect was preset to have a value of −1 W m−2 in 2000 in the CMIP5 simulations (Miller et al., 2014), while conﬁgurations with aerosol microphysics have free latitude to produce whatever forcing is calculated. Note that the above description applies only to AM3 and CM3. For CM4 (Zhao et al., 2016), development is ongoing, and the speciﬁc tuning practices will be documented in future papers. In simulations with interactive atmospheric composition, there are two speciﬁc tunings for ozone chemistry: the pho- tolysis rate in the atmospheric window region for incom- ing solar radiation and the temperature threshold for the for- mation of polar stratospheric clouds (and hence the hetero- geneous chemistry associated with them) (Shindell et al., 2013). The former is tuned so that N2O and O3 ﬁelds in the lower tropical stratosphere match observations, while the lat- ter can be used to ensure that the polar ozone hole timing is correct despite potential biases in polar vortex tempera- tures. With respect to dust aerosols, emissions are tuned so that the model can match retrieved aerosol optical depths for the present day (Miller et al., 2006); similarly, tuning of the lightning parameterization (and associated source for NOx) is done against modern observations of ﬂash rate and tropo- spheric ozone amounts. 3.3 NASA GISS Tuning strategies in GISS ModelE2 are described in Schmidt et al. (2014). In the atmosphere-only simulation under 1850 preindustrial conditions, the parameters in the cloud schemes that control the threshold relative humidity and the critical ice mass for condensate conversion are used to achieve global ra- diative balance and a global-mean albedo of between 29 and 30 %. Additionally, parameters in the gravity wave drag are chosen to optimize the simulation to the lower stratospheric seasonal zonal wind ﬁeld and the minimum tropopause tem- perature. This also impacts high-latitude sea level pressure. In ocean-only simulations as described in the CORE proto- col (Grifﬁes et al., 2009), mixing parameters are chosen to minimize drift from observations in the basin-averaged tem- perature and salinity. For the E2.1 model and subsequent CMIP6 submissions, all tuning is being done with preindustrial and present- day fully interactive simulations (including chemistry and aerosols and indirect effects) and the noninteractive versions will use the composition derived from those simulations and the same tuning. Upon coupling the ocean and atmosphere models, there is an initial drift to a quasi-stable equilibrium which is judged on overall terms for realism, including the overall skill in the climatological metrics for zonal mean temperature, sur- face temperatures, sea level pressure, short- and longwave radiation ﬂuxes, precipitation, lower stratospheric water va- por, and seasonal sea-ice extent. For the conﬁguration to be acceptable, drifts have to be relatively small, and quasi- stable behavior of the North Atlantic meridional circulation and other ocean metrics, including the Antarctic Circumpolar Circulation, are required. While ENSO metrics are also mon- itored, they are not speciﬁcally tuned for. In practice, longer spin-up integrations help reduce drift, and the model state, once stabilized, can be assessed for suitability. Large drifts at the start of an integration have often been reduced by dif- ferent tuning choices that either affect surface atmospheric ﬂuxes or (more usually) ocean mixing. G. A. Schmidt et al.: US climate model tuning Optimal tuning for mean state and vari- ability in some cases conﬂicted. In AM3, this was particu- larly evident for the tropical wave spectrum, including the Madden–Julian Oscillation (MJO). Deep convective closures and triggers which produced a realistic mean simulation did so at the expense of the tropical wave spectrum (Benedict et al., 2013). AM3 includes prognostic aerosols based on emissions, transport, chemical processes, and dry and wet removal. An important aerosol tuning parameter is the strength of wet scavenging. In-cloud condensate fractions were prescribed to provide a reasonable simulation of the global mean and regional distribution of aerosol optical depth. These conden- sate fractions maintain relative solubilities among the various aerosols in AM3. AM3 is the ﬁrst GFDL model to include cloud–aerosol interactions. At the outset of this aspect of AM3 develop- ment, estimates of climate forcing by cloud–aerosol interac- The CM3 coupled model was initialized from present-day ocean conditions and allowed to adjust to a preindustrial, quasi-steady state with a small TOA energy imbalance (0.2– Geosci. Model Dev., 10, 3207–3223, 2017 www.geosci-model-dev.net/10/3207/2017/ www.geosci-model-dev.net/10/3207/2017/ 3213 G. A. Schmidt et al.: US climate model tuning the model validation suite. The tuning suite includes present- day (AMIP-style) climate simulations, “replay” experiments at different resolutions (similar to nudging towards a re- analysis), coupled atmosphere–ocean experiments, coupled atmosphere–chemistry simulations, short-term forecasts, and data assimilation experiments. Tropospheric Emission Spectrometer: TES) are used in the GMAO to validate the simulated turbulent and convective transport. Data from MERRA-2, for example, include an aerosol assimilation to assess errors in turbulent and convec- tive transport (Gelaro et al., 2016). Although not all the mis- matches between observed and modeled aerosol and CO are attributable to transport, particular events or locations are iso- lated where those processes dominate over others. These esti- mates are largely used to constrain the tuning of the GEOS-5 surface and atmospheric turbulence parameterizations. The choice of the turbulent length scale and the choice of param- eters that govern the entrainment into buoyantly rising turbu- lent parcels of air are made so as to constrain the turbulent transport of aerosol. The extent of vertical mixing as well as the advective transport out of the source regions is governed by this choice of tuning parameters. The tuning of the current version of the GEOS-5 AGCM is described in Molod et al. (2015), which shows the results of a series of sensitivity experiments demonstrating the impact of each change in tuning. The substantial majority of the tun- ing is focused on the behavior of the moisture turbulence pa- rameterizations, and also includes a parameter change in the gravity-wave drag scheme. For the lower-resolution appli- cations and uses, systematic comparisons of seasonal mean prognostic ﬁelds with different reanalysis estimates and com- parisons of cloud properties with satellite based estimates are used to identify errors in the mean present-day climate. It- erative 30-year simulations at low resolution (100 km) and repeated comparisons ensure that a change in tuning to ame- liorate one bias does not inadvertently exacerbate another. Key metrics include the mean and variance of the spatial dis- tribution of CERES observations of all sky TOA longwave and shortwave radiative ﬂuxes, together with the daily TOA longwave and shortwave distributions, which are monitored to ensure that performance does not degrade through the de- velopment or tuning process. The GEOS-5 AGCM includes some resolution-dependent parameters that govern the behavior of the moist processes. G. A. Schmidt et al.: US climate model tuning The two most important parameters that are speciﬁed to change with resolution in an ad hoc manner are chosen based on physical arguments and based on results from GEOS- 5 global mesoscale simulations. The ﬁrst of these is the critical relative humidity for condensation and evaporation, which accounts for subgrid-scale variations of total water. Critical RH increases with resolution based on the expecta- tion and evidence from global mesoscale model results that subgrid-scale variations of total water decrease with increas- ing resolution (Molod, 2012). The second of the resolution- dependent parameters is the so-called Tokioka limit in the convective parameterization. Again based on the expectation that the larger convective motions are resolved explicitly and on evidence from global mesoscale model results, the param- eters that govern the stochastic Tokioka limit changes so as to restrict parameterized deep convection at higher resolutions. The contribution of cloudy effects is approached by ad- justing the parameters that describe the cloud radiative ef- fect (cloud particle size and autoconversion rates). The clear- sky portion of the TOA ﬂuxes is matched by tuning the pa- rameters that govern the mean atmospheric humidity and surface albedo over ice-covered surfaces. The free atmo- sphere speciﬁc humidity is quite sensitive to the “critical rel- ative humidity” speciﬁed in the cloud macrophysical scheme (Molod, 2012), and so although this parameter is largely dictated by observed subgrid-scale moisture variations, the ﬁne-tuning and the details of the vertical proﬁle are tuned to match a consensus of reanalysis estimates of speciﬁc and relative humidity and Special Sensor Microwave Imager (SSM/I) total precipitable water. At the higher resolutions (25 km and better) the tuning pa- rameters are chosen based on short-term forecasts and the behavior as part of the data assimilation system. Forecast skill scores, the ﬁdelity of the spin-up of tropical cyclones and the innovation vector for data assimilation (observation- forecast statistics) are critical relevant metrics for new tuning choices, and any new choices of tuning parameters are evalu- ated with an ensemble of forecasts. The analysis increments during both data assimilation and replay experiments provide the key guidance for choosing the parameters to tune. 3.4 NASA GMAO The Goddard Earth Observing System model is currently in use at the NASA GMAO at a wide range of resolutions and for a wide range of applications. The range of resolu- tions and applications for the atmospheric model includes global mesoscale simulations and forecasts at approximately 7 km, atmospheric data assimilation and forecasts at 12 km (with ensemble members running at 50 km), seasonal cou- pled atmosphere–ocean forecasts at approximately 50 km, present-day climate simulations at 100 km, and present-day coupled chemistry climate simulations at resolutions from 12 to 100 km. The tuning of the GEOS-5 AGCM physical parameterizations, therefore, is designed to allow the model to function across this range of uses and requires ﬁdelity in many aspects of the simulation. The tuning also includes ap- propriate resolution dependence. Tuning targets differ among the many types of experiments that are conducted as part of Subsequent to CMIP5, further tuning exercises and de- velopment has occurred for the production of the E2.1 ver- sion of the model. One important tuning success was due to the adjustments made to the convection scheme in order to allow for the simulation of the Madden–Julian Oscillation (Kim et al., 2012). A combination of greater entrainment and www.geosci-model-dev.net/10/3207/2017/ Geosci. Model Dev., 10, 3207–3223, 2017 3214 www.geosci-model-dev.net/10/3207/2017/ G. A. Schmidt et al.: US climate model tuning scribed observed SSTs and sea-ice distributions. Initial de- velopment testing is performed using SSTs of the climato- logical period centered around the year 2000 for 5–10-year periods. This length of simulation is necessary due to the high Arctic variability. The ﬁrst key measure of a simula- tion that will be appropriate to the fully coupled simulation is the TOA energy balance. Estimates of the observed present- day energy imbalance are on the order of 0.5–1.0 W m−2 (Loeb et al., 2009), and the aim is to achieve close to that through modiﬁcation of cloud-related ﬁelds that have an im- pact both on the shortwave and longwave components of the energy budget. The ﬁrst quantitative assessment of simula- tion ﬁdelity is given by summary RMSE and bias scores for a number of variables key to the fully coupled system, includ- ing surface stresses, precipitation, temperature, cloud forc- ings, and surface pressure. A secondary assessment involves “preindustrial minus present day” simulations to determine the aerosol indirect effects that would be expected in his- torical coupled simulations. This involves ensuring that the net aerosol forcing is not greater in magnitude than about (negative) 1.5 W m−2, based on guidance from Myhre et al. (2013). rameters which are adjusted to meet the tuning targets are the autoconversion rates, ice-fall rates, and the cloud droplet size. In addition to these parameters, high-resolution tun- ing also includes adjustments of the Tokioka limit and the timescale of adjustment in the convective parameterization. As an aside, we note that resolution decisions almost always affect tunings (and development), and the goal that parame- terized physics or models can be independent of resolution, while a noble aim, is not yet a reality. The ability to spin up tropical cyclones and match the cor- rect track was found to be quite sensitive to the magnitude of low-level drag. Based on theoretical considerations and the results of laboratory experiments, the model’s function which relates surface stress to roughness height over the oceans (the “Charnock coefﬁcient”) was adjusted to decrease the drag at high wind speeds and resulted in substantial improvements in the simulation of tropical cyclones (Molod et al., 2013). In addition to the tuning based on physical reasoning and diagnosis of errors using comparisons with observations, some tuning choices are based on trial-and-error experimen- tation. G. A. Schmidt et al.: US climate model tuning These include parameters that govern the magnitude of the different types of surface drag (more drag increases forecast skill score) and the adjustment timescale of mid- latitude parameterized convection (more mid-latitude con- vection increases forecast skill score). In parallel to the atmosphere component activities, the ocean and ice working groups perform equivalent “stand- alone experiments” with forcing provided by multiple cy- cles of the CORE forcing protocol (Grifﬁes et al., 2009). The phenomena of key importance are the meridional over- turning circulation (particularly in the North Atlantic), Gulf Stream separation, Drake passage ﬂow, equatorial ther- mocline depth, and SSTs in the Paciﬁc. The land tun- ing approach uses land-only conﬁgurations forced by bias- corrected reanalysis-based meteorological forcing products. Metrics of performance are generally assessed for leaf area index, gross primary productivity, river discharge, latent heat ﬂux, and vegetation and soil carbon stocks. Other physical components of the coupled system, including land ice and biogeochemistry, will also be developed and tuned in paral- lel within their respective working groups. The suite of different types of experiments with the GEOS- 5 GCM at different resolutions are run iteratively as part of the overall tuning process, and the result is a model which meets the variety of tuning targets described here. The trade- offs among the parameter choices to meet the different tar- gets exist, and necessitate prioritization of the tuning targets, but in general this process results in a robust model that func- tions well in the various applications needed to fulﬁll the GMAO’s goals and mission. G. A. Schmidt et al.: US climate model tuning Under the general assumption that the mean analysis increments in- dicate systematic errors in the model physics (which is not always valid), correlations between the tendency term from any individual physical parameterization and the analysis in- crement reveals errors due to the behavior of that parameter- ization, and parameters of that scheme are adjusted so as to minimize the mean analysis increments. The boreal winter mean circulation, compared to reanal- yses (as seen by the 200 hPa eddy height or by the 300 hPa velocity potential), was found to be quite sensitive to the in- tensity of the hydrological cycle, largely dictated by the rates of re-evaporation or sublimation of rain and snow. These pa- rameters are chosen so as to ensure agreement of the seasonal mean circulation with reanalysis, the seasonal mean precip- itation with observations from GPCP and TRMM, and the agreement of the cloud radiative effects with CERES and with SRB at the surface. The behavior of the atmosphere– ocean coupled system is particularly sensitive to the geo- graphical distribution of the surface shortwave cloud radia- tive forcing in the tropics. High-resolution forecasts are also evaluated and tuned based on comparisons with spatial and temporal variability of high-resolution top-of-the-atmosphere ﬂuxes and radar- derived precipitation. As with the lower resolutions, the pa- Additional observations of aerosol optical depth (from MODIS) and other chemical constituents (e.g., CO from the www.geosci-model-dev.net/10/3207/2017/ Geosci. Model Dev., 10, 3207–3223, 2017 3215 G. A. Schmidt et al.: US climate model tuning G. A. Schmidt et al.: US climate model tuning used. The ﬁrst is to use an observed Levitus temperature and salinity state with the ocean at rest. The second approach is to initialize from an ocean state of a previously run simula- tion. This has the advantages of a spun-up ocean state, and in particular the deep ocean, that is more “familiar” with the overlying atmosphere component. However, it is undesirable from the perspective of simulation provenance. A combina- tion of the two are used. If the equilibrium energy imbalance is greater than 0.1–0.2 W m−2 then the system will need to be retuned, again most commonly through minor adjustments of cloud radiative impact parameters. If the energy imbal- ance and surface temperature drifts are observed to be small in short decadal runs, then longer 50–100-year simulations are performed to determine whether the performance of the ocean-ice-only simulations translate to the fully coupled sys- tem. ifying the forecasts, from the initial conditions provided by these data assimilation systems, with independent observa- tions. A new CFS is built by taking a snapshot of the latest state- of-the-art GFS as its atmospheric component, along with state-of-the-art ocean, sea-ice, and land models which are available at that time. In developing CFSv1 in 2002, a “large” (≈10) number of candidate coupled ocean–atmosphere models were constructed, which were then run on a limited number of test cases, with differing vertical and horizontal resolutions, as well as with different physics parameteriza- tions, such as convection and radiation schemes. The results were then judged, along with the normal veriﬁcation metrics, on whether the 9-month predictions produced skillful ENSO predictions. Our goal at that time was to be competitive with the statistical models that were predominantly being used for ENSO predictions. After initial testing, the model version that gave the best ENSO predictions was used to make retro- spective forecasts over a period of 20+ years (going back to 1982) in order to calibrate (remove the systematic bias in) the model forecasts and to make a priori skill assessments. These were then used in subsequent real-time operational forecasts made by the CFS. Since it is very expensive to make retro- spective forecasts over long periods (20–30 years) for every imaginable model conﬁguration, the preliminary test over a set of limited cases was extremely important. G. A. Schmidt et al.: US climate model tuning The dominant changes that improved skill were associated with tunings in the convection as model vertical layers were increased from 28 to 64 levels. For the coupled simulations to be considered successful, they have to satisfy many of the requirements outlined above in addition to the dominant ENSO mode of variability – also a very challenging task. For instance, the initial imple- mentation of more advanced convection parameterizations in CAM6 gave rise to a degradation in ENSO performance, but with some tuning to those schemes, ENSO performance skill was enhanced. Another example of coupled issues that arose in constructing the CMIP6 version of the code were a per- sistent cold bias and excessive sea ice in the Labrador Sea, which was mitigated by more accurate routing of local river runoff. In previous versions (such as CCSM4), there were evaluations of the coupled model in historical transient mode, speciﬁcally of the September Arctic sea-ice trend from 1979, which was improved after adjustments to the sea-ice albedo formulation to affect the PI ice thickness (Gent et al., 2011). A “reasonable” historical temperature trend remains the pri- mary metric of success, but no attempts are made to tune for it explicitly. Having achieved some success in the prediction of ENSO in seasonal forecasts up to 9 months ahead in CFSv1, the goal for CFSv2 was to tackle subseasonal predictions, mainly of the MJO in the tropics. Prediction of the MJO from 5 days was successfully extended to nearly 21 days by improving model physics and having a high-resolution state-of-the-art data assimilation system to assimilate direct satellite radiance data. Also, greenhouse gas concentration changes were im- plemented in the NCEP forecast system. While the NCEP focus is short-term (seasonal) climate prediction, it has been recognized that even for these predictions, the forecast needs to be warmer than a “normal” that, by necessity, is based on past data. The increase in GHGs also played an important role in improving the data assimilation of satellite radiance data. Each satellite over the 1979–present history was cali- brated using GHG concentrations observed at the time these satellites were operational. The result was a reasonable up- ward temperature trend over the 1979–present period, much better than at the time of CFSv1, when the upward trend over land was brought about only by the warming in initial global ocean conditions. As described in Saha et al. G. A. Schmidt et al.: US climate model tuning (2014), the sea- sonal prediction model may not be exactly the same as the model used for weather forecasts. In the absence of data as- similation, coupled ocean–atmosphere models can drift and produce, for instance, a very cold Paciﬁc Ocean due to a boundary layer parameterization change in a weather model 3.5 NCAR The Community Earth System model (Hurrell et al., 2013) is a joint NCAR and university-wide activity and gover- nance takes place through a working group structure. Work- ing groups are teams of scientists that contribute to the de- velopment of each individual component (atmosphere, land, ocean, sea ice, land ice, chemistry, and biogeochemistry) and relevant topics (such as climate variability or climate change). Ideally, an atmosphere that was well-tuned in a conﬁgu- ration with SSTs, sea ice, and land conditions relevant to the preindustrial period would in principle translate well to a coupled system close to energy balance, i.e., with no net increase or decrease in energy into the whole coupled sys- tem. However, coupled-system biases in the surface distribu- tion of SSTs and sea ice mean that tuning also needs to be performed in the fully coupled system. Tuning begins as a generally separate activity for each component within the working groups. During this initial phase of tuning, periodic preindustrial control coupled simu- lations are performed as a check on the impact of each com- ponents’ developments to date on the whole coupled system, and to ensure features of the simulation have not signiﬁcantly degraded. Coupled model tuning brings together the individual fully active “tuned” components and their associated working groups to perform a series of preindustrial climate experi- ments. The same performance metrics that are applied in at- mosphere AMIP simulations apply to the coupled simula- tion, namely top-of-atmosphere zero-energy imbalance. An equilibrium energy imbalance is the most challenging task in coupled CESM tuning. The difﬁculty lies in spin-up and drift of the system. Two ocean initialization approaches are The atmosphere model tuning strategy initially performs “stand-alone” experiments using the AMIP protocol with interactive land and atmosphere components and with pre- www.geosci-model-dev.net/10/3207/2017/ Geosci. Model Dev., 10, 3207–3223, 2017 3216 G. A. Schmidt et al.: US climate model tuning 4 Commonalities and differences As might be expected, the broad picture of tuning across the climate model groups is consistent. The key adjustable parameters are those associated with uncertain and poorly constrained processes such as clouds, convection, gravity wave drag, and ocean-mixing parameters. Common too are the broad array of targets against which skill of the mod- els are judged, e.g., the TOA shortwave and longwave radi- ation, 500 hPa geopotential height, surface temperatures, sea level pressure, and precipitation. However, it is also abun- dantly clear that the procedures at each group are quite dis- tinct and can reasonably be surmised to reﬂect different sci- entiﬁc priorities and missions and thus will produce different outcomes. The descriptions above suggest that increasing knowledge over time about the current radiative imbalance has clearly inﬂuenced model development. Developers prior to CMIP3 (circa 2004) had a general expectation that net radiative forc- ing over the 20th century was positive, but they were not able to use a speciﬁc value for the present-day energy imbalance because oceanic analyses were not accurate enough: compare Levitus et al. (2000) to Allan et al. (2014) for instance. Thus a posterior quantitative test of the model imbalance in cou- pled runs compared to (improving) observations was a valid test of skill (Hansen et al., 2005). This may not be true for a large fraction of simulations in CMIP6. The model groups also differ in whether they focus on preindustrial conditions or present-day simulations. The for- mer has the beneﬁt of being closer to climate stability, while the latter has substantially more observational data. The groups focusing on the preindustrial are judging (mostly cor- rectly) that the errors in the control simulation (whether run for preindustrial or present-day periods) are larger than the trends between those periods. A stark difference does exist between the models that have operational data assimilation products (NCEP and GMAO) and those that do not. The abil- ity to assess improvements in fast physics based on short forecasts is an excellent resource that, even if the climate models were not run operationally in this way, should be- come a more widely used methodology (e.g., Hurrell et al., 2006). Recent experience with this mode of testing in the GISS model has shown very positive results for representa- tion of the MJO and tropical convection (Del Genio et al., 2015). G. A. Schmidt et al.: US climate model tuning mospheric diagnostics use these experiments which have the advantage of fast equilibration times and reduced computa- tional load. Tunings for ocean components can be done in stand-alone experiments, but often are done within the full coupled framework, with at least some model groups tuning sea-ice and ocean-mixing parameterizations to produce ac- ceptable sea-ice cover and ocean circulation metrics. that produced more marine stratus clouds, but which became excessive in the fully coupled runs. This change was thus re- versed in the seasonal simulations. Development is now underway for the next model, CFSv3. NCEP/EMC has a strategic plan to unify the global fore- cast systems and develop a Uniﬁed Global Coupled Sys- tem (UGCS) for both weather and seasonal climate predic- tion. This system will have six fully coupled model compo- nents, namely the atmosphere, ocean, sea ice, land, waves and aerosols. It will also have a strongly coupled data assim- ilation system in each of these six components. 4.1 Use of recent trends and present-day radiative imbalance Because of the high importance and visibility of climate models’ simulation of the historical period (PI to PD), model groups have to be particularly clear in how information that reﬂects the ongoing trends in temperature and ocean heat content have been used in the tuning process. 3.6 NCEP In recent history two fully coupled climate models have be- come operational at NCEP, the CFS version 1 (Saha et al., 2006) and CFS version 2 (Saha et al., 2010, 2014). For the most part, the CFS and its predecessors (since there have been global climate models at NCEP since 1995) have been developed in the same way as weather prediction models. In- deed, the atmospheric component of the CFS is taken from the Global Forecast System (GFS), which is the NCEP ﬂag- ship that makes weather forecasts from day 1 to 15. Veriﬁca- tion against independent future reality (the weather happen- ing worldwide every day) shows the GFS and similar opera- tional models elsewhere steadily improving their skill scores on independent data over the last 50 years. The daily veriﬁcation skill scores are the dominant source for tracking model improvement. This is a powerful target for tuning which confronts the model with real-time obser- vations in evolving data assimilation systems and then ver- Geosci. Model Dev., 10, 3207–3223, 2017 www.geosci-model-dev.net/10/3207/2017/ 3217 G. A. Schmidt et al.: US climate model tuning G. A. Schmidt et al.: US climate model tuning Similarly, for the GFDL AM3 AMIP runs from 1981 to 2000, the radiative imbalance is 0.8 W m−2, while the im- balance for the same period in the coupled historical runs has an ensemble mean of 0.4 W m−2. The differences de- pend critically on the patterns of SST and sea ice, related to both the rectiﬁcation of interannual variability and the offsets in the coupled model climatology compared to observations. The question that is raised by this is whether, given the in- crease in forcings over the historical period and the sensitiv- ity each model has, tuning the present-day imbalance (how- ever deﬁned) determines (even to zeroth order) the coupled imbalance, the “committed warming” (at constant concen- trations) for the model, or the historical trend. With a perfect coupled model, and perfect knowledge of the forcings, this might be the case, but the imperfections in both imply that tuning to the PD imbalance is less of a constraint than might be assumed. good agreement according to standard GCM validation met- rics (e.g., Taylor diagrams, climate state mean biases, spatial correlations). If these distinct and separate volumes of tuning parameter space lead to simulations that exhibit similarly good agree- ment with observations, there is no clear scientiﬁc reason to prefer one over another. But will our decisions on parameter combinations today have a noticeable impact on the simu- lated climate several centuries from now or to climate sen- sitivity more broadly? Speciﬁcally, does choosing different local minima in parameter phase space “matter”? With more combinations, is there room for improving re- gional biases in simulations while simultaneously making the tuning process more automated? These questions have mo- tivated an effort, using the GISS model as a test bed, for developing a more robust framework for assessing the true existence of local minima in a multidimensional space (see also Hourdin et al., 2017). This is being explored by incorpo- rating situational or regime-dependent errors in observations or regional biases in GCM ﬁelds in weighted cost functions that deﬁne model “goodness”. We hope that this endeavor will increase the objectivity for deciding on the most appro- priate tuning parameters and either lead to improved met- rics for diagnosing the ﬁdelity of a particular model or reveal the spread in simulated climate sensitivity arising from set- tling on very different, but seemingly optimal, combinations of tuning parameters. G. A. Schmidt et al.: US climate model tuning 3218 Table 2. Use of historical period trends and imbalances during the tuning process. 2. Use of historical period trends and imbalances during the tuning process. Modeling Historical Radiative Radiative Aerosol forcing Aerosol group temp. trend balance (PI) imbalance (PD) (as tunable parameter) indirect effect (AIE) DOE Yes1 Yesa 0.5–1.0 W m−2a < 1.5 W m−2 Yes GFDL No No Yes, < 1.0 W m−2a No2 Yes GISS No Yesa No Yes/No3 Yes GMAO No n/a No No No (pending) NCAR Yes/No4 Yesc 0.5–1.0 W m−2a < 1.5 W m−2 Yes NCEP No n/a No No No a Using atmosphere-only or AMIP simulations. c Using coupled ocean–atmosphere simulations. 1 PD has to be warmer than PI. 2 However, sensitivity and forcing were jointly constrained with respect to the previous model. 3 Set in simulations with noninteractive composition only. 4 It was a necessary criteria for CCSM4, but not speciﬁcally tuned for. n/a = not applicable a Using atmosphere-only or AMIP simulations. c Using coupled ocean–atmosphere simulations. 1 PD has to be warmer than PI. 2 However, sensitivity and forcing were jointly constrained with respect to the previous model. 3 Set in simulations with noninteractive composition only. 4 It was a necessary criteria for CCSM4, but not speciﬁcally tuned for. n/a = not applicable a Using atmosphere-only or AMIP simulations. c Using coupled ocean–atmosphere simulations. 1 PD has to be warmer than PI. 2 However, sensitivity and forcing were jointly constrained with respect to the previous model. 3 Set in simulations with noninteractive composition only. 4 It was a necessary criteria for CCSM4, but not speciﬁcally tuned for. n/a = not applicable model simulations are unclear. For example in the GISS-E2 model, the decadal mean imbalance (1996–2005) in AMIP simulations, including all forcings and annually varying ob- served SST and sea ice, is 1.25 W m−2 (for 1981–2000 it is 0.6 W m−2). However, using the decadal mean SST and sea ice for the same period and constant year-2000 forc- ings, the imbalance is much larger, at 1.74 W m−2. Further- more, the decadal mean imbalance in coupled simulations with the same forcings is ≈1.0 ± 0.1 W m−2 (Miller et al., 2014). G. A. Schmidt et al.: US climate model tuning More generally, the large variety of approaches demon- strated among just these six models indicates that the doc- umentation of tuning procedures across a multimodel en- semble like CMIP6 will be quite challenging. What role should the degree of tuning matter when assessing the cou- pled model skill? Should simulations be up-weighted in the ensemble because of a closer climatology to observations, or down-weighted because this is partly due to accommoda- tion? Should models that are tuned differently but have sim- ilar physics be treated as independent or not? (Annan and Hargreaves, 2017; Knutti et al., 2017). These questions play into more fundamental issues related to how one should think 4 Commonalities and differences Groups also differ on which metrics they monitor for whether they are within an acceptable range, or if a speciﬁc value is tuned for directly (for instance, as for the present-day energy balance for some groups in Table 2). We summarize the results in Table 2. None of the models described here use the temperature trend over the historical period directly as a tuning target, nor are any of the mod- els tuned to set climate sensitivity to some preexisting as- sumption. However, NCAR, GFDL, and DOE do tune for a global radiative imbalance at near-present-day conditions. For instance, GFDL AM3 with observed SSTs was tuned to have a positive imbalance, with a magnitude less than about 1 W m−2 for 1981–2000. As discussed above, the radiative imbalance can be af- fected in two ways: by adjusting internal parameters (mostly associated with clouds) and/or by using a different historical forcing. Four models adjust their historical aerosol forcing: GISS, though only in its noninteractive runs, aims for an in- direct aerosol forcing of −1 W m−2 (Schmidt et al., 2014); NCAR CESM and DOE ACME tune for a substantive posi- tive effective radiative forcing at near-present conditions (im- plying a limit of −1.5 W m−2 for aerosols); GFDL AM3 con- strained its ratio of Cess sensitivity to RFP to be close to its value in its prior-generation coupled model, which implied an aerosol forcing around −1.6 W m−2 in AM3. At least three of the model groups discussed here ﬁnd a difference between the energy imbalance using year 2000 forcings together with observed SST and sea ice, and the transient coupled simulations for the same time period and forcings. However, the differences in how this calculation is done can be important, and the implications for the coupled There are also some clear commonalities in approaches. All groups focus on atmospheric models at ﬁrst either in an AMIP-style mode (annually varying modern SST and sea ice), using a climatological approach (decadal mean ob- served ocean conditions and forcings), or in weather fore- cast mode. Tunings for atmospheric composition and key at- www.geosci-model-dev.net/10/3207/2017/ Geosci. Model Dev., 10, 3207–3223, 2017 G. A. Schmidt et al.: US climate model tuning 5 Discussion and future approaches As models are continually evaluated at the process-level against an increasing number of observations, analyses often show that existing parameterizations lack enough ﬂexibility to represent the coupling between the subgrid scale and the environment in all relevant climate regimes. The response is often to increase the complexity of a parameterization, which comes at the cost of an increased number of tunable param- eters. With that increase, the challenges faced by the devel- opers also rise, as does the potential for “local minima” to occur, i.e., different parameter combinations have similarly www.geosci-model-dev.net/10/3207/2017/ Geosci. Model Dev., 10, 3207–3223, 2017 www.geosci-model-dev.net/10/3207/2017/ G. A. Schmidt et al.: US climate model tuning 3219 Table 3. Code availability at each center. Model Code URL and collaboration policies (if relevant) DOE ACME https://climatemodeling.science.energy.gov/projects/accelerated-climate-modeling-energy https://climatemodeling.science.energy.gov/sites/default/ﬁles/publications/ACME_collaboration_30Jun15.pdf GFDL CM3 https://www.gfdl.noaa.gov/modeling-systems-group-public-releases/ GISS ModelE2 https://www.giss.nasa.gov/tools/modelE/ GMAO GEOS-5 http://geos5.org/wiki/index.php?title=GEOS-5_public_AGCM_Documentation_and_Access NCAR CESM1 http://www.cesm.ucar.edu/models/cesm1.2 http://www.cesm.ucar.edu/models/cesm1.2/copyright.html NCEP CFSv2 http://cfs.ncep.noaa.gov/cfsv2/downloads.html 3219 G. A. Schmidt et al.: US climate model tuning Table 3. Code availability at each center. about an unstructured multimodel ensemble (see, e.g., Knutti et al., 2010, 2013). Center for Climate Simulation (NCCS) at Goddard Space Flight Center. Work at LLNL was supported by the U.S. Department of Energy, Ofﬁce of Science E3SM project under contract DE-AC52- 07NA27344. NCEP is a division of the National Weather Service in NOAA within the Department of Commerce. Discussions at the US Climate Modeling Summit, convened by USGCRP in February 2016, were instrumental for putting this paper together. Manuscript reviews by Larry Horowitz and Levi Silvers (GFDL) are appreciated. We would like to thank Steve Sherwood and an anonymous reviewer for constructive comments on the discussion version of the paper. At the minimum, we recommend that all future model de- scription papers (or systematic documentation projects such as ES-DOC http://es-doc.org) include a list of tuned-for tar- gets and monitored diagnostics and describe clearly (as in Table 2) their use of historical trends and imbalances in the development process. While we have only discussed tuning in the context of his- torical and modern simulations, it is vital to assess the cred- ibility of models by examining their performance in out-of- sample situations. This is easy for the models with an oper- ational weather forecast mode (at least for some aspects of the climate system), and participation in paleoclimate model tests by NCAR and GISS are also invaluable. 5 Discussion and future approaches Medium-term climate forecasts based on anticipated changes in forcings (such as the eruption of Mount Pinatubo (1991) or the rise in greenhouse gases) have been shown to have skill (Hansen et al., 1988, 1992; Hargreaves, 2010). The importance (or lack thereof) of tuning always needs to be seen within that context. This paper alone cannot hope to answer all of the above questions, but we hope that it can contribute to a more transparent and more widely usable discussion. Edited by: James Annan Reviewed by: Steven Sherwood and one anonymous referee Edited by: James Annan Reviewed by: Steven Sherwood and one anonymous referee Edited by: James Annan Edited by: James Annan Reviewed by: Steven Sherwood and one anonymous referee Edited by: James Annan Reviewed by: Steven Sherwood and one anonymous referee References Alexander, M. and Dunkerton, T.: A spectral parameterization of mean-ﬂow forcing due to breaking gravity waves, J. Atmos. Sci., 56, 4167–4182, 1999. Allan, R. P., Liu, C., Loeb, N. G., Palmer, M. D., Roberts, M., Smith, D., and Vidale, P.-L.: Changes in global net radiative imbalance 1985–2012, Geophys. Res. Lett., 41, 5588–5597, https://doi.org/10.1002/2014gl060962, 2014. Annan, J. D. and Hargreaves, J. C.: On the meaning of inde- pendence in climate science, Earth Syst. Dynam., 8, 211–224, https://doi.org/10.5194/esd-8-211-2017, 2017. Code and data availability. The availability of code for the models discussed in this paper is laid out in Table 3, though note that some decisions and/or tunings mentioned above are associated with cur- rent development which may only be accessible with a collaborative agreement in place. Benedict, J., Maloney, E., Sobel, A., Frierson, D., and Don- ner, L.: Tropical intraseasonal variability in Version 3 of the GFDL atmosphere model, J. Climate, 26, 426–449, https://doi.org/10.1175/JCLI-D-12-00103.1, 2013. Bi, D., Dix, M., Marsland, S., O’Farrell, S., Rashid, H., Uotila, P., Hirst, A., Kowalczyk, E., Golebiewski, M., Sullivan, A., Yan, H., Hannah, N., Franklin, C., Sun, Z., Vohralik, P., Watterson, I., Zhou, X., Fiedler, R., Collier, M., Ma, Y., Noonan, J., Stevens, L., Uhe, P., Zhu, H., Grifﬁes, S., Hill, R., Harris, C., and Puri, K.: The ACCESS coupled model: description, control climate and evaluation, Aust. Met. Oceanogr. J., 63, 41–64, 2013. Competing interests. The authors declare that they have no conﬂict of interest. Acknowledgements. Climate modeling at GISS and GMAO is supported by the NASA Modeling, Analysis and Prediction program and resources supporting this work were provided by the NASA High-End Computing (HEC) Program through the NASA Bogenschutz, P., Gettelman, A., Morrison, H., Larson, V., Craig, C., and Schanen, D.: Higher-order closure and its impact on climate simulations in the Community Atmosphere Model, Geosci. Model Dev., 10, 3207–3223, 2017 G. A. Schmidt et al.: US climate model tuning L., Ghan, S. J., Kiehl, J. T., Lacis, A. A., Le Treut, H., Li, Z.-X., Liang, X.-Z., McAvaney, B. J., Meleshko, V. P., Mitchell, J. F. B., Morcrette, J.-J., Randall, D A Rik L R k E R J F S hl U Sh i i Cess, R., Potter, G., Blanchet, J., Boer, G., Genio, A. D., Déqué, M., Dymnikov, V., Galin, V., Gates, W. L., Ghan, S. J., Kiehl, J. T., Lacis, A. A., Le Treut, H., Li, Z.-X., Liang, X.-Z., McAvaney, B. J., Meleshko, V. P., Mitchell, J. F. B., Morcrette, J.-J., Randall, D. A., Rikus, L., Roeckner, E., Royer, J. F., Schlese, U., Sheinin, D. A., Slingo, A., Sokolov, A. P., Taylor, K. E., Washington, W. M., Wetherald, R. T., Yagai, I., and Zhang, M.-H.: Intercompari- son and interpretation of climate feedback processes in 19 atmo Lacis, A. A., Le Treut, H., Li, Z.-X., Liang, X.-Z., McAvaney, B. J., Meleshko, V. P., Mitchell, J. F. B., Morcrette, J.-J., Randall, J., Meleshko, V. P., Mitchell, J. F. B., Morcrette, J.-J., Randall, D. A., Rikus, L., Roeckner, E., Royer, J. F., Schlese, U., Sheinin, D. A., Rikus, L., Roeckner, E., Royer, J. F., Schlese, U., Sheinin, D. A., Slingo, A., Sokolov, A. P., Taylor, K. E., Washington, W. M., Wetherald, R. T., Yagai, I., and Zhang, M.-H.: Intercompari- son and interpretation of climate feedback processes in 19 atmo- spheric general circulation models, J. Geophys. Res., 95, 16601– 16615, https://doi.org/10.1029/JD095iD10p16601, 1990. Gent, P. R., Danabasoglu, G., Donner, L. J., Holland, M. M., Hunke, E. C., Jayne, S. R., Lawrence, D. M., Neale, R. B., Rasch, P. J., Vertenstein, M., Worley, P. H., Yang, Z.-L., and Zhang, M.: The Community Climate System Model Version 4, J. Climate, 24, 4973–4991, https://doi.org/10.1175/2011jcli4083.1, 2011. Church, J. A., White, N. J., Konikow, L. F., Domingues, C. M., Cog- ley, J. G., Rignot, E., Gregory, J. M., van den Broeke, M. R., Monaghan, A. J., and Velicogna, I.: Revisiting the Earth’s sea- level and energy budgets from 1961 to 2008, Geophys. Res. Lett., 38, L18601, https://doi.org/10.1029/2011GL048794, 2011. The GFDL Global Atmospheric Model Development Team: The new GFDL global atmosphere and land model AM2-LM2: Eval- uation with prescribed SST simulations, J. Climate, 17, 4641– 4673, https://doi.org/10.1175/JCLI-3223.1, 2004. Del Genio, A. D., Wu, J., Wolf, A. B., Chen, Y., Yao, M.-S., and Kim, D.: Constraints on Cumulus Parameterization from Sim- ulations of Observed MJO Events, J. G. A. Schmidt et al.: US climate model tuning Climate, 28, 6419–6442, https://doi.org/10.1175/jcli-d-14-00832.1, 2015. Golaz, J.-C., Salzmann, M., Donner, L., Horowitz, L., Ming, Y., and Zhao, M.: Sensitivity of the Aerosol Indirect Effect to Subgrid Variability in the Cloud Parameterization of the GFDL Atmo- sphere General Circulation Model AM3, J. Climate, 24, 3145– 3160, https://doi.org/10.1175/2010JCLI3945.1, 2011. Dessler, A. E. and Davis, S. M.: Trends in tropospheric humid- ity from reanalysis systems, J. Geophys. Res., 115, D19127, https://doi.org/10.1029/2010jd014192, 2010. Donner, L.: A cumulus parameterization including mass ﬂuxes, vertical momentum dynamics, and mesoscale effects, J. Atmos. Sci., 50, 889–906, https://doi.org/10.1175/1520- 0469(1993)050<0889:ACPIMF>2.0.CO;2, 1993. Golaz, J.-C., Horowitz, L., and Levy II, H.: Cloud tuning in a cou- pled climate model: Impact on 20th century warming, Geophys. Res. Lett., 40, 2246–2251, https://doi.org/10.1002/grl.50232, 2013. Donner, L. J., Wyman, B., Hemler, R., Horowitz, L., Ming, Y., Zhao, M., Golaz, J.-C., Ginoux, P., Lin, S., Schwarzkopf, M. D., Austin, J., Alaka, G., Cooke, W. F., Delworth, T. L., Freidenre- ich, S. M., Gordon, C. T., Grifﬁes, S. M., Held, I. M., Hurlin, W. J., Klein, S. A., Knutson, T. R., Langenhorst, A. R., Lee, H., Lin, Y., Magi, B. I., Malyshev, S. L., Milly, P. C., Naik, V., Nath, M. J., Pincus, R., Ploshay, J. J., Ramaswamy, V., Se- man, C. J., Shevliakova, E., Sirutis, J. J., Stern, W. F., Stouffer, R. J., Wilson, R. J., Winton, M., Wittenberg, A. T., and Zeng, F.: The dynamical core, physical parameterizations, and basic simulation characteristics of the atmospheric component of the GFDL global coupled model CM3, J. Climate, 24, 3484–3519, https://doi.org/10.1175/2011JCLI3955.1, 2011. Grifﬁes, S., Winton, M., Donner, L., Horowitz, L., Downes, S., Far- neti, R., Gnanadesikan, A., Hurlin, W. J., Lee, H., Liang, Z., Pal- ter, J. B., Samuels, B. L., Wittenberg, A. T., Wyman, B. L., Yin, J., and Zadeh, N.: GFDL’s CM3 coupled climate model: Char- acteristics of the ocean and sea ice simulations, J. Climate, 24, 3520–3544,https://doi.org/10.1175/2011JCLI3964.1, 2011. Grifﬁes, S. M., Biastoch, A., Böning, C., Bryan, F., Danabasoglu, G., Chassignet, E. P., England, M. H., Gerdes, R., Haak, H., Hallberg, R. W., Hazeleger, W., Jungclaus, J., Large, W. G., Madec, G., Pirani, A., Samuels, B. L., Scheinert, M., Gupta, A. S., Severijns, C. A., Simmons, H. L., Treguier, A. M., Winton, M., Yeager, S., and Yin, J.: Coordinated Ocean-ice Reference Experiments (COREs), Ocean Model., 26, 1–46, https://doi.org/10.1016/j.ocemod.2008.08.007, 2009. Feingold, G.: Modeling of the ﬁrst indirect effect: Analysis of measurement requirements, Geophys. Res. G. A. Schmidt et al.: US climate model tuning 3220 J. Climate, 26, 9655–9676, https://doi.org/10.1175/JCLI-D-13- 00075.1, 2013. J. Climate, 26, 9655–9676, https://doi.org/10.1175/JCLI-D-13- 00075.1, 2013. Qin, D., Plattner, G.-K., Tignor, M., Allen, S. K., Boschung, J., Nauels, A., Xia, Y., Bex, V., and Midgley, P., Cambridge Uni- versity Press, Cambridge, United Kingdom and New York, NY, USA, 2013. Bosilovich, M. G.: Regional Climate and Variability of NASA MERRA and Recent Reanalyses: U.S. Summertime Precipita- tion and Temperature, J. Appl. Meteorol. Clim., 52, 1939–1951, https://doi.org/10.1175/jamc-d-12-0291.1, 2013. Forster, P. M., Andrews, T., Good, P., Gregory, J. M., Jackson, L. S., and Zelinka, M.: Evaluating adjusted forcing and model spread for historical and future scenarios in the CMIP5 gener- ation of climate models, J. Geophys. Res.-Atmos., 118, 1139– 1150, https://doi.org/10.1002/jgrd.50174, 2013. Bretherton, C., McCaa, J., and Grenier, H.: A new parameteriza- tion for shallow cumulus parameterization and its application to marine subtropical cloud-topped boundary layers, Mon. Weather Rev., 132, 864–882, 2004. Gelaro, R., McCarty, W., Suárez, M. J., Todling, R., Molod, A., Takacs, L., Randles, C. A., Darmenov, A., Bosilovich, M. G., Re- ichle, R., Wargan, K., Coy, L., Cullather, R., Draper, C., Akella, S., Buchard, V., Conaty, A., da Silva, A. M., Gu, W., Kim, G., Koster, R., Lucchesi, R., Merkova, D., Nielsen, J. E., Partyka, G., Pawson, S., Putman, W., Rienecker, M., Schubert, S. D., Sienkiewicz, M., and Zhao, B.: The Modern-Era Retrospective Analysis for Research and Applications, Version-2 (MERRA-2), J. Climate, 30, 5419–5454, https://doi.org/10.1175/JCLI-D-16- 0758.1, 2016. Cess, R., Potter, G., Blanchet, J., Boer, G., Genio, A. D., Déqué, M., Dymnikov, V., Galin, V., Gates, W. L., Ghan, S. J., Kiehl, J. T., Lacis, A. A., Le Treut, H., Li, Z.-X., Liang, X.-Z., McAvaney, B. J., Meleshko, V. P., Mitchell, J. F. B., Morcrette, J.-J., Randall, D. A., Rikus, L., Roeckner, E., Royer, J. F., Schlese, U., Sheinin, D. A., Slingo, A., Sokolov, A. P., Taylor, K. E., Washington, W. M., Wetherald, R. T., Yagai, I., and Zhang, M.-H.: Intercompari- son and interpretation of climate feedback processes in 19 atmo- spheric general circulation models, J. Geophys. Res., 95, 16601– 16615, https://doi.org/10.1029/JD095iD10p16601, 1990. Cess, R., Potter, G., Blanchet, J., Boer, G., Genio, A. D., Déqué, M., Dymnikov, V., Galin, V., Gates, W. L., Ghan, S. J., Kiehl, J. T., Lacis, A. A., Le Treut, H., Li, Z.-X., Liang, X.-Z., McAvaney, B. Cess, R., Potter, G., Blanchet, J., Boer, G., Genio, A. D., Déqué, M., Dymnikov, V., Galin, V., Gates, W. G. A. Schmidt et al.: US climate model tuning 3221 The tropical subseasonal variability simulated in the NASA GISS general circulation model, J. Climate, 25, 4641–4659, https://doi.org/10.1175/JCLI-D-11-00447.1, 2012. Hansen, J., Nazarenko, L., Ruedy, R., Sato, M., Willis, J., Del Ge- nio, A., Koch, D., Lacis, A., Lo, K., Menon, S., Novakov, T., Perl- witz, J., Russell, G., Schmidt, G. A., and Tausnev, N. L.: Earth’s Energy Imbalance: Conﬁrmation and Implications, Science, 308, 1431–1435, https://doi.org/10.1126/science.1110252, 2005. Knutti, R.: Should we believe model predictions of future cli- mate change?, Philos. T. Roy. Soc. A, 366, 4647–4664, https://doi.org/10.1098/rsta.2008.0169, 2008. Hansen, J. E., Russell, G. L., Rind, D., Stone, P., Lacis, A., Ruedy, R., and Travis, L.: Efﬁcient three- dimensional models for climatic studies, Mon. Weather Rev., 111, 609–662, https://doi.org/10.1175/1520- 0493(1983)111<0609:ETDGMF>2.0.CO;2, 1983. Knutti, R., Abramowitz, G., Collins, M., Eyring, V., Gleckler, P. J., Hewitson, B., and Mearns, L.: Good Practice Guidance Paper on Assessing and Combining Multi Model Climate Projections, Tech. rep., IPCC Working Group I Technical Support Unit, Uni- versity of Bern, Bern, Switzerland, in: Meeting Report of the Intergovernmental Panel on Climate Change Expert Meeting on Assessing and Combining Multi Model Climate Projections, edited by: Stocker, T. F., Qin, D., Plattner, G.-K., Tignor, M., and Midgley, P. M., 2010. Hargreaves, J. C.: Skill and uncertainty in climate models, Wi- ley Interdisciplinary Reviews: Climate Change, 1, 556–564, https://doi.org/10.1002/wcc.58, 2010. Hawkins, E., Ortega, P., Suckling, E., Schurer, A., Hegerl, G., Jones, P., Joshi, M., Osborn, T. J., Masson-Delmotte, V., Mignot, J., Thorne, P., and van Oldenborgh, G. J.: Estimating changes in global temperature since the pre-industrial period, B. Am. Meteorol. Soc., https://doi.org/10.1175/bams-d-16-0007.1, on- line ﬁrst, 2017. Knutti, R., Masson, D., and Gettelman, A.: Climate model geneal- ogy: Generation CMIP5 and how we got there, Geophys. Res. Lett., 40, 1194–1199, https://doi.org/10.1002/grl.50256, 2013. Knutti, R., Sedláˇcek, J., Sanderson, B. M., Lorenz, R., Fis- cher, E. M., and Eyring, V.: A climate model projec- tion weighting scheme accounting for performance and interdependence, Geophys. Res. Lett., 44, 1909–1918, https://doi.org/10.1002/2016gl072012, 2017. Hourdin, F., Mauritsen, T., Gettelman, A., Golaz, J.-C., Balaji, V., Duan, Q., Folini, D., Klocke, D. J. D., Qian, Y., Rauser, F., Rio, C., Tomassini, L., Watanabe, M., and Williamson, D.: The art and science of climate model tuning, B. Am. Meteorol. Soc., 98, 589–602, https://doi.org/10.1175/BAMS-D-15-00135.1, 2017. Levitus, S., Antonov, J. I., Boyer, T. P., and Stephens, C.: Warming of the world ocean, Science, 287, 2225–2228, https://doi.org/10.1126/science.287.5461.2225, 2000. Hurrell, J. G. A. Schmidt et al.: US climate model tuning W., Bader, D., Delworth, T., Kirtman, B., Meehl, J., Pan, H.-L., and Wielicki, B.: Seamless Prediction. White paper for the U.S. Inter-Agency Working Group on Climate Modeling, available at: http://www.cgd.ucar.edu/staff/jhurrell/ docs/SeamlessModellingDraft03302007.pdf (last access: 21 Au- gust 2017), 2006. Lin, J.-L.: The Double-ITCZ Problem in IPCC AR4 Coupled GCMs: Ocean–Atmosphere Feedback Analysis, J. Climate, 20, 4497–4525, https://doi.org/10.1175/jcli4272.1, 2007. Locarnini, R. A., Mishonov, A. V., Antonov, J. I., Boyer, T. P., Gar- cia, H. E., Baranova, O. K., Zweng, M. M., Paver, C. R., Reagan, J. R., Johnson, D. R., Hamilton, M., and Seidov, D.: World Ocean Atlas 2013, Volume 1: Temperature, NOAA Atlas NESDIS 73, U.S. Government Printing Ofﬁce, Washington, D.C., edited by: Levitus, S. and Mishonov, A., 2013. Hurrell, J. W., Holland, M. M., Gent, P. R., Ghan, S., Kay, J. E., Kushner, P. J., Lamarque, J.-F., Large, W. G., Lawrence, D., Lindsay, K., Lipscomb, W. H., Long, M. C., Mahowald, N., Marsh, D. R., Neale, R. B., Rasch, P., Vavrus, S., Vertenstein, M., Bader, D., Collins, W. D., Hack, J. J., Kiehl, J., and Mar- shall, S.: The Community Earth System Model: A Framework for Collaborative Research, B. Am. Meteorol. Soc., 94, 1339– 1360, https://doi.org/10.1175/bams-d-12-00121.1, 2013. Loeb, N., Wielicki, B., Doelling, D., Smith, G., Keyes, D., Kato, S., Matalo-Smith, N., and Wong, T.: Toward optimal closure of the Earth’s top-of-atmosphere radiation budget, J. Climate, 22, 748–766, https://doi.org/10.1175/2008JCLI2637.1, 2009. Intemann, K.: Distinguishing between legitimate and illegitimate values in climate modeling, European Journal for Philosophy of Science, 5, 217–232, https://doi.org/10.1007/s13194-014-0105- 6, 2015. Lohmann, U. and Feichter, J.: Global indirect aerosol ef- fects: a review, Atmos. Chem. Phys., 5, 715–737, https://doi.org/10.5194/acp-5-715-2005, 2005. IPCC: Summary for Policymakers. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Cli- mate Change, edited by: Stocker, T. F., Qin, D., Plattner, G.- K., Tignor, M., Allen, S. K., Boschung, J., Nauels, A., Xia, Y., Bex, V., and Midgley, P. M., Cambridge University Press, Cam- bridge/New York, 2013. Ma, H.-Y., Chuang, C. C., Klein, S. A., Lo, M.-H., Zhang, Y., Xie, S., Zheng, X., Ma, P.-L., Zhang, Y., and Phillips, T. J.: An im- proved hindcast approach for evaluation and diagnosis of phys- ical processes in global climate models, J. Adv. Model. Earth Syst., 7, 1810–1827, https://doi.org/10.1002/2015ms000490, 2015. Manabe, S. and Bryan, K.: Climate Calculations with a Combined Ocean-Atmosphere Model, J. At- mos. G. A. Schmidt et al.: US climate model tuning Lett., 30, 1997, https://doi.org/10.1029/2003GL017967, 2003. Hansen, J., Fung, I., Lacis, A., Rind, D., Lebedeff, S., Ruedy, R., Russell, G., and Stone, P.: Global climate changes as forecast by Goddard Institute for Space Studies three-dimensional model, J. Geophys. Res., 93, 9341–9364, 1988. Flato, G., Marotzke, J., Abiodun, B., Braconnot, P., Chou, S. C., Collins, W., Cox, P., Driouech, F., Emori, S., Eyring, V., Forest, C., Gleckler, P., Guilyardi, E., Jakob, C., Kattsov, V., Reason, C., and Rummukaine, M.: Evaluation of Climate Models, in: Cli- mate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergov- ernmental Panel on Climate Change, edited by: Stocker, T. F., Hansen, J., Lacis, A., Ruedy, R., and Sato, M.: Potential climate im- pact of Mount Pinatubo eruption, Geophys. Res. Lett., 19, 215– 218, 1992. Geosci. Model Dev., 10, 3207–3223, 2017 www.geosci-model-dev.net/10/3207/2017/ G. A. Schmidt et al.: US climate model tuning 3222 global model, J. Adv. Model. Earth Syst., 4, M00A01, https://doi.org/10.1029/2012MS000154, 2012. versity Press, Cambridge, United Kingdom and New York, NY, USA, 2013. versity Press, Cambridge, United Kingdom and New York, NY, USA, 2013. Oueslati, B. and Bellon, G.: The double ITCZ bias in CMIP5 models: interaction between SST, large-scale cir- culation and precipitation, Clim. Dynam., 44, 585–607, https://doi.org/10.1007/s00382-015-2468-6, 2015. McFiggans, G., Artaxo, P., Baltensperger, U., Coe, H., Facchini, M. C., Feingold, G., Fuzzi, S., Gysel, M., Laaksonen, A., Lohmann, U., Mentel, T. F., Murphy, D. M., O’Dowd, C. D., Snider, J. R., and Weingartner, E.: The effect of physical and chemical aerosol properties on warm cloud droplet activation, Atmos. Chem. Phys., 6, 2593–2649, https://doi.org/10.5194/acp-6-2593- 2006, 2006. p g Park, S.: A unﬁed convection scheme (UNICON), Part I: Formula- tion, J. Atmos. Sci., 71, 3902–3930, https://doi.org/10.1175/JAS- D-13-0233.1, 2014. Meier, W. N., Stroeve, J., Barrett, A., and Fetterer, F.: A simple approach to providing a more consistent Arctic sea ice extent time series from the 1950s to present, The Cryosphere, 6, 1359– 1368, https://doi.org/10.5194/tc-6-1359-2012, 2012. Pawlowska, H. and Brenguier, J.-L.: An observational study of driz- zle formation in stratocumulus clouds for general circulation model (GCM) parameterizations, J. Geophys. Res., 108, 8630, https://doi.org/10.1029/2002JD002679, 2003. Phillips, N. A.: The general circulation of the atmosphere: A numerical experiment, Q. J. Roy. Meteor. Soc., 82, 123–164, https://doi.org/10.1002/qj.49708235202, 1956. Miller, R. L., Cakmur, R. V., Perlwitz, J., Geogdzhayev, I. V., Ginoux, P., Kohfeld, K. E., Koch, D., Prigent, C., Ruedy, R., Schmidt, G. A., and Tegen, I.: Mineral dust aerosols in the NASA Goddard Institute for Space Sciences ModelE at- mospheric general circulation model, J. Geophys. Res., 111, D06208, https://doi.org/10.1029/2005JD005796, 2006. Pithan, F., Ackerman, A., Angevine, W. M., Hartung, K., Ickes, L., Kelley, M., Medeiros, B., Sandu, I., Steeneveld, G.-J., Sterk, H., Svensson, G., Vaillancourt, P. A., and Zadra, A.: Se- lect strengths and biases of models in representing the Arctic winter boundary layer: The Larcform 1 single column model intercomparison, J. Adv. Model. Earth Syst., 8, 1345–1357, https://doi.org/10.1002/2016ms000630, 2016. Miller, R. L., Schmidt, G. A., Nazarenko, L. S., Tausnev, N., Ruedy, R., Kelley, M., Lo, K. K., Aleinov, I., Bauer, M., Bauer, S., Bleck, R., Canuto, V., Cheng, Y., Clune, T. L., Del Genio, A., Faluvegi, G., Hansen, J. E., Healy, R. J., Kiang, N. Y., Koch, D., Lacis, A. A., LeGrande, A. N., Lerner, J., Menon, S., Oinas, V., Perlwitz, J., Puma, M. G. A. Schmidt et al.: US climate model tuning J., Rind, D., Romanou, A., Russell, G. L., Sato, M., Shindell, D. T., Sun, S., Tsigaridis, K., Unger, N., Voulgarakis, A., Yao, M.- S., and Zhang, J.: CMIP5 Historical Simulations (1850–2012) With GISS ModelE2, J. Adv. Model. Earth Syst., 6, 441–477, https://doi.org/10.1002/2013MS000266, 2014. Reichler, T. and Kim, J.: How Well do Coupled Models Simu- late Today’s Climate?, B. Am. Meteorol. Soc., 89, 303–311, https://doi.org/10.1175/BAMS-89-3-303, 2008. Richardson, M., Cowtan, K., Hawkins, E., and Stolpe, M. B.: Rec- onciled climate response estimates from climate models and the energy budget of Earth, Nature Climate Change, 6, 931–935, https://doi.org/10.1038/nclimate3066, 2016. Ming, Y., Ramaswamy, V., Ginoux, P. A., Horowitz, L. W., and Russell, L.: Geophysical Fluid Dynamics Laboratory general cir- culation model investigation of the indirect radiative effects of anthropogenic sulphate aerosol, J. Geophys. Res., 110, D22206, https://doi.org/10.1029/2005JD006161, 2005. Rienecker, M., Suarez, M., Todling, R., J. Bacmeister, L. T., Liu, H.-C., Gu, W., Sienkiewicz, M., Koster, R. D., Gelaro, R., Sta- jner, I., and Nielsen, J.: The GEOS-5 Data Assimilation System – Documentation of versions 5.0.1 and 5.1.0, and 5.2.0., Tech. rep., 2008. Ming, Y., Ramaswamy, V., Donner, L., and Phillips, V.: A new pa- rameterization of cloud droplet activation applicable to general circulation models, J. Atmos. Sci., 63, 1348–1356, 2006. Ringler, T., Petersen, M., Higdon, R. L., Jacobsen, D., Jones, P. W., and Maltrud, M.: A multi-resolution approach to global ocean modeling, Ocean Model., 69, 211–232, https://doi.org/10.1016/j.ocemod.2013.04.010, 2013. Molod, A.: Constraints on the Proﬁles of Total Water PDF in AGCMs from AIRS and a High-Resolution Model, J. Climate, 25, 8341–8352, https://doi.org/10.1175/jcli-d-11-00412.1, 2012. Rotstayn, L.: A physically based scheme for the treatment of strat- iform clouds and precipitation in large-scale models. I. Descrip- tion and evaluation of microphysical processes, Q. J. Roy. Me- teor. Soc., 123, 1227–1282, 1997. Molod, A., Suarez, M., and Partyka, G.: The impact of limiting ocean roughness on GEOS-5 AGCM tropi- cal cyclone forecasts, Geophys. Res. Lett., 40, 411–416, https://doi.org/10.1029/2012gl053979, 2013. Saha, S., Nadiga, S., Thiaw, C., Wang, J., Wang, W., Zhang, Q., den Dool, H. M. V., Pan, H.-L., Moorthi, S., Behringer, D., Stokes, D., Peña, M., Lord, S., White, G., Ebisuzaki, W., Peng, P., and Xie, P.: The NCEP Climate Forecast System, J. Climate, 19, 3483–3517, https://doi.org/10.1175/jcli3812.1, 2006. Molod, A., Takacs, L., Suarez, M., and Bacmeister, J.: Development of the GEOS-5 atmospheric general circulation model: evolution from MERRA to MERRA2, Geosci. Model Dev., 8, 1339–1356, https://doi.org/10.5194/gmd-8-1339-2015, 2015. G. A. Schmidt et al.: US climate model tuning Sci., 26, 786–789, https://doi.org/10.1175/1520- 0469(1969)026<0786:ccwaco>2.0.co;2, 1969. Jones, P. D., Lister, D. H., Osborn, T. J., Harpham, C., Salmon, M., and Morice, C. P.: Hemispheric and large-scale land- surface air temperature variations: An extensive revision and an update to 2010, J. Geophys. Res.-Atmos., 117, D05127, https://doi.org/10.1029/2011jd017139, 2012. Masson, D. and Knutti, R.: Climate model genealogy, Geophys. Res. Lett., 38, L08703, https://doi.org/10.1029/2011GL046864, 2011. Kiehl, J. T.: Twentieth century climate model response and climate sensitivity, Geophys. Res. Lett., 34, L22710, https://doi.org/10.1029/2007GL031383, 2007. Mauritsen, T., Stevens, B., Roeckner, E., Crueger, T., Esch, M., Giorgetta, M., Haak, H., Jungclaus, J., Klocke, D., Matei, D., Mikolajewicz, U., Notz, D., Pincus, R., Schmidt, H., and Tomassini, L.: Tuning the climate of a Kim, D., Sobel, A. H., Del Genio, A. D., Chen, Y.-H., Ca- margo, S. J., Yao, M. S., Kelley, M., and Nazarenko, L.: Geosci. Model Dev., 10, 3207–3223, 2017 www.geosci-model-dev.net/10/3207/2017/ G. A. Schmidt et al.: US climate model tuning 3223 cast System Reanalysis, B. Am. Meteorol. Soc., 91, 1015–1057, https://doi.org/10.1175/2010bams3001.1, 2010. cast System Reanalysis, B. Am. Meteorol. Soc., 91, 1015–1057, https://doi.org/10.1175/2010bams3001.1, 2010. Suzuki, K., Golaz, J.-C., and Stephens, G.: Evaluating cloud tun- ing in a climate model with satellite observations, Geophys. Res. Lett., 40, 4464–4468, https://doi.org/10.1002/grl.50874, 2013. Saha, S., Moorthi, S., Wu, X., Wang, J., Nadiga, S., Tripp, P., Behringer, D., Hou, Y.-T., ya Chuang, H., Iredell, M., Ek, M., Meng, J., Yang, R., Mendez, M. P., van den Dool, H., Zhang, Q., Wang, W., Chen, M., and Becker, E.: The NCEP Cli- mate Forecast System Version 2, J. Climate, 27, 2185–2208, https://doi.org/10.1175/jcli-d-12-00823.1, 2014. Thompson, D. W. J., Kennedy, J. J., Wallace, J. M., and Jones, P. D.: A large discontinuity in the mid-twentieth century in ob- served global-mean surface temperature, Nature, 453, 646–649, https://doi.org/10.1038/nature06982, 2008. Tiedtke, M.: Representation of clouds in large-scale models, Mon. Weather Rev., 40, 3040–3061, https://doi.org/10.1175/1520- 0493(1993)121<3040:ROCILS>2.0.CO;2, 1993. Schmidt, G. A. and Sherwood, S.: A practical philosophy of com- plex climate modelling, European Journal for Philosophy of Science, 5, 149–169, https://doi.org/10.1007/s13194-014-0102- 9, 2014. von Schuckmann, K., Palmer, M. D., Trenberth, K. E., Cazenave, A., Chambers, D., Champollion, N., Hansen, J., Josey, S. A., Loeb, N., Mathieu, P.-P., Meyssignac, B., and Wild, M.: An imperative to monitor Earth’s energy imbalance, Nature Cli- mate Change, 6, 138–144, https://doi.org/10.1038/nclimate2876, 2016. Schmidt, G. A., Ruedy, R., Hansen, J. E., Aleinov, I., Bell, N., Bauer, M., Bauer, S., Cairns, B., Canuto, V., Cheng, Y., Del Ge- nio, A., Faluvegi, G., Friend, A. D., Hall, T. M., Hu, Y., Kelley, M., Kiang, N. Y., Koch, D., Lacis, A. A., Lerner, J., Lo, K. K., Miller, R. L., Nazarenko, L., Oinas, V., Perlwitz, J., Perlwitz, J., Rind, D., Romanou, A., Russell, G. L., Sato, M., Shindell, D. T., Stone, P. H., Sun, S., Tausnev, N., Thresher, D., and Yao, M.-S.: Present day atmospheric simulations using GISS ModelE: Com- parison to in-situ, satellite and reanalysis data, J. Climate, 19, 153–192, https://doi.org/10.1175/JCLI3612.1, 2006. Walsh, J. E., Fetterer, F., Stewart, J. S., and Chapman, W. L.: A database for depicting Arctic sea ice variations back to 1850, Geogr. Rev., 107, 89–107, 7 https://doi.org/10.1111/j.1931- 0846.2016.12195.x, 2017. Winsberg, E.: Values and Uncertainties in the Predictions of Global Climate Models, Kennedy Inst. Ethic. J., 22, 111–137, https://doi.org/10.1353/ken.2012.0008, 2012. Schmidt, G. A., Kelley, M., Nazarenko, L., Ruedy, R., Russell, G. L., Aleinov, I., Bauer, M., Bauer, S., Bhat, M. G. A. Schmidt et al.: US climate model tuning Saha, S., Moorthi, S., Pan, H.-L., Wu, X., Wang, J., Nadiga, S., Tripp, P., Kistler, R., Woollen, J., Behringer, D., Liu, H., Stokes, D., Grumbine, R., Gayno, G., Wang, J., Hou, Y.-T., Chuang, H.- Y., Juang, H.-M. H., Sela, J., Iredell, M., Treadon, R., Kleist, D., Delst, P. V., Keyser, D., Derber, J., Ek, M., Meng, J., Wei, H., Yang, R., Lord, S., Dool, H. V. D., Kumar, A., Wang, W., Long, C., Chelliah, M., Xue, Y., Huang, B., Schemm, J.-K., Ebisuzaki, W., Lin, R., Xie, P., Chen, M., Zhou, S., Higgins, W., Zou, C.-Z., Liu, Q., Chen, Y., Han, Y., Cucurull, L., Reynolds, R. W., Rutledge, G., and Goldberg, M.: The NCEP Climate Fore- Myhre, G., Shindell, D., Bréon, F.-M., Collins, W., Fuglestvedt, J., Huang, J., Koch, D., Lamarque, J.-F., Lee, D., Mendoza, B., Nakajima, T., Robock, A., Stephens, G., Takemura, T., and Zhang, H.: Anthropogenic and Natural Radiative Forcing, in: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Inter- governmental Panel on Climate Change, edited by: Stocker, T. F., Qin, D., Plattner, G.-K., Tignor, M., Allen, S. K., Boschung, J., Nauels, A., Xia, Y., Bex, V., and Midgley, P., Cambridge Uni- D., Delst, P. V., Keyser, D., Derber, J., Ek, M., Meng, J., Wei, Geosci. Model Dev., 10, 3207–3223, 2017 www.geosci-model-dev.net/10/3207/2017/ G. A. Schmidt et al.: US climate model tuning K., Bleck, R., Canuto, V., Chen, Y., Cheng, Y., Clune, T. L., Del Ge- nio, A., de Fainchtein, R., Faluvegi, G., Hansen, J. E., Healy, R. J., Kiang, N. Y., Koch, D., Lacis, A. A., LeGrande, A. N., Lerner, J., Lo, K. K., Matthews, E. E., Menon, S., Miller, R. L., Oinas, V., Oloso, A. O., Perlwitz, J., Puma, M. J., Putman, W. M., Rind, D., Romanou, A., Sato, M., Shindell, D. T., Sun, S., Syed, R. A., Tausnev, N., Tsigaridis, K., Unger, N., Voulgarakis, A., Yao, M.-S., and Zhang, J.: Conﬁgura- tion and assessment of the GISS ModelE2 contributions to the CMIP5 archive, J. Adv. Model. Earth Syst., 6, 141–184, https://doi.org/10.1002/2013MS000265, 2014. Yokohata, T., Annan, J. D., Collins, M., Jackson, C. S., Tobis, M., Webb, M. J., and Hargreaves, J. C.: Reliability of multi-model and structurally different single-model ensembles, Clim. Dy- nam., 39, 599–616, https://doi.org/10.1007/s00382-011-1203-1, 2012. Zhao, M., Golaz, J.-C., Held, I. M., Ramaswamy, V., Lin, S.-J., Ming, Y., Ginoux, P., Wyman, B., Donner, L. J., Paynter, D., and Guo, H.: Uncertainty in Model Climate Sensitivity Traced to Representations of Cumulus Precipitation Microphysics, J. Climate, 29, 543–560, https://doi.org/10.1175/jcli-d-15-0191.1, 2016. Zweng, M. M., Reagan, J. R., Antonov, J. I., Locarnini, R. A., Mis- honov, A. V., Boyer, T. P., Garcia, H. E., Baranova, O. K., John- son, D. R., Seidov, D., and Biddle, M. M.: World Ocean Atlas 2013, Volume 2: Salinity, NOAA Atlas NESDIS 74, U.S. Gov- ernment Printing Ofﬁce, Washington, D.C., edited by: Levitus, S. and Mishonov, A., 2013. Shindell, D. T., Pechony, O., Voulgarakis, A., Faluvegi, G., Nazarenko, L., Lamarque, J.-F., Bowman, K., Milly, G., Ko- vari, B., Ruedy, R., and Schmidt, G. A.: Interactive ozone and methane chemistry in GISS-E2 historical and future climate simulations, Atmos. Chem. Phys., 13, 2653–2689, https://doi.org/10.5194/acp-13-2653-2013, 2013. Siebesma, A. P. and Cuijpers, J. W. M.: Evaluation of Parametric Assumptions for Shallow Cumulus Convection, J. Atmos. Sci., 52, 650–666, https://doi.org/10.1175/1520- 0469(1995)052<0650:eopafs>2.0.co;2, 1995. Geosci. Model Dev., 10, 3207–3223, 2017 www.geosci-model-dev.net/10/3207/2017/
https://openalex.org/W2155601949	https://hal.inria.fr/hal-01583635/document	English	null	Storage-Enabled Access Points for Improved Mobile Performance: An Evaluation Study	Lecture notes in computer science	2,011	cc-by	5,302	To cite this version: Efthymios Koutsogiannis, Lefteris Mamatas, Ioannis Psaras. Storage-Enabled Access Points for Im- proved Mobile Performance: An Evaluation Study. 9th Wired/Wireless Internet Communications (WWIC), Jun 2011, Vilanova i la Geltrú, Spain. pp.116-127, 10.1007/978-3-642-21560-5_10. hal- 01583635 Distributed under a Creative Commons Attribution 4.0 International License ⋆The research leading to these results has received funding from the European Com- munity’s Seventh Framework Programme (FP7/2007-2013 FP7-REGPOT-2010-1, SP4 Capacities, Coordination and Support Actions) under grant agreement no 264226 (project title: Space Internetworking Center-SPICE ). Storage-Enabled Access Points for Improved Mobile Performance: An Evaluation Study ⋆ Efthymios Koutsogiannis1, Lefteris Mamatas1, and Ioannis Psaras2 1 Democritus University of Thrace, Dept. of Electrical and Computer Engineering/Space Internetworking Center, 12 Vas. Soﬁas Str., 67100 Xanthi, Greece {ekoutsog, emamatas}@ee.duth.gr 2 University College London, Dept. of Electronic and Electrical Engineering, WC1E 7JE, Torrington Place, London, UK i.psaras@ee.ucl.ac.uk 1 Democritus University of Thrace, Dept. of Electrical and Computer Engineering/Space Internetworking Center, 12 Vas. Soﬁas Str., 67100 Xanthi, Greece {ekoutsog, emamatas}@ee.duth.gr 2 University College London, Dept. of Electronic and Electrical Engineering, WC1E 7JE, Torrington Place, London, UK i.psaras@ee.ucl.ac.uk Abstract. Due to existing networking paradigms and people communi- cation habits, Internet is characterized by: (i) information that associates more and more with network edges, while corresponding traﬃc doesn’t, (ii) a considerable amount of unexplored network resources that resides near mobile users, while their connectivity is usually poor and expensive in terms of ﬁnancial cost and resource consumption. To address the above challenges, we introduce a cooperative networking scheme where Home Access Points are equipped with storage capabilities and oﬀer connectivity to mobile users. Whenever connectivity sharing is unavailable, the access points store the pending data and transmit it once connectivity becomes available. We explore how storage-enabled opportunistic routing algorithms can handle periods of intermittent con- nectivity. We experiment with realistic scenarios based on real network conﬁgurations and mobility maps. Our results demonstrate the poten- tial of the proposed paradigm as well as guide us to deﬁne its follow-up evolution steps. Keywords: user-centric networking, delay-tolerant networking, oppor- tunistic routing HAL Id: hal-01583635 https://inria.hal.science/hal-01583635v1 Submitted on 7 Sep 2017 L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. Distributed under a Creative Commons Attribution 4.0 International License 1 Introduction Research on Opportunistic Mobile Networks and DTNs has been mainly con- cerned with storage and energy constraints for mobile devices and optimiza- tion of connectivity opportunities between encounters. The more the encoun- ters and data-exchanges between the mobile nodes, the more the energy and storage needed. The Delay-Tolerant and Opportunistic Networking research has focused mainly on mobile-only devices. That is, researchers have been investigat- ing ways to optimize communication opportunities between the mobile devices themselves3. This is indeed the case in rural areas, for instance, where connectiv- ity points to the network are rare, or in case of applications that do not require Internet connectivity (e.g., distributed online social networks [1]). In this study, we focus on metropolitan environments, where there are plenty of connectivity points to the Internet. We extend the operational spectrum of Opportunistic Networks and DTNs and consider Home Access Points (APs or HAPs) as integral part of these networks. We consider storage a cheap resource that can easily be integrated in future APs and evaluate the performance of existing DTN protocols in this setting. In particular, given that connectivity opportunities to APs/HAPs and therefore the Internet now increase, we inves- tigate whether and to which extend the storage and energy constraints can now be somewhat relaxed. We motivate our study based on the following facts: – New applications and people communication habits call for approaches that go beyond the traditional client-server model. Users are increasingly active in uploading content to social networking sites, for instance, or hosting data, e.g., in P2P ﬁle-sharing platforms. Moreover, it is shown that Internet traﬃc is growing faster at the network edge than at the core [2]. In other words, future networks call for user-centricity, where applications demand that con- tent is closer to the user. – Internet traﬃc forecasts show that mobile users are expected to double-up every year through 20144. Similar studies show that the amount of user- generated content is going to increase massively5 and furthermore, that most of it will be generated by mobile devices, e.g., smart-phones et simila. This urges the need for easy access to the Internet for mobile users, since content has to be readily available. – Statistics show that the vast majority of households (e.g., 70% in UK) own a broadband connection6, which remains unexploited most of the day. 6 http://www.statistics.gov.uk/cci/nugget.asp?ID=8 5 http://newsroom.cisco.com/dlls/2009/prod 102109.html 6 1 Introduction It is obvious that an area of dense connectivity can be formed if the HAPs’ network and storage resources are being shared between home- and mobile- users. – User-Provided Networks and connectivity sharing schemes have already been investigated and implemented (e.g., FON [3]). In our previous studies [4], [5] we have presented algorithms for connectivity sharing among home- and guest-users and proved that careful design can guarantee seamless connec- tivity sharing for the home-user. 3 Vehicular DTNs [6] and Infostations [7] can be considered as exceptions here, since these deployments are not constrained by energy and/or storage depletion in most cases 4 http://www.cisco.com/en/US/netsol/ns827/networking solutions sub solution.html 5 5 http://newsroom.cisco.com/dlls/2009/prod 102109.htm In this study, we explore the performance beneﬁts of extending the users’ connectivity graph to include neighbor HAPs, exploiting this way local unused resources. An occupied HAP that temporarily cannot forward packets from guest- users to the Internet may store the packets until a connection opportunity is available. The stored packets can be forwarded to another HAP in range, or relayed to a mobile node that may ﬁnd another connectivity point. We focus on the evaluation of DTN routing protocols, when the opportunistic network infrastructure is extended to also include HAPs, whenever and wherever these are available. We have built a realistic scenario, where mobile users walk around the center of London. We assume that existing FON subscribers (i.e., we used the database from maps.btfon.com) employ the proposed paradigm. We apply diﬀerent usage scenarios to the home-users (e.g., periodic use, away for the day, away for work) in order to emulate their on-oﬀbehavior in terms of connectivity oﬀering. We assume that whenever the home-user is transmit- ting/receiving data the HAP is not available to the guest user. Therefore, the connection is available for guests whenever the home connection is idle. At this stage, we explore how existing opportunistic routing algorithms behave in this particular context; we evaluate the following well-known algorithms: Epidemic [8], First contact [9], Spray and Focus [10], Spray and Wait [11] and MaxProp [12]. We elaborate on these results in order to draw design guidelines for new algorithms, especially designed and optimized for the proposed paradigm. 1 Introduction Our target is to identify the necessary properties that need to be integrated into op- portunistic routing protocols in order to operate as eﬃciently as possible under this new setting, i.e., when storage resources are part of the HAP and are oﬀered to mobile users. We leave these designs as our immediate future work task. Our ﬁndings indicate that: (i) integrating storage capabilities to HAPs sub- stantially increases connectivity opportunities for mobile nodes, (ii) service qual- ity for mobile users increases with the number of available HAPs and their respective storage resources, (iii) there is a lot of space for improvement for opportunistic routing protocols, performance-wise, (iv) among the evaluated al- gorithms Spray and Focus is the most eﬃcient, since it balances better between performance and overhead, and (v) an eﬃcient routing algorithm for this new networking paradigm could oﬄoad routing sophistication or routing state info from mobile devices to HAPs, in order to reduce energy consumption at mobile users, without compromising performance. In Section 2, we discuss the related work in terms of other broadband access sharing schemes and opportunistic routing algorithms. In Section 3, we describe our experimental setup and the associated methodology and in Section 4 we present our experimental results. We conclude the paper in Section 5, where we also discuss our future work. 2.1 State of the Art User-Provided Networks. UPNs have been recently proposed [13] as an um- brella framework for Broadband Access Sharing. Citywide Ubiquitous Wi-Fi Access sharing issues have been investigated in the past [14] and have shown that commercial deployment of such initiatives is indeed possible. p y p Famous commercial deployments include the FON Community, the OpenSpark Community and the Wiﬁ.com7. Each of these deployments has a diﬀerent start- ing point, incentive-wise, and therefore, comprises diﬀerent scalability and ef- ﬁciency properties. For instance, Wiﬁ.com is based on a social networking ap- proach, while FON is based on a business model that explicitly allocates re- sources to home- and guest-users. Recent studies have also investigated the wireless interface’s capabilities to share resources [15]. The success of the above initiatives clearly shows that users are keen on giving away a portion of their bandwidth in order to have access when out of home or oﬃce. In our opinion, whatever the Broadband Access Sharing scheme, it has to balance between two fragile service-points: (i) the guarantee to the home- user for seamless sharing, performance-wise and (ii) the mobile-users’ quality of experience. The former point has been our focus in previous studies [4], [5]. The latter constitutes our target in the present study, where we elaborate on the functionality of an eﬃcient routing mechanism in order to achieve lower latency and reduced transmission redundancy. Hence, our target is increased responsiveness, eﬃcient storage management and longer battery life for mobile devices. Opportunistic Networks and DTNs. A lot of research has taken place recently to deal with routing issues in intermittently connected networks. Here, we brieﬂy present the most prevalent DTN routing algorithms. In Epidemic routing algorithm [8], a node copies a message to every other node it encounters. Variations of this algorithm include Randomized ﬂooding and Utility-based ﬂooding. Randomized ﬂooding or Gossiping (random-ﬂood) is quite similar to Epidemic although each message is copied with probability p < 1. In the Utility-based ﬂooding, a message is copied only if the node encountered has a utility value higher than the current by some threshold Uth. MaxProp [12] is another ﬂooding-based mechanism which removes the mes- sages once a copy is delivered. Moreover, MaxProp prioritizes both messages to be transmitted and dropped based on previous encounters. Resource Allocation Protocol for Intentional DTN (RAPID) [16] treats DTN routing as a resource allocation problem. 2 Related work Broadly speaking, our study balances between User-Provided Networks and Op- portunistic or Delay-Tolerant-Networks. Below, we brieﬂy provide background, state of the art information on these areas and highlight how the present study advances research beyond the state of the art. 7 URLs: http://www.btfon.com, http://open.spark.ﬁ, http://www.wiﬁ.com 2.2 Progress Beyond the State of the Art To release the boundaries that keep these two research ﬁelds separated and unleash the communication opportunities of the corresponding framework, we highlight the following: – UPNs exploit connectivity sharing opportunities only. Broadband access is only one of the available resources that can be shared between home and mo- bile users. In this study, we explore the potential of sharing storage resources as well, given that storage is available at the HAPs. That is, we extend the sharing capabilities of UPNs to include storage resources as well. – DTN routing research has been mainly concerned with encounters between mobile nodes only. Although in case of DTN nodes, both connectivity and storage resources are shared between the nodes, they do not consider collab- orating with ﬁxed points of the infrastructure. Here, we extend this trend to a boundless framework, where mobile nodes can also take advantage of both connectivity and storage resources of HAPs. 2.1 State of the Art It uses a utility function that: (i) as- signs a value based on the metric being optimized to every packet and (ii) ﬁrst replicates packets that increase the utility function. Other approaches use a controlled replication or spraying. In Spray and Wait [11], a small ﬁxed number of copies are distributed (sprayed) to diﬀerent relays (i.e., the ﬁrst few encountered). Each relay waits until it encounters the destina- tion itself. A more sophisticated variation is Spray and Focus [10] which operates like Spray and Wait during spraying. Each relay, instead of waiting, can forward its copy to a potentially more appropriate relay, using a utility-based scheme. For a complete survey on DTN routing protocols, we refer the reader to [17] and [18]. In addition, related work on exploiting WLAN hotspots in order to enhance opportunistic web access and oﬄoad cellular traﬃc can be found in [19] and [20], respectively. 3 Experimental Setup and Methodology We simulate a realistic mobility setup. We have chosen a geographical area in the center of London sized 2160m x 1600m, see Fig. 1. This area covers 349 diﬀerent streets and 1876 landmarks. We have gathered detailed information from [3] regarding the exact position and the number of BTFON users in this particular area (i.e., 118 users). We parsed all this information into the ONE simulator [21]. We extended the ONE simulator with better support for wired networks and two new types of nodes: (i) the home-user that owns a HAP, and (ii) a server node, which hosts data for the mobile users (e.g., social proﬁle photos). We assume that all home-users are permanently connected to a server node (e.g., a server farm) situated in Gower Street (see Fig. 1). Mobile nodes are moving in the area and transmit periodically data (every 60 to 80 seconds, according to a uniform distribution) to the server node. Message sizes follow the uniform distribution [100kB, 200kB] (e.g., a typical social proﬁle photo). Their movement is restricted to the streets in the map, according to the map route movement model and the shortest path algorithm in ONE simulator. All nodes (mobile and static) have storage capabilities, but static nodes, i.e., HAPs, comprise extensive storage resources. In our setup, mobile nodes and HAPs have 30MB of storage available. These are only indicative ﬁgures and serve as pointers for future settings. As already mentioned, the HAP is only available to mobile users, whenever the home-user is idle. This practically means that when the home-user is using his Internet connection, only storage capabilities are oﬀered to the mobile nodes within range. We assume three representative home-user proﬁles: (i) users that are idle for 8 hours per day only (e.g., during working hours), (ii) users that are permanently idle (e.g., they are absent), and (iii) users that are only periodically using their connection throughout the day (i.e., according to a normal probability distribution: 1-3 hours at home and 1- 2 hours away). Nodes periodically transmit beacons to recognize each other’s presence. Fig. 1, shows our experimental setup. The server is referenced as s- node, mobile-guests as g-nodes and home-users as h-nodes. At this stage of investigation, we evaluate how existing opportunistic routing algorithms behave in this particular scenario, where HAPs oﬀer storage resources apart from only connectivity. 3 Experimental Setup and Methodology We selected ﬁve representative storage-enabled algorithms, namely: Epidemic [8], First contact [9], Spray and Wait [11], Spray and Focus [10] and MaxProp [12]. Our target is to identify key requirements and design guidelines for a customized algorithm that exploits eﬃciently not only connection opportunities, but storage resources as well. Since a crucial aspect is to maximize energy eﬃciency of the mobile nodes, a maximum data delivery ratio should be combined with minimum communication / storage overhead and latency. In this context, we measure the following: – OverheadRatio = P acketsRelayed−P acketsReceived P acketsReceived , which captures packet de- livery ratio with respect to the number of packets relayed. – OverheadRatio = P acketsRelayed−P acketsReceived P acketsReceived , which captures packet de- livery ratio with respect to the number of packets relayed. – AverageLatency, which captures the average packet latency that increases with the number of hops the packet crosses. – AverageBufferTime, which reﬂects the average time the packets are buﬀered, in order to evaluate protocol storage eﬃciency. 8 Epidemic protocol was selected for these scenarios and more speciﬁcally Epidemi- cOracle version, which upon message delivery removes that message from all nodes. 4.1 Scenario 1: Impact of Home-Users’ Number. The purpose is to show the advantages of the proposed framework, i.e., the ex- tension of mobile network’s boundaries to also include storage-equipped HAPs. In this ﬁrst scenario, we explore the performance of a representative oppor- tunistic routing protocol and more speciﬁcally Epidemic with 100 mobile-users, where the availability of HAPs is gradually increased (0-100). Home-users can be perceived as points of guaranteed message delivery to the destination server. We observe in Fig. 2(a), that as the number of home-users increases, the aver- age Latency declines, since less time is needed to reach a connectivity point to the Internet. Consequently, the average buﬀering time declines as well, as is shown in Fig. 2(b), owing to timely message delivery. The peak in measurements (i.e., Figures 2(a), 2(b)) for 5 home-users can be justiﬁed as follows: most of the ﬁrst ﬁve HAPs in our database are situated in peripheral streets. Hence, they have low interaction probability with mobile nodes and consequently they do not get involved in the replication process. Therefore, average buﬀering time increases (Fig. 2(b)), since mobile nodes need to carry data for more time. Clearly, there are still more delivery probabilities than in the case of zero home-users (Fig. 2(a)), where the server is the only data delivery point. In conclusion, increasing the availability of storage-equipped HAPs enhances network’s connectivity, thus average Latency and Buﬀering time decrease. The purpose is to show the advantages of the proposed framework, i.e., the ex- tension of mobile network’s boundaries to also include storage-equipped HAPs. In this ﬁrst scenario, we explore the performance of a representative oppor- tunistic routing protocol and more speciﬁcally Epidemic with 100 mobile-users, where the availability of HAPs is gradually increased (0-100). Home-users can be perceived as points of guaranteed message delivery to the destination server. We observe in Fig. 2(a), that as the number of home-users increases, the aver- age Latency declines, since less time is needed to reach a connectivity point to the Internet. Consequently, the average buﬀering time declines as well, as is shown in Fig. 2(b), owing to timely message delivery. The peak in measurements (i.e., Figures 2(a), 2(b)) for 5 home-users can be justiﬁed as follows: most of the ﬁrst ﬁve HAPs in our database are situated in peripheral streets. Hence, they have low interaction probability with mobile nodes and consequently they do not get involved in the replication process. 4 Experimental Results The experimental scenarios presented in the next section were designed in order to allow for: – Studying the impact of the proposed extended networking scheme on the performance of a representative opportunistic routing algorithm. In this con- text, the Epidemic8 protocol was evaluated in various deployment levels by increasing the number of HAPs in Scenario 1. In Scenario 2, evaluation is based on diﬀerent usage proﬁles that home-users adopt. 8 Epidemic protocol was selected for these scenarios and more speciﬁcally Epidemi- cOracle version, which upon message delivery removes that message from all nodes. Fig. 1. Experimental Scenario Fig. 1. Experimental Scenario – Evaluating well-known DTN routing protocols’ functionality used for com- munication between mobile nodes and static infrastructure. More speciﬁ- cally, we study their performance while scaling the proposed scheme with respect to mobile nodes’ number, in Scenario 3. In scenario 4, we focus on the eﬃciency of the protocols with regards to guest-users’ speed. – Evaluating well-known DTN routing protocols’ functionality used for com- munication between mobile nodes and static infrastructure. More speciﬁ- cally, we study their performance while scaling the proposed scheme with respect to mobile nodes’ number, in Scenario 3. In scenario 4, we focus on the eﬃciency of the protocols with regards to guest-users’ speed. 4.1 Scenario 1: Impact of Home-Users’ Number. Therefore, average buﬀering time increases (Fig. 2(b)), since mobile nodes need to carry data for more time. Clearly, there are still more delivery probabilities than in the case of zero home-users (Fig. 2(a)), where the server is the only data delivery point. In conclusion, increasing the availability of storage-equipped HAPs enhances network’s connectivity, thus average Latency and Buﬀering time decrease. (a) Average Latency (b) Average Buﬀering Time Fig. 2. Scenario 1: Impact of Home-Users’ Number (a) Average Latency (b) Average Buﬀering Time (b) Average Buﬀering Time (a) Average Latency Fig. 2. Scenario 1: Impact of Home-Users’ Number 4.2 Scenario 2: Impact of Home-Users’ Usage-Proﬁles. Here, each home-user adopts one of the three usage-proﬁles described in Section 3.1. This scenario intends to study the behavior of a representative opportunistic routing protocol in a topology consisting of 118 home-users with diﬀerent usage- proﬁles. We discriminate the cases according to home-users’ percentage that follows each of the three usage proﬁles: A) away for 8 hours, B) always away, C) periodically away (1-3 hours at home, then 1-2 hours away). Moreover, we simulated the cases where (A, B, C) percentages are: i) (0, 100, 0), ii) (20, 40, 40), iii) (10, 10, 80), iv) (0, 0, 100), while guest-users vary (10-100). We conducted a series of experiments where increasing home users with usage proﬁles (i) and (iii), causes lack of connectivity, while increasing the number of guest-users allows for more routes and improved message delivery in terms of time. In these experiments, we observed that data remains stored at home- users available for: i) delivery to the server when connection will be up, ii) transmission to a guest-user in the vicinity, iii) transmission to a home-user in the vicinity. Fig. 3(b) presents average buﬀering time, which decreases as mobile nodes increase, while the four graphs coincide. Clearly, usage-proﬁle has small impact on buﬀering time, since the number of home-users is suﬃciently large. On the contrary, average Latency graphs vary, and their values increase when periodic nodes increase, while increasing guest-users decreases average Latency as depicted in Fig. 3(a). We report, although not present it here due to space limitations, that Overhead ratio is not inﬂuenced by usage-proﬁle. Usage-proﬁles appear to have signiﬁcant impact on average Latency, while Epidemic protocol manages to deliver messages, though with a high overhead. (a) Average Latency (b) Average Buﬀering Time Fig. 3. Scenario 2: Impact of Home-Users’ Usage-Proﬁles (b) Average Buﬀering Time (a) Average Latency (b) Average Buﬀering Time Fig. 3. Scenario 2: Impact of Home-Users’ Usage-Proﬁles 4.3 Scenario 3: Impact of Guest-Users’ Number. In this scenario, we aim to compare the functionality of all ﬁve protocols and highlight the most eﬃcient strategy, in the extended opportunistic networking paradigm introduced herein. The topology consists of 118 HAPs constantly avail- able, while mobile nodes vary (10-100). Spray and Wait introduces the highest average Buﬀering time (see Fig. 4(b)), which is not aﬀected by the number of guest-users, while Spray and Focus achieves the lowest. Both algorithms were conﬁgured with 6 message copies, therefore their behavior at spraying phase is identical; consequently, the focus phase is responsible for the signiﬁcant reduction of buﬀering time, as opposed to the wait phase. Fig. 4(a), depicts the average Latency, where Spray and Focus is again outperforming the rest of the protocols. First contact is the worst per- forming algorithm, since the ﬁrst encountering node has no guarantee to deliver the message immediately. The Overhead ratio in Fig. 4(c) evinces the eﬃciency of Spray and Focus mechanism compared to the others. As expected Epidemic did not perform well with regards to overhead owing to mere ﬂooding. Therefore, increasing guest-users’ number improves the average latency, ex- cept from First Contact. Furthermore, it also reduces average buﬀering time at the expense of overhead. In summary, regarding the impact of guest-users’ number, Spray and Focus is the most eﬃcient of the tested algorithms. 5 Conclusions We have proposed and investigated a cooperative networking scheme, in the context of User-Provided Networks, where storage resources are oﬀered to mo- bile users, until bandwidth at HAPs becomes available to transmit mobile users’ data. Storage-enabled nodes and access points are a key feature of our oppor- tunistic networking paradigm. We evaluated the impact of this scheme on the performance of opportunistic routing and compared the performance of ﬁve well- known DTN routing protocols in a realistic simulation setting. The proposed networking scheme avails in terms of latency and buﬀering time, although over- head ratio is sometimes increasing. Since mobile devices are involved, an eﬃcient routing protocol that achieves high delivery ratio and low overhead is crucial. We have concluded that Spray and Focus [10] is the most suitable for the afore- mentioned scheme, amongst the evaluated protocols. These preliminary results indicate that a new DTN routing protocol suitable for this environment can achieve improved overall performance. Its functionality should allow for reduced latency, communication and storage overhead, in con- junction with high delivery ratio, even in case of opportunistic routing (few and sparse HAPs). We argue that the basic guidelines for the design of the algorithm are: (i) Route data through HAPs with a higher probability than mobile nodes in order to alleviate the latter’s storage and energy constraints. (ii) Exploit- ing HAP density and oﬄoading sophistication or routing state info from mobile nodes to HAPs will further enhance performance e.g., via encounter prediction. (iii) Integration of Spray and Focus functionality in order to perform eﬃciently when mobile nodes are dominant in the topology. (iv) Dynamic conﬁguration of message redundancy according to feedback from encounters and time scales [22], will allow for low overhead in conjunction with the desired delivery rate. As a next step we intend to integrate these observations in a new protocol. 4.4 Scenario 4: Impact of Guest-Users’ Speed. The speed of mobile nodes, together with connectivity opportunities, the mes- sage size and links’ bandwidth are critical to protocol’s eﬃciency. Increased speed allows for more encounters with other nodes, although it reduces the commu- nication time. Using realistic settings, we evaluate the performance of the ﬁve (a) Average Latency (b) Average Buﬀering Time (c) Overhead Ratio (d) Average Latency (e) Average Buﬀering Time (f) Overhead Ratio Fig. 4. Scenario 3: (a), (b), (c), Scenario 4: (d), (e), (f) (a) Average Latency (b) Average Buﬀering Time (a) Average Latency (a) Average Latency (c) Overhead Ratio ( ) A B ﬀ i Ti (c) Overhead Ratio (d) Average Latency (d) Average Latency (d) Average Latency (f) Overhead Ratio (e) Average Buﬀering Time (f) Overhead Ratio Fig. 4. Scenario 3: (a), (b), (c), Scenario 4: (d), (e), (f) protocols with regards to speed diﬀerentiation, for 10 guest- (maximum sup- ported number for the speciﬁc conﬁguration) and 118 home-users available. Increasing speed (1 to 20 m/s) allows for increased data relaying due to more encounters and reduces the average buﬀering time (Fig. 4(e)). Spray and Focus has the best performance, even though it is the only algorithm whose buﬀering time increases with speed; this is due to utility function’s inability to adapt to increased encounters. This result shows the potential improvement of the speciﬁc algorithm in this context. Average latency in Fig. 4(d) is highly decreased for low speeds (5 m/s), while further speed increase does not aﬀect the results. Fig. 4(f) depicts the overhead ratio, where Epidemic routing has the highest overhead ratio in contrast to Spray and Focus. In conclusion, Spray and Focus is the best performing algorithm in this scenario with regards to all evaluation parameters. References 1. Xu, T., Chen, Y., Fu, X., Hui, P.: Twittering by cuckoo: decentralized and socio- aware online microblogging services. SIGCOMM Computer Communication Review 40 (4), 473–474 (August 2010). ( ) ( ) 2. Leavitt, N.: Network-Usage Changes Push Internet Traﬃc to the Edge. IEEE Com- puter Magazine 43(10), 13–15 (Oct. 2010). p g ( ), ( ) 3. BT FON FON Wireless Ltd and British Telecommunications. http://www.btfon.com. // 4. Psaras, I., Mamatas, L.: On Demand Connectivity Sharing: Queuing Manage- ment and Load Balancing for User-Provided Networks. Elsevier Computer Networks 4. Psaras, I., Mamatas, L.: On Demand Connectivity Sharing: Queuing Manage- ment and Load Balancing for User-Provided Networks. Elsevier Computer Networks (COMNET), Special Issue on Wireless for the Future Internet, accepted (2010). g U C p (COMNET), Special Issue on Wireless for the Future Internet, accepted (2010). ( ), p , p ( ) 5. Mamatas, L., Psaras, I., Pavlou, G.: Incentives and Algorithms for Broadband Ac- cess Sharing. 1st ACM Sigcomm Workshop on Home Networks (HomeNets), New Delhi, India (August 2010). , ( g ) 6. Bychkovsky, V. et al.: A Measurement Study of Vehicular Internet Access Using in Situ Wi-Fi Networks. In Proceedings of MobiCom, 50-61, New York, USA, (2006). 7. Goodman, D.J., Borras, J., Mandayam, N.B., Yates, R.D.: INFOSTATIONS: a new system model for data and messaging services. In IEEE VTC’97, volume 2, pages 969–973 (May 1997). ( y ) 8. Vahdat, A., Becker, D.: Epidemic routing for partially connected ad hoc networks. Technical Report CS-200006, Duke University (2000). 9. Fall, K.: A delay-tolerant network architecture for challenged internets. SIG- COMM ’03, ACM, New York, NY, USA, 27-34. DOI=10.1145/863955.863960 http://doi.acm.org/10.1145/863955.863960 10. Spyropoulos, T., Psounis, K., Raghavendra, C.S.: Spray and Focus: Eﬃcient Mobility-Assisted Routing for Heterogeneous and Correlated Mobility. Fifth Annual IEEE International Conference on Pervasive Computing and Communications (Per- Com), 79–85, White Plains, NY (19-23 March 2007). ) ( ) 11. Spyropoulos, T., Psounis, K., Raghavendra, C. S.: Spray and wait: Eﬃcient routing in intermittently connected mobile networks. In Proceedings of ACM SIGCOMM workshop on Delay Tolerant Networking (WDTN) (2005). y g ( ) ( ) 12. Burgess, J., Gallagher, B., Jensen, D., Levine, B. N.: MaxProp: Routing for Vehicle- Based Disruption-Tolerant Networks. In Proc. of IEEE INFOCOM (April 2006). 12. Burgess, J., Gallagher, B., Jensen, D., Levine, B. N.: MaxProp: Routing for Vehicle- Based Disruption-Tolerant Networks. In Proc. of IEEE INFOCOM (April 2006). 13. References Soﬁa, R., Mendes, P.: User-provided Networks: Consumer as Provider. IEEE Com- 13. Soﬁa, R., Mendes, P.: User-provided Networks: Consumer as Provider. IEEE Com- munications Magazine 46(12), 86–91 (December 2008). 13. Soﬁa, R., Mendes, P.: User-provided Networks: Consumer as Provide munications Magazine 46(12), 86–91 (December 2008). g ( ), ( ) 14. Thraves, C. et al.: Driving the Deployment of Citywide Ubiquitous WiFi Access. In SimulWorks (2008). 15. Solarski, M. et al.: An Experimental Evaluation of Urban Networking using IEEE 802.11 Technology. In Wireless Mesh, 1–10 (2006). 16. Balasubramanian, A., Levine, B.N., Venkataramani, A.: DTN routing as a resource allocation problem. In Proc. ACM SIGCOMM (August 2007). ( g ) 17. Small, T., Haas, Z.: Resource and performance tradeoﬀs in delay-tolerant wireless networks. In Proceedings of ACM SIGCOMM WDTN (2005). 18. Psaras, I., Wood, L., Tafazolli, R.: Delay-/Disruption-Tolerant Networking: State of the Art and Future Challenges Technical Report (2009), available at: http://www.ee.ucl.ac.uk/ uceeips/dtn-srv-ipsaras.pdf p // / p / p p 19. Pitk¨anen, M., K¨arkk¨ainen, T., Ott, J.: Opportunistic Web Access via WLAN Hotspots. In Proc. of PERCOM ’10, Mannheim, Germany, March 2010. 20. Han, B. et al.: Cellular traﬃc oﬄoading through opportunistic communications: a case study. In Proceedings of ACM CHANTS, 2010. 20. Han, B. et al.: Cellular traﬃc oﬄoading through oppor case study. In Proceedings of ACM CHANTS, 2010. y g , 21. Ker¨anen, A., Ott, J., K¨arkk¨ainen, T.: The ONE Simulator for DTN Protocol Eval- uation. SIMUTools ’09, Rome, Italy (2009). uation. SIMUTools ’09, Rome, Italy (2009). y ( ) 22. Karvo, J., Ott, J.: Time scales and delay-tolerant routing protocols. In Proceedings of the 3rd ACM CHANTS ’08, ACM, New York, NY, USA, 33–40 (2008).
https://openalex.org/W2800334041	https://www.zora.uzh.ch/id/eprint/170568/1/s12870-019-1705-2.pdf	English	null	Plant trichomes and a single gene<i>GLABRA1</i>contribute to insect community composition on field-grown<i>Arabidopsis thaliana</i>	bioRxiv (Cold Spring Harbor Laboratory)	2,018	cc-by	11,048	Zurich Open Repository and Archive University of Zurich University Library Strickhofstrasse 39 CH-8057 Zurich www.zora.uzh.ch University of Zurich University Library Strickhofstrasse 39 CH-8057 Zurich www.zora.uzh.ch Year: 2019 © The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Plant trichomes and a single gene GLABRA1 contribute to insect community composition on field-grown Arabidopsis thaliana Plant trichomes and a single gene GLABRA1 contribute to insect community composition on field-grown Arabidopsis thaliana Sato, Yasuhiro ; Shimizu-Inatsugi, Rie ; Yamazaki, Misako ; Shimizu, Kentaro K ; Nagano, Atsushi J DOI: https://doi.org/10.1186/s12870-019-1705-2 DOI: https://doi.org/10.1186/s12870-019-1705-2 DOI: https://doi.org/10.1186/s12870-019-1705-2 Posted at the Zurich Open Repository and Archive, University of Zurich ZORA URL: https://doi.org/10.5167/uzh-170568 Journal Article Published Version The following work is licensed under a Creative Commons: Attribution 4.0 Internation ollowing work is licensed under a Creative Commons: Attribution 4.0 International (CC BY 4.0) License. Originally published at: Sato, Yasuhiro; Shimizu-Inatsugi, Rie; Yamazaki, Misako; Shimizu, Kentaro K; Nagano, Atsushi J (2019). Plant trichomes and a single gene GLABRA1 contribute to insect community composition on field-grown Arabidopsis thaliana. BMC Plant Biology, 19:163. DOI: https://doi.org/10.1186/s12870-019-1705-2 Sato et al. BMC Plant Biology (2019) 19:163 https://doi.org/10.1186/s12870-019-1705-2 Open Access Abstract Background: Genetic variation in plants alters insect abundance and community structure in the field; however, little is known about the importance of a single gene among diverse plant genotypes. In this context, Arabidopsis trichomes provide an excellent system to discern the roles of natural variation and a key gene, GLABRA1, in shaping insect communities. In this study, we transplanted two independent glabrous mutants (gl1–1 and gl1–2) and 17 natural accessions of Arabidopsis thaliana to two localities in Switzerland and Japan. Results: Fifteen insect species inhabited the plant accessions, with the insect community composition significantly attributed to variations among plant accessions. The total abundance of leaf-chewing herbivores was negatively correlated with trichome density at both field sites, while glucosinolates had variable effects on leaf chewers between the sites. Interestingly, there was a parallel tendency for the abundance of leaf chewers to be higher on gl1–1 and gl1–2 than on their different parental accessions, Ler-1 and Col-0, respectively. Furthermore, the loss of function in the GLABRA1 gene significantly decreased the resistance of plants to the two predominant chewers; flea beetles and turnip sawflies. Conclusions: Overall, our results indicate that insect community composition significantly varies among A. thaliana accessions across two distant field sites, with GLABRA1 playing a key role in altering the abundance of leaf-chewing herbivores. Given that such a trichome variation is widely observed in Brassicaceae plants, the present study exemplifies the community-wide effect of a single plant gene on crucifer-feeding insects in the field. Keywords: Brassicaceae, Community genetics, GL1, Herbivory, In natura, Plant-insect interaction Background association of particular genetic polymorphisms with leaf damage [9, 10], insect abundance [3, 11], and community composition [3] in the field. In comparison, other studies have focused on how single genes affect the insect commu- nity using transformed plants [12, 13]. These lines of evidence from diverse plant species suggest that both quan- titative genetic variation and single genes contribute to the community genetics of plant-insect interactions. Plants develop various resistance traits, such as spines and toxins, to deter herbivory [1]. A growing number of studies on community genetics has revealed that genetic variation in plant resistance traits exerts cascading effects on insect abundance and community composition [2–5]. These in- sect indices projected on individual plants, called extended phenotype [5], can be explained by variation among plant genotypes [6–8]. Some researchers have reported the Arabidopsis thaliana (L.) Heynh. (Brassicaceae) is well-studied as a model system of Brassicaceae-insect in- teractions [14], within which intensive genomic and phenotypic information is available for the world-wide collection of natural accessions [15]. In Arabidopsis-her- bivore interactions, plant trichomes (epidermal hairs) * Correspondence: kentaro.shimizu@ieu.uzh.ch; anagano@agr.ryukoku.ac.jp 3Department of Evolutionary Biology and Environmental Studies, University of Zurich, Winterthurerstrasse 190, 8057 Zurich, Switzerland 5Department of Plant Life Sciences, Faculty of Agriculture, Ryukoku University, Yokotani 1-5, Seta Oe-cho, Otsu, Shiga 520-2194, Japan Full list of author information is available at the end of the article * Correspondence: kentaro.shimizu@ieu.uzh.ch; anagano@agr.ryukoku.ac.jp 3Department of Evolutionary Biology and Environmental Studies, University of Zurich, Winterthurerstrasse 190, 8057 Zurich, Switzerland 5Department of Plant Life Sciences, Faculty of Agriculture, Ryukoku University, Yokotani 1-5, Seta Oe-cho, Otsu, Shiga 520-2194, Japan Full list of author information is available at the end of the article Plant trichomes and a single gene GLABRA1 contribute to insect community composition on field-grown Arabidopsis thaliana Yasuhiro Sato1,2, Rie Shimizu-Inatsugi3, Misako Yamazaki3, Kentaro K. Shimizu3,4* and Atsushi J. Yasuhiro Sato1,2, Rie Shimizu-Inatsugi3, Misako Yamazaki3, Kentaro K. Shimizu3,4* and Atsushi J. Nagano5* Plant materials and defense traits Plant materials and defense traits Arabidopsis thaliana (L.) Heynh., commonly known as thale cress or mouse-ear cress, is an annual weed native to Eurasia and Africa and naturalized in North America and East Asia [15]. This species is predominantly self-fertilizing [50] and when plants are collected from wild populations or when mutants are isolated by muta- genesis, selfed seeds can be maintained as an inbred line called an “accession”. Weak dormancy and early-flowering accessions, such as Col-0 and Ler-1 [51, 52], form both the spring and summer cohort owing to their rapid life-cycles [53]. The spring cohort flowers and sets seeds in spring, and the summer cohort germi- nates in early summer and flowers in autumn [53]. The accessions with strong dormancy, such as Cvi-0 and Shahdara [52], pass the summer as seeds. The accessions with a strong dormancy and late-flowering phenotype, such as Kas-2, are predominantly winter-annuals with only one generation within a calendar year [54]. These different life-cycles of A. thaliana accessions depend on the level of seed dormancy, which can be attributed to the allelic status of the DELAY OF GERMINATION1 (DOG1) and DOG6 genes [51, 52, 54] and the duration to flower development, which is determined by FRI- GIDA, FLOWERING LOCUS C and several other genes [54, 55]. In wild populations within Europe, generalist slugs and seed weevils feed on A. thaliana during late spring, while more diverse herbivores, such as Phyllo- treta beetles, green peach aphids Myzus persicae and diamondback moths Plutella xylostella, occur during summer [56, 57]. In wild populations near our study lo- cations, we discovered A. thaliana plants during early summer (Fig. 1), though the population size seemed smaller than those that overwintered. We also observed flowering and vegetative A. thaliana co-occurring during early summer and the plants displayed leaf damage from insect herbivores (Fig. 1b). Therefore, to investigate the diverse herbivores on A. thaliana we simulated the sum- mer cohort using accessions with different life-cycles and defense traits. y p Complex genetic mechanisms underlie anti-herbivore defenses when plants are exposed to multiple biotic and abiotic stresses [26, 27]; consequently, the phenotype under constant laboratory conditions might not be adequate for understanding how genes function in the field [28–32]. Page 2 of 12 Page 2 of 12 Sato et al. BMC Plant Biology (2019) 19:163 Sato et al. BMC Plant Biology (2019) 19:163 function as a mechanical barrier against feeding and ovi- position by insect herbivores [11, 16–18]. Glucosinolates (GSLs) are major secondary metabolites of Brassicales that act as toxic chemicals against generalists [19, 20], but can be detoxified by specialist herbivores [14, 20, 21]. For example, previous studies on A. thaliana focused on how these physical and chemical traits confer resistance against specific herbivore species, such as the cabbage butterfly Pieris rapae [22, 23], the diamond back moth Plutella xylostella [16, 20] and the green peach aphid Myzus persicae [24, 25]. However, knowledge remains lim- ited about (i) how many insect species occupy A. thaliana in the field, (ii) whether plant defense traits contribute to insect abundance and community composition and (iii) the host genes that are responsible for community mem- bers and overall community composition. is there significant variation in herbivore abundance and community composition among the A. thaliana accessions; (ii) which plant traits (physical, chemical or other life-his- tory traits) influence herbivore abundance and community composition; (iii) does the loss of function of a single gene, GL1, alter insect abundance and community composition? Plant materials and defense traits thaliana with small leaf holes on their leaves, observed on 5 July 2018 near Zurich (47° 21′ N, 8° 38′ E) and c Flowering A. thaliana observed on 2 June 2018 near Otsu (34° 57′ N, 135° 56′ E). A voucher of the wild A. thaliana from near Zurich has been deposited in the United Herbaria Z + ZT (https://www.herbarien.uzh.ch/en/herbarien-zzt.html), Zurich, Switzerland, with the reference code of Z-000164966, Z-000164967 and Z-000164968 Fig. 1 Photographs of Arabidopsis thaliana occurring in wild populations during early summer. a Flowering A. thaliana observed on 25 July 2018 near Zurich (47° 21′ N, 8° 38′ E), b Vegetative A. thaliana with small leaf holes on their leaves, observed on 5 July 2018 near Zurich (47° 21′ N, 8° 38′ E) and c Flowering A. thaliana observed on 2 June 2018 near Otsu (34° 57′ N, 135° 56′ E). A voucher of the wild A. thaliana from near Zurich has been deposited in the United Herbaria Z + ZT (https://www.herbarien.uzh.ch/en/herbarien-zzt.html), Zurich, Switzerland, with the reference code of Z-000164966, Z-000164967 and Z-000164968 pairwise genetic distance was 5.7% in median, which is comparable to that of all accessions analyzed by the 1001 Genome Consortium [15]. These 17 accessions include both early- and late-flowering accessions (e.g., Col-0 and Kas-2 analyzed by Taylor et al. [54]), such that the flower- ing time under a long-day laboratory conditions ranged from 23 (Ws-2 accession) to 92 days (Br-0 accession) [35]. To examine the effects of plant life-history traits on insect community composition, we measured and incorporated the plant size and presence of flowering stems (see ‘Com- mon garden experiment’ and ‘Statistical analysis’ below). Table 1 Arabidopsis thaliana accessions used in this study Accession ID Locality Trichome (no./cm2) Bay-0 N22633 Germany 26.3 Br-0 N22628 Czech Republic 0 C24 N22620 Portugal 2.5 Col-0 N22625 USA 32.5 Col (gl1–2) CS3126† USA 4.0‡ Cvi-0 N22614 Cape Verde 104.3 Est-1 N22629 Russia 39.3 Kas-2 CS6751 India 9 Kin-0 N22654 USA 14 Ler-1 N22618 Germany 14.3 Ler (gl1–1) CS64* Germany 0 Mr-0 N22640 Italy 23.3 Ms-0 N22655 Russia 43.6‡ Nd-1 N22619 Switzerland 47 Se-0 N22646 Spain 30.5 Shahdara N22652 Tajikistan 55.5 Tsu-1 N22641 Japan 11.3 Van-0 N22627 Canada 20.8 Ws-2 N22659 Russia 33.3 The table shows the stock ID, locality and trichome density (no./cm2: Atwell et al. [35]) Obtained through Kiyotaka Okada Laboratory of Kyoto University, Japan †Obtained through Dr. M. Plant materials and defense traits In this concept of in natura study [28, 30–32], Arabidopsis thaliana provides an ideal opportunity to test the effects of single plant genes on herbivores because its molecular mechanisms of defense have been studied using a variety of accessions, with respect to trichomes [33–36] and second- ary metabolites [20, 36, 37]. For example, loss of function mutations in a few transcription factor genes (including GLABRA1 (GL1), GLABRA2, GLABRA3, and TRANSPAR- ENT TESTA GLABRA1) result in glabrous phenotypes in A. thaliana [38–42] and related species [9, 43, 44]. While the loss of function of the latter three genes results in pleio- tropic defects in root hairs, the loss of function of GL1 does not affect root hairs, due to the subfunctionalization of GL1 and its homolog WEREWOLF [40–42, 45]. Laboratory experiments have shown that loss of function in GL1 decreases resistance against leaf-chewing herbivores [22] and improves plant growth by saving on the cost of defense [17, 25]. However, these genetic effects remain unexplored in the field. Common garden experiments using single-gene mutants provide a powerful tool to determine the causal links be- tween a particular gene and its phenotypes [46–49]. In this study, we transplanted two glabrous mutants and 17 nat- ural accessions of A. thaliana. In particular, we focused on gl1–1 and gl1–2 accessions, of which the former is a null trichome mutant derived from the Ler-1 accession and the latter is a hypomorphic mutant from the Col-0 accession [38, 45]. In addition, the natural accessions were selected to cover variations in trichome density and GSLs content [35, 37, 43]. Common garden experiments with these plants were conducted at two field sites in Switzerland and Japan, to identify common patterns between the two insect communities. Three specific questions were addressed: (i) To cover the wide variation in trichome density (phys- ical defense) and GSL accumulation (chemical defense) with early- and late-flowering cycles, we selected 17 nat- ural accessions and two glabrous mutants (Table 1). The natural accessions selected in this study should represent world-wide genetic variation, because the genome-wide Sato et al. BMC Plant Biology (2019) 19:163 Page 3 of 12 Fig. 1 Photographs of Arabidopsis thaliana occurring in wild populations during early summer. a Flowering A. thaliana observed on 25 July 2018 near Zurich (47° 21′ N, 8° 38′ E), b Vegetative A. Plant materials and defense traits 500 m) and the Center for Ecological Research, Kyoto University (Otsu, Japan: 35° 06′ N, 134° 56′ E, alt. ca. 200 m) (Fig. 2). The Zurich site is close to a deciduous forest and the surroundings of the common garden are covered with concrete tiles to prevent weeds. The Otsu site is a suburb of cultivated fields and the ground of the study site is covered with short grasses. In the Otsu site, the grass weeds were mown, and the sur- roundings were covered with agricultural sheets before the experiment. No large Brassica cultivars were grown at either site during early summer. Average air temperature and total precipitation was 19 °C and 198 mm in Zurich (during July 2016; MeteoSwiss, http:// www.meteoswiss.admin.ch/home.html) and 22 °C and 321 mm in Otsu (during June 2016; Japan Meteoro- logical Agency, http://www.jma.go.jp/jma/index.html). data in Chan et al. [37] on 21 GSLs of 96 A. thaliana ac- cessions. They used a mature leaf at 35-days post germin- ation from a plant grown under short-day laboratory conditions without herbivory. As they performed two tri- als to quantify GSL, we used the average GSL contents (nmol/mg flesh weight). We focused on variation in ali- phatic GSLs and their chain-length, because these param- eters play a major role in preventing above-ground herbivory [48, 60, 61]. Regarding the data of Chan et al. [37], we applied a principal component analysis (PCA) to total C3-, C4-, C5-, C7-, and C8- aliphatic GSLs. The first and second principal components explained 44 and 33% variation in the GSL profiles among our 17 accessions, re- spectively (Additional file 1); therefore, these two compo- nents were used in our statistical analyses. Plant materials and defense traits Ohto ‡Estimated from the relative trichome density to Col-0 accession presented in previous publications (Hauser et al. [43] and Yoshida et al. [39] for Ms-0 and gl1–2, respectively) To test the functional advantages of the GL1 gene in pro- ducing trichomes, we added two glabrous mutants, gl1–1 and gl1–2, to the set of natural accessions (Table 1). The former mutant, gl1–1, has the background of Ler accession with a null mutation due to a 6.5-kb deletion on GL1 and lacks leaf surface trichomes. The latter, gl1–2, has the back- ground of Col accession with the deletion of 27 amino acids induced by X-ray radiation, showing a hypomorphic mutation with a lower density of trichomes on leaf surfaces [38, 45]. Out of the 17 natural accessions, Br-0 and C24 have no or few trichomes due to a frameshift mutation and one amino acid change in the myb DNA binding domain of GL1, respectively [34]. We compiled the data on leaf trichome density (no./cm2) from the GWA-portal (https:// gwas.gmi.oeaw.ac.at/: [35]). All the natural accessions were included in previous quantitative genetic studies of GSL. Seven were used as the parental genotypes of recombinant inbred lines (e.g., Col × Ler and Cvi × Ler [20]; Bay × Sha [58]; Kas × Tsu [59]) and the other accessions were used in genome-wide association mapping [37]. To test whether genetic poten- tials in GSL profiles explain herbivory rates, we used the Page 4 of 12 Sato et al. BMC Plant Biology (2019) 19:163 Page 4 of 12 8° 33′ E, alt. ca. 500 m) and the Center for Ecological Research, Kyoto University (Otsu, Japan: 35° 06′ N, 134° 56′ E, alt. ca. 200 m) (Fig. 2). The Zurich site is close to a deciduous forest and the surroundings of the common garden are covered with concrete tiles to prevent weeds. The Otsu site is a suburb of cultivated fields and the ground of the study site is covered with short grasses. In the Otsu site, the grass weeds were mown, and the sur- roundings were covered with agricultural sheets before the experiment. No large Brassica cultivars were grown at either site during early summer. Average air temperature and total precipitation was 19 °C and 198 mm in Zurich (during July 2016; MeteoSwiss, http:// www.meteoswiss.admin.ch/home.html) and 22 °C and 321 mm in Otsu (during June 2016; Japan Meteoro- logical Agency, http://www.jma.go.jp/jma/index.html). 8° 33′ E, alt. ca. Statistical analysis Response variables - Community indices were examined at three levels (i.e., component species, guilds and entire com- munities) as response variables in the following analyses. At the species level, we analyzed the number of individuals of each herbivorous species. We analyzed species for which more than 20 individuals were observed at each site, be- cause statistical tests were difficult to apply to rare species. For the Zurich data, we analyzed the number of leaf holes as an indicator of damage by flea beetles. At the guild level, we classified herbivorous species into those feeding on ex- ternal leaf tissues (i.e., leaf chewers) and those feeding on internal plant tissues (including sap suckers and leaf miners). We also separated herbivorous species into spe- cialists on Brassicaceae (e.g., cabbage butterflies, cabbage sawflies and turnip flea beetles) and generalists on multiple plant families (some species of aphids and thrips) (Table 2; Additional file 2). The total number of insect individuals in each category were analyzed as guild level statistics. At the entire community level, we calculated species richness (i.e., number of species), Shannon’s diversity index H′ and the total number of insect individuals on individual plants. All of the response variables were ln(x + 1)-transformed to im- prove normality before statistical analyses. All statistical analyses were conducted using R version 3.2.0 [63]. We utilized the rda function (in the vegan package: [64]) to perform the redundancy analysis. We used the lme function (in the nlme package: [65]) to estimate heritability, as described below. y y Insects and herbivorous collembola on individual plants were visually counted every 2–3 days. These species were identified ocularly with a magnifying glass. Dwelling traces and mummified aphids were also counted as a proxy of the number of leaf miners and parasitoid wasps, respect- ively. Eggs, larvae, and adults were counted for all species, as long as they could be observed by the naked eye. The abundance of each species was evaluated by the cumula- tive number of individuals over the experimental period to reflect the herbivory load on the plants [62]. Small holes made by flea beetles were counted at the Zurich site and the maximum number throughout the experiment was used as an indicator of damage by flea beetles; however, this phenotyping was difficult in Japan, due to heavier and simultaneous infestation by sawflies. All counting was conducted by a single observer during the daytime (08:00–17:00). Common garden experiment We prepared 10 replicates of 19 accessions (= 190 plants in total) for each experiment. To minimize un- necessary variation due to competition and consequent We used the experimental gardens of the University of Zurich at Irchel campus (Zurich, Switzerland: 47° 23′ N, Fig. 2 Photographs of plants and insects in the study field. a Leaf holes made by flea beetles (arrows), b a striped flea beetle Phyllotreta striolata, c a turnip flea beetle Phyllotreta atra, d mustard aphids Lipaphis erysimi, e a larva of the turnip sawfly Athalia rosae, f a newly hatched larva of the cabbage butterfly Pieris rapae, g a larva of the diamond back moth Plutella xylostella, h a western flower thrips Frankliniella occidentalis, i the field site in Zurich, Switzerland, and j the field site in Otsu, Japan Fig. 2 Photographs of plants and insects in the study field. a Leaf holes made by flea beetles (arrows), b a striped flea beetle Phyllotreta striolata, c a turnip flea beetle Phyllotreta atra, d mustard aphids Lipaphis erysimi, e a larva of the turnip sawfly Athalia rosae, f a newly hatched larva of the cabbage butterfly Pieris rapae, g a larva of the diamond back moth Plutella xylostella, h a western flower thrips Frankliniella occidentalis, i the field site in Zurich, Switzerland, and j the field site in Otsu, Japan Sato et al. BMC Plant Biology (2019) 19:163 Page 5 of 12 Sato et al. BMC Plant Biology (2019) 19:163 Page 5 of 12 Page 5 of 12 size bias, experimental plants were grown in separate pots under uniform conditions before they were exposed to the field environment in the common garden. We cultivated plants using mixed soils of agricultural com- posts (Profi Substrat Classic CL ED73, Einheitserde Co. in Switzerland; MetroMix 350, SunGro Co. in Japan) and perlites with a compost to perlite ratio of 3:1 litter volume. No additional fertilizers were supplied because the agricultural soils contain fertilizers. Seeds were sown on the soil and stratified under constant dark conditions at 4–5 °C air temperature for a week. Plants were then grown under short-day conditions (8 h:16 h light:dark [L:D], 20 °C air temperature, and 60% relative humidity) for 1 month to prevent flowering before the field experi- ment. The plant positions were rotated every week to minimize growth bias from light conditions. Common garden experiment These experiments were conducted from June 18 to July 1, 2016 in Otsu and from July 13 to August 3, 2016 in Zurich. Plants were watered every 3 days in Otsu and every day in Zurich. The mortality of plants during insect monitoring was moderate, 4% at Zurich, and 9% at Otsu. The field experi- ment at Otsu ended on July 1, 2016 because one third of the plants died due to temporary high air temperatures on July 2 and 3, 2016. We recorded the initial plant size and the presence/ab- sence of flowering stems to incorporate the effects of plant life-history traits on insect abundance. Initial plant size was evaluated by the length of the largest rosette leaf (mm) at the beginning of the field experiment, because this parameter represents plant size at the growth stage. The presence/absence of flowering stems was recorded 2 weeks after transplanting the plants. The final leaf size could not be evaluated because some large leaves received herbivore damage or started drying due to senescence at the end of the experiment; therefore, only the initial size was incorporated into our statistical analyses. Common garden experiment Each indi- vidual plant was moved to a plastic pot (7.5-cm diam- eter, with 6.5-cm depth in Japan; 6.0 × 6.0 × 6.0 cm in Switzerland) and acclimated for 3 days in a shaded out- door place before the field experiments. The potted plants were randomly placed among three blocks in each common garden: 68, 69, and 53 plants were assigned to each block in Zurich; and 76, 76, and 38 plants were assigned to each block in Otsu. The potted plants were set on water-permeable plastic sheets in a checkered manner within the blocks, without being embedded in the ground. The blocks were more than 1.0 m apart. These experiments were conducted from June 18 to July 1, 2016 in Otsu and from July 13 to August 3, 2016 in Zurich. Plants were watered every 3 days in Otsu and every day in Zurich. size bias, experimental plants were grown in separate pots under uniform conditions before they were exposed to the field environment in the common garden. We cultivated plants using mixed soils of agricultural com- posts (Profi Substrat Classic CL ED73, Einheitserde Co. in Switzerland; MetroMix 350, SunGro Co. in Japan) and perlites with a compost to perlite ratio of 3:1 litter volume. No additional fertilizers were supplied because the agricultural soils contain fertilizers. Seeds were sown on the soil and stratified under constant dark conditions at 4–5 °C air temperature for a week. Plants were then grown under short-day conditions (8 h:16 h light:dark [L:D], 20 °C air temperature, and 60% relative humidity) for 1 month to prevent flowering before the field experi- ment. The plant positions were rotated every week to minimize growth bias from light conditions. Each indi- vidual plant was moved to a plastic pot (7.5-cm diam- eter, with 6.5-cm depth in Japan; 6.0 × 6.0 × 6.0 cm in Switzerland) and acclimated for 3 days in a shaded out- door place before the field experiments. The potted plants were randomly placed among three blocks in each common garden: 68, 69, and 53 plants were assigned to each block in Zurich; and 76, 76, and 38 plants were assigned to each block in Otsu. The potted plants were set on water-permeable plastic sheets in a checkered manner within the blocks, without being embedded in the ground. The blocks were more than 1.0 m apart. Statistical analysis Early-flowering accessions of A. thaliana start reproduction under long-day field conditions and terminate their life-cycle within a month. Thus, to avoid the disturbance during analysis due to plant senescence, the duration of field observations was for 3 weeks, after the beginning of the field experiment. We rotated the order of census plants to avoid any bias of census time. Variation in insects on plant accessions - To quantify variation in insect communities among plant accessions and study sites, we performed a redundancy analysis to partition sources of variation in community composition into the plant accession, study sites and accession-by-site effects. The accession-by-site interaction was first ana- lyzed 999-times by permutation tests and then the main Page 6 of 12 Sato et al. BMC Plant Biology (2019) 19:163 Table 2 Insect species observed on field-grown Arabidopsis thaliana Common name Scientific name Feeding habit Host range Abundance † Zurich Otsu Cabbage looper Trichoplusia ni Leaf chewer Generalist 0 – Diamond back moth Plutella xylostella Leaf chewer Specialist ++ ++ Garden springtail Bourletiella hortensis Leaf chewer Generalist – + Piggyback grasshopper Atractomorpha lata Leaf chewer Generalist – – Cabbage butterfly** Pieris rapae Leaf chewer Specialist + + Striped flea beetle Phyllotreta striolata Leaf chewer Specialist ++ – Turnip flea beetle Phyllotreta atra Leaf chewer Specialist +++ 0 (Leaf holes made by flea beetles) Phyllotreta spp. Leaf chewer Specialist +++ – Turnip sawfly Athalia rosae Leaf chewer Specialist 0 +++ (Dwelling traces) NA Internal feeder Generalist – 0 Green peach aphid Myzus persicae Internal feeder Generalist + + Mustard aphids Lipaphis erysimi Internal feeder Specialist +++ + Onion thrip Thrips tabaci Internal feeder Generalist + + Western flower thrip Frankliniella occidentalis Internal feeder Generalist ++ ++ (Parasitoid wasp indicated by mummified aphids) NA Carnivore + + Seven-spot ladybird Coccinella septempunctata Carnivore 0 – Detailed information about the abundance of insects at the Zurich and Otsu sites is provided in the supplementary material (Additional file 2) Only this species is a non-insect arthropod *The abundance of Pieris rapae was evaluated by the total number of eggs and larvae †Abundance level: +++ (abundant), ++, +, −(rare), and 0 (not found) NA indicates not applicable pairs). We considered the difference of experimental block as a covariate. Because trichome density had a highly skewed distribution due to completely glabrous pheno- types, this variable was ln(x + 1)-transformed before the analysis. Statistical analysis First, we tested the effects of plant traits on each response variable without the dataset on glabrous mutants. When detecting significant effects of trichomes on a par- ticular herbivore, we then compared the two glabrous mu- tants and their parental accessions to test how the GL1 genes affect guild and community indices encompassing the focal herbivore. Multiple regression with linear mixed models was used to analyze trichome production, initial plant size (mm), and the presence/absence of flowering stems as explanatory variables. The difference in parental background (i.e., Ler-1 or Col-0) was considered as a ran- dom effect. We used the lme function with the maximum likelihood method for these mixed models. All of the con- tinuous response and explanatory variables were standard- ized following a normal distribution, with zero mean and one variance, to make coefficients comparable between the linear models. P-values were corrected by FDR [68]. effects of accessions and sites were examined without the interaction term. Then, we estimated broad-sense herit- ability H2 in a focal response, as the proportion of variance attributable to plant accessions. We used linear mixed models, in which the accession ID was assigned as a ran- dom effect. This variance component of random effect was estimated by the restricted maximum likelihood method [66, 67]. The significance of heritability was exam- ined by likelihood ratio tests by comparing the linear models with or without the random effect of accession ID. This estimation of heritability was separately performed for the data from Zurich and Otsu. P-values were cor- rected by the false discovery rate (FDR) of multiple testing [68]. Another option of estimating heritability is to incorp- orate a genetic distance matrix among natural accessions, as used in genome-wide association studies [e.g., 26, 27]. However, it was difficult to apply the same approach to single-gene mutants and the limited number of accessions; thus, we adopted the linear mixed models without the dis- tance matrix to estimate broad-sense heritability. Effects of plant traits - To address whether particular plant traits contributed to community members and com- position, we used multiple regressions that considered the trichome density, PC1 and PC2 of aliphatic GSLs, the pres- ence/absence of flowering stems, and initial plant size (mm) as explanatory variables. No explanatory variables were heavily correlated with each other (\|r\| < 0.6 for all Plant traits underlying the abundance and communities of insects We examined whether the species, guild, and community structure of insects were affected by trichome density, glucosinolates, and life-history traits among natural acces- sions (Fig. 4). Significant effects of trichomes at the species and guild levels were observed (Fig. 4; Additional file 4). Two predominant leaf chewers, the striped flea beetle in Zurich and the turnip sawfly in Otsu, occurred less on hairy accessions than on accessions that produced low quantities of trichomes (Figs. 4 and 5b; Additional file 4). The number of leaf holes was lower on hairy plants com- pared to glabrous plants at the Zurich site (Figs. 4 and 5a; Additional file 4), indicating that trichomes have a resist- ance function against flea beetles. At the Otsu site, the abundance of the cabbage butterfly Pieris rapae was also lower on hairy plants (Fig. 4; Additional file 4). At the guild level, trichomes had significant negative effects on Abundance and communities of insects among plant accessions and study sites We observed 15 insect species including flea beetles, saw- flies, butterflies, moths, aphids and thrips on A. thaliana Page 7 of 12 Sato et al. BMC Plant Biology (2019) 19:163 Page 7 of 12 the leaf chewers at both sites (Figs. 4 and 5c, d; Additional file 4). In contrast to trichomes, aliphatic GSLs did not have any consistent effects on herbivore abundance. The first principal component of GSLs had negative effects on leaf chewers, specialist herbivores, species richness and total abundance at the Zurich site but no significant ef- fects on these indices at the Otsu site (Fig. 4). The second principal component of GSLs was positively correlated with the abundance of turnip sawflies at the Otsu site and western flower thrips at the Zurich site (Fig. 4: Additional file 4). These effects of trichomes and GSLs were variable between the two sites with respect to insect richness, Shannon diversity and total abundance (Figs. 4 and 5e, f; Additional file 4). Initial plant size or the presence of flowering stems significantly increased insect richness, diversity and total abundance at both sites (Fig. 4; Add- itional file 4). The result that Kas-2 in Otsu and C24 in Zurich were less likely to be occupied by leaf chewers (Fig. 5) was due to the small plant size of these acces- sions (Additional file 5). in the two field experiments (Table 2; Fig. 2). Of these in- sects, five and three species were specific to the Otsu and Zurich sites, respectively. Redundancy analysis and per- mutation tests confirmed that the plant accession, study site and accession-by-site effects exhibited significant sources of variation in the community composition (Ac- cession, Sum of Squares (SS) = 0.99, F = 1.57, P < 0.001; Site, SS = 1.31, F = 36.2, P < 0.001; Accession-by-site, SS = 0.96, F = 1.54, P < 0.001 with 999 permutations; Fig. 3). Based on the sum of squares, the redundancy analysis in- dicated that 13% of variation in the insect community composition can be explained by the plant accessions and accession-by-site effects. We also found significant broad- sense heritability in species richness, Shannon diversity and total abundance of insects on A. thaliana, and its magnitude varied between the two study sites (10–11% and 15–30% heritability in Zurich and Otsu, respectively: Additional file 3). Insect species diversity on field-grown Arabidopsis thaliana A number of researchers have studied leaf herbivory [17, 48, 60, 69], herbivorous fauna [19, 56, 57, 70], and plant fit- ness [17, 48, 60, 69] in Arabidopsis thaliana under field conditions; however, quantitative evidence remains limited in relation to insect community composition on this plant species. In the present study, we found that the community composition of 15 arthropod species significantly varied among A. thaliana accessions in the two distant field sites. Consistent with our results in the Zurich site, Harvey et al. [57] found that the summer cohort of A. thaliana were heavily attacked by Phyllotreta beetles and also harbored diamondback moths and aphids in a wild population of Europe. While winter-annual is considered the major life-cycle of A. thaliana [56, 57], the short generation time and seed bank may allow them to form alternative life-cy- cles in the wild. Near the Zurich site, in addition to the overwintering cohort, flowering and vegetative plants co-occurred during July and more importantly, the wild A. thaliana had many holes in their leaves (Fig. 1). Our evi- dence of a summer cohort suggests that the present field experiments using a variety of accessions should represent a part of the Arabidopsis-herbivore interactions observed in nature, and thereby highlights a further need to study insect communities across the multiple life stages of A. thaliana. Comparing glabrous mutants and parental hairy accessions We examined the effects of a single gene GLABRA1 on herbivory, guild and community indices encompassing two predominant leaf chewers, flea beetles and sawflies. At the species level, compared to parental accessions, the two glabrous mutants had significantly more leaf holes made by the flea beetles and the larvae of the turnip sawfly (Fig. 5a, b; Additional file 6). At the guild level, leaf chewers tended to occur more often on the two glabrous mutants than on each of the parental accessions (Fig. 5c, d), although this difference was not statistically significant in Zurich (Additional file 6). Among community indices, total abundance at the Otsu site was significantly lower on the glabrous mutants than on the hairy parents (coef. ± SE = −0.46 ± 0.16, Z = −2.86, P < 0.01; Additional file 6). Fig. 3 Redundancy analysis summarizing the community composition among 19 accessions of A. thaliana in Zurich, Switzerland and Otsu, Japan. White and grey circles indicate the accession mean and its standard error projected on the first and second RDA dimension. Arrows represent the contributions of each species. Permutation tests confirmed significant variation in the community composition among plant accessions and study sites (see the Results section) Fig. 3 Redundancy analysis summarizing the community composition among 19 accessions of A. thaliana in Zurich, Switzerland and Otsu, Japan. White and grey circles indicate the accession mean and its standard error projected on the first and second RDA dimension. Arrows represent the contributions of each species. Permutation tests confirmed significant variation in the community composition among plant accessions and study sites (see the Results section) Sato et al. BMC Plant Biology (2019) 19:163 Page 8 of 12 Fig. 4 A heat map showing FDR-corrected p-values (Pfdr) for the effects of plant traits on insect species, guild and community indices among 17 natural accessions in Zurich, Switzerland and Otsu, Japan. Shown are the effects of trichome density (Trchm), PC1 and PC2 of aliphatic glucosinolates (GSL1 and GSL2), presence of flowering stems (Flower) and initial leaf length (Leaf), on the diamond back moth (Dbm), striped flea beetle (Sfb), turnip flea beetle (Tfb), mustard aphid (Ma), western flower thrip (Wft), turnip sawfly (Ts), cabbage butterfly (Cb), green peach aphid (Gpa), leaf chewers (Chwr), internal feeders (Intrnl), specialists (Spcl), generalists (Gnrl), species richness (Rchns), Shannon diversity (Shnn) and total abundance (Total). Comparing glabrous mutants and parental hairy accessions Colors represent the sign and significance of trait effects: (dark blue square), −coef. With Pfdr < 0.01; (blue square), −coef. With Pfdr < 0.05; (dark red square), + coef. With Pfdr < 0.01; (red square), + coef. With Pfdr < 0.05; (white square), not significant at Pfdr > 0.05 Fig. 4 A heat map showing FDR-corrected p-values (Pfdr) for the effects of plant traits on insect species, guild and community indices among 17 natural accessions in Zurich, Switzerland and Otsu, Japan. Shown are the effects of trichome density (Trchm), PC1 and PC2 of aliphatic glucosinolates (GSL1 and GSL2), presence of flowering stems (Flower) and initial leaf length (Leaf), on the diamond back moth (Dbm), striped flea beetle (Sfb), turnip flea beetle (Tfb), mustard aphid (Ma), western flower thrip (Wft), turnip sawfly (Ts), cabbage butterfly (Cb), green peach aphid (Gpa), leaf chewers (Chwr), internal feeders (Intrnl), specialists (Spcl), generalists (Gnrl), species richness (Rchns), Shannon diversity (Shnn) and total abundance (Total). Colors represent the sign and significance of trait effects: (dark blue square), −coef. With Pfdr < 0.01; (blue square), −coef. With Pfdr < 0.05; (dark red square), + coef. With Pfdr < 0.01; (red square), + coef. With Pfdr < 0.05; (white square), not significant at Pfdr > 0.05 wide association mapping [26]. In the context of commu- nity genetics, plant architectures are a key predictor of in- sect community composition on perennial herbs [6] and woody plants [2]. In the present study, we also found that the presence of flowering stems and larger plant size in- creased insect species richness and diversity on A. thaliana (Fig. 4; Additional file 4). Even though similar accessions produced flowering stems between the two study sites (Additional file 5), only at the Otsu site did the pres- ence of flowering stems contribute to insect species diversity. This result could be partly because the flow- ering stems attracted three herbivore species, the cab- bage butterfly, green peach aphid and western flower thrip, at the Otsu site (Fig. 4). Effects of trichomes and the single gene GLABRA1 on insect abundance In reverse genetic analysis, multiple independent mu- tants with a consistent phenotype are required to prove the function of a particular gene. In addition, using mul- tiple genetic backgrounds of parent can also give a strong indication of gene function. In previous studies, the roles of single genes in modulating herbivory were quantified using mutants derived from a single parental accession [12, 13, 48]. Our common garden experiments illustrate the function of the GL1 gene against herbivory in natura using two distinct lines, Ler (gl1–1) and Col (gl1–2). Notably, the Br-0 and C24 accessions were the most susceptible accessions to leaf chewers at each site (Fig. 5c, d), and these two accessions also have disruptive mutations on GL1 [34, 43]. Such associations between The plant apparency hypothesis assumed the importance of plant size and architectural traits in anti-herbivore defense [71], and this hypothesis has been supported by a meta-analysis [72], comparative study [73], and genome- Sato et al. BMC Plant Biology (2019) 19:163 Page 9 of 12 Sato et al. BMC Plant Biology Fig. 5 Variation in insect species, guild and community on 19 A. thaliana accessions (mean ± SE) in Zurich, Switzerland (left panels) and Otsu, Japan (right panels). Given the significant effects of trichomes on flea beetles P. striolata and sawflies A. rosae (Fig. 4), these panels show herbivory, guild and community indices comprising the flea beetles and sawflies. White bars of plant accessions represent the sparse density of less than 10 leaf trichomes/cm2. Connected lines highlight pairs between a glabrous mutant and its parental accession, where solid and dashed lines indicate significant and non-significant differences between the mutants and parental accessions at Pfdr < 0.05 Fig. 5 Variation in insect species, guild and community on 19 A. thaliana accessions (mean ± SE) in Zurich, Switzerland (left panels) and Otsu, Japan (right panels). Given the significant effects of trichomes on flea beetles P. striolata and sawflies A. rosae (Fig. 4), these panels show herbivory, guild and community indices comprising the flea beetles and sawflies. White bars of plant accessions represent the sparse density of less than 10 leaf trichomes/cm2. Connected lines highlight pairs between a glabrous mutant and its parental accession, where solid and dashed lines indicate significant and non-significant differences between the mutants and parental accessions at Pfdr < 0.05 GL1 polymorphism and anti-herbivore functions have also been reported in field populations of A. Effects of trichomes and the single gene GLABRA1 on insect abundance lyrata [9, 10] and A. halleri [11, 18]. Plant trichomes also prevent herbivory by sawflies [74] and flea beetles [75, 76] on Brassica cultivars. Together with these results, our present results indicate that plant trichomes and a single gene, GL1, play a key role in physical defense against crucifer feeders. detected in field-grown A. halleri [11]. In addition, we could not detect any significant effects of trichomes and GL1 on the abundance of larval P. xylostella, even though trichomes prevent adult moths ovipositing on A. thaliana under laboratory conditions [16]. Handley et al. [16] focused on several northern accessions of A. thali- ana, whereas the current experimental setting covers a wider geographical range of natural accessions. A recent genome-wide association study using 350 natural acces- sions also found no significant association between GL1 polymorphism and herbivory by P. xylostella [27]. Com- bined with the previous studies, our present results from field-grown Arabidopsis exemplify the importance of testing diverse accessions and environmental conditions. Laboratory experiments on single-gene mutants and natural accessions of A. thaliana suggested that plants with high trichome density resisted infestation by aphids [25, 77]. Under the two tested field conditions, trichomes had no significant effect on the abundance of aphids, possibly because aphids primarily occurred on flowering stems, on which the trichome density is low. In fact, the presence of flowering stems was positively correlated with the abundance of aphids (Additional file 4). These results support the limited associations between aphid abundance and GL1 polymorphism Varying effects of glucosinolates on specialist herbivores Some components of GSLs had negative effects whereas others had no or positive effects on the specialist herbi- vores. A possible explanation for these varying effects is Sato et al. BMC Plant Biology (2019) 19:163 Page 10 of 12 that GSL data quantified under laboratory conditions might be insufficient to reflect the GSL accumulation in the field, due to phenotypic plasticity and the induced re- sponse of GSLs to herbivory [23, 24, 48, 78]. For example, the cabbage butterfly P. rapae can modify the expression level of the MAM1 gene, which involves a chain elong- ation of aliphatic GSL [23, 24]. It is also possible that some specialist herbivores overcome GSLs [20, 21, 61, 79, 80] and have thus obscured the effects of GSLs on the com- munity composition. Specifically, the striped flea beetle P. striolata efficiently sequesters 4-methylthiobutyl from A. Competing interests Competing interests The authors declare that they have no competing interests. Additional file 2: Cumulative number of each insect species in Zurich, Switzerland (left chart) and Otsu, Japan (right chart) throughout the experiments. See Table 2 for the name of the arthropod species. Notes: 1Total number of parasitoid wasps and mummified aphids; 2This species is a non-insect arthropod; 3Only a dwelling trace was observed. (PNG 319 kb) Authors’ contributions YS, RSI and MY. performed the field experiment. YS analyzed the data. YS, RSI, KKS and AJN designed the project and wrote the manuscript with input from all co-authors. All the authors have read and approved the manuscript. Abbreviation FDR F l di FDR: False discovery rate; GL1: GLABRA1; GSL: Glucosinolate; PCA: Principal component analysis; RDA: Redundancy analysis Funding This study was supported by JSPS Postdoctoral Fellowship (Grant Number, 16 J30005) and JST PRESTO (JPMJPR17Q4) to Y.S., JST CREST (JPMJCR15O2) and JSPS KAKENHI (JP16H06171 and JP16H01473) to A.J.N., Swiss National Foundation, URPP Global Change and Biodiversity of the University of Zurich, JST CREST (JPMJCR16O3), Japan, MEXT KAKENHI (18H04785) to K.K.S. The field experiment in Japan was supported by the Joint Usage/Research Grant of Center for Ecological Research, Kyoto University. The funders have no role in study design, data analysis, or decision to publish. Availability of data and materials The data and R source code are included in the additional files (Additional files 5 and 7). The data and R source code are included in the additional files Effects of trichomes and the single gene GLABRA1 on insect abundance thaliana, a short-chain aliphatic GSL [80]. The larvae of A. rosae sawflies also sequester GSLs [21], whereas adults utilize isothiocyanates, which are breakdown products of aliphatic GSLs, to find host plants [79]. Unlike the case of trichomes, an on-site quantification of GSLs is needed to resolve their effects on herbivore abundance in the field. and community composition for 17 natural accessions of Arabidopsis thaliana. Standardized coefficient (Coef.), standard error (SE), t-value, and FDR-corrected P-values are shown for each explanatory variable from Zur- ich, Switzerland and Otsu, Japan. Bold values highlight significant effects at Pfdr < 0.05. (XLSX 14 kb) Additional file 5: Insect abundance data. (XLS 137 kb) Additional file 6: Effects of trichome density, presence of flowering stem and initial plant size on insect abundance and community composition in a comparison between glabrous mutants and their parental accessions. Standardized coefficient (Coef.), standard error (SE), t- value, and FDR-corrected P-values are shown for each explanatory vari- able. Bold values highlight significant effects at Pfdr < 0.05. NA means no data available. The trichome density represents differences between the null and hypomorphic mutants; as the trichome density had a significant effect on flea beetles and sawflies (Additional file 4), we focused on spe- cies, guild and community indices comprising these two leaf chewers. (XLSX 11 kb) Additional file 7: R source code for statistical analyses (TXT 10 kb) Additional file 7: R source code for statistical analyses (TXT 10 kb) Acknowledgements Our field investigation showed a genetic basis in the in- sect community assemblage on A. thaliana, and the ad- vantages of the functional allele of GL1 in avoiding leaf chewers. In Brassicaceae plants, evidence is accumulat- ing to suggest that genetic variation within a plant spe- cies alters insect community composition and, in turn, exerts selection on plant defense [4, 81, 82]. Variation in the trichome density is also observed across Brassicaceae plants [74–76], where GL1 orthologs affect the trichome density [83]. In the context of community genetics, the present study on GL1 provides evidence of a key gene affecting the community composition of crucifer-feeding insects. Future studies should assess the relative import- ance of single genes and quantitative genetic variation towards a comprehensive understanding of plant genetic effects on insect community assembly. The authors thank Jordi Bascompte and Matthew Barbour for their valuable discussions, Satoshi Oda and Aki Morishima for helping with plant cultivation and fieldwork, and three anonymous reviewers for their comments on the manuscript. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Additional file 3: Likelihood ratio tests for estimating broad-sense herit- ability H2. Likelihood ratio, LR-χ2, was tested by comparing the models with and without a random effect of the accession ID. P-values were based on a χ2 distribution with one degree of freedom and corrected by false dis- covery rate (FDR) [68]. Bold values indicate significant H2 at Pfdr < 0.05. Bars indicate no information available due to low abundance. (XLSX 11 kb) Consent for publication Not applicable. Consent for publication Not applicable. Competing interests The authors declare that they have no competing interests. Additional files Ethics approval and consent to participate Not applicable. Ethics approval and consent to participate Ethics approval and consent to participate Ethics approval and consent to participate Not applicable. Additional file 1: The first and second principal component (PC1 and PC2) summarizing the total amount (nmol/mg flesh weight) of C3-, C4-, C5-, C7- and C8-Aliphatic glucosinolates for 17 accessions of A. thaliana (compiled from Chan et al. [37]). Red arrows indicate contributions of each glucosinolate to PC1 and PC2. (PNG 145 kb) Not applicable. Additional file 1: The first and second principal component (PC1 and PC2) summarizing the total amount (nmol/mg flesh weight) of C3-, C4-, C5-, C7- and C8-Aliphatic glucosinolates for 17 accessions of A. thaliana (compiled from Chan et al. [37]). Red arrows indicate contributions of each glucosinolate to PC1 and PC2. (PNG 145 kb) Additional file 2: Cumulative number of each insect species in Zurich, Switzerland (left chart) and Otsu, Japan (right chart) throughout the experiments. See Table 2 for the name of the arthropod species. Notes: 1Total number of parasitoid wasps and mummified aphids; 2This species is a non-insect arthropod; 3Only a dwelling trace was observed. (PNG 319 kb) Additional file 3: Likelihood ratio tests for estimating broad-sense herit- ability H2. Likelihood ratio, LR-χ2, was tested by comparing the models with and without a random effect of the accession ID. P-values were based on a χ2 distribution with one degree of freedom and corrected by false dis- covery rate (FDR) [68]. Bold values indicate significant H2 at Pfdr < 0.05. Bars indicate no information available due to low abundance. (XLSX 11 kb) Additional file 4: Effects of trichome density, the first and second principal component (PC1 and PC2) of aliphatic glucosinolates (GSLs), presence of flowering stems and initial plant sizes on insect abundance References 28. Shimizu KK, Kudoh H, Kobayashi MJ. Plant sexual reproduction during climate change: gene function in natura studied by ecological and evolutionary systems biology. Ann Bot. 2011;108:777–87. 1. Schoonhoven LM, van Loon JJA, Dicke M. Insect-plant biology. 2nd ed. Oxford, UK: Oxford University Press; 2005. 1. Schoonhoven LM, van Loon JJA, Dicke M. Insect-plant biology. 2nd ed. Oxford, UK: Oxford University Press; 2005. 2. Barbour MA, Rodriguez-Cabal MA, Wu ET, et al. Multiple plant traits shape the genetic basis of herbivore community assembly. Funct Ecol. 2015;29:995–1006. 2. Barbour MA, Rodriguez-Cabal MA, Wu ET, et al. Multiple plant traits shape the genetic basis of herbivore community assembly. Funct Ecol. 2015;29:995–1006. 29. Nagano AJ, Sato Y, Mihara M, et al. Deciphering and prediction of transcriptome dynamics under fluctuating field conditions. Cell. 2012;151:1358–69. 3. Bernhardsson C, Robinson KM, Abreu IN, et al. Geographic structure in metabolome and herbivore community co-occurs with genetic structure in plant defence genes. Ecol Lett. 2013;16:791–8. 3. Bernhardsson C, Robinson KM, Abreu IN, et al. Geographic structure in metabolome and herbivore community co-occurs with genetic structure in plant defence genes. Ecol Lett. 2013;16:791–8. 30. Kudoh H. Molecular phenology in plants: in natura systems biology for the comprehensive understanding of seasonal responses under natural environments. New Phytol. 2016;210:399–412. 4. Poelman EH, van Dam NM, van Loon JJA, Vet LE, Dicke M. Chemical diversity in Brassica oleracea affects biodiversity of insect herbivores. Ecology. 2009;90:1863–77. 4. Poelman EH, van Dam NM, van Loon JJA, Vet LE, Dicke M. Chemical diversity in Brassica oleracea affects biodiversity of insect herbivores. Ecology. 2009;90:1863–77. 31. Yamasaki E, Altermatt F, Cavender-Bares J, et al. Genomics meets remote sensing in global change studies: monitoring and predicting phenology, evolution and biodiversity. Curr Opin Environ Sustain. 2017;29:177–86. 5. Whitham TG, Bailey JK, Schweitzer JA, et al. A framework for community and ecosystem genetics: from genes to ecosystems. Nat Rev Genet. 2006;7:510–23. 5. Whitham TG, Bailey JK, Schweitzer JA, et al. A framework for community and ecosystem genetics: from genes to ecosystems. Nat Rev Genet. 2006;7:510–23. 32. Zaidem M, Groen SC, Purugganan MD. Evolutionary and ecological functional genomics, from lab to the wild. Plant J. 2018; https://doi.org/10.1111/tpj.14167. 6. Johnson MTJ, Agrawal AA. Plant genotype and environment interact to shape a diverse arthropod community on evening primrose (Oenothera biennis). Ecology. 2005;86:874–85. 33. Symonds VV, Godoy AV, Alconada T, et al. References Kallenbach M, Bonaventure G, Gilardoni PA, et al. Empoasca leafhoppers attack wild tobacco plants in a jasmonate-dependent manner and identify jasmonate mutants in natural populations. Proc Natl Acad Sci U S A. 2012;109:E1548–57. 40. Hülskamp M. Plant trichomes: a model for cell differentiation. Nat Rev Mol Cell Biol. 2004;5:471–80. 41. Ishida T, Kurata T, Okada K, Wada T. A genetic regulatory network in the development of trichomes and root hairs. Annu Rev Plant Biol. 2008;59:365–86. van Poecke RMP. Arabidopsis-insect interactions. Arabidopsis Book. 2007;5: 15. The 1001 Genomes Consortium. 1,135 genomes reveal the global pattern of polymorphism in Arabidopsis thaliana. Cell. 2016;166:481–91. 42. Grebe M. The patterning of epidermal hairs in Arabidopsis - updated. Cur Opin Plant Biol. 2012;15:31–7. 16. Handley R, Ekbom B, Ågren J. Variation in trichome density and resistance against a specialist insect herbivore in natural populations of Arabidopsis thaliana. Ecol Entomol. 2005;30:284–92. 43. Hauser MT, Harr B, Schlötterer C. Trichome distribution in Arabidopsis thaliana and its close relative Arabidopsis lyrata: molecular analysis of the candidate gene GLABROUS1. Mol Biol Evol. 2001;18:1754–63. 17. Mauricio R. Costs of resistance to natural enemies in field populations of the annual plant Arabidopsis thaliana. Am Nat. 1998;151:20–8. 17. Mauricio R. Costs of resistance to natural enemies in field populations of the annual plant Arabidopsis thaliana. Am Nat. 1998;151:20–8. 44. Kawagoe T, Shimizu KK, Kakutani T, Kudoh H. Coexistence of trichome variation in a natural plant population: a combined study using ecological and candidate gene approaches. PLoS One. 2011;6:e22184. 18. Sato Y, Kudoh H. Herbivore-mediated interaction promotes the maintenance of trichome dimorphism through negative frequency dependent selection. Am Nat. 2017;190:E67–77. 45. Larkin JC, Walker JD, Bolognesi-Winfield AC, et al. Allele-specific interactions between ttg and gl1 during trichome development in Arabidopsis thaliana. Genetics. 1999;151:1591–604. 19. Falk KL, Kästner J, Bodenhausen N, et al. The role of glucosinolates and the jasmonic acid pathway in resistance of Arabidopsis thaliana against molluscan herbivores. Mol Ecol. 2014;23:1188–203. 46. Shimizu KK. Ecology meets molecular genetics in Arabidopsis. Popul Ecol. 2002;44:221–33. 20. Kliebenstein D, Pedersen D, Barker B, Mitchell-Olds T. Comparative analysis of quantitative trait loci controlling glucosinolates, myrosinase and insect resistance in Arabidopsis thaliana. Genetics. 2002;161:325–32. 47. Wilczek AM, Roe JL, Knapp MC, et al. Effects of genetic perturbation on seasonal life history plasticity. Science. 2009;323:930–4. 21. Müller C. Interactions between glucosinolate- and myrosinase-containing plants and the sawfly Athalia rosae. Phytochem Rev. 2009;8:121–34. 48. Received: 10 March 2018 Accepted: 11 March 2019 Received: 10 March 2018 Accepted: 11 March 2019 27. Davila Olivas NH, Kruijer W, Gort G, et al. Genome-wide association analysis reveals distinct genetic architectures for single and combined stress responses in Arabidopsis thaliana. New Phytol. 2017;213:838–51. Author details 1 1PRESTO, Japan Science and Technology Agency, Kawaguchi 332-0012, Japan. 2Research Institute for Food and Agriculture, Ryukoku University, Yokotani 1-5, Seta Oe-cho, Otsu, Shiga 520-2194, Japan. 3Department of Evolutionary Biology and Environmental Studies, University of Zurich, Winterthurerstrasse 190, 8057 Zurich, Switzerland. 4Kihara Institute for Biological Research, Yokohama City University, 641-12 Maioka, 244-0813 Totsuka-ward, Yokohama, Japan. 5Department of Plant Life Sciences, Faculty Additional file 4: Effects of trichome density, the first and second principal component (PC1 and PC2) of aliphatic glucosinolates (GSLs), presence of flowering stems and initial plant sizes on insect abundance Additional file 4: Effects of trichome density, the first and second principal component (PC1 and PC2) of aliphatic glucosinolates (GSLs), presence of flowering stems and initial plant sizes on insect abundance Page 11 of 12 Page 11 of 12 Sato et al. BMC Plant Biology (2019) 19:163 Sato et al. BMC Plant Biology (2019) 19:163 of Agriculture, Ryukoku University, Yokotani 1-5, Seta Oe-cho, Otsu, Shiga 520-2194, Japan. 25. Züst T, Joseph B, Shimizu KK, et al. Using knockout mutants to reveal the growth costs of defensive traits. Proc R Soc Lond B. 2011;278:2598–603. 26. Davila Olivas NH, Frago E, Thoen MPM, et al. Natural variation in life history strategy of Arabidopsis thaliana determines stress responses to drought and insects of different feeding guilds. Mol Ecol. 2017;26:2959–77. Received: 10 March 2018 Accepted: 11 March 2019 Received: 10 March 2018 Accepted: 11 March 2019 References Mapping quantitative trait loci in multiple populations of Arabidopsis thaliana identifies natural allelic variation for trichome density. Genetics. 2005;169:1649–58. 7. Keith AR, Bailey JK, Whitham TG. A genetic basis to community repeatability and stability. Ecology. 2010;91:3398–406. 34. Bloomer RH, Juenger TE, Symonds VV. Natural variation in GL1 and its effects on trichome density in Arabidopsis thaliana. Mol Ecol. 2012;21:3501–15. 8. Maddox GD, Root RB. Resistance to 16 diverse species of herbivorous insects within a population of goldenrod, Solidago altissima: genetic variation and heritability. Oecologia. 1987;72:8–14. 35. Atwell S, Huang YS, Vilhjálmsson BJ, et al. Genome-wide association study of 107 phenotypes in Arabidopsis thaliana inbred lines. Nature. 2010;465:627–31. 36. Roux F, Bergelson J. The genetics underlying natural variation in the biotic interactions of Arabidopsis thaliana: the challenges of linking evolutionary genetics and community ecology. Curr Top Dev Biol. 2016;119:111–56. 9. Kivimäki M, Kärkkäinen K, Gaudeul M, et al. Gene, phenotype and function: GLABROUS1 and resistance to herbivory in natural populations of Arabidopsis lyrata. Mol Ecol. 2007;16:453–62. 9. Kivimäki M, Kärkkäinen K, Gaudeul M, et al. Gene, phenotype and function: GLABROUS1 and resistance to herbivory in natural populations of Arabidopsis lyrata. Mol Ecol. 2007;16:453–62. 37. Chan EKF, Rowe HC, Kliebenstein DJ. Understanding the evolution of defense metabolites in Arabidopsis thaliana using genome-wide association mapping. Genetics. 2010;185:991–1007. 10. Løe G, Toräng P, Gaudeul M, Ågren J. Trichome production and spatiotemporal variation in herbivory in the perennial herb Arabidopsis lyrata. Oikos. 2007;116:134–42. 10. Løe G, Toräng P, Gaudeul M, Ågren J. Trichome production and spatiotemporal variation in herbivory in the perennial herb Arabidopsis lyrata. Oikos. 2007;116:134–42. 38. Oppenheimer DG, Herman PL, Sivakumaran S, Esch J, Marks MD. A myb gene required for leaf trichome differentiation in Arabidopsis is expressed in stipules. Cell. 1991;67:483–93. 11. Sato Y, Kudoh H. Tests of associational defence provided by hairy plants for glabrous plants of Arabidopsis halleri subsp. gemmifera against insect herbivores. Ecol Entomol. 2015;40:269–79. 39. Yoshida Y, Sano R, Wada T, et al. Jasmonic acid control of GLABRA3 links inducible defense and trichome patterning in Arabidopsis. Development. 2009;136:1039–48. 12. Kessler A, Halitschke R, Baldwin IT. Silencing the jasmonate cascade: induced plant defenses and insect populations. Science. 2004;305:665–8. 13. Kallenbach M, Bonaventure G, Gilardoni PA, et al. Empoasca leafhoppers attack wild tobacco plants in a jasmonate-dependent manner and identify jasmonate mutants in natural populations. Proc Natl Acad Sci U S A. 2012;109:E1548–57. 13. References Kerwin R, Feusier J, Corwin J, et al. Natural genetic variation in Arabidopsis thaliana defense metabolism genes modulates field fitness. eLife. 2015;4:e05604. 49. Heidel AJ, Clarke JD, Antonovics J, Dong X. Fitness costs of mutations affecting the systemic acquired resistance pathway in Arabidopsis thaliana. Genetics. 2004;168:2197–206. 22. Reymond P, Bodenhausen N, van Poecke RMP, et al. A conserved transcript pattern in response to a specialist and a generalist herbivore. Plant Cell. 2004;16:3132–47. 50. Shimizu KK, Tsuchimatsu T. Evolution of selfing: recurrent patterns in molecular adaptation. Ann Rev Ecol Evol Syst. 2015;46:593–622. 23. Mewis I, Tokuhisa JG, Schultz JC, Appel HM, Ulrichs C, Gershenzon J. Gene expression and glucosinolate accumulation in Arabidopsis thaliana in response to generalist and specialist herbivores of different feeding guilds and the role of defense signaling pathways. Phytochemistry. 2006;67:2450–62. 51. Chiang GC, Barua D, Dittmar E, et al. Pleiotropy in the wild: the dormancy gene DOG1 exerts cascading control on life cycles. Evolution. 2013;67:883–93. 24. Kempema LA, Cui X, Holzer FM, Walling LL. Arabidopsis transcriptome changes in response to phloem-feeding silverleaf whitefly nymphs: similarities and distinctions in responses to aphids. Plant Physiol. 2007;143:849–65. 52. Bentsink L, Hanson J, Hanhart CJ, et al. Natural variation for seed dormancy in Arabidopsis is regulated by additive genetic and molecular pathways. Proc Natl Acad Sci U S A. 2010;107:4264–9. Page 12 of 12 Page 12 of 12 Page 12 of 12 Sato et al. BMC Plant Biology (2019) 19:163 Sato et al. BMC Plant Biology (2019) 19:163 53. Thompson L. The spatiotemporal effects of nitrogen and litter on the population dynamics of Arabidopsis thaliana. J Ecol. 1994;82:63–88. 79. Barker AM, Molotsane R, Müller C, et al. Chemosensory and behavioural responses of the turnip sawfly, Athalia rosae, to glucosinolates and isothiocyanates. Chemoecology. 2006;16:209–18. 54. Taylor MA, Cooper MD, Sellamuthu R, et al. Interacting effects of genetic variation for seed dormancy and flowering time on phenology, life history, and fitness of experimental Arabidopsis thaliana populations over multiple generations in the field. New Phytol. 2017;216:291–302. 80. Beran F, Pauchet Y, Kunert G, et al. Phyllotreta striolata flea beetles use host plant defense compounds to create their own glucosinolate-myrosinase system. Proc Natl Acad Sci U S A. 2014;111:7349–54. 55. Shindo C, Aranzana MJ, Lister C, et al. Role of FRIGIDA and FLOWERING LOCUS C in determining variation in flowering time of Arabidopsis. Plant Physiol. 2005;138:1163–73. 56. Arany AM, de Jong TJ, van der Meijden E. References Herbivory and abiotic factors affect population dynamics of Arabidopsis thaliana in a sand dune area. Plant Biol. 2005;7:549–56. 83. Li F, Kitashiba H, Nishio T. Association of sequence variation in brassica GLABRA1 orthologs with leaf hairiness. Mol Breeding. 2011;28:577–84. 57. Harvey JA, Witjes LMA, Benkirane M, Duyts H, Wagenaar R. Nutritional suitability and ecological relevance of Arabidopsis thaliana and Brassica oleracea as food plants for the cabbage butterfly, Pieris rapae. Plant Ecol. 2007;189:117–26. 58. Wentzell AM, Rowe HC, Hansen BG, Ticconi C, Halkier BA, Kliebenstein DJ. Linking metabolic QTLs with network and cis-eQTLs controlling biosynthetic pathways. PLoS Genet. 2007;3:e162. 59. Joseph B, Corwin JA, Zust T, et al. Hierarchical nuclear and cytoplasmic genetic architectures for plant growth and defense within Arabidopsis. Plant Cell. 2013;25:1929–45. 60. Brachi B, Meyer CG, Villoutreix R, et al. Coselected genes determine adaptive variation in herbivore resistance throughout the native range of Arabidopsis thaliana. Proc Natl Acad Sci U S A. 2015;112:4032–7. 61. Okamura Y, Sawada Y, Hirai MY, Murakami M. Effects of different secondary metabolite profiles in plant defense syndromes on specialist and generalist herbivores. Entomol Sci. 2016;19:97–103. 62. Ruppel RF. Cumulative insect-days as an index of crop protection. J Econ Entomol. 1983;76:375–7. 63. R Core Team. R: A language and environment for statistical computing. Vienna: R Foundation for Statistical Computing; 2015. http://www.R-project. org/;. 64. Oksanen J, Blanchet FG, Kindt R, et al. vegan: community ecology package. R package version 2.3–4. http://CRAN.R-project.org/package=vegan; 2015. 65. Pinheiro J, Bates D, DebRoy S, et al. nlme: linear and nonlinear mixed effects models. R package version 3.1–126. URL: http://CRAN.R-project.org/ package=nlme; 2016. 66. Patterson HD, Thompson R. Maximum likelihood estimation of variances. Biometrika. 1971;58:545–54. 67. Falconer DS. Introduction to quantitative genetics. In: Pearson Education India. 3rd ed; 1989. 68. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J Royal Stat Soc B. 1995;57:289–300. 69. Mauricio R, Rausher MD. Experimental manipulation of putative selective agents provides evidence for the role of natural enemies in the evolution of plant defense. Evolution. 1997;51:1435–44. 70. Bidart-Bouzat MG, Kliebenstein DJ. Differential levels of insect herbivory in the field associated with genotypic variation in glucosinolates in Arabidopsis thaliana. J Chem Ecol. 2008;34:1026–37. 71. Feeny P. Plant apparency and chemical defense. In: Wallace JW, Mansell RL, editors. Biochemical interaction between plants and insects. Boston, MA: Springer; 1976. p. 1–40. 72. Carmona D, Lajeunesse MJ, Johnson MTJ. 79. Barker AM, Molotsane R, Müller C, et al. Chemosensory and behavioural responses of the turnip sawfly, Athalia rosae, to glucosinolates and isothiocyanates. Chemoecology. 2006;16:209–18. 80. Beran F, Pauchet Y, Kunert G, et al. Phyllotreta striolata flea beetles use host plant defense compounds to create their own glucosinolate-myrosinase system. Proc Natl Acad Sci U S A. 2014;111:7349–54. 81. Lankau R, Strauss S. Community complexity drives patterns of natural selection on a chemical defense of Brassica nigra. Am Nat. 2008;171:150–61. 82. Lankau RA, Kliebenstein DJ. Competition, herbivory and genetics interact to determine the accumulation and fitness consequences of a defence metabolite. J Ecol. 2009;97:78–88. 83. Li F, Kitashiba H, Nishio T. Association of sequence variation in brassica GLABRA1 orthologs with leaf hairiness. Mol Breeding. 2011;28:577–84. References Plant traits that predict resistance to herbivores. Funct Ecol. 2011;25:358–67. 72. Carmona D, Lajeunesse MJ, Johnson MTJ. Plant traits that predict resistance to herbivores. Funct Ecol. 2011;25:358–67. 73. Schlinkert H, Westphal C, Clough Y, et al. Feeding damage to plants increases with plant size across 21 Brassicaceae species. Oecologia. 2015;179:455–66. 74. Travers-Martin N, Müller C. Matching plant defence syndromes with performance and preference of a specialist herbivore. Funct Ecol. 2008;22:1033–43. 75. Agerbirk N, Ørgaard M, Nielsen JK. Glucosinolates, flea beetle resistance, and leaf pubescence as taxonomic characters in the genus Barbarea (Brassicaceae). Phytochemistry. 2003;63:69–80. 76. Soroka JJ, Holowachuk JM, Gruber MY, Grenkow LF. Feeding by flea beetles (Coleoptera: Chrysomelidae; Phyllotreta spp.) is decreased on canola (Brassica napus) seedlings with increased trichome density. J Econ Entomol. 2011;104:125–36. 77. Züst T, Heichinger C, Grossniklaus U, et al. Natural enemies drive geographic variation in plant defenses. Science. 2012;338:116–9. 77. Züst T, Heichinger C, Grossniklaus U, et al. Natural enemies drive geographic variation in plant defenses. Science. 2012;338:116–9. 78. Shirakawa M, Hara-Nishimura I. Specialized vacuoles of myrosin cells: chemical defense strategy in Brassicales plants. Plant Cell Physiol. 2018;59: 1309–16. 78. Shirakawa M, Hara-Nishimura I. Specialized vacuoles of myrosin cells: chemical defense strategy in Brassicales plants. Plant Cell Physiol. 2018;59: 1309–16.
https://openalex.org/W3090517585	https://hal.science/hal-03602835/file/Chiara%20Barri%20et%20al%202021%20Nanotechnology32%20025303.pdf	English	null	Engineering of the spin on dopant process on silicon on insulator substrate	Nanotechnology	2,020	cc-by	6,895	Engineering of the spin on dopant process on silicon on insulator substrate Chiara Barri, Erfan Mafakheri, Luca Fagiani, Giulio Tavani, Andrea Barzaghi, Daniel Chrastina, Alexey Fedorov, Jacopo Frigerio, Mario Lodari, Francesco Scotognella, et al. To cite this version: Chiara Barri, Erfan Mafakheri, Luca Fagiani, Giulio Tavani, Andrea Barzaghi, et al.. Engineering of the spin on dopant process on silicon on insulator substrate. Nanotechnology, 2020, 32 (2), pp.025303. hal-03602835 HAL Id: hal-03602835 https://hal.science/hal-03602835v1 Submitted on 9 Mar 2022 L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. Nanotechnology PAPER • OPEN ACCESS Abstract We report on a systematic analysis of phosphorus diffusion in silicon on insulator thin ﬁlm via spin-on-dopant process (SOD). This method is used to provide an impurity source for semiconductor junction fabrication. The dopant is ﬁrst spread into the substrate via SOD and then diffused by a rapid thermal annealing process. The dopant concentration and electron mobility were characterized at room and low temperature by four-probe and Hall bar electrical measurements. Time-of-ﬂight-secondary ion mass spectroscopy was performed to estimate the diffusion proﬁle of phosphorus for different annealing treatments. We ﬁnd that a high phosphorous concentration (greater than 1020 atoms cm−3) with a limited diffusion of other chemical species and allowing to tune the electrical properties via annealing at high temperature for short time. The ease of implementation of the process, the low cost of the technique, the possibility to dope selectively and the uniform doping manufactured with statistical process control show that the methodology applied is very promising as an alternative to the conventional doping methods for the implementation of optoelectronic devices. Keywords: spin on dopant, four-probe, SOI doping, P diffusion, ToF-SIMS Keywords: spin on dopant, four-probe, SOI doping, P diffusion, ToF-SIMS (Some ﬁgures may appear in colour only in the online journal) Chiara Barri1,2, Erfan Mafakheri2, Luca Fagiani1,2, Giulio Tavani1,2, Andrea Barzaghi1, Daniel Chrastina1, Alexey Fedorov2, Jacopo Frigerio1 , Mario Lodari3, Francesco Scotognella1, Elisa Arduca4, Marco Abbarchi5 , Michele Perego4 and Monica Bollani2 1 L-NESS, Department of Physics, Politecnico di Milano, Via Anzani 42, I-22100 Como, Italy 2 IFN-CNR, LNESS laboratory, Como, Italy 3 QuTech and Kavli Institute of Nanoscience, Delft University of Technology, PO Box 5046, 2600 Delft, The Netherlands 4 CNR-IMM, Unit of Agrate Brianza, via Olivetti 2, I-20864 Agrate Brianza, Italy 5 Aix Marseille Univ, Université de Toulon, CNRS, IM2NP Marseille, France 1 L-NESS, Department of Physics, Politecnico di Milano, Via Anzani 42, I-22100 Como, Italy 2 IFN-CNR, LNESS laboratory, Como, Italy 3 QuTech and Kavli Institute of Nanoscience, Delft University of Technology, PO Box 5046, 2600 GA Delft, The Netherlands 4 CNR-IMM, Unit of Agrate Brianza, via Olivetti 2, I-20864 Agrate Brianza, Italy 5 Aix Marseille Univ, Université de Toulon, CNRS, IM2NP Marseille, France 2 IFN-CNR, LNESS laboratory, Como, Italy 3 QuTech and Kavli Institute of Nanoscience, Delft University of Technology, PO Box 5046, 2600 GA Delft The Netherlands 4 CNR-IMM, Unit of Agrate Brianza, via Olivetti 2, I-20864 Agrate Brianza, Italy 5 Aix Marseille Univ, Université de Toulon, CNRS, IM2NP Marseille, France E-mail: monica.bollani@ifn.cnr.it Received 11 July 2020, revised 19 September 2020 Accepted for publication 2 October 2020 Published 20 October 2020 Abstract Received 11 July 2020, revised 19 September 2020 Accepted for publication 2 October 2020 Published 20 October 2020 Received 11 July 2020, revised 19 September 2020 Accepted for publication 2 October 2020 Published 20 October 2020 Engineering of the spin on dopant process on silicon on insulator substrate To cite this article: Chiara Barri et al 2021 Nanotechnology 32 025303 View the article online for updates and enhancements. This content was downloaded from IP address 147.94.135.1 on 17/11/2020 at 15:04 This content was downloaded from IP address 147.94.135.1 on 17/11/2020 at 15:04 his content was downloaded from IP address 147.94.135.1 Nanotechnology https://doi.org/10.1088/1361-6528/abbdda Nanotechnology 32 (2021) 025303 (9pp) © 2020 The Author(s). Published by IOP Publishing Ltd Printed in the UK Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. © 2020 The Author(s). Published by IOP Publishing Ltd Printed in the UK 1 1. Introduction Moreover, this approach lacks spatial control as it affects homogeneously the full wafer surface. In doping furnaces, doping is achieved by exposing the wafers to a ﬂux of dopant atoms from the sublimation/evaporation of a solid/liquid source or from a gas source. Dopant diffusion into the substrate is promoted by heating the wafers for the time needed to achieve the desired doping proﬁle. Typically, doping furnaces can be hosted only in large facilities and can accommodate a large quantity of substrates. However, due to the hazardous nature of the dopant sources, this step often requires a very long time in a controlled environ- ment [11, 12]. proﬁle, high yields, and possibility to introduce both n- and p-type doping [26, 27]. The advantage of SOD over the aforementioned methods relies in its low cost, simple control of the process and its adaptability to different platforms, achieving a uniformity of the coated ﬁlm on large substrates. Moreover, it eliminates the need for dangerous gases and their related safety precautions, it is more environmentally friendly, and reduces manufacturing costs. Additionally, unlike CVD method and furnace-based doping approaches, the SOD process allows selective doping via lithographically patterned areas, similarly to ion implantation [28]. The possibility of diffusing the SOD only in speciﬁc areas is very interesting for new nanoelectronic applications: for instance, in the ﬁn ﬁeld effect transistor the channel has not to be doped, giving the gate higher ability to control threshold voltage, which is a very important property when the transistor is scaling down, while the source and drain regions can be doped locally by SOD process. Here, we have targeted high doping levels -even though perfectly localized- to address the needs of FinFET devices for plasmonic and quantum applications [29, 30]. Clearly, such devices will not be the only perspec- tive potentials of applications, but the necessity of having a high level of localized doping, makes them one of most challenging devices. To overcome the difﬁculties of conventional technologies and their correlated costs, tremendous research efforts have been taken in recent years to develop new strategies for introducing dopants into semiconductor materials such as monolayer doping [13–16] or spin on dopant (SOD) [17–23]. 1. Introduction device fabrication, the doping is carried out using mainly three methods: ion implantation, in situ co-deposition, or doping furnaces. Ion implantation is an excellent process that allows high level of control on both the dose of implanted impurity atoms and the implantation depth [4–6]. However, the conventional ion implantation process, which relies on the bombardment of semiconductors with energetic ions, is incompatible with nanostructured materials, such as one- dimensional (1D) nanowires. Furthermore, it induces severe crystal damage which requires also an annealing step to reconstruct the crystal and activate the implanted impurities. Very high temperature are usually needed, limiting its applicability to platforms that can substain high thermal The functionality of semiconductors relies strongly on the impurities that can be added to the intrinsic materials to change their electrical, physical, and optical properties. Uni- form doping to realize ultrashallow junctions at transistor-like device source and drain has been a key point of the effort toward device scaling [1–3]. In traditional semiconductor Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. 0957-4484/21/025303+09$33.00 1 C Barri et al Nanotechnology 32 (2021) 025303 Figure 1. Representation of the doping treatment using SOD P508. (a) SOI wafer sketch; (b) SOI wafer spin coated with SOD P508; (c) annealing steps on a hot plate; (d) activation of the diffusion of P using the RTA treatment; (e) doped sample after the cleaning procedure. Figure 1. Representation of the doping treatment using SOD P508. (a) SOI wafer sketch; (b) SOI wafer spin coated with SOD P508; (c) annealing steps on a hot plate; (d) activation of the diffusion of P using the RTA treatment; (e) doped sample after the cleaning procedure. budget processes. These aspects make ion implantation a delicate and costly process. In co-deposition techniques, such as chemical vapor deposition (CVD), the dopant is directly introduced in the growth chamber by a carrier gas [7–10]. This approach offers an excellent control of the dopant con- centration and solubility, which can be tuned by changing the vapor pressure of the source and the temperature of the sub- strate during the thin ﬁlms deposition. However, in CVD method the presence of toxic and explosive gases requires many safety precautions, making it an expensive and dan- gerous technique. 1. Introduction Speciﬁcally, the adhesion promoter TI PRIME was spin-coated at 3000 rpm for 5 s, then the positive resist AZ 5214E was spin-coated in two different steps at 750 rpm for 5 s and at 4000 rpm for 40 s. After the optical lithography, the samples were dipped for 15 s in dilute HF to remove the native oxide layer on the surface right before the contacts deposition. Finally, to obtain ohmic contacts we performed the same procedure reported for the four-probe devices. To improve the quality of the contacts, by lowering the contacts resistivity, the samples were annealed in a furnace at 300 °C for 180 s [31]. Samples were characterized electrically according to the Van der Pauw method [32], and carrier density and Hall mobility were extracted from Hall effect measurements. This electrical characterization was carried out using a permanent magnet creating a 0.204 T magnetic ﬁeld at room temperature, and an electromagnet which applied a magnetic ﬁeld ranging from 0 to 1 T at 3 K with an accuracy of 3 mK. The temperature has been measured with a Cernox® negative temperature coefﬁcient RTD made of ceramic oxy- nitride. The phosphorous proﬁle was obtained by ToF-SIMS analyses operating in negative mode and sputtering an area of 500×500 μm2 using Cs+ ions at 1 keV and an area of 50×50 μm2 using Ga+ ions at 25 keV. Phosphorous quan- tiﬁcation was performed according to a calibration procedure that was widely discussed in a previous paper [33]. Finally, by a combination of electron beam lithography and reactive ion etching a Si-based nanodevice using the same SOI sub- strate before descripted is realized: a SiO2 hard mask is 1. Introduction In this work we applied a SOD treatment with rapid thermal annealing process (RTA) to drive-in the dopants from the SOD polymeric ﬁlm into the silicon substrate, without mod- iﬁng the surface energy, requiring covalent attachment methods, such in surface-initiated polymerization [2] and grafting of functionalized polymers [24, 25]. The SOD ﬁlm provides an impurity source for semiconductor junction fab- rication, and due to the convenience and simplicity of use, it greatly reduces the reliance on the traditional facilities for doping diffusion. SOD based treatments have become increasingly popular due to their simple, uniform doping In this work we systematically investigate the diffusion of phosphorous from a SOD source in silicon on insulator (SOI) wafers, RTA treatments are performed using different temperatures and process times, leading to the optimization of the desired ﬁnal doping concentration. By a systematic ana- lysis of the electrical characterizations, the values of P dif- fusion electrically activated, the doping and mobility were extracted and compared with the chemical P diffusion obtained through time of ﬂight secondary ion mass spectro- metry (ToF-SIMS) analyses measurements, demonstrating a successful control of dopant diffusion on SOI susbtrates. At the end, a localized doped Si-based nanodevice is tested as validation of the SOD procedure. 2 C Barri et al Nanotechnology 32 (2021) 025303 Table 1. Parameters of the rapid thermal annealing treatments of all the samples analyzed. Series A–B–C–D–E samples used for four-probe measurements: for a ﬁxed temperature, different annealing times have been characterized, keeping the ramp up (9 s) constant. Series G samples used in ToF-SIMS analyses. Series H samples used for mobility measurements. Table 1. Parameters of the rapid thermal annealing treatments of all the samples analyzed. Series A–B–C–D–E samples used for four-probe measurements: for a ﬁxed temperature, different annealing times have been characterized, keeping the ramp up (9 s) constant. Series G samples used in ToF-SIMS analyses. Series H samples used for mobility measurements. evaporation of 7 nm of Ti, 50 nm of Au and 100 nm of Al followed by a lift-off process (ﬁgures 2(g), (h)). The Ti was used as a prime to increase the adhesion of Au to the Si substrate, Al allowed us to increase the ﬁnal thickness of the metals to facilitate the bonding. The Hall bar shaped devices consist of a mesa, patterned by optical lithography and transferred to the substrate by dry etching (CF4 plasma), (ﬁgure 2(b)). 2. Methods The optimization of the SOD treatments was carried out using a (001) SOI substrate, featuring a 125 nm thick Si layer on a 1 μm buried oxide (BOX) (ﬁgure 1(a)). After a cleaning process (ultrasonic treatment with acetone, isopropanol, and N-methyl-2-pyrrolidone) the SOI wafers were spin-coated with the commercial SOD P508 (composition: 8% P and 5% SiO2) for 30 s at speed of 3000 rpm (ﬁgure 1(b)). This process was carried out in a temperature and humidity-controlled environment to prevent cloudiness of the SOD layer. To evaporate the volatile solvent contained in P508, the samples were baked on a hot plate at 55 °C for 10 min and then at 120 °C for 15 min (ﬁgure 1(c)). The two baking steps were necessary to avoid cracking of the SOD ﬁlm. The diffusion of the dopant in silicon was performed by RTA. Different annealing temperatures and times were tested to study the behavior of the diffusion proﬁle and the resulting doping level as reported in table 1. In each process the RTA system ramped up in 9 s to the annealing temperature, which was maintained for process times in the 10–90 s range. The samples were cooled down in a controlled way by ﬂowing the N2 in the chamber (ﬁgure 1(d)). The remaining SOD layer was removed by rinsing in a HF solution (5% vol) for 1 min, followed by an oxygen plasma asher treatment for 30 min at 1000 W to remove any organic contamination. A ﬁnal HF rinse (5% vol) for 30 s (ﬁgure 1(e)) is performed to remove any SOD residues. The reproducibility of the process was validated by processing three different samples batches. After the doping treatment by SOD and the cleaning of the samples, devices for four-probe measurements and Hall bar shaped devices were fabricated (ﬁgure 2). The four-probe devices fabrication includes just one step of optical lithography, where a positive resist is patterned before the metal deposi- tion. The metallic contacts were obtained by electron beam 3 C Barri et al Nanotechnology 32 (2021) 025303 d and the SOD spincoated in selective areas. The I/V terization at room temperatures has been carried out to m the localized doping optimized by SOD treatment. four-point probe method [34]. The two resistances RA and RB associated with the corresponding contacts are reported in ﬁgure 3. 2. Methods From these characteristic resistances the sheet resistance RS and the electrical resistivity r can be obtained 2. Representation of the fabrication steps. (a) UV exposure of the spin-coated sample; (b) after development, the Si layer is etched by asma; (c) second exposure using a negative resist (d) metal evaporation. (e),(f) 3D representation and plan view of the Hall bar es: contacts 2–4 and 3–5 are used for longitudinal voltage, contacts 2–3 and 4–5 for transverse voltage measures; (g), (h) 3D ntation and plan view of the contacts used for four-probe. gy ( ) re 2 Representation of the fabrication steps (a) UV e pos re of the spin coated sample; (b) after de elopment the Si la Figure 2. Representation of the fabrication steps. (a) UV exposure of the spin-coated sample; (b) after development, the Si layer is etched by CF4 plasma; (c) second exposure using a negative resist (d) metal evaporation. (e),(f) 3D representation and plan view of the Hall bar structures: contacts 2–4 and 3–5 are used for longitudinal voltage, contacts 2–3 and 4–5 for transverse voltage measures; (g), (h) 3D representation and plan view of the contacts used for four-probe. four-point probe method [34]. The two resistances RA and RB associated with the corresponding contacts are reported in ﬁgure 3. From these characteristic resistances the sheet resistance RS and the electrical resistivity r can be obtained by: realized and the SOD spincoated in selective areas. The I/V characterization at room temperatures has been carried out to conﬁrm the localized doping optimized by SOD treatment. p = + R R R 2 ln 2 2 , 1 S A B ( ) ( ) Table 2. Parameters used in equation (5) [30]. Table 2. Parameters used in equation (5) [30]. A0 A1 A2 A3 −3.0951 −3.2303 −1.2024 −0.13679 B0 B1 B2 B3 1 1.0205 0.38382 0.041338 where xSi is the thickness of the conductive layer [35]. In this analysis, the top Si layer (xSi=125 nm) is considered com- pletely and homogeneously doped, neglecting that P atoms have a limited penetration depth as detected by the ToF-SIMS analyses reported below. Since the contacts were positioned in the middle of the sample and not in the corners of a square sample, a correction factor of 2 (in the equation (1)) is reported in the numerator in the classical Van der Pauw conﬁguration [36]. The model described by Thurber et al was used to ﬁnd concentration using the measured resistivity [37]. According to this model it is possible to deﬁne a parameter x as the logarithm of the ratio between the measured resistivity r and the normalization value r = W 1 cm 0 ò p = = µ +¥ N x c x t x x c Dt t 1 , d 1 2 , 7 D Si 0 Si 0 ( ) ( ) 7 ( ) where = c x t c , erfc , x Dt 0 2( ) ( ) / where = c x t c , erfc , x Dt 0 2( ) ( ) c0 being the doping con- centration at the SOD/silicon interface, and it is assumed to have a solubility limit of c0=3.5×1020 atoms cm−3 for phosphorus [39]. The doping concentration as a function of the square root of the annealing time is reported in ﬁgure 4, where different temperatures have been considered. From the angular coefﬁcient of linear ﬁt of the data m , fit the diffusivity of phosphorus in Si can be calculated: where = c x t c , erfc , x Dt 0 2( ) ( ) c0 being the doping con- centration at the SOD/silicon interface, and it is assumed to have a solubility limit of c0=3.5×1020 atoms cm−3 for phosphorus [39]. The doping concentration as a function of the square root of the annealing time is reported in ﬁgure 4, where different temperatures have been considered. Table 2. Parameters used in equation (5) [30]. From the angular coefﬁcient of linear ﬁt of the data m , fit the diffusivity of phosphorus in Si can be calculated: r r = x log 3 0 ( ) ⎛ ⎝ ⎜ ⎞ ⎠ ⎟ 3 ( ) 3 ( ) and the quantity P as the product between the electronic charge q, the measured resistivity r and the electrically active doping density ND: p = D m x c 2 . 8 fit Si 0 2 ( ) ⎛ ⎝⎜ ⎞ ⎠⎟ 8 ( ) From the experimental data, the diffusivity values for samples annealed at 860 °C, 880 °C, 900 °C, 915 °C and 920 °C are 1.73×10−14 cm2 s−1, 3.15×10−14 cm2 s−1, 7.10×10−14 cm2 s−1, 9.13×10−14 cm2 s−1 and 1.34×10−13 cm2 s−1 respectively. From the data of diffu- sivity taken from the four-probe measurements, it is possible to extrapolate the value of the activation energy (Ea) of phosphorus diffusion in Si according to the deﬁnition of the diffusivity by the Arrhenius equation = - D D e 0 E k T a B ( ) where kB is the Boltzmann constant and T the annealing temperature. From the linear ﬁt of the natural logarithm of diffusivity = - D D ln ln E k T 0 a B ( ) ( ) ( ) in function of T−1), the activation energy of 3.88 eV has been obtained. This value is in agreement with the phosphorus activation energy obtained by others using a similar SOD product but other annealing treatments [40, 41]. To characterize the chemical diffusion proﬁle of phosphorous, some ToF-SIMS analyses have been carried out (G series in table 1). The chemical proﬁles for different species diffused by RTA with various annealing From the experimental data, the diffusivity values for samples annealed at 860 °C, 880 °C, 900 °C, 915 °C and 920 °C are 1.73×10−14 cm2 s−1, 3.15×10−14 cm2 s−1, 7.10×10−14 cm2 s−1, 9.13×10−14 cm2 s−1 and 1.34×10−13 cm2 s−1 respectively. From the data of diffu- sivity taken from the four-probe measurements, it is possible to extrapolate the value of the activation energy (Ea) of phosphorus diffusion in Si according to the deﬁnition of the diffusivity by the Arrhenius equation = - D D e 0 E k T a B ( ) where kB is the Boltzmann constant and T the annealing temperature. r = P q N . 3. Results 1 ( ) The doping concentration ND was obtained by measuring the resistivity ρ of each sample with the dual conﬁguration r = R x , 2 S Si ( ) 2 ( ) r = R x , S Si 4 Nanotechnology 32 (2021) 025303 C Barri et al Figure 3. Schematic representation of the electrical measurement of the four-probe conﬁguration for the characteristic resistance (a) RA and (b) RB. possible to perform four-probe measurements at the various depths levels. By increasing the annealing time and temper- ature it was possible to obtain a high level of doping in sili- con. For an annealing time below 30 s a doping level of 1019 atoms cm−3 can be achieved at an annealing temperature of at least 880 °C, while on increasing the annealing time to 90 s, the same level can be obtained for annealing temperatures just at 860 °C. The maximum reached dopant density measured by four-probe method is about 7.8×1019 atoms cm−3 for 90 s annealing time at 915 °C. Figure 3. Schematic representation of the electrical measurement of the four-probe conﬁguration for the characteristic resistance (a) RA and (b) RB. Considering the samples annealed in the range between 860 °C and 915 °C and characterized by four-probes mea- surements, the doping concentration is a linear function of the square root of the annealing time as reported in ﬁgure 4. This same behavior is predicted by the diffusion model assuming that (1) the Si thickness is few times larger than the diffusion length of the dopant, and (2) the SOD acts as an inﬁnite source of dopant in contact with a semi-inﬁnite medium [38]. Following this model, the doping level is the integral of the diffusion proﬁle of the concentration of dopant normalized with the depth of the Si layer (xSi) in which the dopant dif- fuses. By applying this model the doping concentration has a linear dependence on the square root of the annealing time: Table 2. Parameters used in equation (5) [30]. 4 D ( ) 4 ( ) Then, solving equation (5) using the parameters reported in table 2 and the normalization factor P0 equal to 1 V s cm−2: = + + + + + + P P A A x A x A x B B x B x B x log10 0 0 1 2 2 3 3 0 1 2 2 3 3 ⎛ ⎝⎜ ⎞ ⎠⎟ 5 ( ) the doping concentration ND is obtained from equations (4) and (5): r = N P q 10 . 6 D P P log 0 0 ( ) ⎜ ⎟ ⎛ ⎝ ⎞ ⎠ From the linear ﬁt of the natural logarithm of diffusivity = - D D ln ln E k T 0 a B ( ) ( ) ( ) in function of T−1), the activation energy of 3.88 eV has been obtained. This value is in agreement with the phosphorus activation energy obtained by others using a similar SOD product but other annealing treatments [40, 41]. To characterize the chemical diffusion proﬁle of phosphorous, some ToF-SIMS analyses have been carried out (G series in table 1). The chemical proﬁles for different species diffused by RTA with various annealing 6 ( ) The ND values should be considered as a lower bound of the doping concentration. In fact in the evaluation of the resistivity, we considered the whole thickness of the SOI, 125 nm in our case. Although the Tof- SIMS characterizations (shown below) conﬁrm a gradient of the chemical diffusion of phosphorus, in this analysis we consider a uniform distribu- tion of the electrically active phosphorus because it was not 5 C Barri et al Nanotechnology 32 (2021) 025303 Figure 4. Linear ﬁtting of the experimental data obtained from four-probe measurements of samples annealed at 860 °C (R-squared 0.96) blue line at 880 °C (R squared 0 98) red line at 900 °C (R squared 0 98) gray line at 915 °C (R squared 0 99) green line and at 920 °C (R squared Figure 4. Linear ﬁtting of the experimental data obtained from four-probe measurements of samples annealed at 860 °C (R-squared 0.96) blue line, at 880 °C (R-squared 0.98) red line, at 900 °C (R squared 0.98) gray line, at 915 °C (R squared 0.99) green line and at 920 °C (R squared 0.97) red line. Figure 4. Table 2. Parameters used in equation (5) [30]. Sem image of the inner part of the four contacts with nanometric intrinsic channel (width 150 nm) before the SOD treatment; (a) ohmic behavior of the device measuring I in function of V between contact 1 and 3. The current reaches value of hundreds of μA. In order to continue the characterization of the samples, Hall bar measurements were performed in order to obtain the doping concentration and the electron mobility. The Hall mobility is obtained from the ratio between the Hall coefﬁcient of the material RH and the resistivity ρ. By exploiting the six-contact Hall bar geometry, an average value of the Hall coefﬁcient RH can be obtained in order to limit the negative effects caused by the non-uni- formities of the sample. The resistivity ρ, instead, is obtained from: r = V I Wx L , 9 Si ( ) 9 ( ) where V is the longitudinal voltage (ﬁgure 2(f)), W and L are respectively the width of the conductive channel of the Hall bar and the length between the voltage probes. The mea- sured data are reported in table 3. Figure 7. Sem image of the inner part of the four contacts with nanometric intrinsic channel (width 150 nm) before the SOD treatment; (a) ohmic behavior of the device measuring I in function of V between contact 1 and 3. The current reaches value of hundreds of μA. The doping concentrations are similar at 300 and 3 K, while the mobility values are higher at low temperature. The main factors that affect the mobility in an extrinsic semi- conductor are impurity scattering, phonon scattering, and electron–electron scattering: the role of impurity scattering becomes more important as the doping density increases or when the temperature decreases [42]. According to [43], samples characterized by high doping levels are more dif- ﬁcult to measure and are not particularly affected by the variation of the temperature. The measured mobility values are in agreement with those of silicon samples doped with P using different techniques [44, 45], conﬁrming the validity of doping mechanism with fast and low-cost SOD process. Finally, it is shown the capability of doping selective area through a Si-based nanodevice. From an SOI with a device layer of 145 nm are fabricated four Si pads linked through a 200 nm wide channel as reported in ﬁgure 7. Table 2. Parameters used in equation (5) [30]. Linear ﬁtting of the experimental data obtained from four-probe measurements of samples annealed at 860 °C (R-squared 0.96) blue line, at 880 °C (R-squared 0.98) red line, at 900 °C (R squared 0.98) gray line, at 915 °C (R squared 0.99) green line and at 920 °C (R squared 0.97) red line. Figure 5. Doping proﬁles for different chemical species (Si black line—P red line—SiO3 green line) diffused by RTA with annealing time of 5 s, 10 s and 20 s at the annealing temperature of 900 °C (a) and 915 °C (b). Figure 5. Doping proﬁles for different chemical species (Si black line—P red line—SiO3 green line) diffused by RTA with annealing time of 5 s, 10 s and 20 s at the annealing temperature of 900 °C (a) and 915 °C (b). time and temperatures were compared in ﬁgure 5. The longer the annealing time, the higher the concentration of P in the Si substrate was detected. However, by increasing the temper- ature, oxygen impurities (detected as SiO3) diffused from the SOD ﬁlms into the ﬁrst nm of the silicon ﬁlm, resulting in the formation of a partially oxidized SiOx layer in each sample, with a progressively increasing thickness as the annealing time increases, without limiting the electric transport 6 Nanotechnology 32 (2021) 025303 C Barri et al Figure 6. Linear ﬁtting of the natural logarithm of the diffusivity values extrapolated from the experimental data to obtain the phosphorous activation energy. C Barri et al Table 3. Experimental value of the doping density level and the electron mobility of 3 samples from Hall bar measurements carried out at 300 K and at 3 K. Table 3. Experimental value of the doping density level and the electron mobility of 3 samples from Hall bar measurements carried out at 300 K and at 3 K. Figure 6. Linear ﬁtting of the natural logarithm of the diffusivity values extrapolated from the experimental data to obtain the phosphorous activation energy. properties as reported by the mobility values in table 3. A phosphorous signal is detected in ﬁrst 15–30 nm with a relatively long tail extending towards the Si/BOX interface, conﬁrming a gradient of the phosphorous chemical diffusion. Figure 7. Table 2. Parameters used in equation (5) [30]. To selectively doped two of the four pads (respectively contacts 1 and 3) and to leave the inner nanometric intrinsic channel, a SiO2 mask is created ﬁgure 8. After the SOD doping treatment at 880 °C, an electrical catherization is performed between all contacts. The I–V curve reported in ﬁgure 7(a) shows an ohmic proﬁle between the two SOD doped contacts (con- tacts 1 and 3), while no current is observed between the undoped ones (pad 2 and pad4), conﬁrming the selective doping validation procedure. Figure 8. Sem characterization of the SiO2 hard mask that cover the nanometric channel and all the structures except for the contact 1 and contact 3. Figure 8. Sem characterization of the SiO2 hard mask that cover the nanometric channel and all the structures except for the contact 1 and contact 3. References [1] Peercy P S 2000 The drive to miniaturization Nature 406 1023–6 [19] Liu Y, Tanaka H, Umeyama N, Koga K, Khumpuang S, Nagao M, Matsukawa T and Hara S 2018 Investigation of piezoresistive effect in p-channel metal-oxide- semiconductor ﬁeld-effect transistors fabricated on circular silicon-on-insulator diaphragms using cost-effective minimal-fab process Japan. J. Appl. Phys. 57 06HD03 [2] Ho J C, Yerushalmi R, Jacobson Z A, Fan Z, Alley R L and Javey A 2008 Controlled nanoscale doping of semiconductors via molecular monolayers Nat. Mater. 7 62–7 [3] Blase X, Bustarret E, Chapelier C, Klein T and Marcenat C 2009 Superconducting group-IV semiconductors Nat. Mater. 8 375–82 [20] Sharp J, Lee W J, Ploog K, Umana-Membreno G A, Faraone L and Dell J M 2013 A novel technique for degenerate p-type doping of germanium Solid-State Electron. 89 146–52 [4] Yamamoto K and Itoh H 2006 XPS study of silicon surface after ultra-low-energy ion implantation Surf. Sci. 600 3753–6 [21] Boldrini V, Carturan S M, Maggioni G, Napolitani E, Napoli D R, Camattari R and De Salvador D 2017 Optimal process parameters for phosphorus spin-on-doping of germanium Appl. Surf. Sci. 392 1173–80 [5] Tetelbaum D I, Gorshkov O N, Trushun S A, Revin D G, Gaponova D M and Eckstein W 2000 The enhancement of luminescence in ion implanted Si quantum dots in SiO2matrix by means of dose alignment and doping Nanotechnology 11 295–7 g pp f [22] Li J, Cheng R, Liu C, Zhang P, Lu J, Chen K, Zhang R and Zhao Y 2017 High performance Ge ultra-shallow junctions fabricated by a novel formation technique featuring spin-on dopant and laser annealing for sub-10nm technology applications Microelectron. Eng. 168 1–4 [6] Tetelbaum D I, Trushin S A, Mikhaylov A N, Vasil’ev V K, Kachurin G A, Yanovskaya S G and Gaponova D M 2003 The inﬂuence of the annealing conditions on the photoluminescence of ion-implanted SiO2:Si nanosystem at additional phosphorus implantation Physica E 16 410–3 [23] Mathiot D, Lachiq A, Slaoui A, Noël S, Muller J C and Dubois C 1998 Phosphorus diffusion from a spin-on doped glass (SOD) source during rapid thermal annealing Mater. Sci. Semicond. Process. 1 34231236 [7] von Känel H, Chrastina D, Rössner B, Isella G, Hague J P and Bollani M 2004 High mobility SiGe heterostructures fabricated by low-energy plasma-enhanced chemical vapor deposition Microelectron. Eng. ORCID iDs [16] Perego M, Caruso F, Seguini G, Arduca E, Mantovan R, Sparnacci K and Laus M 2020 Doping of silicon by phosphorus end-terminated polymers: drive-in and activation of dopants J. Mater. Chem. C 8 10229 Jacopo Frigerio https://orcid.org/0000-0003-2117-2744 Marco Abbarchi https://orcid.org/0000-0002-2760-4766 Monica Bollani https://orcid.org/0000-0002-0078-5085 Jacopo Frigerio https://orcid.org/0000-0003-2117-2744 Marco Abbarchi https://orcid.org/0000-0002-2760-4766 Monica Bollani https://orcid.org/0000-0002-0078-5085 [17] O’Connell J, Biswas S, Duffy R and Holmes J D 2016 Chemical approaches for doping nanodevice architectures Nanotechnology 27 342002–37 [18] Liang D et al 2019 Spin-on doping of phosphorus on Ge with a 9 nm amorphous Si capping layer to achieve n+/p shallow junctions through rapid thermal annealing J. Phys. D: Appl. Phys. 52 195101 4. Conclusions In this work, we exploited the spin-on dopant process on SOI wafers. By tuning the RTA treatment parameters, we are able to diffuse the phosphorous into the silicon layer with different 7 Nanotechnology 32 (2021) 025303 C Barri et al depth proﬁles. In contrast with the typical doping processes, the SOD diffusion is a low cost process and it does not damage the crystal structure. By electrical characterization at room and low temperatures, the activated phosphorous con- centration, the carrier mobility, and P diffusivity have been measured. We show that by SOD treatment, a high level of n-type doping greater than 1×1020 atoms cm−3 can be reached. The Tof-SIMS analyses reveal that the phosphorous proﬁles depends to the RTA time. The possibility of doping with high doses, in conﬁned areas and at a low cost is of great interest for quantum devices, Fin-FET and similar devices. [9] Kawano A, Ishiwata H, Iriyama S, Okada R, Yamaguchi T, Takano Y and Kawarada H 2010 Superconductor-to- insulator transition in boron-doped diamond ﬁlms grown using chemical vapor deposition Phys. Rev. B 82 085318 [10] Kalyadin A E, Sobolev N A, Strel’chuk A M, Aruev P N, Zabrodskiy V V and Shek E I 2016 Effect of the fabrication conditions of SiGe LEDs on their luminescence and electrical properties Semiconductors 50 249–51 [11] Lee B W, Choi I, Jung C U and Lee S I 2008 Doping in MgB2 superconductors using a high-pressure furnace J. Magn. Magn. Mater. 320 e484–6 [12] Suzuki K, Miyata N and Kawamura K 1995 Resistivity of heavily doped polycrystalline silicon subjected to furnace annealing Japan. J. Appl. Phys. 34 1748–52 [13] Park C J, Jung S M, Kim J H and Shin M W 2020 Conformal doping strategy for ﬁn structures: tailoring of dopant proﬁle through multiple monolayer doping and capping layer control Semicond. Sci. Technol. 35 055028 Acknowledgments We acknowledge the EU H2020 FET-OPEN Narciso (grant no. 828890), the EU ERC Paideia (grant no. 816313), the PRIN NOMEN project (grant no. 2017MP7F8F) and the TEINVEIN project, Call Accordi per la Ricerca e l’Innova- zione, co-funded by POR FESR 2014–2020 (ID: 242092). [14] Gao X, Kolevatov I, Chen K, Guan B, Mesli A, Monakhov E and Dan Y 2020 Full activation of boron in silicon doped by self-assembled molecular monolayers ACS Appl. Electron. Mater. 2 268–74 [15] Katsumata R, Limary R, Zhang Y, Popere B C, Heitsch A T, Li M, Trefonas P and Segalman R A 2018 Mussel-inspired strategy for stabilizing ultrathin polymer ﬁlms and its application to spin-on doping of semiconductors Chem. Mater. 30 5285–92 References 76 279–84 [24] Perego M, Seguini G, Arduca E, Nomellini A, Sparnacci K, Antonioli D, Gianotti V and Laus M 2018 Control of doping level in semiconductors via self-limited grafting of phosphorus end-terminated polymers ACS Nano 12 178–86 [8] Chrastina D, Isella G, Rössner B, Bollani M, Müller E, Hackbarth T and von Känel H 2004 High quality SiGe electronic material grown by low energy plasma enhanced chemical vapour deposition Thin Solid Films 459 37–40 [25] Hoarfrost M L, Takei K, Ho V, Heitsch A, Trefonas P, Javey A and Segalman R A 2013 Spin-on organic polymer dopants for silicon J. Phys. Chem. Lett. 4 3741–6 8 Nanotechnology 32 (2021) 025303 C Barri et al [36] Miccoli I, Edler F, Pfnür H and Tegenkamp C 2015 The 100th anniversary of the four-point probe technique: the role of probe geometries in isotropic and anisotropic systems J. Phys.: Condens. Matter 27 223201 [26] Sadhu J S, Tian H, Spila T, Kim J, Azeredo B, Ferreira P and Sinha S 2014 Controllable doping and wrap-around contacts to electrolessly etched silicon nanowire arrays Nanotechnology 25 375701 [37] Thurber W R 1981 The Relationship between Resistivity and Dopant Density for Phosphorus-and Boron-doped Silicon vol 400 (US Department of Commerce, National Bureau of Standards) [27] Jamil M, Mantey J, Onyegam E U, Carpenter G D, Tutuc E and Banerjee S K 2011 High-Performance Ge nMOSFETs with n+ -p junctions formed by ‘Spin-On Dopant’ IEEE Electron Device Lett. 32 1203–5 [28] Bollani M et al 2019 Templated dewetting of single-crystal sub-millimeter-long nanowires and on-chip silicon circuits Nat. Commun. 10 5632 [38] Shaw D 2017 Diffusion in semiconductors Springer Handbook of Electronic and Photonic Materials ed S Kasap and P Capper (Cham: Springer) [29] Frigerio J et al 2016 Tunability of the dielectric function of heavily doped germanium thin ﬁlms for mid-infrared plasmonics Phys. Rev. B 94 085202 [39] Jaeger R C 2002 Introduction to Microelectronic Fabrication vol 2 (Upper Saddle River, NJ: Prentice-Hall) [40] Mar K M 1979 Phosphorus diffusion in silicon from a spin-on P-doped silicon oxide ﬁlm J. Electrochem. Soc. 126 1252 [30] Yang J M et al 2019 Phosphorus doping of Si nanosheets by spin-on dopant proximity Electron. Mater. Lett. 15 208215 [30] Yang J M et al 2019 Phosphorus doping of Si nanosheets by spin-on dopant proximity Electron. Mater. Lett. References 15 208215 [31] Li S S 2006 Semiconductor Physical Electronics (New York: Springer) [41] Teh S T and Chuah D G S 1989 Diffusion proﬁle of spin-on dopant in silicon substrate Sol. Energy Mater. 19 237–47 [31] Li S S 2006 Semiconductor Physical Electronics (New York: Springer) p gy [42] Li S S and Thurber W R 1977 The dopant density and temperature dependence of electron mobility and resistivity in n-type silicon Solid-State Electron. 20 609–16 [32] Price W L V 1973 Electric potential and current distribution in a rectangular sample of anisotropic material with application to the measurement of the principal resistivities by an extension of van der Pauw’s method Solid-State Electron. 16 753–62 yp [43] Ayele Y, Yohannes Z and Benor A 2016 Effect of phosphorus dopant concentration on the carrier mobility in crystalline silicon Eth. J. Sci. Technol. 9 113 [33] Perego M et al 2015 Thermodynamic stability of high phosphorus concentration in silicon nanostructures Nanoscale 7 14469–75 [44] Finetti M, Negrini P, Solmi S and Nobili D 1981 Electrical properties and stability of supersaturated phosphorus-doped silicon layers J. Electrochem. Soc. 128 1313 [34] Thorsteinsson S, Wang F, Petersen D H, Hansen T M, Kjær D, Lin R, Kim J-Y, Nielsen P F and Hansen O 2009 Accurate microfour-point probe sheet resistance measurements on small samples Rev. Sci. Instrum. 80 053902 [45] Masetti G, Severi M and Solmi S 1983 Modeling of carrier mobility against carrier concentration in arsenic-, phosphorus-, and boron-doped silicon IEEE Trans. Electron Devices 30 764–9 p [35] Schroder D K 2015 Semiconductor Material and Device Characterization (New York: Wiley) 9
https://openalex.org/W4234411195	https://zenodo.org/record/4885461/files/Perlukah%20Sumut%20Dimekarkan.pdf	Indonesian	null	Perlukah Sumut Dimekarkan?	null	2,021	cc-by	1,667	Oleh : Pardomuan Gultom Pada 2 Mei 2011 yang lalu Panitia Khusus (Pansus) Pemekaran Propinsi Sumatera Utara - yang dibentuk oleh DPRD Propinsi Sumatera Utara - membacakan laporan Pansus Pemekaran atas kajian, hasil studi banding (Propinsi Kepulauan Riau dan Propinsi Bangka Belitung), hasil kunjungan kerja ke daerah, dan aspirasi sejumlah kepala daerah, pimpinan DPRD Kabupaten/Kota, serta ormas terhadap usulan daerah yang akan dimekarkan. Agenda pembacaan laporan Pansus Pemekaran Rapat paripurna DPRD Propinsi Sumatera Utara sempat tertunda dari yang dijadwalkan pada tanggal 25 April 2011. Laporan tersebut mencakup usulan 3 (tiga) daerah pemekaran, yakni: Propinsi Sumatera Tenggara, Propinsi Kepulauan Nias, dan Propinsi Tapanuli. Keterlambatan rapat paripurna DPRD Propinsi Sumatera Utara dengan agenda pembacaan laporan Pansus Pemekaran disebabkan belum maksimalnya laporan pansus dan juga perbedaan pendapat mengenai materi laporan pansus (Harian Seputar Indonesia, 03/05). Dan wacana ini menjadi perhelatan politik baru di Sumatera Utara pasca 'tragedi Protap' (baca: Provinsi Tapanuli). Yang perlu dikaji disini adalah bagaimana perkembangan dalam pengusulan pembentukan Provinsi Sumatera Tenggara dan Provinsi Nias; dan untuk kasus tertentu, seperti usulan pembentukan Provinsi Tapanuli, apa perbedaan antara masa sebelum 'tragedi Protap' dan pasca tragedi tersebut serta seberapa besar peluangnya. Untuk mengetahui prosesnya, maka yang menjadi landasan berpikir untuk menganalisa situasi yang berkembang dalam upaya pemekaran Provinsi Sumatera Utara, yakni dasar konstitusi yang menjadi landasan sistematika bagi pembentukan daerah otonom baru dan respon kebijakan. Dasar konstitusi maksudnya adalah aturan hukum (rule of law) dan aturan main (rule of game) yang diatur dalam Undang-Undang No. 32 Tahun 2004 tentang Pemerintahan Daerah dan Peraturan Pemerintah No. 78 Tahun 2007 tentang Tata Cara Pembentukan, Penghapusan, dan Penggabungan Daerah. dan respon kebijakan dimaksudkan pada proses politik yang dilalui dalam upaya pembentukan daerah otonom baru. Sesuai ketentuan yang diatur dalam UU No. 32 Tahun 2004 dan PP No. 78 Tahun 2007, maka daerah yang mengusulkan pembentukan daerah otonom baru (provinsi atau kabupaten/kota) diwajibkan memenuhi 3 (tiga) syarat, yaitu: syarat administratif, syarat teknis, dan syarat fisik kewilayahan. Perlukah Sumut Dimekarkan ? Oleh : Pardomuan Gultom Persyaratan Isu pemekaran atau pembentukan daerah otonom baru bukanlah agenda politik yang baru di Sumatera Utara. Dalam kurun waktu satu dekade terakhir (sejak 1999 hingga 2009), Sumatera Utara telah memekarkan 14 kabupaten/kota (lihat Daftar Daerah Otonom Baru (DOB) Pemekaran (1999 - 2009) oleh KPPOD). Pada tahun 2001, Kota Padang Sidempuan - melalui UU No. 4 Tahun 2001 - telah dimekarkan dari Kabupaten Tapanuli Selatan. Tahun 2003, ada 5 (lima) daerah otonomi baru, yakni: Kabupaten Pakpak Bharat , Kabupaten Nias Selatan, Kabupaten Humbang Hasundutan (UU No. 9 Tahun 2003); Kabupaten Serdang Bedagai dan Kabupaten Samosir (UU No. 36 Tahun 2003). Tahun 2007, ada 3 (tiga) daerah otonomi baru, yakni: Kabupaten Batubara (UU No. 5 Tahun 2007), Kabupaten Padang Lawas Utara (UU No. 37 Tahun 2007), dan Kabupaten Padang Lawas (UU No. 38 Tahun 2007). Tahun 2008, ada 5 (lima) daerah otonom baru, yakni: Kabupaten Labuhan Batu Utara (UU No. 23 Tahun 2008), Kabupaten Labuhan Batu Selatan (UU No. 22 Tahun 2008), Kabupaten Nias Barat (UU No. 46 Tahun 2008), Kabupaten Nias Utara (UU No. 45 Tahun 2008), dan Kota Gunung Sitoli (UU No. 47 Tahun 2008). Kehadiran 14 kabupaten/kota yang telah dimekarkan ternyata berpeluang dalam membentuk provinsi baru. UU No. 32 Tahun 2004 telah memberi ruang bagi setiap kabupaten/kota untuk membentuk provinsi baru sesuai dengan ketentuan yang diatur dalam PP No. 78 Tahun 2007. Peluang tersebut jadi momentum untuk memekaran Provinsi Sumatera Utara dalam 3 (tiga) usulan provinsi, yakni: Propinsi Sumatera Tenggara (Kabupaten Padang Lawas, Kabupaten Padang Lawas Utara, Kota Padang Sidempuan, Kabupaten Mandailing Natal, dan Kabupaten Tapanuli Selatan), Propinsi Kepulauan Nias (Kabupaten Nias, Kabupaten Nias Utara, Kabupaten Nias Barat, Kabupaten Nias Selatan, dan Kota Gunung Sitoli), dan Propinsi Tapanuli (Kabupaten Tapanuli Utara, Kabupaten Toba Samosir, Kabupaten Humbang Hasundutan, dan Kabupaten Samosir). Jika merujuk pada UU No. 32 Tahun 2004 Pasal 5 angka (5) dan PP No. 78 Tahun 2007 Pasal 8 huruf a, maka usulan pembentukan provinsi yang dianggap layak sesuai syarat fisik kewilayahan - selain syarat administratif dan syarat teknis - adalah Propinsi Sumatera Tenggara dan Propinsi Kepulauan Nias. Terlepas dari segala bentuk polemik yang pernah terjadi dalam upaya mengusung pembentukan Propinsi Tapanuli, ternyata usulan pembentukan Propinsi Tapanuli tidak mampu memenuhi syarat fisik kewilayahan karena hanya diusung oleh 4 (empat) kabupaten. Hal ini disebabkan oleh berkurangnya dukungan kabupaten/kota lain yang sebelumnya memberikan dukungan karena sikap yang tidak elegan dalam proses penyampaian aspirasi. Pasang Surut Dalam perjalanannya, sejarah pembentukan Provinsi Tapanuli bisa menjadi sebagai bahan pembelajaran dalam upaya pembentukan provinsi baru lainnya. Artinya, dinamika politik pembentukan Provinsi Tapanuli mengalami perubahan dalam bentuk dukungan kabupaten/kota, metode artikulasi kepentingan, dan problem elitisme aktor-aktor pengusungnya. Dalam perjalanannya, sejarah pembentukan Provinsi Tapanuli bisa menjadi sebagai bahan pembelajaran dalam upaya pembentukan provinsi baru lainnya. Artinya, dinamika politik pembentukan Provinsi Tapanuli mengalami perubahan dalam bentuk dukungan kabupaten/kota, metode artikulasi kepentingan, dan problem elitisme aktor-aktor pengusungnya. Dari segi mobilisasi dukungan, ada 3 (tiga) fase yang telah dilalui dalam upaya pembentukan Provinsi Tapanuli, yakni: pertama, fase embrio (2002-2004), yang diawali dengan usulan 4 (empat) kabupaten/kota untuk membentuk Provinsi Tapanuli, yang terdiri dari Kabupaten Tapanuli Utara, Kabupaten Tapanuli Tengah, Kota Sibolga, dan Kabupaten Toba Samosir; kedua, fase penggalangan dukungan (2006), yaitu fase dimana penggalangan terhadap daerah lainnya sebagai daerah tambahan yang disebabkan oleh adanya perubahan jumlah syarat minimum kabupaten/kota dalam pembentukan provinsi (akibat perubahan atas UU No. 22 Tahun 1999 tentang Pemerintahan Daerah menjadi UU No. 32 Tahun 2004, dari minimum 3 (tiga) kabupaten/kota menjadi 5 (lima) kabupaten/kota). Ada 3 (tiga) daerah yang memberikan dukungan tambahan, antara lain: Kabupaten Humbang Hasundutan, Kabupaten Samosir, dan Kabupaten Nias Selatan. Hingga tahun 2008, proses konsolidasi antarkabupaten/kota yang mengusung pembentukan Provinsi Tapanuli tidak solid lagi. Ini bisa dilihat dari keputusan bupati/walikota dan keputusan DPRD kabupaten/kota perihal cakupan wilayah, penetapan nama, penetapan lokasi ibukota, alokasi dana, penyerahan aset, dan pemindahan personil pembentukan calon Provinsi Tapanuli - yang hanya diterbitkan oleh 6 (enam) kabupaten/kota. Hal ini disebabkan oleh Keputusan DPRD Kabupaten Tapanuli Tengah yang secara resmi mencabut dukungan terhadap pembentukan Provinsi Tapanuli pada tanggal 21 Nopember 2008. Penarikan dukungan tersebut dilatarbelakangi oleh persetujuan Kabupaten Tapanuli Tengah dalam pembentukan Provinsi Tapanuli Barat. Ketiga, fase finalisasi, yaitu fase dimana upaya pembentukan Provinsi Tapanuli telah mencapai tahapan akhir, yang ditandai dengan diterbitkannya surat presiden RI Nomor R.04/Pres/02/2008 tentang Pembentukan Kabupaten/Kota dan Pembentukan Provinsi Tapanuli - yang merupakan amanat presiden kepada Dewan Perwakilan Rakyat (DPR) untuk membahas Rancangan Undang-Undang pembentukan 14 kabupaten/kota dan Provinsi Tapanuli - dan Keputusan Gubernur Sumatera Utara tentang Persetujuan Pembentukan Provinsi Tapanuli dan Cakupan Wilayah Kabupaten /Kota Calon Provinsi Tapanuli. Persyaratan Selain persyaratan fisik kewilayahan yang dibutuhkan dalam pembentukan provinsi baru, dua syarat lainnya - administratif dan teknis - penting untuk disimak. Syarat administrasi meliputi 5 (lima) tahapan, yakni: (1) Keputusan masing-masing DPRD kabupaten/kota yang akan menjadi cakupan wilayah calon provinsi berdasarkan hasil rapat paripurna; (2) Keputusan bupati/walikota yang ditetapkan dalam keputusan bersama bupati/walikota wilayah calon provinsi tentang persetujuan pembentukan calon provinsi; (3) Keputusan DPRD provinsi induk tentang persetujuan pembentukan calon provinsi berdasarkan hasil Rapat Paripurna - yang berdasarkan aspirasi sebagian besar masyarakat setempat yang dituangkan dalam keputusan DPRD kabupaten/kota yang akan menjadi cakupan wilayah calon provinsi; (4) Keputusan gubernur tentang persetujuan pembentukan calon provinsi; dan (5) Rekomendasi menteri (PP No. 78 Tahun 2007, Pasal 5 Ayat (1) huruf (a), (b), (c), (d), (e) dan Ayat (4) ). Syarat ketiga adalah syarat teknis yang meliputi faktor kemampuan ekonomi, potensi daerah, sosial budaya, sosial politik, kependudukan, luas daerah, pertahanan, keamanan, kemampuan keuangan, tingkat kesejahteraan masyarakat, dan rentang kendali penyelenggaraan pemerintahan daerah - yang dinilai berdasarkan hasil kajian daerah terhadap indikator (PP No. 78 Tahun 2007 Pasal 6 Ayat (1). Untuk mengetahui lebih jauh proses penyelenggaraan kebijakan di bidang pembentukan daerah otonom baru, khususnya pemekaran di Provinsi Sumatera Utara, dapat kita perhatikan dari proses yang telah berjalan, baik dari sisi lokalitas, seperti: aspirasi publik, kemauan politik pemerintah daerah dan DPRD kabupaten/kota, aspek persyaratan, maupun respon kebijakan pemerintah pusat. Pasang Surut Keempat, fase kehancuran (patahan sejarah), yaitu fase dimana proses yang selama ini telah ditempuh dalam pembentukan Provinsi Tapanuli, akhirnya hancur dan justru menimbulkan trauma politik bagi masyarakat akibat sikap anarkisme gerakan kelompok pendukung Provinsi Tapanuli yang dikaitkan dengan kematian mendadak Abdul Azis Angkat (Ketua DPRD Sumatera Utara) saat Rapat Paripurna DPRD Sumatera Utara (tanggal 3 Februari 2009). Selain itu, tokoh-tokoh yang menjadi penggerak pembentukan Provinsi Tapanuli yang tersangkut masalah hukum juga turut menambah catatan hitam perjalanan politik pemekaran daerah di Sumatera Utara. Fase-fase yang telah dipaparkan diatas merupakan kompleksitas masalah yang terjadi dalam sejarah politik pemekaran di Sumatera Utara, khususnya kebijakan di seputar pembentukan Provinsi Tapanuli, sepanjang proses pencapaiannya tidak elegan, elitis, dan kental nuansa politik kepentingan antardaerah. Instrumen Pembangunan Di tengah kebijakan moratorium pemekaran yang didengungkan oleh Pemerintah Pusat, namun pemekaran daerah - sebagai amanat UU No. 32 Tahun 2004 - tidak serta merta terhambat oleh kebijakan yang tidak populis itu. Kebijakan tersebut justru kontraproduktif dengan upaya percepatan pembangunan di seluruh wilayah Republik Indonesia. Sinergisitas pelaksanaan program pembangunan yang dimandatkan oleh pusat terhadap daerah tidak akan berjalan jika aspirasi dan prakarsa masyarakat tidak mendapat tempatnya dalam upaya pemekaran daerah. Sebagai contoh, beberapa daerah (Kabupaten Tapanuli Tengah, Kabupaten Tapanuli Utara, Kabupaten Tapanuli Selatan, dan Kabupaten Madina) yang berada di posisi Pantai Barat Sumatera Utara mengalami kesulitan dalam porsi anggaran dari tingkat provinsi karena masih ada anggapan bahwa wilayah Pantai Barat Sumatera Utara merupakan halaman belakang Provinsi Sumatera Utara. Jadi dengan demikian, prasyarat yang dibutuhkan untuk mendorong kemajuan suatu daerah tidak berjalan, misalnya kondisi jalan provinsi dan jalan kabupaten/kota yang rusak parah di Kawasan Pantai Barat Sumatera Utara. Untuk itu, jika pemekaran dijadikan sebagai salah satu instrumen pembangunan daerah dalam rangka percepatan pembangunan, maka setiap daerah pemekaran baru dapat berkreasi sesuai dengan potensi alam, sumber daya manusia, dan kemampuan anggaran yang dimiliki. Hal ini dapat menjadi tenaga pendorong pemajuan suatu daerah untuk tujuan mendukung kepentingan pembangunan nasional. Pemekaran bukanlah jalan satu-satunya dalam tujuan mencapai kemakmuran daerah, tapi tanpa hak otonom bagi daerah dalam pengelolaan rumah tangganya, maka jebakan kemiskinan tetap menjadi warisan bagi masa depan anak bangsa. *** *Tulisan ini pernah dimuat di Harian ANALISA (Edisi Cetak), Senin, 20 Juni 2011, dan dimuat juga di URL: (http://www.analisadaily.com/index.php?option=com_content&view=article&id=99021:perlukah- sumut-dimekarkan&catid=78:umum&Itemid=131)
https://openalex.org/W2784131604	https://pubs.rsc.org/en/content/articlepdf/2018/dt/c7dt03848h	English	null	Tin(<scp>iv</scp>) chalcogenoether complexes as single source precursors for the chemical vapour deposition of SnE<sub>2</sub> and SnE (E = S, Se) thin films	Dalton transactions	2,018	cc-by	8,092	†Electronic supplementary information (ESI) available: 1H, 13C{1H}, 77Se{1H} and 119Sn NMR and IR spectroscopic data for each of the new complexes, together with their TGA profiles, EDX analyses for the SnE2 and SnE films and SEM, XRD and Raman spectra for the films obtained by LP CVD using (3) at elev- ated temperature. See DOI: 10.1039/c7dt03848h aSchool of Natural Sciences, Mahindra Ecole Centrale, Hyderabad, India bChemistry, University of Southampton, Southampton SO17 1BJ, UK. E-mail: G.Reid@soton.ac.uk cElectronic and Computer Science, University of Southampton, Southampton SO17 1BJ, UK dISIS Neutron and Muon Source, Rutherford Appleton Laboratory, Harwell Science and Innovation Campus, Didcot, OX11 0QX, UK Open Access Article. Published on 02 February 2018. Downloaded on 10/24/2024 This article is licensed under a Creative Commons Attribution 3.0 U The molecular Sn(IV) complexes, [SnCl4{nBuS(CH2)3SnBu}] (2), [SnCl4(nBu2S)2] (3) and [SnCl4(nBu2Se)2] (4) have been prepared in good yield from reaction of SnCl4 with the appropriate chalcogenoether ligand in anhydrous hexane and, together with the known [SnCl4{nBuSe(CH2)3SenBu}] (1), employed as single source precursors for the low pressure chemical vapour deposition of the corresponding tin dichalcogen- ide thin ﬁlms. At elevated temperatures the bidentate ligand precursors, (1) and (2), also form the tin monochalcogenides, SnSe and SnS, respectively. In contrast, (3) gave a mixture of phases, SnS2, Sn2S3 and SnS and (4) gave SnSe2 only. The morphologies, elemental compositions and crystal structures of the resulting ﬁlms have been determined by scanning electron microscopy, energy dispersive X-ray spec- troscopy, grazing incidence X-ray diﬀraction and Raman spectroscopy. Van der Pauw measurements on the SnS2, SnS and SnSe2 ﬁlms conﬁrm their resistivities to be 2.9(9), 266(3) and 4.4(3) Ω cm, respectively. Received 12th October 2017, Accepted 10th November 2017 DOI: 10.1039/c7dt03848h rsc.li/dalton Dalton Transactions Open Access Article. Published on 02 February 2018. Downloaded on 10/24/2024 6:47:22 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. This journal is © The Royal Society of Chemistry 2018 Tin(IV) chalcogenoether complexes as single source precursors for the chemical vapour deposition of SnE2 and SnE (E = S, Se) thin ﬁlms† Cite this: Dalton Trans., 2018, 47, 2628 Chitra Gurnani,a Samantha L. Hawken,b Andrew L. Hector, b Ruomeng Huang, c Marek Jura,d William Levason, b James Perkins,b Gillian Reid b and Gavin B. G. Stenningd Precursor preparation and characterisation 77Se{1H} NMR (CD2Cl2, 298 K) δ/ppm: 284 (br), 297 (br); (183 K) δ/ppm: 299 (1JSnSe = 530 Hz), 323 (1JSnSe = 478 Hz). IR (Nujol) ν/cm−1: 324 (s), 302 (sh), 292 (s) Sn–Cl. y nBuS(CH2)3SnBu: was obtained by dissolving HS(CH2)3SH (6.05 g, 56 mmol) in anhydrous EtOH (60 mL) with addition of elemental sodium (2.8 g, 112 mmol) cut into small pieces. After dissolution of the sodium, the reaction mixture was refluxed for 1 h, cooled to ambient temperature and nBuBr (15.34 g, 12.03 mL, 112 mmol) was added slowly. The reaction mixture was then refluxed for 1 h, forming a white solid. After cooling, saturated aqueous NaCl and NaHCO3 solutions were added. After filtering, the product was extracted from the aqueous filtrate with Et2O (4 × 50 mL), dried overnight (Na2SO4) and, after removal of the drying agent, the solvent was removed in vacuo, leaving a colourless oil. Yield: 6.0 g, 48%. 1H NMR (CDCl3, 298 K) δ/ppm: 0.92 (t, [6H], CH3), 1.41 (m, [4H], –CH2CH3), 1.57 (m, [4H], –SCH2CH2CH2CH3), 1.86 (m, [2H], –SCH2CH2CH2S–), 2.52 (t, [4H], SCH2–), 2.62 (t, [4H], Open Access Article. Published on 02 February 2018. Downloaded on 10/24/202 This article is licensed under a Creative Commons Attribution 3.0 U [SnCl4(nBu2S)2] (3). A solution of nBu2S (0.944 g, 6.45 mmol) in anhydrous hexane (5 mL) was added slowly to a solution of SnCl4 (0.840 g, 3.22 mmol) in anhydrous hexane (10 mL) under constant stirring at room temperature to aﬀord a colour- less solution. It was stirred for 20 minutes. It was concentrated in vacuo, yielding a white precipitate. The solution was de- canted and the precipitate washed with anhydrous hexane and dried in vacuo. Yield: 1.57 g, 86%. Anal. calcd for C16H36Cl4S2Sn: C 34.8, H 6.6. Found: C 34.7, H 6.8%. 1H NMR (CDCl3) δ/ppm: 0.97 (t, [3H], CH3), 1.49 (m, [2H], SCH2CH2CH2CH3), 1.76 (br, [2H], SCH2CH2CH2CH3), 3.00 (vbr, [2H], SCH2–). 13C{1H} NMR (CDCl3) δ/ppm: 13.56 (CH3), 21.89 (CH2), 30.47 (CH2), 43.22 (SCH2). 119Sn NMR (CH2Cl2/CD2Cl2, 298 K) δ/ppm: −594 (br), −603 (br); (223 K) δ/ppm: −581 (s), −589 (s). IR (Nujol) ν/cm−1: 330 (sh), 318 (s), 300 (sh) Sn–Cl. [SnCl4(nBu2S)2] (3). A solution of nBu2S (0.944 g, 6.45 mmol) in anhydrous hexane (5 mL) was added slowly to a solution of SnCl4 (0.840 g, 3.22 mmol) in anhydrous hexane (10 mL) under constant stirring at room temperature to aﬀord a colour- less solution. It was stirred for 20 minutes. It was concentrated in vacuo, yielding a white precipitate. The solution was de- canted and the precipitate washed with anhydrous hexane and dried in vacuo. Yield: 1.57 g, 86%. Anal. calcd for C16H36Cl4S2Sn: C 34.8, H 6.6. Found: C 34.7, H 6.8%. 1H NMR (CDCl3) δ/ppm: 0.97 (t, [3H], CH3), 1.49 (m, [2H], SCH2CH2CH2CH3), 1.76 (br, [2H], SCH2CH2CH2CH3), 3.00 (vbr, [2H], SCH2–). 13C{1H} NMR (CDCl3) δ/ppm: 13.56 (CH3), 21.89 (CH2), 30.47 (CH2), 43.22 (SCH2). 119Sn NMR (CH2Cl2/CD2Cl2, 298 K) δ/ppm: −594 (br), −603 (br); (223 K) δ/ppm: −581 (s), −589 (s). IR (Nujol) ν/cm−1: 330 (sh), 318 (s), 300 (sh) Sn–Cl. Precursor preparation and characterisation All reactions were conducted using Schlenk, vacuum line and glove box techniques under a dry nitrogen atmosphere. The reagents were stored and manipulated using a glove box. Hexane was dried by distillation over sodium wire. SnCl4, nBuLi (1.6 mol dm−3 in diethyl ether) and nBu2S were obtained from Sigma-Aldrich. nBu2S was dried over sieves, SnCl4 and nBuLi were used as received. nBu2Se,21 nBuSe(CH2)3SenBu20 and [SnCl4{nBuSe(CH2)3SenBu}]20 (1) were prepared according to the literature methods. IR spectra were recorded as Nujol mulls between CsI plates using a PerkinElmer Spectrum 100 instrument. 1H and 13C{1H} NMR spectra were recorded from solutions in CDCl3 or CD2Cl2 on a Bruker AV400 spectrometer, 77Se{1H} and 119Sn NMR spectra on a Bruker AV400-II spectro- meter and referenced to external neat SeMe2 and SnMe4 respectively. A small amount of [Cr(acac)3] was added as a relaxation agent to the samples for 119Sn NMR measurements. Microanalytical results were obtained from Medac Ltd. [SnCl4(nBu2Se)2] (4). A solution of nBu2Se (1.25 g, 6.45 mmol) in anhydrous hexane (5 mL) was added slowly to a solution of SnCl4 (0.840 g, 3.22 mmol) in anhydrous hexane (10 mL) under constant stirring at room temperature to aﬀord a pale yellow solution. It was stirred for 20 minutes. It was con- centrated in vacuo, yielding a pale orange precipitate. The solu- tion was decanted and the precipitate washed with anhydrous hexane and dried in vacuo. Yield: 0.924 g, 52%. Anal. calcd for C16H36Cl4Se2Sn: C 29.71, H 5.61. Found: C 29.41, H 5.61%. [SnCl4(nBu2Se)2] (4). A solution of nBu2Se (1.25 g, 6.45 mmol) in anhydrous hexane (5 mL) was added slowly to a solution of SnCl4 (0.840 g, 3.22 mmol) in anhydrous hexane (10 mL) under constant stirring at room temperature to aﬀord a pale yellow solution. It was stirred for 20 minutes. It was con- centrated in vacuo, yielding a pale orange precipitate. The solu- tion was decanted and the precipitate washed with anhydrous hexane and dried in vacuo. Yield: 0.924 g, 52%. Anal. calcd for C16H36Cl4Se2Sn: C 29.71, H 5.61. Found: C 29.41, H 5.61%. 1H NMR (CDCl3) δ/ppm: 0.98 (t, [3H], CH3), 1.50 (m, [2H], SeCH2CH2CH2CH3), 1.86 (br, [2H], SeCH2CH2–), 3.10 (br, [2H], SeCH2–). 13C{1H} NMR (CDCl3) δ/ppm: 13.48 (CH3), 22.83 (CH2), 30.42 (CH2), 35.64 (SeCH2). 119Sn NMR (CD2Cl2, 298 K) δ/ppm: −717 (br, s); (183 K) δ/ppm: −719 (1JSnSe = 530 Hz, minor), −718 (1JSnSe = 478 Hz, major). Introduction interest as a potentially eﬀective phase change material for data storage applications. For this application fast recrystallis- ation times are key and SnSe2, with a recrystallisation time of 20 ns,9 is a promising material in this field. Tin chalcogenides (SnEx) (E = S, Se; x = 1, 2) are layered semi- conducting materials that are becoming widely acknowledged as important for a wide variety of electronic applications; specifically optoelectronics,1–5 thermoelectrics6,7 and phase change devices.8,9 SnS2 also shows potential for use in photo- detectors, with a high external quantum eﬃciency of 3.2 × 105%;1 high performance sodium storage, with a high revers- ible capacity reported of 576 mA h g−1;10 and phototransistors with a high on/oﬀratio reported of >106.2 SnS, on the other hand, has a very low thermal conductivity ∼0.29 W mK−1 and a reported ZT value of 0.41,7 which are of particular interest for thermoelectric devices, whilst ZT values as high as 2.6 have been claimed for single crystals of SnSe.7 SnSe2 has attracted Tin chalcogenide thin films can be produced by a variety of methods. Using atmospheric pressure chemical vapour depo- sition and a dual source precursor approach, SnCl4 with either H2S or Et2Se can produce SnS2 or SnS,11,12 and SnSe2 or SnSe,13 respectively, depending on the conditions employed. A single source precursor approach, in which the metal and chalcogen are incorporated into one molecular compound, can lead to easier to handle precursors (since they are coordi- natively saturated), potentially easier control of stoichiometry and, in some cases, the ability to deposit the binary material selectively onto specific regions of a patterned substrate on the micro-14 or nano-scale.15 While this substrate selectivity is rather atypical of CVD, it can be highly advantageous for certain applications. In previous work, SnS and SnS2 have been produced from the tin(IV) thiolate complex, [Sn(SCH2CH2S)2], using aerosol assisted chemical vapour deposition in the presence of H2S gas,16 or using alkyl tin dithiocarbamates17 and SnS nanosheets by liquid exfoliation.18 Single source precursor approaches to tin sulfide and selenide thin films by us have used the thio- and seleno-ether complexes, [SnCl4{o- C6H4(CH2EMe)2}] (E = S, Se)19 to produce SnS2 and small amounts of SnS and SnSe, or; [SnCl4(SeEt2)2]19 and [SnCl4{nBuSe This journal is © The Royal Society of Chemistry 2018 2628 \| Dalton Trans., 2018, 47, 2628–2637 View Article Online Dalton Transactions Paper SCH2–). Introduction 13C{1H} NMR (CDCl3, 298 K) δ/ppm 13.65 (CH3), 21.98, 29.43, 30.96, 31.74, 31.81 (all CH2). SCH2–). 13C{1H} NMR (CDCl3, 298 K) δ/ppm 13.65 (CH3), 21.98, 29.43, 30.96, 31.74, 31.81 (all CH2). SCH2–). 13C{1H} NMR (CDCl3, 298 K) δ/ppm 13.65 (CH3), 21.98, 29.43, 30.96, 31.74, 31.81 (all CH2). (CH2)xSenBu}] (x = 2, 3)20 to form SnSe2. Within the chalcogen- oether- and chalcogenolate-based precursors, terminal butyl substituents have been shown to be well-suited for the growth of high quality binary chalcogenide thin films. This is attribu- ted to their clean, low energy β-hydride elimination pathway. [SnCl4{nBuS(CH2)3SnBu}] (2). A solution of nBuS(CH2)3SnBu (0.24 g, 1.09 mmol) in anhydrous hexane (5 mL) was added slowly to a solution of SnCl4 (0.15 mL, 1.25 mmol) in anhy- drous hexane (10 mL) under constant stirring at room temp- erature and under a dinitrogen atmosphere. A white solid immediately precipitated out of solution. It was collected by filtration after 20 minutes and dried in vacuo to aﬀord a white solid. Yield: 0.46 g, 88%. Anal. calcd for C11H24Cl4S2Sn: C 27.5, H 5.0. Found: C 26.9, H 5.3%. 1H NMR (CDCl3, 298 K) δ/ppm: 0.98 (t, [6H], CH3), 1.49 (m, [4H], CH2CH3), 1.76 (m, [4H], CH2), 2.39 (m, [2H], –SCH2CH2CH2S–), 3.17 (m, [4H], SCH2), 3.32 (m, [4H], SCH2). 13C{1H} NMR (CDCl3) δ/ppm: 13.50 (CH3), 21.87 (CH2), 24.02 (CH2), 29.40 (CH2), 34.19 (CH2), 36.74 (CH2). 119Sn NMR (CH2Cl2/CD2Cl2, 298 K): −575 (br); (200 K) δ/ppm: −567, −585. IR (Nujol) ν/cm−1: 324 (s br) Sn–Cl. Here we report the synthesis, characterisation and evalu- ation of three novel single source precursors, [SnCl4{nBuS (CH2)3SnBu}] (2) and [SnCl4(nBu2E)2] (E = S, (3); E = Se, (4)), and their use for the low pressure chemical vapour deposition (LPCVD) of SnS2 and SnSe2 thin films, as well as variation of the deposition conditions to produce SnSe and SnS films. High temperature LPCVD using [SnCl4{nBuSe(CH2)3SenBu}], (1), is also explored as a route to SnSe films. Energy dispersive X-ray (EDX), scanning electron microscopy (SEM), Raman spectroscopy and grazing incidence X-ray diﬀraction (XRD) data on the resulting films are pre- sented, together with resistivities determined by van der Pauw measurements. LPCVD onto PVD SiO2 substrates Physical vapour deposited (PVD) SiO2 on Si substrates were prepared as described previously.20 In a typical LPCVD experi- ment the reagent (70–80 mg) and PVD SiO2 substrates were loaded into a closed end silica tube in a glove box (precursor Dalton Trans., 2018, 47, 2628–2637 \| 2629 This journal is © The Royal Society of Chemistry 2018 Dalton Transactions View Article Online View Article Online Paper Dalton Transactions [SnCl4(Et2Se)2]19 and [SnCl4{nBuSe(CH2)3SenBu}],20 have been shown to produce polycrystalline SnSe2 thin films, whilst LPCVD of [SnCl4{o-C6H4(CH2SeMe)2}] at 600 °C resulted in SnSe,19 and, despite the S–C bonds in thioethers being stron- ger than Se–C bonds,25 [SnCl4{o-C6H4(CH2SMe)2}] has been shown to be suitable as a single source precursor for the depo- sition of tin sulfide thin films, forming SnS2 at 650 °C, with some SnS forming further into the hot zone.19 However, depo- sition of the monochalcogenides, SnE (E = S, Se), led to rather poor, non-continuous substrate coverage in all cases, and with SnS2 typically deposited on the same substrate.19 This may occur because the temperature required for the sublimation of the [SnCl4{o-C6H4(CH2SMe)2}] precursor is very close to that of the boundary between deposition of SnS2 and SnS. By modify- ing the precursor to be more volatile, as well as incorporating nBu substituents in place of Me groups (the former have been shown to lead to cleaner growth of metal chalcogenides in other work20), we proposed that the temperature diﬀerence between sublimation of precursor and deposition of the metal chalcogenide would be increased and better film quality achieved. Additionally, precursors containing the monodentate nBu2E (E = S, Se) ligands would be more convenient to syn- thesise and therefore easier to access than bidentate ligands. Finally, it was anticipated that these precursors may facilitate growth of both SnE2 and SnE depending upon the deposition conditions used (with any excess nBu2E being distilled oﬀ readily under the low pressure CVD conditions employed or acting as a source of chalcogen to maintain stoichiometry). in bulb at closed end and followed by four substrate tiles (each 20 × 8 × 1 mm) positioned adjacently along the tube). The tube was placed horizontally in the furnace such that the precursor was outside of the heated zone. The tube was evacuated to 0.02 mmHg. The furnace was then heated to the desired temp- erature and the temperature was allowed to stabilise. Thin film characterisation X-ray diﬀraction (XRD) patterns were collected in grazing inci- dence on a Rigaku SmartLab system, using CuKα X-rays, a 2θ scan range of 10–80° and an ω-oﬀset of 0.7 or 1°. The crystal- line phase of the films was determined by matching to a litera- ture XRD pattern22 and lattice parameters calculated by further optimisation of the fit using PDXL.23 Scanning electron microscopy (SEM) images and energy dispersive X-ray (EDX) measurements were obtained using a JEOL JSM 6500 F Field Emission Scanning Electron Microscope with an Oxford INCA x-sight 7418 EDX probe. An accelerating voltage of 15 kV was used and samples were calibrated against standards. The novel complexes, [SnCl4(nBu2E)2] (E = S, Se) and [SnCl4{nBuS(CH2)3SnBu}] were synthesized in good yield by the direct reaction of SnCl4 and the appropriate ligand in anhy- drous hexane at room temperature (Scheme 1). The complexes were readily isolated as powdered solids by filtration. They are somewhat hydrolytically sensitive and therefore were stored and handled in a dry N2-purged glove-box. They have been characterized by 1H, 13C{1H} and variable temperature 77Se{1H} and 119Sn NMR spectroscopy, IR spectroscopy, and by micro- analysis, as appropriate. The IR and microanalytical data are fully consistent with the formulations, while solution NMR data for [SnCl4(nBu2E)2] reveal similar chemical shift ranges and dynamic behaviour in solution, fully consistent with that of isomeric cis and trans distorted octahedral complexes, which could not be separated. The NMR data on [SnCl4(nBu2E)2] are very similar to those reported for [SnCl4(Me2E)2],24,26 in particular, the broad lines in the room temperature 119Sn and 77Se NMR spectra which sharpen on cooling the solutions, demonstrate rapid exchange between the two isomers as discussed in detail by Knight and Merbach.26 As with the previously reported tin(IV) halide com- plexes with dithioether ligands,24 the low temperature 119Sn NMR spectra from [SnCl4{nBuS(CH2)3SnBu}] show two singlets representing the meso and DL diastereoisomers. Raman scattering spectra of the deposited films were measured at room temperature on a Renishaw InVia Micro Raman Spectrometer using a helium–neon laser with a wave- length of 632.8 nm. The incident laser power was adjusted to 0.1 mW for all samples. Van der Pauw measurements were performed at room temperature on a Nanometrics HL5500PC at 300 K. For each measurement, four copper probes with diameter of ca. 1 mm were carefully placed on the sample corners. Thin film characterisation Extra care was taken to ensure linear contact was obtained between each probe and the sample before each measurement. LPCVD onto PVD SiO2 substrates The tube was then repositioned, if necessary, to move the precursor closer to the heated zone to a point where sublimation could be observed. At this point the position of the tube was main- tained until the deposition was complete (i.e. the precursor was completely sublimed). This typically took 1–3 hours. Once complete, the tube was allowed to cool to room temperature and transferred to the glove box where the substrates were removed and stored under N2 before characterization. Open Access Article. Published on 02 February 2018. Downloaded on 10/24/2024 6:47:22 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. The LPCVD experiments produced black or silver films. Precursors 1 and 4 typically gave SnSe2 at all temperatures up to 1023 K. Precursor 2 gave SnS2 at all temperatures and 3 gave SnS2 at low temperatures and Sn2S3 at higher temperatures. These results were confirmed to be reproducible. It was noted that on the occasions when precursors 1 and 2 travelled further into the heated zone, SnSe and SnS films, respectively, were obtained. Results and discussion Distorted octahedral tin(IV) halide complexes with a range of neutral thio-, seleno- and telluro-ether ligands have been reported by us previously, with the chloro complexes proving rather less hydrolytically sensitive than either those of the heavier halide analogues or indeed SnCl4 itself.24 Furthermore, LPCVD studies using the selenoether complexes, This journal is © The Royal Society of Chemistry 2018 LP CVD of tin sulfide thin films LP CVD experiments using precursors (2) and (3) at 286 and 372 °C (0.02 mmHg), respectively, resulted in the deposition of SnS2 in both cases, without the need for an additional chalco- gen source. Film deposition tended to occur mostly at the edge of the heated zone nearest to the precursor, typically leading to ca. 50% coverage of the substrate. Film properties are sum- marised in Table 1. The 001 reflection is very intense in both samples, which is typical for thin films of these crystal types and indicates sig- nificant preferred orientation. Pole figure measurements were also acquired for the key reflections in the film grown from (2) in order to further investigate the preferred orientation further (Fig. 3). The pole figure for the 001 reflection reveals a sharp spot in the projection and a narrow peak in the integral at α = 90°. This is consistent with strong orientation of the c axis per- pendicular to the substrate and the crystallites lying parallel. The SEM images show the edges of crystallites which suggests SEM images of the thin films (Fig. 1 and 2) revealed hexag- onal platelets formed from both precursors, with higher uni- formity and larger crystallites observed in the deposit formed from (2). In the deposition from (3), larger truncated crystal- lites are also observed. EDX analysis revealed composition of 36% Sn and 64% S and 34% Sn, 66% S for the films deposited from (2) (at T = 286 °C) and (3) (at T = 372 °C), respectively, consistent within experimental error of SnS2, with no residual Cl observed. Grazing incidence XRD patterns also show that Table 1 Lattice parameters and resistivities measured for the tin chalcogenide thin ﬁlms Precursor Deposition temperature/°C Phase Lattice parameters/Å Resistivity/Ω cm Reference (1) 480–500 2H-SnSe2 a = 3.81–3.83 210 × 10−3 20 c = 6.14–6.16 (1) 588 SnSe a = 11.459(17) Not measureda This work b = 4.166(8) c = 4.433(6) (2) 286 2H-SnS2 a = 3.664(3) 2.9(9) This work c = 5.935(8) (2) 558 SnS a = 11.19(16) 266(3) This work b = 3.98(3) c = 4.24(3) (3) 372 2H-SnS2 a = 3.642(3) — This work c = 5.930(9) and 4H-SnS2 a = 3.643(7) — c = 11.84(4) (3) 470 Mixture of SnS2/Sn2S3/SnS — — This work (4) 325 2H-SnSe2 a = 3.8051(17) 4.4(3) This work c = 6.187(4) a Not measured since discontinuous film. ccess Article. Published on 02 February 2018. Downloaded on 10/24/2024 6:47:22 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. Open Access Article. Published on 02 February 2018. Downloaded on 10/24/2024 6:47:22 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. limation of [SnCl4{nBuS(CH and [SnCl4(nBu2Se)2] (4) at 19 ively (see ESI†). LP CVD of tin sulfide thin film LP CVD experiments using p 372 °C (0.02 mmHg), respecti SnS2 in both cases, without t gen source. Film deposition t of the heated zone nearest to to ca. 50% coverage of the su marised in Table 1. SEM images of the thin fil onal platelets formed from b formity and larger crystallites from (2). In the deposition f lites are also observed. EDX 36% Sn and 64% S and 34% from (2) (at T = 286 °C) and consistent within experiment Cl observed. Grazing inciden Scheme 1 Preparations of the tin Table 1 Lattice parameters and r Precursor Deposition temp (1) 480–500 (1) 588 (2) 286 (2) 558 (3) 372 (3) 470 (4) 325 a Not measured since discontinuo This journal is © The Royal Society of Open Access Article. Published on 02 February 2018. Downloaded on 10/24/2024 6:47:22 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. Scheme 1 Preparations of the tin(IV) single source precursors used in this work. Scheme 1 Preparations of the tin(IV) single source precursors used in this work. limation of [SnCl4{nBuS(CH2)3SnBu}] (2), [SnCl4(nBu2S)2] (3) and [SnCl4(nBu2Se)2] (4) at 195 °C, 138 °C and 204 °C, respect- ively (see ESI†). crystalline SnS2 is formed in both cases, and are consistent with the space group P3ˉm1 (2H-SnS2). In the deposition from (3), a second phase of SnS2, in space group P63mc, is also present (4H-SnS2) and SEM analysis shows larger truncated crystallites. The refined lattice parameters for the films de- posited from (2) (Table 1) also match well with literature values of Berndtite-2H SnS2 (a = 3.62–3.70, c = 5.66–5.90 Å).12,27 Lattice parameters for the two films de- posited from (3) match well with those of Berndtite-2H (a = 3.64, c = 5.88 Å)28 and Berndtite-4H (a = 3.64, c = 11.80 Å) SnS2,29 respectively. Low pressure CVD experiments Thermogravimetric analysis (TGA) of all of the precursors under an argon atmosphere showed clean and complete sub- 2630 \| Dalton Trans., 2018, 47, 2628–2637 This journal is © The Royal Society of Chemistry 2018 Dalton Transactions Scheme 1 Preparations of the tin(IV) single source precursors used in this work. p Scheme 1 Preparations of the tin(IV) single source precursors used in this work. limation of [SnCl4{nBuS(CH2)3SnBu}] (2), [Sn and [SnCl4(nBu2Se)2] (4) at 195 °C, 138 °C and ively (see ESI†). LP CVD of tin sulfide thin films LP CVD experiments using precursors (2) and 372 °C (0.02 mmHg), respectively, resulted in t SnS2 in both cases, without the need for an ad gen source. Film deposition tended to occur m of the heated zone nearest to the precursor, t to ca. 50% coverage of the substrate. Film pro marised in Table 1. SEM images of the thin films (Fig. 1 and 2) onal platelets formed from both precursors, w formity and larger crystallites observed in the from (2). In the deposition from (3), larger tr lites are also observed. EDX analysis revealed 36% Sn and 64% S and 34% Sn, 66% S for the from (2) (at T = 286 °C) and (3) (at T = 372 ° consistent within experimental error of SnS2, w Cl observed. Grazing incidence XRD patterns Scheme 1 Preparations of the tin(IV) single source pre Table 1 Lattice parameters and resistivities measured Precursor Deposition temperature/°C Ph (1) 480–500 2H (1) 588 Sn (2) 286 2H (2) 558 Sn (3) 372 2H 4H (3) 470 M (4) 325 2H a Not measured since discontinuous film. This journal is © The Royal Society of Chemistry 2018 Open Access Article. Published on 02 February 2018. Downloaded on 10/24/2024 6:47:22 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. limation of [SnCl4{nBuS(CH2)3SnBu}] (2), [SnCl4(nBu2S)2] (3) and [SnCl4(nBu2Se)2] (4) at 195 °C, 138 °C and 204 °C, respect- ively (see ESI†). LP CVD of tin sulfide thin films LP CVD experiments using precursors (2) and (3) at 286 and 372 °C (0.02 mmHg), respectively, resulted in the deposition of SnS2 in both cases, without the need for an additional chalco- gen source. Film deposition tended to occur mostly at the edge of the heated zone nearest to the precursor, typically leading to ca. 50% coverage of the substrate. Film properties are sum- marised in Table 1. Low pressure CVD experiments SEM images of the thin films (Fig. 1 and 2) revealed hexag- onal platelets formed from both precursors, with higher uni- formity and larger crystallites observed in the deposit formed from (2). In the deposition from (3), larger truncated crystal- lites are also observed. EDX analysis revealed composition of 36% Sn and 64% S and 34% Sn, 66% S for the films deposited from (2) (at T = 286 °C) and (3) (at T = 372 °C), respectively, consistent within experimental error of SnS2, with no residual Cl observed. Grazing incidence XRD patterns also show that crystall with th (3), a presen crystall posited values 5.66–5. posited 3.64, c SnS2,29 The typical nifican also ac in orde (Fig. 3) spot in 90°. Th pendic The SE Scheme 1 Preparations of the tin(IV) single source precursors used in this work. Table 1 Lattice parameters and resistivities measured for the tin chalcogenide thin Precursor Deposition temperature/°C Phase (1) 480–500 2H-SnSe2 (1) 588 SnSe (2) 286 2H-SnS2 (2) 558 SnS (3) 372 2H-SnS2 4H-SnS2 (3) 470 Mixture of SnS2/Sn2S3/SnS (4) 325 2H-SnSe2 a Not measured since discontinuous film. This journal is © The Royal Society of Chemistry 2018 Open Access Article. Published on 02 February 2018. Downloaded on 10/24/2024 6:47:22 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. LP CVD of tin sulfide thin films This journal is © The Royal Society of Chemistry 2018 Dalton Trans., 2018, 47, 2628–2637 \| 2631 Table 1 Lattice parameters and resistivities measured for the tin chalcogenide thin ﬁlms a Not measured since discontinuous film. This journal is © The Royal Society of Chemistry 2018 Dalton Trans., 2018, 47, 2628–2637 \| 2631 Fig. 1 (Left) Grazing incidence XRD pattern from the SnS2 ﬁlm deposited from [SnCl4{nBuS(CH2)3SnBu}] (2), with stick pattern from the literature data for bulk 2H-SnS2;28 (right) SEM image of SnS2 deposited from [SnCl4{nBuS(CH2)3SnBu}] (2). Paper Dalton Transactions tion 3.0 Unported Licence. View Article Online View Article Online Fig. 1 (Left) Grazing incidence XRD pattern from the SnS2 ﬁlm deposited from [SnCl4{nBuS(CH2)3SnBu}] (2), with stick pattern from the literature data for bulk 2H-SnS2;28 (right) SEM image of SnS2 deposited from [SnCl4{nBuS(CH2)3SnBu}] (2). Paper Dalton Transactions View Article Online Dalton Transactions Paper Access Article. Published on 02 February 2018. Downloaded on 10/24/2024 6:47:22 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. Open Access Article. Published on 02 February 2018. Downloaded on 10/24/2024 6:47:2 This article is licensed under a Creative Commons Attribution 3.0 Unported Fig. 1 (Left) Grazing incidence XRD pattern from the SnS2 ﬁlm deposited from [SnCl4{nBuS(CH2)3SnBu}] (2), with stick pattern from the literature data for bulk 2H-SnS2;28 (right) SEM image of SnS2 deposited from [SnCl4{nBuS(CH2)3SnBu}] (2). Fig. 2 (Left) Grazing incidence XRD pattern from SnS2 ﬁlm deposited from [SnCl4(nBu2S)2] (3) and stick patterns for bulk 2H-SnS2 28 and 4H-SnS2.29 The reﬂections marked are from the TiN layer underlying the top SiO2 on the substrate, whilst is from the Si; (right) SEM image of the SnS2 ﬁlm grown from [SnCl4(nBu2S)2] (3). Fig. 2 (Left) Grazing incidence XRD pattern from SnS2 ﬁlm deposited from [SnCl4(nBu2S)2] (3) and stick patterns for bulk 2H-SnS2 28 and 4H-Sn Fig. 2 (Left) Grazing incidence XRD pattern from SnS2 ﬁlm deposited from [SnCl4(nBu2S)2] (3) and stick patterns for bulk 2H-SnS2 28 and 4H-SnS2.29 The reﬂections marked are from the TiN layer underlying the top SiO2 on the substrate, whilst * is from the Si; (right) SEM image of the SnS2 ﬁlm grown from [SnCl4(nBu2S)2] (3). Fig. 3 3D pole ﬁgure projections with α integral of the SnS2 thin ﬁlm de- posited from (2) at 286 °C, showing the (001) plane (top) and (011) plane (bottom) where α is the tilt angle from the sample surface normal direction. This journal is © The Royal Society of Chemistry 2018 LP CVD of tin sulfide thin films the c-axis in the film plane, so the pole figure shows the bulk of the underlying material is lying flat. For the 011 reflection, a ring is observed at α = 30°. The expected value for this reflec- tion with crystallites in the [00l] orientation is calculated as α = 27.9° (tan α = a cos 30/c). Raman spectra (Fig. 4) recorded for the thin films deposited from (2) showed an intense peak at ∼320 cm−1 and a small broad feature at ∼200 cm−1. This is consistent with values reported in the literature of 313 cm−1 and 205 cm−1, represent- ing the A1g and Eg modes for 2H-SnS2.30,31 It is noted that an additional peak positioned at ∼190 cm−1 was also detected in some regions (red * in Fig. 4 (left)). This is possibly the E2 peak in the 4H-SnS2 phase, although this phase is not evident in the XRD data. The broad feature at ∼200 cm−1, present in the Raman spectrum of the film deposited from (3) is attribu- ted to the overlapping peaks (200–214 cm−1) associated with the 4H-SnS2 and 2H-SnS2 phases.31 The resistivity of the SnS2 thin film grown from (2) at 286 °C was determined by van der Pauw measurements. The film shows a low resistivity, ρ = 2.9(9) Ω cm, compared with Fig. 3 3D pole ﬁgure projections with α integral of the SnS2 thin ﬁlm de- posited from (2) at 286 °C, showing the (001) plane (top) and (011) plane (bottom) where α is the tilt angle from the sample surface normal direction. 2632 \| Dalton Trans., 2018, 47, 2628–2637 This journal is © The Royal Society of Chemistry 2018 This journal is © The Royal Society of Chemistry 2018 Fig. 4 Raman spectra of the SnS2 ﬁlms deposited from (2) (left) and (3) (right) at 286 °C. Dalton Transactions Paper View Article Online View Article Online Paper Dalton Transactions Paper Fig. 4 Raman spectra of the SnS2 ﬁlms deposited from (2) (left) and (3) (right) at 286 °C. other reported values.32,33 This could be attributed to the devi- ation from precise 1 : 2 stoichiometry. Similar behaviour was also reported by Kourtakis et al. in which a low n-type resis- tivity of ρ = 4.5(5) Ω cm was observed due to a deficit of sulfur in the SnS2 film. LP CVD of tin sulfide thin films Downloaded on 10/24/2024 6:47:22 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. Open Access Article. Published on 02 February 2018. Downloaded on 10/24/2024 6:47:22 This article is licensed under a Creative Commons Attribution 3.0 Unported Fig. 7 SEM images of SnSe2 deposited from [SnCl4(nBu2Se)2] at 325 °C top view (a); cross section (b); a thinner region of the ﬁlm deposited at 325 °C (c); top view of a SnSe2 ﬁlm deposited at 470 °C (d). osited from [SnCl4(nBu2Se)2] at 325 °C top view (a); cross section (b); a thinner region of the ﬁlm deposited at deposited at 470 °C (d). The SnS film deposited in this work exhibits a resistivity of 266(3) Ω cm, which is comparable to SnS films deposited via other techniques.35,36 respectively. These were single phase SnSe2 as identified by XRD. SEM images (Fig. 7) showed regular hexagonal platelets, typical of this space group, constituting the film deposited at In contrast, whilst SnS2 films were obtained at 372 °C using precursor (3) (containing the monodentate nBu2S ligand), LP CVD experiments at higher temperature (470 °C) resulted in mainly Sn2S3 (EDX and XRD evidence, see ESI†) and this phase was observed down to 395 °C, where both SnS2 and Sn2S3 were observed. SEM images (ESI†) show that the crystal- lites of Sn2S3 form clusters of small florets. However, there was also significant Cl incorporation evident from EDX analysis and the Sn : S ratio varies across the film, consistent with inhomogeneity. In the film obtained at 395 °C, several mor- phologies can be observed by SEM analysis and the XRD pattern obtained from this film showed a mixture of SnS, SnS2 and Sn2S3. Fig. 8 Grazing incidence XRD patterns of SnSe2 ﬁlms deposited from [SnCl4(nBu2Se)2] at 325 °C (top) and 470 °C (middle), together with the stick pattern from bulk SnSe2 (bottom). The reﬂections marked * are from crystalline TiN underlying the top SiO2 layer on the substrate. LP CVD of tin sulfide thin films In that case annealing in a sulfur atmosphere led to the film resistivity increasing to 2 × 105 Ω cm, demon- strating that non-stoichiometry was predominantly responsible for the low SnS2 resistivity.34 Fig. 5 SEM image of the polycrystalline SnS thin ﬁlm deposited in the hotter region of the furnace (558 °C) from precursor (2). Depositions were also carried out at an increased tempera- ture (558 °C) in an eﬀort to promote the formation of SnS over SnS2. Using the hotter furnace temperature, LPCVD of (2) still gave mainly SnS2 (at the region nearer the precursor), but there was also evidence of some SnS formation. The mor- phology of the SnS film appears similar to that of SnS2 (Fig. 5), but EDX analysis show a 49% : 51% Sn : S stoichiometry, within experimental error of that expected for SnS (no Cl was observed). The orthorhombic SnS phase is further confirmed by XRD analysis (Fig. 6) and the refined lattice parameters (Table 1) match well with the literature values (a = 11.20, b = 3.99, c = 4.30).12 The Raman spectrum (Fig. 6) of the film also clearly shows SnS with peaks at 220, 190 and 165 cm−1. The 165 cm−1 peak represents the B3g mode, whilst the 190 cm−1 and 220 cm−1 peaks are associated with the Ag modes and the 290 cm−1 peak is ascribed to the B2g mode.3 Fig. 5 SEM image of the polycrystalline SnS thin ﬁlm deposited in the hotter region of the furnace (558 °C) from precursor (2). Fig. 6 XRD pattern (left) and Raman spectrum (right) of SnS deposited from precursor (2) in the hotter region of the furnace (558 °C). The sharp reﬂection marked * is from the Si in the substrate. Fig. 6 XRD pattern (left) and Raman spectrum (right) of SnS deposited from precursor (2) in the hotter region of the furnace (558 °C). The sharp reﬂection marked * is from the Si in the substrate. Dalton Trans., 2018, 47, 2628–2637 \| 2633 This journal is © The Royal Society of Chemistry 2018 View Article Online Paper Dalton Transactions Fig. 7 SEM images of SnSe2 deposited from [SnCl4(nBu2Se)2] at 325 °C top view (a); cross section (b); a thinner region of the ﬁlm depos 325 °C (c); top view of a SnSe2 ﬁlm deposited at 470 °C (d). Open Access Article. Published on 02 February 2018. Tin selenide thin films Fig. 8 Grazing incidence XRD patterns of SnSe2 ﬁlms deposited from [SnCl4(nBu2Se)2] at 325 °C (top) and 470 °C (middle), together with the stick pattern from bulk SnSe2 (bottom). The reﬂections marked * are from crystalline TiN underlying the top SiO2 layer on the substrate. LPCVD experiments were also conducted using the tin(IV) chloride selenoether complex, [SnCl4(nBu2Se)2] (4). At 325 °C and 470 °C, a reflective silver and a black film were obtained, This journal is © The Royal Society of Chemistry 2018 2634 \| Dalton Trans., 2018, 47, 2628–2637 View Article Online View Article Online Dalton Transactions Paper 325 °C. A thinner region of this film, however, showed a more disordered morphology, and it was apparent that the platelets initially lie parallel to the substrate, before more material grows upon its surface. At an elevated temperature, the crystal- lites appear to be a lot smaller and rougher though still ordered. EDX analysis showed a 34 : 66 and 35 : 65 Sn : Se stoi- chiometry for the films deposited at 325 and 470 °C respect- ively. There is no evidence for incorporation of Cl into the films. Fig. 10 SEM image of the polycrystalline SnSe deposited further into the hot zone of the furnace (at 588 °C) using precursor (1). Fig. 9 Raman spectrum of SnSe2 ﬁlm deposited from (4) at 325 °C. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. Fig. 9 Raman spectrum of SnSe2 ﬁlm deposited from (4) at 325 °C. XRD analysis shows that the SnSe2 crystallises in space group of P3ˉm1 in all cases. The decrease in intensity of the 011 reflection relative to the 001 reflection (Fig. 8) suggests that the film deposited at 470 °C is thinner (and hence more strongly [001] oriented) than that deposited at 325 °C. This is borne out by cross-sectional SEM imaging and further sup- ported by the appearance of reflections belonging to the sub- strate. Lattice parameters of determined for the films de- posited at 325 °C and 470 °C (Table 1) are consistent with those from films deposited from our previously reported [SnCl4{nBuSe(CH2)nSenBu}] (n = 2, 3) (a = 3.81–3.83 and c = 6.14–6.16 Å).20 Open Access Article. Published on 02 February 2018. Downloaded on 10/24/2024 6:47:22 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. Open Access Article. Published on 02 February 2018. There are no conflicts to declare. 19 S. D. Reid, A. L. Hector, W. Levason, G. Reid, B. J. Waller and M. Webster, Dalton Trans., 2007, 4, 4769. Open Access Article. Published on 02 February 2018. Downloaded on 10/24/2024 This article is licensed under a Creative Commons Attribution 3.0 U Several new molecular tin(IV) chloride complexes with mono- dentate and bidentate thio- and selenoether ligands bearing n-butyl terminal substituents have been prepared and evalu- ated as single source precursors for the LPCVD of both tin dichalcogenide (Sn(IV)) and tin monochalcogenide (Sn(II)) thin films. All four complexes have been shown to be suitable as precursors for the deposition of crystalline SnE2 thin films, the identities of which were established by grazing incidence XRD, SEM, EDX, Raman spectroscopy, and, where a single phase was obtained, van der Pauw resisitivity measurements. Using higher deposition temperatures, precursors (1) and (2) have also been shown to form tin monochalcogenide thin films, which were characterised similarly. In contrast, higher temp- erature LPCVD experiments using the thioether complex, (3), produces multiple phases (SnS2, Sn2S3, SnS). 10 P. Zhou, X. Wang, W. Guan, D. Zhang, L. Fang and Y. Jiang, ACS Appl. Mater. Interfaces, 2017, 9, 6979. 11 L. S. Price, I. P. Parkin, T. G. Hibbert and K. C. Molloy, Chem. Vap. Deposition, 1998, 4, 222. 12 L. S. Price, I. P. Parkin, A. M. E. Hardy, R. J. H. Clark, T. G. Hibbert and K. C. Molloy, Chem. Mater., 1999, 11, 1792. 13 N. D. Boscher, C. J. Carmalt, R. G. Palgrave and I. P. Parkin, Thin Solid Films, 2008, 516, 4750. 14 S. L. Benjamin, C. H. de Groot, C. Gurnani, A. L. Hector, R. Huang, E. Koukharenko, W. Levason and G. Reid, J. Mater. Chem. A, 2014, 2, 4865. 15 R. Huang, S. L. Benjamin, C. Gurnani, Y. Wang, A. L. Hector, W. Levason, G. Reid and C. H. de Groot, Sci. Rep., 2016, 6, 27593. The resistivities measured for the SnS2 and SnS films are comparable to those reported for similar films produced from other deposition processes, whilst that for SnSe2 is at the lower end of the values reported. 16 I. P. Parkin, L. S. Price, T. G. Hibbert and K. C. Molloy, J. Mater. Chem., 2001, 11, 1486. 17 K. Ramasamy, V. L. Kuznetsov, K. Gopal, M. A. Malik, J. Raftery, P. P. Edwards and P. O’Brien, Chem. Mater., 2013, 25, 266. Acknowledgements 20 C. H. de Groot, C. Gurnani, A. L. Hector, R. Huang, M. Jura, W. Levason and G. Reid, Chem. Mater., 2012, 24, 4442. We thank STFC for funding via ST/P00007X/1, STFC and EPSRC for a CASE studentship to S. L. H. (EP/M50662X/1) and EPSRC for equipment funding (EP/K009877/1). C. G. also thanks the Royal Society for Newton Alumnus Funding. 21 D. J. Gulliver, E. G. Hope, W. Levason, S. G. Murray, D. M. Potter and G. L. Marshall, J. Chem. Soc., Perkin Trans. 2, 1984, 429. 22 ICSD: Inorganic Crystal Structure Database (ICSD), Fachinformationszentrum Karlsruhe (FIZ), accessed via the EPSRC funded National Chemical Database Service hosted by the Royal Society of Chemistry. s Article. Published on 02 February 2018. Downloaded on 10/24/2024 6:47:22 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. 7 C. Wang, Y. Chen, J. Jiang, R. Zhang, Y. Niu, T. Zhou, J. Xia, H. Tian, J. Hu and P. Yang, RSC Adv., 2017, 7, 16795. Open Access Article. Published on 02 February 2018. Downloaded on 10/24/2024 6:47:22 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. 8 K.-M. Chung, D. Wamwangi, M. Woda, M. Wuttig and W. Bensch, J. Appl. Phys., 2008, 103, 83523. Conﬂicts of interest 18 J. R. Brent, D. J. Lewis, T. Lorenz, E. A. Lewis, N. Savjani, S. J. Haigh, G. Seifert, B. Derby and P. O’Brien, J. Am. Chem. Soc., 2015, 137, 12689. There are no conflicts to declare. Dalton Transactions Dalton Transactions (Fig. 11a) show the material to be orthorhombic SnSe. The refined lattice parameters (Table 1) match well with the litera- ture values (a = 11.48–11.53, b = 4.06–4.09, c = 4.25–4.28).13 The Raman spectrum (Fig. 11b) shows peaks at 150, 135 and 115 cm−1, consistent with SnSe. The weak band at 185 cm−1 suggests a small SnSe2 component is present, although this is not evident in the XRD data. The peak at 115 cm−1 can be assigned to the B3g mode, 135 cm−1 the A2 g mode and 150 cm−1 the A3 g mode.37 The discontinuous SnSe film cover- age precluded electrical measurements. 4 F. K. Butt, M. Mirza, C. Cao, F. Idrees, M. Tahir, M. Safdar, Z. Ali, M. Tanveer and I. Aslam, CrystEngComm, 2014, 16, 3470. 5 M. S. Mahdi, K. Ibrahim, A. Hmood, N. M. Ahmed, F. I. Mustafa and S. A. Azzez, Mater. Lett., 2017, 200, 10. 6 L.-D. Zhao, S.-H. Lo, Y. Zhang, H. Sun, G. Tan, C. Uher, C. Wolverton, V. P. Dravid and M. G. Kanatzidis, Nature, 2014, 508, 373. Conclusions 9 R. Y. Wang, M. A. Caldwell, R. G. D. Jeyasingh, S. Aloni, R. M. Shelby, H. S. P. Wong and D. J. Milliron, J. Appl. Phys., 2011, 109, 113506. Tin selenide thin films Downloaded on 10/24/2024 6:47:22 This article is licensed under a Creative Commons Attribution 3.0 Unported Raman spectra (Fig. 9) show an intense peak at 185 cm−1 and a weaker feature at 115 cm−1. These peaks can be assigned to the A1g and Eg modes of SnSe2, respectively.30 Fig. 9 Raman spectrum of SnSe2 ﬁlm deposited from (4) at 325 °C. Fig. 9 Raman spectrum of SnSe2 ﬁlm deposited from (4) at 325 °C. A resistivity of 4.4(3) mΩ cm was obtained from this SnSe2 film. The value is lower than for the SnSe2 films deposited via spin-coating and annealing (∼30 mΩ cm),9 molecular beam epitaxy (22 mΩ cm)8 and CVD 210 mΩ cm.20 We also under- took further LPCVD studies to investigate film growth at higher temperatures using precursor (1), containing the biden- tate diselenoether ligand.20 Similarly to the high temperature depositions using (2) described above, these showed that increasing the furnace temperature still resulted in significant SnSe2 film growth at the edge of the hot zone as before. However, where the precursor travelled further into the hotter region of the furnace, crystalline SnSe was observed. In con- trast, we found no evidence for SnSe deposition using precur- sor (4). Fig. 10 SEM image of the polycrystalline SnSe deposited further into the hot zone of the furnace (at 588 °C) using precursor (1). The SEM images (Fig. 10) of the SnSe deposited reveal a non-continuous, orthorhombic structure. EDX analysis of the film showed a 53 : 47 Sn : Se (with no Cl evident). XRD patterns Fig. 10 SEM image of the polycrystalline SnSe deposited further into the hot zone of the furnace (at 588 °C) using precursor (1). Fig. 11 XRD pattern (a) and Raman spectrum (b) of SnSe deposited from (1). The sharp reﬂection marked * is from the Si underlying the SiO2 on the substrate. Fig. 11 XRD pattern (a) and Raman spectrum (b) of SnSe deposited from (1). The sharp reﬂection marked * is from the Si underlying the SiO2 on the substrate. Dalton Trans., 2018, 47, 2628–2637 \| 2635 This journal is © The Royal Society of Chemistry 2018 View Article Online Paper References 1 Y. Fu, G. Gou, X. Wang, Y. Chen, Q. Wan, J. Sun, S. Xiao, H. Huang, J. Yang and G. Dai, Appl. Phys. A, 2017, 123, 299. by the Royal Society of Chemistry. 23 S. Grazulis, D. Chateigner, R. T. Downs, A. F. Yokochi, M. Quiros, L. Lutterotti, E. Manakova, J. Butkus, P. Moeck and A. Le Bail, J. Appl. Crystallogr., 2009, 42, 726. 2 Y. Huang, H.-X. Deng, K. Xu, Z.-X. Wang, Q.-S. Wang, F.-M. Wang, F. Wang, X.-Y. Zhan, S.-S. Li, J.-W. Luo and J. He, Nanoscale, 2015, 7, 14093. 24 S. E. Dann, A. R. J. Genge, W. Levason and G. Reid, J. Chem. Soc., Dalton Trans., 1996, 4471; S. E. Dann, A. R. J. Genge, W. Levason and G. Reid, J. Chem. Soc., 3 I. Y. Ahmet, M. S. Hill, A. L. Johnson and L. M. Peter, Chem. Mater., 2015, 27, 7680. This journal is © The Royal Society of Chemistry 2018 2636 \| Dalton Trans., 2018, 47, 2628–2637 View Article Online Dalton Transactions Paper 31 A. J. Smith, P. E. Meek and W. Y. Liang, J. Phys. Chem. C, 1977, 1321, 1321. Dalton Trans., 1997, 2207; A. R. J. Genge, W. Levason and G. Reid, J. Chem. Soc., Dalton Trans., 1997, 4549. 32 A. Voznyi, V. Kosyak, A. Opanasyuk, N. Tirkusova, L. Grase, A. Medvids and G. Mezinskis, Mater. Chem. Phys., 2016, 173, 52. 25 L. Batt, in The Chemistry of Organic Selenium and Tellurium compounds, ed. S. Patai and Z. Rappoport, Wiley, NY, 1986, vol. 1, ch. 4. 26 C. T. Knight and A. E. Merbach, Inorg. Chem., 1985, 24, 576. 33 L. Amalraj, C. Sanjeeviraja and M. Jayachandran, J. Cryst. Growth, 2002, 234, 683. 34 K. Kourtakis, J. Di Carlo, R. Kershaw, K. Dwight and A. Wold, J. Solid State Chem., 1988, 76, 186. 27 Y. Kumagai, L. A. Burton, A. Walsh and F. Oba, Phys. Rev. Appl., 2016, 6, 1. 28 R. M. Hazen and L. W. Finger, Am. Mineral., 1978, 63, 289. 35 K. T. Ramakrishna Reddy, P. Purandhara Reddy, P. K. Datta and R. W. Miles, Thin Solid Films, 2002, 403–404, 116. 29 J. R. Guenter and H. R. Oswald, Naturwissenschaften, 1968, 55, 177. 36 T. Gotoh, Phys. Status. Solidi A, 2016, 213, 1869. 30 D. Walsh, S. Jandl and J. Harbec, J. Phys. C: Solid State Phys., 1980, 13, 125. 37 S. Luo, X. Qi, H. Yao, X. This journal is © The Royal Society of Chemistry 2018 Dalton Trans., 2018, 47, 2628–2637 \| 2637 References Ren, Q. Chen and J. Zhong, J. Phys. Chem. C, 2017, 121, 4674. Dalton Trans., 2018, 47, 2628–2637 \| 2637
https://openalex.org/W2990949115	https://europepmc.org/articles/pmc6978483?pdf=render	English	null	Women's empowerment and child nutrition: The role of intrinsic agency	SSM, population health	2,019	cc-by	12,686	Rebecca Jones a,, Regine Haard€orfer b, Usha Ramakrishnan a,c, Kathryn M. Yount c,d, Stephanie Miedema d, Amy Webb Girard a,c a Nutrition and Health Sciences Program, Laney Graduate School, Emory University, USA b Department of Behavioral Sciences and Health Education, Emory University, USA c Hubert Department of Global Health, Emory University, USA d Department of Sociology, Emory University, USA A R T I C L E I N F O Keywords: Women’s empowerment Nutritional status Maternal and child health East Africa Women’s empowerment is associated with improved child nutrition, and both underpin the achievement of multiple Sustainable Development Goals (SDGs). We examined pathways by which women’s empowerment influences child nutritional status. We pooled nationally representative data from Demographic and Health Surveys (2011–2016) collected from married women with children aged 6–24 months in Ethiopia, Kenya, Rwanda, Tanzania, and Uganda (n ¼ 13,780). We operationalized child nutritional status using anemia, height- for-age z-score (HAZ), and weight-for-age z-score (WHZ). We operationalized women’s empowerment using a validated measure comprised of three latent domains: social/human assets (“assets”), intrinsic agency (attitudes about intimate partner violence), and instrumental agency (influence in household decision making). We used structural equation models with latent constructs to estimate hypothesized pathways from women’s empower ment to child nutritional status with further mediation by maternal body mass index (BMI) and stratification by wealth. Women’s empowerment domains were directly and positively associated with maternal BMI (estimateSE: assets, 0.17 0.03; intrinsic agency, 0.23 0.03; instrumental agency, 0.03 0.01). Maternal BMI was directly and positively associated with child HAZ (0.08 0.04) and child WHZ (0.35 0.03). Assets were indirectly associated with child HAZ and WHZ through intrinsic agency and maternal BMI. In the lowest wealth category, the direct effects from women’s empowerment to child nutritional status were significant (assets and instrumental agency were associated with anemia; intrinsic agency associated with HAZ). In the highest wealth category, direct effects from women’s empowerment on child nutritional status were significant (intrinsic and instrumental agency associated with WHZ). Improving women’s empowerment, especially intrinsic agency, in East Africa could improve child nutrition directly and via improved maternal nutrition. These findings suggest that efforts to realize SDG 5 may have spillover effects on other SDGs. However, strategies to improve nutrition through empowerment approaches may need to also address household resource constraints. Corresponding author. Atlanta USA 1518 Clifton Rd NE, Atlanta, GA, 30322, USA. E-mail address: rebecca.evelyn.jones@emory.edu (R. Jones). 1 Nutrition-specific interventions aim to directly improve nutrient intakes and diet practices; examples include diet-related behavior change, supplementation, and fortification. Research documents that, on their own, they are insufficient to achieve target reductions in stunting even when delivered at scale and with high (>80%) coverage (Black, Victora, Walker, Bhutta, Christian, De Onis, & al., 2013; Ruel & Alderman, 2013). SSM - Population Health 9 (2019) 100475 SSM - Population Health 9 (2019) 100475 Women’s empowerment and child nutrition: The role of intrinsic agency Rebecca Jones a,*, Regine Haard€orfer b, Usha Ramakrishnan a,c, Kathryn M. Yount c,d, Stephanie Miedema d, Amy Webb Girard a,c Available online 20 November 2019 2352-8273/© 2020 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). https://doi.org/10.1016/j.ssmph.2019.100475 Received 15 September 2018; Received in revised form 8 April 2019; Accepted 22 August 2019 Available online 20 November 2019 2352-8273/© 2020 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Received 15 September 2018; Received in revised form 8 April 2019; Accepted 22 August 2019 Available online 20 November 2019 2352-8273/© 2020 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). 1. Introduction (Abada & Tenkorang, 2012; Balk, 1994; Bloom, Wypij, & Das Gupta, 2001; Carlson et al., 2015; DeRose & Ezeh, 2005; Hadley, Patil, & Nahayo, 2010; Hou & Ma, 2013; James-Hawkins, Peters, VanderEnde, Bardin, & Yount, 2018a; Morgan & Niraula, 1995; Patel et al., 2006; Qadir, Khan, Medhin, & Prince, 2011; Rahman, 2012; Upadhyay & Hindin, 2005; Woldemicael & Tenkorang, 2010; Yount, Dijkerman, Zureick-Brown, & VanderEnde, 2014). i Moreover, women’s empowerment, despite differences in oper ationalization, has been associated with practices that indirectly affect child nutrition. Examples include practices that influence fertility, birth spacing, mental health, and health seeking e.g. (Abada & Tenkorang, 2012; Balk, 1994; Bloom, Wypij, & Das Gupta, 2001; Carlson et al., 2015; DeRose & Ezeh, 2005; Hadley, Patil, & Nahayo, 2010; Hou & Ma, 2013; James-Hawkins, Peters, VanderEnde, Bardin, & Yount, 2018a; Morgan & Niraula, 1995; Patel et al., 2006; Qadir, Khan, Medhin, & Prince, 2011; Rahman, 2012; Upadhyay & Hindin, 2005; Woldemicael & Tenkorang, 2010; Yount, Dijkerman, Zureick-Brown, & VanderEnde, 2014). i Based on this discussion, we propose three hypotheses for the cross- national evaluation of pathways of women’s empowerment to child nutritional status in five countries in East Africa. First, in a pooled sample, we expect that women’s empowerment, comprised of three latent factors (assets, intrinsic agency, and instrumental agency) will be associated with child nutritional status. Second, we expect that this relationship is mediated by women’s nutritional status, measured using BMI. We acknowledge that the window for maternal nutrition to influ ence child nutrition may be narrow; however, this window may be wider than previously thought and may include preconception, with the more recent emphasis on preconception nutrition (Ramakrishan, Grant; Goldenberg et al., 2012; Young, Nguyen, Addo, et al., 2015; Hambidge et al., 2019). Furthermore, the implications are broader, given the importance of direct nutritional transfer and the impact of childcare capabilities. Finally, we evaluate the relationship of household wealth in confounding or modifying these proposed pathways. Evidence in sup port of these hypotheses would provide a theoretically informed, empirical example for using a subset of standard, cross-national survey Despite these significant associations observed between women’s empowerment and child nutrition, the pathways of influence and do mains of women’s empowerment most relevant for child nutrition remain due in large part to the heterogeneity in describing, defining, and operationalizing women’s empowerment (Kishor & Subaiya, 2008; Pratley, 2016). 1. Introduction Understanding the contextually relevant domains through which empowerment influences child nutrition is important for designing and implementing contextually appropriate policies and programs. Identifying those elements of empowerment that are relevant for child nutrition yet are stable or invariant across contexts may provide useful insights for monitoring empowerment and progress towards multiple Sustainable Development Goals SDGs related to both nutrition and gender. g Wealth is consistently associated with the quality of child diets and child nutritional status (Alderman, Haddad, Headey, & Smith, 2014; Alderman & Headey, 2018; Krishna et al., 2015). In keeping with Kab eer’s framework, household wealth may be considered a resource for empowerment if it facilitates increased transfer of resources to women, as has been hypothesized in Kenya (Voronca, Walker, & Egede, 2018). However, as Hanmer and Klugman note (2016, pg 241), there exist “context-specific dimensions to and constraints, on choice—ncluding poverty and access to services—that shape the options that are open to people.” Therefore, there may be a threshold requirement for commu nity, household, or individual wealth before women are able to act on their agency to achieve goals (Field, Jayachandran, & Pande, 2010; Habibov, Barrett, & Chernyak, 2017; Kabeer, 1999). In contrast, Kan diyoti argues that household poverty can weaken patriarchy because it compels women to go out of the home to work (Kandiyoti, 1988). Thus, in addition to examining the pathways by which women’s empower ment is associated with child nutritional status, we explored the extent to which household wealth may confound or modify these pathways. Although we acknowledge the limitations in evaluating household wealth across five separate contexts, this measure is one of the few available to evaluate socioeconomic status within these samples. Furthermore, the Human Development Index (HDI) ranking for these five countries during our time period of study is between 0.463 and 0.590, which signifies a similar level of average achievement in key dimensions of human development across the countries.2 Moreover, women’s empowerment, despite differences in oper ationalization, has been associated with practices that indirectly affect child nutrition. Examples include practices that influence fertility, birth spacing, mental health, and health seeking e.g. 2 Human Development Index is a summary measure of average achievement in key dimensions of human development: a long and healthy life, being knowledgeable, and having a decent standard of living. Website: http://hdr.org /en/content/human-development-index.hdi. 1. Introduction 2013; Ramakrishnan, Imhoff-Kunsch, & Martorell, 2014), these effects often are small in magnitude. Accelerating progress to achieve this new target requires not only expansion of evidence-based nutrition-specific strategies1 but greater prioritization and scale up of approaches that tackle the underlying determinants of child nutrition (Bhutta et al., 2013; Black et al., 2008, 2013; Ruel & Alderman, 2013). In 2015, global leaders pledged to reduce chronic malnutrition in children younger than 5 years by 40 percent before the year 2025 (In ternational Food Policy Research Institute, 2016). Despite global efforts, only three regions achieved at least a 50% decline in stunting from 1990 to 2015, (UNICEF, 2015). Although nutrition-specific interventions to improve maternal and child nutrition are efficacious (Bhutta et al., Women’s empowerment is one such underlying determinant of child nutrition, with substantial evidence that it positively influences child SSM - Population Health 9 (2019) 100475 R. Jones et al. nutrition and growth (Carlson, Kordas, & Murray-Kolb, 2015; van den Bold, Quisumbing, & Gillespie, 2013). Scholars define women’s empowerment as the ability of a woman to claim enabling resources, exercise voice and agency, and act on desires to transform her own life in contexts where this ability has been denied (Kabeer, 1999). In many societies, women are responsible for the procurement, preparation, and allocation of food and ultimately preserving food culture (Hodgson, 1999; Holtzman, 2002). As Holtzman describes, among the Samburu pastoralists of northern Kenya, “women are foremost constituted as food givers, responsible for providing sustenance to children and elders” (Holtzman p.1045). Thus, it is unsurprising that women’s empowerment is associated with improved household and own dietary diversity (Amugsi, Lartey, Kimani, & Mberu, 2016;Galie et al., 2019 ; Malapit, Kadiyala, Quisumbing, Cunningham, & Tyagi, 2015; Yimer & Tadesse, 2015) and improved optimal infant and young-child feeding practices (Bose, 2011; Ickes, Heymsfield, Wright, & Baguma, 2017; Na, Jennings, Telegawker, & Ahmed, 2015; Shroff et al., 2011; Smith et al., 2002). Relative to disempowered women, more empowered women may have a greater influence on budgets allocated to household food procurement and how food is allocated in the household and greater agency in how they choose to feed their infants, contributing to improved diets and nutritional status. 1. Introduction Integrating women’s empowerment into social and behavioral strategies for improving nutrition may shift societal and household norms away from women’s role as “last and least” in food prioritization (Carlson et al., 2015; Gittelsohn, Thapa, & Landman, 1997; Ruel & Alderman, 2013; Smith et al., 2002). Empowerment of women resulting in more equitable distributions of power within the household may contribute to prioritization of women’s own nutrition and health, which may manifest in improved body mass index (BMI) and micronutrient status (Ramakrishnan, Grant, Goldenberg, Zongrone, & Martorell, 2012; Young, Nguyen, Addo, et al., 2015; Hambidge et al., 2019). This shift could enhance children’s nutrition through direct biological transfer of nutrients in utero and in early infancy (Ackerson & Subramanian, 2008; Mason et al., 2014; Ramakrishnan et al., 2014; Shrimpton & Rokx, 2012; Wu, Imhoff-Kunsch, & Girard, 2012) and improved childcare capabilities throughout childhood (Engle, Menon, & Haddad, 1999; Matare, Mbuya, Pelto, Dickin, & Stoltzfus, 2015; Nuss baum, 2001). Hollos, 2003; Njuki, Kaaria, Chamunorwa, & Chiuri, 2011). Therefore, women in East Africa may hold greater autonomy in, for example, economic decision making than their South Asian and Middle Eastern counterparts. However, despite these opportunities for economic au tonomy, persistent gender differentials in the value of what is controlled reflect gendered differences in intrahousehold power and influence household cooperation strategies, especially in rural agricultural households (Njuki et al., 2011). As a result, the process for women to become empowered and the domains most relevant for child nutrition may differ across contexts. Understanding the contextually relevant domains through which empowerment influences child nutrition is important for designing and implementing contextually appropriate policies and programs. Identifying those elements of empowerment that are relevant for child nutrition yet are stable or invariant across contexts may provide useful insights for monitoring empowerment and progress towards multiple Sustainable Development Goals SDGs related to both nutrition and gender. Hollos, 2003; Njuki, Kaaria, Chamunorwa, & Chiuri, 2011). Therefore, women in East Africa may hold greater autonomy in, for example, economic decision making than their South Asian and Middle Eastern counterparts. However, despite these opportunities for economic au tonomy, persistent gender differentials in the value of what is controlled reflect gendered differences in intrahousehold power and influence household cooperation strategies, especially in rural agricultural households (Njuki et al., 2011). As a result, the process for women to become empowered and the domains most relevant for child nutrition may differ across contexts. 1. Introduction Integrating women’s empowerment into social and behavioral strategies for improving nutrition may shift societal and household norms away from women’s role as “last and least” in food prioritization (Carlson et al., 2015; Gittelsohn, Thapa, & Landman, 1997; Ruel & Alderman, 2013; Smith et al., 2002). Empowerment of women resulting in more equitable distributions of power within the household may contribute to prioritization of women’s own nutrition and health, which may manifest in improved body mass index (BMI) and micronutrient status (Ramakrishnan, Grant, Goldenberg, Zongrone, & Martorell, 2012; Young, Nguyen, Addo, et al., 2015; Hambidge et al., 2019). This shift could enhance children’s nutrition through direct biological transfer of nutrients in utero and in early infancy (Ackerson & Subramanian, 2008; Mason et al., 2014; Ramakrishnan et al., 2014; Shrimpton & Rokx, 2012; Wu, Imhoff-Kunsch, & Girard, 2012) and improved childcare capabilities throughout childhood (Engle, Menon, & Haddad, 1999; Matare, Mbuya, Pelto, Dickin, & Stoltzfus, 2015; Nuss baum, 2001). nutrition and growth (Carlson, Kordas, & Murray-Kolb, 2015; van den Bold, Quisumbing, & Gillespie, 2013). Scholars define women’s empowerment as the ability of a woman to claim enabling resources, exercise voice and agency, and act on desires to transform her own life in contexts where this ability has been denied (Kabeer, 1999). In many societies, women are responsible for the procurement, preparation, and allocation of food and ultimately preserving food culture (Hodgson, 1999; Holtzman, 2002). As Holtzman describes, among the Samburu pastoralists of northern Kenya, “women are foremost constituted as food givers, responsible for providing sustenance to children and elders” (Holtzman p.1045). Thus, it is unsurprising that women’s empowerment is associated with improved household and own dietary diversity (Amugsi, Lartey, Kimani, & Mberu, 2016;Galie et al., 2019 ; Malapit, Kadiyala, Quisumbing, Cunningham, & Tyagi, 2015; Yimer & Tadesse, 2015) and improved optimal infant and young-child feeding practices (Bose, 2011; Ickes, Heymsfield, Wright, & Baguma, 2017; Na, Jennings, Telegawker, & Ahmed, 2015; Shroff et al., 2011; Smith et al., 2002). Relative to disempowered women, more empowered women may have a greater influence on budgets allocated to household food procurement and how food is allocated in the household and greater agency in how they choose to feed their infants, contributing to improved diets and nutritional status. 2.1. Framework and rationale g We operationalized women’s empowerment using three latent domains—enabling resources, intrinsic agency and instrumental agen cy—identified through exploratory and confirmatory factor analysis in random split-half samples (Miedema et al., 2018). The invariance of these domains across countries was then examined through multigroup confirmatory factor analysis. The final three latent domains included three items focused on human and social assets that related to women’s enabling resources (“assets”) and reflected levels of women’s repro ductive and marital power over their lifetime. These included ages at first sex, first cohabitation, and first birth.3 The latent domain of intrinsic agency captured whether women’s expression of gender atti tudes rejected or reflected normative beliefs of gendered violence (Yount, Krause & VanderEnde, 2016); it included five items related to women’s attitudes about justification of wife beating under various conditions. The latent domain of instrumental agency measured women’s ability to exercise decision making in the household. This domain included four items related to influence in decisions about large purchases, visiting family and friends, respondent’s earnings, and re spondent’s health. The response options were respondent, respondent’s partner/husband, both the respondent and her partner/husband, or someone else. Responses were coded 1 for women who reported sole or joint decision making with husband and 0 otherwise. More specifics about each WE domain and individual indicators are found in the Ap pendix (Supplemental Table 1). i We adapted the framework proposed by Kabeer (1999) to depict enabling human and social assets as well as selected dimensions of women’s agency as determinants of their achievements, in this case child nutritional status (Fig. 1). Using this adapted framework, we developed and psychometrically validated a three-dimensional model for women’s empowerment in East Africa (Miedema, Haard€orfer, Gir ard, & Yount, 2018) that was measurement invariant across five coun tries in East Africa: Kenya, Tanzania, Uganda, Rwanda, and Ethiopia. Although geographically contiguous and often assumed to be culturally homogenous, these five countries differ substantially in their historical trajectories, political systems, and sociocultural contexts, all of which influence gender norms and empowerment processes (Miedema et al., 2018). The lack of research on women’s empowerment in East Africa relative to other regions and the heterogeneity in gender dynamics across this region provide strong rationale for testing an invariant model of women’s empowerment across this region (Miedema et al., 2018) and assessing associations with child nutrition. 2.1. Framework and rationale The women’s empowerment model developed through this process consisted of women’s human and social assets, intrinsic agency to reject violence against women, and instrumental agency in household decisions. To fill this gap, we used a women’s empowerment measure previously developed and validated by our team to quantify the influence of different empowerment domains on child anemia and anthropometry in five East African countries and the role of women’s nutritional status as a mediator or confounder of this relationship. Height-for-age z-score (HAZ) is a measure of chronic un dernutrition and captures linear growth in utero and early childhood (De Onis et al., 2007; WHO Multicentre Growth Reference Study Group, 2006). HAZ is among the preferred indicators of childhood nutrition and is a strong predictor of human capital (Victora et al., 2008). Weight-for-age z-score (WHZ) is an indicator of acute malnutrition. Anemia, measured using blood hemoglobin, is used as a proxy for iron deficiency in large-scale surveys. We hypothesized that each of these three domains of women’s empowerment would be positively associated with child nutritional status and hemoglobin (defined as achievements). We previously identified positive and direct associations between domains of women’s empowerment and women’s nutritional status (Jones et al., 2019). Thus, we treated maternal nutritional status as a mediator measured as BMI (defined as weight in kg/[height in m]2). i g g g We used the DHS household wealth index as a measure of socio economic status. The DHS household wealth index is a composite measure of a household’s cumulative living standard. The wealth index is calculated using data on a household’s ownership of selected assets, such as televisions and bicycles; materials used for housing construction; and types of water access and sanitation facilities (Macro International Inc, 2006). The index is generated through principle components anal ysis, and the derived score from the first component ranks households on a continuous scale of relative wealth. The DHS then separates all interviewed households into five wealth quintiles to allow for the ability to compare the influence of wealth. For ease of interpretation, we recalculated these quintiles into three categories—low, middle, and high household wealth—to evaluate differences in hypothesized paths across household wealth groups and ran analyses stratified on these categories. 3 Education, commonly considered an enabling resource in Kabeer’s frame work and participation in economic activities were each tested in EFA but dropped in subsequent analyses due to inconsistent and low factor loadings across countries. 2.2. Data resources and study population We conducted a secondary analysis of data from the most recent Demographic and Health Surveys (DHSs) (2011–2016) for married women with children aged 6–24 months in Ethiopia, Kenya, Rwanda, Tanzania, and Uganda (n ¼ 13,780). DHSs are nationally representative, cross-sectional surveys conducted in low- and middle-income countries since 1985 (ICF International, 2012b) and typically are based on a stratified, multistage probability sample design (Rutstein & Rojas, 2006). DHSs routinely include data on health and nutritional status based on anthropometric measurements and blood hemoglobin levels for women and children. Women were included in the present analysis if they were aged 15–49 years, nonpregnant, married or living with a partner at the time of the survey, and had a child aged 6–24 months with anthropometry measurements. (The sample size for each country was as follows: Ethiopia, n ¼ 2664; Kenya, n ¼ 4918; Rwanda, n ¼ 1925; Tanzania, n ¼ 2490; and Uganda, n ¼ 1783). Overall, response rates for women eligible to take part in the DHS ranged from 93.8% to 99.1% across the DHSs in this study. 2.1. Framework and rationale We acknowledge that based on Kabeer’s framing, wealth can also be a resource for empowerment; however, because of the cross-sectional nature of the data, we are restricted to evaluating wealth as a modi fier. The overarching aim of this study was to evaluate a previously created measurement-invariant three-factor construct of women’s empowerment and its associations with health outcomes (Miedema et al., 2018). Given this focus, we evaluated household wealth in the pooled sample. To strengthen our results, we further evaluated indi vidual country paths by controlling first for household wealth and then stratifying by household wealth (Appendix). 1. Introduction Studies assessing associations between WE and child nutrition have operationalized this construct using myriad direct and indirect measures of women’s empowerment. Moreover, scholars generally agree that women’s empowerment is a process that operates in context-specific ways (Yount, Peterman, & Cheong, 2018). Heteroge neity in patriarchal family systems and the societal conditions that subordinate women (Kandiyoti, 1988) contributes to context specificity. For example, compared with South Asia and the Middle East, which historically are more classically patriarchal, women in East Africa participate to a greater extent in the household economy through in come generation and often control low-revenue commodities such as vegetables, milk, or eggs (Dolan, 2001; Kandiyoti, 1988; Larsen & 2 SSM - Population Health 9 (2019) 100475 R. Jones et al. items in multicountry assessments of the consequences of women’s empowerment, in alignment with SDG 5. nutritional status (HAZ and WHZ) using the World Health Organization (WHO) Child Growth Standard. Following altitude adjustment of he moglobin, anemia status was categorized as either anemic or not anemic based on DHS cutoffs (<11.0 g/dL) (ICF International, 2012a; Pullum, Collison, Namaste, & Garrett, 2017). The 2014 Kenya DHS did not collect data on hemoglobin. 2.3. Variables Three indicators of child’s nutritional status were used in these an alyses: HAZ, WHZ, and anemia status. We estimated children’s 3 R. Jones et al. SSM - Population Health 9 (2019) 100475 SSM - Pop Fig. 1. Conceptual tested path models from Women’s empowerment to Children’s nutritional status. Fig. 1. Conceptual tested path models from Women’s empowerment to Children’s nutritional status. are less than \|0.05\| were considered trivial (Hatcher, 1994; Muthen & Muthen, 2001). Indirect effects were considered significant based on bootstrapping methods (Bollen & Stine, 1990). Each test was resampled 1000 times, and an accelerated 95% confidence interval (95% CI) was determined. When the 95% CI did not include zero, the indirect asso ciation was significant. The hypothesized model was first evaluated for all five countries individually (women’s empowerment to child nutri tional status) and adjusted for household wealth. Country-level samples were then merged and used for the full five-country path analysis. The model accounted for country-level sample clusters, using the DHS cluster variable by country. Next, we evaluated the five-country model mediated by maternal BMI for outcomes HAZ and WHZ. Finally, we evaluated a multigroup model stratified by household wealth (low, middle, and high) for the full five-country sample mediated by maternal BMI. We further evaluated these stratified household wealth models by each individual country and saw no large differences in trends (Sup plemental Fig. 3) Of note, the path analysis for Rwanda was unique with respect to associations between empowerment domain and indicators of 2.4. Statistical analysis 2.4. Statistical analysis are less than \|0.05\| were considered trivial (Hatcher, 1994; Muthen & Muthen, 2001). Indirect effects were considered significant based on bootstrapping methods (Bollen & Stine, 1990). Each test was resampled 1000 times, and an accelerated 95% confidence interval (95% CI) was determined. When the 95% CI did not include zero, the indirect asso ciation was significant. The hypothesized model was first evaluated for all five countries individually (women’s empowerment to child nutri tional status) and adjusted for household wealth. Country-level samples were then merged and used for the full five-country path analysis. The model accounted for country-level sample clusters, using the DHS cluster variable by country. Next, we evaluated the five-country model mediated by maternal BMI for outcomes HAZ and WHZ. Finally, we evaluated a multigroup model stratified by household wealth (low, middle, and high) for the full five-country sample mediated by maternal BMI. We further evaluated these stratified household wealth models by each individual country and saw no large differences in trends (Sup plemental Fig. 3) Of note, the path analysis for Rwanda was unique with respect to associations between empowerment domain and indicators of We calculated means and standard deviation for all women’s empowerment and health items for the overall sample (Tables 1 and 2) as well as by individual country (Supplemental Table 2). The relative frequencies of all variables were estimated to assess completeness and distributions. Spearman’s rho correlations (Kline, 1998) were estimated to measure bivariate associations between items. For continuous vari ables, correlations were also assessed. STATA 14.1 software (College Station, Texas) was used to run descriptive analyses. Data were trans ferred to Mplus 8 statistical software (Los Angeles, California) for all model building, path analyses, and estimation. We used structural equation models (SEM) with latent constructs to estimate the strength of hypothesized pathways from enabling resources to child nutritional status, directly and through women’s intrinsic and instrumental agency. We present models with standardized path co efficients using weighted least squares with mean and variance-adjusted estimation for categorical data, and standardized path coefficients that a Abbreviation: SD, standard deviation. 3.2. Reduced-form latent structural equation model b Anemia status excluding Kenya (n ¼ 4918). b Anemia status excluding Kenya (n ¼ 4918). c In intrinsic agency, the proportion represents the women who disagree with wife-beating for various reasons listed. For our reduced-form five country model (Fig. 2), the standardized direct association of intrinsic agency with child HAZ was 0.05 (95% CI, 0.01, 0.09). Each unit increase in intrinsic agency was associated with an increase of 0.08 in HAZ (Table 3). Intrinsic agency and assets also were directly associated with WHZ (0.11 [0.03, 0.18] and 0.09 [0.01,0.17], respectively). Each unit increase in intrinsic agency was associated with an increase of 0.09 in WHZ. Instrumental agency was directly associated with WHZ (0.05 [0.02, 0.08]) and Anemia (0.08 [0.05, 0.11]). The in direct associations between assets and HAZ (0.01 [0.003, 0.02]) and WHZ (0.03 [0.007, 0.05]) through intrinsic agency were significant. child nutritional status (Supplemental Fig. 4). A sensitivity analysis was run for the full country model, with wealth stratification excluding the Rwandan sample (n ¼ 11,855). There were some differences between country models, specifically with regard to associations between empowerment domains; however, because of the theoretical reasoning and the focus on empowerment’s association with nutritional status, we present full five-country results. Because of low correlations between child nutritional outcomes and maternal age in all five countries, we did not control for age. i i When evaluating the associations between empowerment domains across countries, the standardized associations between assets and intrinsic agency (0.24 [0.19, 0.28]) and instrumental agency (0.16 [0.07, 0.25]) were significant (Fig. 2). Country-specific path analyses differed somewhat in the strengths of associations and some pathways (See Supplemental Fig. 1). For example, in Rwanda, assets were not associated with either intrinsic or instrumental agency, while in Kenya, assets were associated with both intrinsic agency (0.22 [0.17, 0.25]) and instrumental agency (0.13 [0.09, 0.16]). Our interpretations of the adequacy of model fit are based on theo retical interpretation and three measures of goodness-of-fit: the Comparative Fit Index (CFI), Tucker Lewis Index (TLI), and root mean square error of approximation (RMSEA) (Kline, 1998). Acceptable threshold levels for fit indices were CFI greater than 0.95, TLI greater than 0.95, and RMSEA less than 0.6 (Hooper, Coughlan, & Mullen, 2008). Chi-square fit indices consider acceptable threshold level as low χ2 relative to degrees of freedom and a nonsignificant P value (Hooper et al., 2008). 3.2. Reduced-form latent structural equation model However, because of the sample sizes required for such models, these fit indices are rarely informative and we do not report them (Bandalos & Finney, 2010). All SEM analyses were conducted using Mplus software, and direct associations, indirect associations, and standard errors were generated for each pathway. Table 1 SSM - Population Health 9 (2019) 100475 R. Jones et al. R. Jones et al. 3.1. Sample characteristics Table 1 shows selected demographic characteristics for women with children aged 6–24 months by country. Not all variables in Table 1 are used for further analyses but rather for a picture of the sample within each country. Mean age of the women ranged between 28 years (Kenya, Uganda) and 30 years (Rwanda). Mean education (in years) of the women ranged from 4.7 to 6.4 years, with the exception of Ethiopia, in which the mean education was 1.4 years. On average, women had had between 2 and 3 children. The percentage of women living in an urban area ranged from 18% in Ethiopia to 32% in Kenya. Table 2 shows selected characteristics for women in the full sample by household wealth categories (n ¼ 13,780). The measures of enabling resources, mean ages at first sex, first cohabitation, and first birth, increased with increasing household wealth. With regard to the mea sures of intrinsic agency, the percentage of respondents who did not justify wife-beating under certain conditions varied by condition and household wealth category; 43% of respondents in the low wealth category did not justify wife beating if a woman neglected her children compared with 49% and 61% in the middle and high wealth categories, respectively. If a woman burned food, 66%–80% of all respondents did not believe this was grounds for abuse. In relation to instrumental agency, 84% or more of respondents were responsible for decisions about their own incomes. For the other three items, the proportion of women reporting sole or joint decision making increased with increasing household wealth category (P < .05). The mean BMI of women was in the normal range across all wealth categories, ranging from 21.0 3.1 kg/m2 in the low category to 23.5 1.7 kg/m2 in the high category. Across wealth categories, mean HAZ and WHZ scores ranged from 1.0 to 1.4 and from 0.1 to 0.4, respectively. The prevalence of anemia ranged from 39% to 45%. a Abbreviation: SD, standard deviation. Table 1 Table 1 Selected characteristics of East African women and children sampled by demographic health surveys (2011–2016) (n ¼ 13,780). Variable Ethiopia 2011 Kenya 2014 Rwanda 2014–2015 Tanzania 2015–2016 Uganda 2011 (n ¼ 2664) (n ¼ 4918) (n ¼ 1925) (n ¼ 2490) (n ¼ 1783) Mean /% SD Mean /% SD Mean /% SD Mean /% SD Mean /% SD Mother Age 28.1 6.37 28.0 6.24 30.1 5.93 28.5 6.90 28.0 6.49 Education No education 71% 24% 13% 24% 16% Primary 25% 51% 74% 58% 62% Secondary 3% 19% 11% 17% 17% Higher 2% 7% 3% 1% 5% Parity 2.3 1.16 2.4 1.25 2.2 0.78 2.6 0.82 3.1 0.51 Urban 18% 32% 22% 22% 20% Body mass index Underweight 26% 13% 4% 8% 14% Normal weight 68% 61% 70% 70% 71% Overweight 5% 18% 21% 15% 11% Obese 1% 7% 5% 7% 4% Household wealth Low 48% 56% 45% 43% 47% Middle 16% 17% 18% 19% 18% High 36% 27% 36% 38% 36% Child Male 50% 51% 50% 49% 50% Note. a Abbreviation: SD standard deviation n women and children sampled by demographic health surveys (2011–2016) (n ¼ 13,780). cted characteristics of East African women and children sampled by demographic health surveys (2011–2016) (n ¼ 13,780). Table 2 Women’s Empowerment, BMI, and, Child Nutritional Status of East African Women and Children Sampled by Demographic Health Surveys (2011–2016) (n ¼ 13,780). Variable Low Middle High (n ¼ 9210) (n ¼ 3068) (n ¼ 5800) Mean /% SD Mean /% SD Mean /% SD Domain: Human and Social Assets (Assets) Age at first sex 16.5 3.15 17.0 3.25 18.1 3.65 Age at first cohabitation 17.6 3.42 18.2 3.54 19.4 4.02 Age at first birth 18.8 3.44 19.2 3.21 20.5 3.87 Domain: Intrinsic Agency Goes out without spouse 50% 57% 68% If neglects child 43% 49% 61% If argues with husband 51% 56% 69% If refuses sex 57% 63% 74% If burns food 66% 70% 80% Domain: Instrumental Agency For respondent’s earnings 83% 88% 92% For respondent’s health 68% 73% 75% For large purchases 52% 59% 61% For family/friend’s visits 63% 68% 73% Mediator: Women’s BMI 21.0 3.01 21.7 1.95 23.5 1.74 Outcomes: Child Nutritional Status Height-for-age z-Score 1.4 1.78 1.3 1.34 1.0 1.64 Weight-for-age z- Score 0.4 1.15 0.3 1.21 0.1 1.18 Anemia status 39% 43% 45% Note. a Abbreviation: SD, standard deviation. R. Jones et al. 3.4. Mediation model with maternal BMI ii a Model fit statistics: df ¼ 78; CFI ¼ 0.995; TLI: 0.993; RMSEA: 0.004 (0.001,0.006). ii b Significant paths based upon 95% confidence intervals. i b Significant paths based upon 95% confidence intervals. i c All nonsignificant paths were dropped for simplicity. c All nonsignificant paths were dropped for simplicity. d Anemia model excludes Kenya (n ¼ 4918). e Abbreviations: HAZ, height-for-age z-score; WHZ, weight-for-age z-score. df, degrees of freedom; CFI, comparative fit index; TLI, tucker lewis index; RMSEA, root-mean square error of approximation. e Abbreviations: HAZ, height-for-age z-score; WHZ, weight-for-age z-score. df, degrees of freedom; CFI, comparative fit index; TLI, tucker lewis index; RMSEA, root-mean square error of approximation. Women’s BMI was positively and directly associated with WHZ (0.35 [0.26, 0.43]) and child HAZ (0.08 [0.05, 0.10]). The indirect path from assets to intrinsic agency to maternal BMI was significant (0.06 [0.03, 0.080]). The extension of the indirect path from intrinsic agency through women’s BMI to WHZ was significant (0.08 [0.04, 0.11]). anemia. Associations between empowerment domains and nutritional outcomes differed by wealth categories. For anthropometric status, but not anemia, the pathways from the women’s empowerment domains were mediated through their effects on maternal BMI. Prior research documented similar associations between maternal age at pivotal life events (the domain of assets), attitudes about intimate partner violence (IPV) (the domain of intrinsic agency), and indicators of child stunting (HAZ) and wasting (WHZ) (Carlson et al., 2015). Our use of a path analysis enabled deeper understanding of how these do mains function. We observed that when maternal BMI was included as a mediator of the relationship, direct effects diminished and were replaced by pathways that operated through maternal BMI. This finding has several potential explanations and implications. First, women’s empowerment serves as a foundational path to improve child nutrition. These improvements may derive from women’s increased agency to ensure adequate food and care not only for their child but also for themselves (Richards et al., 2013). Research documents women’s ten dency to eat last and least—a phenomenon resulting from the in tersections of societal norms and women’s limited agency to advocate for their own needs (Gittelsohn et al., 1997). 3.4. Reduced form latent structural equation model by household wealth Paths from assets to intrinsic and to instrumental agency were higher in magnitude with greater household wealth, particularly from middle to high wealth category (Fig. 4). In examining associations between domains of women’s empowerment and CNS, we noted that domains of women’s empowerment operated differently depending on wealth categories. Assets were directly associated with anemia in the lowest wealth category (0.05 [0.008, 0.09]) and with WHZ in the middle wealth category (0.05 [0.06, 0.15]). Intrinsic agency had a direct positive as sociation with child HAZ in the low wealth group (0.03 [0.01, 0.05]), with anemia in the middle wealth group (0.03 [0.01, 0.05]), and with WHZ in the high wealth group (0.05 [0.03, 0.18]). Instrumental agency was associated only with anemia status in the lowest household wealth group (0.10 [0.08, 0.13]). In the low wealth group, intrinsic agency were associated with maternal BMI (0.11 [0.08, 0.14]). BMI was subsequently associated with HAZ (0.04 (0.02, 0.06]) and WHZ (0.30 [0.27, 0.33]). In the middle wealth group, intrinsic agency (0.13 [0.10, 0.15]) and instrumental agency (0.04 [0.02, 0.06]) were directly associated with maternal BMI. In the high wealth group, assets (0.40 [0.37, 0.42] and intrinsic agency (0.24 [0.21, 0.26]) were both directly associated with maternal BMI. Indirect effects differed by household wealth, with the majority of effects coming through intrinsic agency. The path from attitudes through BMI to WHZ was significant in the low wealth group (0.03 [0.01, 0.06]). This same path stayed consistent in the middle (0.007 [0.004, 0.01]) and high wealth groups (0.07 [0.04, 0.09]). An indirect path from intrinsic agency through BMI to HAZ was significant in the middle wealth group (0.003 [0.002, 0.005]). Goodness-of-fit statistics across models exceeded the threshold. Our work noted the importance of intrinsic agency for child nutri tion. This domain comprised items related to attitudes about IPV and may reflect women’s views about the right to bodily integrity (Miedema et al., 2018; Yount, VanderEnde, Dodell, & Cheong, 2016; Yount et al. under review). Hanmer and Klugman highlight attitudes about IPV as a critical domain of women’s sense of empowerment, claiming that beliefs in regressive gender norms, such as IPV, minimize women’s individual agency and ability to act in their own best interests. In Bangladesh, toddlers of women who reported that IPV was justified had significantly higher odds of stunting than women who rejected these attitudes. 3.4. Reduced form latent structural equation model by household wealth Beyond intrinsic agency, exposure to IPV negatively impacts child growth. In their review of the impacts of IPV on child nutrition, Yount, DiGirolamo, and Ramakrishnan (2011) hypothesized that IPV directly affects young children via a dysregulated stress response and indirectly via impacts on women’s physical, mental, and emotional well-being. i 3.4. Mediation model with maternal BMI Biologic mechanisms whereby the adequate nutrition of women before, during, and after pregnancy optimizes biologic transfer of nutrition to the child in utero and early infancy contribute to improved nutritional status of the child by establishing the foundations for adequate child growth trajectories (Shrimpton et al., 2001; Victora, De Onis, Hallal, Blossner, & Shrimpton, 2010; Young et al., 2017). Additional follow-on achievements of empowerment for women’s nutrition could indirectly improve child nutrition through pathways related to enhanced caring capabilities, women’s self-care (Pratley, 2016), and mental health (James-Hawkins et al., 2018b; Qutteina et al., 2017; K. M. Yount, S. Dijkerman, S. Zureick-Brown, & K. E. VanderEnde, 2014). 3.4. Mediation model with maternal BMI When accounting for mediation by maternal BMI, assets were no longer directly associated with any child nutritional outcomes (Fig. 3). Instrumental agency was the only direct association which stayed sig nificant after incorporation of maternal BMI (0.08 [0.05, 0.10]). All three domains of women’s empowerment were significantly associated with maternal BMI, with attitudes being the strongest estimate (assets: 0.17 [0.14, 0.20]; intrinsic agency: 0.23 [0.20, 0.26]; instrumental agency: 0.03 [0.01, 0.05]). 5 Women’s BMI was positively and directly associated with WHZ (0.35 [0.26, 0.43]) and child HAZ (0.08 [0.05, 0.10]). The indirect path from assets to intrinsic agency to maternal BMI was significant (0.06 [0.03, 0.080]). The extension of the indirect path from intrinsic agency through women’s BMI to WHZ was significant (0.08 [0.04, 0.11]). 3.4. Reduced form latent structural equation model by household wealth anemia. Associations between em outcomes differed by wealth categ not anemia, the pathways from t were mediated through their effec Prior research documented si age at pivotal life events (the dom partner violence (IPV) (the domai Fig. 2. ized pa ment (N ¼ 13 Note. a Mode 0.993; b Signif interval c All n simplic d Anem e Abbr WHZ, freedom lewis in approxi R. Jones et al. SSM - Population Health 9 (2019) 100475 R. Jones et al. Fig. 2. Diagram of direct and indirect standard ized path coefficients from Women’s empower ment to Children’s nutritional status (N ¼ 13,780). Note. a Model fit statistics: df ¼ 78; CFI ¼ 0.995; TLI: 0.993; RMSEA: 0.004 (0.001,0.006). b Significant paths based upon 95% confidence intervals. c All nonsignificant paths were dropped for simplicity. d Anemia model excludes Kenya (n ¼ 4918). e Abbreviations: HAZ, height-for-age z-score; WHZ, weight-for-age z-score. df, degrees of freedom; CFI, comparative fit index; TLI, tucker lewis index; RMSEA, root-mean square error of approximation. Fig. 2. Diagram of direct and indirect standard ized path coefficients from Women’s empower ment to Children’s nutritional status (N ¼ 13,780). Note. i Fig. 2. Diagram of direct and indirect standard ized path coefficients from Women’s empower ment to Children’s nutritional status (N ¼ 13,780). Note. i Fig. 2. Diagram of direct and indirect standard ized path coefficients from Women’s empower ment to Children’s nutritional status (N ¼ 13,780). Note. i a Model fit statistics: df ¼ 78; CFI ¼ 0.995; TLI: 0.993; RMSEA: 0.004 (0.001,0.006). Table 3 i Anemia in children represents a serious public health problem globally, with implications for infant development and child school performance (Allali, Brousse, Sacri, Chalumeau, & de Montalembert, 2017; Harika et al., 2017). In many countries, the prevalence of anemia exceeds the prevalence of stunting, and efforts to reduce anemia burden have been limited (Schümann & Solomons, 2017). Although studies have noted that schooling of women reduces anemia in their offspring (Harding, Aguayo, Masters, & Webb, 2018), ours is the first study, to our knowledge, to include anemia in the women’s empowerment–child nutrition nexus. We noted significant and direct associations between women’s instrumental agency in household decisions and child hemo globin concentrations. Intrinsic agency (attitudes about IPV) was not directly associated with anemia. These findings contrast research in India that found that the odds of child anemia were significantly higher in households in which women experienced IPV (Ackerson & Sub ramanian, 2008). Differences in our findings may result from contextual differences or from differences in measurement. Although the study in India evaluated IPV experiences, our measure utilized items related to attitudes about rather personal experiences of IPV in terms of questions around the multiple scenarios where wife-beating is considered justi fied. Secondly, unlike other measures of nutritional status, associations between women’s empowerment and anemia were not mediated through maternal nutrition. This finding suggests that effects may operate through pathways other than the biologic transfer of nutrition or the effects of a shared nutritional environment. Like stunting, the eti ology of anemia is complex, with nutrition representing only one of myriad causal pathways (Ngure et al., 2014; Schümann & Solomons, 2017; Stoltzfus, 2003); a more significant pathway in some contexts involves malaria and other parasitic infections. Thus, instrumental agency may be a more relevant domain for anemia such that women have agency to seek and apply preventive and curative measures for their children. A review of women’s empowerment and child nutritional status has documented a positive association between direct and indirect measures of women’s empowerment and child anthropometry (Carlson et al., 2015), although the authors note that their findings are less consistent in sub-Saharan Africa compared to other parts of the globe. Scholars argue that the contextualized nature of empowerment limits cross-context comparisons and may contribute to inconsistent findings; heterogene ity in how researchers operationalize empowerment also contributes to inconsistencies. Table 3 Table 3 Nonstandardized Path Coefficients for Direct, Indirect (where applicable), and Total Effects From Women’s Empowerment to Children’s Nutritional Status (N ¼ 13,780). domains were positively associated with each other (Voronca et al., 2018). Our study documents effect modification of the empowerment-nutrition relationship by household wealth. Although a Nonstandardized Path Coefficients for Direct, Indirect (where applicable), and Total Effects From Women’s Empowerment to Children’s Nutritional Status (N ¼ 13,780). Path and Dependent Outcomes Latent Predictors Direct Effect (β) Indirect Effect (β) Total Effect Path 1: Intrinsic Agency Assets 0.22 (0.20, 0.25) Path 1: HAZ Intrinsic Agency 0.05 (0.01, 0.08) Through intrinsic agency Assets 0.01 (0.003, 0.02) 0.06 (0.01,0.07) Path 1: WHZ Intrinsic Agency 0.09 (0.07, 0.12) Through intrinsic agency Assets 0.02 (0.015,0.03) 0.11 (0.04,0.13) Path 1: Anemia Intrinsic Agency 0.01 (0.02, 0.04) Through intrinsic agency Assets 0.002 (0.005, 0.009) 0.01 (0.02,0.05) Path 2: HAZ Assets 0.04 (0.01, 0.07) Path 2: WHZ Assets 0.08 (0.06, 0.11) Path 2: Anemia Assets 0.06 (0.03, 0.08) Path 3: Instrumental Agency Assets 0.14 (0.12, 0.16) Path 3: HAZ Instrumental Agency 0.02 (0.06, 0.01) Through instrumental agency Assets 0.003 (0.008, 0.001) 0.02 (0.05,0.01) Path 3: WHZ Instrumental agency 0.01 (0.03, 0.02) Through instrumental agency Assets 0.000 (0.004, 0.003) 0.003 (0.03,0.03) Path 3: Anemia Instrumental agency 0.09 (0.06, 0.13) Through instrumental agency Assets 0.01 (0.007, 0.02) 0.10 (0.05,0.14) Note. a Model fit statistics: df ¼ 78; CFI ¼ 0.995; TLI: 0.993; RMSEA: 0.004 (0.001,0.006). b 95% confidence intervals based on bootstrapping methods. d Anemia model excludes Kenya (n ¼ 4918). e Abbreviations: HAZ, height-for-age z-score; WHZ, weight-for-age z-score. df, degrees of freedom; CFI, comparative fit index; TLI, tucker lewis index; RMSEA, root-mean square error of approximation. We recognize that the relationship between wealth and empower ment is not straightforward. Research in Bangladesh, for example, notes that women’s engagement in decision making is lowest in the highest wealth groups potentially as a result of the male breadwinner having a more dominant role or the sole role in generating income (Mahmud, Shah, & Becker, 2012). Strategies that aim to empower women but are not contextualized to a community’s or household’s socioeconomic conditions and gender norms may inadvertently introduce inequity such that certain groups benefit while others are harmed. Thus, the efforts of government and others to mitigate poverty and economic disparity support an enabling environment in which women and households may benefit from interventions to increase agency. 4. Discussion Using a three-domain construct of women’s empowerment grounded in Kabeer’s assets, agency, and achievements framework, we observed that empowerment domains were associated differently with child nutrition in East Africa. Two general patterns emerged from the find ings. First, the latent domain for human and social assets, represented as age at pivotal life events, and that for intrinsic agency, represented as attitudes to intimate partner violence, appeared more relevant for anthropometric status, while women’s instrumental agency, measured as participation in household decision making was more relevant for Researchers in Kenya observed significant increases in household wealth and women’s empowerment over time, including engagement in decision making and disapproval of IPV, and that wealth and these 6 SSM - Population Health 9 (2019) 100475 R. Jones et al. R. Jones et al. household. If the household or community does not have financial or infrastructure resources, then the ability to act on these decisions is limited. Table 3 Our work, focused on five countries in eastern sub-Saharan Africa, makes several novel contributions to this body of evidence. Other studies using DHS data operationalized women’s empowerment in myriad ways, including the simple summation of re sponses, principle components analysis, the development of ad hoc cut-points, or use of single-item indicators as proxies (Pratley, 2016; Richardson, 2017; Yount et al., 2018). Although many of these variables are theoretically informed, they rarely are validated across contexts using systematic approaches. We used a theoretically informed approach and cross-contextually validated the invariance of our women’s empowerment measure. The use of a measure that is e Abbreviations: HAZ, height-for-age z-score; WHZ, weight-for-age z-score. df, degrees of freedom; CFI, comparative fit index; TLI, tucker lewis index; RMSEA, root-mean square error of approximation. domains were positively associated with each other (Voronca et al., 2018). Our study documents effect modification of the empowerment-nutrition relationship by household wealth. Although a clear pattern failed to emerge, it is plausible that a certain level of re sources is needed for women to act on their agency. For example, women may be empowered to participate in the decisions of or solely decide on how income in the household is to be used, when and how to seek care for themselves or children, or which foods to purchase for the 7 SSM - Population Health 9 (2019) 100475 psychometrically comparable across contexts allows for greater confi dence in both cross-contextual comparisons and in the measures of as sociation between women’s empowerment and child nutrition (Miedema et al., 2018). Lastly, the application of path analysis allowed our study. Our approach was limited by those questions available in the DHS. Other measures not captured in the DHS may be salient domains of women’s empowerment for child nutrition (Moghadam & Senftova, 2005) including, for example, women’s time-use (Johnston, Stevano,i Fig. 3. Diagram of direct and indirect standard ized path coefficients from Women’s empower ment to Children’s nutritional status mediated by Women’s BMI (N ¼ 13,780). Note. a Model fit statistics: df ¼ 88; CFI ¼ 0.994; TLI: 0.991; RMSEA: 0.004 (0.002,0.006). b Significant paths based upon 95% confidence intervals. c All nonsignificant paths were dropped for simplicity. d Anemia model excludes Kenya (n ¼ 4918). e Abbreviations: BMI, body mass index; HAZ, height-for-age z-score; WHZ, weight-for-age z- score. Table 3 d Anemia model excludes Kenya (n ¼ 4918). e Abbreviations: BMI, body mass index; HAZ, height-for-age z-score; WHZ, weight-for-age z- score. df, degrees of freedom; CFI, comparative fit index; TLI, tucker lewis index; RMSEA, root- mean square error of approximation. , y height-for-age z-score; WHZ, weigh score. df, degrees of freedom; CFI, c fit index; TLI, tucker lewis index; RM mean square error of approximation Fig. 4. Diagram of direct and indire ized path coefficients from Women’ ment to Children’s nutritional status m Women’s BMI, by wealth groups (N A. Low. B. Middle. C. High. Note. a Model fit statistics: df ¼ 288; CFI ¼ 0.988; RMSEA: 0.008 (0.006,0.009). b Significant paths based upon 95% intervals. c All nonsignificant paths were d Fig. 4. Diagram of direct and indirect standard ized path coefficients from Women’s empower ment to Children’s nutritional status mediated by Women’s BMI, by wealth groups (N ¼ 13,780). A. Low. b Significant paths based upon 95% confidence intervals. i c All nonsignificant paths were dropped for simplicity. c d e h m simplicity. c Anemia mo d Wealth cate Health Survey e Abbreviatio height-for-age score. df, deg fit index; TLI mean square c Anemia model excludes Kenya (n ¼ 4918). d d Wealth categories based on Demographic and Health Survey household wealth index. e Abbreviations: BMI, body mass index; HAZ, height-for-age z-score; WHZ, weight-for-age z- score. df, degrees of freedom; CFI, comparative fit index; TLI, tucker lewis index; RMSEA, root- mean square error of approximation. psychometrically comparable across contexts allows for greater confi dence in both cross-contextual comparisons and in the measures of as sociation between women’s empowerment and child nutrition (Miedema et al., 2018). Lastly, the application of path analysis allowed for the additional examination of maternal BMI as a key mediator that lies on the pathway between women’s empowerment and child nutrition. psychometrically comparable across contexts allows for greater confi dence in both cross-contextual comparisons and in the measures of as sociation between women’s empowerment and child nutrition (Miedema et al., 2018). Lastly, the application of path analysis allowed for the additional examination of maternal BMI as a key mediator that lies on the pathway between women’s empowerment and child nutrition. our study. Our approach was limited by those questions available in the DHS. Table 3 df, degrees of freedom; CFI, comparative fit index; TLI, tucker lewis index; RMSEA, root- mean square error of approximation. Fig. 4. Diagram of direct and indirect standard ized path coefficients from Women’s empower ment to Children’s nutritional status mediated by Women’s BMI, by wealth groups (N ¼ 13,780). A. Low. B. Middle. C. High. Note. a Model fit statistics: df ¼ 288; CFI ¼ 0.990; TLI: 0.988; RMSEA: 0.008 (0.006,0.009). b Significant paths based upon 95% confidence intervals. c All nonsignificant paths were dropped for simplicity. c Anemia model excludes Kenya (n ¼ 4918). d Wealth categories based on Demographic and Health Survey household wealth index. e Abbreviations: BMI, body mass index; HAZ, height-for-age z-score; WHZ, weight-for-age z- score. df, degrees of freedom; CFI, comparative fit index; TLI, tucker lewis index; RMSEA, root- mean square error of approximation. R. Jones et al. SSM - Population Health 9 (2019) 100475 Fig. 3. Diagram of direct and indirect standard ized path coefficients from Women’s empower ment to Children’s nutritional status mediated by Women’s BMI (N ¼ 13,780). Note. a Model fit statistics: df ¼ 88; CFI ¼ 0.994; TLI: 0.991; RMSEA: 0.004 (0.002,0.006). b Significant paths based upon 95% confidence intervals. c All nonsignificant paths were dropped for simplicity. d Anemia model excludes Kenya (n ¼ 4918). e Abbreviations: BMI, body mass index; HAZ, height-for-age z-score; WHZ, weight-for-age z- score. df, degrees of freedom; CFI, comparative fit index; TLI, tucker lewis index; RMSEA, root- mean square error of approximation. Fig. 4. Diagram of direct and indirect standard ized path coefficients from Women’s empower ment to Children’s nutritional status mediated by Women’s BMI, by wealth groups (N ¼ 13,780). A. Low. B. Middle. C. High. Note. a Model fit statistics: df ¼ 288; CFI ¼ 0.990; TLI: 0.988; RMSEA: 0.008 (0.006,0.009). b Significant paths based upon 95% confidence intervals. c All nonsignificant paths were dropped for i li i R. Jones et al. SSM - Population Health 9 (2019) 100475 R. Jones et al. R. Jones et al. Fig. 3. Diagram of direct and indirect standard ized path coefficients from Women’s empower ment to Children’s nutritional status mediated by Women’s BMI (N ¼ 13,780). Note. i Fig. 3. Diagram of direct and indirect standard ized path coefficients from Women’s empower ment to Children’s nutritional status mediated by Women’s BMI (N ¼ 13,780). Note. i c All nonsignificant paths were dropped for simplicity. Table 3 Other measures not captured in the DHS may be salient domains of women’s empowerment for child nutrition (Moghadam & Senftova, 2005) including, for example, women’s time-use (Johnston, Stevano, Malapit, Hull, & Kadiyala, 2015; Njuki et al., 2016), self-efficacy, and social capital (Crandall, Rahim, & Yount, 2015; Ickes et al., 2017). Similarly, nutritional outcomes such as BMI and anemia have complex etiologies, including dietary patterns, disease exposure, and access to Despite these novel contributions, several limitations are evident in 8 SSM - Population Health 9 (2019) 100475 R. Jones et al. The authors declare they have no conflicts of interest. The authors declare they have no conflicts of interest. ( ) p g p // p g Dolan, C. S. (2001). The ’good wife’: Struggle over resources in the Kenyan horticultural sector. Journal of Development Studies, 37(3), 39–70. f p Engle, P. L., Menon, P., & Haddad, L. J. (1999). Care and nutrition: Concepts and measurement. World Development, 27(8), 1309–1337. Supplementary data to this article can be found online at https://doi. org/10.1016/j.ssmph.2019.100475. Supplementary data to this article can be found online at https://doi. org/10.1016/j.ssmph.2019.100475. References Abada, T., & Tenkorang, E. Y. (2012). Women’s autonomy and unintended pregnancies in the Philippines. Journal of Biosocial Science, 44(6), 703–718. https://doi.org/ 10.1017/s0021932012000120. Ackerson, L. K., & Subramanian, S. V. (2008). Domestic violence and chronic malnutrition among women and children in India. American Journal of Epidemiology, 167(10), 1188–1196. Alderman, H., Haddad, L., Headey, D. D., & Smith, L. (2014). Association between economic growth and early childhood nutrition. Lancet Glob Health, 2(9), e500. https://doi.org/10.1016/S2214-109X(14)70266-9. Alderman, H., & Headey, D. (2018). The timing of growth faltering has important implications for observational analyses of the underlying determinants of nutrition outcomes. PLoS One, 13(4), e0195904. https://doi.org/10.1371/journal. pone.0195904. Allali, S., Brousse, V., Sacri, A. S., Chalumeau, M., & de Montalembert, M. (2017). Anemia in children: Prevalence, causes, diagnostic work-up, and long-term consequences. Expert Review of Hematology, 10(11), 1023–1028. https://doi.org/ 10.1080/17474086.2017.1354696. 5. Conclusion Using structural equation modeling, we observed that a previously developed, three-domain latent construct of women’s empowerment grounded in Kabeer’s assets, agency, and achievements framework functioned through maternal BMI to influence child growth indicators but not anemia. The paths of influence from empowerment to child nutrition indicators differed by household wealth status. These findings suggest that to achieve impacts on child nutrition, designers of empowerment strategies should consider a strategy’s potential to affect change in women’s nutrition and be cognizant of the resource con straints in the household or community. Amugsi, D. A., Lartey, A., Kimani, E., & Mberu, B. U. (2016). Women’s participation in household decision-making and higher dietary diversity: Findings from nationally representative data from Ghana. Journal of Health, Population and Nutrition, 35(1), 16. https://doi.org/10.1186/s41043-016-0053-1. 16. https://doi.org/10.1186/s41043-016-0053-1. Balk, D. (1994). Individual and community aspects of women’s status and fertility in rural Bangladesh. Population Studies, 48(March), 21–45. i Bandalos, D. L., & Finney, S. J. (2010). Factor Analysis: Exploratory and confirmatory. In G. R. Hancock, & R. O. Mueller (Eds.), The reviewer’s GUide to quantitative methods in the social sciences (pp. 93–114). New York: Taylor & Francis. Bhutta, Z. A., Das, J. K., Rizvi, A., Gaffey, M. F., Walker, N., Horton, S., et al. (2013). Evidence-based interventions for improvement of maternal and child nutrition: What can be done and at what cost? Lancet, 382(9890), 452–477. https://doi.org/ 10.1016/s0140-6736(13)60996-4. i Black, R., Allen, L., Bhutta, Z., Caulfield, L., De Onis, M., Ezzati, M., et al. (2008). Maternal and child undernutrition: Global and regional exposures and health outcomes. Lancet, 371(9608), 243–260. Contributions Rebecca Jones, Usha Ramakrishnan, and Amy Webb Girard conceptualized the research question. Rebecca Jones, Regine Haard€or fer, Amy Webb Girard, Usha Ramakrishnan and Kathryn Yount conceptualized the analysis plan; Regine Haard€orfer and Kathryn Yount guided analysis. Stephanie Miedema constructed the empowerment measure. All authors contributed to interpretation of results. Rebecca Jones and Amy Webb Girard wrote the first and subsequent drafts of the article. All authors contributed to critically revising the article and gave final approval of the version to be published. Black, R., Victora, C., Walker, S., Bhutta, Z., Christian, P., De Onis, M., et al. (2013). Maternal and child undernutrition and overweight in low-income and middle- income countries. Lancet, 382, 427–451. https://doi.org/10.1016/S0140-6735(13) 60937-X. Bloom, S. S., Wypij, D., & Das Gupta, M. (2001). Dimensions of women’s autonomy and the influence on maternal health care utilization in a north Indian city. Demography, 38(1), 67–78. van den Bold, M., Quisumbing, A., & Gillespie, S. R. (2013). Women’s empowerment and nutrition: An evidence review. Retrieved from Washington, D.C. Retrieved from http s://pdfs.semanticscholar.org/e43b/1432ddc7f87327fc0d086893da802df7a61a.pdf. Bollen, K. A., & Stine, R. (1990). Direct and indirect effects: Classical and bootstrap estimates of variability. Sociological Methodology, 20, 115–140. Bose, S. (2011). The effect of women’s status and community on the gender differential in children’s nutrition in India. Journal of Biosocial Science, 43(5), 513–533. https:// doi.org/10.1017/s002193201100006x. Statement of ethics Carlson, G. J., Kordas, K., & Murray-Kolb, L. E. (2015). Associations between women’s autonomy and child nutritional status: A review of the literature. Maternal and Child Nutrition, 11, 452–582. i This secondary analysis was conducted using Demographic and Health Survey data, for which the protocol was reviewed and approved by the ethics committee for each individual country. Written informed consent was obtained from all study participants before interview. The analysis presented here used deidentified secondary data, and additional human subjects review was not required. Crandall, A., Rahim, H. A., & Yount, K. M. (2015). Validation of the general self-efficacy scale among Qatari young women. Eastern Mediterranean Health Journal, 21(12), 891. De Onis, M., Onyango, A., Borghi, E., Siyam, A., Nishida, C., & Siekmann, J. (2007). Development of a WHO growth reference for school-aged children and adolescents. Bulletin of the World Health Organization, 85(9), 660–667. l DeRose, L. F., & Ezeh, A. C. (2005). Men’s influence on the onset and progress of fertility decline in Ghana, 1988-98. Population Studies, 59(2), 197–210. https://doi.org/ 10.1080/00324720500099496. Declaration of competing interest DHS. (2016). DHS program. Retrieved from http://dhsprogram.com. l ( ) h d if l i h DHS. (2016). DHS program. Retrieved from http://dhsprogram.com. Dolan, C. S. (2001). The ’good wife’: Struggle over resources in the Kenyan horticultural sector. Journal of Development Studies, 37(3), 39–70. Appendix A. Supplementary data general and reproductive health services. We attempted to include other indicators as mediators in the path analysis related to child diets (diet diversity, minimum acceptable diets); however, the categorical nature of these data and limited variation in responses prevented convergence and reduced model fit. Further, cross-sectional survey data limits our ability to test how resources for empowerment may influence women’s agency and then nutritional status (Kishor & Subaiya, 2008). The use of cross-sectional data fails to capture the inherent nature of empowerment as a process. Finally, the four-year range across surveys may result in bias due to unforeseeable events that occurred across settings during the duration of data collection. Because of data limitations, we also were unable to include Burundi, another East African country with DHS data. Although we focused specifically on East Africa, we recommend repli cation of this analysis in other regions and cross-regionally to assess direct and indirect effects of women’s empowerment on individual health across multiple contexts. Acknowledgements The effect of women’s decision-making power on maternal health services uptake: Evidence from Pakistan. Health Policy and Planning, 28(2), 176–184. https://doi.org/10.1093/heapol/czs042. i Ngure, F. M., Reid, B. M., Humphrey, J. H., Mbuya, M. N., Pelto, G., & Stoltzfus, R. J. (2014). Water, sanitation, and hygiene (WASH), environmental enterpathy, nutrition, and early child development: Making the links. Annals of the New York Academy of Sciences, 1308, 118–128. https://doi.org/10.1111/nyas.12330. j ki i h hi i ( ) i ki llh ld ICF International. (2012a). Measure DHS biomarker field manual. Retrieved from Calverton, Maryland. Retrieved from https://dhsprogram.com/pubs/pdf/DHSM7/D HS6_Biomarker_Manual_9Jan2012.pdf. Njuki, J. M., Kaaria, S., Chamunorwa, A., & Chiuri, W. (2011). Linking smallholder farmers to markets, gender, and intra-household dynamics: Does the choice of commodity matter. The European Journal of Development Research, 23(3), 426–443. Nj ki J M W A B l k I Y K M N ll C R k i h U l p ICF International. (2012b). Survey organization manual for demographic and health surveys. Retrieved from Calverton, Maryland. Retrieved from https://dhsprogram. com/pubs/pdf/DHSM10/DHS6_Survey_Org_Manual_7Dec2012_DHSM10.pdf. i Njuki, J. M., Wyatt, A., Baltenweck, I., Yount, K. M., Null, C., Ramakrishnan, U., et al. (2016). An exploratory study of dairying intensification, women’s decision making, and time use and implications for child nutrition in Kenya. European Journal of Development Research, 28(4), 722–740. Ickes, S. B., Heymsfield, G. A., Wright, T. W., & Baguma, C. (2017). “Generally the young mom suffers much:” Socio-cultural influences of maternal capabilities and nutrition care in Uganda. Maternal and Child Nutrition, 13(3). https://doi.org/10.1111/ mcn.12365. Nussbaum, M. C. (2001). Women and human development: The capabilities approach. Cambridge, United Kingdom: Cambridge University Press. International Food Policy Research Institute. (2016). Global nutrition report 2016: From promise to impact: Ending malnutrition by 2030. Washington, DC. Retrieved from http://www.ifpri.org/publication/global-nutrition-report-2016-promise-impact-en ding-malnutrition-2030. Patel, V., Kirkwood, B. R., Pednekar, S., Pereira, B., Barros, P., Fernandes, J., et al. (2006). Gender disadvantage and reproductive health risk factors for common mental disorders in women: A community survey in India. Archives of General Psychiatry, 63(4), 404–413. James-Hawkins, L., Peters, C., VandeEnde, K. E., Bardin, L., & Yount, K. M. (2018a). Women’s agency and its relationship to current contraceptive use in lower- and middle-income countries: A systematic review of the literature. Global Public Health, 13(7), 843–858. https://doi.org/10.1080/17441692.2016.1239270. Pratley, P. (2016). Acknowledgements Field, E., Jayachandran, S., & Pande, R. (2010). Do traditional institutions constrain female entrepeneurship? A field experiment on business training in India. The American Economic Review, 100(2), 125–129. i The authors thank the Competitive Research Grants to Develop Innovative Methods and Metrics for Agriculture and Nutrition Actions (IMMANA Grants) and UK aid from the UK government for supporting this work. We gratefully acknowledge three anonymous reviewers, who provided valuable feedback during the preparation of this manuscript. Finally, we thank the study participants and the DHS staff. Galie, A., Teufil, N., Webb Girard, A., Baltenweck, I., Dominguez-Salas, P., Jones, R., et al. (2019). Women’s empowerment, food security and forage in pastoral communities of Tanzania. Global Food Security, 23, 125–134. https://doi.org/ 10.1016/j.gfs.2019.04.005. Gittelsohn, J., Thapa, M., & Landman, L. T. (1997). Cultural factors, caloric intake and micronutrient sufficiency in rural. Nepali Households, 44(11), 1739–1749, 1997. i Habibov, N., Barrett, B. J., & Chernyak, E. (2017). Understanding women’s empowerment and its determinants in post-communist countries: Results of 9 R. Jones et al. SSM - Population Health 9 (2019) 100475 Mahmud, S., Shah, N. M., & Becker, S. (2012). Measurement of women’s empowerment in rural Bangladesh. World Development, 40(3), 610–619. https://doi.org/10.1016/j. worlddev.2011.08.003. Azerbaijan national survey. Women’s Studies International Forum, 62, 125–135. https://doi.org/10.1016/j.wsif.2017.05.002. i Hadley, C., Patil, C., & Nahayo, D. (2010). Difficulty in the food environment and the experience of food insecurity among refugees resettled in the United States. Ecology of Food and Nutrition, 49, 390–407. Malapit, H. J., Kadiyala, S., Quisumbing, A., Cunningham, K., & Tyagi, P. (2015). Women’s empowerment mitigates the negative effects of low production diversity on maternal and child nutrition in Nepal. Journal of Development Studies, 51(8). https:// doi.org/10.1080/00220388.2014.1018904. Hambidge, K. M., Westcott, J. E., Garces, A., Figueroa, L., Goudar, S. S., Dhaded, S. M., et al. (2019). A multicountry randomized controlled trial of comprehensive maternal nutrition supplementation initiated before conception: The women first trial. American Journal of Clinical Nutrition, 109(2), 457–469. https://doi.org/10.1093/ ajcn/nqy228. l l d Mason, J. B., Shrimpton, R., Saldanha, L. S., Ramakrishnan, U., Victora, C. G., Girard, A. W., et al. (2014). The first 500 days of life: Policies to support maternal nutrition. Global Health Action, 7, 23623. https://doi.org/10.3402/gha.v7.23623. b l k l f ( ) Hanmer, L., & Klugman, J. (2016). Exploring women’s agency and empowerment in developing countries: Where do we stand? Feminist Economics, 22(1), 237–263. https://doi.org/10.1080/13545701.2015.1091087. Matare, C. R., Mbuya, M. N. Acknowledgements N., Pelto, G. H., Dickin, K. L., & Stoltzfus, R. J. (2015). Assessing maternal capabilities in the SHINE trial: Highlighting a hidden link in the causal pathway to child health. Clinical Infectious Diseases, 61(S7), S745–S751. https://doi.org/10.1093/cid/civ851. Harding, K. L., Aguayo, V. M., Masters, W. A., & Webb, P. (2018). Education and micronutrient deficiencies: An ecological study exploring interactions between women’s schooling and children’s micronutrient status. BMC Public Health, 18. https://doi.org/10.1186/s12889-018-5312-1. Miedema, S. S., Haard€orfer, R., Girard, A. W., & Yount, K. M. (2018). Women’s empowerment in east Africa: Development of a cross-country comparable measure. World Development, 110, 453–464. https://doi.org/10.1016/j. worlddev.2018.05.031. https://doi.org/10.1186/s12889-018-5312-1. Harika, R., Faber, M., Samuel, F., Mulugeta, A., Kimiywe, J., & Eilander, A. (2017). Are low intakes and deficiencies in iron, vitamin A, zinc, and iodine of public health concern in Ethiopian, Kenyan, Nigerian, and South African children and adolescents? Food and Nutrition Bulletin, 38(3), 405–427. https://doi.org/10.1177/ 0379572117715818. Moghadam, V. M., & Senftova, L. (2005). Measuring women’s empowerment: Participation and rights in civil, political, social, economic, and cultural domains. International Social Science Journal, 57(2), 389. https://doi.org/10.1111/j.1468- 2451.2005.00557.x. Hatcher, L. (1994). A step-by-step approach to using the SAS system for factor Analysis and structural equation modeling (Cary, NC). Morgan, S. P., & Niraula, B. (1995). Gender inequality and fertility in two Nepali villages. Population and Development Review, 21, 541–561. Muthen, L. K., & Muthen, B. O. (2001). MPlus user’s guide. Retrieved from Los Angeles, CA. Retrieved from https://www.statmodel.com/download/usersguide/Mplus% 20user%20guide%20Ver_7_r3_web.pdf. Hodgson, D. L. (1999). Pastoralism, Patriarchy and History: Changing gender relations among Maasai in tanganyika, 1890-1940. The Journal of African History, 40, 41–65. Hodgson, D. L. (1999). Pastoralism, Patriarchy and History: Changing gender relations among Maasai in tanganyika, 1890-1940. The Journal of African History, 40, 41–65. Holtzmann, J. (2002). Politics and gastropolitics: Gender and the power of food in two african pastoralist societies. J Royal Anthro Instit, 8(2), 259–278. Holtzmann, J. (2002). Politics and gastropolitics: Gender and the power of food in two african pastoralist societies. J Royal Anthro Instit, 8(2), 259–278. Na, M., Jennings, L., Telegawker, S. A., & Ahmed, S. (2015). Association between women’s empowerment and infant and child feeding practices in sub-saharan Africa: An analysis of demographic and health surveys. Public Health Nutrition, 18(17), 3155–3165. Hooper, D., Coughlan, J., & Mullen, M. R. (2008). Structural equation modeling: Guidelines for determining model fit. Electronic Journal of Business Research Methods, 6(1), 53–60. Hou, X., & Ma, N. (2013). Acknowledgements Associations between quantitative measures of women’s empowerment and access to care and health status for mothers and their children: A systematic review of evidence from the developing world. Social Science & Medicine, 169, 119–131. https://doi.org/10.1016/j.socscimed.2016.08.001. 169, 119–131. https://doi.org/10.1016/j.socscimed.2016.08.001. James-Hawkins, L., Shaltout, E., Nur, A. A., Nasrallah, C., Qutteina, Y., Abdul Rahim, H. F., et al. (2018b). Human and economic resources for empowerment and pregnancy-related mental health in the arab Middle East: A systematic review. Archives of Women’s Mental Health. https://doi.org/10.1007/s00737-018-0843-0. Pullum, T. W., Collison, D. K., Namaste, S., & Garrett, D. (2017). Hemoglobin data in DHS surveys: Intrinsic variation and measurement error. Retrieved from Rockville, Maryland. Retrieved from https://dhsprogram.com/publications/publication-mr 18-methodological-reports.cfm. g p Qadir, F., Khan, M. M., Medhin, G., & Prince, M. (2011). Male gender preference, female gender disadvantage as risk factors for psychological morbidity in Pakistani women of childbearing age - a life course perspective. BMC Public Health, 11, 745. https:// doi.org/10.1186/1471-2458-11-745. Johnston, D., Stevano, S., Malapit, H. J. L., Hull, E., & Kadiyala, S. (2015). Agriculture, gendered time use, and nutritional outcomes: A systematic review. Retrieved from Washington, D.C http://ebrary.ifpri.org/cdm/ref/collection/p15738coll2/id/1 29389. Qutteina, Y., Nasrallah, C., James-Hawkins, L., Nur, A. A., Yount, K. M., Hennink, M., et al. (2017). Social resources and arab women’s perinatal mental health: A systematic review. Women and Birth. https://doi.org/10.1016/j. wombi.2017.11.005. Jones, R. E., Haard€orfer, R., Ramakrishnan, U., Yount, K. M., Miedema, S. S., Roach, T. D., et al. (2019). Domains of women’s empowerment are directly and positively associated with nutritional status of East African women. under review. l Kabeer, N. (1999). Resources, agency, achievements: Reflections on the measurement of women’s empowerment. Development and Change, 30(3), 435–464. https://doi.org/ 10.1111/1467-7660.00125. Rahman, M. (2012). Women’s autonomy and unintended pregnancy among currently pregnant women in Bangladesh. Maternal and Child Health Journal, 16(6), 1206–1214. https://doi.org/10.1007/s10995-011-0897-3. Kandiyoti, D. (1988). Bargaining with patriarchy. Gender & Society, 2(3), 274–290. https://doi.org/10.1177/089124388002003004. Ramakrishnan, U., Grant, F., Goldenberg, T., Zongrone, A., & Martorell, R. (2012). Effect of women’s nutrition before and during early pregnancy on maternal and infant outcomes: A systematic review. Paediatric & Perinatal Epidemiology, 26(S1), 285–301. Kishor, S., & Subaiya, L. (2008a). Understanding women’s empowerment: A comparative analysis of demographic and health surveys (DHS) data. Comparative Reports No. 20. Retrieved from Calverton, Maryland https://dhsprogram.com/publications/publ ication-cr20-comparative-reports.cfm. Ramakrishnan, U., Imhoff-Kunsch, B., & Martorell, R. (2014). Maternal nutrition interventions to improve maternal, newborn, and child health outcomes. Nestle Nutrition Institute Workshop Ser, 78, 71–80. https://doi.org/10.1159/000354942. Acknowledgements p p Kline, R. (1998). Principles and practices of structural equation modeling. New York: Retrieved from. Richardson, R. (2017). Measuring women’s empowerment: A critical review of current practices and recommendations for researchers. Social Indicators Research. https:// doi.org/10.1007/s11205-017-1622-4. Krishna, A., Oh, J., Lee, J. K., Lee, H. Y., Perkins, J. M., Heo, J., et al. (2015). Short-term and long-term associations between household wealth and physical growth: A cross- comparative analysis of children from four low- and middle-income countries. Global Health Action, 8, 26523. https://doi.org/10.3402/gha.v8.26523. Richards, E., Theobald, S., George, A., Kim, J. C., Rudert, C., Jehan, K., et al. (2013). Going beyond the surface: Gendered intra-household bargaining as a social determinant of child health and nutrition in low and middle income countries. Social Science & Medicine, 95, 24–33. Larsen, U., & Hollos, M. (2003). Women’s empowerment and fertility decline among the Pare of Kilimanjaaro region, Northern Tanzania. Social Science & Medicine, 57(6), 1099–1115. Ruel, M. T., & Alderman, H. (2013). Nutrition-sensitive interventions and programmes: How can they help to accelerate progress in improving maternal and child nutrition? Lancet, 382(9891), 536–551. https://doi.org/10.1016/s0140-6736(13)60843-0. Macro International Inc. (2006). Demographic and health survey interviewer’s manual. Retrieved from Calverton, MD. Retrieved from https://dhsprogram.com/pubs/pdf/ DHSM1/DHS7-Interviewer’s-Manual-EN-12Jun2017-DHSM1.pdf. Rutstein, S. O., & Rojas, G. (2006). Guide to DHS statistics: Demographic and health surveys methodology. Retrieved from Calverton, MA https://www.dhsprogram. com/publications/publication-dhsg1-dhs-questionnaires-and-manuals.cfm. 10 R. Jones et al. SSM - Population Health 9 (2019) 100475 Schümann, K., & Solomons, N. W. (2017). Perspective: What makes it so difficult to mitigate worldwide anemia prevalence? Advance Nutrition, 8(3), 401–408. https:// doi.org/10.3945/an.116.013847. WHO Multicentre Growth Reference Study Group. (2006). WHO Child Growth. Standards based on length/height, weight and age. Acta Paediatrica, 45(Suppl), 76–85. Woldemicael, G., & Tenkorang, E. Y. (2010). Women’s autonomy and maternal health- seeking behavior in Ethiopia. Maternal and Child Health Journal, 14(6), 988–998. https://doi.org/10.1007/s10995-009-0535-5. Shrimpton, R., & Rokx, C. (2012). The doulbe burden of malnutrition: A review of global evidence. Washington, DC: Paper presented at the The International Bank for Reconstruction and Development. Wu, G., Imhoff-Kunsch, B., & Girard, A. W. (2012). Biological mechanisms for nutritional regulation of maternal health and fetal development. Paediatric & Perinatal Epidemiology, 26(Suppl 1), 4–26. https://doi.org/10.1111/j.1365-3016.2012.01291. x. Suppl 1. Shrimpton, R., Victora, C. G., De Onis, M., Lima, R. C., Bl€ossner, M., & Clugston, G. (2001). Worldwide timing of growth faltering: Implications for nutritional interventions. Pediatrics, 107(5). i Shroff, M. R., Griffiths, P. Acknowledgements L., Suchindran, C., Nagalla, B., Vazir, S., & Bentley, M. E. (2011). Does maternal autonomy influence feeding practices and infant growth in rural India? Social Science & Medicine, 73(3), 447–455. https://doi.org/10.1016/j. socscimed.2011.05.040. Yimer, F., & Tadesse, F. (2015). Women’s empowerment in agriculture and dietary diversity in Ethiopia. Retrieved from International Food Policy Research Institute. Young, M. F., Hong Nguyen, P., Addo, O. Y., Pham, H., Nguyen, S., Martorell, R., et al. (2017). Timing of gestational weight gain on fetal growth and infant size at birth in vietnam. PLoS One, 12(1), e0170192. https://doi.org/10.1371/journal. pone.0170192. Smith, L. C., Ramakrishnan, U., Ndiaye, A., Haddad, L. J., Martorell, R., Ait-Kadi, M., et al. (2002). The importance of women’s status for children in developing countries. Washington, D.C: Retrieved from. i Stoltzfus, R. J. (2003). Iron deficiency: Global prevalence and consequences. Food and Nutrition Bulletin, 24(4 Suppl), S99–S103. Young, M. F., Nguyen, P. H., Addo, O. Y., Hao, W., Nguyen, H., Pham, H., et al. (2015). The relative influences of maternal nutritional status before and during pregnancy on birth outcomes in Vietnam. European Journal of Obstetrics & Gynecology and Reproductive Biology, 194, 223–227. pp UNICEF. (2015). Progress for children: Beyond averages - learning from the MDGs. Retrieved from New York, NY https://www.unicef.org/publications/index_82231.ht ml. Yount, K. M., DiGirolamo, A. M., & Ramakrishnan, U. (2011). Impacts of domestic violence on child growth and nutrition: A conceptual review of the pathways of influence. Social Science & Medicine, 72(9), 1534–1554. https://doi.org/10.1016/j. socscimed.2011.02.042. Upadhyay, U. D., & Hindin, M. J. (2005). Do higher status and more autonomous women have longer birth intervals? Results from cebu, Philippines. Social Science & Medicine, 60(11), 2641–2655. https://doi.org/10.1016/j.socscimed.2004.11.032. Yount, K. M., Dijkerman, S., Zureick-Brown, S., & VanderEnde, K. E. (2014). Women’s empowerment and generalized anxiety in Minya, Egypt. Social Science & Medicine, 106, 185–193. https://doi.org/10.1016/j.socscimed.2014.01.022. Victora, C. G., Adair, L., Fall, C., Hallal, P. C., Martorell, R., Richter, L., et al. (2008). Maternal and child undernutrition: Consequences to adult health and human capital. Lancet, 371, 340–357. Victora, C. G., De Onis, M., Hallal, P. C., Blossner, M., & Shrimpton, R. (2010). Worldwide timing of growth faltering: Revisiting implications for interventions. Pediatrics, 125, e473–e480. Yount, K. M., Krause, K. H., & VanderEnde, K. E. (2016a). Economic coercion and partner violence against wives in vietnam: A unified framework? Journal of Interpersonal Violence, 31(20), 3307–3331. https://doi.org/10.1177/0886260515584350. Yount, K. M., Peterman, A., & Cheong, Y. F. (2018). Acknowledgements Measuring women’s empowerment: A need for context and caution. The Lancet Global Health, 6(1), e29. Voronca, D., Walker, R. J., & Egede, L. E. (2018). Relationship between empowerment and wealth: Trends and predictors in Kenya between 2003 and 2008-2009. International Journal of Public Health, 63(5), 641–649. https://doi.org/10.1007/ s00038-017-1059-1. Yount, K. M., VanderEnde, K. E., Dodell, S., & Cheong, Y. F. (2016b). Measurement of women’s agency in Egypt: A national validation study. Social Indicators Research, 128 (3), 1171–1192. https://doi.org/10.1007/s11205-015-1074-7. 11 11
https://openalex.org/W2516033724	https://www.aclweb.org/anthology/W16-3105.pdf	English	null	HPI Question Answering System in BioASQ 2016	null	2,016	cc-by	4,110	2http://bioasq.org/ 1 Introduction The deluge of scientiﬁc publication in biomedicine requires tools for processing and searching precise information in real time. Question answering (QA) comes as an alternative to standard search engines system, e.g. PubMed1, and provides precise and short answers for questions in natural language (Athenikos and Han, 2010; Neves and Leser, 2015). One of the advantages of QA systems is that the user does not need to be proﬁcient in formulating queries in a way that the system can understand. Instead, a user may simply enter a question as they would pose it to another person and receive a answer in return. Thus, no formal training is required to use QA systems. The next section presents a short description of our the HPI2 system, followed by the preliminary results that we obtained in the challenge and a short discussion about our performance and meth- ods. QA is one of the more complex applications of natural language processing (NLP) (Jurafsky and Martin, 2013). This is usually achieved through a three-steps architecture: (1) the users question Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics, pages 38–44, Berlin, Germany, August 7-12, 2016. c⃝2016 Association for Computational Linguistics Abstract must be processed so that a query can be gener- ated; (2) this query is then used to ﬁnd all relevant text passages from a large document collection; and (3) ﬁnally, the system generates the exact an- swer to the users question and/or a summary of the facts from these passages. Some QA systems al- ready exist for the biomedical domain (Bauer and Berleant, 2012). However, none of them are ca- pable of answering questions in real time, in part due to the large collections of documents involved in the task. Question answering (QA) systems are cru- cial when searching for exact answers for natural language questions in the biomed- ical domain. Answers to many of such questions can be extracted from the 26 millions biomedical publications currently included in MEDLINE when relying on appropriate natural language processing (NLP) tools. In this work we describe our participation in the task 4b of the BioASQ challenge using two QA systems that we developed for biomedicine. Preliminary results show that our systems achieved ﬁrst and second positions in the snippet re- trieval sub-task and for the generation of ideal answers. We describe our participation in the fourth edi- tion of the BioASQ challenge2 (Tsatsaronis et al., 2015), a community-based shared task which aims to evaluate the current solutions for a variety of QA sub-tasks. We submitted runs from two QA systems which were speciﬁcally developed for the biomedical domain. One of the system (HPI1) successfully participated in the previous editions of the BioASQ challenge (Neves, 2015) and our second system (HPI2) is described in this work. We relied on existing NLP functionality from a in- memory database (IMDB) and we extend it with new procedures tailored speciﬁcally to QA. We participated in the task 4b (Biomedical Semantic QA) which is split in two phases: (a) phase A: concept mapping and document, passage and RDF triples retrieval; and (b) phase B: exact and ideal (short summary) answers. 1http://www.ncbi.nlm.nih.gov/pubmed 2.1 MEDLINE MEDLINE3 is the main source for biomedical publications and grows continuously. We down- loaded the publications from MEDLINE and inte- grated them into our local database. For the pur- poses of our QA system, an article consists of a title, an abstract and the main text. In this paper we refer only to titles and abstracts, as full pa- pers are not considered in the current edition of the BioASQ challenge. Figure 1: Work-ﬂow of our question answering system. 2 Data We relied on two main resources when developing our QA system: the MEDLINE and the Uniﬁed 38 Medical Language System (UMLS). In this sec- tion, we give a short overview on both resources. answer processing. The later includes a two-step phase: exact answer extraction (not included in this paper) and summarization. Details for each component are described below. 2.2 Uniﬁed Medical Language System Extracting meaning out of biomedical documents is usually supported by manually curated dic- tionaries. These dictionaries contain words and phrases which are common to the biomedical do- main. such dictionaries are used to map synonyms and abbreviations of terms to a common base term. Extracting meaning out of biomedical documents is usually supported by manually curated dic- tionaries. These dictionaries contain words and phrases which are common to the biomedical do- main. such dictionaries are used to map synonyms and abbreviations of terms to a common base term. Often, they also contain information to assign cat- egories to terms. There are various terminolo- gies for the biomedical domain, such as UMLS, SNOMED CT or MeSH. Figure 1: Work-ﬂow of our question answering system. Figure 1: Work-ﬂow of our question answering system. UMLS4 is a comprehensive database that com- bine various sources into a single knowledge base. It includes vocabularies mapping words and phrases onto a set of concepts. Each concept has an associated semantic type and group, which classiﬁes the category of the concept, such as gene or disease. 3.1 Architecture Our system was developed on top of a IMDB (SAP HANA database) (Plattner, 2013), which allows fast access of data directly from main memory, in contrast to processing data from ﬁles that re- side on disk space, thus requiring loading data into main memory. The IMDB we used comes with built-in text analysis features, such as language detection, sentence splitting, tokenization, stem- ming, part-of-speech (POS) tagging, NER based on pre-compiled dictionaries, information extrac- tion based on manually crafted rules, document indexing, approximate searching and sentiment analysis. In our QA system, UMLS was mainly used for named-entity recognition (NER), i.e., for extract- ing named-entities both in the question and in the document collection. Also in the context of NER, we used the UMLS semantic types to map the named-entities to their corresponding types. Fi- nally, we also rely on UMLS to resolve synonyms, thus avoiding to miss important passages which include synonyms to the words in the questions. Abbreviations, in particular, are very frequent in biomedical documents. All textual resources (documents and ques- tions) were added to the database and dictionar- ies of biomedical terms were created based on the UMLS terminology. Then we created the so- called full text index (FTI), i.e., an additional table which can be created for columns which contain text. Such an index can be created in many ways, we opted for two of them, namely: (a) a linguis- tic index, which contains all words from the origi- nal documents, as well as corresponding POS tags; and (b) a NER index, which contains all entities that were found based on the dictionary that was 3https://www.nlm.nih.gov/pubs/ factsheets/medline.html 4https://www.nlm.nih.gov/research/ umls/ 4https://www.nlm.nih.gov/research/ umls/ 3https://www.nlm.nih.gov/pubs/ factsheets/medline.html 3https://www.nlm.nih.gov/pubs/ 3.3 Document and Passage Retrieval The ﬁrst step in a question answering system is to analyze the input question. This step is composed of three components in our QA system: (a) ques- tion type detection, (b) target extraction, and (c) query building. The query that was built in the previous step was used in this step to ﬁnd relevant documents and passages within the millions abstracts. We relied on the tf-idf method (Manning et al., 2008) as a ba- sis and we adjusted it by various means to better ﬁt the biomedical domain. We opted for the weighted tf-idf approach since our experiments showed that it provided up to 10% more recall than an equally weighted approach. We used a proximity measure to boosts a documents relevancy rating when it contains words from the query which appear close together. This measure searches for each possible word pair that appears in the query and applies a ﬁxed rating increase for each such pair that is sep- arated by a maximum of two words anywhere in the document. Question Type Detection. The question type can be either ”yes/no”, ”factoid”, ”list” or ”sum- mary”. It deﬁnes which kind of the answer the system needs to return. In this step, we split the question into words and and apply special rules to ﬁnd the correct type, by considering question words and the structure (POS tags) of the ques- tion. Our approach is based on regular expres- sions, for instance, a questions beginning with an auxiliary verb is classiﬁed as yes/no-question. Al- though our QA system includes a component for detecting the question type, this step is not neces- sary in the BioASQ challenge because all question types are given. We also consider the documents title in our ap- proach. A documents titles relevancy was added to the documents relevancy in a weighted sum, thereby increasing the relevancy of documents with relevant titles. We also utilized a Jaccard- based word overlap measure between sentences in the document and in the question for the pas- sage retrieval step. Our system ﬁrst retrieves the 100.000 most relevant documents and then checks their sentences. This way we achieve a signiﬁcant speed-up compared to calculating relevancy scores for all sentences in all documents. The document’s total proximity score and the best sentence’s word overlap score are then used to boost the initial tf- idf score. 3.3 Document and Passage Retrieval Their inﬂuence was tuned empirically on a test set of BioASQ questions and answers. Finally, our document and passage retrieval al- gorithms return a list of documents or passages, sorted by their relevance score. Target Extraction. The second component of our question processing step extracts the target of the question, in case of factoid questions, and clas- siﬁes it according to the UMLS semantic types5, e.g, whether the question asks for a disease or a gene. This is an important information for the answer extraction step. We extract the headword using simple rules, for instance, the ﬁrst noun af- ter the question word (e.g., ”what”, ”which”). For classifying the headwords according to the many UMLS semantic types, and inspired by (Huang et al., 2008), we relied on a machine learning (ML) approach based on the implementation of the Support Vector Machine (SVM) algorithm in the IMDB database. The features that we use were the headwords and the questions words. All head- words in the factoid questions were manually clas- siﬁed into the semantic types by one of the authors (MN) and this is the training data that was used in our experiments. During the process, several different features were evaluated, but they did not improved our results. 3 Methods Our QA is composed of many components (cf. Figure 1) which are included in three main steps, i.e., question processing, document retrieval, and 39 previously built. In summary, from the linguistic FTI it is possible to retrieve information about sen- tence splitting, tokenization, stemming and POS tags, while the NER provides the named-entities. Query Building. Good query terms are impor- tant features when relying on a keyword-based search to ﬁnd relevant documents for the question. For this purpose, we use all words, except for stop- words and question words (e.g., ”what”, ”which”). Query Building. Good query terms are impor- tant features when relying on a keyword-based search to ﬁnd relevant documents for the question. For this purpose, we use all words, except for stop- words and question words (e.g., ”what”, ”which”). 5https://metamap.nlm.nih.gov/ SemanticTypesAndGroups.shtml 6http://www.bioasq.org/participate/ third-challenge-winners 3.4 Answer extraction We only submitted ideal answers, i.e. short sum- maries, for the BioASQ challenge. Our approach is described in details below. For the generation of summaries, we used an algorithm that is based on LexRank (Erkan and Radev, 2004), but that solely used the named- entities for the similarity function. In other words, 40 instead of using tf/idf values to rate the importance of each word, we use the named-entities instead. 4. Repeat the steps 2 and 3 until enough sen- tences are in set A. The ﬁrst step was to build a sentence graph. Therefore we calculated the cosine similarity of each sentence with each other sentences, i.e., a vector representation of each sentence. However, instead of using each word as dimension for the vector, we only use the named-entities. After the construction of the vectors, we calculate the co- sine similarity (cf. equation 1) between each two of these: In a ﬁnal step, we order the sentences from set A according to their occurrence in the original docu- ments. Thus, we tried to roughly keep the sentence at the position that the author intended. 4.1 BioASQ challenge We participated on the Task 4b, which is com- posed of two phases: A and B. During phase A, the participating teams received a test set of 100 ques- tions along with their question type, i.e., whether yes/no, factoid, list or summary, and had 24 hours to submit their predictions for concepts, docu- ments, passages and RDF triples. When phase A was over, the organizers released the the test set for phase B which contained the same questions previously released for phase A along with gold- standard annotations. During phase B, the partici- pating teams had 24 hours to submit their predic- tions for exact and ideal answers. where Ai and Bi are the dimensions of the vectors representing the sentences. Afterwards, we create the sentence graph by adding a ver- tex for each sentence. Then we create edges be- tween those vertices whose corresponding sen- tences have a similarity score above 0.2. For calculating the ranking, we used the exact round based formula (cf. equation 2) that is used in LexRank and that originates from PageRank (Page et al., 1999): score(si) = d N +(1−d) X sj∈adj[si] score(sj) deg(sj) (2) The BioASQ organizers released ﬁve bathes of around 100 questions every two weeks. Although our QA systems are capable to output results for most of the tasks covered in BioASQ, we did not submit runs for every sub-task due to problems with the systems, which are still under develop- ment. where N is the total number of vertices in the graph, adj[s] are all adjacent vertices of the vertex s. Additionally, we have the parameter d, a ’damp- ing factor’, which is typically set to 0.2 (Page et al., 1999). Subsequently, we ranked all sentences accord- ing to their centrality in the set of related abstracts. We need a last step to generate a summary by removing redundant sentences and we follow the following process: 4 Results and Discussion In this section, we present the preliminary results we obtained in the fourth edition of the BioASQ challenge. We introduce the details of the BioASQ challenge and then present our results for the two systems with which we participated this year. cosine = nP i=1 AiBi s nP i=1 A2 i s nP i=1 B2 i (1) (1) 7http://participants-area.bioasq. org/results/4b/phaseA/ and http:// participants-area.bioasq.org/results/ 4b/phaseB/ 4.2 Systems ”na” indicated that results are still not available for this task, while ”-” indicated that we did not sub- mit any run for the task. The values inside parame- ters indicate our current rank and the total number of submissions for the task. 4.2 Systems ”na” indicated that results are still not available for this task, while ”-” indicated that we did not sub- mit any run for the task. The values inside parame- ters indicate our current rank and the total number of submissions for the task. HPI1 HPI2 Concepts MAP MAP batch1 na - batch2 - - batch3 na - batch4 na - batch5 na - Documents MAP MAP batch1 0.0474 (12/15) 0.0028 (15/15) batch2 - - batch3 0.0674 (16/18) 0.0006 (18/18) batch4 - - batch5 0.434 (16/21) - Snippets MAP MAP batch1 0.0481 (1/7) - batch2 - - batch3 0.0715 (4/14) - batch4 - - batch5 0.0510 (5/16) - Ideal Answ. Rouge-2 Rouge-2 batch1 - 0.2231 (1/2) batch2 - 0.2240 (6/7) batch3 - 0.2559 (6/7) batch4 - 0.2280 (4/4) batch5 - 0.3233 (6/7) The document and passage retrieval of the HPI1 system did not make use of our local copy of MEDLINE but it queries PubMed instead. For each question, we generate two queries based on its tokens: (1) by using the ”OR” operator and words in the question, except stopwords, and (2) by using the ”AND” operator and using all words in the question, except stopwords and words in list of common English words (cf. (Neves, 2015)). We retrieve up to 200 PubMed documents for each of the queries and index these in the IMDB. We rank the sentences for each question based on an approximate similarity between the words in the question and the ones in the document, while a score is automatically calculate between those. Fi- nally, we rank the sentences according to the sum of scores of the matching words and select the top 10 sentences. The list of up to 10 documents is de- rived from the list to top 10 sentences, i.e., the cor- responding documents of these sentences, in the same order. Table 1: Preliminary results in the BioASQ task 4b. Scores for concepts, documents and snippets are in terms of MAP (Mean Average Precision). ”na” indicated that results are still not available for this task, while ”-” indicated that we did not sub- mit any run for the task. The values inside parame- ters indicate our current rank and the total number of submissions for the task. Table 1: Preliminary results in the BioASQ task 4b. Scores for concepts, documents and snippets are in terms of MAP (Mean Average Precision). 4.2 Systems We participated this year with two QA systems, as identiﬁed by their run names: 1. HPI1: our previous system that participated in the BioASQ challenge last year (Neves, 2015); 1. Initialize two sets: (a) an empty set A and a set B that contains all extracted sentences. 2. HPI2: our new QA system, which is de- scribed in this work. 2. Order the sentences in set B by decreasing order of their score. HPI1 is exactly the same system that partici- pated in the BioASQ 2015 and that was one of the winners systems6. We made no changes in the 3. Move the top sentence si from set B to set A. Then penalize all sentences sj whose similar- ity to si is greater than a threshold of 0.3 by multiplying their score with the penalty fac- tor of 0.5. 41 system and details on the methods can be found in our previous publication (Neves, 2015). This system was used this year for concept matching and for document and snippet retrieval. The only change made to this system was on the dictio- naries which are used in the concept matching task of Phase A. The dictionaries were re-created based on newer versions of the ﬁve terminologies speciﬁed in the guidelines of the BioASQ chal- lenge: DO, MeSH, Jochem, GO and Uniprot. We downloaded the original ﬁles from the respective web sites and compiled dictionaries for each of the terminologies. The dictionaries include var- ious names and synonyms for each concept and was used by the built-in NER functionality of the database to match concepts to the questions. HPI1 HPI2 Concepts MAP MAP batch1 na - batch2 - - batch3 na - batch4 na - batch5 na - Documents MAP MAP batch1 0.0474 (12/15) 0.0028 (15/15) batch2 - - batch3 0.0674 (16/18) 0.0006 (18/18) batch4 - - batch5 0.434 (16/21) - Snippets MAP MAP batch1 0.0481 (1/7) - batch2 - - batch3 0.0715 (4/14) - batch4 - - batch5 0.0510 (5/16) - Ideal Answ. Rouge-2 Rouge-2 batch1 - 0.2231 (1/2) batch2 - 0.2240 (6/7) batch3 - 0.2559 (6/7) batch4 - 0.2280 (4/4) batch5 - 0.3233 (6/7) Table 1: Preliminary results in the BioASQ task 4b. Scores for concepts, documents and snippets are in terms of MAP (Mean Average Precision). Acknowledgments Snippets. Our system HPI1 performed well again and it a good candidate for obtaining ﬁrst and second position in the challenge. This proves that the IMDB could effectively match the key- words in the queries to the documents and rank the sentences. However, we see much room for im- provement in our approach as named-entities are still not being used in this component, a step which can certainly improve both document and passage retrieval. We thank David Heller, Thomas Hille and Fabian Eckert for the interesting discussions during the project and the HPI Future Soc Lab Team for pro- viding us the access to the IMDB. Finally, we also would like to thank technical support from the stu- dents of our current Bachelor Project: Maximilian Goetz, Marcel Jankrift, Julian Niedermeier, Toni Stachewicz and Soeren Tietboehl. Ideal Answers. Our results for ideal answers, i.e., short summaries, provided by system HPI2 also obtained either ﬁrst or second positions in the all of the batches, when considering results by teams, instead of each individual run. References Soﬁa J. Athenikos and Hyoil Han. 2010. Biomedical question answering: A survey. Computer Methods and Programs in Biomedicine, 99(1):1 – 24. MichaelA Bauer and Daniel Berleant. 2012. Usability survey of biomedical question answering systems. Human Genomics, 6(1):1–4. 4.3 Evaluation Currently, only preliminary results are available for some of the tasks of the BioASQ challenge. We summarize them in Table 1. More details on the results can be found in the BioASQ web site 7. We present in this section a discussion on the preliminary results that we obtained in the BioASQ challenge, on the limitation of our meth- ods and improvements for future versions of our QA system. 42 Documents. Curiously, although the strategy used for the document retrieval is exactly the same one used for the snippet retrieval, we obtained much better results for the later, in term of po- sition in the ranking, also in previous editions of the BioASQ challenge. As gold-standard and not available, we can only try to guess the reasons for our performance. When comparing our two systems, HPI2 performed much worse than HPI1, which proves that we still have to need to be im- proved to deal with large document collections, while HPI1 rely on up to 200 previously retrieved from PubMed. system are close to 100%. We should per- form a comprehensive evaluation of the ques- tion processing step, specially the target iden- tiﬁcation step, and properly integrate further components which can potentially boost our results, such as NER, chunking and semantic role labeling. Finally, we should perform a comprehensive evaluation on biomedical corpora for the many built-in NLP components of the IMDB, such as NER and POS tagging, as mistakes returned by these can be propagated throughout the system. Mariana Neves and Ulf Leser. 2015. Question answer- ing for biology. Methods, 74:36 – 46. 5 Conclusions and Future Work In this work we present our results for our two QA systems that participated in task 4b of the BioASQ challenge. The preliminary results show that our approaches are obtained top positions for the snip- pet retrieval and for the ideal answers. Regarding future work, we envisage much room for improve- ment for our HPI2 system, the one which is cur- rently under development in our group: G¨unes Erkan and Dragomir R Radev. 2004. Lexrank: Graph-based lexical centrality as salience in text summarization. Journal of Artiﬁcial Intelligence Research, pages 457–479. Zhiheng Huang, Marcus Thint, and Zengchang Qin. 2008. Question classiﬁcation using head words and their hypernyms. In Proceedings of the Conference on Empirical Methods in Natural Language Pro- cessing, EMNLP ’08, pages 927–936, Stroudsburg, PA, USA. Association for Computational Linguis- tics. • Both the document and snippet retrieval steps performed much worse than the HPI1 sys- tem, which rely on PubMed API. Future work should aim at improving our current ranking algorithms. Daniel Jurafsky and James H. Martin. 2013. Speech and Language Processing. Prentice Hall Interna- tional, 2 revised edition. Christopher D. Manning, Prabhakar Raghavan, and Hinrich Sch¨utze. 2008. Introduction to Information Retrieval. Cambridge University Press, Cambridge, UK. • We did not submit runs for factoid and list questions because our system could not re- turn any answer for most of the answers. We did submit one run for yes/no questions but MAP value was of only 25%, while other 43 Mariana Neves. 2015. HPI question answering system in the bioasq 2015 challenge. In Working Notes of CLEF 2015 - Conference and Labs of the Evaluation forum, Toulouse, France, September 8-11, 2015. Lawrence Page, Sergey Brin, Rajeev Motwani, and Terry Winograd. 1999. The pagerank citation rank- ing: bringing order to the web. Hasso Plattner. 2013. A Course in In-Memory Data Management: The Inner Mechanics of In-Memory Databases. Springer, 1st edition. George Tsatsaronis, Georgios Balikas, Prodromos Malakasiotis, Ioannis Partalas, Matthias Zschunke, Michael R Alvers, Dirk Weissenborn, Anastasia Krithara, Sergios Petridis, Dimitris Polychronopou- los, et al. 2015. An overview of the bioasq large-scale biomedical semantic indexing and ques- tion answering competition. BMC bioinformatics, 16(1):138. 44 44 44
https://openalex.org/W4389519307	https://aclanthology.org/2023.findings-emnlp.128.pdf	English	null	Transformer-Based Language Model Surprisal Predicts Human Reading Times Best with About Two Billion Training Tokens	null	2,023	cc-by	4,494	Abstract been evaluated on their ability to predict measures of processing difficulty. Recent psycholinguistic studies have drawn conflicting conclusions about the relationship between the quality of a language model and the ability of its surprisal estimates to predict human reading times, which has been specu- lated to be due to the large gap in both the amount of training data and model capacity across studies. The current work aims to consol- idate these findings by evaluating surprisal esti- mates from Transformer-based language model variants that vary systematically in the amount of training data and model capacity on their ability to predict human reading times. The results show that surprisal estimates from most variants with contemporary model capacities provide the best fit after seeing about two bil- lion training tokens, after which they begin to diverge from humanlike expectations. Addi- tionally, newly-trained smaller model variants reveal a ‘tipping point’ at convergence, after which the decrease in language model perplex- ity begins to result in poorer fits to human read- ing times. These results suggest that the mas- sive amount of training data is mainly respon- sible for the poorer fit achieved by surprisal from larger pre-trained language models, and that a certain degree of model capacity is nec- essary for Transformer-based language models to capture humanlike expectations. Recent studies in computational psycholinguis- tics have provided conflicting evidence with regard to the relationship between LM quality (i.e. next- word prediction accuracy) and goodness-of-fit to human reading times. Earlier work using newly- trained LMs showed a negative relationship be- tween LM perplexity and predictive power of sur- prisal estimates (Goodkind and Bicknell, 2018; Wilcox et al., 2020; Merkx and Frank, 2021), but more recent work using large pre-trained Transformer-based LMs (e.g. GPT-2; Radford et al., 2019) show a robust positive relationship between the two variables (Oh et al., 2022; Oh and Schuler, 2023). While Oh and Schuler (2023) conjecture that these studies capture two distinct regimes, it remains less clear where the reversal in this rela- tionship happens. The main challenge in answering this question lies in the massive difference in terms of both the amount of training data and the model capacity of LMs that were studied. Abstract The current study aims to cover this conceptual middle ground by evaluating, on their ability to pre- dict human reading times, surprisal estimates from Transformer-based LM variants that vary system- atically in the amount of training data and model capacity. Results from regression analyses show that surprisal from most LM variants with contem- porary model capacities make the biggest contri- bution to regression model fit after seeing about two billion tokens of training data, after which ad- ditional training data result in surprisal estimates that continue to diverge from humanlike expecta- tions. Additionally, surprisal estimates from newly- trained smaller LM variants reveal a ‘tipping point’ at convergence, after which the decrease in perplex- ity begins to result in poorer fits to human read- ing times. Taken together, these results suggest that the vast amount of training data is mainly re- sponsible for the poorer fit achieved by surprisal Transformer-Based Language Model Surprisal Predicts Human Reading Times Best with About Two Billion Training Tokens Byung-Doh Oh Department of Linguistics The Ohio State University oh.531@osu.edu William Schuler Department of Linguistics The Ohio State University schuler.77@osu.edu Findings of the Association for Computational Linguistics: EMNLP 2023, pages 1915–1921 December 6-10, 2023 ©2023 Association for Computational Linguistics 1 Introduction The predictability of upcoming linguistic material has long been considered a crucial factor underly- ing difficulty in human sentence processing (Hale, 2001; Levy, 2008), and has received empirical support from numerous studies showing surprisal (Shannon, 1948) to be highly predictive of relevant behavioral and neural measures (e.g. Demberg and Keller, 2008; Smith and Levy, 2013; Hale et al., 2018; Shain et al., 2020). Since language models (LMs) are trained to estimate a conditional proba- bility distribution of a word given its context, sur- prisal estimates calculated from them have often 1915 Model #L #H dmodel #Parameters Pythia 70M 6 8 512 ∼70M Pythia 160M 12 12 768 ∼160M Pythia 410M 24 16 1024 ∼410M Pythia 1B 16 8 2048 ∼1B Pythia 1.4B 24 16 2048 ∼1.4B Pythia 2.8B 32 32 2560 ∼2.8B Pythia 6.9B 32 32 4096 ∼6.9B Pythia 12B 36 40 5120 ∼12B Table 1: Model capacities of Pythia variants whose surprisal estimates were examined in this work. #L, #H, and dmodel refer to number of layers, number of attention heads per layer, and embedding size, respectively. from larger Transformer-based pre-trained LMs (Oh et al., 2022; Oh and Schuler, 2023), and that a certain degree of model capacity is necessary for Transformer-based LMs to capture humanlike expectations that manifest in reading times. 2.1 Response Data The Natural Stories Corpus contains reading times from 181 subjects that read 10 naturalistic English stories consisting a total of 10,245 tokens. The data points were filtered to remove those for sentence- initial and final words, those from subjects who answered three or fewer comprehension questions correctly, and those shorter than 100 ms or longer than 3000 ms, which resulted in a total of 384,905 observations in the exploratory set. The Dundee Corpus contains eye-gaze durations from 10 sub- jects that read 67 newspaper editorials consisting a total of 51,501 tokens. The data points were fil- tered to exclude those for unfixated words, words following saccades longer than four words, and sentence-, screen-, document-, and line-initial and final words, which resulted in a total of 98,115 ob- servations in the exploratory set.1 All observations were log-transformed prior to model fitting. Each article of the Natural Stories Corpus and each article of the Dundee Corpus was tokenized by Pythia’s byte-pair encoding (BPE; Sennrich et al., 2016) tokenizer and provided as input to each model variant. For each model variant, all publicly available intermediate model weights were used to calculate surprisal estimates on the two corpora. In cases where each story or article was longer than a single context window of 2,048 tokens, surprisal estimates for the remaining tokens were calculated by using the second half of the previous context window as the first half of a new context window. 2 Experiment 1: Influence of Training Data Size Table 1: Model capacities of Pythia variants whose surprisal estimates were examined in this work. #L, #H, and dmodel refer to number of layers, number of attention heads per layer, and embedding size, respectively. Table 1: Model capacities of Pythia variants whose surprisal estimates were examined in this work. #L, #H, and dmodel refer to number of layers, number of attention heads per layer, and embedding size, respectively. The first experiment examines the influence of training data size on the predictive power of Transformer-based LM surprisal by evaluating LM variants at various points in training on self-paced reading times from the Natural Stories Corpus (Futrell et al., 2021) and go-past eye-gaze durations from the Dundee Corpus (Kennedy et al., 2003). Crucially for this experiment, all eight Pythia variants were trained using identical batches of training examples that were presented in the same order. These training examples come from the Pile (Gao et al., 2020), which is a collection of English language datasets that consist of around 300 bil- lion tokens. Batches of 1,024 examples with a sequence length of 2,048 (i.e. 2,097,152 tokens) were used to train the eight variants for 143,000 steps, which amounts to about one epoch of the entire Pile dataset. Model parameters that were saved during early training stages (i.e. after 1, 2, 4, ..., 256, 512 steps) as well as after every 1,000 steps are publicly available. 2.2 Predictors This experiment evaluates surprisal estimates from eight variants of Pythia LMs (Biderman et al., 2023), whose intermediate parameters were saved at various points during training. Pythia LMs are decoder-only autoregressive Transformer-based models2 whose variants differ primarily in their ca- pacity. The model capacities of the Pythia variants are summarized in Table 1. 2Technical details such as the parallelization of self- attention/feedforward computations and the separation of em- bedding/projection matrices differentiate Pythia LMs from other large language model families. 1The held-out set of each corpus, which have a comparable number of observations, is reserved for statistical significance testing and therefore was not analyzed in this work. 1The held-out set of each corpus, which have a comparable number of observations, is reserved for statistical significance testing and therefore was not analyzed in this work. 2Technical details such as the parallelization of self- attention/feedforward computations and the separation of em- bedding/projection matrices differentiate Pythia LMs from other large language model families. 4Results from after 2,000 steps were selectively omitted for clarity, as they were consistent with the general trend. As pointed out by a reviewer, including a frequency-based pre- dictor in the regression models may change the exact location of this peak. However, this work avoids potential confounds introduced by the corpus used for frequency estimation by evaluating surprisal estimates on their own following the pro- tocols of Oh and Schuler (2023). 3‘Spillover’ predictors were not included in the regression models to avoid convergence issues. 2.3 Regression Modeling Subsequently, following previous work (Oh et al., 2022; Oh and Schuler, 2023), a ‘baseline’ linear mixed-effects (LME) model that contains baseline predictors for low-level cognitive processing, and ‘full’ LME models that additionally contain each LM surprisal predictor, were fit to self-paced read- ing times and go-past durations using lme4 (Bates et al., 2015). These baseline predictors are word length in characters and index of word position in each sentence (Natural Stories and Dundee), as well as saccade length and whether or not the pre- 1916 Figure 1: Increase in regression model log-likelihood due to including each surprisal estimate from Pythia variants as a function of training steps (top) and perplexity (middle; the stars indicate the fully trained versions after 143,000 steps), as well as perplexity as a function of training steps (bottom) on the exploratory set of Natural Stories (left) and Dundee data (right). Figure 1: Increase in regression model log-likelihood due to including each surprisal estimate from Pythia variants as a function of training steps (top) and perplexity (middle; the stars indicate the fully trained versions after 143,000 steps), as well as perplexity as a function of training steps (bottom) on the exploratory set of Natural Stories (left) and Dundee data (right). gression model. Finally, the perplexity of each LM variant was calculated on the two corpora. vious word was fixated (Dundee only). All pre- dictors were centered and scaled,3 and the LME models included by-subject random slopes for all fixed effects and random intercepts for each subject. In addition, a random intercept for each subject- sentence interaction was included for self-paced reading times collected from 181 subjects, and a random intercept for each sentence was included for eye-gaze durations collected from a smaller number of 10 subjects. Once the regression mod- els were fit, the increase in regression model log- likelihood (∆LL) was calculated for each regres- sion model by subtracting the log-likelihood of the baseline regression model from that of a full re- from larger Transformer-based LMs. from larger Transformer-based LMs. optimum,’ after which surprisal estimates begin to diverge from humanlike expectations as training continues. At this point in training, there appears to be no systematic relationship between model ca- pacity and predictive power of surprisal estimates. However, after all 143,000 training steps (i.e. after about 300 billion tokens), the eight model variants show a strictly monotonic and negative relation- ship, which directly replicates the findings of Oh and Schuler (2023).5 Taken together, these results indicate that the vast amount of training data is re- sponsible for the poorer fit achieved by surprisal 5The best-fitting line between log perplexity and ∆LL of these variants had a slope significantly greater than 0 at p < 0.05 level according to a one-tailed t-test on both corpora. 3 Experiment 2: Influence of Model Capacity The second experiment further examines the re- lationship between model capacity and predic- tive power of surprisal estimates by evaluating Transformer-based LM variants smaller than the Pythia variants at various points in training, follow- ing similar procedures as Experiment 1. 2.4 Results The results in Figure 1 show that across both cor- pora, surprisal from most LM variants made the biggest contribution to regression model fit after 1,000 training steps (i.e. after about two billion tokens).4 This seems to represent a ‘humanlike 1917 Figure 2: Increase in regression model log-likelihood due to including each surprisal estimate from newly-trained LM variants as a function of training steps (top) and perplexity (middle; the stars indicate the fully trained versions after 10,000 steps), as well as perplexity as a function of training steps (bottom) on the exploratory set of Natural Stories (left) and Dundee data (right). The variants are labeled using their number of layers, number of attention heads per layer, and embedding size, in that order. Figure 2: Increase in regression model log-likelihood due to including each surprisal estimate from newly-trained LM variants as a function of training steps (top) and perplexity (middle; the stars indicate the fully trained versions after 10,000 steps), as well as perplexity as a function of training steps (bottom) on the exploratory set of Natural Stories (left) and Dundee data (right). The variants are labeled using their number of layers, number of attention heads per layer, and embedding size, in that order. 6Refer to Appendix A for the model capacities of these variants as well as further details on their training procedures. Code and trained weights are available at https://github. com/byungdoh/slm_surprisal. 3.2 Results The results in Figure 2 show that surprisal from the two largest variants made the biggest contribution to regression model fit after 1,000 training steps on both corpora, replicating the results of Experiment 1. In contrast, the smaller variants such as the 2- 2-128 and 2-3-192 variants seem to peak later at around 2,000 training steps and stabilize afterward. After all 10,000 training steps, the model variants show a reversal in the relationship between LM perplexity and fit to reading times; the 2-3-192 variant seems to represent a ‘tipping point,’ after which the decrease in perplexity starts to result in poorer fits to human reading times. Additionally, variants that are smaller than this yield surprisal estimates that are less predictive of reading times when sufficiently trained. These results suggest that a certain degree of model capacity is necessary for Transformer-based LMs to capture humanlike expectations that manifest in reading times. Finally, across both experiments, model capacity does not seem to modulate the relationship between perplexity and fit to human reading times, with data points from different LM variants forming a con- tinuous curve between log perplexity and ∆LL. This suggests that Transformer-based LMs of dif- ferent capacities share a similar inductive bias that initially improves the fit of surprisal estimates to human reading times but begins to have an adverse effect on it with large amounts of training data. 3.1 Procedures Surprisal estimates from eight smaller LM vari- ants were evaluated at various points during train- ing in this experiment. The largest of these vari- 1918 ants has the same model capacity as the smallest Pythia 70M variant, and the smaller variants were designed to have fewer layers and attention heads, as well as smaller embeddings. These variants were trained closely following the training proce- dures of the Pythia variants, including the size and order of training batches. For computational ef- ficiency, these variants were trained for the first 10,000 training steps, based on the observation that ∆LL on both corpora did not change substantially after 8,000 steps for the smallest Pythia variant.6 The predictive power of resulting surprisal esti- mates was evaluated following identical procedures as Experiment 1. helpful for learning e.g. selectional preferences that align well with humanlike prediction and process- ing difficulty. However, as the models see more training data, they are able to achieve ‘superhuman’ prediction, which makes their surprisal estimates diverge more and more from human reading times as training continues. The words for which predic- tion by LMs improves with massive amounts of training data are likely to be open-class words like nouns and adjectives, whose reading times were identifed as being most severely underpredicted by their surprisal estimates (Oh and Schuler, 2023). Moreover, at the end of training, these model variants show a strictly monotonic and negative relationship between perplexity and fit to human reading times. This directly replicates the find- ings of Oh et al. (2022) and adds to a growing body of research reporting an inverse correlation be- tween model size and regression model fit (Kurib- ayashi et al., 2022; Shain et al., 2022; de Varda and Marelli, 2023). The current results demonstrate that this relationship emerges with large amounts of training data and becomes stronger as training continues. The bottleneck posed by the limited model capacity of the smaller variants appears to prevent them from learning to make excessively accurate predictions that cause the divergence be- tween surprisal and human reading times. How- ever, newly-trained LM variants that are smaller than those of contemporary standards reveal a ‘tip- ping point’ at convergence, which indicates that a certain amount of model capacity is necessary for LMs to correctly learn humanlike expectations. 4 Discussion and Conclusion This work aims to consolidate conflicting findings about the relationship between LM quality and the predictive power of its surprisal estimates by systematically manipulating the amount of train- ing data and model capacity. Experimental re- sults show that surprisal from most contemporary Transformer-based LM variants provide the best fit to human reading times with about two billion training tokens, after which they begin to diverge from humanlike expectations. It is conjectured that early training data up to about two billion tokens is Acknowledgments We thank the reviewers and the area chair for their helpful comments. This work was supported by the National Science Foundation grant #1816891. All views expressed are those of the authors and do not necessarily reflect the views of the National Science Foundation. 1919 Ethics Statement Experiments presented in this work used datasets from previously published research (Futrell et al., 2021; Kennedy et al., 2003), in which the proce- dures for data collection and validation are outlined. As this work focuses on studying the connection between conditional probabilities of language mod- els and human sentence processing, its potential negative impacts on society seem to be minimal. Adam Goodkind and Klinton Bicknell. 2018. Predic- tive power of word surprisal for reading times is a linear function of language model quality. In Pro- ceedings of the 8th Workshop on Cognitive Modeling and Computational Linguistics, pages 10–18. John Hale. 2001. A probabilistic Earley parser as a psy- cholinguistic model. In Proceedings of the Second Meeting of the North American Chapter of the Asso- ciation for Computational Linguistics on Language Technologies, pages 1–8. John Hale, Chris Dyer, Adhiguna Kuncoro, and Jonathan Brennan. 2018. Finding syntax in human encephalography with beam search. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics, pages 2727–2736. Limitations Richard Futrell, Edward Gibson, Harry J. Tily, Idan Blank, Anastasia Vishnevetsky, Steven Piantadosi, and Evelina Fedorenko. 2021. The Natural Stories corpus: A reading-time corpus of English texts con- taining rare syntactic constructions. Language Re- sources and Evaluation, 55:63–77. The connection between conditional probabilities of Transformer-based language models and human sentence processing drawn in this work is based on language model variants trained on English text and data from human subjects that are native speakers of English. Therefore, the connection made in this work may not generalize to other languages. Leo Gao, Stella Biderman, Sid Black, Laurence Gold- ing, Travis Hoppe, Charles Foster, Jason Phang, Horace He, Anish Thite, Noa Nabeshima, Shawn Presser, and Connor Leahy. 2020. The Pile: An 800GB dataset of diverse text for language modeling. arXiv preprint, arXiv:2101.00027. References Alex Andonian, Quentin Anthony, Stella Biderman, Sid Black, Preetham Gali, Leo Gao, Eric Hallahan, Josh Levy-Kramer, Connor Leahy, Lucas Nestler, Kip Parker, Michael Pieler, Shivanshu Purohit, Tri Songz, Phil Wang, and Samuel Weinbach. 2021. GPT-NeoX: Large scale autoregressive language modeling in Py- Torch. Alan Kennedy, Robin Hill, and Joël Pynte. 2003. The Dundee Corpus. In Proceedings of the 12th Euro- pean Conference on Eye Movement. Douglas Bates, Martin Mächler, Ben Bolker, and Steve Walker. 2015. Fitting linear mixed-effects models using lme4. Journal of Statistical Software, 67(1):1– 48. Diederik P. Kingma and Jimmy Ba. 2015. Adam: A method for stochastic optimization. In Proceedings of the 3rd International Conference on Learning Rep- resentations, pages 1–15. Tatsuki Kuribayashi, Yohei Oseki, Ana Brassard, and Kentaro Inui. 2022. Context limitations make neural language models more human-like. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 10421–10436. Stella Biderman, Hailey Schoelkopf, Quentin Gregory Anthony, Herbie Bradley, Kyle O’Brien, Eric Hal- lahan, Mohammad Aflah Khan, Shivanshu Purohit, USVSN Sai Prashanth, Edward Raff, Aviya Skowron, Lintang Sutawika, and Oskar van der Wal. 2023. Pythia: A suite for analyzing large language mod- els across training and scaling. In Proceedings of the 40th International Conference on Machine Learning, volume 202, pages 2397–2430. Roger Levy. 2008. Expectation-based syntactic compre- hension. Cognition, 106(3):1126–1177. Danny Merkx and Stefan L. Frank. 2021. Human sen- tence processing: Recurrence or attention? In Pro- ceedings of the Workshop on Cognitive Modeling and Computational Linguistics, pages 12–22. Tri Dao, Daniel Y. Fu, Stefano Ermon, Atri Rudra, and Christopher Ré. 2022. FlashAttention: Fast and memory-efficient exact attention with IO-awareness. arXiv preprint, arXiv:2205.14135. Byung-Doh Oh, Christian Clark, and William Schuler. 2022. Comparison of structural parsers and neural language models as surprisal estimators. Frontiers in Artificial Intelligence, 5:777963. Andrea de Varda and Marco Marelli. 2023. Scaling in cognitive modelling: a multilingual approach to human reading times. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics, pages 139–149. Byung-Doh Oh and William Schuler. 2023. Why does surprisal from larger Transformer-based language models provide a poorer fit to human reading times? Transactions of the Association for Computational Linguistics, 11:336–350. Vera Demberg and Frank Keller. 2008. Data from eye- tracking corpora as evidence for theories of syntactic processing complexity. Cognition, 109(2):193–210. 7The only minor difference is that the FlashAttention (Dao et al., 2022) implementation of scaled dot-product attention could not be used during training due to a mismatch in GPU hardware specifications. References 2020) were provided to each variant in the same or- der as the Pythia variants. The variants were trained using the Zero Redundancy Optimizer (ZeRO; Ra- jbhandari et al., 2020) implementation of Adam (Kingma and Ba, 2015) with a learning rate of 0.001. The learning rate was warmed up linearly over the first 1% of training steps (i.e. 1,430 steps) and were subsequently lowered to a minimum of 0.0001 following a cosine annealing schedule over the remainder of the 143,000 training steps. How- ever, for computational efficiency, training was stopped after the first 10,000 training steps. For comparability with the Pythia variants, intermedi- ate parameters were saved during early training stages (i.e. after 1, 2, 4, ..., 256, 512 steps) as well as after every 500 steps from step 1,000 onward. Cory Shain, Idan Asher Blank, Marten van Schijn- del, William Schuler, and Evelina Fedorenko. 2020. fMRI reveals language-specific predictive coding dur- ing naturalistic sentence comprehension. Neuropsy- chologia, 138:107307. Cory Shain, Clara Meister, Tiago Pimentel, Ryan Cot- terell, and Roger Levy. 2022. Large-scale evidence for logarithmic effects of word predictability on read- ing time. PsyArXiv preprint. Claude Elwood Shannon. 1948. A mathematical theory of communication. Bell System Technical Journal, 27:379–423. Nathaniel J. Smith and Roger Levy. 2013. The effect of word predictability on reading time is logarithmic. Cognition, 128:302–319. Ethan Gotlieb Wilcox, Jon Gauthier, Jennifer Hu, Peng Qian, and Roger P. Levy. 2020. On the predictive power of neural language models for human real- time comprehension behavior. In Proceedings of the 42nd Annual Meeting of the Cognitive Science Society, pages 1707–1713. Ethan Gotlieb Wilcox, Jon Gauthier, Jennifer Hu, Peng Qian, and Roger P. Levy. 2020. On the predictive power of neural language models for human real- time comprehension behavior. In Proceedings of the 42nd Annual Meeting of the Cognitive Science Society, pages 1707–1713. References 1920 Model #L #H dmodel #Parameters Repro 1-1-64 1 1 64 ∼6M Repro 1-2-128 1 2 128 ∼13M Repro 2-2-128 2 2 128 ∼13M Repro 2-3-192 2 3 192 ∼20M Repro 2-4-256 2 4 256 ∼27M Repro 3-4-256 3 4 256 ∼28M Repro 4-6-384 4 6 384 ∼46M Repro 6-8-512 6 8 512 ∼70M Table 2: Model capacities of newly-trained LM variants whose surprisal estimates were examined in this work. #L, #H, and dmodel refer to number of layers, number of attention heads, and embedding size, respectively. Alec Radford, Jeff Wu, Rewon Child, David Luan, Dario Amodei, and Ilya Sutskever. 2019. Language models are unsupervised multitask learners. OpenAI Technical Report. Alec Radford, Jeff Wu, Rewon Child, David Luan, Dario Amodei, and Ilya Sutskever. 2019. Language models are unsupervised multitask learners. OpenAI Technical Report. Samyam Rajbhandari, Jeff Rasley, Olatunji Ruwase, and Yuxiong He. 2020. ZeRO: Memory optimiza- tions toward training trillion parameter models. In Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis, pages 1–16. Table 2: Model capacities of newly-trained LM variants whose surprisal estimates were examined in this work. #L, #H, and dmodel refer to number of layers, number of attention heads, and embedding size, respectively. Table 2: Model capacities of newly-trained LM variants whose surprisal estimates were examined in this work. #L, #H, and dmodel refer to number of layers, number of attention heads, and embedding size, respectively. Rico Sennrich, Barry Haddow, and Alexandra Birch. 2016. Neural machine translation of rare words with subword units. In Proceedings of the 54th Annual Meeting of the Association for Computational Lin- guistics, pages 1715–1725. 2020) were provided to each variant in the same or- der as the Pythia variants. The variants were trained using the Zero Redundancy Optimizer (ZeRO; Ra- jbhandari et al., 2020) implementation of Adam (Kingma and Ba, 2015) with a learning rate of 0.001. The learning rate was warmed up linearly over the first 1% of training steps (i.e. 1,430 steps) and were subsequently lowered to a minimum of 0.0001 following a cosine annealing schedule over the remainder of the 143,000 training steps. How- ever, for computational efficiency, training was stopped after the first 10,000 training steps. For comparability with the Pythia variants, intermedi- ate parameters were saved during early training stages (i.e. after 1, 2, 4, ..., 256, 512 steps) as well as after every 500 steps from step 1,000 onward. A Model Capacities and Training Procedures of Smaller LM Variants The eight LM variants that were trained as part of Experiment 2 are decoder-only autoregressive Transformer-based models that share the same ar- chitecture as the Pythia LM variants (Biderman et al., 2023). Their model capacities are summa- rized in Table 2. These variants were trained using the GPT-NeoX library (Andonian et al., 2021) closely following the training procedures of the Pythia LM variants.7 Identical training batches of 1,024 examples with a sequence length of 2,048 from the Pile (Gao et al., 1921
https://openalex.org/W4393093208	https://applbiolchem.springeropen.com/counter/pdf/10.1186/s13765-024-00884-9	English	null	Development and application of a high-performance liquid chromatography diode-array detection (HPLC–DAD) method for the simultaneous quantification of phenolic compounds in the aerial part of Glehnia littoralis	Applied biological chemistry	2,024	cc-by	9,213	†Yun Ji Park and Yeong Bin Choi contributed equally to this work. Glehnia littoralis belongs to the Apiaceae family and is a perennial herb distributed in coastal dunes throughout East Asia [1]. The G. littoralis root, Radix Glehniae, has been traditionally used in medicine for diaphoretic, anti- pyretic, and analgesic purposes [2]. It is also listed in the Korean, Japanese, and Chinese pharmacopeia and is pre- scribed as a tonic and mucolytic for the treatment of gas- trointestinal and respiratory disorders [3]. Furthermore, Correspondence: Sang Min Kim kimsm@kist.re.kr 1 Smart Farm Research Center, Korea Institute of Science and Technology (KIST), Gangneung Institute of Natural Products, 679, Saimdang‑Ro, Gangneung, Gangwon‑Do 25451, Republic of Korea 2 Department of Bio‑Medical Science & Technology, University of Science and Technology, Seoul 02792, Republic of Korea © The Author(s) 2024. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. Development and application of a high‑performance liquid chromatography diode‑array detection (HPLC–DAD) method for the simultaneous quantification of phenolic compounds in the aerial part of Glehnia littoralis Development and application of a high‑performance liquid chromatography diode‑array detection (HPLC–DAD) method for the simultaneous quantification of phenolic compounds in the aerial part of Glehnia littoralis Abstract Glehnia littoralis, a medicinal herb employed in traditional practices for alleviating fatigue, cough, and a dry throat, is recognized for its beneficial properties due to a diverse array of active compounds found in its extracts. For exam‑ ple, the G. littoralis roots (Radix Glehniae) mainly contain coumarins and phenolic acids, serving as the primary focus of this study. Despite the widespread use of the tools in various industries and the development of multiple analytical methods for their examination, the edible aerial parts have industrial potential, and there is currently no analytical method available to identify their key components. In this study, a high-performance liquid chromatography method combined with diode array detection (HPLC–DAD) was developed to simultaneously detect 16 phenolic compounds previously reported to be present in the edible aerial parts of G. littoralis. The proposed approach included using gra‑ dient elution to change the solvent system from water/acetonitrile to water/methanol. Furthermore, the method vali‑ dation was conducted, assessing its linearity, limit of detection, limit of quantification, precision, accuracy, and recov‑ ery, all of which demonstrated satisfactory results. Subsequently, the developed method was applied to quantify the phenolic compounds in various G. littoralis samples obtained from different organs, solvent extraction processes, and processing methods. Moreover, the online HPLC-ABTS (2,2ʹ-azino-bis(3-ethylbenzthiazoline-6-sulfonic acid) assay was used to evaluate the antioxidant capacities of individual constituents, identifying four important antioxidants and estimate the overall antioxidant capacity of the G. littoralis extract. Keywords High-performance liquid chromatography, Diode-array detection, Phenolic compound, Glehnia littoralis, Antioxidant activity Correspondence: Sang Min Kim kimsm@kist.re.kr 1 Smart Farm Research Center, Korea Institute of Science and Technology (KIST), Gangneung Institute of Natural Products, 679, Saimdang‑Ro, Gangneung, Gangwon‑Do 25451, Republic of Korea 2 Department of Bio‑Medical Science & Technology, University of Science and Technology, Seoul 02792, Republic of Korea Park et al. Applied Biological Chemistry (2024) 67:34 https://doi.org/10.1186/s13765-024-00884-9 © The Author(s) 2024. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. Chemicals and preparation of the standard solutionsl ( ) Chemicals and preparation of the standard solutions Nicotiflorin (≥ 98.8%, CAS No. 17650-84-9), hyperoside (≥ 95.8%, CAS No. 482-36-0), and phellopterin (≥ 98.7%, CAS No. 2543-94-4) were purchased from Chromadex (Irvine, CA, USA). Chlorogenic acid (≥ 95.0%, CAS No. 327-97-9), caffeic acid (≥ 98.0%, CAS No. 331-39-5), sco- poletin (≥ 99.0%, CAS No. 92-61-5), rutin (≥ 95.0%, CAS No. 153-18-4), isoquercetin (≥ 98.2%, CAS No. 482-35- 9), psoralen (≥ 99.0%, CAS No. 66-97-7), xanthotoxin (≥ 98.0%, CAS No. 298-81-7), bergapten (≥ 99.0%, CAS No. 484-20-8), isopimpinellin (≥ 95.0%, CAS No. 482-27- 9), oxypeucedanin (≥ 98.0%, CAS No. 737-52-0), imper- atorin (≥ 98.0%, CAS No. 482-44-0), isoimperatorin (≥ 98.0%, CAS No. 482-45-1), bergamottin (≥ 98.0%, CAS No. 7380-40-7), ABTS (CAS No. 30931-67-0), ( ±)-6-hydroxy-2,5,7,8-tetramethylchromane-2-carbox- ylic acid (Trolox, CAS No. 53188-07-1), potassium per- sulfate (CAS No. 7727-21-1), and formic acid (CAS No. 64-18-6) were supplied by Sigma-Aldrich (St. Louis, MO, USA). HPLC-grade water, methanol, and acetonitrile were purchased from Thermo Fisher Scientific (Waltham, MA, USA). j In several countries, the fresh leaves and stems of G. lit- toralis are commonly consumed and used in processed foods, such as rice cakes and tea [7]. Although the major- ity of phytochemicals in this plant are derived from the underground organs (i.e., the roots and rhizomes), a few studies have explored the phytochemical compositions of its aerial parts. For example, G. littoralis leaf extracts sig- nificantly inhibited tyrosinase and elastase, implying pos- sible application in cosmetic products for skin whitening and anti-wrinkle qualities [5]. Moreover, extracts from the aerial parts of this plant have been shown to regulate humoral immunity and potently inhibit the development of human cancer cells. This also raises the possibility that the antiproliferative action of the crude extracts of G. lit- toralis may be partially attributed to the coumarins and polyacetylenes [8]. Moreover, the whole plant extract demonstrated a noteworthy inhibitory activity against nitric oxide production, suggesting that angular-type dihydropyranocoumarins possessing ester groups at the C-3′ and C-4′ positions hold considerable therapeutic promise as anti-inflammatory agents [9]. Despite ongo- ing studies into G. littoralis for application in pharma- ceuticals, health-functional foods, and cosmetics, there have been very few studies on the edibility of the aerial parts, and research has predominantly focused on the root, which is used solely in traditional medicine [1]. As a result, there are no established methods to assess the quality of the different parts of G. Park et al. Applied Biological Chemistry (2024) 67:34 Page 2 of 12 Page 2 of 12 the dried roots are valuable as food ingredients and medicinal materials for healthy foods. For examination, the roots are often added to soups, porridge, medicinal wines, and teas [1]. To date, 186 chemical constituents have been identified for G. littoralis, including terpe- noids, organic acids, phenylpropanoids, coumarins, lig- nans, flavonoids, steroids, volatile oils, polysaccharides, and polyols [1, 4]. As a result, this herb exhibits a range of pharmaceutical properties, including antibacterial, anti- fungal, anti-inflammatory, antioxidant, antitumor, anal- gesic, hepatoprotective, and immunoregulatory effects [4, 5]. Among the identified compounds present in G. lit- toralis, 69 major components were identified in the root, including the coumarins psoralen, imperatorin, isoimper- atorin, and bergapten. Research into the pharmacological effects of these coumarin components has revealed their anti-carcinogenic effects on liver and gastric cancer cells, as well as anti-inflammatory properties [4]. In addition, flavonoids, including quercetin, isoquercetin, rutin, chlo- rogenic acid, and caffeic acid have been identified in the roots of G. littoralis as major antioxidants [6]. analytical method is developed by applying the optimal solvent conditions to an HPLC–DAD (diode array detec- tion) system. To validate this method, the specificity, lin- earity, limit of detection (LOD), limit of quantification (LOQ), precision, accuracy, and quantification range are evaluated. Furthermore, the phenolic content is quanti- fied based on the developed method, considering the various processing methods and extraction solvents used to obtain the G. littoralis aerial extract. Finally, the anti- oxidant properties of the individual phenolic compounds present in the aerial parts of the plant are determined using an online HPLC-ABTS (2,2′-azino-bis(3-ethylben- zthiazoline-6-sulfonic acid)) method. Chemicals and preparation of the standard solutionsl ( ) littoralis and the qual- ity of any processed foods containing this herb. The 16 standard compounds, including chlorogenic acid, caffeic acid, scopoletin, rutin, hyperoside, iso- quercetin, nicotiflorin, psoralen, xanthotoxin, isop- impinellin, bergapten, oxypeucedanin, imperatorin, phellopterin, isoimperatorin, and bergamottin were individually weighed (1 mg each) and dissolved in pure methanol (1 mL) to prepare the standard stock solutions. Subsequently, working solutions were produced from the standard stock solutions by dilution with water until reaching the appropriate concentrations for construc- tion of the calibration curves. All solutions were stored at − 20 °C until required for further use. LC–MS LC–MS analysis was done by using Liquid Chroma- tography/Mass Spectrometer at Gyeongnam Bio and Anti-aging Core Facility Center. To identify the 16 phe- nolic compounds, a LC–MS in positive ESI mode was employed. The ESI parameters were set as follows: cap- illary temperature = 350 °C, nebulizer gas pressure = 35 psi, drying gas flow = 9 L min−1 (N2), and capillary volt- age = 4000 V. Full-scan MS spectra were acquired in the mass-to-charge (m/z) range of 150–1500. Data analysis was performed using the Agilent Chem Station software. A Luna 5U C18 column (150 mm × 4.60 mm, 5 μm, Phe- nomenex Inc., USA.) was used for the chromatographic Plant materials and sample preparationh In this study, 16 phenolic compounds present in the G. littoralis aerial parts are identified using liquid chroma- tography–mass spectrometry (LC–MS). Additionally, an The aerial parts and roots of G. littoralis were pur- chased from Wild-Crafted Sources (Gangneung, Page 3 of 12 Page 3 of 12 Park et al. Applied Biological Chemistry (2024) 67:34 Park et al. Applied Biological Chemistry (2024) 67:34 Gangwon-do, Korea), and the materials were harvested from three-year-old plants (Fig. 1A). The samples were lyophilized, ground into fine powders, and stored in a freezer until required for further use. The sample powder (100 mg) was then dissolved in 80% methanol (1 mL) and extracted for 90 min via ultrasonication at room temperature. Each extract solution was sub- sequently subjected to centrifugation at 3500 rpm for 20 min prior to filtration with a 0.45 μm syringe filter. The analytical method was optimized and the 16 phe- nolic compounds present in G. littoralis were quanti- fied using these prepared samples. LC–MS LOD = 3.3 × δ/C Thermal treatment on the G. littoralis aerial partsh The fresh G. littoralis aerial parts were subjected to ther- mal treatment as follows. Shade drying was carried out under a natural airflow and at ~ 25 °C for 2–3 d. For oven drying, the aerial parts were dried in trays in a venti- lated oven at 70 °C for 8 h. Roasting conditions were also examined by heating at 110 °C for 20 min without burn- ing. To steam the materials, the aerial part was added into a steaming basket and placed in a large pot with an integral stainless-steel strainer for 40 min at 73–76 °C. After the completion of each thermal treatment, the materials were allowed to cool to room temperature for 30 min. Subsequently, they were frozen at − 80 °C for 24 h and then lyophilized at − 120 °C for 3 d. The freeze-dried samples were ground into fine powders, and the phenolic compounds present in each sample were quantified on a dry weight basis. Fig. 1 A Photographic images of the aerial part (left) and root (right) from G. littoralis, and B representative chromatograms analyzed using different mobile phases. 1: Chlorogenic acid, 2: caffeic acid, 3: scopoletin, 4: rutin, 5: hyperoside, 6: isoquercetin, 7: nicotiflorin, 8: psoralen, 9: xanthotoxin, 10: isopimpinellin, 11: bergapten, 12: oxypeucedanin, 13: imperatorin, 14: phellopterin, 15: isoimperatorin, and 16: bergamottin Fig. 1 A Photographic images of the aerial part (left) and root (right) from G. littoralis, and B representative chromatograms analyzed using different mobile phases. 1: Chlorogenic acid, 2: caffeic acid, 3: scopoletin, 4: rutin, 5: hyperoside, 6: isoquercetin, 7: nicotiflorin, 8: psoralen, 9: xanthotoxin, 10: Fig. 1 A Photographic images of the aerial part (left) and root (right) from G. littoralis, and B representative chromatograms analyzed using different mobile phases. 1: Chlorogenic acid, 2: caffeic acid, 3: scopoletin, 4: rutin, 5: hyperoside, 6: isoquercetin, 7: nicotiflorin, 8: psoralen, 9: xanthotoxin, 10: isopimpinellin, 11: bergapten, 12: oxypeucedanin, 13: imperatorin, 14: phellopterin, 15: isoimperatorin, and 16: bergamottin Park et al. Applied Biological Chemistry (2024) 67:34 Page 4 of 12 concentrations were used to generate a calibration curve. The following equations were used to calculate the LOD and LOQ based on the standard deviation of the inter- cept (δ) and the slope of the calibration curve (C). separation. Gradient elution was performed using mobile phase A (0.3% formic acid in water) and B (0.3% for- mic acid in acetonitrile). The gradient conditions were as follows: 15% B, increasing to 23% B at 5 min, 25% B at 12 min, 46% B at 15 min, 48% B at 25 min, 50% B at 35 min, 65% B at 45 min, and 90% B at 55 min. The flow rate was adjusted to 0.7 mL min−1, maintaining a column temperature of 35 °C, and a 10 μL injection volume was employed. separation. Gradient elution was performed using mobile phase A (0.3% formic acid in water) and B (0.3% for- mic acid in acetonitrile). The gradient conditions were as follows: 15% B, increasing to 23% B at 5 min, 25% B at 12 min, 46% B at 15 min, 48% B at 25 min, 50% B at 35 min, 65% B at 45 min, and 90% B at 55 min. The flow rate was adjusted to 0.7 mL min−1, maintaining a column temperature of 35 °C, and a 10 μL injection volume was employed. Processing procedure for G. littoralis In preparation for appropriate processing of the G. lit- toralis aerial part, the effects of thermal treatment and different solvent extraction processes were examined to compare the resulting contents of the various phenolic compounds. Method validationh The method used to quantify the 16 phenolic compounds was validated in terms of its linearity, LOD, LOQ, preci- sion, and accuracy. Linearity An external standard calibration method was used. The stock solutions were serially diluted to prepare seven solutions with various concentrations of the phenolic compounds. The ranges of linearity were obtained for the 16 compounds as follows: chlorogenic acid and rutin, 0.5–1000 μg mL−1; isoquercetin and nicotiflorin, 0.5–100 μg mL−1; remaining 12 compounds, 0.25– 50 μg mL−1. Each standard was determined using the developed analytical method, and calibration curves were constructed using linear regression. The linearity of each standard was confirmed by evaluating the correlation coefficient of the calibration curve (R2 ≥ 0.999). HPLC–DAD analysis The precision of the proposed technique was assessed based on the intra- and inter-day precisions. The intraday precision was assessed by calculating the retention time and peak area of each standard solution at three different concentrations within a single day, and is expressed as the relative standard deviation (RSD%). The RSD values of the standard solutions across a three-day period were used to evaluate the inter-day precision. Recovery tests were conducted by spiking the extracts with three dif- ferent concentrations (50, 100, and 250 μg mL−1) of the standard mixture. By calculating the recovery percentage from these tests that were run, the accuracy of the HPLC method was assessed. HPLC–DAD analysis was conducted using an Agilent 1200 series instrument (Agilent Technologies, USA) equipped with a quaternary pump, an ALS autosampler, a DADVL DAD, and a TTC column compartment. The col- umn used was a Hydrosphere C18 column (4.6 × 250 mm, 5 µm) manufactured by YMC (Tokyo, Japan). The oven temperature was held constant at 35 °C with 10 µL injection volume, and the flow rate set to 1 mL min−1. Mobile phase A was composed of water with 0.3% for- mic acid (v/v), while mobile phase B contained acetoni- trile with 0.3% formic acid, and 100% methanol made up mobile phase C. The 16 phenolic compounds were sepa- rated via gradient elution as follows (time, %A/%B/%C): 0 min, 88/12/0%; 8 min, 88/12/0%; 12 min, 82/18/0%; 22 min, 82/18/0%; 26 min, 77/23/0%; 27 min, 50/0/50%; 37 min, 50/0/50%; 56 min, 9/0/91%; 65 min 9/0/91%. All compounds exhibited a maximum absorption band at 254 nm. DAD was employed for method development. Recovery(%) = Detected concentration Nominal concentration × 100 On‑line HPLC‑ABTS radical scavenging activityh The antioxidant activities of the components represent- ing the individual HPLC peaks were evaluated using an online HPLC antioxidant detector system employing the trolox equivalent antioxidant capacity (TEAC) assay, following the method outlined by Koleva et al. [10]. The separation conditions were identical to those described above for the HPLC method. The injection volume was 10 μL, and the detection wavelengths were set at 254 and 734 nm. The mobile phase was mixed with the ABTS solution (5.5 μM, prepared in a 10% solution of MeOH (v/v) in phosphate-buffered saline), which was delivered by a second pump at a flow rate of 0.5 mL min−1. Using a multiwavelength detector (MWD), the reaction products were determined to show a negative peak at 734 nm. Trolox was used as a standard compound to measure the radical scavenging activity, and the Trolox equiva- lent (TE) values were calculated using a calibration curve plotted from 0.25 to 62.58 μg mL−1 (y = 53.101x + 3.9368; R2 = 0.9988). The sum of antioxidant activities of all rel- evant molecules was used to determine the total anti- oxidant activity (TAA). In this case, the antioxidant compounds were chlorogenic acid, caffeic acid, rutin, and isoquercetin. To quantify the radical scavenging capacity of each compound, the TEAC was used, and the results were expressed as the concentration of Trolox (μM) that exhibits the same activity as the compound of interest at a concentration of 1 μM. TEAC = Compound(µM) Trolox(µM) TEAC = Compound(µM) Trolox(µM) LOD and LOQh The LOD and LOQ were calculated to assess the sensitiv- ity of the method. The mean values from three replicated measurements of each standard solution at different Park et al. Applied Biological Chemistry (2024) 67:34 Page 5 of 12 Park et al. Applied Biological Chemistry (2024) 67:34 Optimization of the chromatographic separation process and identification of the phenolic compounds present in the aerial parts of G. littoralis Based on previous literature [1], a total of 16 phenolic compounds, including ten coumarins [scopoletin (3), psoralen (8), xanthotoxin (9), isopimpinellin (10), ber- gapten (11), oxypeucedanin (12), imperatorin (13), phel- lopterin (14), isoimperatorin (15), and bergamottin (16)], four flavonoids [rutin (4), hyperoside (5), isoquercetin (6), and nicotiflorin (7)], and two hydroxycinnamic acids [chlorogenic acid (1) and caffeic acid (2)] were selected as the major phenolic compounds found in the G. littora- lis aerial part. To establish the optimal HPLC conditions for these compounds, LC–MS analysis was conducted on the aerial part extract obtained using 80% methanol. As a result, 16 major peaks were obtained and subsequently analyzed using MS. For each detected compound, the mass value of the [M+H]+ ion was confirmed, and the compound was identified by referencing the reported compounds present in G. littoralis. The absorption wave- lengths and mass values of these compounds are listed in Additional file 1: Table S1, and their chemical structures are shown in Additional file 1: Fig. S1. It was confirmed that the major MS peaks of representative compounds from G. littoralis appeared with high intensity and exactly matched the [M+H]+ values (Additional file 1: Fig. S2). The mobile phase elution conditions were then opti- mized to separate the 16 phenolic compounds from the 80% methanol extract. The use of acetonitrile and metha- nol individually did not provide efficient separation. For example, when acetonitrile was employed, peaks 10 and 11 co-eluted as a single peak. Furthermore, methanol alone did not adequately separate peaks 4, 5, and 6. As a result, the elution conditions were adjusted for the sepa- ration of peaks 1–6 using an acetonitrile/water system, and subsequently, a methanol/water system was used as the elution solvent for the separation of peaks 8–16. The chromatograms obtained under these elution conditions are shown in Fig. 1B, while the optimal analytical condi- tions are listed in Table 1. Optimization of the extraction solvent p < 0.05. The results, obtained from three biological repli- cates, are presented as the mean ± standard deviation. Dried powders (60 g) from the G. littoralis aerial parts and roots were combined with the desired extraction solvent (95% ethanol or water, 600 mL). The mixtures were incubated in round-bottomed flasks at 80 °C for 3 h. Subsequently, each extract was filtered using No. 2 filter paper (8 μm, Whatman, Maidstone, UK) and con- centrated at 40 °C for 3 h using a rotary evaporator. Sub- sequently, a sample of each extract (10 mg) was dissolved in 80% methanol (1 mL) and quantification of phenolic compounds was carried out. Statistical analysis littoralis using a combina- tion of HPLC and tandem MS [11]. A method based on high-performance liquid chromatography–electrospray ionization–tandem mass spectrometry (HPLC–ESI– MS/MS) has been also established for the simultaneous identification of 15 bioactive compounds, including cou- marins, phenolics, and adenosine, which are present in the G. littoralis roots [12]. This method employs a binary mobile phase consisting of acetonitrile and water in gra- dient mode. The detection and identification of peaks at 310 nm revealed the presence of bergapten, cnidilin, imperatorin, isoimperatorin, psoralen, and xanthotoxin. In addition, using HPLC coupled with photodiode array detection at 254 nm, a water/methanol mobile phase suc- cessfully separated seven specific bioactive compounds Statistical analysis littoralis using a combina- tion of HPLC and tandem MS [11]. A method based on high-performance liquid chromatography–electrospray ionization–tandem mass spectrometry (HPLC–ESI– MS/MS) has been also established for the simultaneous identification of 15 bioactive compounds, including cou- marins, phenolics, and adenosine, which are present in the G. littoralis roots [12]. This method employs a binary mobile phase consisting of acetonitrile and water in gra- dient mode. The detection and identification of peaks at 310 nm revealed the presence of bergapten, cnidilin, imperatorin, isoimperatorin, psoralen, and xanthotoxin. In addition, using HPLC coupled with photodiode array detection at 254 nm, a water/methanol mobile phase suc- cessfully separated seven specific bioactive compounds from the G. littoralis roots [13]. However, little research has been carried out into the edible potential of the aerial parts, and as a result, there are no defined procedures for assessing the quality of any samples obtained from these parts. In the simultaneous analysis of multiple ana- lytes, the co-elution of isomers and analytes with simi- lar physicochemical properties typically results in poor selectivities and/or longer analysis times when DAD or photodiode array detection are employed [14]. It is there- fore essential to optimize and resolve these coelution problems, potentially through the use of three different mobile phases. However, to date, this has received little research attention for the development of unique separa- tion methods. As one example, a method was established for the simultaneous separation of 12 impurities through gradient elution using three mobile phases, namely a disodium hydrogen phosphate solution, acetonitrile, and methanol [15]. With this in mind, a water/acetonitrile/ methanol system was examined here in, and it was found that the good separation and precise quantification of 16 phenolic compounds could be achieved using extracts obtained from the G. littoralis aerial parts. the flavonoids, which are typically present in the form of glycosides or carbohydrates, primarily accumulate in the aerial parts [4]. Growing knowledge of the biological and pharmacological effects of bioactive compounds has spurred the development of various analytical approaches for their qualitative and quantitative evaluation. Indeed, several chromatographic techniques have been used to analyze the phenolic compounds present in G. littoralis. To date, numerous analytical methods have been estab- lished for the quantitative examination of the bioactive compounds present in G. littoralis roots. As an example, 10 coumarins, four phenolic acids, and adenosine were detected in the root of G. Statistical analysis Various bioactive compounds, including coumarins, flavonoids, lignanoids, polyacetylenes, and organic acids, have been identified in different organs of G. littora- lis. Among these, the coumarins, which are character- ized by a structure containing two fused six-membered rings (one benzene ring and the other containing an alk- ene functionality and an ester group) bearing a phenyl, alkoxyl, hydroxyl, or isopentenyl substituent, are promi- nent components of the G. littoralis root. In contrast, All statistical analyses were performed using the SPSS software (version 26.0; SPSS Inc., Chicago, IL, USA). Mean comparisons of the G. littoralis organs and extrac- tion solvents were performed using the t-test at a level of p < 0.05. To evaluate any significant differences based on the different thermal treatment approaches, a one-way analysis of variance (ANOVA) was conducted, followed by Tukey’s multiple range test at a significance level of Page 6 of 12 Park et al. Applied Biological Chemistry (2024) 67:34 Table 1 Proposed analytical method for determination of 16 phenolic compounds in G. littoralis Table 1 Proposed analytical method for determination of 16 phenolic compounds in G. littoralis HPLC unit Agilent 1260 series Column YMC Hydrosphere C18 (4.6 × 250 mm, i.d., 5 µm particle size) Detector Diode array detector (DAD) Detection wavelength (nm) 254 nm Flow rate (mL min−1) 1.0 mL min−1 Injection volume 10 µL Column temperature 35 ℃ Mobile phase A: 0.3% formic acid in Water B: 0.3% formic acid in Acetonitrile C: Methanol Time (min) Mobile phase A (%) Mobile phase B (%) Mobile phase C (%) HPLC gradient profile 0 88 12 0 8 88 12 0 12 82 18 0 22 82 18 0 26 77 23 0 27 50 0 50 37 50 0 50 56 9 0 91 65 9 0 91 Agilent 1260 series Mobile phase A (%) the flavonoids, which are typically present in the form of glycosides or carbohydrates, primarily accumulate in the aerial parts [4]. Growing knowledge of the biological and pharmacological effects of bioactive compounds has spurred the development of various analytical approaches for their qualitative and quantitative evaluation. Indeed, several chromatographic techniques have been used to analyze the phenolic compounds present in G. littoralis. To date, numerous analytical methods have been estab- lished for the quantitative examination of the bioactive compounds present in G. littoralis roots. As an example, 10 coumarins, four phenolic acids, and adenosine were detected in the root of G. Quantification of the different G. littoralis samples using the proposed HPLC methodh Method precision was evaluated through both intra- and inter-day experiments. The intraday precision was evaluated by performing three repetitive injections of the standard solutions containing 16 phenolic compounds in a single day, whereas the three consecutive days were used to evaluate inter-day accuracy. The intra- and inter- day precisions of the peak areas obtained for these 16 compounds were calculated and expressed as RSDs (Additional file 1: Table S2), giving values of 1.38–6.06% and 1.23–8.24%, respectively. The accuracy was deter- mined by expressing the mean concentration of each quantified sample as a percentage of the nominal concen- tration, resulting in values of 90.35–116.40%. The preci- sion and accuracy of the method were therefore clearly within the acceptable limits of a 15% RSD (20% close to the LOQ) [17]. The method recovery was determined The phenolic compounds present in the G. littoralis aer- ial parts and roots were quantified using the analytical method described herein. As shown in Table 3, the total phenolic contents were 18.32 and 0.19 mg g−1 dry weight (DW) in the G. littoralis aerial parts and roots, respec- tively. In the aerial parts, the majority of target coumarins detected were quantified in large amounts, whereas only four coumarins were detected in the roots, and these were present in very small quantities. Among the 10 cou- marins of interest, psoralen (8) and bergamottin (16) were not detected in either the aerial parts or the roots. The highest flavonoid content was found in the aerial parts of the plant, and was determined to be 13.16 mg g−1 DW; among these, rutin (4) was the most abundant (i.e., 10.01 mg g−1 DW), accounting for more than half of total Table 2 Results for the parameters of method validation of simultaneous determination for 16 phenolic compounds in G. Method validationh The established HPLC method was validated in terms of its linearity, specificity, precision, accuracy, LOD, and LOQ based on ICH guidelines [16]. Calibration curves Page 7 of 12 Park et al. Applied Biological Chemistry (2024) 67:34 Park et al. Applied Biological Chemistry (2024) 67:34 Park et al. Applied Biological Chemistry (2024) 67:34 for the 16 phenolic compounds were obtained using five serial dilutions of mixed standard solutions, and plot- ting their peak areas (y) against their concentrations (x) to provide linear equations (y = ax + b). These equations, along with the corresponding correlation coefficients (R2), LODs, and LOQs are summarized in Table 2. Excel- lent linearities were shown by all calibration curves, with correlation coefficients (R2) > 0.999, thereby satisfying the criterion of R2 > 0.998 for a satisfactory linearity. For the 16 phenolic compounds examined herein, the LOD ranged from 0.31 to 1.93 μg mL−1, while the LOQ ranged from 0.95 to 5.86 μg mL−1, indicating the high sensitiv- ity of the method for detecting low concentrations of the analytes. by the addition of specific concentrations of the various standards, followed by subsequent sample recovery. For this purpose, 16 standard solutions were added to the G. littoralis extracts, and the average recoveries were calcu- lated by comparing the detected quantity with the added quantity (Additional file 1: Table S3). As a result, the recoveries were determined to be 86.99–109.26%, with RSDs of 0.07–4.89%, thereby indicating that the devel- oped method was accurate within the specified criteria. The above validation results confirm the reliability of the optimized analytical method for the simultaneous analy- sis of 16 phenolic compounds in G. littoralis. Quantification of the different G. littoralis samples using the proposed HPLC methodh Significant differences were assessed by *(P < 0.001), (P < 0.01), (P < 0.05) 1 n.d., not detected Classification Peak Compound Aerial Root Coumarin 3 Scopoletin 0.01 ± 0.00* 0.01 ± 0.00 8 Psoralen n.d.1 n.d 9 Xanthotoxin 0.05 ± 0.00* 0.04 ± 0.00 10 Isopimpinellin 0.16 ± 0.02 n.d 11 Bergapten 0.14 ± 0.02 n.d 12 Oxypeucedanin 0.02 ± 0.00* n.d 13 Imperatorin 0.41 ± 0.01* 0.02 ± 0.00 14 Phellopterin 0.03 ± 0.01* n.d 15 Isoimperatorin 0.09 ± 0.00* 0.02 ± 0.00 16 Bergamottin n.d n.d Total 0.93 ± 0.04* 0.09 ± 0.01 Flavonoid 4 Rutin 10.01 ± 0.09*** n.d 5 Hyperoside 0.02 ± 0.01* n.d 6 Isoquercetin 0.94 ± 0.01* n.d 7 Nicotiflorin 2.19 ± 0.04* n.d Total 13.16 ± 0.11 n.d Hydroxycinnamic acid 1 Chlorogenic acid 4.23 ± 0.02* 0.10 ± 0.01 2 Caffeic acid 0.02 ± 0.00* n.d Total 4.25 ± 0.02* 0.10 ± 0.01 Fig. 2 Coumarin, flavonoid, and hydroxycinnamic acid contents following A different thermal treatment approaches, and B extraction using different solvents. Additional file 1: Tables S4 and S5 indicate the quantification results for the 16 individual phenolic compounds. SD, shade drying; OD, oven drying; RO, roasting; and ST, steaming Table 3 Contents (mg g−1 dry weight) of 16 phenolic compounds in the G. littoralis aerial parts and roots Each data was represented as mean ± standard deviation (n = 3) and statistically analyzed using one-tailed t-test. Significant differences were assessed by (P < 0.001), (P < 0.01), (P < 0.05) Each data was represented as mean ± standard deviation (n = 3) and statistically analyzed using one-tailed t-test. Significant differences were assessed by *(P < 0.001), (P < 0.01), (P < 0.05) 1 n.d., not detected phenolic content. However, no flavonoids were detected in the roots. In terms of the hydroxycinnamic acids, chlo- rogenic acid (1) was detected at levels of 4.23 mg g−1 DW in the aerial parts, while the caffeic acid (2) content was 0.02 mg g−1 DW. In contrast, in the root, no caffeic acid (2) was detected, and the chlorogenic acid (1) levels were extremely low (i.e., 0.10 mg g−1 DW). Thus, among the 16 target compounds, only five were detected in the roots using the analytical method described herein, thereby indicating that this method is suitable for quantitative analysis of the G. littoralis aerial parts. Fig. Quantification of the different G. littoralis samples using the proposed HPLC methodh littoralis by HPLC–DAD Peak Compound Calibration range (μg mL−1) Calibration curve Correlation coefficient (R2) LOD (μg mL−1) LOQ (μg mL−1) 1 Chlorogenic acid 5–1000 y = 33.176x + 22.421 0.9996 1.93 5.86 2 Caffeic acid 2.5–50 y = 54.162x − 11.175 0.9994 0.31 0.95 3 Scopoletin 2.5–50 y = 39.628x − 5.7914 0.9993 0.34 1.04 4 Rutin 5–1000 y = 20.75x − 50.604 0.9999 1.39 4.23 5 Hyperoside 2.5–50 y = 28.517x − 5.9583 0.9992 0.63 1.90 6 Isoquercetin 5–100 y = 45.605x − 21.084 0.9992 1.27 3.85 7 Nicotiflorin 5–100 y = 16.772x − 4.5759 0.9993 1.26 3.82 8 Psoralen 2.5–50 y = 84.526x − 25.319 0.9996 0.63 1.92 9 Xanthotoxin 2.5–50 y = 46.756x − 7.2072 0.9993 0.58 1.76 10 Isopimpinellin 2.5–50 y = 23.361x − 3.2631 0.9993 0.59 1.78 11 Bergapten 2.5–50 y = 47.866x − 12.234 0.9995 0.71 2.15 12 Oxypeucedanin 2.5–50 y = 30.84x − 2.5168 0.9993 0.69 2.08 13 Imperatorin 2.5–50 y = 41.322x − 34.968 0.9997 1.00 3.03 14 Phellopterin 2.5–50 y = 25.084x − 1.9087 0.9993 0.60 1.82 15 Isoimperatorin 2.5–50 y = 24.793x − 3.4156 0.9993 0.62 1.87 16 Bergamottin 2.5–50 y = 29.773x − 5.1171 0.9993 0.59 1.80 parameters of method validation of simultaneous determination for 16 phenolic compounds in G. littoralis by Park et al. Applied Biological Chemistry (2024) 67:34 Page 8 of 12 Table 3 Contents (mg g−1 dry weight) of 16 phenolic compounds in the G. littoralis aerial parts and roots Each data was represented as mean ± standard deviation (n = 3) and statistically analyzed using one-tailed t-test. Determination of the antioxidant activity using on‑line HPLC‑ABTSh Through the use of parallel chemical detection methods, online HPLC-ABTS analyses enable the simultaneous identification and quantification of active compounds [23]. Thus, using this approach, the analysis of a stand- ard mixture of 16 phenolic compounds (Fig. 3A) revealed five components with ABTS radical-scavenging activity, namely chlorogenic acid (1), caffeic acid (2), rutin (4), hyperoside (5), and isoquercetin (6). A representative chromatogram of the 95% ethanol extract from the G. littoralis aerial parts is shown in Fig. 3B, wherein it can be seen that peaks of chlorogenic acid (1), rutin (4), and isoquercetin (6) were clearly visible at 734 nm, suggest- ing a significant ABTS radical-scavenging activity. Fur- thermore, a small peak was detected for caffeic acid (2). According to previous literature, the TEAC values for ascorbic acid, butylated hydroxyl anisole, and butylated hydroxyl toluene were found to be 1.08, 1.02, and 0.47 as representative antioxidant compounds, respectively [24]. In comparison, the TEAC values from G. littoralis was shown that among the four antioxidant compounds detected, rutin exhibited the highest TEAC (2.34), fol- lowed by isoquercetin, chlorogenic acid, and caffeic acid, with values of 1.56, 1.43, and 1.37, respectively (Table 4). g y g [ ] During the extraction of organic compounds from plant materials, the process is largely affected by the diverse compounds present in the plant matrix, the choice of extraction solvent, the active ingredient solubil- ity, and the interactions of these compounds with other solutes. In addition, the concentrations of bioactive natu- ral products in natural sources are typically low, and they are often embedded within the plant matrices. The effec- tiveness of the extraction process therefore also depends on the properties of the plant matrix, and the extraction solvent, temperature, pressure, and duration [21]. Thus, the phenolic compounds present in the extracts of the G. littoralis aerial parts and roots were extracted using the two most common extraction solvents, namely 95% ethanol and water (Fig. 2B). The extracts obtained from G. littoralis roots using these two extraction solvents did not contain coumarins or flavonoids, and only a small amount of hydroxycinnamic acid was detected. How- ever, it appeared that 95% ethanol was more suitable as an extraction solvent, giving 43.82 mg g−1 extract (ex) compared to the 13.73 mg g−1 ex obtained using water for total phenolic compound contents. Quantification of the different G. littoralis samples using the proposed HPLC methodh 2 Coumarin, flavonoid, and hydroxycinnamic acid contents following A different thermal treatment approaches, and B extraction using different solvents. Additional file 1: Tables S4 and S5 indicate the quantification results for the 16 individual phenolic compounds. SD, shade drying; OD, oven drying; RO, roasting; and ST, steaming y Given that the aerial parts (young leaves and stems) of G. littoralis are currently registered as food ingredients by the Korean Food and Drug Administration and hold potential for use as functional ingredients, it is crucial to employ appropriate processing procedures for the fur- ther utilization of this plant. Among the various process- ing procedures available for use with such specimens, thermal treatment was initially considered. Thus, the G. littoralis samples were subjected to a range of thermal conditions, including shade drying, oven drying, roast- ing, and steaming. Upon evaluation of the total phenolic contents after treatment, it was apparent that shade dry- ing led to the highest content (7.74 mg g−1 DW), followed by steaming (6.65 mg g−1 DW), oven drying (3.86 mg g−1 DW), and roasting at (2.27 mg g−1 DW) (see Fig. 2A). Scopoletin (3), psoralen (8), bergamottin (16), and caf- feic acid (2) were not detected after any thermal treat- ment, and the total coumarin levels remained consistent regardless of the thermal treatment approach employed (Additional file 1: Table S4). In contrast, the flavonoid and hydroxycinnamic acid contents varied, and with the exception of nicotiflorin (7), the contents of these com- pounds decreased upon increasing the harshness of the DW), and roasting at (2.27 mg g−1 DW) (see Fig. 2A). Scopoletin (3), psoralen (8), bergamottin (16), and caf- feic acid (2) were not detected after any thermal treat- ment, and the total coumarin levels remained consistent regardless of the thermal treatment approach employed (Additional file 1: Table S4). In contrast, the flavonoid and hydroxycinnamic acid contents varied, and with the exception of nicotiflorin (7), the contents of these com- pounds decreased upon increasing the harshness of the Park et al. Applied Biological Chemistry (2024) 67:34 Page 9 of 12 and this is consistent with previous studies highlighting the significant impact of employing an ethanol/water- based solvent system for phenolic compound extraction from plant materials. For example, a higher the amount of total phenolic compounds was extracted from peanut skin using 80% ethanol instead of water [22]. treatment method (i.e., shade drying > steaming > oven drying > roasting). Quantification of the different G. littoralis samples using the proposed HPLC methodh Thus, roasting led to the most signifi- cant reduction in the phenolic content, whereas shade drying was deemed most the suitable approach for fur- ther investigations. It is well known that different thermal treatment condi- tions can significantly alter the phenolic contents of plant materials owing to the polymerization and degradation of these compounds [18]. More specifically, it has been reported that thermal treatment may lead to an increase in bound phenolic compounds, whilst promoting the degradation of free phenolic compounds [18, 19]. As indicated above, thermal treatments, such as oven dry- ing, roasting, and steaming, negatively affected the levels of rutin (4) and chlorogenic acid (1) in the G. littoralis specimens. This aligns with previous research indicating that the levels of rutin and chlorogenic acid in avocado leaves decrease to varying degrees under different drying methods, including oven drying at 50 and 100 °C [20]. Determination of the antioxidant activity using on‑line HPLC‑ABTSh DAD, Diode array detector; MWD, Multiple wavelength detector Fig. 3 Chromatographic fingerprints and ABTS inhibition profiles for A the standards containing the 16 phenolic compounds, and B the 95% ethanol extract of the G. littoralis aerial parts. Peak detection was carried out at 734 nm (red peak, negative) for the antioxidant activity and at 330 nm (blue, positive) for quantification. 1: Chlorogenic acid, 2: caffeic acid, 3: scopoletin, 4: rutin, 5: hyperoside, 6: isoquercetin, 7: nicotiflorin, 8: psoralen, 9: xanthotoxin, 10: isopimpinellin, 11: bergapten, 12: oxypeucedanin, 13: imperatorin, 14: phellopterin, 15: isoimperatorin, and 16: bergamottin. DAD, Diode array detector; MWD, Multiple wavelength detector Table 4 TE (mg g−1 ex) and TEAC values of major antioxidant compounds in G. littoralis extracts and their antioxidant capacities All value was calculated using Trolox calibration curve (y = 53.101x + 3.9368, R2 = 0.9988) and represented as mean ± SD (n = 3). Each data was represented as mean ± standard deviation (n = 3) and statistically analyzed using one-tailed t-test. significant differences were assessed by (P < 0.001), (P < 0.01), (P < 0.05) 1 TE, Trolox equivalent 2 TAA, Percentage of the total antioxidant activity of all identified compounds 3 TEAC, Trolox equivalent antioxidant capacity, concentration of Trolox (μM) that exhibits the same activity as 1 μM each identified compound 4 n.d., not detected Peak Compound TEAC3 TE from aerial part TE from root 95% EtOH extract Water extract 95% EtOH extract Water extract 1 Chlorogenic acid 1.43 4.34 ± 0.91 1.44 ± 0.20 0.45 ± 0.01 * 0.01 ± 0.00 2 Caffeic acid 1.37 0.15 ± 0.01 0.14 ± 0.02 0.19 ± 0.03 0.14 ± 0.01 4 Rutin 2.34 3.86 ± 0.17 * 1.51 ± 0.04 n.d.4 n.d 6 Isoquercetin 1.56 1.07 ± 0.08 0.31 ± 0.01 n.d n.d Total TE1 9.36 ± 1.20 3.39 ± 0.09 0.65 ± 0.03 0.15 ± 0.00 TAA (%)2 74.8 56.9 48.6 64.1 offering the advantage of real-time measurement of the antioxidant activity of a compound immediately after its constituent analysis [25]. Accordingly, the cur- rent findings confirmed that four compounds of the in G. littoralis specimen exhibited antioxidant activity, namely chlorogenic acid (1), caffeic acid (2), rutin (4), and isoquercetin (6). Additionally, the aerial parts have higher antioxidant activity than the roots. Determination of the antioxidant activity using on‑line HPLC‑ABTSh In addition, the 95% ethanol extract contained various coumarins, includ- ing isopimpinellin (10), bergapten (11), imperatorin (13), and isoimperatorin (15); no such coumarins were detected in the water extract (Additional file 1: Table S5). Furthermore, the majority of flavonoids and hydroxycin- namic acids were present in higher quantities in the 95% ethanol extract than in the water extract. Among these, rutin (4) represented the most abundant phenolic com- pound, followed by chlorogenic acid (1), nicotiflorin (7), and isoquercetin (6). These findings therefore demon- strate the successful enrichment of phenolic compounds in the 95% ethanol extract obtained from the aerial parts, Furthermore, the radical-scavenging activities of the G. littoralis extracts from the aerial parts and the roots were measured. The concentration of each antioxidant compound was determined using an appropriate stand- ard reference and the antioxidant activity was calculated based on the TE values. More specifically, for the four antioxidant compounds, the TE was highest in the aerial parts extracted with 95% ethanol (9.36 ± 1.20 mg TE g−1 ex), followed by the water extract of the aerial parts (3.39 ± 0.09 mg TE g−1 ex), the ethanol extract of the roots (0.65 ± 0.03 mg TE g−1 ex), and the water extract of the roots (0.15 ± 0.00 mg TE g−1 ex). The antioxidant effects of the total antioxidants obtained from each extract were also measured, and the highest TAA was found for the 95% ethanol extract of the aerial parts. These results suggest that the G. littoralis aerial parts exhibit a higher antioxidant activity than the roots, and that the ethanol extract exhibited a superior antioxidant potential than the water extract. The online HPLC-ABTS approach is known to com- plement the conventional ABTS method, whilst also Park et al. Applied Biological Chemistry (2024) 67:34 Park et al. Applied Biological Chemistry Page 10 of 12 Fig. 3 Chromatographic fingerprints and ABTS inhibition profiles for A the standards containing the 16 phenolic compounds, and B the 95% ethanol extract of the G. littoralis aerial parts. Peak detection was carried out at 734 nm (red peak, negative) for the antioxidant activity and at 330 nm (blue, positive) for quantification. 1: Chlorogenic acid, 2: caffeic acid, 3: scopoletin, 4: rutin, 5: hyperoside, 6: isoquercetin, 7: nicotiflorin, 8: psoralen, 9: xanthotoxin, 10: isopimpinellin, 11: bergapten, 12: oxypeucedanin, 13: imperatorin, 14: phellopterin, 15: isoimperatorin, and 16: bergamottin. Supplementary Information The online version contains supplementary material available at https://doi. org/10.1186/s13765-024-00884-9. The online version contains supplementary material available at https://doi. org/10.1186/s13765-024-00884-9. In conclusion, the establishment of an appropriate analytical method is required to set specific standards to facilitate the industrial application of the edible aerial parts of the G. littoralis plant. Using liquid chromatog- raphy–mass spectrometry (LC-MS), 16 major phenolic compounds were identified in the aerial parts of G. litto- ralis, namely 10 coumarins, 4 flavonoids, and 2 hydrox- ycinnamic acids. Subsequently, an optimal analytical method was established using a combination of three dif- ferent mobile phases (water, acetonitrile, and methanol) to simultaneously separate the 16 phenolic compounds. The reliability of the developed method was validated using various parameters. Importantly, quantification of the aerial parts and roots of G. littoralis using the pro- posed analytical method showed that the majority of phenolic compounds were more abundant in the aerial parts than in the roots. Thus, to determine an appropri- ate processing procedure for the edible G. littoralis aerial parts to yield functional ingredients for foods, thermal treatment and solvent extraction were investigated. It was found that when the aerial parts were subjected to ther- mal treatment, roasting led to the most notable decline in phenolic compounds, whereas shade drying retained the greatest quantities of active compounds. Furthermore, extraction with 95% ethanol was found to be favorable, giving an extract enriched with phenolic compounds. The proposed analytical method was then applied to an online HPLC-ABTS assay to assess the antioxidant activities of the 16 phenolic compounds. As a result, four antioxidant components were determinded, namely chlo- rogenic acid, caffeic acid, rutin, and isoquercetin. Moreo- ver, the 95% ethanol extract exhibited the highest total Trolox equivalent value and the highest total antioxidant ability, thereby indicating its superior antioxidant proper- ties compared to the root and water extracts. Additional file 1: Fig. S1. Chemical structures of targeted compounds in G. littoralis. Fig. S2. Positive ion ESI/MS spectrum of representative com‑ pounds including chlorogenic acid (1), caffeic acid (2), rutin (4), nicotiflorin (7), isopimpinellin (10), and isoimperatorin (15) in G. littoralis. Table S1. ESI/MS data of 16 phenolic compounds in G. littoralis. Table S2. Results of accuracy and precision of 16 phenolic compounds in G. littoralis by HPLC– DAD. Table S3. Results of recovery test of 16 phenolic compounds in G. littoralis by HPLC–DAD method. Table S4. Determination of the antioxidant activity using on‑line HPLC‑ABTSh Similarly, a radical scavenging assay identified caffeic acid, chlo- rogenic acid, isoquercetin, quercetin and rutin as the principal antioxidant constituents of this plant [6]. It should be noted here that the ABTS radical-scavenging Park et al. Applied Biological Chemistry (2024) 67:34 Page 11 of 12 potentials of various G. littoralis extracts, including the leaves, stems, fruits, whole plants, and roots, have been previously evaluated [26]. Availability of data and materials The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request. Supplementary Information Contents (mg g−1 dry weight) of 16 phenolic compounds dissolved in different extraction solvents from aerial part and root of G. littoralis. Table S5. Contents (mg g−1 dry weight) of 16 phenolic compounds depending on different thermal treatments of G. littoralis. Author contributions YJP and YBC visualized and analyzed data. SBO investigated experiments. YJP, YBC, JYM, TQT, and PKH wrote the original manuscript. YJP and SMK edited the manuscript. SMK conceptualized the research and supervised all processes. All authors read and approved the final manuscript. Funding This study was supported by Ministry of Agriculture, Food and Rural Affairs (MAFRA), Ministry of Science and ICT (MSIT), and Rural Development Adminis‑ tration (RDA) (421034-04). Acknowledgements This work was supported by the Korea Institute of Planning and Evaluation for Technology in Food, Agriculture and Forestry (IPET) and Korea Smart Farm R&D Foundation (KosFarm) through the Smart Farm Innovation Technology Development Program, funded by Ministry of Agriculture, Food and Rural Affairs (MAFRA), Ministry of Science and ICT (MSIT), and Rural Development Administration (RDA) (421034-04). 5. Choe S-Y, Hong J-H, Gu Y-R, Kim I-D, Dhungana S, Moon K-D (2019) Hot water extract of Glehnia littoralis leaf showed skin-whitening and anti- wrinkle properties. S Afr J Bot 127:104–109 Competing interests Th h d l h The authors declare that they have no competing interests. Received: 17 December 2023 Accepted: 3 March 2024 Received: 17 December 2023 Accepted: 3 March 2024 Abbreviations ABTS 2.2ʹ-Azino-bis(3-ethylbenzthiazoline-6-sulfonic acid ANOVA One-way analysis of variance DAD Diode-array detection DW Dry weight ESI Electrospray ionization HPLC High-performance liquid chromatograph LC–MS Liquid chromatography-mass spectrometry LOD Limit of detection LOQ Limit of quantification MWD Multi-wavelength detector RSD Relative standard deviations TAA Total antioxidant activity TE Trolox equivalent TEAC Trolox equivalent antioxidant capacity Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub‑ lished maps and institutional affiliations. 7. Hong SJ, Park NI, Hwang DK, Yi TG, Eum HL (2022) Comparison of yield and metabolites according to the types of hilling materials utilized dur‑ ing Glehnia littoralis sprout vegetable cultivation. Food Sci Biotechnol 31:669–679 8. Kong C-S, Um YR, Lee JI, Kim YA, Yea SS, Seo Y (2010) Constituents isolated from Glehnia littoralis suppress proliferations of human cancer cells and MMP expression in HT1080 cells. Food Chem 120:385–394 9. Lee JW, Lee C, Jin Q, Yeon ET, Lee D, Kim S-Y, Han SB, Hong JT, Lee MK, Hwang BY (2014) Pyranocoumarins from Glehnia littoralis inhibit the LPS- induced NO production in macrophage RAW 264.7 cells. Bioorganic Med Chem Lett 24:2717–2719 10. Koleva II, Niederländer HA, van Beek TA (2001) Application of ABTS radical cation for selective on-line detection of radical scavengers in HPLC elu‑ ates. Anal Chem 73:3373–3381 11. Yang W, Ye M, Liu M, Kong D, Shi R, Shi X, Zhang K, Wang Q, Lantong Z (2010) A practical strategy for the characterization of coumarins in Radix Glehniae by liquid chromatography coupled with triple quadrupole- linear ion trap mass spectrometry. J Chromatogr A 1217:4587–4600 12. Yang W, Feng C, Kong D, Shi X, Zheng X, Cui Y, Liu M, Zhang L, Wang Q (2010) Simultaneous determination of 15 components in Radix Glehniae by high performance liquid chromatography–electrospray ionization tandem mass spectrometry. Food Chem 120:886–894 13. Seo UM, Zhao BT, Kim YH, Kang JS, Son JK, Woo M-H (2016) Simultane‑ ous analysis of seven marker compounds from Saposhnikoviae Radix, Glehniae Radix and Peucedani Japonici Radix by HPLC/PDA. Arch Pharm Res 39:695–704 14. Zhao XJ, Guo PM, Pang WH, Zhang YH, Zhao QY, Jiao BN, Kilmartin PA (2020) A rapid UHPLC-QqQ-MS/MS method for the simultaneous qualita‑ tion and quantitation of coumarins, furocoumarins, flavonoids, phenolic acids in pummelo fruits. Food Chem 325:126835 15. Chen D, Shen D, Huang L, Hou L, Xiong Y, Zeng L, Wang J (2022) Develop‑ ment of a novel and stable indicating RP-HPLC method for the simultane‑ ous analysis of 12 impurities in midazolam and midazolam injection products. Chromatographia 85:949–958 16. Borman P, Elder D (2017) Q2(R1) Validation of analytical procedures: text and methodology. ICH quality guidelines: an implementation guide. Wiley, New York, pp 127–166 17. Peters FT, Drummer OH, Musshoff F (2007) Validation of new methods. References 1. Yang M, Li X, Zhang L, Wang C, Ji M, Xu J, Zhang K, Liu J, Zhang C, Li M (2019) Ethnopharmacology, phytochemistry, and pharmacology of the genus glehnia: a systematic review. Evid Based Complement Alternat Med. https://doi.org/10.1155/2019/1253493 1. Yang M, Li X, Zhang L, Wang C, Ji M, Xu J, Zhang K, Liu J, Zhang C, Li M (2019) Ethnopharmacology, phytochemistry, and pharmacology of the genus glehnia: a systematic review. Evid Based Complement Alternat Med. https://doi.org/10.1155/2019/1253493 2. Shao C, Wang G, Ding X, Yang C, Yan M (2021) Physiological and bio‑ chemical characteristics of cold stratification to overcome morphophysi‑ ological dormancy in Glehnia littoralis seed. Seed Sci Technol 49:19–24 3 Y T Ch B Ch M L D Ch i G L A Ki H (2008) Ph 4. Li S, Xu N, Fang Q, Cheng X, Chen J, Liu P, Li L, Wang C, Liu W (2023) Glehnia littoralis Fr. Schmidtex Miq.: a systematic review on ethnophar‑ macology, chemical composition, pharmacology and quality control. J Ethnopharmacol 317:116831 5. Choe S-Y, Hong J-H, Gu Y-R, Kim I-D, Dhungana S, Moon K-D (2019) Hot water extract of Glehnia littoralis leaf showed skin-whitening and anti- wrinkle properties. S Afr J Bot 127:104–109 Page 12 of 12 Park et al. Applied Biological Chemistry (2024) 67:34 Park et al. Applied Biological Chemistry (2024) 67:34 6. Yuan Z, Tezuka Y, Fan W, Kadota S, Li X (2002) Constituents of the under‑ ground parts of Glehnia littoralis. Chem Pharm Bull 50:73–77 Springer Nature remains neutral with regard to jurisdictional claims in pub‑ lished maps and institutional affiliations. Publisher’s Note Forensic Sci Int 165:216–224 18. Li M, Chen X, Deng J, Ouyang D, Wang D, Liang Y, Chen Y, Sun Y (2020) Effect of thermal processing on free and bound phenolic compounds and antioxidant activities of hawthorn. Food Chem 332:127429 19. Ngamsuk S, Huang T-C, Hsu J-L (2019) Determination of phenolic compounds, procyanidins, and antioxidant activity in processed Coffea arabica L. leaves. Foods 8:389 20. Loh ZH, Lim YY (2018) Drying effects on antioxidant activity, enzyme activity, and phytochemicals of avocado (Persea americana) leaves. J Food Process Preserv 42:e13667 21. Jha AK, Sit N (2022) Extraction of bioactive compounds from plant mate‑ rials using combination of various novel methods: a review. Trends Food Sci Technol 119:579–591 22. Yu J, Ahmedna M, Goktepe I (2005) Effects of processing methods and extraction solvents on concentration and antioxidant activity of peanut skin phenolics. Food Chem 90:199–206 23. Karaçelik AA, Küçük M, Iskefiyeli Z, Aydemir S, De Smet S, Miserez B, Sandra P (2015) Antioxidant components of Viburnum opulus L. deter‑ mined by on-line HPLC–UV–ABTS radical scavenging and LC–UV–ESI-MS methods. Food Chem 175:106–114 24. Çelik SE, Özyürek M, Güçlü K, Apak R (2010) Solvent effects on the anti‑ oxidant capacity of lipophilic and hydrophilic antioxidants measured by CUPRAC, ABTS/persulphate and FRAP methods. Talanta 81:1300–1309 25. Lee HJ, Pan C-H, Kim E-S, Kim CY (2012) Online high performance liquid chromatography (HPLC)-ABTS+ based assay and HPLC-electrospray ioni‑ zation mass spectrometry analysis of antioxidant phenolic compounds in Salsola komarovii. J Korean Soc Appl Biol Chem 55:317–321 26. Kim CJ, Ghimire BK, Choi SK, Yu CY, Lee JG (2023) Sustainable bioactive composite of Glehnia littoralis extracts for osteoblast differentiation and bone formation. Processes 11:1491
https://openalex.org/W3211847913	https://www.frontiersin.org/articles/10.3389/fmed.2021.681172/pdf	English	null	Psoriasis to Psoriatic Arthritis: The Application of Proteomics Technologies	Frontiers in medicine	2,021	cc-by	7,185	Edited by: Edited by: Saumya Panda, KPC Medical College and Hospital, India Reviewed by: Unni Samavedam, University of Cincinnati, United States Philippe Lefrançois, McGill University, Canada Correspondence: Yanling He heyanlingdermatology@gmail.com †These authors have contributed equally to this work Reviewed by: Unni Samavedam, University of Cincinnati, United States Philippe Lefrançois, McGill University, Canada Reviewed by: Unni Samavedam, University of Cincinnati, United States Philippe Lefrançois, McGill University, Canada Keywords: psoriasis, psoriatic disease, psoriatic arthritis, proteomics, biomarkers MINI REVIEW MINI REVIEW published: 16 November 2021 doi: 10.3389/fmed.2021.681172 published: 16 November 2021 doi: 10.3389/fmed.2021.681172 Psoriasis to Psoriatic Arthritis: The Application of Proteomics Technologies Fei Qi †, Yaqi Tan †, Amin Yao, Xutong Yang and Yanling He Department of Dermatology, Capital Medical University Afﬁliated Beijing Chaoyang Hospital, Beijing, China Psoriatic disease (PsD) is a spectrum of diseases that affect both skin [cutaneous psoriasis (PsC)] and musculoskeletal features [psoriatic arthritis (PsA)]. A considerable number of patients with PsC have asymptomatic synovio-entheseal inﬂammations, and approximately one-third of those eventually progress to PsA with an enigmatic mechanism. Published studies have shown that early interventions to the very early-stage PsA would effectively prevent substantial bone destructions or deformities, suggesting an unmet goal for exploring early PsA biomarkers. The emergence of proteomics technologies brings a complete view of all involved proteins in PsA transitions, offers a unique chance to map all potential peptides, and allows a direct head-to-head comparison of interaction pathways in PsC and PsA. This review summarized the latest development of proteomics technologies, highlighted its application in PsA biomarker discovery, and discussed the possible clinical detectable PsA risk factors in patients with PsC. INTRODUCTION Psoriatic disease (PsD), as an umbrella term, describes a systemic inﬂammatory disease that predominantly aﬀects the skin [cutaneous psoriasis (PsC)] and musculoskeletal features [psoriatic arthritis (PsA)], with ∼125 million patients worldwide (1, 2). The concept of PsD indicates the realization of the common inﬂammatory and metabolic pathways working on the skin and synovium (3). Although it is still controversial whether PsC and PsA shared the same immunological factors or belonged to the same spectrum of diseases, studies from genetic and proteomics conﬁrmed the overlap between PsC and PsA (4–8). †These authors have contributed equally to this work Specialty section: This article was submitted to Dermatology, a section of the journal Frontiers in Medicine Received: 16 March 2021 Accepted: 18 October 2021 Published: 16 November 2021 Specialty section: This article was submitted to Dermatology, a section of the journal Frontiers in Medicine Psoriatic arthritis is characterized by multiple joints stiﬀness, pain, and swelling with insidious onset (1, 9). Poor prognosis with debilitating joint destruction brings a tremendously negative impact on the life quality of all patients (10). It aﬀects one in ﬁve people who have a psoriasis diagnosis, while only 15% of PsA cases get cutaneous lesions after arthritis onset (11, 12). After the initiation of psoriasis, the prevalence of PsA grows over time, hitting 20% after 30 years (13, 14). It is signiﬁcant to identify patients who are at risk for PsA and enable targeting therapies to prevent and intercept the joint involvement at a very early stage of the psoriatic arthropathy (15, 16). A 6-month delay in joint destruction detection is linked to a signiﬁcantly lower treatment response (17). Received: 16 March 2021 Accepted: 18 October 2021 Published: 16 November 2021 Keywords: psoriasis, psoriatic disease, psoriatic arthritis, proteomics, biomarkers Citation: Qi F, Tan Y, Yao A, Yang X and He Y (2021) Psoriasis to Psoriatic Arthritis: The Application of Proteomics Technologies. Front. Med. 8:681172. doi: 10.3389/fmed.2021.681172 Psoriatic arthritis was strongly associated with nail, scalp, skinfold, elbow/knee involvement, the severity, early onset age, and total disease time of the cutaneous presentation (18–20). Symptoms like arthralgia in female psoriasis patients indicated a high chance of developing PsA (21). Although November 2021 \| Volume 8 \| Article 681172 Frontiers in Medicine \| www.frontiersin.org Proteomics in Psoriatic Disease Qi et al. Qi et al. not all PsO patients with joints pain have PsA, a longitudinal study conﬁrmed that compared with psoriatic patients without joint complaints (PsO), those with arthralgia (PsOAr) were more likely to develop PsA in the subsequent follow-up period (22). better understanding of its eﬀects at the protein level (43–45). With the help of a series of peptide-to-spectra matches (PSM) by assigning fragment ion mass spectra to peptide sequences, which is similar to proteomics, proteogenomics query the search engines with a customized protein FASTA, which contain both genomes- and protein-modiﬁed sequence (46). More recently, an integrated proteomics pipeline (IPP) was established to combine a variety of search engines to improve the sensitivity of novel peptide identiﬁcations with a novel “cascade search” method, which maximizes the accuracy and reliability of new candidate biomarker discovery. The current proteogenomics application mainly focuses on precision oncology, which assists in diﬀerentiating the subtypes and relevant pathways of tumors (47–53). Although no studies have shown its application on rheumatic diseases, proteogenomics is now the primary suggestion for PsC/PsA biomarker discovery (2, 30). Psoriatic arthropathy, an early stage of joint involvement that may not fulﬁll the PsA diagnostic criteria, is more common than PsA in PsO patients (23). For those with asymptomatic joint abnormalities, early synovio-entheseal inﬂammation or bone erosion can be detected by imaging features like ultrasonography or MRI (24, 25). However, with these predictors, it is still hard to foresee the possibility of the transition to PsA (26). Unlike rheumatoid arthritis (RA), the absence of serum diagnostic biomarkers impedes the identiﬁcation of very early PsA from PsC patients (8, 9, 27). “Omic” technologies have achieved enormous progress in their development and application over the past decades, which provided an unprecedented opportunity to decipher the entire genes (genomics), mRNA (transcriptomics), proteins (proteomics), and metabolites (metabolomics) of a speciﬁc biological sample (28, 29). Citation: Notably, advances in proteomics have made it possible for the head-to-head comparisons of potential biomarkers in the heterogeneity of PsD (8, 30). The present article reviewed the latest development of proteomics technologies, summarized its application in PsA biomarker discovery, and discussed the possible clinical detectable PsA risk factors in PsC patients. Proximity extension assay is a novel technology with up to 96-plex immune assays invented by Olink Proteomics (Uppsala, Sweden), which consolidates quantitative real-time PCR (54, 55). It was based on a dual recognition of selected antibodies with which biomarker-speciﬁc DNA “barcodes” oligonucleotides were labeled. The unique DNA will be merged by high-throughput relative quantiﬁcation microﬂuidic qPCR for up to 1,161 human proteins in the plasma (54, 55). Compared with LC-MS/MS, PEA covers a broader dynamic range with higher sensitivity, which provides sensitive and speciﬁc detection of low-abundant proteins in human blood and other body ﬂuid samples (55–58). Moreover, PEA also tends to be less inﬂuenced by multiplex ELISA technical problems, such as antibody crossreactivity and interassay variability (59). PEA has been widely applied in non-clinical biomedical research to decipher minute protein concentrations in minute sample volumes. In contrast, current studies have seen more applications of PEA in exploring both diagnostic markers and inﬂammation key components (60, 61). RECENT DEVELOPMENTS IN PROTEOMICS TECHNOLOGIES Proteome, as the ultimate goal for biomarker discovery, is the analysis of the whole protein materials of a disease or a biological sample, which oﬀers possibilities to track the changes in protein expression under diﬀerent conditions (31, 32). Present proteomic technologies could be addressed either as system- wide and unbiased tools such as antibody-based assay, aptamer- based assay, and mass spectrometry (MS) or a highly sensitive targeted immunoassay, such as the proximity extension assay (PEA) (33–35). PROTEOMICS IN POTENTIAL BIOMARKER DISCOVERY OF PSA TRANSITION Identifying early asymptomatic PsA in patients with PsO has been recognized as a historically complex issue with no exact serum diagnostic biomarkers used in daily clinical practice (8). Proteomics is extensively adopted in biomarker exploration. The emergence of proteomic technologies allows deciphering the changes in protein expression under diseased conditions. The following session of this work will review the detected possible predictors that may indicate early preclinical and subclinical PsA under the novel proteomics technologies (62) (Figure 1). Mass spectrometry is a powerful and ﬂexible instrument for characterizing proteins in their entirety (36–38). Of note, the introduction of high-throughput and high sensitivity protein identiﬁcation and quantiﬁcation methods to the single-cell proteomics and multi omics technologies help identify the candidate biomarkers in a protein-centric molecular way (29, 39–41). Ample studies have shown that the protein expression proﬁle in the serum of patients with PsC or PsA can be illustrated via multiple MS approaches, including data-dependent methods (such as label-based, label-free, MuDPIT, and shotgun proteomics) and targeted data- independent approaches (such as SWATH and MSE, multiple reaction monitoring, phospho-, and ubiquitinoylation-targeted proteomics) (35, 42). Furthermore, an emerging concept of “proteogenomics” produced fused the insights of proteomic and genomic, in which genomic events, such as SNPs, mutations, insertions, deletions, and substitutions and be detected with a Frontiers in Medicine \| www.frontiersin.org Proteomics in Peripheral Blood p Plasma and serum are extensively applied for proteomics-based biomarker discovery (63). Plenty of studies highlighted both PsA diagnostic and prognostic biomarkers with the help of MS proteomics technology (64). Serum proteome can also be obtained by PEA, an emerging technology previously explored in immune-mediated diseases of the skin, such as atopic dermatitis (65, 66). In a head-to-head comparison of serum biomarkers November 2021 \| Volume 8 \| Article 681172 Frontiers in Medicine \| www.frontiersin.org 2 Qi et al. Proteomics in Psoriatic Disease FIGURE 1 \| Overview of the potential biomarkers of PsA found through proteomic technologies from four different kinds of biological samples: Biomarkers from synovial ﬂuid and skin tissues are presented in the ﬁgure. No possible biomarkers are found from urine and serum. PsA, psoriatic arthritis; PEA, proximity extension assay. FIGURE 1 \| Overview of the potential biomarkers of PsA found through proteomic technologies from four different kinds of biological samples: Biomarkers from synovial ﬂuid and skin tissues are presented in the ﬁgure. No possible biomarkers are found from urine and serum. PsA, psoriatic arthritis; PEA, proximity extension assay. potential biomarkers of PsA found through proteomic technologies from four different kinds of biological samples: Biomarkers from s are presented in the ﬁgure. No possible biomarkers are found from urine and serum. PsA, psoriatic arthritis; PEA, proximity between PsC and PsA, Leijten et al. chose a high-throughput serum biomarker platform (Olink) to evaluate the concentrations of 951 serum proteins in both patients with PsA and PsC. Although no biomarkers with a signiﬁcant diﬀerence were found between PsC and PsA, PASI scores were found most strongly correlated to the proteins PI3, IL-17 receptor A, MMP-1, and SERPING8, when patients with PsA and PsO belonged to one group. When analyzing PsA patients as one group separately, PASI score was found correlated to Gal-4 and IGFBPL1. Four proteins including Intercellular adhesion molecule-1 (ICAM1), CC chemokine ligand 18 (CCL18), Dipeptidyl-peptidase 4 (DPP4), Vascular endothelial growth factor D (VEGFD), were found correlate to arthritis activity evaluated by swollen joint count (SJC), among which ICAM-1 and CCL18 were reported relevant to synovial tissue in rheumatoid arthritis activity. The swollen joint count (SJC) was identiﬁed, among which ICAM- 1 and CCL18 were reported relevant to synovial tissue in RA, whereas VEGFD was proposed to participate in the pathogenesis of arthritis. Frontiers in Medicine \| www.frontiersin.org Proteomics in Peripheral Blood DPP4 was only found to be related to type 2 diabetes mellitus rather than in arthritis development (8, 67–71). between PsC and PsA, Leijten et al. chose a high-throughput serum biomarker platform (Olink) to evaluate the concentrations of 951 serum proteins in both patients with PsA and PsC. Although no biomarkers with a signiﬁcant diﬀerence were found between PsC and PsA, PASI scores were found most strongly correlated to the proteins PI3, IL-17 receptor A, MMP-1, and SERPING8, when patients with PsA and PsO belonged to one group. When analyzing PsA patients as one group separately, PASI score was found correlated to Gal-4 and IGFBPL1. Four proteins including Intercellular adhesion molecule-1 (ICAM1), CC chemokine ligand 18 (CCL18), Dipeptidyl-peptidase 4 (DPP4), Vascular endothelial growth factor D (VEGFD), were found correlate to arthritis activity evaluated by swollen joint count (SJC), among which ICAM-1 and CCL18 were reported relevant to synovial tissue in rheumatoid arthritis activity. The swollen joint count (SJC) was identiﬁed, among which ICAM- 1 and CCL18 were reported relevant to synovial tissue in RA, whereas VEGFD was proposed to participate in the pathogenesis of arthritis. DPP4 was only found to be related to type 2 diabetes mellitus rather than in arthritis development (8, 67–71). is diﬃcult to ﬁnd a simple diagnostic protein from the serum to discriminate patients with PsA from patients with PsO, there is still a scarcity of serum proteome with PEA technology, and the mentioned study was completed with a small number of samples. Besides the 11 selected platforms encompassing only inﬂammatory proteins, more proteins reﬂected bone turnover and tissue biological changes, such as matrix metalloproteinase (MMPs) (72). Although human plasma is believed to be a feasible and less invasive source with a rich proteome, potential biomarkers secreted by the targeted tissues may be diluted in the blood with an undetectable concentration by current MS methods (73). In addition, many coexisting factors in the peripheral blood may interfere with the candidate soluble potential proteins. Thus, other biological samples, such as synovial ﬂuid (SF) and skin, have drawn more interest to be analyzed (74). Besides, some authorities recommended a more speciﬁc method to ﬁnding serum markers after the proteomics of inﬂamed synovial biomarkers (75). Proteomics in Synovial Samples Proteomics in Synovial Samples Synovium is the primary aﬀected site in most inﬂammatory arthritis (74). Many pathological modiﬁcations in inﬂamed synovial tissue are mirrored in the SF, which was more easily It was found that there were 20 dysregulated proteins, which specially existed in the serum of patients with PsA, which showed at the normal range in the PsO group when compared with the health control (8). Though the published research suggested, it Synovium is the primary aﬀected site in most inﬂammatory arthritis (74). Many pathological modiﬁcations in inﬂamed synovial tissue are mirrored in the SF, which was more easily November 2021 \| Volume 8 \| Article 681172 3 Qi et al. Proteomics in Psoriatic Disease accessible and widely studied (76). SF is a versatile source for proteins from the synovial membrane, cartilage, and plasma, depicting the pathophysiological issues that cause arthritis (77). A previously performed label-free MS quantitation of SF proteomics identiﬁed and veriﬁed 12 candidate PsA markers, including MPO, M2BP, DEFA1, H4, H2AFX, ORM1, CD5L, PFN1, and C4BP, as well as the top three upregulated proteins: MMP3, S100A9, and CRP (78). In another age-matched study, 10 SF samples from patients with PsA who were examined by using liquid chromatography-tandem MS quantitation revealed that Periostin (POSTN) and phosphoglycerate kinase 1 (PGK1) were upregulated with folded ratio compared with healthy controls (79). Although both studies showed a promising direction in SF proteome, no available data compared SF biomarkers between PsA and PsC samples. accessible and widely studied (76). SF is a versatile source for proteins from the synovial membrane, cartilage, and plasma, depicting the pathophysiological issues that cause arthritis (77). A previously performed label-free MS quantitation of SF proteomics identiﬁed and veriﬁed 12 candidate PsA markers, including MPO, M2BP, DEFA1, H4, H2AFX, ORM1, CD5L, PFN1, and C4BP, as well as the top three upregulated proteins: MMP3, S100A9, and CRP (78). In another age-matched study, 10 SF samples from patients with PsA who were examined by using liquid chromatography-tandem MS quantitation revealed that Periostin (POSTN) and phosphoglycerate kinase 1 (PGK1) were upregulated with folded ratio compared with healthy controls (79). Although both studies showed a promising direction in SF proteome, no available data compared SF biomarkers between PsA and PsC samples. on a rabbit tendon injury model to compare protein expression in intrasynovial tendon grafts and extra synovial tendon grafts, which oﬀered a possible substitute for the hard-to-access human samples (91). Proteomics in Skin Lesion Biopsy p y Skin manifestations, which include psoriasis Vulgaris or plaque psoriasis, were strongly associated with PsA (82). One hypothetical model for PsA transition was a systematic expansion of inﬂammation from the skin to synovio-entheseal tissues (62, 83). Factors that caused cutaneous diseases in the skin were released to promote a systemic dysregulated immune-mediated response and to develop musculoskeletal lesions after a second hit, such as trauma, infection, etc. (84, 85). Hence, it is of great need to explore the skin proteome in patients with PsA and PsC. Label-free quantitation of skin proteins veriﬁed 47 diﬀerent peptides between samples in the two groups. After validation in serum by ELISA, integrin β5 (ITGB5), a group of transmembrane receptors function on cell adhesion, increased signiﬁcantly in the PsA group when compared with the PsC group. Besides POSTN, a secreted extracellular matrix protein originally derived from the osteoblasts, was believed as a potential serum biomarker with a slightly higher concentration in PsA patients than in PsC patients (86). Another latest research using isobaric tags for relative and absolute quantitation (iTRAQ), a labeled MS technology, found 2-5-oligoadenylate synthase levels in both serum and psoriatic epidermis that were positively correlated with the severity of psoriasis through PASI and BSA (87, 88). As some data suggest, severe psoriasis can account for another cutaneous feature with a higher risk and prevalence of psoriatic arthropathy. The plasma membrane ATPase (derived from the OSA2 gene) might become another possible predictor for early joint inﬂammations in psoriatic patients (89, 90). Although these results are promising, limitations such as small sample numbers and the absence of further repetitive investigations in skin proteome impede the uncover of candidate PsA biomarkers, as well as the understanding of the underlying mechanism. There is no published research involving synovial tissue proteome in patients with PsA or PsC. Farnebo et al. performed MS analysis Proteomics in Urine Urine is another excellent source for both systemic and renal inﬂammatory biomarker exploration for its non-invasive sample collection approach as well as the low dynamic analytes range (92). Most proteins identiﬁed in urine are ﬁltered from the plasma or generated by inﬂammatory renal cells, contributing to a relatively small number of proteins appearing in the urine in patients with normal kidney function (93). Meanwhile, active proteases in the urine limit the degradation of biomarkers, leading urinary proteomics with MS-based analysis to become one of the most attractive directions in disease biomarker discovery (94, 95). Most published literature utilized urine proteome as a target for detecting biomarkers to kidney and cardiovascular diseases, with only a few describing urine proteomics technologies on inﬂammatory arthritis (64, 96, 97). In research exploring urine biomarkers in four diﬀerent arthritis [RA; PsA; osteoarthritis (OA); and inﬂammatory bowel diseases (IBD)], 50 most signiﬁcant peptides, including 80% speciﬁc for one group only, and a minor overlap were found through urinary proteomics (98). However, the most detectable peptide markers in this study were collagen fragments previously derived from proteins functionally diﬀerent from arthritis, which may be due to the ﬁltration of the glomerulus or the limited uncovered nature of the peptides in the urine (98). The result indeed showed the potent application of urine proteomics and peptidomics in the future (99). More longitude cohort studies in a large number of samples should be carried out in the future. The acquisition of SF is more feasible than synovial tissue, but it is undeniable that SF sometimes provides only indirect biomarkers (80). In the study of RA, the analysis of synovial tissue samples oﬀered great insights into both epigenetic and proteomic changes in patients with very early-stage RA. Therefore, synovial tissue might also be helpful and become a more precise target source in investigating PsA (74, 81). Frontiers in Medicine \| www.frontiersin.org REFERENCES Genome- wide association analysis of psoriatic arthritis and cutaneous psoriasis reveals diﬀerences in their genetic architecture. Am J Hum Genet. (2015) 97:816– 36. doi: 10.1016/j.ajhg.2015.10.019 18. Wilson FC, Icen M, Crowson CS, Mcevoy MT, Gabriel SE, Kremers HM. Incidence and clinical predictors of psoriatic arthritis in patients with psoriasis: a population-based study. Arthritis Rheum. (2009) 61:233– 9. doi: 10.1002/art.24172 6. Sakkas LI, Bogdanos DP. Are psoriasis and psoriatic arthritis the same disease? The IL-23/IL-17 axis data. Autoimmun Rev. (2017) 16:10– 5. doi: 10.1016/j.autrev.2016.09.015 19. Choi JW, Kim BR, Seo E, Youn SW. Could psoriatic arthritis be easily diagnosed from current suspicious physical ﬁndings in the dermatology clinic? Ann Dermatol. (2017) 29:48–54. doi: 10.5021/ad.2017.29.1.48 7. Veale DJ, Fearon U. The pathogenesis of psoriatic arthritis. Lancet. (2018) 391:2273–84. doi: 10.1016/S0140-6736(18)30830-4 20. Yan D, Ahn R, Leslie S, Liao W. Clinical and genetic risk factors associated with psoriatic arthritis among patients with psoriasis. Dermatol Ther. (2018) 8:593–604. doi: 10.1007/s13555-018-0266-x 8. Leijten E, Tao W, Pouw J, Van Kempen T, Olde Nordkamp M, Balak D, et al. Broad proteomic screen reveals shared serum proteomic signature in patients with psoriatic arthritis and psoriasis without arthritis. Rheumatology. (2021) 60:751–61. doi: 10.1093/rheumatology/keaa405 21. Eder L, Polachek A, Rosen CF, Chandran V, Cook R, Gladman DD. The development of psoriatic arthritis in patients with psoriasis is preceded by a period of nonspeciﬁc musculoskeletal symptoms: a prospective cohort study. Arthritis Rheumatol. (2017) 69:622–9. doi: 10.1002/art.39973 9. Ritchlin CT, Colbert RA, Gladman DD. Psoriatic arthritis. N Engl J Med. (2017) 376:957–70. doi: 10.1056/NEJMra1505557 Arthritis Rheumatol. (2017) 69:622–9. doi: 10.1002/art.39973 10. Kavanaugh A, Gottlieb A, Morita A, Merola JF, Lin CY, Birt J, et al. The contribution of joint and skin improvements to the health-related quality of life of patients with psoriatic arthritis: a post hoc analysis of two randomised controlled studies. Ann Rheum Dis. (2019) 78:1215– 9. doi: 10.1136/annrheumdis-2018-215003 22. Zabotti A, Mcgonagle DG, Giovannini I, Errichetti E, Zuliani F, Zanetti A, et al. Transition phase towards psoriatic arthritis: clinical and ultrasonographic characterisation of psoriatic arthralgia. RMD Open. (2019) 5:e001067. doi: 10.1136/rmdopen-2019-001067 23. Helliwell PS, Taylor WJ. Classiﬁcation and diagnostic criteria for psoriatic arthritis. Ann Rheum Dis. (2005) (64 Suppl. 2):ii3–8. doi: 10.1136/ard.2004.032318 11. Gladman DD, Antoni C, Mease P, Clegg DO, Nash P. Psoriatic arthritis: epidemiology, clinical features, course, and outcome. Ann Rheum Dis. (2005) 64(Suppl. 2):ii14–7. doi: 10.1136/ard.2004.032482 24. FUNDING This project was funded by Beijing Municipal Science and Technology Project (Project) No. Z171100001017058 and National Natural Science Foundation of China No. 81773314. CONCLUSION Over the past two decades, PsD is gradually considered a systematic inﬂammation that causes multiple associated comorbidities across the body rather than a simple disease cutaneous lesion (100). The emergence of skin presentation of psoriasis oﬀers a unique opportunity for early management for those at high-risk systematic progression (101). Although existing reviews have already pointed out that imaging methods, such as ultrasound and MRI, can also become a valuable method to detect early the inﬂammatory lesions of joints, the expensive costs of exam fees and related equipment, and the long waiting time are limitations. Examination time and hard-interpreted imaging results for non-professional clinicians were all hurdles that hamper the prevalence of application on imaging examinations on patients with PsC (22, 25, 102). Consequently, a fast exam kit with an accessible kit becomes more necessary, suggesting an imperative need to explore a possible biomarker. The immense development and utilization in proteomics have provided an extraordinary chance to detail the molecular and mechanistic understanding of PsD pathways, decode the potential biomarkers, and investigate more eﬀective intervention therapies (103, 104). November 2021 \| Volume 8 \| Article 681172 4 Proteomics in Psoriatic Disease Qi et al. AUTHOR CONTRIBUTIONS This review summarized the current approaches applied in the early PsA proteome. Compared with the traditional LC-MS/MS methods in proteogenomics, PEA provides more sensitive and speciﬁc detection for a more considerable range of low-abundant proteins in human blood and other body ﬂuid samples (55– 58). However, the need for the custom panel of biomarkers also restricted the exploration of the unknown proteins. Only a few studies that focused on psoriatic arthropathy ﬁnished their study with PEA technology. It highlighted the great need to perform high-throughput analyses in serum and tissues and other possible samples to discover PsA precursors. The future work on performing extensive integrative analysis will be undoubtedly challenging. Still, the increasing recognition of human proteome and consistent progression on proteomics technologies will become the most supportive foundation for challenging tasks. YH: did the project administration, conceptualization, and methodology. FQ and YT: did the investigation and formal analysis. FQ, YT, AY, and XY: oﬀered the resources. FQ wrote the original draft. YT: reviewed and edited the draft. YH: visualized the whole project and supervised the whole project. All authors contributed to the article and approved the submitted version. REFERENCES of psoriasis among patients seen in European dermatology clinics. J Eur Acad Dermatol Venereol. (2010) 24:548–54. doi: 10.1111/j.1468-3083.2009.03463.x 1. Armstrong AW, Read C. Pathophysiology, clinical presentation, and treatment of psoriasis: a review. JAMA. (2020) 323:1945– 60. doi: 10.1001/jama.2020.4006 1. Armstrong AW, Read C. Pathophysiology, clinical presentation, and treatment of psoriasis: a review. JAMA. (2020) 323:1945– 60. doi: 10.1001/jama.2020.4006 14. Ogdie A, Weiss P. The epidemiology of psoriatic arthritis. Rheum Dis Clin North Am. (2015) 41:545–68. doi: 10.1016/j.rdc.2015.07.001 15. Kampylafka E, Simon D, D’oliveira I, Linz C, Lerchen V, Englbrecht M, et al. Disease interception with interleukin-17 inhibition in high-risk psoriasis patients with subclinical joint inﬂammation- data from the prospective IVEPSA study. Arthritis Res Ther. (2019) 21:178. doi: 10.1186/s13075-019-1957-0 2. Jadon DR, Stober C, Pennington SR, Fitzgerald O. Applying precision medicine to unmet clinical needs in psoriatic disease. Nat Rev Rheumatol. (2020) 16:609–27. doi: 10.1038/s41584-020-00507-9 2. Jadon DR, Stober C, Pennington SR, Fitzgerald O. Applying precision medicine to unmet clinical needs in psoriatic disease. Nat Rev Rheumatol. (2020) 16:609–27. doi: 10.1038/s41584-020-00507-9 3. Belasco J, Louie JS, Gulati N, Wei N, Nograles K, Fuentes-Duculan J, et al. Comparative genomic proﬁling of synovium versus skin lesions in psoriatic arthritis. Arthritis Rheumatol. (2015) 67:934–44. doi: 10.1002/art.38995 16. Zabotti A, Tinazzi I, Aydin SZ, Mcgonagle D. From psoriasis to psoriatic arthritis: insights from imaging on the transition to psoriatic arthritis and implications for arthritis prevention. Curr Rheumatol Rep. (2020) 22:24. doi: 10.1007/s11926-020-00891-x 4. Pollock RA, Abji F, Liang K, Chandran V, Pellett FJ, Virtanen C, et al. Gene expression diﬀerences between psoriasis patients with and without inﬂammatory arthritis. J Invest Dermatol. (2015) 135:620– 3. doi: 10.1038/jid.2014.414 4. Pollock RA, Abji F, Liang K, Chandran V, Pellett FJ, Virtanen C, et al. Gene expression diﬀerences between psoriasis patients with and without inﬂammatory arthritis. J Invest Dermatol. (2015) 135:620– 3. doi: 10.1038/jid.2014.414 17. Haroon M, Gallagher P, Fitzgerald O. Diagnostic delay of more than 6 months contributes to poor radiographic and functional outcome in psoriatic arthritis. Ann Rheum Dis. (2015) 74:1045–50. doi: 10.1136/annrheumdis-2013-204858 5. Stuart PE, Nair RP, Tsoi LC, Tejasvi T, Das S, Kang HM, et al. Genome- wide association analysis of psoriatic arthritis and cutaneous psoriasis reveals diﬀerences in their genetic architecture. Am J Hum Genet. (2015) 97:816– 36. doi: 10.1016/j.ajhg.2015.10.019 5. Stuart PE, Nair RP, Tsoi LC, Tejasvi T, Das S, Kang HM, et al. REFERENCES Cheung P, Khatri P, Utz PJ, Kuo AJ. Single-cell technologies – studying rheumatic diseases one cell at a time. Nat Rev Rheumatol. (2019) 15:340–54. doi: 10.1038/s41584-019- 0220-z 50. Zhang B, Wang J, Wang X, Zhu J, Liu Q, Shi Z, et al. Proteogenomic characterization of human colon and rectal cancer. Nature. (2014) 513:382–7. doi: 10.1038/nature 13438 31. Chandramouli K, Qian PY. Proteomics: challenges, techniques and possibilities to overcome biological sample complexity. Hum Genomics Proteomics. (2009) 2009:239204. doi: 10.4061/2009/ 239204 51. Li H, Joh YS, Kim H, Paek E, Lee SW, Hwang KB. Evaluating the eﬀect of database inﬂation in proteogenomic search on sensitive and reliable peptide identiﬁcation. BMC Genomics. (2016) 17:1031. doi: 10.1186/s12864-016-3327-5 32. Whiteaker JR, Halusa GN, Hoofnagle AN, Sharma V, Maclean B, Yan P, et al. CPTAC Assay Portal: a repository of targeted proteomic assays. Nat Methods. (2014) 11:703–4. doi: 10.1038/nmeth.3002 52. Mertins P, Mani DR, Ruggles KV, Gillette MA, Clauser KR, Wang P, et al. Proteogenomics connects somatic mutations to signalling in breast cancer. Nature. (2016) 534:55–62. doi: 10.1038/nature 18003 33. Mann M. Origins of mass spectrometry-based proteomics. Nat Rev Mol Cell Biol. (2016) 17:678. doi: 10.1038/nrm.2016.135 34. Smith JG, Gerszten RE. Emerging aﬃnity-based proteomic technologies for large-scale plasma proﬁling in cardiovascular disease. Circulation. (2017) 135:1651–64. doi: 10.1161/CIRCULATIONAHA.116. 025446 53. Park GW, Hwang H, Kim KH, Lee JY, Lee HK, Park JY, et al. Integrated proteomic pipeline using multiple search engines for a proteogenomic study with a controlled protein false discovery rate. J Proteome Res. (2016) 15:4082– 90. doi: 10.1021/acs.jproteome.6b00376 35. Mcardle A, Pennington S, Fitzgerald O. Clinical features of psoriatic arthritis: a comprehensive review of unmet clinical needs. Clin Rev Allergy Immunol. (2018) 55:271–94. doi: 10.1007/s12016-017- 8630-7 54. Assarsson E, Lundberg M, Holmquist G, Bjorkesten J, Thorsen SB, Ekman D, et al. Homogenous 96-plex PEA immunoassay exhibiting high sensitivity, speciﬁcity, and excellent scalability. PLoS ONE. (2014) 9:e95192. doi: 10.1371/journal.pone. 0095192 36. Gstaiger M, Aebersold R. Applying mass spectrometry-based proteomics to genetics, genomics and network biology. Nat Rev Genet. (2009) 10:617– 27. doi: 10.1038/nrg2633 55. Petrera A, Von Toerne C, Behler J, Huth C, Thorand B, HilgendorﬀA, et al. Multiplatform approach for plasma proteomics: complementarity of olink proximity extension assay technology to mass spectrometry-based protein proﬁling. J Proteome Res. (2021) 20:751–62. doi: 10.1021/acs.jproteome. 0c00641 37. Cox J, Mann M. Quantitative, high-resolution proteomics for data-driven systems biology. Annu Rev Biochem. (2011) 80:273–99. doi: 10.1146/annurev-biochem-061308- 093216 56. REFERENCES Lundberg M, Eriksson A, Tran B, Assarsson E, Fredriksson S. Homogeneous antibody-based proximity extension assays provide sensitive and speciﬁc detection of low-abundant proteins in human blood. Nucl Acids Res. (2011) 39:e102. doi: 10.1093/nar/gkr424 38. Kim MS, Pinto SM, Getnet D, Nirujogi RS, Manda SS, Chaerkady R, et al. A draft map of the human proteome. Nature. (2014) 509:575– 81. doi: 10.1038/nature13302 39. Bock C, Farlik M, Sheﬃeld NC. Multi-omics of single cells: strategies and applications. Trends Biotechnol. (2016) 34:605– 8. doi: 10.1016/j.tibtech.2016.04.004 57. Csosz E, Toth N, Deak E, Csutak A, Tozser J. Wound-healing markers revealed by proximity extension assay in tears of patients following glaucoma surgery. Int J Mol Sci. (2018) 19. doi: 10.3390/ijms19 124096 40. Yates JR 3rd. Innovation: structural proteomics goes global. J Proteome Res. (2018) 17:3613. doi: 10.1021/acs.jproteome.8b00698 58. Sjoqvist S, Otake K, Hirozane Y. Analysis of cerebrospinal ﬂuid extracellular vesicles by proximity extension assay: a comparative study of four isolation kits. Int J Mol Sci. (2020) 21. doi: 10.3390/ijms21249425 41. Marx V. A dream of single-cell proteomics. Nat Methods. (2019) 16:809– 12. doi: 10.1038/s41592-019-0540-6 42. Butt AQ, Mcardle A, Gibson DS, Fitzgerald O, Pennington SR. Psoriatic arthritis under a proteomic spotlight: application of novel technologies to advance diagnosis and management. Curr Rheumatol Rep. (2015) 17:35. doi: 10.1007/s11926-015- 0509-0 59. Tighe PJ, Ryder RR, Todd I, Fairclough LC. ELISA in the multiplex era: potentials and pitfalls. Proteomics Clin Appl. (2015) 9:406–22. doi: 10.1002/prca.201400130 60. Mikacic I, Beluzic R, Vugrek O, Plavljanic D. A Proximity Extension Assay (PEA)-based method for quantiﬁcation of bevacizumab. J Pharmacol Toxicol Methods. (2018) 92:20–3. doi: 10.1016/j.vascn.2018.02.008 43. Marx H, Hahne H, Ulbrich SE, Schnieke A, Rottmann O, Frishman D, et al. Annotation of the domestic pig genome by quantitative proteogenomics. J Proteome Res. (2017) 16:2887–98. doi: 10.1021/acs.jproteome.7b00184 61. Jabbari E, Woodside J, Guo T, Magdalinou NK, Chelban V, Athauda D, et al. Proximity extension assay testing reveals novel diagnostic biomarkers of atypical parkinsonian syndromes. J Neurol Neurosurg Psychiatry. (2019) 90:768–73. doi: 10.1136/jnnp-2018-320151 44. Ruggles KV, Krug K, Wang X, Clauser KR, Wang J, Payne SH, et al. Methods, tools and current perspectives in proteogenomics. Mol Cell Proteomics. (2017) 16:959–81. doi: 10.1074/mcp.MR117. 000024 62. Scher JU, Ogdie A, Merola JF, Ritchlin C. Preventing psoriatic arthritis: focusing on patients with psoriasis at increased risk of transition. Nat Rev Rheumatol. (2019) 15:153–66. doi: 10.1038/s41584-019 -0175-0 45. REFERENCES Simon D, Faustini F, Kleyer A, Haschka J, Englbrecht M, Kraus S, et al. Analysis of periarticular bone changes in patients with cutaneous psoriasis without associated psoriatic arthritis. Ann Rheum Dis. (2016) 75:660– 6. doi: 10.1136/annrheumdis-2014-206347 12. Alinaghi F, Calov M, Kristensen LE, Gladman DD, Coates LC, Jullien D, et al. Prevalence of psoriatic arthritis in patients with psoriasis: a systematic review and meta-analysis of observational and clinical studies. J Am Acad Dermatol. (2019) 80:251–65 e219. doi: 10.1016/j.jaad.2018. 06.027 25. Qi F, Yao A, He Y. Medical imaging examination in psoriasis and early psoriatic arthritis patients: an updated systematic review and meta-analysis. Int J Dermatol. (2021) 60:1354–62. doi: 10.1111/ijd.15384 25. Qi F, Yao A, He Y. Medical imaging examination in psoriasis and early psoriatic arthritis patients: an updated systematic review and meta-analysis. Int J Dermatol. (2021) 60:1354–62. doi: 10.1111/ijd.15384 13. Christophers E, Barker JN, Griﬃths CE, Dauden E, Milligan G, Molta C, et al. The risk of psoriatic arthritis remains constant following initial diagnosis November 2021 \| Volume 8 \| Article 681172 Frontiers in Medicine \| www.frontiersin.org 5 Qi et al. Proteomics in Psoriatic Disease (NGS) and mass spectrometry-based proteomics to precision medicine. Clin Chim Acta. (2019) 498:38–46. doi: 10.1016/j.cca.2019. 08.010 (NGS) and mass spectrometry-based proteomics to precision medicine. Clin Chim Acta. (2019) 498:38–46. doi: 10.1016/j.cca.2019. 08.010 26. Chandran V. Pathway to biomarker discovery in psoriatic arthritis. Expert Rev Clin Immunol. (2020) 16:471–8. doi: 10.1080/1744666X.2020. 1752667 47. Jagtap P, Goslinga J, Kooren JA, Mcgowan T, Wroblewski MS, Seymour SL, et al. A two-step database search method improves sensitivity in peptide sequence matches for metaproteomics and proteogenomics studies. Proteomics. (2013) 13:1352–7. doi: 10.1002/pmic.201200352 27. Taguchi A, Hanash SM. Unleashing the power of proteomics to develop blood-based cancer markers. Clin Chem. (2013) 59:119–26. doi: 10.1373/clinchem.2012.184572 28. Horgan RP, Kenny LC. ‘Omic’ technologies: genomics, transcriptomics, proteomics and metabolomics. Obstetrician Gynaecol. (2011) 13:189– 95. doi: 10.1576/toag.13.3.189.27672 48. Alfaro JA, Sinha A, Kislinger T, Boutros PC. Onco- proteogenomics: cancer proteomics joins forces with genomics. Nat Methods. (2014) 11:1107–13. doi: 10.1038/nmeth .3138 29. Olivier M, Asmis R, Hawkins GA, Howard TD, Cox LA. The need for multi- omics biomarker signatures in precision medicine. Int J Mol Sci. (2019) 20. doi: 10.3390/ijms20194781 49. Rivers RC, Kinsinger C, Boja ES, Hiltke T, Mesri M, Rodriguez H. Linking cancer genome to proteome: NCI’s investment into proteogenomics. Proteomics. (2014) 14:2633–6. doi: 10.1002/pmic.201400193 30. REFERENCES The activity of CCL18 is principally mediated through interaction with glycosaminoglycans. Front Immunol. (2013) 4:193. doi: 10.3389/ﬁmmu.2013.00193 88. Zhou Y, Wang P, Yan BX, Chen XY, Landeck L, Wang ZY, et al. Quantitative proteomic proﬁle of psoriatic epidermis identiﬁes oas2 as a novel Biomarker for Disease Activity. Front Immunol. (2020) 11:1432. doi: 10.3389/ﬁmmu.2020.01432 69. Rohrborn D, Wronkowitz N, Eckel J. DPP4 in diabetes. Front Immunol. (2015) 6:386. doi: 10.3389/ﬁmmu.2015. 00386 70. Balogh E, Biniecka M, Fearon U, Veale DJ, Szekanecz Z. Angiogenesis in inﬂammatory arthritis. Isr Med Assoc J. (2019) 21:345–52. 89. Rouzaud M, Sevrain M, Villani AP, Barnetche T, Paul C, Richard MA, et al. Is there a psoriasis skin phenotype associated with psoriatic arthritis? Systematic literature review. J Eur Acad Dermatol Venereol. (2014) (28 Suppl. 5):17–26. doi: 10.1111/jdv. 12562 71. Muller N. The role of intercellular adhesion molecule- 1 in the pathogenesis of psychiatric disorders. Front Pharmacol. (2019) 10:1251. doi: 10.3389/fphar.2019. 01251 90. Eder L, Haddad A, Rosen CF, Lee KA, Chandran V, Cook R, et al. The incidence and risk factors for psoriatic arthritis in patients with psoriasis: a prospective cohort study. Arthritis Rheumatol. (2016) 68:915–23. doi: 10.1002/art .39494 72. Penkava F, Velasco-Herrera MDC, Young MD, Yager N, Nwosu LN, Pratt AG, et al. Single-cell sequencing reveals clonal expansions of pro-inﬂammatory synovial CD8 T cells expressing tissue- homing receptors in psoriatic arthritis. Nat Commun. (2020) 11:4767. doi: 10.1038/s41467-020-18513-6 91. Farnebo S, Wiig M, Holm B, Ghafouri B. Diﬀerentially expressed proteins in intra synovial compared to extra synovial ﬂexor tendon grafts in a rabbit tendon transplantation model. Biomedicines. (2020) 8. doi: 10.3390/biomedicines8100408 73. Cretu D, Diamandis EP, Chandran V. Delineating the synovial ﬂuid proteome: recent advancements and ongoing challenges in biomarker research. Crit Rev Clin Lab Sci. (2013) 50:51– 63. doi: 10.3109/10408363.2013.802408 92. Klein J, Bascands JL, Mischak H, Schanstra JP. The role of urinary peptidomics in kidney disease research. Kidney Int. (2016) 89:539–45. doi: 10.1016/j.kint.2015. 10.010 74. Orr C, Vieira-Sousa E, Boyle DL, Buch MH, Buckley CD, Canete JD, et al. Synovial tissue research: a state-of-the-art review. Nat Rev Rheumatol. (2017) 13:463–75. doi: 10.1038/nrrheum.2017.115 93. Aljaberi N, Bennett M, Brunner HI, Devarajan P. Proteomic proﬁling of urine: implications for lupus nephritis. Expert Rev Proteomics. (2019) 16:303–13. doi: 10.1080/14789450.2019. 1592681 75. Choi IY, Gerlag DM, Holzinger D, Roth J, Tak PP. REFERENCES Zhu Y, Orre LM, Johansson HJ, Huss M, Boekel J, Vesterlund M, et al. Discovery of coding regions in the human genome by integrated proteogenomics analysis workﬂow. Nat Commun. (2018) 9:903. doi: 10.1038/s41467-018-03311-y 63. Lee PY, Osman J, Low TY, Jamal R. Plasma/serum proteomics: depletion strategies for reducing high-abundance proteins for biomarker discovery. Bioanalysis. (2019) 11:1799–812. doi: 10.4155/bio-2019-0145 46. Ang MY, Low TY, Lee PY, Wan Mohamad Nazarie WF, Guryev V, Jamal R. Proteogenomics: from next-generation sequencing November 2021 \| Volume 8 \| Article 681172 Frontiers in Medicine \| www.frontiersin.org 6 Qi et al. Proteomics in Psoriatic Disease 64. Mahendran SM, Chandran V. Exploring the psoriatic arthritis proteome in search of novel biomarkers. Proteomes. (2018) 6. doi: 10.3390/proteomes 6010005 82. Ogdie A, Coates LC, Gladman DD. Treatment guidelines in psoriatic arthritis. Rheumatology. (2020) 59:i37– 46. doi: 10.1093/rheumatology/kez383 65. He H, Del Duca E, Diaz A, Kim HJ, Gay-Mimbrera J, Zhang N, et al. Mild atopic dermatitis lacks systemic inﬂammation and shows reduced nonlesional skin abnormalities. J Allergy Clin Immunol. (2020) 147:1369–80. doi: 10.1016/j.jaci.2020. 08.041 83. Fitzgerald O, Winchester R. Editorial: emerging evidence for critical involvement of the interleukin-17 pathway in both psoriasis and psoriatic arthritis. Arthritis Rheumatol. (2014) 66:1077–80. doi: 10.1002/art.38370 84. Leonardi C, Matheson R, Zachariae C, Cameron G, Li L, Edson-Heredia E, et al. Anti-interleukin-17 monoclonal antibody ixekizumab in chronic plaque psoriasis. N Engl J Med. (2012) 366:1190–9. doi: 10.1056/NEJMoa1109997 66. He H, Li R, Choi S, Zhou L, Pavel A, Estrada YD, et al. Increased cardiovascular and atherosclerosis markers in blood of older patients with atopic dermatitis. Ann Allergy Asthma Immunol. (2020) 124:70–8. doi: 10.1016/j.anai.2019 .10.013 85. Taams LS, Steel KJA, Srenathan U, Burns LA, Kirkham BW. IL-17 in the immunopathogenesis of spondyloarthritis. Nat Rev Rheumatol. (2018) 14:453–66. doi: 10.1038/s41584-018-0044-2 86. Cretu D, Liang K, Saraon P, Batruch I, Diamandis EP, Chandran V. Quantitative tandem mass-spectrometry of skin tissue reveals putative psoriatic arthritis biomarkers. Clin Proteomics. (2015) 12:1. doi: 10.1186/1559-0275-12-1 67. Baldwin ME, Halford MM, Roufail S, Williams RA, Hibbs ML, Grail D, et al. Vascular endothelial growth factor D is dispensable for development of the lymphatic system. Mol Cell Biol. (2005) 25:2441–9. doi: 10.1128/MCB.25.6.2441-2449. 2005 87. Wiese S, Reidegeld KA, Meyer HE, Warscheid B. Protein labeling by iTRAQ: a new tool for quantitative mass spectrometry in proteome research. Proteomics. (2007) 7:340–50. doi: 10.1002/pmic.200 600422 68. Krohn S, Garin A, Gabay C, Proudfoot AE. REFERENCES From synovial tissue to peripheral blood: myeloid related protein 8/14 is a sensitive biomarker for eﬀective treatment in early drug development in patients with rheumatoid arthritis. PLoS ONE. (2014) 9:e106253. doi: 10.1371/journal.pone.0106253 94. Havanapan PO, Thongboonkerd V. Are protease inhibitors required for gel-based proteomics of kidney and urine? J Proteome Res. (2009) 8:3109– 17. doi: 10.1021/pr900015q 76. Levick JR. Permeability of rheumatoid and normal human synovium to speciﬁc plasma proteins. Arthritis Rheum. (1981) 24:1550–60. doi: 10.1002/art.1780241215 77. Hui AY, Mccarty WJ, Masuda K, Firestein GS, Sah RL. A systems biology approach to synovial joint lubrication in health, injury, and disease. Wiley Interdiscip Rev Syst Biol Med. (2012) 4:15–37. doi: 10.1002/wsbm.157 95. Thomas S, Hao L, Ricke WA, Li L. Biomarker discovery in mass spectrometry-based urinary proteomics. Proteomics Clin Appl. (2016) 10:358–70. doi: 10.1002/prca.201500102 78. Cretu D, Prassas I, Saraon P, Batruch I, Gandhi R, Diamandis EP, et al. Identiﬁcation of psoriatic arthritis mediators in synovial ﬂuid by quantitative mass spectrometry. Clin Proteomics. (2014) 11:27. doi: 10.1186/1559-0275-11-27 96. Santucci L, Bruschi M, Candiano G, Lugani F, Petretto A, Bonanni A, et al. Urine proteome biomarkers in kidney diseases. Limits I, perspectives, and ﬁrst focus on normal urine. Biomark Insights. (2016) 11:41–8. doi: 10.4137/BMI.S26229 79. Mahendran SM, Keystone EC, Krawetz RJ, Liang K, Diamandis EP, Chandran V. Elucidating the endogenous synovial ﬂuid proteome and peptidome of inﬂammatory arthritis using label-free mass spectrometry. Clin Proteomics. (2019) 16:23. doi: 10.1186/s12014-019- 9243-3 97. Baetta R, Pontremoli M, Martinez Fernandez A, Spickett CM, Banﬁ C. Proteomics in cardiovascular diseases: unveiling sex and gender diﬀerences in the era of precision medicine. J Proteomics. (2018) 173:62– 76. doi: 10.1016/j.jprot.2017.11.012 98. Siebert S, Porter D, Paterson C, Hampson R, Gaya D, Latosinska A, et al. Urinary proteomics can deﬁne distinct diagnostic inﬂammatory arthritis subgroups. Sci Rep. (2017) 7:40473. doi: 10.1038/srep 40473 80. Bresnihan B, Tak PP, Emery P, Klareskog L, Breedveld F. Synovial biopsy in arthritis research: ﬁve years of concerted European collaboration. Ann Rheum Dis. (2000) 59:506–11. doi: 10.1136/ard.59.7.506 99. Latosinska A, Siwy J, Mischak H, Frantzi M. Peptidomics and proteomics based on CE-MS as a robust tool in clinical application: the past, the present, and the future. Electrophoresis. (2019) 40:2294–308. doi: 10.1002/elps.201900091 81. Van Kuijk AW, Tak PP. Synovitis in psoriatic arthritis: immunohistochemistry, comparisons with rheumatoid arthritis, and eﬀects of therapy. Curr Rheumatol Rep. (2011) 13:353– 9. Frontiers in Medicine \| www.frontiersin.org November 2021 \| Volume 8 \| Article 681172 REFERENCES doi: 10.1007/s11926-011-0181-y November 2021 \| Volume 8 \| Article 681172 Frontiers in Medicine \| www.frontiersin.org 7 Qi et al. Proteomics in Psoriatic Disease Conﬂict of Interest: The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. 100. Yeung H, Takeshita J, Mehta NN, Kimmel SE, Ogdie A, Margolis DJ, et al. Psoriasis severity and the prevalence of major medical comorbidity: a population-based study. JAMA Dermatol. (2013) 149:1173– 9. doi: 10.1001/jamadermatol.2013.5015 101. Tillett W, Charlton R, Nightingale A, Snowball J, Green A, Smith C, et al. Interval between onset of psoriasis and psoriatic arthritis comparing the UK Clinical Practice Research Datalink with a hospital-based cohort. Rheumatology. (2017) 56:2109–13. doi: 10.1093/rheumatology/kex323 Publisher’s Note: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their aﬃliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. 102. Artandi MK, Stewart RW. The outpatient physical examination. Med Clin North Am. (2018) 102:465–73. doi: 10.1016/j.mcna.2017.12.008 103. Ciuﬀa R, Caron E, Leitner A, Uliana F, Gstaiger M, Aebersold R. Contribution of mass spectrometry-based proteomics to the understanding of TNF-alpha signaling. J Proteome Res. (2017) 16:14–33. doi: 10.1021/acs.jproteome.6b00728 Copyright © 2021 Qi, Tan, Yao, Yang and He. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. 104. Sharma S, Gioia L, Abe B, Holt M, Costanzo A, Kain L, et al. Using single cell analysis for translational studies in immune mediated diseases: opportunities and challenges. Mol Immunol. (2018) 103:191– 9. doi: 10.1016/j.molimm.2018.09.020 November 2021 \| Volume 8 \| Article 681172 Frontiers in Medicine \| www.frontiersin.org 8
https://openalex.org/W4391690698	https://www.frontiersin.org/articles/10.3389/fphy.2024.1340505/pdf?isPublishedV2=False	English	null	Wavelength dependent transmission in multimode graded-index microstructured polymer optical fibers	Frontiers in physics	2,024	cc-by	5,358	OPEN ACCESS OPEN ACCESS EDITED BY Jitendra Bahadur Maurya, National Institute of Technology Patna, India REVIEWED BY Satyendra Kumar Mishra, Centre Tecnologic De Telecomunicacions De Catalunya, Spain Sushank Chaudhary, Chulalongkorn University, Thailand CORRESPONDENCE Zhuo Wang, zhuowang@bnu.edu.cn RECEIVED 18 November 2023 ACCEPTED 26 January 2024 PUBLISHED 09 February 2024 CITATION Simović A, Savović S, Wang Z, Drljača B, Kovačević MS, Kuzmanović L, Djordjevich A, Aidinis K and Chen C (2024), Wavelength dependent transmission in multimode graded- index microstructured polymer optical ﬁbers. Front. Phys. 12:1340505. doi: 10.3389/fphy.2024.1340505 OPEN ACCESS EDITED BY Jitendra Bahadur Maurya, National Institute of Technology Patna, India REVIEWED BY Satyendra Kumar Mishra, Centre Tecnologic De Telecomunicacions De Catalunya, Spain Sushank Chaudhary, Chulalongkorn University, Thailand CORRESPONDENCE Zhuo Wang, zhuowang@bnu.edu.cn RECEIVED 18 November 2023 ACCEPTED 26 January 2024 PUBLISHED 09 February 2024 CITATION Simović A, Savović S, Wang Z, Drljača B, Kovačević MS, Kuzmanović L, Djordjevich A, Aidinis K and Chen C (2024), Wavelength dependent transmission in multimode graded- index microstructured polymer optical ﬁbers. Front. Phys. 12:1340505. doi: 10.3389/fphy.2024.1340505 Ana Simović1, Svetislav Savović1,2, Zhuo Wang 3*, Branko Drljača4, Milan S. Kovačević1, Ljubica Kuzmanović1, Alexandar Djordjevich 2, Konstantinos Aidinis5,6 and Chen Chen 7 1Faculty of Science, University of Kragujevac, Kragujevac, Serbia, 2Department of Mechanial Enginering, City University of Hong Kong, Hong Kong, China, 3Center for Cognition and Neuroergonomics, State Key Laboratory of Cognitive Neuroscience and Learning, Beijing Normal University at Zhuhai, Zhuhai, China, 4Faculty of Sciences and Mathematics, University of Priština in Kosovska Mitrovica, Kosovska Mitrovica, Serbia, 5Department of Electrical Engineering, Ajman University, Ajman, United Arab Emirates, 6Center of Medical and Bio-Allied Health Sciences Research, Ajman University, Ajman, United Arab Emirates, 7 7School of Microelectronics and Communication Engineering, Chongqing University, Chongqing, China Up to now, there have been no commercial simulation tools accessible for researching the transmission properties of multimode microstructured optical ﬁbers (MOFs). In order to avoid this problem, this study uses the time- independent power ﬂow equation (TI PFE) numerical solution to examine the wavelength dependency of the equilibrium mode distribution (EMD) and steady state distribution (SSD) in multimode graded-index microstructured polymer optical ﬁbers (GI mPOF) with a solid core. We showed that the lengths zs at which an SSD is obtained in GI mPOF and the coupling length Lc necessary to create an EMD are shorter at λ = 568 nm than they are found to be at λ = 633 nm. TYPE Original Research PUBLISHED 09 February 2024 DOI 10.3389/fphy.2024.1340505 TYPE Original Research PUBLISHED 09 February 2024 DOI 10.3389/fphy.2024.1340505 KEYWORDS polymer optical ﬁber, graded-index optical ﬁber, microstructured optical ﬁber, power ﬂow equation, wavelength dependent transmission frontiersin.org COPYRIGHT © 2024 Simović, Savović, Wang, Drljača, Kovačević, Kuzmanović, Djordjevich, Aidinis and Chen. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. polymer optical ﬁber, graded-index optical ﬁber, microstructured optical ﬁber, power ﬂow equation, wavelength dependent transmission OPEN ACCESS The lengths Lc and zs stay constant when the wavelength decreases further from λ = 568 to 522 and then to 476 nm. As a result, it is anticipated that a faster bandwidth enhancement in the tested GI mPOF will take place at wavelengths around λ = 568 nm as opposed to λ = 633 nm. Such a bandwidth improvement is not brought about by additional wavelength reduction. The study’s ﬁndings can be used in communication and sensory systems that use multimode GI mPOFs at different wavelengths. y doi: 10.3389/fphy.2024.1340505 1 Introduction Recent years have seen a signiﬁcant increase in research interest in high-speed short- range signal transmission across polymer optical ﬁbers (POFs) [1–3]. The assets of POFs, such as a large core and simple connection, may offer a cost-effective solution for the in- home network. Polymethyl methacrylate (PMMA) [4, 5], polydimethylsiloxane (PDMS) [6, 7], polystyrene (PS) [8, 9], polycarbonate (PC) [10, 11], perﬂuorinated polymer (CYTOP®) [12, 13], cyclooleﬁn polymer (ZEONEX®) [14, 15], and cyclooleﬁn copolymer (TOPAS®) [16, 17] are just a few of the materials used to fabricate POFs. Due to the ﬂexibility of POF material, it is feasible to produce POFs that meet the requirements of various applications by Frontiers in Physics 01 frontiersin.org Simović et al. 10.3389/fphy.2024.1340505 FIGURE 1 (A) The cross-section of the multimode GI MOF. The pitch Λ determines the position of air holes in a triangular lattice. The air holes that make the four rings of the core have the following diameters: d1, d2, d3, and d4. Two rings of air holes of the same diameter as those in the outermost core ring form the cladding (d4 = d5 = d6). (B) The dashed blue line shows the referent multimode GI MOF’s RI distribution. The solid black line represents the RI distribution in the core based on Eq. 1, when g = 2.0, and at λ = 633 nm. FIGURE 1 (A) The cross-section of the multimode GI MOF. The pitch Λ determines the position of air holes in a triangular lattice. The air holes that make the four rings of the core have the following diameters: d1, d2, d3, and d4. Two rings of air holes of the same diameter as those in the outermost core ring form the cladding (d4 = d5 = d6). (B) The dashed blue line shows the referent multimode GI MOF’s RI distribution. The solid black line represents the RI distribution in the core based on Eq. 1, when g = 2.0, and at λ = 633 nm. Up to now, there have been no commercial simulation tools available for studying the transmission characteristics of multimode MOFs. To circumvent this problem, the time-independent power ﬂow equation (TI PFE) is numerically solved in this study to investigate the wavelength dependent light transmission in GI mPOF. 2 GI mPOF design The GI mPOF that was examined in this study is depicted in Figure 1. This GI mPOF is made up of six air-hole rings, numbered 1, 2, ... , 6, respectively. Mode coupling is primarily caused by light scattering, which takes place when transient abnormalities in multimode optical ﬁbers transmit power from one mode to another. Modal dispersion can be decreased and transmission bandwidth increased by using mode coupling [30]. Mode coupling prevents measurements of an optical ﬁber’s fundamental optical characteristics, such as attenuation and bandwidth, from being made until the steady state distribution (SSD) has not yet been fully obtained at length zs. Thus, it is essential to comprehend the ﬁber lengths at which an equilibrium mode distribution (EMD) and SSD are established (EMD is achieved at length Lc). It is of particular interest to explore how wavelength affects GI mPOF’s structural and physical parameters and therefore power ﬂow at different ﬁber lengths. A triangular lattice with pitch Λ holds the air holes in the studied polymer ﬁber. The parabolic RI distribution in the core is a result of the appropriate choice of the air-hole diameters in the four inner air-hole rings. The air-hole diameters in rings 5 and 6 are equal to those in ring 4 (d4 = d5 = d6). This system was simulated using the TI PFE. 1 Introduction We calculated the lengths for achieving the EMD and SSD for multimode GI mPOF with a solid core using launch beam distributions with different radial offsets Δr at different wavelengths λ (the low attenuation windows of POFs). As shown in Figure 1, we proposed that the air holes in the core and cladding be arranged in a grid of triangles with regular pitch Λ. The shorter the GI mPOF’s length at which EMD is attained, the sooner the functional dependency of bandwidth changes from of 1/z to of 1/z1/2 (slower bandwidth decline) [16]. This study is the ﬁrst to examine how wavelength affects power ﬂow in GI mPOF to the best of our knowledge. The numerical results reported in this work are very useful in communication and sensory systems that use multimode GI mPOFs at different wavelengths. using alternative speciﬁcations or materials. PMMA is the material that has been used to make POFs the most frequently up until this time [18–22]. The RI distribution of GI multimode POF gradually decreases from the core axis to the cladding. The POF’s bandwidth and transmission distance can both be increased using this type of RI distribution. To create GI POF, however, advanced doping techniques are required. MOF, often referred to as photonic crystal ﬁber, was successfully proposed in the 1990s [23]. The ﬂexibility of the optical ﬁber is considerably increased by the microstructure of MOFs. Numerous relevant MOF features have been investigated by changing the microstructure [24–27]. Eijkelenborg and associates created the ﬁrst PMMA mPOF in 2001 [28]. The various applications of mPOF then attracted scientiﬁc attention [29, 30]. The core and/or cladding layer of a typical mPOF design can be changed by altering the placement and/ or size (d) of air holes within a concentric ring-like region, as shown in Figure 1. In Figure 1, an mPOF that mimics a GI optical ﬁber features a core with different-sized air holes. GI mPOF offers more latitude in changing the air-hole diameters and pitch than typical GI POF, which calls for complex doping methods. Additionally, for communication purposes, it has been found that GI mPOF has a wider bandwidth and less loss than traditional GI POF [31]. 3 Time-independent power ﬂow equation The GI optical ﬁbers have the following RI proﬁle: Frontiers in Physics 02 frontiersin.org frontiersin.org Simović et al. 10.3389/fphy.2024.1340505 where aeff = Λ/ 3 √ [33, 34], and nfsm is the effective RI for various core and cladding layers, as determined by combining the Equation 5 with the effective V parameter [33, 34]: TABLE 1 Refractive index n0, n1, n2, n3, n4, n5, and n6, the relative index difference Δ, the core index exponent g, and the maximum principal mode number M at different wavelengths. λ [nm] 633 568 522 476 n0 1.5220 1.5240 1.5260 1.5280 n1 1.5201 1.5232 1.5253 1.5274 n2 1.5145 1.5231 1.5252 1.5273 n3 1.5050 1.5223 1.5246 1.5268 n4 1.4920 1.5099 1.5140 1.5178 n5 1.4920 1.5099 1.5140 1.5178 n6 1.4920 1.5099 1.5140 1.5178 Δ 0.0197 0.0093 0.0079 0.0067 g 2.0 4.5 4.7 5.0 M 24 22 22 22 V λ Λ, d Λ A1 + A2 1 + A3 exp A4λΛ (6) (6) The ﬁtting parameters Ai (i = 1–4) are given as: Ai ai0 + ai1 d Λ bi1 + ai2 d Λ bi2 + ai3 d Λ bi3 (7) (7) The coefﬁcients ai0 to ai3 and bi1 to bi3 (i = 1–4) are given in our previous work [34]. The coefﬁcients ai0 to ai3 and bi1 to bi3 (i = 1–4) are given in our previous work [34]. We employed our approach Eq. 2 on the GI mPOF with the core radius a = 4Λ = 16 μm, where Λ = 4 μm, and the diameter of the ﬁber b = 1 mm. Table 1 displays the core’s refractive index nco at different wavelengths when measured along the ﬁber axis. For Λ = 4 μm and air-hole diameters of the four air-hole rings in the core d1 = 0.6 μm, d2 = 0.7 μm, d3 = 1.3 μm, and d4 = 3.1 μm, the refractive indices n1, n2, n3, and n4, respectively, calculated using Eqs 6, 7 for different wavelengths, are given in Table 1. Parabolic RI distribution Eq. 1 in the core with g = 2.0, 4.5, 4.7, and 5.0 is achieved at λ = 633, 568, 522, and 476 nm, respectively. 3 Time-independent power ﬂow equation The air- hole diameter in the cladding rings 5 and 6 is d4 = d5 = d6, and therefore the refractive index of the cladding is n4 = n5 = n6 = ncl. Table 1, for the GI mPOF under investigation, at various wavelengths, provides the maximum principal mode number M (Eq. 3). The coupling coefﬁcient is D = 1482 1/m [29]. The typical values of D that deﬁne a standard GI POF can be used when modeling the GI mPOF due to the fact that the intensity of mode coupling in all types of POFs is correlated with the polymer core material. An analogous foundation was used to model a silica MOF [31]. n r, λ ( ) nco λ ( ) 1 −2Δ λ ( ) r a g 1/2 0 ≤r ≤a ( ) nco λ ( ) 1 −2Δ λ ( ) ( )1/2 ncl λ ( ) r > a ( ) ⎧ ⎪ ⎪ ⎨ ⎪ ⎪ ⎩ (1) (1) Here nco(λ) is the core’s highest index (measured at the ﬁber axis), ncl(λ) is the cladding’s index, Δ [nco(λ) −ncl(λ)]/nco(λ) is the relative index difference, g is the core index exponent, and a is the core radius. he TI PFE for GI optical ﬁber is [32]: ∂P m, λ, z ( ) ∂z D m ∂P m, λ, z ( ) ∂m + D ∂P2 m, λ, z ( ) ∂m2 (2) The TI PFE for GI optical ﬁber is [32]: (2) where P(m, λ, z) is power in the m-th principal mode (modal group), z is the coordinate along the ﬁber axis, and D is a constant mode coupling coefﬁcient. The maximum principal mode number M(λ) can be calculated as [32]: As an illustration, for λ = 568 nm, Figure 2 shows the development of the normalized output modal power distribution P (m,λ,z), which depends on the length of the ﬁber. Eq. 2 assumes a Gaussian beam P (θ,z) launched with 〈θ〉= 0o for numerical calculations. Results are displayed for radial offsets of Δr = 0, 4, 8, and 12 µm. It can be seen from Figure 2A that at short ﬁber lengths, due to mode coupling, only lower-order modes shift their midpoints of the power distributions to zero (m = 0). 3 Time-independent power ﬂow equation With increasing ﬁber length, higher order modes start to couple, shifting their distributions to m = 0 (Figure 2B). The EMD is obtained by shifting the midpoints of the power distributions of all modes to m = 0 at the coupling length of Lc = 6 m (Figure 2C). Figure 2D shows that SSD is established at z≡zs = 30 m. The lengths Lc and zs at various wavelengths are displayed in Table 2. It can be seen that the maximum principal mode number M drops with decreasing wavelength from λ = 633 to 568 nm, which causes the lengths Lc and zs to decrease. Shorter lengths are required to accomplish EMD and SSD due to the smaller wavelength and fewer propagating modes. The maximum principal mode number maintains the constant value M = 22 with subsequent wavelength reduction from λ = 568 to 522 and ﬁnally to 476 nm, leading to the same lengths Lc = 6 m and zs = 30 m. It is also worth noting that increasing the parameter g with decreasing the wavelength λ, i.e., modiﬁcation of the GI M λ ( ) gΔ λ ( ) g + 2 aknco λ ( ) (3) (3) where k = 2π/λ. The principal mode m excited at the input ﬁber end is [32]: m M Δr a g + θ2 2Δ g+2 ( )/2g (4) (4) where Δr is the radial offset of the launch beam and θ is the launch beam angle. In this work, Equation 2 is solved using the explicit ﬁnite difference method [32]. 5 Conclusion TABLE 2 Lengths Lc (for achieving emd) and Zs (for achieving ssd) at different wavelengths. TABLE 2 Lengths Lc (for achieving emd) and Zs (for achieving ssd) at different wavelengths. In this study, the power ﬂow along a GI mPOF at various wavelengths is examined using the TI PFE. We have demonstrated that the lengths Lc and zs needed to achieve an EMD and an SSD, respectively, in GI mPOF are shorter at λ = 568 nm than they are at λ = 633 nm. The lengths Lc and zs stay constant when the wavelength decreases further from λ = 568 to 522 and then to 476 nm. Therefore, the shorter Lc causes a quicker changeover to the slower bandwidth decrease regime. As a result, a faster bandwidth enhancement in the tested GI mPOF is only anticipated to take place at wavelengths λ = 568 nm as opposed to that at λ = 633 nm. Such a bandwidth improvement is not brought about by additional wavelength reduction. The study’s ﬁndings can be used in communication and sensory systems that use multimode GI mPOFs at different wavelengths, i.e., at different low attenuation windows. Calculating the modal distribution of the GI mPOF used as a component of the optical ﬁber sensory system at a speciﬁc length at different wavelengths is also important. The future research on this type of optical ﬁber should be calculations of bandwidth at different wavelengths, which can be realized by numerically solving the time- dependent power ﬂow equation. distribution toward a step-index distribution, does not lead to longer lengths Lc and zs. This is a consequence of the larger inﬂuence of wavelength λ and maximum principal number M on these two characteristic ﬁber lengths. It is important to notice that mode coupling behavior controls how the GI MOF bandwidth varies with length. A length less than the coupling length Lc has an inversely linear effect on the bandwidth. Beyond this equilibrium length LC, it has a Z −1/2 dependence, though. As a result, a shorter LC would lead to a more rapid transition to a slower bandwidth drop phase [30, 35, 36]. The investigated GI mPOF is predicted to experience a faster bandwidth enhancement at a wavelength of λ = 568 nm than at λ = 633 nm. 5 Conclusion Such an improvement in bandwidth is not achieved by further reducing the wavelength from λ = 568 to 522 and subsequently to 476 nm. 4 Numerical simulation results Light transmission was examined in a multimode GI mPOF with a solid core (Figure 1). The effective V parameter for such a ﬁber is given as: V 2π λ aeff n2 0 −n2 fsm (5) (5) Frontiers in Physics Frontiers in Physics 03 frontiersin.org Simović et al. 10.3389/fphy.2024.1340505 FIGURE 2 Normalized output modal power distribution P (m,λ,z) acquired by numerically solving the TI PFE (2) over a range of radial offsets Δr = 0, 4, 8, and 12 μm at different ﬁber lengths (A) z = 0.2 m, (B) z = 1 m, (C) z = 6 m, and (D) z = 30 m at λ = 568 nm. FIGURE 2 Normalized output modal power distribution P (m,λ,z) acquired by numerically solving the TI PFE (2) over a range of radial offsets Δr = 0, 4, 8, and 12 μm at different ﬁber lengths (A) z = 0.2 m, (B) z = 1 m, (C) z = 6 m, and (D) z = 30 m at λ = 568 nm. Frontiers in Physics Author contributions Science, Technological Development, and Innovations (Agreement No. 451-03-47/2023-01/200122); and a grant from Guangdong Basic and Applied Basic Research Foundation (2021A1515011997). AS: Conceptualization, Methodology, Software, Writing–original draft. SS: Conceptualization, Methodology, Software, Supervision, Writing–original draft, Writing–review and editing. ZW: Funding acquisition, Project administration, Writing–review and editing. BD: Methodology, Software, Writing–original draft. MK: Conceptualization, Methodology, Software, Writing–original draft. LK: Methodology, Software, Writing–original draft. AD: Writing–original draft, Writing–review and editing. KA: Writing–original draft, Writing–review and editing. CC: Writing–original draft, Writing–review and editing, Conceptualization. Data availability statement In contrast to the GI mPOF that we focused on in this investigation, silica MOFs have much weaker mode coupling, resulting in a length Lc between 1.45 and 1.65 km at which an EMD is achieved, and a length zs between 3.30 and 3.80 km for the establishment of an SSD [34]. The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author. Frontiers in Physics 04 frontiersin.org frontiersin.org Simović et al. 10.3389/fphy.2024.1340505 References 13. Chapalo I, Chah K, Gusarov A, Ioannou A, Pospori A, Nan Y-G, et al. Gamma- radiation enhancement of sensing properties of FBGs in a few-mode polymer CYTOP ﬁber. Opt Lett (2023) 48(5):1248–51. doi:10.1364/ol.482936 13. Chapalo I, Chah K, Gusarov A, Ioannou A, Pospori A, Nan Y-G, et al. Gamma- radiation enhancement of sensing properties of FBGs in a few-mode polymer CYTOP ﬁber. Opt Lett (2023) 48(5):1248–51. doi:10.1364/ol.482936 1. Barbio C, Cunha T, Linnartz J, Huijskens F, Koonen T, Tangdiongga E. Passive OFE WDM-over-POF Gigabits per second performance comparison of spatial diversity and spatial multiplexing. J Lightwave Technol (2023) 41(11):3567–76. doi:10.1109/jlt. 2023.3253622 14. Qu H, Chen Z, Gao S, Min R, Woyessa G, Bang O, et al. Femtosecond laser line-by-line tilted Bragg grating inscription in single-mode step-index TOPAS/ ZEONEX polymer optical ﬁber. Opt Lett (2023) 48(6):1438–41. doi:10.1364/ol. 482598 14. Qu H, Chen Z, Gao S, Min R, Woyessa G, Bang O, et al. Femtosecond laser line-by-line tilted Bragg grating inscription in single-mode step-index TOPAS/ ZEONEX polymer optical ﬁber. Opt Lett (2023) 48(6):1438–41. doi:10.1364/ol. 482598 2. Huang O, Shi J, Chi N. Performance and complexity study of a neural network post- equalizer in a 638-nm laser transmission system through over 100-m plastic optical ﬁber. Opt Eng (2022) 61(12):126108. doi:10.1117/1.oe.61.12.126108 3. Koike Y, Muramoto K. Error-free PAM-4 transmission by a high-speed plastic optical ﬁber without forward error correction. Opt Lett (2021) 46:3709–12. doi:10.1364/ ol.433885 3. Koike Y, Muramoto K. Error-free PAM-4 transmission by a high-speed plastic optical ﬁber without forward error correction. Opt Lett (2021) 46:3709–12. doi:10.1364/ ol.433885 15. Pereira L, Marques C, Min R, Woyessa G, Bang O, Varum H, et al. Bragg gratings in ZEONEX microstructured polymer optical ﬁber with 266-nm Nd:YAG laser. IEEE Sensors J (2023) 23(9):9308–16. doi:10.1109/jsen.2023.3259546 4. Wang J, Sun G, Ning Y, Li Z, Zhao L, Wang J, et al. Innovative sensor for refractive index monitoring based on a plastic optical ﬁber side-polished in a ‘V’-bent structure. Meas Sci Technol (2023) 34(5):055114. doi:10.1088/1361-6501/acbab5 16. Inglev R, Woyessa G, Bang O, Janting J. Polymer optical ﬁber modiﬁcation by etching using Hansen solubility parameters—a case study of TOPAS, Zeonex, and PMMA. J Lightwave Technol (2019) 37:4776–83. doi:10.1109/jlt.2019.2919798 17. Woyessa G, Fasano A, Stefani A, Markos C, Nielsen K, Rasmussen HK, et al. Single mode step-index polymer optical ﬁber for humidity insensitive high temperature ﬁber Bragg grating sensors. Opt Express (2016) 24(2):1253–60. doi:10.1364/oe.24.001253 5. Funding The author(s) declare ﬁnancial support was received for the research, authorship, and/or publication of this article. This research was funded by the National Natural Science Foundation of China (62003046, 6211101138); a grant from Ajman University (Grant No. 2023-IRG-ENIT-14); a grant from City University of Hong Kong (Project No. CityU 7004600); a grant from the Serbian Ministry of All claims expressed in this article are solely those of the authors and do not necessarily represent those of their afﬁliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. Conﬂict of interest The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. The author(s) declared that they were an editorial board member of Frontiers, at the time of submission. This had no impact on the peer review process and the ﬁnal decision. References Wang K, Wang Q, Zou L, Su Y, Liu K, Li W, et al. Study on thermal protection and temperature of PMMA plastic optical ﬁber for concentrated sunlight transmission in daylighting. Solar Energy (2023) 253:127–36. doi:10.1016/j. solener.2023.02.015 6. Leal-Junior A, Guo J, Min R, Fernandes AJ, Frizera A, Marques C. Photonic smart bandage for wound healing assessment. Photon Res (2021) 9(3):272–80. doi:10.1364/prj. 410168 6. Leal-Junior A, Guo J, Min R, Fernandes AJ, Frizera A, Marques C. Photonic smart bandage for wound healing assessment. Photon Res (2021) 9(3):272–80. doi:10.1364/prj. 410168 18. He R, Teng C, Kumar S, Marques C, Min R. Polymer optical ﬁber liquid level sensor: a Review. IEEE Sensors J (2022) 22(2):1081–91. doi:10.1109/jsen.2021.3132098 19. Cheng X, Gunawardena DS, Pun CJ, Bonefacino J, Tam H. Single nanosecond- pulse production of polymeric ﬁber Bragg gratings for biomedical applications. Opt Express (2020) 28(22):33573–83. doi:10.1364/oe.408744 7. Zha B, Wang Z, Li L, Hu X, Ortega B, Li X, et al. Wearable cardiorespiratory monitoring with stretchable elastomer optical ﬁber. Biomed Opt. Express (2023) 14: 2260–75. doi:10.1364/boe.490034 7. Zha B, Wang Z, Li L, Hu X, Ortega B, Li X, et al. Wearable cardiorespiratory monitoring with stretchable elastomer optical ﬁber. Biomed Opt. Express (2023) 14: 2260–75. doi:10.1364/boe.490034 20. Kuang R, Ye Y, Chen Z, He R, Savović I, Djordjevich A, et al. Low-cost plastic optical ﬁber integrated with smartphone for human physiological monitoring. Opt Fiber Technol (2022) 71:102947. doi:10.1016/j.yofte.2022.102947 8. Liang G, Tang Y, Chen P, Li J, Yuan Y, Yu S, et al. Polystyrene-ﬁber-rod hybrid composite structure for optical enhancement in quantum-dot-converted light-emitting diodes. ACS Appl Polym Mater. (2022) 4(1):91–99. doi:10.1021/acsapm.1c01022 8. Liang G, Tang Y, Chen P, Li J, Yuan Y, Yu S, et al. Polystyrene-ﬁber-rod hybrid composite structure for optical enhancement in quantum-dot-converted light-emitting diodes. ACS Appl Polym Mater. (2022) 4(1):91–99. doi:10.1021/acsapm.1c01022 21. Hu X, Chen Z, Cheng X, Min Qu H, Caucheteur C, et al. Femtosecond laser point- by-point Bragg grating inscription in BDK-doped step-index PMMA optical ﬁbers. Opt Lett (2022) 47(2):249–52. doi:10.1364/ol.450047 9. Makino K, Akimoto Y, Koike K, Kondo A, Inoue A, Koike Y. Low loss and high bandwidth polystyrene-based graded index polymer optical ﬁber. J Light Technol (2013) 31(14):2407–12. doi:10.1109/jlt.2013.2266671 9. Makino K, Akimoto Y, Koike K, Kondo A, Inoue A, Koike Y. Low loss and high bandwidth polystyrene-based graded index polymer optical ﬁber. J Light Technol (2013) 31(14):2407–12. doi:10.1109/jlt.2013.2266671 22. Frontiers in Physics References Theodosiou A, Min R, Leal-Junior AG, Ioannou A, Frizera A, Pontes MJ, et al. Long period grating in a multimode cyclic transparent optical polymer ﬁber inscribed using a femtosecond laser. Opt Lett (2019) 44(21):5346–9. doi:10. 1364/ol.44.005346 10. Dadabayev R, Malka D. A visible light RGB wavelength demultiplexer based on polycarbonate multicore polymer optical ﬁber. Opt Laser Technol (2019) 116:239–45. doi:10.1016/j.optlastec.2019.03.034 10. Dadabayev R, Malka D. A visible light RGB wavelength demultiplexer based on polycarbonate multicore polymer optical ﬁber. Opt Laser Technol (2019) 116:239–45. doi:10.1016/j.optlastec.2019.03.034 23. Knight JC, Birks TA, Russell P, Atkin DM. All-silica single-mode optical ﬁber with photonic crystal cladding. Opt Lett (1996) 21(19):1547. doi:10.1364/ol.21.001547 11. Zubel MG, Fasano A, Woyessa GT, Min R, Leal AG, Theodosiou A, et al. Bragg gratings inscribed in solid-core microstructured single-mode polymer optical ﬁber drawn from a 3D-printed polycarbonate preform. IEEE Sens J (2020) 20(21):12744–57. doi:10.1109/jsen.2020.3003469 11. Zubel MG, Fasano A, Woyessa GT, Min R, Leal AG, Theodosiou A, et al. Bragg gratings inscribed in solid-core microstructured single-mode polymer optical ﬁber drawn from a 3D-printed polycarbonate preform. IEEE Sens J (2020) 20(21):12744–57. doi:10.1109/jsen.2020.3003469 24. Stefańska K, Majchrowska S, Gemza K, Soboń G, Sotor J, Mergo P, et al. Soliton trapping and orthogonal Raman scattering in a birefringent photonic crystal ﬁber. Opt Lett (2022) 47(16):4183–6. doi:10.1364/ol.463643 12. Leal-Junior A, Theodosiou A, Biazi V, Macedo L, Marques C, Kalli K, et al. Temperature-insensitive curvature sensor with plane-by-plane inscription of off-center tilted bragg gratings in CYTOP ﬁbers. IEEE Sensors J (2022) 22(12):11725–31. doi:10. 1109/jsen.2022.3171039 12. Leal-Junior A, Theodosiou A, Biazi V, Macedo L, Marques C, Kalli K, et al. Temperature-insensitive curvature sensor with plane-by-plane inscription of off-center tilted bragg gratings in CYTOP ﬁbers. IEEE Sensors J (2022) 22(12):11725–31. doi:10. 1109/jsen.2022.3171039 25. Zhao S, Guo R, Zeng Y. Effects of frequency-dependent Kerr nonlinearity on higher-order soliton evolution in a photonic crystal ﬁber with one zero- 25. Zhao S, Guo R, Zeng Y. Effects of frequency-dependent Kerr nonlinearity on higher-order soliton evolution in a photonic crystal ﬁber with one zero- Frontiers in Physics 05 frontiersin.org Simović et al. 10.3389/fphy.2024.1340505 10.3389/fphy.2024.1340505 dispersion wavelength. Phys Rev A (2022) 106(3):033516. doi:10.1103/physreva. 106.033516 31. Lwin R, Barton G, Harvey L, Harvey J, Hirst D, Manos S, et al. Beyond the bandwidth-length product: graded index microstructured polymer optical ﬁber. Appl Phys Lett (2007) 91(19):191119. doi:10.1063/1.2805216 dispersion wavelength. Phys Rev A (2022) 106(3):033516. doi:10.1103/physreva. 106.033516 31. 28. Eijkelenborg MA, Large MCJ, Argyros A, Zagari J, Manos S, Issa NA, et al. Microstructured polymer optical ﬁbre. Opt Express (2001) 9(7):319–27. doi:10.1364/oe. 9.000319 26. Wang C, Lin K, Cao S, Feng G, Wang J, Abdalla AN. Polarized supercontinuum generation in CS2-core all-normal dispersion photonic crystal ﬁber. IEEE Photon J (2022) 14(6):1–7. doi:10.1109/jphot.2022.3223534 30. Min R, Pereira L, Paixao T, Woyessa G, Hu X, Antunes P, et al. Chirped POF Bragg grating production utilizing UV cure adhesive coating for multiparameter sensing. Opt Fiber Technol (2021) 65:102593. doi:10.1016/j.yofte.2021.102593 References Lwin R, Barton G, Harvey L, Harvey J, Hirst D, Manos S, et al. Beyond the bandwidth-length product: graded index microstructured polymer optical ﬁber. Appl Phys Lett (2007) 91(19):191119. doi:10.1063/1.2805216 26. Wang C, Lin K, Cao S, Feng G, Wang J, Abdalla AN. Polarized supercontinuum generation in CS2-core all-normal dispersion photonic crystal ﬁber. IEEE Photon J (2022) 14(6):1–7. doi:10.1109/jphot.2022.3223534 32. Savović S, Simović A, Drljača B, Djordjevich A, Stepniak G, Bunge CA, et al. Power ﬂow in graded index plastic optical ﬁbers. J Lightwave Technol (2019) 37(19):4985–90. doi:10.1109/jlt.2019.2926700 32. Savović S, Simović A, Drljača B, Djordjevich A, Stepniak G, Bunge CA, et al. Power ﬂow in graded index plastic optical ﬁbers. J Lightwave Technol (2019) 37(19):4985–90. doi:10.1109/jlt.2019.2926700 27. Heydarian K, Nosratpour A, Razaghi M. Computational study of wavelength conversion based on XGM by photonic crystal semiconductor optical ampliﬁer. Opt Laser Technol (2022) 156(1):108531. doi:10.1016/j.optlastec.2022.108531 33. Saitoh K, Koshiba M. Empirical relations for simple design of photonic crystal ﬁbers. Opt Express (2005) 13(1):267–74. doi:10.1364/opex.13.000267 28. Eijkelenborg MA, Large MCJ, Argyros A, Zagari J, Manos S, Issa NA, et al. Microstructured polymer optical ﬁbre. Opt Express (2001) 9(7):319–27. doi:10.1364/oe. 9.000319 34. Savović S, Kovačević MS, Simović A, Drljača B, Kuzmanović L, Djordjevich A. Method for investigation of mode coupling in multimode step-index silica photonic crystal ﬁbers. Optik (2021) 246:167728. doi:10.1016/j.ijleo.2021.167728 29. Woyessa G, Pedersen JKM, Fasano A, Nielsen K, Markos C, Rasmussen HK, et al. Zeonex-PMMA microstructured polymer optical FBGs for simultaneous humidity and temperature sensing. Opt Lett (2017) 42(6):1161–4. doi:10.1364/ol.42.001161 29. Woyessa G, Pedersen JKM, Fasano A, Nielsen K, Markos C, Rasmussen HK, et al. Zeonex-PMMA microstructured polymer optical FBGs for simultaneous humidity and temperature sensing. Opt Lett (2017) 42(6):1161–4. doi:10.1364/ol.42.001161 35. Drljača B, Savović S, Kovačević MS, Simović A, Kuzmanović L, Djordjevich A, et al. Theoretical investigation of bandwidth in multimode step-index silica photonic crystal ﬁbers. Photonics (2022) 9(4):214. doi:10.3390/photonics9040214 35. Drljača B, Savović S, Kovačević MS, Simović A, Kuzmanović L, Djordjevich A, et al. Theoretical investigation of bandwidth in multimode step-index silica photonic crystal ﬁbers. Photonics (2022) 9(4):214. doi:10.3390/photonics9040214 30. Min R, Pereira L, Paixao T, Woyessa G, Hu X, Antunes P, et al. Chirped POF Bragg grating production utilizing UV cure adhesive coating for multiparameter sensing. Opt Fiber Technol (2021) 65:102593. doi:10.1016/j.yofte.2021.102593 30. Min R, Pereira L, Paixao T, Woyessa G, Hu X, Antunes P, et al. 27. Heydarian K, Nosratpour A, Razaghi M. Computational study of wavelength conversion based on XGM by photonic crystal semiconductor optical ampliﬁer. Opt Laser Technol (2022) 156(1):108531. doi:10.1016/j.optlastec.2022.108531 29. Woyessa G, Pedersen JKM, Fasano A, Nielsen K, Markos C, Rasmussen HK, et al. Zeonex-PMMA microstructured polymer optical FBGs for simultaneous humidity and temperature sensing. Opt Lett (2017) 42(6):1161–4. doi:10.1364/ol.42.001161 31. Lwin R, Barton G, Harvey L, Harvey J, Hirst D, Manos S, et al. Beyond the bandwidth-length product: graded index microstructured polymer optical ﬁber. Appl Phys Lett (2007) 91(19):191119. doi:10.1063/1.2805216 References Chirped POF Bragg grating production utilizing UV cure adhesive coating for multiparameter sensing. Opt Fiber Technol (2021) 65:102593. doi:10.1016/j.yofte.2021.102593 36. Savović S, Simović A, Drljača B, Kovačević MS, Kuzmanović L, Djordjevich A, et al. High bandwidth performance of multimode graded-index microstructured polymer optical ﬁbers. Res Phys (2023) 50:106548. doi:10.1016/j.rinp.2023.106548 06 06 Frontiers in Physics frontiersin.org
https://openalex.org/W2885402161	https://www.revistaspp.org/index.php/pediatria/article/download/440/398, https://dialnet.unirioja.es/descarga/articulo/6958334.pdf	es		Absceso pulmonar primario en pediatría. Reporte de casos y revisión de la literatura	Pediatría	2,018	cc-by	3,419	REPORTE DE CASOS Absceso pulmonar primario en pediatría. Reporte de casos y revisión de la literatura Primary pulmonary abscess in pediatrics. Case reports and a review of the literature Gloria Celeste Samudio Domínguez¹, Lidia María Ortiz Cuquejo¹, Julia Coronel², Irene Centurión², Domingo Pérez Regalado3 RESUMEN ABSTRACT Introducción: el absceso pulmonar es una patología infrecuente en la edad pediátrica, y de etiología diversa. El tratamiento es médico - quirúrgico. Objetivo: revisar la presentación clínica, la bacteriología, las imágenes y la evolución de abscesos pulmonares en la edad pediátrica. Material y Métodos: estudio observacional, descriptivo, realizado de enero de 2010 a setiembre de 2016. Se incluyeron todos los pacientes menores de 16 años, con diagnóstico de absceso pulmonar verificado por tomografía. Se recabaron datos clínicos, hematológicos, bacteriológicos, el tratamiento y la evolución. Resultados: se incluyeron 10 pacientes, con un promedio de edad de 6 años, más frecuente el sexo masculino. Un 80% se desarrolló como complicación de neumonía adquirida en la comunidad (NAC) y el 100% fue primario. El diagnóstico se confirmó por TAC de tórax. El germen aislado fue Staphylococcus aureus en 30%. El pulmón más afectado fue el derecho. Todos curaron con tratamiento médico solamente, utilizándose cefalosporina de tercera generación, unida a un antibiótico antiestafilocócico por más de 4 semanas. No se registraron óbitos ni reingresos. Conclusión: los abscesos pulmonares son poco frecuentes. Pueden presentarse como complicación de NAC o septicemia. Responden bien al tratamiento médico prolongado y pueden presentarse en población pediátrica sin patologías de base. Introduction: Pulmonary abscess is an infrequent pathology in the pediatric age with a diverse etiology. The treatment is medical - surgical. Objective: To review the clinical presentation, bacteriology, images and outcomes of pulmonary abscesses in pediatric patients. Materials and Methods: This was an observational, descriptive study, carried out from January 2010 to September 2016. All patients under 16 years of age, with a diagnosis of pulmonary abscess verified by computerized axial tomography (CT), were included. We collected the relevant clinical, hematological, bacteriological, treatment and outcome information for each case. Results: 10 patients were included. They had an average age of 6 years, more frequently male. 80% developed as a complication of community-acquired pneumonia (CAP) and 100% were primary. The diagnosis was confirmed by chest CT. Staphylococcus aureus was isolated in 30% of cases. The right lung was the most frequently affected. All cases were cured with medical treatment only, using a third-generation cephalosporin, combined with an antistaphylococcal antibiotic, with a treatment course of over 4 weeks. There were no deaths or readmissions. Conclusions: Pulmonary abscesses are rare. They can present as a complication of CAP or septicemia. They respond well to prolonged medical treatment and can occur in pediatric population without underlying pathologies. Palabras clave: absceso pulmonar, neumonía bacteriana, Staphylococcus Aureus, patología asociada. Keywords: Pulmonary abscess, bacterial pneumonia, Staphylococcus aureus, underlying pathology. ¹ Hospital Nacional de Itauguá, Departamento de Pediatría. Itauguá, Paraguay. ² Hospital General de Barrio Obrero, Servicio de Pediatría. Asunción, Paraguay. 3 Universidad del Norte. Asunción, Paraguay. Correspondencia: Gloria Celeste Samudio Domínguez; Correo: gsamudio.samudio@gmail.com Conflicto de Interés: Los autores declararan que no poseer conflicto de interés Recibido: 29/01/2018. Aceptado: 22/04/2018 DOI: https://doi.org/10.31698/ped.45012018008 Pediatr. (Asunción). 2018; 45(1):59-64 (enero - abril) 59 Samudio Domínguez GC, Ortiz Cuquejo LM, Coronel J, Centurión I, Pérez D INTRODUCCIÓN El absceso pulmonar se define como una necrosis licuefactiva del parénquima pulmonar, que mide más de 2 centímetros, y puede afectar a una o varias áreas del parénquima pulmonar. El drenaje de pus a través de una fístula broncopleural, permite que se visualice como una cavidad con nivel hidroaéreo. (1,2) Los abscesos primarios son aquellos que cursan sin lesión pulmonar previa, observándose con más frecuencia en menores de 3 años de edad. Los secundarios ocurren en presencia de situaciones s u b ya c e n t e s t a l e s c o m o f i b r o s i s q u í s t i c a , inmunodeficiencias primarias, malformaciones, entre otras (3,4). Mc Cracken reporta 30 casos en 20 años, pudiendo presentarse por aspiración o diseminación hematógena, y con peor pronóstico (5-7). Hasta 90% son polimicrobianos. Las bacterias anaerobias gram negativas encontradas con más frecuencia son Bacteroides fragilis, Fusobacterium capsulatum y necrophorum, además de anaerobios gram positivos tales como Peptostreptococcus y estreptococos microaerofílicos. Entre las bacterias aeróbicas se pueden citar Pseudomonas aeruginosa, Haemophilus influenza tipo b, Acinetobacter spp, Escherichia coli, y Legionella (8-12); se citan como más frecuentes en niños Staphylococcus aureus, Streptococcus pyogenes, Streptococcus pneumoniae, Klebsiella pneumoniae (13-14). Es una afección rara en pediatría. Tiene predilección por el lóbulo derecho, y su asociación con neumonía es frecuente (15). Actualmente existen pocos reportes de esta patología en la edad pediátrica. Las grandes series de casos en este grupo etáreo corresponden a publicaciones de las décadas del 60 y 70 (5-6). El objetivo de este trabajo es describir las características clínicas y microbiológicas de casos de abscesos pulmonares en edad pediátrica, desde enero 2010 a septiembre de 2016. MATERIALES Y MÉTODOS Estudio retrospectivo, observacional, descriptivo. Se examinaron las historias clínicas de pacientes pediátricos menores de 16 años de edad, de enero de 2010 a septiembre de 2016 en el Hospital Nacional de Itaugua y en el Hospital Barrio Obrero, pertenecientes a la red del Ministerio de Salud Pública y Bienestar 60 Social, egresados con diagnóstico de absceso pulmonar corroborado por Tomografía Axial Computada (TAC) de tórax. Se consideró absceso a toda cavidad intraparenquimatosa de más de 2 centímetros de diámetro, con bordes gruesos, de contenido líquido, con o sin niveles hidroaéreos, acompañado de síntomas compatibles con proceso infeccioso agudo. Los cultivos se realizaron en medios para gérmenes aerobios, no se utilizaron cultivos para anaerobiosis. Criterios de inclusión: todos los pacientes en edad pediátrica con cuadro clínico e imágenes compatibles con absceso pulmonar. Se procedió a recabar de cada ficha clínica las siguientes variables: edad, sexo, síntomas, patología subyacente o asociada, cuadro clínico, etiología bacteriana, y datos referentes a su tratamiento y evolución. RESULTADOS Durante el periodo de 5 años y 9 meses (enero de 2010 a septiembre de 2016) hubo 48.727 egresos hospitalarios y se identificaron 10 pacientes con absceso pulmonar, de los cuales 7 fueron de sexo masculino. La mediana de edad de los pacientes fue de 6 años con rango de 11 meses a 15 años. (DS 4,21). Dos pacientes presentaron antecedentes de patología subyacente; uno con acondroplasia y otro con desnutrición severa (Tabla 1). Tabla 1. Datos generales de los pacientes. N=10 Masculino Femenino Patología de base Antecedentes de infecciones a repetición NAC Choque séptico N 7 3 2 0 8 2 % 70 30 20 0 80 20 NAC: neumonía adquirida en la Comunidad El tiempo promedio de diagnóstico fue de 16 días (15–30 días), dependiendo del motivo de internación. En todos los casos se realizó TAC de tórax, encontrándose imágenes correspondientes a abscesos: paredes gruesas, contenido líquido dentro del parénquima pulmonar, con nivel hidroaéreo (Figuras 1 y 2). El lóbulo pulmonar más afectado fue el derecho, en 80% de los casos. Pediatr. (Asunción), Vol. 45; N° 1; (enero - abril) 2018 Absceso pulmonar primario en pediatría. Reporte de casos y revisión de la literatura Tabla 2. Signos y síntomas más frecuentes en pacientes con diagnóstico de absceso pulmonar. N=10. Fiebre Disnea Tos Dificultad respiratoria Vómitos Distensión abdominal Compromiso del estado general Figura 1. TAC de tórax donde se aprecia absceso pulmonar en lóbulo superior derecho. Nótese nivel hidroaéreo. Número 9 3 8 7 2 1 1 % 90 30 80 70 20 10 10 Los estudios de laboratorio se realizaron en todos los pacientes, al igual que el cultivo de sangre y líquido pleural en medios aeróbicos. Todos los cultivos fueron tomados previo al inicio del tratamiento y se repitieron durante el tratamiento antibiótico. Los resultados se resumen en la Tabla 3. Tabla 3. Descripción laboratorial y topográfica de los casos de abscesos pulmonares. N=10. Caso Hb 1 9,9 2 11 3 10 4 9,1 5 10 6 9,3 7 6,7 8 12,3 9 8 10 11,6 GB 27200 3400 26800 14900 4800 5600 22600 16500 23880 14600 PCR 9,3 29 30 0,8 0,6 0,6 5,5 5 Positivo 19,8 Ph 7,58 7,35 7,44 7,51 7,31 NR 7,52 7,48 NR 7,37 Cultivos* Negativos SAMS -HMC Negativo Negativo SAMR -HMC Negativo SAMR -LP Negativo Negativo Negativo Localización LSD LSD LII LSD LID LSD LID LID LID LII Cultivos*: en sangre y líquido pleural HMC: hemocultivo LP: líquido plural LSD: lóbulo superior derecho LII: lóbulo inferior izquierdo LID: lóbulo inferior derecho SAMS: Staphylococcus aureus meticilino sensible SAMR: Staphylococcus aureus meticilino resistente Llama la atención la leucocitosis (70%) y la leucopenia (20%) observada en los casos estudiados. Los cultivos fueron positivos sólo en 3 casos, correspondiendo todos ellos a Staphylococcus aureus; dos meticilino resistentes y uno meticilino sensible. Figura 2. TAC de tórax donde se aprecia absceso pulmonar en lóbulo superior derecho. Nótese predominancia de componente parietal, sin contenido Los signos y síntomas más frecuentes fueron fiebre, tos y dificultad respiratoria. Ningún paciente presentó vómica, ni hemoptisis. Los signos y síntomas encontrados se detallan en la Tabla 2. El tratamiento fue exclusivamente médico, con la combinación antibiótica empírica de una cefalosporina de tercera generación asociada a clindamicina o vancomicina, con pasaje a vía oral para completar un tiempo promedio de 6 semanas de tratamiento. Todos los pacientes con buena evolución al alta. Ninguno requirió tratamiento quirúrgico. No se registraron óbitos ni reingresos. DISCUSIÓN Los abscesos pulmonares son una entidad poco frecuente en pediatría. La sintomatología se refiere Pediatr. (Asunción), Vol. 45; N° 1; (enero - abril) 2018 61 Samudio Domínguez GC, Ortiz Cuquejo LM, Coronel J, Centurión I, Pérez D principalmente al aparato respiratorio, encontrándose fiebre, tos y dificultad respiratoria (7). Es frecuente su asociación con neumonías adquiridas en la comunidad (15). La edad de presentación de abscesos es variable, siendo la mediana, en nuestra serie de 6 años, concordando con la literatura(16). Los sujetos de nuestra serie presentaron el absceso pulmonar como complicación de una neumonía adquirida en la comunidad en su gran mayoría, o por un cuadro séptico con afectación pulmonar en forma menos frecuente. La persistencia de síntomas respiratorios en forma prolongada debe hacer sospechar de que un absceso pudiera estar complicando el cuadro inicial, por lo que debe investigarse esta posibilidad a través de estudios de imágenes, teniendo cuidado en descartar otros diagnósticos diferenciales que deben ser tenidos en cuenta ante un paciente con sospecha de absceso pulmonar, tales como quiste hidatídico, neumonía necrotizante, malformación quística adenomatosa, bronquiectasia infectada, enfermedad de Churg Strauss, secuestro pulmonar, entre otros (17). La tuberculosis, endémica en nuestro país, también debe ser estudiada (18). La radiografía ha demostrado ser de utilidad en el diagnóstico de esta patología, sin embargo es la TAC de tórax el medio auxiliar de mejor rendimiento, ya sea para la localización o para medir el tamaño y volumen del mismo. En los niños, la radiografía detecta hasta el 41% de los casos de neumonía necrotizante, a diferencia de la TAC cuya sensibilidad es muy superior, siendo por lo tanto imprescindible contar con este método auxiliar, de manera a asegurar el diagnóstico (19,20). En nuestros pacientes, el diagnostico se realizó en todos los casos por medio de la TAC de tórax. Los lóbulos más afectados son los del pulmón derecho (7). Tal como corroboramos en nuestra serie, donde un 80% de los abscesos se ubicó en el lado derecho, aunque el mecanismo para ello no fue claro ya que la patogenia en nuestra serie fue diversa (NAC y vía hematógena) (16). Pueden estar implicados diversos tipos de bacterias aerobias y anaerobias. El aislamiento de gérmenes se 62 realiza a través de cultivos de sangre, o secreción purulenta por punción o broncoscopía. Este último procedimiento no se realiza con frecuencia en la edad pediátrica, por lo que los reportes a este respecto son escasos (14). En general, las aplicaciones de este procedimiento con fines terapéuticos aún son limitadas, por el escaso desarrollo del instrumental accesorio necesario, de fibrobroncoscopios ultrafinos con canal de trabajo que permita su empleo en la edad pediátrica (21). Ninguno de nuestros pacientes fue sometido a broncoscopía para aspiración u otro tipo de procedimiento invasivo. Los gérmenes fueron recuperados por medio de cultivos de sangre o líquido pleural. Los cultivos utilizados en muestras de sangre y líquido pleural, fueron exclusivamente en medios aeróbicos, ya que no contamos con medios que permitan crecimiento de bacterias anaeróbicas en ninguno de los centros donde realizamos el trabajo. Este hecho, unido al uso de antibióticos por la mayoría de los pacientes en el momento de la toma del cultivo, limitó la rentabilidad de obtención de gérmenes. De los aislamientos, todos ellos correspondieron a Staphylococcus aureus, por lo que creemos prudente incluir en el manejo empírico inicial de los pacientes con neumonía adquirida en la comunidad complicadas con derrame pleural o en pacientes con choque séptico de origen dérmico o pulmonar, cobertura para este germen, adecuando la elección del antibiótico a los patrones de sensibilidad locales. Actualmente en nuestro país existen reportes de resistencia creciente a la meticilina en los Staphylococcus aureus de la comunidad (SAMR- CA). En Paraguay se reporta que hasta 43% de MRSA-CA son portadores de leucocidina de Panton Valentine (luk-PV) (22,23,24). Entre los análisis necesarios para el manejo de estos pacientes se incluyen los parámetros hematológicos, gasometría y reactantes de fase aguda. En nuestra serie se encontró tanto leucocitosis como leucopenia, unidos a anemia leve, proteína C reactiva (PCR) elevada en la mayor parte de los casos y alcalosis respiratoria. Todos estos hallazgos son consistentes con un cuadro respiratorio de origen infeccioso bacteriano. Pediatr. (Asunción), Vol. 45; N° 1; (enero - abril) 2018 Absceso pulmonar primario en pediatría. Reporte de casos y revisión de la literatura La respuesta al tratamiento médico prolongado fue buena en todos ellos, sin requerir cirugía, inclusive en los que cursaron con leucopenia. El tratamiento de los abscesos pulmonares no está estandarizado y se basa en la experiencia del médico tratante. La mayoría de los médicos prefieren tratamiento conservador, con cuatro a seis semanas de tratamiento antibiótico. El pasaje a la vía oral debe ser lo más temprano posible, dependiendo de la clínica y el laboratorio del paciente (16,25). Las indicaciones de cirugía incluyen falla de respuesta al tratamiento antibiótico luego de 10 días, tamaño del absceso entre 4 a 8 centímetros de diámetro, pacientes inestables para ser sometidos a lobectomía o abscesos con contenido homogéneo, a fin de evitar la descompresión brusca que podría ocasionar el uso de endoscopía en el drenaje de la colección. Existe hasta 4% de mortalidad con uso de drenaje percutáneo, el cual es mucho menor que el de la cirugía (26). Con uso de endoscopía para drenaje, se cita entre 0 a 9% de mortalidad (12). La cirugía con resección de lóbulo pulmonar se reserva para pacientes con evidencias de necrosis pulmonar, sepsis grave, fístula broncopleural de alto débito o insuficiencia respiratoria aguda que no responde al tratamiento médico (27). No se registró ningún óbito ni reingreso hospitalario durante el tiempo de estudio, por lo que podemos concluir que, a pesar de ser una patología de alta morbilidad, tiene baja mortalidad, tal como lo demuestra nuestra serie. CONCLUSIONES Los abscesos pulmonares no son frecuentes en la edad pediátrica, relacionándose con enfermedades infecciosas agudas bacterianas tales como las NAC y las sepsis. El cuadro clínico se caracterizó por fiebre, tos y disnea. La sospecha diagnóstica se instaló ante la persistencia y/o evolución tórpida de manifestaciones del cuadro infeccioso inicial y se confirmó por TAC. El pulmón derecho fue el más afectado y se aisló con más frecuencia el Staphylococcus aureus en muestras de cultivo. El tratamiento antibiótico empírico inicial debe contener al menos una cefalosporina de tercera generación unida a un antibiótico antiestafilocócico, sin olvidar a los anaerobios. La evolución fue buena solo con tratamiento médico. Debilidades del estudio: Si bien ningún paciente presentó signos que hicieran sospechar de inmunodeficiencia primaria o secundaria, no se realizaron estudios de funcionalidad de neutrófilos en los pacientes; lo cual hubiera sido muy interesante para poder explicar el por qué los mismos desarrollaron abscesos como complicación de patologías relativamente frecuentes en los pacientes pediátricos, tales como la NAC y las sepsis por estafilococos. La falta de cultivos para anaerobios impidió la obtención de este tipo de gérmenes. AGRADECIMIENTOS A la Dra Wilma Basualdo, por sus aportes en la revisión del presente artículo. REFERENCIAS 1. Seo H, Cha SI, Shin KM, Lim J, Yoo S, Lee J ,et al. Focal necrotizing pneumonia is a distinct entity from lung abscess. Respirology 2013; 18(7):1095-100. 2. Mi Suk Choi, Ji Hye Chun, Kyung Suk Lee, Yeong Ho Rha, Sun Hee Choi. Clinical characteristics of lung abscess in children: 15-year experience at two university hospitals Korean J Pediatr 2015;58(12):478-83. 3. Vallejo J, Lahoz F. Absceso pulmonar primario: papel jugado por las bacterias anaerobias. Arch Bronconeumol 1978; 14(4):193-7. 4. Moreira JS, Camargo J J, Felicetti JC, Goldenfun PR, Moreira A L S, Porto NS. Lung abscess: analysis of 252 consecutive cases diagnosed between 1968 and 2004. J. Bras. Pneumol. 2006; 32(2):136-43. 5. Mark PH, Turner JAP. Lung abscess in childhood. Thorax. 1968;23(2):216-20. 6. Mc Cracken GH Jr. Lung abscess in childhood. Hosp Pract 1978;13:35-6. 7. Moreira JS, Camargo JJ, Felicetti JC, Goldenfun PR, Pediatr. (Asunción), Vol. 45; N° 1; (enero - abril) 2018 63 Samudio Domínguez GC, Ortiz Cuquejo LM, Coronel J, Centurión I, Pérez D Moreira ALS, Porto NS. Lung abscess: analysis of 252 consecutive cases diagnosed between 1968 and 2004. J. bras. pneumol. 2006; 32(2):136-43. abscess due to Mycobacterium tuberculosis in an infant. World Journal of Medicine and Medical Science Research 2014;2(2):17-9. 8. Stock CT, Ho VP, Towe C, Pieracci FM, Barie PS. Lung abscess. Surg Infect (Larchmt). 2013;14(3):335-6. 19. Hidalgo A. Radiología de las infecciones fúngicas invasoras del aparato respiratorio. Alberto Hidalgo. Revista Iberoamericana de Micología. 2007;24(1):14–8. 9. Bartlett JG. Anaerobic bacterial infection of the lung. Anaerobe 2012; 18:235-9. 10. Kuhajda I, Zarogoulidis K, Tsirgogianni K, Tsavlis D, Kioumis I, Kosmidis, et al. Lung abscess-etiology, diagnostic and treatment options. Rev Pneumol Clin. 2014;70(6):370-4. 11. Bouyahia O, Jlidi S, Sammoud A. Primitive lung abscess: an unusual situation in children. Paediatr Respir Rev. 2007;8(1):77-84. 12. Marra A, Hillejan L, Ukena D. Management of lung abscess. Zentralbl Chir. 2015; 140 Suppl1:S47-53. 13. Patradoon-Ho P1, Fitzgerald DA. Lung abscess in children. Paediatr Respir Rev. 2007; 8(1):77-84. 14. Chan PC, Huang LM, Wu PS, Chang PY, Yang TT, Lu CY, et al. Clinical management and outcome of childhood lung abscess: a 16-year experience. J Microbiol Immunol Infect. 2005; 38(3):183-8. 15. Krenke K, Krawiec M, Kraj G, Peradzynska J, Krauze A, Kulus M. Risk factors for local complications in children with community-acquired pneumonia. Clin Respir J. 2018; 12(1):253-61. 16. Nombera LJ, Acuña GR, Navarro MH, Caussade L, Zúñiga RS, García BC, et al. Doce casos de absceso pulmonar en pediatría: revisión clínica. Rev. chil. pediatr. 2001; 72(2):128-34. 17. Odev K, Guler İ, Altinok T, Pekcan S, Batur A, Ozbiner H. Cystic and cavitary lung lesions in children: radiologic findings with pathologic correlation. J Clin Imaging Sci. 2013;3:60. 18. Aggarwal A, Aggarwal V, Kumar S, Kumar V. Lung 64 20. Hodina M, Hanquinet S, Cotting J, Schnyder P, Gudinchet F. Imaging of cavitary necrosis in complicated childhood pneumonia. Eur Radiol 2002;12(2):391-6. 21. Cordón MA, Spitaleri G, Pérez Ruiz E, Pérez FJ. Indicaciones de la fibrobroncoscopía en pediatría. En: Broncoscopía pediátrica. 3ra ed. Madrid: Editorial Ergon; 2008 p: 9-17. 22. Guillén R, Carpinelli L, Rodríguez F, Castro H, Quiñónez B, Campuzano A, et al. Staphylococcus aureus adquiridos en la comunidad: caracterización clínica, fenotípica y genotípica de aislados en niños paraguayos.) Rev. Chil Infectol. 2016; 33(6):609-18. 23. Abente S, Carpinelli L, Guillén R, Rodríguez R, Fariña N, Laspina F, et al. Frecuencia de Staphylococcus aureus meticilino resistente y del factor de virulencia PVL en pacientes ambulatorios con infección de piel y partes blandas de Asunción, Paraguay. Mem. Inst. Investig. Cienc. Salud. 2016;14(2):8-16. 24. Samudio-Domínguez GC, Bordón L, D`apollo N, Martínez Bareiro M, Benítez D. Patrones de sensibilidad de Staphylococcus aureus de la comunidad aislados de niños con infecciones de piel y partes blandas. Pediatr. (Asunción) 2015; 42(1):31-6. 25. Rahman A, Rahman M. Clindamycin in Treatment of Lung Abscess in Children. American Journal of Drug Delivery and Therapeutics 2014;1:1001-8. 26. Wali SO. An update on the drainage of pyogenic lung abscesses. Annals of Thoracic Medicine 2012;7(1):3-7. 27. Westphal FL, Lima L C, Netto J C, Tavares E, Andrade E, Silva M. Tratamento cirúrgico de crianças com pneumonia necrosante. J. bras. pneumol. 2010; 36(6):716-23. Pediatr. (Asunción), Vol. 45; N° 1; (enero - abril) 2018
https://openalex.org/W4211138203	https://www.cambridge.org/core/services/aop-cambridge-core/content/view/2977C01EBE408D6CFEA1083D3CCA1331/S0955603600072901a.pdf/div-class-title-court-reports-div.pdf	English	null	Court Reports	Psychiatric bulletin of the Royal College of Psychiatrists/Psychiatric bulletin	1,995	cc-by	1,518	CORRESPONDENCE I would remind collea gues who are just starting out and writing such court reports to remember that they are not only writing their report for use in Court, but also for the benefit of those, who, many years later, have to address the question of safety to release. I should add this letter is not addressed solely to forensic psychiatrists. One of the best reports that I have seen in the last year was written by a general psychiatris clinical experience; members Board praised the report for its useful insights. I regret to say report I have read in the last ye by a forensic psychiatrist! ROBERTREEVES,Fromeside Cli Hill, Stapletan, Bristol BS16 1ED The safest antidepressant(s) Sin This is a subject that d straightforward response, thus, t by Duncan & Taylor (Psychiatric 19, 355-357) is most welcome. In double blind trial, information is cal, gathered from clinical pra pharmacodynamic and pharmaco erties of individual antidepressants. The findings so far are sometim and at best, broad guidelines assist in determining the choice sant for an epileptic. The situati compounded by the fact that take more than one anticonvulsan other psychotropic medication the potential of drug interactio highly variable and unpredictable. In the Psychotropic Drug Dir 1995), amoxapine and maprotili as constituting high risk potentia There is a consensus about th these two antidepressants and MAOIs are probably the safes Rosenstein et al (1993) and Stimmel (1992) and in their authors mention fluoxetine first sant of choice in terms of rela epilepsy. While I am convinced alphabetical rather than a safety it is pertinent to note that, ac Psychotropic Drug Directory, been reported to be associated seizures than other SSRIs. The state that trazodone and sertralin likely to interact with anticonvul also point out that, according turer's information, paroxetine interact with anticonvulsants. Thus, each patient with epilep individualised assessment and ing regarding all relevant issues the paper, recognising that the dependent relationship between sant drugs and seizures, and a Court Reports written by a general psychiatrist with great clinical experience; members of the Parole Board praised the report for its valuable and useful insights. I regret to say that the worst report I have read in the last year was written by a forensic psychiatrist! Sin During the last six years I have been a member of the Parole Board and thus a large number of prisoners' parole dossiers have passed through my hands. In a proportion of these cases psychiatric reports must have been available at the time of sentencing and yet they probably are never used beyond the Court hearing. ROBERTREEVES,Fromeside Clinic, Blackberry Hill, Stapletan, Bristol BS16 1ED ROBERTREEVES,Fromeside Clinic, Blackberry Hill, Stapletan, Bristol BS16 1ED ROBERTREEVES,Fromeside Clinic, Blackberry Hill, Stapletan, Bristol BS16 1ED y g The position with regard to psychiatric reports is this: if there is a discretionary or mandatory life sentence, the Judge sends all the reports in his possession to the Home Office; such reports will then be quoted in the dossier or appear in full. However, where there is no life sentence but a period of years, the reports go no further than the Court. With more sentence planning being undertaken these latter reports could be of value to the prison; undoubtedly they are of value when the Parole Board are deciding whether to release or giving advice to the Home Office (the Parole Board has power to release between 4 and 7 years). h i h The safest antidepressant(s) in epilepsy Sin This is a subject that defies a simple straightforward response, thus, the clarification by Duncan & Taylor (Psychiatric Bulletin, 1995, 19, 355-357) is most welcome. In the absence of double blind trial, information is largely empiri cal, gathered from clinical practice and the pharmacodynamic and pharmacokinetic prop erties of individual antidepressants. p The findings so far are sometimes conflicting and at best, broad guidelines are given to assist in determining the choice of antidepres- sant for an epileptic. The situation is further compounded by the fact that patients often take more than one anticonvulsant and are on other psychotropic medication as well, with the potential of drug interactions that are highly variable and unpredictable. In recent years, when appropriate, I have taken to finishing my defence reports with an impartial sounding sentence along the follow ing lines: "In the event of X being made subject to a probation order, it would be of value if my report could be passed to the Probation Service. On the other hand if the Court decide a term of imprisonment is appropriate, then it would be of value if the defence solicitors, after consulting their client, could send a copy of this report to the Prison Medical Officer and also a copy to the Prison Welfare Officer (Probation Officer)". g y p In the Psychotropic Drug Directory (Bazire, 1995), amoxapine and maprotiline are listed as constituting high risk potential for seizures. There is a consensus about the high risk of these two antidepressants and the fact that MAOIs are probably the safest. CORRESPONDENCE CORRESPONDENCE CORRESPONDENCE Court Reports Sin During the last six years I have been a member of the Parole Board and thus a large number of prisoners' parole dossiers have passed through my hands. In a proportion of these cases psychiatric reports must have been available at the time of sentencing and yet they probably are never used beyond the Court hearing. The position with regard to psychiatric reports is this: if there is a discretionary or mandatory life sentence, the Judge sends all the reports in his possession to the Home Office; such reports will then be quoted in the dossier or appear in full. However, where there is no life sentence but a period of years, the reports go no further than the Court. With more sentence planning being undertaken these latter reports could be of value to the prison; undoubtedly they are of value when the Parole Board are deciding whether to release or giving advice to the Home Office (the Parole Board has power to release between 4 and 7 years). In recent years, when appropriate, I have taken to finishing my defence reports with an impartial sounding sentence along the follow ing lines: "In the event of X being made subject to a probation order, it would be of value if my report could be passed to the Probation Service. On the other hand if the Court decide a term of imprisonment is appropriate, then it would be of value if the defence solicitors, after consulting their client, could send a copy of this report to the Prison Medical Officer and also a copy to the Prison Welfare Officer (Probation Officer)". Colleagues writing reports who do not already do this might consider it worthwhile to insert some similar sentence at the end of their reports. I accept the reports may not always follow the prisoner but I know from my experience they do sometimes. Our profession has an important part to play in making the system work, particularly for life sentence prisoners who are beyond tariff. Deterrence and retribu tion have then been satisified and the issue is solely dangerousness. ROBERTREEVES,Fromeside Clinic, Blackberry Hill, Stapletan, Bristol BS16 1ED In quoting Rosenstein et al (1993) and Showron & Stimmel (1992) and in their summary, the authors mention fluoxetine first as antidepres- sant of choice in terms of relative safety in epilepsy. While I am convinced that this is an alphabetical rather than a safety rating, I think it is pertinent to note that, according to the Psychotropic Drug Directory, fluoxetine has been reported to be associated with more seizures than other SSRIs. The authors also state that trazodone and sertraline may be less likely to interact with anticonvulsants. May I also point out that, according to manufac turer's information, paroxetine also does not interact with anticonvulsants. Th h i i h il i Colleagues writing reports who do not already do this might consider it worthwhile to insert some similar sentence at the end of their reports. I accept the reports may not always follow the prisoner but I know from my experience they do sometimes. Our profession has an important part to play in making the system work, particularly for life sentence prisoners who are beyond tariff. Deterrence and retribu tion have then been satisified and the issue is solely dangerousness. I would remind collea gues who are just starting out and writing such court reports to remember that they are not only writing their report for use in Court, but also for the benefit of those, who, many years later, have to address the question of safety to release. Thus, each patient with epilepsy requires an individualised assessment and close monitor ing regarding all relevant issues mentioned in the paper, recognising that there is a dose- dependent relationship between antidepres- sant drugs and seizures, and also taking into y I should add this letter is not addressed solely to forensic psychiatrists. One of the best reports that I have seen in the last year was 706 Correspondence https://doi.org/10.1192/pb.19.11.706 Published online by Cambridge University Press
https://openalex.org/W4200587438	https://e-archivo.uc3m.es/bitstream/10016/35253/1/Multi-Camera_H_2021.pdf	English	null	Multi-Camera Workflow Applied to a Cultural Heritage Building: Alhambra’s Torre de la Cautiva from the Inside	Heritage	2,021	cc-by	18,125	Anto J. Benítez 1,2 , Xose Prieto Souto 1,2,, Manuel Armenteros 1,3, Esteban M. Stepanian 1, Rubén Cantos 1, Miguel García-Villaraco 3, Jaime Solano 2 and Ángela Gómez Manzanares 4 1 Department of Communication, Universidad Carlos III of Madrid, 28903 Madrid, Spain; abenitez@hum.uc3m.es (A.J.B.); marmente@hum.uc3m.es (M.A.); estepani@hum.uc3m.es (E.M.S.); rcantos@hum.uc3m.es (R.C.) 4 Department of Optics, University Complutense of Madrid, 28040 Madrid, Spain; anggomez@ucm.es Correspondence: japsouto@hum.uc3m.es 4 Department of Optics, University Complutense of Madrid, 28040 Madrid, Spain; anggomez@ucm.es * Correspondence: japsouto@hum.uc3m.es Abstract: It is increasingly necessary to generate accessible and navigable digital representations of historical or heritage buildings. This article explains the workﬂow that was applied to create such a digital component for one of the least accessible areas of the Alhambra palace in Granada, the so-called Torre de la Cautiva (Tower of the Captive). The main goal of this process was to create affordable, photorealistic 3D models that contribute to the dissemination of cultural heritage, the decision making for its conservation and restoration, and public engagement and entertainment. With enough preparation, the time spent gathering data following a Structure from Motion (SfM) approach can be signiﬁcantly reduced by using a multi-camera (low cost DSLR) photogrammetric strategy. Without the possibility of artiﬁcial lighting, it was essential to use RAW images and calibrate the color in the scene for material and texture characterization. Through processing, the amount of data was reduced by optimizing the model’s topology. Thus, a photorealistic result was obtained that could be managed and visualized in immersive Visual Reality (VR) environments, simulating different historical periods and environmental and lighting conditions. The potential of this method allows, with slight modiﬁcations, the creation of HBIMs and the adaptation to VR systems development, whose current visualization quality is below the resolution of actionable models in rendering engines. Citation: Benítez, A.J.; Prieto Souto, X.; Armenteros, M.; Stepanian, E.M.; Cantos, R.; García-Villaraco, M.; Solano, J.; Gómez Manzanares, Á. Multi-Camera Workﬂow Applied to a Cultural Heritage Building: Alhambra’s Torre de la Cautiva from the Inside. Heritage 2022, 5, 21–41. https://doi.org/10.3390/ heritage5010002 Academic Editors: Daniel Vázquez-Molini and Antonio Álvarez Fernández-Balbuena Received: 28 October 2021 Accepted: 18 December 2021 Published: 22 December 2021 Citation: Benítez, A.J.; Prieto Souto, X.; Armenteros, M.; Stepanian, E.M.; Cantos, R.; García-Villaraco, M.; Solano, J.; Gómez Manzanares, Á. Multi-Camera Workﬂow Applied to a Cultural Heritage Building: Alhambra’s Torre de la Cautiva from the Inside. Heritage 2022, 5, 21–41. https://doi.org/10.3390/ heritage5010002 Keywords: cultural heritage; photogrammetry; workﬂow; Alhambra’s Tower of the Captive heritage heritage Article Anto J. Benítez 1,2 , Xose Prieto Souto 1,2,*, Manuel Armenteros 1,3, Esteban M. Stepanian 1, Rubén Cantos 1, Miguel García-Villaraco 3, Jaime Solano 2 and Ángela Gómez Manzanares 4 heritage heritage 1. Introduction A fundamental concern in the ﬁeld of Cultural Heritage (CH) is the need to document historical monuments as accurately as possible. Before the advent and popularization of digital culture, it was proposed to carry out extensive suitable inventories of each site [1] with photographs and drawings used to capture the nature of these environments, and to guide conservation, restoration, or excavation activities [2,3]. Technological advances have expanded the possibilities of documenting these CH sites, using not only natural light, but also monochromatic light, ultraviolet light, and infrared rays [4]. Subsequently, photogrammetry was speciﬁcally mentioned as a standard procedure to document the properties of cultural interest in the 1987 Charter for the Conservation of Historic Towns and Urban Areas [5]. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional afﬁl- iations. Copyright: © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). From the time atoms began to be converted into bits [6], digitization has become increasingly widespread, allowing not only for assets to be documented for their con- servation, preservation [7] or restoration [8], but also for digital versions to be created. With these, which constitute the Digital Cultural Heritage (DCH), dissemination actions are promoted, play is proposed [9] and access is allowed [7] through virtual tours even in complicated circumstances such as those experienced in recent times, marked by the COVID-19 pandemic [10]. https://www.mdpi.com/journal/heritage Heritage 2022, 5, 21–41. https://doi.org/10.3390/heritage5010002 Heritage 2022, 5 22 Some arguments have been presented against the creation of DCH models. For example, their cost, the complexity of the processes or the specialization required to obtain them [11,12]. These are arguments that have been losing strength with the simpliﬁcation of software platforms, the increase in the power of hardware equipment and the lowering of the cost of both elements [9,13]. At the same time, the need for a multidisciplinary approach has been recognized (The Norms of Quito, 1967). New approaches have emerged that make it possible to produce more complete documentation and expand the possible uses of the 3D models that were obtained. An increasing number of heritage buildings have an informative model that was obtained through Historic Building Information Modeling (HBIM) procedures [9,14]. 1. Introduction Such models are based on a complex and precise methodology [15,16] originating from architectural functionalities. These approaches encourage the construction of parametric models with the integration of many disciplines [16]. Irrespective of their origin, 3D models have been employed in the interest of the conservation and restoration of CH sites, but also for other objectives. Education [12] or entertainment in the form of video game scenarios or Virtual Reality (VR) experiences [17] are examples. Their use with the holistic intention of gathering as much knowledge as possible about a building and reaching their recipients in the form of ‘edutaintment’ is also remarkable [9]. Moreover, the data collected as a part of these processes can help in the restoration of irreplaceable sites after unexpected catastrophes, as recently highlighted by the ﬁre at the Notre Dame cathedral in Paris [18]. Thus, this article is part of the exploration of the possibilities of digital modeling for historical heritage. Workﬂows (or pipelines) that are applicable to photogrammetry are being simpliﬁed, although they are still not fully standardized. Some of them are becoming clearer and simpler, making them accessible to the uninitiated, but they coexist with others that do require specialization. Some of the possible results can still be achieved in very different ways, sometimes by retracing routes that have already been experienced and commented on. The case that is discussed in this article has tried to avoid this as much as possible by using audiovisual procedures and using only digital photography cameras in the capture, with the aim of obtaining a useful model that works as a video game scenario as well as in a conservation and restoration environment. It was launched before some initiatives were published, several of whose elements we agree with [17], and with which we share the view that it will be necessary to continue looking for procedures that are easy to implement, effective, efﬁcient, cheap, and that can be standardized or become universal. In an attempt to gain a wider picture by drawing on a multidisciplinary environment, more than a few cinematography techniques are also suitable for projects of this type [19]. Minimum standards of precision and accuracy or ﬁdelity must be achieved in the multi- spectral characterization of the surfaces in order to obtain the right data for HBIM [3,20]. 1. Introduction During shooting, cinematographic photography is similarly concerned with the on-screen results, which are achieved through adequate attention to exposure and color calibration. As such, in both cases there is a need to take extreme care in the preparation procedures for shooting. DHCs can and should be used for dissemination as they are part of a collective heritage [7,9,21]. The main goal of the audiovisual sector is that the stories engage the audience. Since the same is true for the dissemination of DCH, it would be logical to use the methods and procedures adopted by the former, i.e., deﬁning and following a script prepared for that purpose [17]. Indeed, some of the processes that will be discussed in this text are speciﬁc to the audiovisual industry, such as: pre-production, multi-camera shooting in rigs, care in exposure, the use of RAW ﬁles, re-topology and the integration of lighting and environment. g g In this case, the aim was to achieve immersion and presence, with full awareness of having discarded the real environment, in the sense attributed to it by [22]. The aim was to provide a simple viewing in VR with a level of quality that allows the user to experience photorealistic sensations [23]. Heritage 2022, 5 23 There are ever more platforms available for this type of experience, although the systems still suffer from many limitations, as their manufacturers warn [17]. It is important to emphasize that at the time of starting this project, the most widespread VR devices were still unable to correctly reproduce the resolution and quality of the models that are being obtained through photogrammetry. However, given the vertiginous advances of the last few years, this team has no doubt that within a short time the performance of the systems that allow VR to be experienced will surpass that of most existing scenes and models. The objective should be independent of the platform used. Rather, it will have to refer to the appearance of real materiality of the content, not to the predisposition to which the technology subjects the viewer. It will have more to do with contemplation, with the amazement of being in the middle of a story that leaves a feeling of photorealism. 1.1. La Torre de la Cautiva (The Tower of the Captive) The Alhambra-Generalife complex in Granada, Spain, ﬁgures on the World Heritage List since 1984, according to UNESCO records. The Alhambra has been the ideal setting for artists and writers of Romanticism, inspired on numerous occasions by folklore and medieval historical plots. According to legend, Doña Isabel de Solís, an Andalusian noblewoman, was kidnapped and fell in love with the king of Granada, Muley Hacén. She converted to Islam and adopted the name Zoraya. The outlines of this story were part of Washington Irving’s book ‘A Chronicle of the conquest of Granada’ [25]. A few years later, Francisco Martínez de la Rosa published the historical novel ‘Doña Isabel de Solís, Reina de Granada’ [26], which delves into this story and is the one that justiﬁes the name of the tower. y j The tower is part of the north-northeast walled enclosure and is located between the Torre del Cadí (Tower of the Judge) and the Torre de las Infantas (Tower of the Princesses). It was built in the mid-fourteenth century, in the time of Yusuf I. Its exterior has a sober defensive aspect and a very simple geometry, but inside it is richly decorated. The main hall measures about 5 m × 5 m and its walls contain a good number of epigraphic poems composed by Ibn al-Yayyab (1274–1349). p y yy In addition to the plans drawn up by architects or draftsmen, which are preserved in the Archives of The Alhambra and Generalife Trust, there is also a physical model of part of the interior of the tower, which was made by Enrique Linares in the last quarter of the 19th century, in the Victoria and Albert Museum in London [27]. It weighs 100 kg and is almost one meter high. 1. Introduction It will have to take advantage of the immersive capacity of the narrative device in order to overcome the resistance that the spectator may oppose to letting him or herself be carried away by the story, in the sense that whoever experiences it may reach a state of pathos such as that described by Sergei M. Eisenstein, whereby one is taken out of oneself and immersed in the reality presented on the screen [24]. To achieve this immersion, it was essential to work with lighting and the setting of scenarios with techniques that achieve an optimal degree of integration [12,17]. This article describes all of the processes that were carried out in order to obtain the model of the interior of the Torre de la Cautiva (The Tower of the Captive) of the Alhambra in Granada, to achieve a stereoscopic piece that can be experienced in VR 360◦, according to a script that was set in simulations throughout different epochs. For a proper analysis of the experience, we will ﬁrst present a brief approach to the selected case study and the challenges it posed. Subsequently, the methodology and workﬂows will be explained in greater detail. Finally, the conclusions of this study will be presented and a series of future lines of research will be proposed. 1.2. Case Study (a) (b) Figure 1. (a,b) Surveys with delineation techniques of the profile and floor plans of the Torre de la Cautiva. The main room has balconies on the walls, while the courtyard or secondary room has no direct lateral exits to the exterior. Compiled by the authors based on documents and plans maintained in the collection of the Patronato de la Alhambra y El Generalife (The Alhambra and Generalife Trust): P-1158_2 and P-4615. Figure 1. (a,b) Surveys with delineation techniques of the proﬁle and ﬂoor plans of the Torre de la Cautiva. The main room has balconies on the walls, while the courtyard or secondary room has no direct lateral exits to the exterior. Compiled by the authors based on documents and plans maintained in the collection of the Patronato de la Alhambra y El Generalife (The Alhambra and Generalife Trust): P-1158_2 and P-4615. A script was drafted to create a dramatic scene that would take place in different epochs and, therefore, different architectural and decorative moments would be shown in the same piece. The action was meant to observe a certain historical consistency. Some characters inhabited the stage, but they were not overly prominent. In this case, it was not so important to obtain data that were faultless from the architectural point of view, that is to say, the result did not have to consist of an informative HBIM model. A script was drafted to create a dramatic scene that would take place in different epochs and, therefore, different architectural and decorative moments would be shown in the same piece. The action was meant to observe a certain historical consistency. Some characters inhabited the stage, but they were not overly prominent. In this case, it was not so important to obtain data that were faultless from the architectural point of view, that is to say, the result did not have to consist of an informative HBIM model. After becoming familiar with the first draft of the script (handling and division into temporal sequences), three strategies were combined, one for each line of work. The first consisted of the application of multi-camera photogrammetry using a vertical rig (see Figure 2) for interiors and walls in their current state of preservation. The second consisted of photogrammetry with a giratutto [29] to obtain the characters. Finally, in the third, conventional 3D modeling was used for the setting and assets. 1.2. Case Study The Torre de la Cautiva is one of the parts of the Alhambra that is not usually open to the public. In the summer of 2018, the challenge was posed to produce a pilot piece of the interior of the tower. Its goal was to explore the possibilities of 3D models obtained through Heritage 2022, 5 24 pen affordable passive-sensor data-acquisition techniques, not only as a simulation platform for conservation and restoration, but also as a vehicle for dissemination and engagement. Two rooms were to be covered: the main room, with profuse decoration and ambient light, and the secondary room or courtyard, with simpler decoration and little natural light (see Figure 1b). The result would be used in stereoscopic 3D in the Head Mount Display (HMD). In Milgram and Kishino’s Reproduction Fidelity dimension, the aim was to achieve what they call the “‘ultimate’ graphic rendering”, referred to as “real-time, high-ﬁdelity 3D animation” [28]. g p p through affordable passive-sensor data-acquisition techniques, not only as a simulation platform for conservation and restoration, but also as a vehicle for dissemination and engagement. Two rooms were to be covered: the main room, with profuse decoration and ambient light, and the secondary room or courtyard, with simpler decoration and little natural light (see Figure 1b). The result would be used in stereoscopic 3D in the Head Mount Display (HMD). In Milgram and Kishino’s Reproduction Fidelity dimension, the aim was to achieve what they call the “‘ultimate’ graphic rendering”, referred to as “real- time, high-fidelity 3D animation” [28]. (b) (a) (b) Figure 1. (a,b) Surveys with delineation techniques of the profile and floor plans of the Torre de la Cautiva. The main room has balconies on the walls, while the courtyard or secondary room has no direct lateral exits to the exterior. Compiled by the authors based on documents and plans maintained in the collection of the Patronato de la Alhambra y El Generalife (The Alhambra and Generalife Trust): P-1158_2 and P-4615. Figure 1. (a,b) Surveys with delineation techniques of the proﬁle and ﬂoor plans of the Torre de la Cautiva. The main room has balconies on the walls, while the courtyard or secondary room has no direct lateral exits to the exterior. Compiled by the authors based on documents and plans maintained in the collection of the Patronato de la Alhambra y El Generalife (The Alhambra and Generalife Trust): P-1158_2 and P-4615. axes (pan, tilt, roll), interaction conditions 1.3. Choice of Strategy freedom of use [30]. It would also have a sound program to guide and complement what the viewer discovers. An additional challenge was the time that was available for data capture, only ten hours spread over one afternoon and the following morning, which was a key factor in choosing the method with which to carry out this task. To complicate matters further, surveying methods to secure control points could not be used, nor lighting supports for A wide repertoire of methods is available with which to obtain digital models [12,21], yet the unique nature of each project implies that the solutions employed in earlier strategies are not always the best approaches to adopt, and that the advances in hardware or software are not the only factors to consider [31]. Likely the only generalization that can be made for digital-model-acquisition projects is that there is not yet a methodology that is applicable to the majority of cases. y g p , g g pp photography. 1.3. Choice of Strategy A wide repertoire of methods is available with which to obtain digital models [12,21], yet the unique nature of each project implies that the solutions employed in earlier strategies are not always the best approaches to adopt and that the advances in hardware j y Among the known strategies, Structure from Motion (SfM) [7,11,13,29,32,33] using only digital camera photography is best adapted to the circumstances associated with the data collection in this study, and it has been used without a set order of shooting [7] or even without checkpoints. This feature constituted a good point of support, since we could not count on having active sensors or Total Stations in our work. Moreover, it was known that stereoscopically conﬁgured work reduces the need for a high number of shots [11,34]. strategies are not always the best approaches to adopt, and that the advances in hardware or software are not the only factors to consider [31]. Likely the only generalization that can be made for digital-model-acquisition projects is that there is not yet a methodology that is applicable to the majority of cases. Among the known strategies, Structure from Motion (SfM) [7,11,13,29,32,33] using only digital camera photography is best adapted to the circumstances associated with the data collection in this study, and it has been used without a set order of shooting [7] or even without checkpoints. 1.2. Case Study y After becoming familiar with the ﬁrst draft of the script (handling and division into temporal sequences), three strategies were combined, one for each line of work. The ﬁrst consisted of the application of multi-camera photogrammetry using a vertical rig (see Figure 2) for interiors and walls in their current state of preservation. The second consisted of photogrammetry with a giratutto [29] to obtain the characters. Finally, in the third, conventional 3D modeling was used for the setting and assets. g g Another of the project’s goals, which was aimed at distinguishing this work from those that used similar approaches, was to achieve a look of veracity, or photorealism, by capturing images with low-cost DSLR cameras. The result would be experienced by the audience on a platform that allowed interaction with a freedom of observation on all three axes (pan, tilt, roll), but on principle, without the freedom of displacement. These interaction conditions are known as three degrees of freedom (3DoF) on the scale of Heritage 2022, 5 25 freedom of use [30]. It would also have a sound program to guide and complement what the viewer discovers. W 5 freedom of use [30]. It would also have a sound program to guide and complement what the viewer discovers. W 5 freedom of use [30]. It would also have a sound program to guide and com the viewer discovers. W Figure 2 Diagram of the vertical rig with three cameras Image by the authors Figure 2. Diagram of the vertical rig with three cameras. Image by the authors. Fi 2 Di f th ti l i ith th I b th th Figure 2. Diagram of the vertical rig with three cameras. Image by the authors Another of the project’s goals, which was aimed at distinguishing this work from those that used similar approaches, was to achieve a look of veracity, or photorealism, by capturing images with low-cost DSLR cameras. The result would be experienced by the audience on a platform that allowed interaction with a freedom of observation on all three ( il ll) b i i l i h h f d f di l Th An additional challenge was the time that was available for data capture, only ten hours spread over one afternoon and the following morning, which was a key factor in choosing the method with which to carry out this task. 1.2. Case Study To complicate matters further, surveying methods to secure control points could not be used, nor lighting supports for photography. 2. Methodology 2. Methodology Thi j t This project was approached in three phases that are not necessarily sequential or consecutive: pre-production, production, and post-production. Based on approaches used in the audiovisual industry, these phases were applied to a script and focused on different aspects: inside the tower, obtaining the characters, and the creation of assets. While it is clear that careful preparation is required, as reﬂected elsewhere in the literature [42,43], phases such as Data Capture, Data Processing and Output Presentation can sometimes be found, which either do not consider shot preparation as a relevant step or do not include it in their ﬁrst stage [17,29–31,33]. Due to the importance of the decisions that had to be taken in the preparation of the workﬂow, it is established here as a singular phase, while the presentation of the ﬁnal result is included in post-production as one of the processes that usually make it up [44]. This project was approached in three phases that are not necessarily sequential or consecutive: pre-production, production, and post-production. Based on approaches used in the audiovisual industry, these phases were applied to a script and focused on different aspects: inside the tower, obtaining the characters, and the creation of assets. While it is clear that careful preparation is required, as reflected elsewhere in the literature [42,43], phases such as Data Capture, Data Processing and Output Presentation can sometimes be found, which either do not consider shot preparation as a relevant step or do not include it in their first stage [17,29–31,33]. Due to the importance of the decisions that had to be taken in the preparation of the workflow, it is established here as a singular phase, while the presentation of the final result is included in post-production as one of the processes that usually make it up [44]. The general outline of the procedures (Figure 3) will be broken down below, with special attention to the fact that one of the objectives was speciﬁcally the development of the workﬂow. The general outline of the procedures (Figure 3) will be broken down below, with special attention to the fact that one of the objectives was specifically the development of the workflow. Figure 3. Scheme of the processes employed in the tower project. All the preparative decisions are grouped together in the pre-production phase. Compiled by the authors. Figure 3. Scheme of the processes employed in the tower project. 1.4. Photorrealism 1.4. Photorrealism The great cine The great cinematographer Néstor Almendros used to argue that “we must learn to transcend the model, while respecting it” and that “the stylization of an image can sometimes be more important than history and logic” [38]. Obviously, this philosophy cannot be applied to obtaining data for HBIM, yet it is valid to immerse the public in the illusion of photorealism within an environment deﬁned by a script. Similarly, it is not valid to simply confront the user with a 3D graphic model [23] via an HDM that fails to provide the user with any other immediate references to reality. Rather, it is necessary to go further, to create an impression of authenticity and accuracy in the user [20,39], but in this case one that is diachronic. It must be borne in mind that to do so implies moving with the imagination to the past, and the capacity to compare the result of a photograph of that time [40,41] with what is being experienced at that moment. g g p g transcend the model, while respecting it” and that “the stylization of an image can sometimes be more important than history and logic” [38]. Obviously, this philosophy cannot be applied to obtaining data for HBIM, yet it is valid to immerse the public in the illusion of photorealism within an environment defined by a script. Similarly, it is not valid to simply confront the user with a 3D graphic model [23] via an HDM that fails to provide the user with any other immediate references to reality. Rather, it is necessary to go further, to create an impression of authenticity and accuracy in the user [20,39], but in this case one that is diachronic. It must be borne in mind that to do so implies moving with the imagination to the past, and the capacity to compare the result of a photograph of that time [40,41] with what is being experienced at that moment. axes (pan, tilt, roll), interaction conditions 1.3. Choice of Strategy This feature constituted a good point of support, since we could not count on having active sensors or Total Stations in our work. Moreover, it was known that stereoscopically configured work reduces the need for a high number of shots [11 34] Although positive results had been reported with mobile devices [35,36] (which is later than our data collection), much of the success of the work depended on the ﬂexibility and reliability of the cameras in capturing, and therefore the cell-phone option was ruled out. We considered cell phones as offering insufﬁcient and unreliable control over exposure in the conditions in which we had to work, which we would still argue today [36]. The aim was to interfere as little as possible with the physical environment, so the motto was “get in, do it and get out”, just as in photogrammetry projects that involve the police or forensics, which cannot alter the environment in any way whatsoever [37]. With this remit it was clear that the background preparation for the data capture had to be as thorough as possible. Heritage 2022, 5 26 as 1.4. Photorrealism 1.4. Photorrealism The great cine 2. Methodology 2. Methodology Thi j t All the preparative decisions are grouped together in the pre-production phase. Compiled by the authors. 2.1. Pre-Production Stage The pre-production phase of the pipeline had to ensure that the restrictions that were imposed on the shooting inside the tower were not insurmountable, since the rest of the Heritage 2022, 5 27 lines of work could be readily controlled. After a process of documentation and deliberation about the different possibilities, it was concluded that the only certain method of covering the surfaces in such a short time and without lighting aids was to use a vertical rig of cameras shooting simultaneously (See Figure 2). The preparation of each shot would take more time, but if properly synchronized, the total time required could be divided by the number of cameras used. Initially, calculations were made for a vertical rig consisting of four mid-range DSLR cameras anchored to a rigid, telescoping pole on a tripod. The path of the rig would be parallel to the walls in different passes at different distances, and the relative position of the cameras on the rig would be changed to obtain segments at different heights. According to the specialist literature, this approach has rarely been used for photogrammetry projects and it is much more common to ﬁnd stereoscopic approaches using two cameras [3,32]. g The cameras that were used were four Canon EOS 1200D with 25–80 mm lenses for the main room, and one additional Canon EOS 60D with a 24–70 mm lens for the second room, bringing the total to ﬁve simultaneous cameras. The cameras were set to a resolution of 5184 × 3456 pixels, about 18 MP (megapixels). The sensor of both rooms was identical, and according to the manufacturer had a total of 18.7 MP, but 18 effective MP and a size of 22.3 mm × 14.9 mm with an aspect ratio of 3:2. This means that the sensor captured 232.47 pixels per mm in width and 231.95 in height (dividing the number of pixels by the number of millimeters). In area: 53,919.7 pixels2/mm2. If photographs were taken at 80 cm from the walls, at a focal length of 25 mm, one would be working with just over three pixels/mm. p The problems of color characterization and exposure could be aggravated by the need to use natural light. The fact that the spaces were either diaphanous or almost diaphanous worked in our favor. 2. Methodology 2. Methodology Thi j t The reference for our work was a color chart, such as Color Checker [45], which is often used in ﬁlm shoots. In order to reduce the possibility of occlusions caused by over- or under-exposures, a ﬁxed aperture was used to ensure a sufﬁcient depth of ﬁeld and low sensitivity in order to avoid problems with noise reading, and RAW ﬁles were also used in order to have the possibility of correcting the exposure values later [29,39,45]. Regarding the characters, the script conceived them as static ﬁgures, establishing situations and suggesting activities, but keeping them static so that the scenery would be the most prominent feature. The animation was applied to environmental or weather conditions, so that it would have an impact on the appearance of the surfaces. Research was done on the clothing, styling, and tools with which to dress the models, and they were asked to pose continuously in order to be captured by with photogrammetry methods, but on a giratutto and in controlled and constant lighting conditions. In this phase, research was carried out on the assets that would be used to decorate the main hall. The initial documentation was collected precisely in the Alhambra’s own museum, but a number of paintings set in the Nasrid period were also used. 2.2. Production Phase. Shooting Inside the Torre de la Cautiva Due to the prevailing weather conditions on the days of shooting, i.e., cloudy with clear spells, it was decided to use a vertical rig of three cameras for the main room and four cameras in the courtyard, given the more complicated geometry but simpler decoration of the latter (see Figure 1b). A route was established that ran parallel to the walls, with stops to take shots every 15 cm. At each height, four passes were made in both rooms, starting almost four meters from the parament that was being portrayed and moving closer. To compose the ﬁrst horizontal band, the cameras were placed on a vertical pole on a tripod, with the camera axis parallel to the ground and perpendicular to the wall at approximately 80, 130 and 170 cm. In the second band, the cameras were settled at approximately two meters, 2.35 m and 2.70 m. Another band was not established due to the risk of excessive camera shaking causing a blurring of the pictures at the exposure speeds that were used. Furthermore, the user experience with 3DoF VR indicates that Heritage 2022, 5 28 d at e to the starting point for observation should be at the height of the person experiencing the room, which enables emphasis to be placed on the strong areas in the ﬁnal result. This way of working requires coordination between the teams that are in charge of the script and photogrammetry, as demanded by some authors [17]. p , p starting point for observation should be at the height of the person experiencing the room, which enables emphasis to be placed on the strong areas in the final result. This way of working requires coordination between the teams that are in charge of the script and photogrammetry, as demanded by some authors [17]. The courtyard needed more time to adjust the exposures and shots, and a Canon 5D camera was used to take cover photographs in hand-held mode at the brightest times of the day. These photographs focused on the most awkward details and the most difﬁcult angles, considering that the main hall had arches with muqarnas and other ﬁligree decora- tions. This camera was also used to take photographs of the coffered ceiling, which was particularly difﬁcult due to its inverted boat hull shape, making it very difﬁcult to achieve a sufﬁcient depth of ﬁeld. 2.2. Production Phase. Shooting Inside the Torre de la Cautiva This situation was aggravated by the minimal relief and the homogeneity of the surface. p g y y The courtyard needed more time to adjust the exposures and shots, and a Canon 5D camera was used to take cover photographs in hand-held mode at the brightest times of the day. These photographs focused on the most awkward details and the most difficult angles, considering that the main hall had arches with muqarnas and other filigree decorations. This camera was also used to take photographs of the coffered ceiling, which was particularly difficult due to its inverted boat hull shape, making it very difficult to achieve a sufficient depth of field. This situation was aggravated by the minimal relief and the homogeneity of the surface. The shots of the main room were taken with a focal length of 25mm, at 100 ISO, an aperture of f/8, a manual exposure at 1/6 of a second, and were corrected according to the lighting conditions; manual white balance with presets for dense and light clouds and Color Checker images were taken in both situations (Figure 4). g y The shots of the main room were taken with a focal length of 25mm, at 100 ISO, an aperture of f/8, a manual exposure at 1/6 of a second, and were corrected according to the lighting conditions; manual white balance with presets for dense and light clouds and Color Checker images were taken in both situations (Figure 4). Figure 4. A Color Checker shot near the epigraphic poems on the wall was used to quickly and automatically adjust the color, and to ensure fidelity of the textures and models. Figure 4. A Color Checker shot near the epigraphic poems on the wall was used to quickly and automatically adjust the color, and to ensure ﬁdelity of the textures and models. Figure 4. A Color Checker shot near the epigraphic poems on the wall was used to quickly and automatically adjust the color, and to ensure fidelity of the textures and models. Figure 4. A Color Checker shot near the epigraphic poems on the wall was used to quickly and automatically adjust the color, and to ensure ﬁdelity of the textures and models. In the inner room, being less accessible to natural light, 20 s exposures were taken. A fourth Canon 60D camera with a 25 mm focal length was attached. 2.2. Production Phase. Shooting Inside the Torre de la Cautiva The material was simultaneously saved as 30 Mb RAW files and 8 Mb JPEGs. In the inner room, being less accessible to natural light, 20 s exposures were taken. A fourth Canon 60D camera with a 25 mm focal length was attached. The material was simultaneously saved as 30 Mb RAW ﬁles and 8 Mb JPEGs. To obtain the images of the characters, a setup was made with a turntable with tickers to mark control points, a lighting installation with four flash-type fixtures (4 Linkstar DL- 500D) and three vertical rigs of five Canon EOS 1200D cameras each [29]. Once characterized with period costumes and styling, the models were asked to pose as still as possible while the giratutto was set in motion. After photographing the Color Checker, a shot was taken every 30 degrees (Figure 5a), synchronized to the photographic flash lights To obtain the images of the characters, a setup was made with a turntable with tickers to mark control points, a lighting installation with four ﬂash-type ﬁxtures (4 Linkstar DL-500D) and three vertical rigs of ﬁve Canon EOS 1200D cameras each [29]. Once charac- terized with period costumes and styling, the models were asked to pose as still as possible while the giratutto was set in motion. After photographing the Color Checker, a shot was taken every 30 degrees (Figure 5a), synchronized to the photographic ﬂash lights through an Arduino ad hoc setup (Figure 5b) and triggering the 15 cameras simultaneously. This was done with two poses of the six different models, sometimes changing the characterization. Finally, the shots of three characters were used, one of them in duplicate. Heritage 2022, 5 29 mes em m (a) (b) Figure 5. (a,b). Image of one of the models (a) on the giratutto with the control points and the color checker. In (b), the Arduino device used that is connected to a small turntable is shown. Figure 5. (a,b). Image of one of the models (a) on the giratutto with the control points and the color checker. In (b), the Arduino device used that is connected to a small turntable is shown. (a) (b) igure 5. (a,b). Image of one of the models (a) on the giratutto with the control points and the color hecker. In (b), the Arduino device used that is connected to a small turntable is shown. 2.2. Production Phase. Shooting Inside the Torre de la Cautiva 3 P t P d ti Ph (a) (a) gur 5 (a b) I f f th d (b) b) h (a) a) f Figure 5. (a,b). Image of one of the models (a) on the giratutto with the control points and the color checker. In (b), the Arduino device used that is connected to a small turntable is shown. Figure 5. (a,b). Image of one of the models (a) on the giratutto with the control points and the color checker. In (b), the Arduino device used that is connected to a small turntable is shown. gure 5. (a,b). Image of one of the models (a) on the giratutto with the control points and the color ecker. In (b), the Arduino device used that is connected to a small turntable is shown. 2.3. Post-Production Phase 2.3. Post-Production Phase 3. Post-Production Phase Th fi t t f th Image before and after color correction from the RAW ﬁle. Note how the walls are much darker in the image on the left ((a), before applying corrections) owing to the contrast and how the edges of the column on the balcony are barely visible. When using a RAW ﬁle, the EVs can be modiﬁed to optimize the image before (b) copying it to the photogrammetry software. Meanwhile, camera alignment tests were carried out to check if the software, Reality Capture 1.3, recognized the camera positions well and provided enough homologous points, which in this case was a number close to 22 million. Through adjustments and corrections, a corrected data set was obtained and redundant points were eliminated from the dense point cloud in order to achieve one that was as faithful as possible to the shots. This was necessary because the software is asked to propose a geometric model that is built with triangles, and this model is usually disproportionately large with a huge number of triangles and is very difficult to handle. In this case, about 50 million triangles were counted in the first version of the mesh. Meanwhile, camera alignment tests were carried out to check if the software, Reality Capture 1.3, recognized the camera positions well and provided enough homologous points, which in this case was a number close to 22 million. Through adjustments and corrections, a corrected data set was obtained and redundant points were eliminated from the dense point cloud in order to achieve one that was as faithful as possible to the shots. This was necessary because the software is asked to propose a geometric model that is built with triangles, and this model is usually disproportionately large with a huge number of triangles and is very difﬁcult to handle. In this case, about 50 million triangles were counted in the ﬁrst version of the mesh. With careful planning of the photographic sessions, the photogrammetry software can produce a geometry with sufficient detail and with textures that are true to the original surfaces. Once the cloud of points was obtained through photogrammetry, simplifying the mesh proved to be useful for the subsequent tasks, as recently noted elsewhere [17,29]. However, there are programs available that specialize in this type of task, such as Zbrush, which recreates the mesh in a much simpler form through a process known as re- topology. 2.3. Post-Production Phase 2.3. Post-Production Phase 3. Post-Production Phase Th fi t t f th The first part of the processing (Figure 6), photogrammetry, required an initial step to prepare the photographs, using the RAW files [29], with two objectives: to adjust the color from the images that had been taken from the chart and to try to avoid occlusions due to incorrect exposures (Figure 7a,b). The ﬁrst part of the processing (Figure 6), photogrammetry, required an initial step to prepare the photographs, using the RAW ﬁles [29], with two objectives: to adjust the color from the images that had been taken from the chart and to try to avoid occlusions due to incorrect exposures (Figure 7a,b). The first part of the processing (Figure 6), photogrammetry, required an initial step o prepare the photographs, using the RAW files [29], with two objectives: to adjust the olor from the images that had been taken from the chart and to try to avoid occlusions ue to incorrect exposures (Figure 7a,b). Figure 6. Pipeline for the photogrammetry phase. Compiled by the authors. Figure 6. Pipeline for the photogrammetry phase. Compiled by the authors. Heritage 2022, 5 30 (a) (b) Figure 7. (a,b). Image before and after color correction from the RAW file. Note how the walls are much darker in the image on the left ((a), before applying corrections) owing to the contrast and how the edges of the column on the balcony are barely visible. When using a RAW file, the EVs can be modified to optimize the image before (b) copying it to the photogrammetry software. Figure 7. (a,b). Image before and after color correction from the RAW ﬁle. Note how the walls are much darker in the image on the left ((a), before applying corrections) owing to the contrast and how the edges of the column on the balcony are barely visible. When using a RAW ﬁle, the EVs can be modiﬁed to optimize the image before (b) copying it to the photogrammetry software. (a) (a) (b) (a) (b) Figure 7. (a,b). Image before and after color correction from the RAW file. Note how the walls are much darker in the image on the left ((a), before applying corrections) owing to the contrast and how the edges of the column on the balcony are barely visible. When using a RAW file, the EVs can be modified to optimize the image before (b) copying it to the photogrammetry software. Figure 7. (a,b). 2.3. Post-Production Phase 2.3. Post-Production Phase 3. Post-Production Phase Th fi t t f th As a result, the mesh was eventually reduced to about 200,000 polygons. This process was necessary for two reasons: first to restore areas that have been left with little or no detail [17], which in our case affected the corners of the floor and some small interior sectors of the balcony arches; second, it offers the possibility of creating clean UVs in order to obtain masks that extend the resolution of the textures using multi-UDIM. If the program originally generated an 8K texture for each room, through this procedure it was possible to have that resolution for each of the maps that were generated, 27 in total (see With careful planning of the photographic sessions, the photogrammetry software can produce a geometry with sufﬁcient detail and with textures that are true to the original surfaces. Once the cloud of points was obtained through photogrammetry, simplifying the mesh proved to be useful for the subsequent tasks, as recently noted elsewhere [17,29]. However, there are programs available that specialize in this type of task, such as Zbrush, which recreates the mesh in a much simpler form through a process known as re-topology. As a result, the mesh was eventually reduced to about 200,000 polygons. This process was necessary for two reasons: ﬁrst to restore areas that have been left with little or no detail [17], which in our case affected the corners of the ﬂoor and some small interior sectors of the balcony arches; second, it offers the possibility of creating clean UVs in order to obtain masks that extend the resolution of the textures using multi-UDIM. If the program originally generated an 8K texture for each room, through this procedure it was possible to have that resolution for each of the maps that were generated, 27 in total (see Figure 8a). Heritage 2022, 5 Heritage 2022, 4 FO 31 11 (a) (b) Figure 8. (a,b). UDIMs with textures (a) and masks (b) prepared for back-projection on the optimized model once the re-topology process had been completed. Figure 8. (a,b). UDIMs with textures (a) and masks (b) prepared for back-projection on the optimized model once the re-topology process had been completed. (a) (a) ( ) (b) (b) Figure 8. (a,b). UDIMs with textures (a) and masks (b) prepared for back-projection on the optimized model once the re-topology process had been completed. Figure 8. (a,b). 2.3. Post-Production Phase 2.3. Post-Production Phase 3. Post-Production Phase Th fi t t f th UDIMs with textures (a) and masks (b) prepared for back-projection on the optimized model once the re-topology process had been completed. The textures that were obtained in this way were albedo (from photogrammetry with color and shadow or exposure correction), tile mask, wooden coffered ceiling mask, stone surfaces mask, floor mask, albedo for the artisan scene (because this scene is conceptualized at the time before the room decoration and therefore needs to be colorless), a displacement map, a normal map, a roughness map and a specular map [17,30]. The textures that were obtained in this way were albedo (from photogrammetry with color and shadow or exposure correction), tile mask, wooden coffered ceiling mask, stone surfaces mask, ﬂoor mask, albedo for the artisan scene (because this scene is conceptualized at the time before the room decoration and therefore needs to be colorless), a displacement map, a normal map, a roughness map and a specular map [17,30]. A few days after the capture, the photogrammetry-generated 3D model visualization tests were carried out with the Clarisse iFX 3.5 program. Re-topology was a complex and laborious task, especially the geometry of the arches, in which there were many small details that required significant precision. It was also necessary to intervene in the wooden coffered ceiling, which was difficult to capture due to the lack of light and the lack of detail, which prevented the software from differentiating between sectors and locating the homologous points. Nowadays, in the same programs, there are new algorithms that perform very accurate automatic re-topology, especially for static models, but at the time of this project they did not exist, so everything was done by hand. A few days after the capture, the photogrammetry-generated 3D model visualization tests were carried out with the Clarisse iFX 3.5 program. Re-topology was a complex and laborious task, especially the geometry of the arches, in which there were many small details that required signiﬁcant precision. It was also necessary to intervene in the wooden coffered ceiling, which was difﬁcult to capture due to the lack of light and the lack of detail, which prevented the software from differentiating between sectors and locating the homologous points. Nowadays, in the same programs, there are new algorithms that perform very accurate automatic re-topology, especially for static models, but at the time of this project they did not exist, so everything was done by hand. 2.3. Post-Production Phase 2.3. Post-Production Phase 3. Post-Production Phase Th fi t t f th Craftsman scene in plan (a) and perspective (b) views in wireframe render. It is crucial to arrange the elements to be able to work interactively with quick previews within the layout of the scenes. The modeling and look-de Zbrush. In total about thirty wer the final scenes. Carpets, curtains medieval chair), the latticewor painter’s tools, lamps, etc., were kept in the Alhambra Museum o As mentioned above, these thus, to save time, the integration gradually, starting by obtaining or are in wireframe. Positions, s enough experience to handle sce (a) of the assets were also performed using Maya and worked on, although not all of them were included in cushions and pillows, the brazier, the jamuga (a type of , vases, the rebec, scaffolding, masonry objects and all inspired or taken directly from originals that were the National Archaeological Museum of Madrid. phases do not have to be carried out consecutively and of the characters and assets in the scenery can be done geometry of materials that are not yet fully worked on zes, scales and points of view can be harmonized with e overlays in this way (see Figure 10a,b). (b) The modeling and look-de Zbrush. In total about thirty wer the final scenes. Carpets, curtains medieval chair), the latticewor painter’s tools, lamps, etc., were kept in the Alhambra Museum o As mentioned above, these thus, to save time, the integratio gradually, starting by obtaining or are in wireframe. Positions, s enough experience to handle sce (a) (b) (a) Figure 10. (a,b). Craftsman scene in plan (a) and perspective (b) views in wireframe render. It is crucial to arrange the elements to be able to work interactively with quick previews within the layout of the scenes. Figure 10. (a,b). Craftsman scene in plan (a) and perspective (b) views in wireframe render. It is crucial to arrange the elements to be able to work interactively with quick previews within the layout of the scenes. a) (b) Figure 10. (a,b). Craftsman scene in plan (a) and perspective (b) views in wireframe render. It is crucial to arrange the elements to be able to work interactively with quick previews within the ayout of the scenes. 2.3. Post-Production Phase 2.3. Post-Production Phase 3. Post-Production Phase Th fi t t f th of this project they did not exist, so everything was done by hand. The shading and look-dev processes (Figure 9), which are common in the work of 3D artists in audiovisuals or video games, for example [17], were directly responsible for the desired photorealism in this study, since they controlled the perception of the surface of the objects and considered their reaction to the light illuminating the scene and the The shading and look-dev processes (Figure 9), which are common in the work of 3D artists in audiovisuals or video games, for example [17], were directly responsible for the desired photorealism in this study, since they controlled the perception of the surface of the objects and considered their reaction to the light illuminating the scene and the atmosphere. the objects and considered their reaction to the light illuminating the scene and the atmosphere. The modeling and look-dev of the assets were also performed using Maya and Zbrush. In total about thirty were worked on, although not all of them were included in the ﬁnal scenes. Carpets, curtains, cushions and pillows, the brazier, the jamuga (a type of medieval chair), the latticework, vases, the rebec, scaffolding, masonry objects and painter’s tools, lamps, etc., were all inspired or taken directly from originals that were kept in the Alhambra Museum or the National Archaeological Museum of Madrid. As mentioned above, these phases do not have to be carried out consecutively and thus, to save time, the integration of the characters and assets in the scenery can be done gradually, starting by obtaining a geometry of materials that are not yet fully worked on or are in wireframe. Positions, sizes, scales and points of view can be harmonized with enough experience to handle scene overlays in this way (see Figure 10a,b). Heritage 2022, 5 Heritage 2022, 4 FOR 32 2 Figure 9. Diagram of the workflow of the last phases of post-production, just before obtaining the final artwork. Own elaboration. The modeling and look-dev of the assets were also performed using Maya and Zbrush. In total about thirty were worked on, although not all of them were included in the final scenes. 2.3. Post-Production Phase 2.3. Post-Production Phase 3. Post-Production Phase Th fi t t f th Carpets, curtains, cushions and pillows, the brazier, the jamuga (a type of medieval chair), the latticework, vases, the rebec, scaffolding, masonry objects and painter’s tools, lamps, etc., were all inspired or taken directly from originals that were kept in the Alhambra Museum or the National Archaeological Museum of Madrid. As mentioned above, these phases do not have to be carried out consecutively and thus, to save time, the integration of the characters and assets in the scenery can be done gradually, starting by obtaining a geometry of materials that are not yet fully worked on or are in wireframe. Positions, sizes, scales and points of view can be harmonized with enough experience to handle scene overlays in this way (see Figure 10a,b). (a) (b) Figure 9. Diagram of the workﬂow of the last phases of post-production, just before obtaining the ﬁnal artwork. Own elaboration. Figure 9. Diagram of the workflow of the last phases of post-production, just before obtaining the final artwork. Own elaboration. The modeling and look-dev of the assets were also performed using Maya and Zbrush. In total about thirty were worked on, although not all of them were included in the final scenes. Carpets, curtains, cushions and pillows, the brazier, the jamuga (a type of medieval chair), the latticework, vases, the rebec, scaffolding, masonry objects and painter’s tools, lamps, etc., were all inspired or taken directly from originals that were kept in the Alhambra Museum or the National Archaeological Museum of Madrid. As mentioned above, these phases do not have to be carried out consecutively and thus, to save time, the integration of the characters and assets in the scenery can be done gradually, starting by obtaining a geometry of materials that are not yet fully worked on or are in wireframe. Positions, sizes, scales and points of view can be harmonized with enough experience to handle scene overlays in this way (see Figure 10a,b). (a) (b) Figure 10. (a,b). Craftsman scene in plan (a) and perspective (b) views in wireframe render. It is crucial to arrange the elements to be able to work interactively with quick previews within the layout of the scenes. However, in order to check if the scene works, it is necessary to test with more or less the final textures and a setting that brings the scene together properly, after ensuring that Figure 10. (a,b). 2.3. Post-Production Phase 2.3. Post-Production Phase 3. Post-Production Phase Th fi t t f th Carpets, curtains, cushions and pillows, the brazier, the jamuga (a type of medieval chair), the latticework, vases, the rebec, scaffolding, masonry objects and painter’s tools, lamps, etc., were all inspired or taken directly from originals that were kept in the Alhambra Museum or the National Archaeological Museum of Madrid. As mentioned above, these phases do not have to be carried out consecutively and thus, to save time, the integration of the characters and assets in the scenery can be done gradually, starting by obtaining a geometry of materials that are not yet fully worked on or are in wireframe. Positions, sizes, scales and points of view can be harmonized with enough experience to handle scene overlays in this way (see Figure 10a,b). (a) (b) Figure 10. (a,b). Craftsman scene in plan (a) and perspective (b) views in wireframe render. It is crucial to arrange the elements to be able to work interactively with quick previews within the layout of the scenes. Figure 10. (a,b). Craftsman scene in plan (a) and perspective (b) views in wireframe render. It is crucial to arrange the elements to be able to work interactively with quick previews within the layout of the scenes. The modeling and look-dev of the assets were also performed using Maya and Zbrush. In total about thirty were worked on, although not all of them were included in the final scenes. Carpets, curtains, cushions and pillows, the brazier, the jamuga (a type of medieval chair), the latticework, vases, the rebec, scaffolding, masonry objects and painter’s tools, lamps, etc., were all inspired or taken directly from originals that were kept in the Alhambra Museum or the National Archaeological Museum of Madrid. As mentioned above, these phases do not have to be carried out consecutively and thus, to save time, the integration of the characters and assets in the scenery can be done gradually, starting by obtaining a geometry of materials that are not yet fully worked on or are in wireframe. Positions, sizes, scales and points of view can be harmonized with enough experience to handle scene overlays in this way (see Figure 10a,b). (a) (b) Figure 10. (a,b). Craftsman scene in plan (a) and perspective (b) views in wireframe render. It is crucial to arrange the elements to be able to work interactively with quick previews within the layout of the scenes. Figure 10. (a,b). 2.3. Post-Production Phase 2.3. Post-Production Phase 3. Post-Production Phase Th fi t t f th Craftsman scene in plan (a) and perspective (b) views in wireframe render. It is crucial to arrange the elements to be able to work interactively with quick previews within the layout of the scenes. However, in order to check if the scene works, it is necessary to test with more or less the ﬁnal textures and a setting that brings the scene together properly, after ensuring that the chosen light interacts correctly with the surfaces and the camera positions and animations are well chosen. In addition to other qualities, the view of the integrated scene must not be disjointed. To achieve this, the ﬁrst rendering tests were performed with Arnold. However, each image took between three and four hours to process at 4K with its optimal parameters (Figure 11), which was unacceptable, especially to observe the result of h i i f li h d d k ﬁ l d i i b i l Figure 9. Diagram of the workflow of the last phases of post-production, just before obtaining the final artwork. Own elaboration. Figure 9. Diagram of the workﬂow of the last phases of post-production, just before obtaining the ﬁnal artwork. Own elaboration. Figure 9. Diagram of the workflow of the last phases of post-production, just before obtaining the final artwork. Own elaboration. The modeling and look-dev of the assets were also performed using Maya and Zbrush. In total about thirty were worked on, although not all of them were included in the final scenes. Carpets, curtains, cushions and pillows, the brazier, the jamuga (a type of medieval chair), the latticework, vases, the rebec, scaffolding, masonry objects and painter’s tools, lamps, etc., were all inspired or taken directly from originals that were kept in the Alhambra Museum or the National Archaeological Museum of Madrid. As mentioned above, these phases do not have to be carried out consecutively and thus, to save time, the integration of the characters and assets in the scenery can be done gradually, starting by obtaining a geometry of materials that are not yet fully worked on or are in wireframe. Positions, sizes, scales and points of view can be harmonized with enough experience to handle scene overlays in this way (see Figure 10a,b). Figure 9. Diagram of the workflow of the last phases of post-production, just before obtaining the final artwork. Own elaboration. Figure 9. 2.3. Post-Production Phase 2.3. Post-Production Phase 3. Post-Production Phase Th fi t t f th Diagram of the workﬂow of the last phases of post-production, just before obtaining the ﬁnal artwork. Own elaboration. Figure 9. Diagram of the workflow of the last phases of post-production, just before obtaining the final artwork. Own elaboration. The modeling and look-dev of the assets were also performed using Maya and Zbrush. In total about thirty were worked on, although not all of them were included in the final scenes. Carpets, curtains, cushions and pillows, the brazier, the jamuga (a type of medieval chair), the latticework, vases, the rebec, scaffolding, masonry objects and painter’s tools, lamps, etc., were all inspired or taken directly from originals that were kept in the Alhambra Museum or the National Archaeological Museum of Madrid. As mentioned above, these phases do not have to be carried out consecutively and thus, to save time, the integration of the characters and assets in the scenery can be done gradually, starting by obtaining a geometry of materials that are not yet fully worked on or are in wireframe. Positions, sizes, scales and points of view can be harmonized with enough experience to handle scene overlays in this way (see Figure 10a,b). Figure 9. Diagram of the workflow of the last phases of post-production, just before obtaining the inal artwork. Own elaboration. Figure 9. Diagram of the workﬂow of the last phases of post-production, just before obtaining the ﬁnal artwork. Own elaboration. gradually, starting by obtaining a geometry of materials that are not yet fully worked on or are in wireframe. Positions, sizes, scales and points of view can be harmonized with enough experience to handle scene overlays in this way (see Figure 10a,b). Figure 9. Diagram of the workflow of the last phases of post-production, just before obtaining the inal artwork. Own elaboration. Figure 9. Diagram of the workﬂow of the last phases of post-production, just before obtaining the ﬁnal artwork. Own elaboration. gradually, starting by obtaining a geometry of materials that are not yet fully worked on or are in wireframe. Positions, sizes, scales and points of view can be harmonized with enough experience to handle scene overlays in this way (see Figure 10a,b). The modeling and look-dev of the assets were also performed using Maya and Zbrush. In total about thirty were worked on, although not all of them were included in the final scenes. 2.3. Post-Production Phase 2.3. Post-Production Phase 3. Post-Production Phase Th fi t t f th This program was also used to perform the ﬁnal rendering at a resolution of 4096 × 4096, with Quick Time output format without YUV 4:2:2 compression. A conversion to the 360◦3D stereoscopic format of the BT (Bottom-Top) type was performed so that it could be experienced with Oculus Rift glasses, with an appearance similar to that shown in Figure 12a–c. g p , p y, g , the captive were composed, while fine-tuning the light and camera animations and creating transitions between them. With a previous low-quality rendering, the sound program was created, and the voiceover, music, effects and Foley were adjusted. This task, together with the sound synchronization, was performed using DaVinci Resolve 16. This program was also used to perform the final rendering at a resolution of 4096 × 4096, with Quick Time output format without YUV 4:2:2 compression. A conversion to the 360° 3D stereoscopic format of the BT (Bottom-Top) type was performed so that it could be experienced with Oculus Rift glasses, with an appearance similar to that shown in Figure 12a–c. (b) (c) Figure 12. (a,b,c) Scenes of the captive (a), the night storm (b) and the craftsman (c) in 360 3D Stereoscopic BT format. (b) (c) Figure 12. (a,b,c) Scenes of the captive (a), the night storm (b) and the craftsman (c) in 360 3D Stereoscopic BT format. Figure 12. (a–c) Scenes of the captive (a), the night storm (b) and the craftsman (c) in 360 3D Stereoscopic BT format. (c) (c) (a) (a) (a) (a) (b) (b) (c) (c) (a) (a) (b) (b) (c) (c) Figure 12. (a,b,c) Scenes of the captive (a), the night storm (b) and the craftsman (c) in 360 3D Stereoscopic BT format. Figure 12. (a,b,c) Scenes of the captive (a), the night storm (b) and the craftsman (c) in 360 3D Stereoscopic BT format. Figure 12. (a–c) Scenes of the captive (a), the night storm (b) and the craftsman (c) in 360 3D Stereoscopic BT format. 2.3. Post-Production Phase 2.3. Post-Production Phase 3. Post-Production Phase Th fi t t f th Follo i g the sc ipt the sce es of the p ese t d the ight sto the tis d ( ) (a) (b) (b) l b h f f d (b) (b) (a) (a) Figure 11. (a,b) Light tests are essential to observe the integration of surfaces and environmental elements in the scene, as well as the suitability of the arrangement of elements and camera positions. For this, it is necessary to launch renders, which in conventional form can take up to four hours using powerful computers for processing. Figure 11. (a,b) Light tests are essential to observe the integration of surfaces and environmental elements in the scene, as well as the suitability of the arrangement of elements and camera positions. For this, it is necessary to launch renders, which in conventional form can take up to four hours using powerful computers for processing. Figure 11. (a,b) Light tests are essential to observe the integration of surfaces and environmental elements in the scene, as well as the suitability of the arrangement of elements and camera positions. For this, it is necessary to launch renders, which in conventional form can take up to four hours using powerful computers for processing. ll h h f h d h h h d Following the script, the scenes of the present day, the night storm, the artisan and the captive were composed, while fine-tuning the light and camera animations and creating transitions between them. With a previous low-quality rendering, the sound program was created, and the voiceover, music, effects and Foley were adjusted. This task, together with the sound synchronization, was performed using DaVinci Resolve 16. This program was also used to perform the final rendering at a resolution of 4096 × 4096, with Quick Time output format without YUV 4:2:2 compression. A conversion to the 360° 3D stereoscopic format of the BT (Bottom-Top) type was performed so that it could be experienced with Oculus Rift glasses, with an appearance similar to that shown in Figure 12a c Following the script, the scenes of the present day, the night storm, the artisan and the captive were composed, while ﬁne-tuning the light and camera animations and creating transitions between them. With a previous low-quality rendering, the sound program was created, and the voiceover, music, effects and Foley were adjusted. This task, together with the sound synchronization, was performed using DaVinci Resolve 16. 3. Results and Discussion 3.1. Multi-Camera Techniques with Rigs 2.3. Post-Production Phase 2.3. Post-Production Phase 3. Post-Production Phase Th fi t t f th However, in order to check if the scene works, it is necessary to test with more or less the final textures and a setting that brings the scene together properly, after ensuring that However, in order to check if the scene works, it is necessary to test with more or less the ﬁnal textures and a setting that brings the scene together properly, after ensuring that the chosen light interacts correctly with the surfaces and the camera positions and animations are well chosen. In addition to other qualities, the view of the integrated scene must not be disjointed. To achieve this, the ﬁrst rendering tests were performed with Arnold. However, each image took between three and four hours to process at 4K with its optimal parameters (Figure 11), which was unacceptable, especially to observe the result of the animation of lights and cameras and make ﬁnal decisions about its placement on screen. It was possible to reduce the time for each image at 4K to 25 min using Houdini 17.5 with Octane render 2019 1.2.0, which is a GPU-based engine, and a computer equipped with two NVIDIA 2080 TI GPUs. Heritage 2022, 5 33 ini ter ter (a) (b) Figure 11. (a,b) Light tests are essential to observe the integration of surfaces and environmental elements in the scene, as well as the suitability of the arrangement of elements and camera positions. For this, it is necessary to launch renders, which in conventional form can take up to four hours using powerful computers for processing. Figure 11. (a,b) Light tests are essential to observe the integration of surfaces and environmental elements in the scene, as well as the suitability of the arrangement of elements and camera positions. For this, it is necessary to launch renders, which in conventional form can take up to four hours using powerful computers for processing. (a) (b) Figure 11. (a,b) Light tests are essential to observe the integration of surfaces and environmental elements in the scene, as well as the suitability of the arrangement of elements and camera positions. For this, it is necessary to launch renders, which in conventional form can take up to four hours using powerful computers for processing. 3.1. Multi-Camera Techniques with Rigs Using a multi-camera data-capture setup ensures the overlapping of images so that each point appears in more than one, which is known since stereoscopy has been used for the same processes [3,11,32,46]. But it allows us to go a step further, covering a larger area in the same amount of time, relative to the number of cameras at each shooting position. A simple technique to prepare the multi-camera shot is to imagine the space in slices (which would be equivalent to the geographical longitude, the vertical sector to be covered), which are represented by each shooting position of a vertical rig, and segments (the geographical Heritage 2022, 5 34 latitude, the horizontal sector to be covered), which are represented by each camera anchor on the tripod. p There is concern about the minimum cost that needs to be addressed for a photogram- metry project [7,11,29,33,45]. While increasing the number of cameras clearly increases the budget, at current prices, it is still much less than covering the costs of using more photographers, a laser scanner or a Total Station [36]. In this project, mid-range DSLR cameras with about 18 Mpix resolutions and conventional optics were used, but mirrorless, medium-format cameras with higher resolution (thus providing more detail from the same number of shots) and faster speeds, because each shot does not involve mechanical dis- placement, are becoming more affordable. When a large number of shots is involved, time savings are multiplied, especially if bracketing is used to ensure correct exposure, which reduces the possibility of occlusion due to under- or over-exposure. In addition, access to better optics also reduces the exposure time, but raises the budget. Nevertheless, this project was less expensive overall as the costs associated with the number of cameras used were much lower than those that would have been imposed by increasing the number of photographers [45]. Another important issue to foresee is the shooting path. It is often necessary to resort to complex algorithms, such as those of the Traveling Salesman Problem for active sensors [47]. In this case, the chosen route was more or less simple and always parallel to the walls with sufﬁcient shots. To generalize the optimization of the route, one should perhaps resort to artiﬁcial intelligence (AI), which considers factors such as the geometry of the object, the available equipment, the environmental conditions and the demands of the script [17]. 3.1. Multi-Camera Techniques with Rigs Any software used must be able to interpret camera positions and locate sufﬁcient homologous points, thereby avoiding occlusions. In our case, it was important to cover all possible angulations in order to obtain a sufﬁcient number of overlapping points, but equal care must be taken not to oversize the shots, as it makes the project unmanageable. On the other hand, when using a multi-camera setup, coordination errors that may exist are multiplied, so extreme care must be taken in this regard. Other factors are focus and exposure, which are explained below. Some 1900 photographs were taken in the tower (7 cameras × 4 passes × 2 heights × 29–35 rig positions), and some 800 without a tripod. Among them, about 2500 were used for the photogrammetric models. For the characters, about 2150 pictures were taken (6 models × 2 poses × 12 shots × 15 cameras) and about 720 were used. The time required to make adjustments before each shot increases when using the multi-camera technique. The exposure and focus must be set manually, even if some pa- rameters are automatic. Again, affordable technology facilitates the operation, as nowadays it is common to control camera viewﬁnders remotely (both the image displayed and the overlays) as they are the main operations for exposure, focus and shooting. y y p p g Another problem was camera shaking; in order to gain height and ensure angles that might otherwise cause occlusions, the cameras had to be raised as high as the pole would allow without swaying. Maintaining stability with light equipment three meters off the ground was difﬁcult, but essential when the shots exceeded 1/30th of a second. This circumstance arose in virtually all of the cases in our work. However, the results were equally dramatic, enabling views from the coffered ceiling with the viewer positioned four meters above the ﬂoor to be included in the ﬁnal experience. There is no doubt that a robotic system that incorporates AI, which is capable of making some decisions regarding settings and routes and also of detecting and solving possible occlusions, would ultimately simplify and reduce the cost of projects, leading to optimal results. 3.2. The Importance of Preparation in Pre-Production 3.2. The Importance of Preparation in Pre-Production Audiovisual projects such as ﬁlm animation or video games are exceptionally complex, and the procedures they have been developing and turning into knowledge with their experience can certainly be beneﬁcial to apply to a photogrammetry challenge that should Heritage 2022, 5 35 have a photorealistic ﬁnal look. With this conviction, this project has assumed some of their techniques. have a photorealistic ﬁnal look. With this conviction, this project has assumed some of their techniques. q A key factor for the success of the project, given the restrictions that were to be experienced during data collection, was to be aware of the relevance of the pre-production- style decision-making process. What had been studied and agreed upon during this phase signiﬁcantly determined the subsequent capacity for action, since it contained an element that was lacking in both the ﬁlming and post-production stages, i.e., time [48]. In the specialized literature, signiﬁcant relevance is always given to the duration and intensity of this phase [49,50]. The ﬁrst decision was to work on a draft script with a ﬁxed number of sequences. It was prepared according to the capabilities that the project wanted to demonstrate: the creation of environments, the simulation of past scenarios with naturalness, and mastery over the scenery. This mastery was especially directed towards the ﬁdelity of the geometry, the variety in the application of textures, and the animation of light sources. Foresight and the careful preparation of the capture allows for decisions to be made ahead of time in terms of overcoming the main challenges, which allows more time to be dedicated to study and optimization of the workﬂow, thereby avoiding confusion. This factor will also help to deﬁne the areas where authenticity and accuracy are more or less important, so that the script does not conﬂict with the pre-production workﬂow of time optimization. On the contrary, storytelling should be another criterion by which to deﬁne technical parameters such as the number of polygons and their layout [17]. It is important to establish ﬂuid communication between those who develop the script and the members of the team responsible for taking the shots and processing the digital models in order to optimize work times and effort. 3.4. Photorealism 3.4. Photorealism In this sense, the audiovisual industry’s way of proceeding consists of trying to make the integration of the different elements invisible, so that even experts find it difficult to reverse engineer and extract the method that was used to achieve this result. h l d h d h d h l h Sometimes a large number of possibilities for conﬁguration or modiﬁcation can be counterproductive in achieving clarity in the ﬁnal results, even making them less convinc- ing. Nevertheless, such tools make it possible to obtain well-integrated scenes that give the illusion of coming close to what historical reality must have been like, or what could now be interpreted as such. In this sense, the audiovisual industry’s way of proceeding consists of trying to make the integration of the different elements invisible, so that even experts ﬁnd it difﬁcult to reverse engineer and extract the method that was used to achieve this result. Photorealism aims to provide the audience with a product that seems natural, such that studies of user experience and acceptance should be considered in order to determine if this is achieved. Strictly speaking, by definition, no image can be photorealistic if it is intended to portray a past era, as it will resemble, at best, what other previous artists have imagined those realities to be. If 3D artists are given the opportunity to work closely with those in charge of the care of monuments and with specialists such as archeologists and architects, many possibilities open up for their conservation, or at least for the simulation of interventions and recreations. Photorealism aims to provide the audience with a product that seems natural, such that studies of user experience and acceptance should be considered in order to determine if this is achieved. Strictly speaking, by deﬁnition, no image can be photorealistic if it is intended to portray a past era, as it will resemble, at best, what other previous artists have imagined those realities to be. If 3D artists are given the opportunity to work closely with those in charge of the care of monuments and with specialists such as archeologists and architects, many possibilities open up for their conservation, or at least for the simulation of interventions and recreations. In this project, the starting point was a narrative script that placed the story in the realm of fiction. 3.4. Photorealism 3.4. Photorealism Shading and look-dev processes are crucial for making the scenery, assets and charac- ters appear photorealistic on screen, but lighting and environmental effects are also equally important. In this project, a detailed study of each character in the environment, light and optics was carried out (see Figure 13a–c), and such a challenge requires a lot of time with conventional methods, so it is recommended, after the experience, to use a rendering engine for testing. Shading and look-dev processes are crucial for making the scenery, assets and characters appear photorealistic on screen, but lighting and environmental effects are also equally important. In this project, a detailed study of each character in the environment, light and optics was carried out (see Figure 13a–c), and such a challenge requires a lot of time with conventional methods, so it is recommended, after the experience, to use a rendering engine for testing. (b) (c) Figure 13. (a,b,c). Some of the lighting, framing and environment tests that were carried out with the captive character. Note the variations in the depth of field due to the use of different optics in the simulation. Figure 13. (a–c). Some of the lighting, framing and environment tests that were carried out with the captive character. Note the variations in the depth of ﬁeld due to the use of different optics in the simulation. (a) (c) (b) (a) (c) (b) Figure 13. (a,b,c). Some of the lighting, framing and environment tests that were carried out with the captive character. Note the variations in the depth of field due to the use of different optics in the simulation. Figure 13. (a–c). Some of the lighting, framing and environment tests that were carried out with the captive character. Note the variations in the depth of ﬁeld due to the use of different optics in the simulation. Sometimes a large number of possibilities for configuration or modification can be counterproductive in achieving clarity in the final results, even making them less convincing. Nevertheless, such tools make it possible to obtain well-integrated scenes that give the illusion of coming close to what historical reality must have been like, or what could now be interpreted as such. 3.3. Exposure and Color Correction Finally, Heritage 2022, 5 36 he thanks to this set of decisions, it was ensured that the results surpassed what the playback or viewing platforms are currently capable of resolving in terms of ﬂuidity and sharpness. what the playback or viewing platforms are currently capable of resolving in terms of fluidity and sharpness. thanks to this set of decisions, it was ensured that the results surpassed what the playback or viewing platforms are currently capable of resolving in terms of ﬂuidity and sharpness. what the playback or viewing platforms are currently capable of resolving in terms of fluidity and sharpness. 3.3. Exposure and Color Correction In cinema, a scientiﬁc approach to questions concerning the control of color and exposure when shooting has long been employed. It is common for all image- and video- editing programs to implement procedures with externally calibrated measurements [45], using very accurate meters and different procedures to control EVs, as well as employing Color Checker. These procedures, which are also described in the literature specialized in photogrammetry [29,45] were used in this project to solve the color-characterization problem and to avoid occlusions due to exposure problems. By obtaining RAW and compressed JPEG ﬁles simultaneously, a safety margin was established in the acquisition that neutralized the risks associated with working with natural light. In fact, during the days of data capture, the sky was very cloudy in Granada and at times quite dark. The use of RAW ﬁles allowed us to vary the exposure values and, to some extent, to correct the ﬂuctuations and prevent over- or under-exposure. p p In terms of focus, there is always a moment when the image is zoomed in when the circle of confusion is no longer small enough and, therefore, when further scaling up means losing sharpness. Obtaining the correct focus in the shot is very important both for the subsequent deﬁnition of the textures and for the photogrammetry program to determine where the points of interest are, and to be able to identify the homologous points. In this project, according to the previsions made in pre-production, we wanted to ensure a minimum aperture of f/8 in order to obtain a sufﬁcient depth of ﬁeld. In this way, the details of the decorations (4 cm fretwork or longer muqarnas) could be captured without the need to correct the focus at each rig position. Such a decision implied a trade-off that increased the exposure time, which ran the risk of blurring the edges and consequently of increasing the time needed for stabilization and exposure. This time can be carefully calculated so that it is less than the time that would be invested in an eventual focus correction. An alternative solution would be to increase the sensitivity, which was kept at 100 ASA, but such an operation would have the consequence of increasing the noise in the images, which is detrimental to the unequivocal obtention of homologous points. 3.5. Re-Topology Coinciding in time with the work presented in this article, but more diligent in publish- ing their results, the text ‘New realities for Canada’s Parliament: A workﬂow for preparing Heritage BIM for game engines and Virtual Reality’ [17] already explains clearly and sufﬁ- ciently the importance of re-topology in the processes of building 3D models and, therefore, in photogrammetry projects. It does not deny that it could always have been photographed better and more effectively and, as is proposed in the mentioned research, that it would have been helpful to have a previous understanding of the script with respect to the points of most interest to the viewers, in order to make decisions during the data capture and processing that would have helped in the composition and the rendering. Without going into the depths of the analysis, the 3D models that were directly produced by the data from photogrammetry platforms far exceeded the dimensions that were manageable by graphics cards and processors, however powerful they may have been, and yet such a profusion of data does not result in improved accuracy or ﬁdelity, nor photorealism, as we have seen. An intermediate process, re-topology, is necessary to simplify the models, retouch possible occlusions or imperfections and produce textures that, once back-projected onto the new model, allow greater detail that is more similar to the real scenarios, and much greater ﬂuidity in the interaction with the scenes. The speciﬁc issue of viewer interaction with the scene also needs to be addressed, as there are speciﬁc methods that can improve the performance of the platforms. At the end of 2015, Nick Kraaman explained in his Headjack blog [51] the progress made by different professionals to develop scenes in which viewers would be immersed and equipped with HMD. One of the points made was the possibility of covering up areas of inactivity with static images instead of video, which would make interaction in the other areas much smoother and take up far fewer system resources. There is no doubt that platforms will continue to improve, as will hardware devices, and it is clear that, by knowing the geography of the scene well and foreseeing the points of greatest interest, it will be possible to make better use of resources. 3.4. Photorealism 3.4. Photorealism However, under certain conditions and scales, the use of photorealism is In this project, the starting point was a narrative script that placed the story in the realm of ﬁction. However, under certain conditions and scales, the use of photorealism is not incompatible with the possibility of obtaining the necessary and sufﬁcient data to make the HBIM of the heritage sites usable by conservators and restorers, from a data-collection method that is more or less accessible and rigorous with the color parameters and with elements of anchorage or location. For the objective of photorealism, accuracy and ﬁdelity are not as important as the feeling of naturalness and coherence with the imagination of the audience. It is true that the desire for photorealism pushes a project towards obtaining an excessive amount of data, thereby generating problems of manageability. For now, not all systems and formats can handle all of these models and sometimes the rendering capacity is insufﬁcient. However, if the scenes have been set up with consistency and are ambitious, the renderings can be improved and adapted to systems with better performance in the near future. Heritage 2022, 5 37 Author Contributions: Conceptualization, A.J.B., X.P.S., M.A.; methodology, A.J.B., X.P.S., M.A.; validation, A.J.B., X.P.S., M.A., E.M.S., J.S., Á.G.M.; investigation, A.J.B., X.P.S., M.A., Á.G.M.; re- sources, M.A., M.G.-V.; data curation, M.G.-V., J.S., R.C.; writing—original draft preparation, A.J.B.; writing—review and editing, A.J.B., X.P.S., E.M.S.; visualization, A.J.B., M.A., M.G.-V., R.C.; supervi- sion, A.J.B., X.P.S., M.A., E.M.S., J.S., Á.G.M.; funding acquisition, Á.G.M. All authors have read and agreed to the published version of the manuscript. 4. Conclusions The Torre de la Cautiva is located in one of the walls of the Alhambra in Granada, a complex named a World Heritage Site in 1984, and is one of its enclaves to which the general public does not usually have access. g p y Bearing in mind its historic value and the fact that it can rarely be enjoyed in person, the challenge of creating a photorealistic 3D model of its interior was posed. This initiative was intended to serve both conservators and restorers, and as a DCH site for public dissemination and entertainment. Throughout this article we have explained the workﬂow that achieves this. Regarding the execution of the project, it should be noted that the data collection was carried out with important restrictions: only ten hours to shoot, using only natural light, and with no cables or control points. The model was built with low-cost photogrammetry techniques, using conventional DSLR cameras as passive sensors. The display was planned on 360◦Stereoscopic VR platforms with interactivity and 3DoF on the scale of freedom of action. The script contemplated several scenes that predominantly focused on the past, al- though one of them was set in the present day. The others recreated a night storm, a typical scene of the period in which the interior of the tower was decorated, and a third evoking the routine atmosphere of the captive princess’s lodging. In addition, three lines of work were deﬁned for the script: that of the stage, which should be given maximum prominence, that of the characters with a narrative sense and that of the assets, fulﬁlling the function of ambience and decoration. A speciﬁc procedure was used to obtain data for each of these lines. Without reaching the demands that HBIM informative models may have, a certain historical and traditional coherence with the period scenes was attempted, as well as a certain level of detail and ﬁdelity to the original decoration. Although in agreement with the academic literature in many points and starting from a script and a desire for photorealism, this project also adopted some procedures commonly employed in the audiovisual industry, which is accustomed to this type of complexity. The pre-production process is therefore analyzed in more detail within its workﬂow. 3.6. Integration and the Problem of Previsualization Logic or common sense establishes an order when it comes to the integration of all the elements of the scenes. Once the geometries have been created, the textures made available, the look-dev has been worked on and the lighting, camera positions and eventual environmental effects and animations have been thought of, it is time to see what the scenes will look like. But working according to a script has advantages: the challenges are already estab- lished, the range of possible decisions is narrower, a narrative structure that makes sense can be followed and the dimension of the result is more or less clear from the beginning. It is also possible to ﬁt the pieces together without all the elements being ﬁnished, so that processes can be followed. For example, in this project it was possible to work in sequences, ﬁtting elements into the layout and studying camera trajectories before the textures were applied. As the time will ultimately arrive to check what the ﬁnal result will look like, rendering must be performed. The usual and most sensible thing to do is to choose some frames ﬁrst and process them, and then do the same with small sequences, using deﬁnitive textures (much more laborious for the machines) or materials depending on what you want to test. Scenes with large amounts of information require a lot of calculation time and, there- fore, necessary decisions must be postponed until the end of the computer processes. The experience of this project before and after using render engines for these tasks is clear: the entire scene must be passed to an engine of these characteristics in order to speed up the decision-making process as much as possible before the ﬁnal calculations. The time required for pre-testing is drastically reduced in this way. This assertion is all the more valid when the ﬁnal layout of the scenes is less studied, i.e., if there is no script, it is more essential to carry out tests with immediate visualizations. Heritage 2022, 5 38 4. Conclusions The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results. Conﬂicts of Interest: The authors have no conﬂict of interest to declare. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results. References Angulo-Fornos, R.; Castellano-Román, M. HBIM as Support of Preventive Conservation Actions in Heritage Architecture. Experience of the Renaissance Quadrant Façade of the Cathedral of Seville. Appl. Sci. 2020, 10, 2428. [CrossRef] g g y 15. Angulo-Fornos, R.; Castellano-Román, M. HBIM as Support of Preventive Conservation Actions in Heritage Architecture. E i f th R i Q d t F d f th C th d l f S ill A l S i 2020 10 2428 [C R f] 15. Angulo-Fornos, R.; Castellano-Román, M. HBIM as Support of Preventive Conservation Actions i Experience of the Renaissance Quadrant Façade of the Cathedral of Seville. Appl. Sci. 2020, 10, 2428. [C 16. Murphy, M.; McGovern, E.; Pavia, S. Historic Building Information Modelling—Adding intelligence to laser and image based surveys of European classical architecture. J. Photogramm. Remote Sens. 2013, 76, 89. [CrossRef] 16. Murphy, M.; McGovern, E.; Pavia, S. Historic Building Information Modelling—Adding intelligenc surveys of European classical architecture. J. Photogramm. Remote Sens. 2013, 76, 89. [CrossRef] 16. Murphy, M.; McGovern, E.; Pavia, S. Historic Building Information Modelling Adding intelligence to laser and image based surveys of European classical architecture. J. Photogramm. Remote Sens. 2013, 76, 89. [CrossRef] p y, ; , ; , g g g g surveys of European classical architecture. J. Photogramm. Remote Sens. 2013, 76, 89. [CrossRef] 17. Pybus, C.; Graham, K.; Doherty, J.; Arellano, N.; Fai, S. New Realities for Canada’S Parliament:a Workﬂow for Preparing Heritage Bim for Game Engines and Virtual Reality. Int. Arch. Photogramm. Remote Sens. Spat. Inf. Sci. 2019, XLII-2/W15, 945. [CrossRef] 17. Pybus, C.; Graham, K.; Doherty, J.; Arellano, N.; Fai, S. New Realities for Canada’S Parliament:a Workﬂow for Preparing Heritage Bim for Game Engines and Virtual Reality. Int. Arch. Photogramm. Remote Sens. Spat. Inf. Sci. 2019, XLII-2/W15, 945. [CrossRef] 18. Al-Muqdadi, F. Assessing the Potentials of Heritage Building Information Modelling (Hbim) in Damaged Heritage Reconstruction. In Proceedings of the 36th Annual ARCOM Conference, Leeds, UK, 7–8 September 2020; Scott, L., Neilson, C.J., Eds.; ARCOM: Leeds, UK, 2020; pp. 25–35. 18. Al-Muqdadi, F. Assessing the Potentials of Heritage Building Information Modelling (Hbim) in Damaged Heritage Reconstruction. In Proceedings of the 36th Annual ARCOM Conference, Leeds, UK, 7–8 September 2020; Scott, L., Neilson, C.J., Eds.; ARCOM: Leeds, UK, 2020; pp. 25–35. 19. Benitez, A.J.; Sánchez Cid, M.; Armenteros, M. Fotogrametría en el audiovisual. Perspectivas de evolución. References 1. The Athens Charter for the Restoration of Historic Monuments—1931—International Council on Monuments and Sites. Avail- able online: https://www.icomos.org/en/resources/charters-and-texts/179-articles-en-francais/ressources/charters-and- standards/167-the-athens-charter-for-the-restoration-of-historic-monuments (accessed on 6 December 2021). 2. Icomos. The Venice Charter. 1964. Available online: https://www.icomos.org/venicecharter2004 omos. The Venice Charter. 1964. Available online: https://www.icomos.org/venicecharter2004/ (accessed on 12 ll d d l d l Charter. 1964. Available online: https://www.icomos.org/venicecharter2004/ (accessed on 12 December 2021). 3. Buill, F.; Jordana, J.J.R.; Andrés, M.A.N. Fotogrametría Arquitectónica; Universitat Politecnica de Catalunya: Barcelona, Spain, 2008. 3. Buill, F.; Jordana, J.J.R.; Andrés, M.A.N. Fotogrametría Arquitectónica; Universitat Politecnica de Catalunya: Barcelona, Spain, 2008. 4. Declaration of Rome (1983)—International Council on Monuments and Sites. Available online: https://www.icomos.org/en/ resources/charters-and-texts/179-articles-en-francais/ressources/charters-and-standards/185-declaration-of-rome (accessed on 6 December 2021). 5. Icomos. Charter for the Conservation of Historic Towns and Urban Areas (The Washington Charter). 1987. Available online: https://www.icomos.org/charters/towns_e.pdf (accessed on 12 December 2021). 6. Negroponte, N. El Mundo Digital; Ediciones B: Barcelona, Spain, 1997. 7. Santagati, C.; Lo Turco, M.; Bocconcino, M.M.; Donato, V.; Galizia, M. 3D Models for All: Low-Cost Acquisition through Mobile Devices in Comparison with Image Based Techniques. Potentialities and Weaknesses in Cultural Heritage Domain. Int. Arch. Photogramm. Remote Sens. Spat. Inf. Sci. 2017, XLII-2/W8, 221. [CrossRef] g p f [ ] 8. Remondino, F.; Rizzi, A. Reality-based 3D documentation of natural and cultural heritage sites—Techniques, problems, and examples. Appl. Geomat. 2010, 2, 85. [CrossRef] 9. Banﬁ, F.; Brumana, R.; Stanga, C. A Content-Based Immersive Experience of Basilica of Sant’ambrogio in Milan: From 3d Survey to Virtual Reality. Int. Arch. Photogramm. Remote Sens. Spat. Inf. Sci. 2019, XLII-2/W11, 159. [CrossRef] 10. El-Said, O.; Aziz, H. Virtual Tours a Means to an End: An Analysis of Virtual Tours’ Role in Tourism Recovery Post COVID-19. J. Travel Res. 2021, 0047287521997567. [CrossRef] 11. Westoby, M.J.; Brasington, J.; Glasser, N.F.; Hambrey, M.J.; Reynolds, J.M. ‘Structure-from-Motion’ photogrammetry: A low-cost, effective tool for geoscience applications. Geomorphology 2012, 179, 300. [CrossRef] 12. Bustillo, A.; Alaguero, M.; Miguel, I.; Saiz, J.M.; Iglesias, L.S. A ﬂexible platform for the creation of 3D semi-immersive environments to teach Cultural Heritage. Digit. Appl. Archaeol. Cult. Herit. 2015, 2, 248. [CrossRef] 13. Inzerillo, L. smart SfM: Salinas Archaeological Museum. Int. Arch. Photogramm. Remote Sens. Spat. Inf. [CrossRef] 14. Zita Sampaio, A.; Gomes, A.M.; Sánchez-Lite, A.; Zulueta, P.; González-Gaya, C. Analysis of BIM Methodology Applied to Practical Cases in the Preservation of Heritage Buildings. Sustainability 2021, 13, 3129. [CrossRef] g g y [ ] 15. 21. Pavlidis, G.; Tsiafakis, D.; Koutsoudis, A.; Arnaoutoglou, F.; Tsioukas, V.; Chamzas, C. Preservation of Architectural Heritage through 3D Digitization. Int. J. Archit. Comput. 2007, 5, 221. [CrossRef] 4. Conclusions Since it is difﬁcult to reach an agreement on the meaning of photorealism, especially if the past is to be shown, it was established as a principle of the project that it could be perceived naturally and coherently by the viewer. In this case, moreover, being accurate with geometry and color characterization was assumed as being of the utmost importance. Another common process in the audiovisual industry that was an important part of the project workﬂow was the multi-camera work, which was carried out with vertical rigs and careful planning. In the post-shooting processes, the importance of re-topology to obtain simpler meshes and of UDIM textures to retain great deﬁnition of detail was clear, as was the importance of the script to be able to work separately on the integration of the elements in the scenes, their lighting, environmental effects, and animation. Two key ideas for future work stem from the experience of this project: the convenience of using robotic devices coupled to AI algorithms to optimize decisions during the data acquisition process, and the adaptation of rendering engines, not only for ﬁnal viewing or creation of VR environments, but also for previsualization and tests, since they radically reduce the time required to perform different procedures. Heritage 2022, 5 39 Funding: This work was funded by of the Ministry of Science and Innovation of Spain, project number RTI2018-097633-A-I00, ‘Photonic restoration applied to cultural heritage: Application to Dali’s painting: Two Figures’. Funding: This work was funded by of the Ministry of Science and Innovation of Spain, project number RTI2018-097633-A-I00, ‘Photonic restoration applied to cultural heritage: Application to Dali’s painting: Two Figures’. Institutional Review Board Statement: Not applicable. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Acknowledgments: The authors wish to thank Miopía FX, GVAM SL, Voxel School of Digital Arts and its students, as well as the Patronato de la Alhambra y El Generalife de Granada (Alhambra and Generalife Trust). Acknowledgments: The authors wish to thank Miopía FX, GVAM SL, Voxel School of Digital Arts and its students, as well as the Patronato de la Alhambra y El Generalife de Granada (Alhambra and Generalife Trust). Conﬂicts of Interest: The authors have no conﬂict of interest to declare. References In Los Nuevos Retos en Ciencias Sociales, Artes y Humanidades; Gedisa: Barcelona, Spain, 2018; pp. 31–45. 20. Roussou, M.; Drettakis, G. Photorealism and Non-Photorealism in Virtual Heritage Representation. In Proceedings of the 4th International Symposium on Virtual Reality, Archaeology and Intelligent Cultural Heritage, Brighton, UK, 5–7 November 2003. [CrossRef] 21. Pavlidis, G.; Tsiafakis, D.; Koutsoudis, A.; Arnaoutoglou, F.; Tsioukas, V.; Chamzas, C. Preservation of Architectural Heritage through 3D Digitization. Int. J. Archit. Comput. 2007, 5, 221. [CrossRef] Heritage 2022, 5 40 22. Slater, M.; Wilbur, S. A Framework for Immersive Virtual Environments (FIVE): Speculations on the Role of Presence in Virtual Environments. Presence Teleoperators Virtual Environ. 1997, 6, 603. [CrossRef] p [ ] 23. Shashi, M.; Jain, K. Use of Photogrammetry in 3D Modeling and Visualization of Buildings. ARPN J. Eng. Appl. Sci. 2007, 2, 37. p 23. Shashi, M.; Jain, K. Use of Photogrammetry in 3D Modeling and Visualization of Buildings. ARPN J. E 24 Ei t i S M T í Té i Ci t áﬁ Edi i Ri l M d id S i 1989 p 23. Shashi, M.; Jain, K. Use of Photogrammetry in 3D Modeling and Visualization of Buildings. ARPN J. Eng. Appl. Sci. 2007, 2, 37. 24. Eisenstein, S.M. Teoría y Técnica Cinematográﬁcas; Ediciones Rialp: Madrid, Spain, 1989. 23. Shashi, M.; Jain, K. Use of Photogrammetry in 3D Modeling and Visualization of Buildings. ARPN 24. Eisenstein, S.M. Teoría y Técnica Cinematográﬁcas; Ediciones Rialp: Madrid, Spain, 1989. y g ﬁ p p 25. Irving, W. Manuscript of The Chronicle of the Conquest of Granada: Isabella Stewart Gardner Museum. Available online: https://www.gardnermuseum.org/experience/collection/28622 (accessed on 4 December 2021). 26. Martínez de la Rosa, F. Doña Isabel de Solís, Reina de Granada. 1837. Available online: http://www.cervantesvirtual.com/obra- visor/dona-isabel-de-solis-reyna-de-granada-novela-historica--0/html/ (accessed on 5 November 2021). 27. Marques, S. Mixed media object: Large and fragile structure—Victoria and Albert Museum. Available online: http://www. vam.ac.uk/content/journals/conservation-journal/issue-48/mixed-media-object-large-and-fragile-structure/ (accessed on 6 December 2021). ) hino, F. A Taxonomy of Mixed Reality Visual Displays. IEICE Trans. Inf. Syst. 1994, E77-D, 1–15. ) 28. Milgram, P.; Kishino, F. A Taxonomy of Mixed Reality Visual Displays. IEICE Trans. Inf. Syst. 1994, E77 29. Apollonio, F.I.; Fantini, F.; Garagnani, S.; Gaiani, M. A Photogrammetry-Based Workﬂow for the Accurate 3D Construction and Visualization of Museums Assets. Remote Sens. 2021, 13, 486. [CrossRef] 30. References Dhanda, A.; Reina Ortiz, M.; Weigert, A.; Paladini, A.; Min, A.; Gyi, M.; Su, S.; Fai, S.; Santana Quintero, M. Recreating Cultural Heritage Environments for Vr Using Photogrammetry. Int. Arch. Photogramm. Remote Sens. Spat. Inf. Sci. 2019, XLII-2/W9, 305. [CrossRef] [ ] 31. Tucci, G.; Bonora, V.; Conti, A.; Fiorini, L. Digital Workﬂow for the Acquisition and Elaboration of 3D Data in a Monumental Complex: The Fortress of Saint John the Baptist in Florence. In Proceedings of the 26th International CIPA Symposium 2017, Ottawa, ON, Canada; 2017; pp. 679–686. pp 32. Moyano, J.; Nieto-Julián, J.E.; Bienvenido-Huertas, D.; Marín-García, D. Validation of Close-Range Photogrammetry for Ar- chitectural and Archaeological Heritage: Analysis of Point Density and 3D Mesh Geometry. Remote Sens. 2020, 12, 3571. [CrossRef] [ ] 33. Rahaman, H.; Champion, E. To 3D or Not 3D: Choosing a Photogrammetry Workﬂow for Cultural Heritage Groups. Heritage 2019, 2, 112. [CrossRef] a, Y.; Ponce, J. Accurate, Dense, and Robust Multiview Stereopsis. IEEE Trans. Pattern Anal. Mach. Intell. 201 f] 34. Furukawa, Y.; Ponce, J. Accurate, Dense, and Robust Multiview Stereopsis. IEEE Trans. Pattern Ana [CrossRef] 34. Furukawa, Y.; Ponce, J. Accurate, Dense, and Robust Multiview Stereopsis. IEEE Trans. Pattern Anal. Mach. Intell. 2010, 32, 1362. [CrossRef] 35 R h N B i A L T C t i Fi D t il I l i L C t S C ti St d I t A h Ph t [CrossRef] 35. Rehany, N.; Barsi, A.; Lovas, T. Capturing Fine Details Involving Low-Cost Sensors –a Comparative Study. Int. Arch. Photogramm. [ ] 35. Rehany, N.; Barsi, A.; Lovas, T. Capturing Fine Details Involving Low-Cost Sensors –a Comparative Study. Int. Arch. Photogramm. Remote Sens. Spat. Inf. Sci. 2017, XLII-2/W8, 213. [CrossRef] 35. Rehany, N.; Barsi, A.; Lovas, T. Capturing Fine Details Involving Low-Cost Sensors –a Comparative Remote Sens. Spat. Inf. Sci. 2017, XLII-2/W8, 213. [CrossRef] p f 36. Khallouﬁ, H.; Azough, A.; Ennahnahi, N.; Kaghat, F.Z. Low-Cost Terrestrial Photogrammetry for 3D Modeling of Historic Sites: A Case Study of the Marinids’ Royal Necropolis City of Fez, Morocco. Mediterr. Archaeol. Archaeom. 2020, 2, 257–272. [CrossRef] A Case Study of the Marinids’ Royal Necropolis City of Fez, Morocco. Mediterr. Archaeol. Archaeom. 2020, 2, 257–272. [CrossRef] 37. Abate, D.; Toschi, I.; Sturdy-Colls, C.; Remondino, F. A Low-Cost Panoramic Camera for the 3D Documentation of Contaminated Crime Scenes. Int. Arch. Photogramm. Remote Sens. Spat. Inf. Sci. 2017, XLII-2/W8, 1. ing, D.K. Producción y Dirección de Cortometrajes y Vídeos; Instituto Oﬁcial de Radio y Televisión: Madrid, Spa 009; p 399 50. Rabiger, M. Dirección Cinematográﬁca: Técnica y Estética; Ediciones Omega: Barcelona, Spain, 2009. 49. Millerson, G.; Owens, J. Television Production, 14th ed.; Focal Press/Elsevier: Amsterdam, The Netherlands; Boston, MA, USA, 2009; p. 399. ; p 50. Rabiger, M. Dirección Cinematográﬁca: Técnica y Estética; Ediciones Omega: Barcelona, Spain, 2009. 51. Kraaman. The Best Encoding Settings for Your 4k 360 3D VR Videos. Available online: https://headjack.io/blog/best-encoding- settings-resolution-for-4k-360-3d-vr-videos/ (accessed on 20 November 2021). 51. Kraaman. The Best Encoding Settings for Your 4k 360 3D VR Videos. Available on settings-resolution-for-4k-360-3d-vr-videos/ (accessed on 20 November 2021). irección Cinematográﬁca: Técnica y Estética; Ediciones Omega: Barcelona, Spain, 2009. References [CrossRef] 38 Al d N Dí d U Cá S i B l B l S i 1990 37. Abate, D.; Toschi, I.; Sturdy-Colls, C.; Remondino, F. A Low-Cost Panoramic Camera for the 3 Crime Scenes. Int. Arch. Photogramm. Remote Sens. Spat. Inf. Sci. 2017, XLII-2/W8, 1. [CrossR 37. Abate, D.; Toschi, I.; Sturdy-Colls, C.; Remondino, F. A Low-Cost Panoramic Camera for the 3D Documentation of Contaminated Crime Scenes. Int. Arch. Photogramm. Remote Sens. Spat. Inf. Sci. 2017, XLII-2/W8, 1. [CrossRef] 38. Almendros, N. Días de Una Cámara; Seix Barral: Barcelona, Spain, 1990. 39. Mueller, P.; Vereenooghe, T.; Vergauwen, M.; Gool, L.V.; Waelkens, M. Photo-Realistic and Detailed 3D Modeling: The Antonine Nymphaeum at Sagalassos (Turkey). In Proceedings of the 32nd Conference on Computer Applications and Quantitative Methods to Archeology. CAA 2004, Prato, Italy, 13–17 April 2004; pp. 1–10. e varieties of realism in computer graphics. In Proceedings of the Electronic Imaging VIII, 2003, Santa Clara 003. [CrossRef] 40. Ferwerda, J.A. Three varieties of realism in computer graphics. In Proceedings of the Electronic Im CA, USA, 17 June 2003. [CrossRef] 41. Tang, R.; Fritsch, D. Correlation Analysis of Camera Self-Calibration in Close Range Photogrammetry. Photogramm. Rec. 2013, 28, 86. [CrossRef] 42. Fraser, C.S. Network design considerations for non-topographic photogrammetry. Photogramm. Eng. Remote Sens. 1984, 50, 1115–1126. 43. James, M.R.; Chandler, J.H.; Eltner, A.; Fraser, C.; Miller, P.E.; Mills, J.P.; Noble, T.; Robson, S.; Lane, S.N. Guidelines on the use of structure-from-motion photogrammetry in geomorphic research. Earth Surf. Processes Landf. 2019, 44, 2081. [CrossRef] 44. Benítez, A.J.; Jiménez, I.; Sánchez Cid, M. Conceptos de Postproducción. In Una Perspectiva Contemporánea; Utray Armenteros, M.F., Benítez, A.J., Eds.; Postproducción digital: Madrid, Spain, 2015; p. 13. 45. Gaiani, M.; Apollonio, F.; Ballabeni, A.; Remondino, F. Securing Color Fidelity in 3D Architectural Heritage Scenarios. Sensors 2017, 17, 2437. [CrossRef] [PubMed] 46. Martínez Rodríguez, S. Evaluación en Campo de Una Nueva Técnica Fotogramétrica de Rango Cercano y Bajo Coste Basada en la Correlación Automática de Imágenes; U de Santiago de Compostela Campus de Lugo: Lugo, Spain, 2014. Available online: https://minerva.usc.es/xmlui/handle/10347/12884 (accessed on 20 November 2021). 47. Díaz-Vilariño, L.; Frías, E.; Balado, J.; González-Jorge, H. Scan Planning and Route Optimization f as-Designed Bim. Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci. 2018, XLII-4, 143. [CrossRef] 48. Rea, P.W.; Irving, D.K. Producción y Dirección de Cortometrajes y Vídeos; Instituto Oﬁcial de Radio y Telev Heritage 2022, 5 41
https://openalex.org/W4352984595	https://libros.univalle.edu.co/index.php/programaeditorial/catalog/download/463/540/3969, https://bibliotecadigital.univalle.edu.co/bitstream/10893/19971/1/Desafios_interes_publico.pdf	es		Desafíos del interés público Identidades y diferencias entre lo público y lo privado	null	2,005	cc-by-sa	134,607	Dirige el Grupo de Investigación “Gestión y Políticas Públicas”, Categoría A Colciencias, y es investigador activo en temas de políticas públicas y teoría organizacional. Desde hace varios años es par evaluador de Colciencias en estas áreas. Ha publicado numerosos libros y artículos en revistas especializadas. Sus más recientes libros son: La soberanía transformada, Bogotá: ECOE editores Universidad del Valle (Febrero 2007); y Gestión y gobernabilidad en Emcali: liberalización de los servicios públicos y vulnerabilidad de las burocracias públicas, Bogotá: ECOE Editores - U n i v e r s i d a d d e l Va l l e Universidad Libre (Seccional Cali), (2007). E ste libro aborda desde perspectivas multidisciplinarias (filosóficas, sociológicas e históricas) la relación tensa y compleja entre lo Público y lo Privado, a través de un hilo conductor: la reforma del Estado y de la función pública, en curso en el mundo entero desde hace varias décadas. En particular el autor pretende suministrar antes que un análisis de casos, un cuadro teórico general, revisando la literatura francófona y anglófona que se ha ocupado del tema. Ofrece un valor agregado al contrastar dichos cuadros teóricos con las especifidades de la evolución del aparato del Estado en América Latina. EDGAR VARELA BARRIOS Estudió FilosofíaHumanidades y adelantó la Maestría en Historia Andina en la Universidad del Valle. Es Doctor en Administración, opción Management de HECUniversidad de Montreal. También se desempeña como profesor titular de la Universidad del Valle a la cual se vinculó desde 1992 como profesor de tiempo completo en la Facultad de Ciencias de la Administración. En la actualidad, ocupa el cargo de Vicerrector Administrativo de la Universidad del Valle, en donde además ha sido Director de la Oficina de Planeación, Representante Profesoral ante el Consejo Académico, Director de los Postgrados en Ciencias de la O r g a n i z a c i ó n y Administración Pública, y Vicedecano de la Facultad de Ciencias de la Administración. ISBN: 978-958-670-397-0 9 789586 703970 http://programaeditorial.univalle.edu.co Telefax: (+57) 2 330 8877 E-mail: editorial@univalle.edu.co Teléfonos: (+57) 2 321 2227 A.A. 025360 Cali, Colombia Universidad del Valle Ciudad Universitaria, Meléndez Vicerrectoría de Investigaciones Universidad del Valle Universidad del Valle Universidad del Valle CONTINÚA EN LA SOLAPA SIGUIENTE Colección Ciencias de la Administración 1 Este libro aborda desde perspectivas multidisciplinarias (filosóficas, sociológicas e históricas) la relación tensa entre lo Público y lo Privado, a través de un hilo conductor: la reforma del Estado y de la función pública, en curso en el mundo entero desde hace varias décadas. En particular el autor pretende suministrar antes que un análisis de casos, un cuadro teórico general, revisando la literatura francófona y anglófona que se ha ocupado del tema. Ofrece un valor agregado al contrastar dichos cuadros teóricos con las especifidades de la evolución del aparato del Estado en América Latina. 2 EDGAR VARELA BARRIOS Colección Ciencias de la Administración 3 Varela Barrios, Edgar, 1955Desafíos del interés público / Edgar Varela Barrios. — Santiago de Cali : Programa Editorial Universidad del Valle, 2008. 364 p. ; 22 cm. Incluye bibliografía e índice. 1. Administración pública 2. Funcionarios públicos 3. Función pública 4. Administración de personal I. Tít. II. Serie. 351 cd 21 ed. A1149923 CEP-Banco de la República-Biblioteca Luis Ángel Arango Universidad del Valle Programa Editorial Título: Desafíos del interés público Identidades y diferencias entre lo público y lo privado Autor: Edgar Varela Barrios ISBN: 978-958-670-397-0 ISBN-PDF: 978-958-5156-75-3 DOI: 10.25100/peu.463 Colección: Ciencias de la Administración Primera Edición Impresa mayo 2005 Rector de la Universidad del Valle: Édgar Varela Barrios Vicerrector de Investigaciones: Héctor Cadavid Ramírez Director del Programa Editorial: Omar J. Díaz Saldaña © Universidad del Valle © Edgar Varela Barrios Diseño de carátula: UV. Media Este libro, o parte de él, no puede ser reproducido por ningún medio sin autorización escrita de la Universidad del Valle. El contenido de esta obra corresponde al derecho de expresión del autor y no compromete el pensamiento institucional de la Universidad del Valle, ni genera responsabilidad frente a terceros. El autor es el responsable del respeto a los derechos de autor y del material contenido en la publicación, razón por la cual la Universidad no puede asumir ninguna responsabilidad en caso de omisiones o errores. Cali, Colombia, octubre de 2020 4 CONTENIDO PREFACIO DE LA SEGUNDA EDICIÓN................................ l RETOS ACTUALES PARA LA DEFENSA DEL INTERÉS PÚBLICO ................................................................................ 1. Identidades entre lo público y lo privado........................ 1.1. Lo público no estatal ............................................... 1.2. ¿existen diferencias entre la gestión de lo público y lo privado?............................................... 2. Los desafíos provenientes de las sociedades civiles....... 2.1. Los difíciles acuerdos teóricos ............................... 2.2. Los conflictos de intereses...................................... 2.3. El auge del individualismo y el problema de las desigualdades ................................................ 3. La preeminencia del estado sobre el interés público ....... 4. El interés público y sus dimenciones ecuménicas .......... 5. La instumentalización del del interés público .................. 11 CAPÍTULO CAPÍTULO 2 GENEALOGÍA MODERNA DE LA FUNCIÓN PÚBLICA ...... 1. Un discurso y una praxis individualista, secular y nacionalista .................................................................. 2. Las premisas rousseaunianas del interes público............ 3. El planteamiento kantiano sobre lo público y lo privado. 4. Hegel: los fundamentos de la sociedad civil y la función de la burocracia estatal ....... ............................... 4.1. El trabajo y el sistema de necesidades..................... 4.2. Hegel y la libertad política....................................... 5 19 20 22 25 30 30 32 35 37 39 43 51 53 61 68 73 74 85 ___ 5. Marx y la filofofía del derecho hegeliano........................ 5.1. La postura política inicial de marx .......................... 5.2. La "inversión" hegeliana.......................................... 5.3. Los límites, según marx, de la dicotomía hegeliana (sociedad civil-estado)............................. 5 .4. Las contradicciones del funcionarismo hegeliano... 6. La concepción marxista del estado y de la burocracia pública .................................................... ...... 6. 1. La crítica marxista al arbitraje estatal del interés general ... .... ........... .. ..... ... ...... ... ........ ......... .... ........ .. . 6.2. La utopía marxista del estado revolucionario de transición y de la sociedad sin clases................. 6.3. Evaluación retrospectiva de la concepción marxista sobre la burocracia................................... 7. A manera de conclución: vigencia del rol del funcionarismo público.................................................... 7.1. ¿cuál es entonces la pertinencia de la visión hegeliana sobre la burocracia? ................................ 7 .2. La continuidad del "fenómeno burocrático" ............ CAPÍTULO 3 92 92 94 97 100 110 111 117 122 127 128 13 1 CONCEPCIONES DE LO PÚBLICO...................................... 147 1. Campos temáticos de lo público..................................... 2. Génesis y desarrollo histórico de la contraposición público-privado............................................................... 2.1 orígenes de la hegemonía de lo público-estatal ........ 2.2. "occidente": fuente emanatoria de la dicotomía público-privado ....................................................... 2.3. Dominio de la privaticidad extendida en el mundo medieval ...................................................... 2.4. Publicidad y vida moderna...................................... 4. Lo público y los desafíos de la modernidad ................... 5. La configuración del espacio público ............................ 6. El sentido político de lo público...................................... 147 6 154 155 158 160 164 170 178 182 CAPÍTULO 4 LA CONTRAPOSICIÓN PÚBLICO-PRIVADO EN SU PERSPECTIVA LATINOAMERICANA.................................... 1. Privatización, patrimonialismo y regalismo ibérico......... 2. Vida familiar y sobredeterminación de lo público ........... 3. El centralismo borbónico en el nuevo mundo................. 4. El rol del militarismo en la vida colonial tardía................ ·5. Lo público y lo privado bajo el proyecto liberal republicano........................................................... 6. Centralización política y busqueda de la unidad nacional . ... ...... ..... .. . ... .... ... .... .... ... .. .... . .. .. ... .. ... . . .. 7. Orden social y libertades republicanas: Un difícil balance ............................................................ 8. La preminencia del intervencionismo estatal sobre lo público........................................................................ 9. La crisis del interés público en el contexto colombiano . 195 196 200 204 207 21 O 219 221 229 236 CAPÍTULO 5 MODERNIZACIÓN INSTITUCIONAL Y NUEVOS ROLES PÚBLICOS Y PRIVADOS........................................... 1. Los términos del debate .................................................. 2. Modernidad, racionalismo y mercantilización de la vida social .... ....... ... .... ......... . .... ..... .. .. ....... . .... ... .. ........ 3. La crisis de la modernidad.............................................. 4. Redefinición del ámbito propio al interés público ........... 5. El predicamento intervencionista.................................... 6. La modernización del estado.......................................... 247 247 25 1 261 265 279 282 CAPÍTULO 6 REFORMA DEL ESTADO Y REDEFINICIÓN DEL CAMPO DE LAS POLÍTICAS PÚBLICAS La experiencia reciente bajo el intl ujo del. global ismo neo! iberal 1. Management y gestión pública ....................................... 2. Los avatares y retrocesos en el cambio de modelos de gestión pública ........................................................... 3. Principales diferencias entre lo público y lo privado....... 4. Dinámicas recientes de la relación público y privado.... 7 293 293 297 301 306 5. Surgimiento de nuevas fuerzas sociales y perspectivas de afirmación del interés público............... 315 CAPÍTULO 7 ¿RENOVAR O DESTRUIR LAS BUROCRACIAS ESTATALES? EL IMPACTO DEL "NEW PUBLIC MANAGEMENT" ............ 1. El ideario del new public management............................ 1.1. La programática del new public management ......... 1.2. Las estrategias de debilitamiento de las burocracias públicas ................................................................... 1.3. Estado, sociedad, mercado ................... .................. 2. Globalismo y desmonte de Las burocracias públicas..... 2. 1. Una radical tranformación de los principios del gasto público ................................... .................. 2.2. La liberalización económica y la competitividad de las fuerzas del trabajo......................................... 3. Tendencias en la transformación de las burocracias públicas: Trabajo alineado, precarización, pauperización 3.1. La alienación laboral en el sector público............... 3.2. La precarización laboral del funcionariado .............. 3.3. La pauperización de las burocracias ....................... 4. Para las teorías manageriales dominantes, ¿es viable la burocracia estatal? ...................................... 5. Entonces, ¿cuál es la funcionalidad actual de las burocracias públicas? ..................................................... 5.1. Neutralidad e imparcialidad ..................................... 5.2. Dinámicas internas en los procesos de burocratización ....................................... ................. 5.3. El futuro de las burocracias públicas...................... BIBLIOGRAFÍA........................................................................ 8 325 327 328 331 333 334 336 337 338 338 340 341 342 346 347 349 351 357 Para Bianney: Que este texto nos ayude en la búsqueda de la felicidad pública, y en los fines de la vida privada. 9 PÁGINA EN BLANCO EN LA EDICIÓN IMPRESA 10 PREFACIO DE LA SEGUNDA EDICIÓN Este libro fue publicado originalmente en Cali, Colombia, en 1998 por el Programa Editorial de la Universidad del Valle, universidad a la cual me encuentro vinculado como profesor de tiempo completo. Al agotarse la primera edición, unos dos años atrás, tuve presente la conveniencia de actualizar, amplificar e incorporar nuevas temáticas, referencias bibliográficas y experiencias sociales. Por supuesto, debo claramente señalar que actualizo también ciertos puntos de vista formulados en 1998 que fueron modificados por la experiencia global de la región en la última década. Estas cuestiones —viejas y nuevas— resultan altamente ilustrativas de los desafíos que enfrentan —primordialmente en el plano de la vida política— la defensa y concreción, más allá de la retórica, del llamado interés público. En virtud de lo anterior, he incluido un par de capítulos adicionales, precisamente el primero y el último. El texto inicial incorporado me permite actualizar el marco teórico de referencia de la relación público-privado para contestar algunas cuestiones que —leído el libro con la distancia que da el tiempo— me parece que estaban esbozadas sólo de manera implícita. En efecto, a pesar del título, no hice explícitos los tipos de desafíos que afronta el interés público, especialmente si tomamos en consideración el ámbito de la reforma de los Estados-naciones y la globalización que debilita sus soberanías. Así mismo, he considerado conveniente precisar los límites y las determinaciones presentes en la dicotomía y en las identidades que, paradójicamente, configuran de manera ontológica las relaciones entre los polos de lo público y lo privado. 11 De otro lado, el capítulo de cierre constituye un balance del impacto de las nuevas teorías manageriales sobre la burocracia estatal que, bajo el influjo del neoliberalismo, han aplicado drásticas políticas públicas de reformulación de la relación de lo público y lo privado. Así, se ha cambiado el statu quo que fue paradigmático en muchas regiones del mundo —principalmente las más industrializadas— de una serie de políticas de Estado en las hoy añoradas épocas del Estado de Bienestar. Las rectificaciones del modelo alternativo se han reclamado desde muchas orillas, incluso desde algunos de sus propios gestores. Lo anterior porque el resultado de las políticas, en buena medida, no obedeció al racionalismo puro de la tecnocracia académica sino al rudo juego de interés de la cara rapaz de la globalización neoliberal (J. Stiglitz, 2003; 2002a). De todas formas el propósito general sigue siendo el mismo: formular, dentro de un plano conceptual interdisciplinario — filosófico, sociohistórico y managerial—, el análisis de la dicotomía público-privado, enfocándola primordialmente en su conexidad con lo Estatal. En pocas palabras, esta investigación de naturaleza teórica la he asumido basado en el convencimiento de que sus logros podrán constituir un espacio que permita explicar o fundamentar, si se quiere verlo así, el decurso posterior de las respectivas teorías políticas de estirpe hegeliana y marxista frente al problema del Estado y la función de la burocracia pública. Soy un convencido de que la clave de muchos de los debates contemporáneos se encuentra en su prístina formulación en términos filosóficos. En especial, este seguimiento me resulta interesante a la luz de los intentos por reinscribir a Hegel en el seno de la tradición liberal-democrática, y por conjugar al joven Marx con el humanismo valorativo del individuo (J. Pinson, 1989). Las teorías de la burocracia y la función pública fueron en general el resultado de un primer balance hecho por las filosofías políticas que en Alemania, Francia e Inglaterra, interpretaron las transformaciones sociales, las revoluciones políticas y los procesos de modernización económica e institucional, como espero poder mostrarlo a lo largo de las siguientes páginas. La investigación aquí abordada pretende, en consecuencia, fundamentar teóricamente el conjunto de categorías con las cuales hoy se interpreta el 12 proceso de reforma gestionaria del aparato público-estatal. Con ello pretendo que los textos incluidos en el presente volumen contribuyan al debate sobre la naturaleza contemporánea —y ello visto desde una perspectiva diacrónica— del Estado, la función pública y la redimensión de los espacios de lo público y lo privado. En otros dos capítulos (tercero y quinto), me propongo estudiar las diversas acepciones de lo público, analizar su configuración en la modernidad y sus actuales tendencias constitutivas. De este modo, discutiré el rol del mercado y la competencia como ámbitos de despliegue de lo privado en el seno de la moderna sociedad civil, evaluando la función mediadora de la publicidad; y reflexionaré, además, sobre la naturaleza de las relaciones sociopolíticas que surgen en dicho contexto mercantil. Tal balance no se puede esclarecer apelando a un simplismo descalificador sólo con base en expresiones tales como las del neoliberalismo o intervencionismo para desdeñar las políticas sin tomarse siquiera el trabajo de estudiarlas y analizar su impacto y sus lógicas constitutivas, puesto que ella no implica neutralidad aséptica alguna. Al contrario, el lector podrá encontrar en estas páginas una posición ideológica, ética y política sobre las transformaciones de las relación público-privado sufridas en las décadas recientes y sobre las tendencias previsibles. En suma, los problemas de política pública son, en realidad, altamente complejos; presentan numerosos matices que corresponden a las especificidades de la cultura política y de la configuración socioeconómica sobre la que ésta se levanta. He redactado, además, un par de capítulos que aterrizan en las ya señaladas consideraciones teóricas dentro del contexto latinoamericano y colombiano. En efecto, el capítulo cuarto (La contraposición público-privado en su perspectiva latinoamericana) tiene como objetivo describir en grandes trazos la forma como ha evolucionado históricamente la dicotomía público-privado en nuestro continente, en tanto que el capítulo sexto (Reforma del Estado y redefinición del campo de las políticas públicas - la experiencia reciente bajo el influjo del globalismo neoliberal)1 busca esclarecer el impacto de las transformaciones de lo público y lo privado en relación con los cambios de nuestras instituciones público-estatales en los decenios más recientes. 13 Conviene subrayar, en esta dirección, que el proceso de modernización del aparato público fue equivalente a lo largo de los últimos dos siglos, a la consolidación del Estado-Nación, e implicó que muchas de las dimensiones de la vida social que en la sociedad tradicional eran propias del ámbito privado resultaran nacionalizadas, estatizadas. En esta perspectiva, el Estado aparecía ante todo como garante de los llamados derechos sociales. La asimilación de lo privado por lo público y la asignación al Estado de los servicios sociales tuvieron una honda relación con el discurso de los Derechos Humanos. Estaba supuesta en el Gobierno la garantía efectiva de los principios de equidad y se veía que la simple competencia privada, el rol distribuidor del mercado, generaba enormes desigualdades. La acumulación de ganancias sin la restricción del Estado implicaba, en esta perspectiva, la aparición de modalidades perversas de capitalismo salvaje. Como una rectificación a dicha situación de desigualdad se le asignó al Estado un papel redistributivo y de justicia social; de atención directa sobre los procesos sociales y económicos, dado que se entendió la igualdad no sólo en términos políticos o formales, sino en términos de la igualdad real: la llamada calidad de vida (D. J. Moon, 1995). Lo anterior implicó, a nivel mundial, un crecimiento desmedido del tamaño del Estado y de los impuestos directos e indirectos que éste percibía de la ciudadanía. Así, el Estado y las elites a éste articuladas se apropiaron de porciones significativas de la riqueza de las naciones. En particular, las burocracias públicas se convirtieron en una medida significativa, tal como lo advirtiera lúcidamente Weber (1976) en un poderoso grupo de interés colocado por encima de las sociedades civiles y legitimado por su parcial función de agente del interés público. Cada vez más le fueron adjudicadas tareas y competencias al gobierno, hasta llegar a límites insospechados: prácticamente ninguna actividad social y económica se descartó como campo de las actividades gubernamentales. Esta situación condujo, desde mediados de los años 70 del siglo pasado, a una crisis recurrente fiscal, social y de legitimidad al modelo tradicional de intervencionismo estatal (A. Cortina, 1998). No sólo la crítica a esta burocratización absorbente provino de la derecha liberalizante que siempre había sido reacia a las políticas redistributivas y al papel del Estado empresarial (B. Levine 14 y otros, 1990). Los excesos de la burocratización fueron incluso lúcidamente analizados en el plano teórico por las propias corrientes del neomarxismo (B. Jessop, 1999; N. Poulantzas, 1973; S. Amin, 1991), la llamada teoría crítica de la sociedad (J. Habermas y otros) y por autores tan diversos como H. Arendt, Marcuse, K. O. Apel y M. Foucault. Sin embargo, dicho avasallamiento de la esfera público-estatal sobre los espacios de la denominada sociedad civil se encontró, desde los 80 y especialmente a lo largo de los años noventa, en un proceso de reversión. Este fue el fondo de las transformaciones significativas que durante las últimas décadas se han producido en relación con el aparato estatal y la correlación entre mercado y gobierno (véanse al respecto, en especial los capítulos sexto y séptimo). El péndulo fue de un extremo al otro. En efecto, se corrigieron muchos de los excesos del intervencionismo de Estado. Éste, en el marco de las sociedades capitalistas, permitió la extensión del bienestar y el asistencialismo a los sectores golpeados por el modelo fordista de acumulación de capital. Empero, el corporativismo fue beneficiario directo principalmente en sus variantes empresariales; la clase política en muchas partes del mundo se apropió, y lo sigue haciendo, de los bienes públicos, y se presentó una agudización de la crisis fiscal. Las debilidades y el agotamiento de dicho modelo de gestión de la vida pública han permitido su reemplazo por otro que ha revivido, en lo fundamental, el capitalismo salvaje y reinstaurado la precarización social y laboral del siglo XIX, que se creía cosa del pasado. La desregulación, el desmantelamiento del Estado y de sus burocracias, resultó peor remedio que las enfermedades tratadas. En América Latina este es un punto que se debate intensamente. Sin mistificar al Estado o creer que constituye una suerte de topos neutro, se valora hoy, desde una perspectiva crítica, su fortalecimiento; se reclama su agenciamiento del interés público frente a las ofensivas neoliberales encaminadas a removerlo como árbitro, desmontar las burocracias públicas, prohijar al máximo las desregulaciones, la mercantilización sin límites de los bienes públicos, asignarle al Estado (veuilleur de nuit) sólo las funciones de seguridad, etc. En este sentido, desde la perspectiva de las políticas públicas es válido, en mi percepción, propiciar que el Estado efectivamente asuma funciones de subsidiaridad, que provea bienes y servicios públicos, más allá de las limitaciones del mercado. 15 Dadas las asimetrías inherentes a los esquemas de mercado, la dimensión política que el Estado encarna es o debe ser un espacio legítimo para proveer la felicidad social. Al menos por eso debemos propugnar. Esa es una parte fundamental de la contradicción con la corriente economicista que pretende separar tajantemente su campo del de la política. Allí es donde la reclamación democrática de economistas críticos de las corrientes dominantes, como Jacques Genereux (2001), encuentra su sentido pleno. El trabajo del profesor Aktouf (2002) igualmente muestra cómo existen además dimensiones sociales, culturales e históricas que son determinantes para valorar las tipologías de relaciones entre lo privado y lo público. Véase su argumento sobre los capitalismos nipon-renano (cfr. también, en ese sentido consultar a M. Albert, 1991) que evidencia la manera cómo las instituciones económicas se imbrican en estos países con la sociedad de manera contraria al mecanismo neoliberal que pretende la universalidad de los métodos y los sistemas, y que justifica además el reduccionismo de las ciencias sociales al imperialismo de la economía-management. *** Este libro, a diferencia de un texto que acabo de publicar (E. Varela, 2005a), no es una compilación de artículos escritos de forma independiente, sino que ha obedecido a un plan de escritura que sigue presente en esta segunda edición2. A pesar de la enorme literatura que en los últimos años se ha ocupado de la relación público-privado, considero que la perspectiva de mi trabajo posee aristas originales en la medida que mi propósito se refiere a la pertinencia de indagar sobre estas categorías en relación con el management y las políticas públicas. Desde luego, y dado que ésta es una edición actualizada, en los capítulos originales he introducido significativos cambios, rectificaciones, correcciones y ajustes, e incorporado nuevas temáticas y referencias bibliográficas. En algunos casos, en realidad, he reescrito gran parte de dichos capítulos porque la perspectiva en sólo unos cuantos años ha cambiado dramáticamente. Esta tarea ha sido, en general, complicada porque ha de tenerse un criterio de selectividad, en mi opinión, determinado por el hilo argumentativo seleccionado y por los propósitos pedagógicos de la obra, puesto que la literatura y las experiencias resultan, en éste como en muchos otros campos, inabarcables para una sola persona. 16 Precisamente, en virtud de tales limitaciones, el esfuerzo de colectivizar la tarea de sistematización e investigación ha implicado la creación del grupo de investigación “Gestión y Políticas Públicas” en la Universidad del Valle, del cual soy actualmente director; y también se ha concretado en el funcionamiento de la Maestría en Políticas Públicas que ya va por la tercera cohorte de estudiantes. Actualmente en Colombia pugnamos por solidificar el trabajo de investigación. Felizmente, el primero de los esfuerzos encaminados a la creación de Doctorados en Ciencias de la Administración se ha concretado con la autorización gubernamental de ofrecer un programa de este tipo a la Universidad Eafit (Medellín). Como profesor visitante de su Maestría en Administración (MSc), no puedo menos que alegrarme, felicitarlos y colaborar en la creación de una red académica nacional, abierta al mundo y al aprendizaje de otras experiencias. Espero que mi propia Universidad, que hace bastantes años busca el mismo objetivo, lo logre prontamente y así comencemos a dar un salto cualitativo en la investigación y la docencia en los distintos campos del saber administrativo. De otro lado, recientemente me he beneficiado del ambiente académico y de las posibilidades de acceso, tanto a los debates como a la literatura contemporánea sobre estos temas, que me facilitó mi estancia en Montreal, donde adelanté un Doctorado en Administración en la HEC de la Universidad de Montreal. Debo reconocer la gentileza y generosidad de los profesores Omar Aktouf y Alain Chanlat, como siempre, con una gran sensibilidad humanista acerca de los dramas de nuestras sociedades que intentan las vías al desarrollo y a la equidad social. Escuchándolos, conversando con ellos y leyendo sus trabajos y los de otros colegas suyos en la HEC, pude percibir nuevas maneras de ver la relación público-privado. Expresadas estas advertencias y salvedades, quiero formular mi agradecimiento a los diversos aportes que colegas, especialmente de mi Grupo de Investigación, y estudiantes de pre y postgrados, han hecho a las versiones preliminares de los diversos capítulos. En efecto, los seminarios sobre Administración de lo Público y lo Privado y sobre Doctrinas Políticas que tengo a mi cargo en el pregrado en administración, así como los seminarios de la 17 Maestría en Políticas Públicas de la Universidad del Valle, Políticas Públicas, Teoría Política y un curso de tipo introductorio al análisis de políticas, han sido escenarios en donde he podido exponer ante estudiantes y colegas mis reflexiones sobre la materia aquí tratada. Especialmente reconozco a la Facultad de Ciencias de la Administración por su ambiente de trabajo académico. Y al Programa Editorial de la Universidad del Valle por acoger nuevamente en sus colecciones esta obra. Debo además agradecer a mis asistentes en la Facultad de Administración dentro del grupo de investigación “Gestión y Políticas Públicas”, Wilson Delgado, Paola Bernal y Yolima Gasca, por su trabajo en la referenciación bibliográfica, revisión literaria y transcripción de una parte de los textos. Finalmente, dedico este libro a mi familia en Cali, en especial a mi esposa, Bianney Arias, por su soporte y apoyo entusiasta. Del mismo modo, hago extensiva esta dedicatoria a mis padres Marco Tulio y Oliva, quienes son ejemplo de la gente honesta y coherente con sus principios; a mi hermano Ray, siempre diligente; y a mi hija, la inquieta y vivaz Juliana. Gracias a todos ellos por soportar las ausencias del Doctorado. Espero que algún día les pueda recompensar. Édgar Varela Barrios 18 Capítulo 1 RETOS ACTUALES PARA LA DEFENSA DEL INTERÉS PÚBLICO En este capítulo inicial me propongo referir ciertos debates claves en la discusión sobre la relación entre lo público y lo privado que, como lo acabo de señalar en el prefacio, estaban sólo postulados de manera implícita en la primera edición. Ello me permite hacer las disgresiones pertinentes sin alterar la estructura narrativa del texto original. Primeramente, me intereso por incorporar ciertas precisiones sobre la célebre dicotomía público-privado (N. Bobbio, 1964; J Weintraub, 1997; J. Calhoun, 1997, 1992), en especial, mostrando sus límites, contextos y “líneas de fuga”; primordialmente hacia el llamado tercer sector. En este punto precisaré además, una serie de lugares comunes sobre la identidad entre los dos términos que son recurrentes en la literatura administrativa y tienen consecuencias prácticas, por ejemplo en la justificación teórica del New Public Management NPM (Cf.: Cap. 7). En realidad existe una serie de grandes desafíos, retos —que dificultan u obstaculizan la concreción efectiva de lo público—, que en la actualidad constituyen elementos recurrentes de la discusión, tanto en el amplio sentido de la opinión ciudadana como en el terreno de los especialistas. Por razones expositivas analizaré estos desafíos en relación con su fuente de origen, a saber: a) aquellos que provienen de las sociedades, incluyendo el mundo de los negocios; b) los que provienen del propio Estado y de las institu19 ciones multilaterales; y c) los que resultan de la creciente mundialización1 con su corolario, esto es, la erosión de las soberanías territoriales (E. Varela, 2005b; Z. Leidy, 2004; D. Held y otros, 1999). Además, en la parte final, habremos de considerar en sus grandes trazos los problemas de tipo instrumental: cómo concretar y hacer efectivas las políticas, asunto que se conecta con los debates actuales sobre la gobernanza y la gobernabilidad, tanto en el plano de los Estados nacionales como en la esfera de lo global o ecuménico. 1. Identidades entre lo público y lo privado Generalmente el presupuesto conceptual de la relación público privado anuda sobre la inevitable dicotomía entre los dos términos. Como tal expresión lo indica, estos serían contradictorios, antinómicos, como la luz y la oscuridad, lo bueno y lo malo, el ser y el no ser, etc. Aunque una definición in extenso sobre esta relación se encuentra detallada más adelante (en especial en un plano abstracto —sociológico— en el capítulo tercero, y de manera más operacional, en los capítulos quinto y sexto), recordemos que estos términos no son unívocos en sí mismos. Al contrario, tomados del llamado lenguaje ordinario poseen ambos una elevada polisemia. Lo público tiene su origen semántico en el pueblo, en aquello que le es propio a la comunidad, a “todos”. El asunto es que siempre el todo remite a un continente delimitado de cosas, de personas, en términos de colectividades micro, meso y macro, tales como una nación o una región geográfica o que se identifican frente a un interés (v. gr., la equidad de los géneros, el acceso a ciertos derechos o la obligación frente a ciertos deberes). Esta es una de las mayores dificultades analíticas y prácticas que enfrentamos2. Una empresa pública del Estado puede ser de propiedad de una municipalidad, de una región o del Estado Nación. Pero siempre la referimos con los mismos descriptores (J. Dewey, 1958). Las cosas se complican mucho más ahora, ya que, con la globalización, numerosas agencias y organizaciones públicas se han transestatalizado de diversas formas. Peor aún, las dificultades taxonómicas se dificultan si trabajamos sobre los ámbitos difusos de lo “público no estatal”. Sean aquí los casos de las organizaciones solidarias mundializadas en el campo de los Derechos Humanos, la 20 ecología, la batalla contra las hambrunas, o instancias multilaterales de fuente no estatal. Como lo mostraré un poco más adelante en este mismo capítulo, el espacio más novedoso pero igualmente avasallador es hoy el de lo público ecuménico, con lo cual alcanzamos por vez primera de forma nítida el nivel más elevado de universalidad en la operacionalidad del concepto de lo público. De tal forma que no nos encontramos aquí frente a una dicotomía pura, como lo concibe en su célebre artículo para la enciclopedia Eunadi el filósofo italiano N. Bobbio (1964). Jurídicamente es cierto, tal como deviene punto de partida en Bobbio, que existe una formalización dicotómica. Ello se expresa, por ejemplo, en relación con los derechos de propiedad. Lo público está abierto a la pertenencia de todos con base en el principio jurisdiccional de la inclusión; en tanto que lo privado se basa en el referente antípoda. Aquello privado (he ahí la genealogía etimológica del término) significa una privación, vale decir una exclusión, una abstinencia. Con lo cual el término privado se define por una vía negativa, mientras que el otro —lo público— se construye en una perspectiva positiva, de afirmación. Pero, al mismo tiempo, el término privado posee el sentido positivo del disfrute continuado y asegurado de la posesión indefinida de aquello que les es negado a los demás. La otra dificultad al contextualizar la noción de privado es la idea arraigada hoy de que se trata de una categoría natural, anclada en la esencia humana. Y la tesis avecindada a la anterior, de que siempre que tratemos con sujetos individuales nos estamos refiriendo al individualismo. Sabemos, por la historiografía contemporánea, que en ambos casos se trata sobre todo de mistificaciones (Cfr. sobretodo F. Braudel, 1966). Aún en nuestros tiempos de globalismo neoliberal, a la mayor parte de los millones y millones de personas marginalizadas nos resulta ilusorio, irreal, imaginárnoslas como agentes egoístas competitivas. Esta no es en modo alguno la condición de quienes a duras penas sobreviven en los márgenes del globalismo mercantil. En esta dirección, las sociedades no deben verse sólo en términos simplistas como sumatorias de individuos privados vs. abstracciones de lo público (naciones, Estados, soberanías, etc.). Las sociedades —antiguas y contemporáneas— han poseído y poseen claramente numerosas instancias de mediación y agregación social. El empresario egoísta de la teoría microeconómica de 21 la firma, nada puede obtener de las oportunidades del mercado, si no posee una organización compleja integrada generalmente por numerosos individuos asalariados y por un corpus administrativo cuyas decisiones no son un simple eco de sus caprichos3. Además, los individuos se asocian en familias, comunidades, asociaciones, gremios, clubes; comparten valores de culto religioso, rituales de integración y exclusión, etc. En todas las sociedades coexisten los intereses solidarios; al menos, entre aquellos que se consideran iguales o pares entres sí. Dicha regularidad corresponde a un primer gran nivel de identidad que sobrepasa el mero individualismo de tipo egoísta y posesivo, caro a las simplificaciones neoliberales. Este es más bien el espacio de las solidaridades limitadas de corte particularista (que son el terreno ontológico predilecto del actual discurso filosófico comunitarista). 1.1. LO PÚBLICO NO ESTATAL En mi percepción, estos niveles de mediación e interacción meso-sociales, aunque no equivalen al individualismo rampante que he señalado arriba, no son tampoco los espacios de constitución del interés general. En tal dirección, las principales mediaciones constitutivas, generatrices del espacio de lo público, lo conforman las enormes potencialidades del llamado tercer sector. Es decir, aquella parte de la sociedad organizada que no busca el ánimo de lucro empresarial ni se articula identitariamente con la función estatal. En esta línea de actuación social, sin idealizar el asunto, el foco de identidad es una solidaridad amplia, universalista. Los Derechos Humanos, el medio ambiente, la protección de la niñez y otras reivindicaciones semejantes, no se dan como banderas de acción sólo para ciertas personas o para un grupo dado de menores de edad o para ciertos problemas ambientales. La legitimidad de dicho tercer sector depende de —se basa en— su consecuencia teórica y práctica con los señalados principios de universalidad. En el fondo buena parte de las ONG y entidades afines, en general, resulta de la concreción pragmática de la vieja utopía del siglo XIX. Permiten vehicular el espíritu de solidaridad desinteresada de vastos núcleos sociales de inspiración humanitaria (N. C. Graw, 1997). Resulta claro que no todas las ONG pertenecen y comparten el terreno, por sí 22 mismas, de concreción de esta idea de la solidaridad incluyente y universalista. Existe además en forma ampliamente extendida el espacio negativo, desde el punto de vista del interés público, de numerosas organizaciones del sector público no estatal que devienen instrumentales a la retirada, fuera de los Estados, de buena parte de sus funciones de tipo redistributivo y justicialista. Como ya lo dije, en muchos casos las ONG son cooptadas por las políticas de Estado o por la mercantilización de lo público (¡¡en las escuelas de management se ofrecen numerosos cursos y postgrados sobre la gestión de este sector, cuyas metodologías se copian de las utilizadas por las empresas!!). Hay allí ciertamente una paradoja: se milita en contra del globalismo neoliberal, pero se le hace el juego a la mercantilización y privatización de los bienes públicos, diseñando, ejecutando y evaluando políticas que, en un contexto realista, habrían de corresponderles a los Estados territoriales4. Trazar la línea divisoria entre la iniciativa ciudadana y social para gestionar la defensa del interés público de manera autónoma a la tutela estatal y a la señalada instrumentalización, es ciertamente una cuestión compleja de resolver a priori. Pongamos un ejemplo reciente a la mano: en Canadá, como en los EE.UU y Europa, la tragedia del 26 de diciembre de 2004 en el Océano Índico (el tsunami que causó decenas de miles de muertos, graves daños y destruyó vastas infraestructuras) ha generado una impresionante ola de solidaridad social, al punto que hay quienes hablan de la “globalización de la generosidad”. Los sentimientos humanitarios difundidos por las imágenes terribles que nos han transmitido los massmedia, impelen al vasto concurso solidario a través de las ONG profesionalizadas en estos asuntos (Médicos sin Fronteras, Care, Fosters Parents, etc) o de la acción ad hoc de las gentes, agrupadas coyunturalmente en las donaciones y el voluntariado. Obviamente, al lado de esta solidaridad social existe el compromiso de los Estados y de la llamada Comunidad Internacional. En general, en estos eventos, la acción de las sociedades civiles se presenta no como un sucedáneo de las responsabilidades solidarias de los regímenes políticos, sino en el ejercicio de la libre iniciativa con la solidaridad ecuménica de corte humanitario. Y así podríamos señalar muchísimos otros ejemplos. Cuando se trabaja en el terreno de los Derechos Humanos o del medio ambiente, se lo hace no para reemplazar a los Estados sino para ir más allá de éstos, 23 prisioneros de lógicas de gobernabilidad o de lo “políticamente correcto”. Incluso una porción significativa de dicho activismo se encamina a hacer eficaces las críticas a las incoherencias e inconsecuencias de los gobiernos en estos asuntos. Se inscribe en la lógica del control social y del Empowerment ciudadano (R. Santana y M. Negron, 1996). Repito, no pretendo idealizar al tercer sector. Entre las ONG existen todo tipo de proyectos, ideologías, funcionalidades, tanto loables como criticables. Pues ellas expresan, de cierta manera, las fuerzas que dimanan de las sociedades y sus formas de articularse o posicionarse respecto del poder político. Cuando hablo de un tercer sector alinderado con lo público no estatal, supongo dos condiciones ontológicas en su constitución: 1) la preeminencia del ámbito de lo público como campo y objeto de acción; y 2) la naturaleza voluntaria, libre, del asociacionismo que lo genera. Mientras la actividad estatal en lo público siempre tiene la fuerza de la coacción universalista de las leyes, en lo público no estatal aparece la fuerza alterna de la solidaridad fundante. Por lo anterior resulta más adecuado hablar de una tricotomía5. Como se supondrá, el primer agente de esta relación múltiple es el espectro de las organizaciones y realidades dimanantes del mercado, como institución central del capitalismo, exacerbado por las derivas actuales del globalismo economicista. En segundo lugar, ocupan su lugar los Estados y las instancias derivadas de los mismos, tanto en lo subnacional como en las relaciones internacionales. Y, finalmente, aparece en escena el tercer sector, cuyo espectro igualmente se ha mundializado. Pero estos tres pilares de la institucionalidad contemporánea no están en pie de igualdad. El dominio, en sus interrelaciones, le corresponde a la relación simbiótica entre los dos primeros. En tanto que el último ocupa un lugar relativamente marginal, o si se quiere ser benévolo, complementario. Esto explica que en numerosas ocasiones la emergencia del tercer sector está lejos de ser funcional a los Estados o a las lógicas del mercado. Es una forma de crítica práctica por la vía de la acción alternativa respecto de estos. Creo que la tricotomía expresa funcionalidades en el sentido que N. Luhmann (1985) le da al concepto. Es decir, en sus interrelaciones sistémicas, como subtotalidades en redes de comunicaciones intrínsecas y extrínsecas. 24 1.2. ¿EXISTEN DIFERENCIAS ENTRE LA GESTIÓN DE LO PÚBLICO Y LO PRIVADO? Otra variante analítica que vale la pena considerar es la recurrente tesis en las Ciencias del Management de que en realidad no hay diferencias de fondo en la gestión de lo público y lo privado. Jacques Ellul (1990), en su pionero trabajo sobre la técnica, alude a este punto. En tal perspectiva, para dicho autor, las técnicas “duras” de todo tipo expresadas en máquinas, redes comunicacionales, infraestructuras, etc.; o las “blandas”, tales como las tecnologías administrativas y organizacionales que permean la vida política y económica, surgen de forma relativamente autónoma y moldean la acción social. De tal modo que no habría distinción de fondo entre los ámbitos público y privado. Las técnicas serían intercambiables y determinantes en su ciega neutralidad (Ibíd.). Más recientemente, Barry Bozeman (1998) argumenta en la misma dirección. Según este teórico americano del Management, no habría diferencias sustanciales entre las organizaciones del gobierno y las empresas económicas que pretenden obtener lucro. Para ello, este último restringe implícitamente la noción de lo privado al círculo de la privacidad o intimidad. Desde luego que todas las organizaciones complejas enfrentan inexorablemente tensiones dinámicas con la sociedad en su sentido extendido. Ya Adam Smith había anotado el carácter público del propio mercado. Este debe ser abierto a los ofertantes y demandantes; la información debe circular sin limitaciones, etc. (J. K. Galbraith, 1989: 84-97; J. Genereux, 2001: 81-88). Empero, Bozeman confunde los planos de análisis, ya que la dicotomía público-privado es necesariamente una disyuntiva de tipo político (J. Freund, 2004: 280). Para entender lo anterior, debemos tener en cuenta que lo público es dúplice al igual que lo privado: a) De una parte tenemos lo público como propiedad de todos (la sociedad políticamente organizada) vs. lo privado como aquello restringido en su uso y definido por el principio de exclusión. b)Además está lo público como espacio abierto, virtualmente a todos (la opinión pública, la publicidad, el derecho a la información, la rendición de cuentas, etc.) vs. lo privado como intimidad (fuero íntimo, libre conciencia, no autoincriminación, presunción de inocencia, desobediencia civil, etc.). 25 Bozeman reduce la dicotomía sólo a su segunda variante y no toma en cuenta a la primera, que es precisamente el ámbito primordial de las políticas públicas y de la relación Estado - Mercado. Si todas las organizaciones fueran verdaderamente públicas, la sociedad se constituiría sobre bases no capitalistas y no liberales (en el sentido filosófico de la expresión, como ámbito de derechos y fueros del individuo). La dicotomía público-privado en el terreno de la gestión se refiere, por lo tanto, a la contraposición ontológica entre el Estado como mediador social y representante (real y/o simbólico) del interés de todos, de la sociedad políticamente organizada vs. el universo de los derechos de propiedad y mercantilización de bienes y servicios, con ánimo de lucro, de la parte del empresariado capitalista. Sin embargo, en el discurso gerencial dominante de esta idea —de anular la separación entre lo privado y lo público— se apoya en la convicción de que el Management, es ante todo, instrumental, técnico, y que por ello, de sus proposiciones, herramientas, etc., puede hacer uso de forma indiferenciada cualquier tipo de organizaciones. Se nos dice, por ejemplo, que las grandes organizaciones empresariales de corte transnacional son semejantes a los Estados; tiene agencias de relaciones exteriores; sus recursos son muchas veces superiores a los de los pequeños Estados y así ad infinitum. Esta argumentación —llevada a sus consecuencias extremas— haría imposible diferenciar sustancialmente a las burocracias estatales de las burocracias empresariales. Las burocracias poseen, desde luego, rasgos comunes, independientemente de su naturaleza pública o privada. La oposición entre los dos términos no es, por ello, absoluta sino relativa. Pero ello no quiere decir que podamos sin más borrar de un trazo sus diferencias, aunque tengan puntos confluentes. La base de tales identidades reside en las lógicas de racionalidad comunes entre ellas. La noción de una valoración central de la ecuación medios/fines (la eficacia y la eficiencia organizacional como normas de actuación) es un elemento central en el análisis de la burocracia y tiene su origen en la sociología weberiana. Allí existe un campo fuerte de identidad de los espacios burocráticos públicos y privados. Recuérdese que las tesis de Weber no enfatizan en tal diferenciación porque él veía en el ascenso irresistible de la burocracia, el dominio de la racionalidad capitalista, una ascesis que marcaría las improntas de la Modernidad (M. Weber, 26 1985). Las diferencias apuntan en general a la mayor discrecionalidad existente en las burocracias privadas respecto del papel de la legalidad estatal. Por ejemplo, la diferencia manifiesta en los regímenes laborales de los Estados territoriales y de las burocracias multilaterales. Igual en los mecanismos de rendición de cuentas o de responsabilización más flexibles en la empresa privada o en el tercer sector que en los gobiernos. La confusión que presenta este tipo de argumentos nos conduce al punto anterior. Una cosa es el ámbito público en el sentido de macrosociabilidades compartidas, entre ellas las técnicas y tecnologías, el tamaño, la magnitud del presupuesto o del personal adscrito, la complejidad de la gestión; y otra muy distinta su naturaleza política. Jamás podremos reconocer en una gran corporación, a pesar de ser éstas generalmente sociedades anónimas abiertas al eventual comprador de derechos de propiedad en las bolsas de valores, el equivalente a un Estado o a una organización público-estatal. Sobre este tema un punto adicional. En un texto reciente, el Premio Nobel de Economía, J. Stiglitz (2002a), al evaluar las diferencias y confluencias entre los dos términos aquí analizados, afirma que la usual frontera según la cual el sector privado se guía por el lucro —en tanto que el sector público no lo hace— es una afirmación falaz. Para ello cita a otro Nobel, Herbert Simon. Según Simon (1997), los gerentes de las grandes corporaciones privadas se comportan como los del sector público. Es decir, no son motivados por las ganancias de dichas corporaciones6. Igual ocurriría con los empleados, cada cual a su escala. Las motivaciones económicas del lucro sólo estarían en la mente de los propietarios; pero no en las personas que integran estas organizaciones, las cuales son asalariadas y, en tanto tales, no son directamente recompensadas en función del lucro obtenido. Este es, como se sabe, un tema recurrente en la literatura administrativa que destaca un rasgo que el propio Marx alcanzó a entrever: la separación entre la administración y la propiedad7. En la medida en la que el control real de las empresas, en particular aquellas bajo formas anónimas, “el arquetipo de la gran corporación” —nos dice Stiglitz—, la gerencia constituye un bien público al igual que en las organizaciones económicas gubernamentales. La teoría económica nos habla de la relación entre la administración y los propietarios privados como un nexo entre Principal y Agente. Un mandante y un mandatario. 27 Por supuesto, lo mismo se puede decir del gobierno o de una gran ONG. La función administrativa —históricamente— se ha constituido siempre en un estadio maduro de su evolución, sobre la relativa autonomía de los gerentes, de las burocracias que se relacionan con sus dueños o jefes políticos, o con la sociedad (en el caso de los bienes públicos) a través de formas de rendición de cuentas en el modelo ideal de gobernabilidad democrática; o, lo que es mas frecuente, en formas rituales en las que el espectro de lo administrativo conserva el control de hecho de las organizaciones y de las formas de propiedad bajo su jurisdicción fáctica. Lo que se presenta —en síntesis— es una privatización o control patrimonial de lo privado (de los managers sobre lo público, en las organizaciones del Estado; o sobre los derechos privados de los accionistas y dueños) en el primer caso. Pero esto no transforma en pública a la gerencia corporativa, sólo nos da cuenta de una tensión esencial entre derechos de propiedad y jurisdicciones formales frente al espacio delegado que constituye la razón de ser del oficio managerial. Tal antítesis es inevitable, en sus diversas escalas, allí donde exista gestión autónoma. ¿En dónde está el espacio de la identidad de lo público y lo privado? Dado que una antinomia total no existe entre estos dos polos, las identidades resultan del proceso de interdependencia, donde la parte fuerte es y ha sido siempre el espacio de lo público. En efecto, sin lo público, sin las sociabilidades macro, la privaticidad no tendría posibilidad alguna de subsistir. De tal suerte que la identidad de lo público frente a lo privado no es explicable mediante el supuesto reiterado en el discurso smithiano del interés general como fruto de la astucia inconsciente del egoísmo. Los intereses privados juegan un papel significativo en la economía mercantil capitalista, pero no han tenido tal rol en las llamadas sociedades tradicionales, aunque en ellas, la relación público-privado ha estado presente8. La privaticidad sólo ha tenido sentido bajo un orden político que presupone como su condición de existencia a la vida pública. El malentendido se explica quizás por la influencia de la tesis de Rousseau, de la que hizo eco Marx, sobre la propiedad privada como fuente originaria de la desigualdad; transformada por este último pensador en una teoría sobre el origen del Estado. Según tal 28 tesis, el móvil constitutivo de éste habría sido el de salvaguardar los derechos privados, la exclusión social de la propiedad. Así, nos explicamos la percepción corriente de que el Estado depende constitutivamente de la institución de la propiedad privada. La paradoja —como puede verse— es que este tipo de argumentos, aunque disímiles, conducen a un punto confluente con el análisis de Smith, que es hoy un lugar común del ultraliberalismo. En otras palabras, debemos tener presente que, en general, lo público subsume a la esfera de lo privado. Si pensamos en el mercado, este es una institución pública, que cumple los principios esenciales de lo público en el segundo sentido que mencioné anteriormente, pues éste implica la inclusión de todos los que deseen participar en él, pero no en tanto propiedad compartida sino, al contrario, en tanto portadores de derechos de propiedad excluyentes que cada cual desea ceder a cambio de un contrato equitativo. Así mismo, recordemos que las reglas del mercado son universalistas: libre acceso al mismo, contratos voluntarios entre las partes, interdependencias, etc. Por supuesto que los mercados no son propiedad de los gobiernos —ni de nadie—. Un mercado en su sentido extendido que ostente un principio de propiedad es un contrasentido semántico y la monopolización del mismo implica su quiebra ontológica (A. Wolfe, 1997). Además, sin Estado no hay mercado. Lo público como sociabilidad requiere un nivel de tutelaje —que es garante de su existencia continuada— de parte de la esfera de lo público-estatal. Incluso en el ámbito internacional el globalismo neoliberal requirió la invención de los controles supraestatales del mismo, tales como el G.A.T.T., la O.M.C., el World Bank y el F.M.I. Estas instancias estatales y transestatales son fundamentos, ciertamente asimétricos, de última instancia, de un segmento fundamental de las transacciones del mercado mundial. Como lo señaló agudamente G.F.W. Hegel (1986), lo universal es —en esencia— siempre, la síntesis de lo particular, la unidad de lo diverso. Así, Hegel nos hablaba de la célebre “unidad de los contrarios”. Por ejemplo, en las guerras, los bandos enfrentados tienden a mantener ciertas simetrías, independientemente de sus motivaciones, intereses e ideologías. Así, se comparten lógicas disciplinarias, rangos, tácticas y estrategias, éticas guerreras, etc., sin que ello implique la pérdida de sus condiciones de antagonismo extremo. 29 Existe por lo tanto, una unidad fundacional entre los oponentes —el fenómeno que denominamos guerra—; pero esta Unidad es al mismo tiempo, el escenario de la confrontación humana llevada a su máximo extremo. En ese sentido filosófico, sustento la unidad de base de las identidades de lo público y lo privado y el papel de mediación entre estos términos por parte de lo público-estatal y de lo público no estatal, según se trate, respectivamente, de una oposición fundada en el primer caso (el Estado), en la política (para ordenar el límite de los derechos de propiedad, primordialmente); y, en el segundo, en la publicidad (la exhibición o libre acceso a la observación general). 2. Los desafíos provenientes de las sociedades civiles En efecto, si del discurso teórico pasamos a las realidades de la correlación público-privado, encontramos de forma manifiesta la preeminencia de su contraposición. En realidad, pese a que en la retórica de la vida pública en apariencia es muy fácil lograr consensos sobre metas del interés público, tales como el logro efectivo de la democratización de la vida social (R. Dahl, 1992, 1988), la lucha por erradicar la pobreza y generar políticas de equidad e integración para hacer frente a la creciente marginalidad (A. Sen, 2000), o la eliminación de la violencia y las guerras, lo cierto es que dichos consensos se fragmentan en tres instancias. 2.1. LOS DIFÍCILES ACUERDOS TEÓRICOS En efecto, la primera dificultad para lograr el mencionado consenso es de orden teórico cuando específicamente se intenta determinar conceptualmente qué entendemos por tales metas. ¿Es la paz simplemente la mera ausencia de la violencia directa (J. Galtung, 1996), o tiene relación inherente con un cierto umbral mínimo de dignidad de la vida humana? Y esto vale para el resto de conceptos fundamentales en los asuntos de interés público. Así, las ideologías políticas y morales, o en términos más explícitos, las diferentes preferencias sobre la “vida buena” multiplican de entrada los disensos. Esto obliga, por lo tanto, a que las distintas opciones se constituyan a sí mismas en función de una fundamentación filosófica determinada. Generalmente este tipo de debates 30 llevan a la cuestión ontológica, es decir, a la preferencia o prioridad de ciertos bienes según las diversas teorías que escalan los valores. Buena parte del propósito de este libro es, precisamente, de naturaleza conceptual. La literatura referida a lo largo del mismo hace referencia a dos planos del problema: los acuerdos sobre las visiones genéricas de lo público y lo privado, y, de otro lado, la construcción de ciertos conceptos operacionales claves donde se espera igualmente un consenso sobre sus usos extendidos (privatización, Modernización, regulación, subsidiaridad, etc.). O sea utilizable en el análisis de políticas sobre la base de acuerdos extendidos en la propia comunidad académica. La filosofía moral contemporánea gira, desde hace unos cuarenta años, en torno de esta relación y se dividen sus aguas en torno a la primacía de los intereses públicos o de los derechos de los individuos en tanto sujetos jurídicos dotados de personalidad autónoma. La polémica entre comunitaristas y liberales en los ee.uu. es una forma específica de tal confrontación. Igual ocurre con la referencia reiterada al modelo de justificación de la sociedad “bien ordenada”, elaborado por el filósofo de Harvard, John Rawls (J. R. Carracedo, 1990). No pretendo entrar a fondo en dicha discusión porque me alejaría demasiado de los objetivos que me fijé al escribir este libro, y además, por su propia extensión. Pero sí quiero precisar, en adelante, algunos de los problemas directamente conectados con la dicotomía que enfrenta al interés privado con el interés público. Volvamos, en primer lugar, sobre la noción de lo público. Ya expliqué los términos dúplices de su constitución. Hablemos ahora de un derivado de esta primera definición: el concepto de bienes públicos 9. Si entendemos con precisión la primera parte, o sea el universo de los bienes, nos estaremos remitiendo a un conjunto tanto tangible como intangible de valores socialmente reconocidos y que se entienden como compartidos por todos los miembros de una sociedad políticamente organizada (o, en su extremo, como lo mostraré más adelante, por la Humanidad entera). Amartya Sen (2000) ha hecho un intento reciente por definir este ámbito. Su crítica a Rawls y su postulado de la prioridad de la libertad, que la constituye en el bien primario por excelencia, 31 nos muestra que la libertad ocupa una plaza central en el orden democrático; pero que es necesario contrabalancear la idea de libertad con la de igualdad de oportunidades. Si a la libertad se le da un peso absoluto, ello nos conduciría al modelo radical de Robert Nozick, que es un cuasiestado de naturaleza (sobre Nozick, véase el capítulo seis, R. Nozick, 1974). El interés público no se apoya en la libertad sino en la igualdad de oportunidades que se expresa filosóficamente como equidad (fairness). El interés público supone que los intereses del colectivo políticamente organizado que llamamos sociedad tienen preeminencia sobre los intereses particulares de sus miembros. Y si se trata efectivamente de garantizar la libertad de todos, ésta es un valor (es decir, un bien público) que debe ser compartido de forma inclusiva, de tal forma que la garantía de la libertad de elección de todos implica restricciones para los excesos de libertad de algunos. Así se evita que se deteriore o afecte la posibilidad de que ésta sea un bien público. El asunto decisivo, en términos de políticas públicas, es la función tutelar del Estado para que todos los miembros de la sociedad tengan acceso asegurado a los bienes públicos. Este universalismo de acceso tendría prioridad y fijaría restricciones relativas al ejercicio de la libertad. En tal medida, ésta no sería el valor central del interés público, lo cual es un punto que diferencia mi argumentación de la propuesta por A. Sen. Utilizo su caso como un importante ejemplo argumentativo, pero no pretendo hacer un listado de bienes públicos ni fijar a priori una jerarquía entre estos. En el análisis de políticas me interesa más indagar, en las realidades concretas, cómo las sociedades asignan estas preferencias y los recursos (medios) para lograrlas. Esto lo digo no por hacer relativismo sino porque lo importante es construir una herramienta interpretativa que sea útil para las diferentes perspectivas de tipo moral. 2. 2. LOS CONFLICTOS DE INTERESES La segunda instancia, es por supuesto, la del reino de los intereses concretos. Pese a lograrse hipotéticamente el mencionado consenso teórico sobre el significado de las metas de interés público, permanece el problema, puesto que las acciones de política pública, pretendiendo concretar tales metas, no son del tipo suma cero, en el que nadie pierde y todos se benefician (el ganant - ganant, según la ciencia política francófona), sino que son más bien del modelo descrito por el economista 32 y sociólogo italiano de comienzos del siglo XX, W. Pareto, en donde la asimetría de la distribución resulta generalmente inevitable. Y siempre que alguien gana es a costa de la pérdida proporcional de otro(s). La actual dinámica del capitalismo mundial, con un elevadísimo proceso de mercantilización de los bienes y servicios públicos, nos muestra la carencia de un límite adecuado, de una frontera conceptual claramente observable, que defina hasta dónde ha de llegar la mercantilización. Así, el principal peligro que enfrenta lo público reside, precisamente, en la tendencia a mercantilizarlo todo. Las experiencias culturales y las formas subjetivas y estéticas de apropiación del mundo son uno de los espacios que han sido arrastrados por la ola mercantilizadora (J. Rifkin, 2000). Las prácticas y ejecutorías gubernamentales han oscilado desde su inicial monopolización estatal de buena parte de los bienes públicos sustrayéndolos al mercado, hasta la actual predominancia de su provisión por la vía del outsourcing. Incluso, los ejércitos contratan la capacitación de su personal; por supuesto, la mayoría de las provisiones y el avituallamiento, pero también el desminado de puertos y zonas de guerra (como en Irak recientemente por parte del ejército de los ee.uu.), la información satelital y las comunicaciones. Además, los Estados delegan frecuentemente en firmas privadas el recaudo de ciertos tributos, la representación del interés nacional, a través de formas múltiples del llamado partenariat. De tal modo que la salud, la educación, la justicia, la seguridad —parcelas significativas de la administración pública al nivel que llaman en las Ciencias del Management “tecno-estructura”— son privatizadas y/o mercantilizadas. Como lo expliqué con detalle en un trabajo reciente (E. Varela, 2005b), no se trata de oponerse per se, por principio, a la mercantilización de todos los bienes públicos10. El problema estriba en trazar una línea divisoria entre aquello que se debe y lo que no se debe mercantilizar. En general, el límite podría demarcarse cuando la mercantilización afecta seriamente los principios de universalidad al acceso o disfrute de tales bienes. Obviamente, cuando el acceso no está determinado en función del peor servicio obtenible o el peor bien disponible en relación con la capacidad adquisitiva de los ciudadanos. Y de otro lado, cuando la oferta mercantilizada implica severas desventajas para porciones de la sociedad, que son simétricas a los beneficios de tipo patrimonial que obtienen los agentes económicos que producen y trafican los bienes y servicios en cuestión. 33 Un ejemplo a la mano: en el Québec se ha debatido duramente el outsourcing en el sistema hospitalario público, pues al privatizarse el aseo de los hospitales, el vestuario y, en general, los ítems de lavandería, las empresas concesionarias, por maximizar ganancias, hacen la tarea por debajo de los estándares mínimos de seguridad para los pacientes y el personal clínico. Como resultado de lo anterior, estalló un escándalo cuando se descubrió que este tipo de deficiencias habían causado la muerte de un número dado de pacientes y de personal; además de la comprobada contaminación de un sector de las gentes de visita en dichas instituciones. El tema de limpieza no es, en modo alguno, algo adjetivo en un hospital (!!). En Colombia podemos encontrar otro ejemplo. En las zonas marginales de Cali, mi ciudad, se contrata desde hace años el servicio educativo de las escuelas primarias con ONG, cooperativas, iglesias, etc., a través del método de capitación, o sea, fijando un precio por el costo de cada alumno por año escolar. Esto se hace mediante una suerte de subasta, y por ende, la administración pública privilegia el otorgamiento de los contratos a los que garanticen mayor cobertura. Como resultado de tales reglas de concurrencia, los concesionarios rebajan los salarios de los maestros, minimizan las dotaciones y útiles escolares, improvisan locales poco o nada adecuados, etc. Así, aunque los más pobres figuran orgullosamente en las estadísticas de ampliación de cobertura, en realidad se les está dando la peor educación, la que no garantiza movilidad social alguna. Si bien es cierto que la educación privada de elite es un negocio en el que, por regla general, al tiempo se satisfacen el ánimo de lucro de los empresarios con la calidad de la oferta, éste no es el caso de la subcontratación en las zonas marginadas de la sociedad. Este tipo de desfases es lo que justifica teóricamente que exista un sistema educativo público, para nada mercantilizado, desde los niveles más básicos hasta la educación superior universitaria de ciclo doctoral y que la limpieza en los grandes hospitales estatales sea asumida por las propias entidades y no se la subcontrate. 34 2.3. EL AUGE DEL INDIVIDUALISMO Y EL PROBLEMA DE LAS DESIGUALDADES Un tercer gran ámbito de desafío al interés público proviene de la apropiación social cada vez más universalizada a escala planetaria de los postulados del individualismo filosófico. Esto es, en gran medida, el efecto directo de la implantación del orden capitalista y mercantil de producción y consumo, con su eje medular: el sacrosanto postulado de la propiedad privada y su corolario, la acumulación indefinida de riquezas. La ofensiva ideológica y práctica contra el Estado de bienestar y las políticas de inclusión social, son los peligros mayores de la actualidad por parte de esta corriente que tiene un influjo significativo en los grandes centros de decisión de naturaleza global. Y se corresponde con los intereses estratégicos de los factores de poder beneficiarios de la actual deriva globalizada. Con esto no pretendo instaurar un alegato moral anticapitalista del tipo ‘no se puede realizar el interés público mientras predomine el orden individualista de las sociedades capitalistas de mercado’. Esta es una manera simplista o quizás elegante de renunciar a la realización de intereses colectivos. Pues es sabido que, al menos en el contexto de nuestra temporalidad, el capitalismo ha sido capaz de sobrevivir a sus crisis y posee ciertamente una gran potencialidad de “reinvención” respecto al socialismo de Estado que fracasó en su experiencia alternativa de eliminar el orden mercantil dentro de las estructuras sociales. Como lo advertía E. Durkheim, la cuestión verdaderamente importante es explicarse por qué y cómo es posible un orden social basado en principios universales de inclusión, cuando en la base de estas sociedades predominan las ideologías utilitarias e individualistas (C. Laval, 2002). En frente de este tipo de interrogantes encuentro un mérito indiscutible en la respuesta largamente formulada por J. Rawls (1971), referida en particular a su noción operacional de “consenso traslapado” (overlapping consensus). Pienso que no es indispensable para utilizar la noción de tal consenso el aceptar la idea básica de Rawls de un neocontrato originario, de clara inspiración kantiana y rousseauniana11. Pero una visión de las diferentes sociedades “realmente existentes” nos muestra que ellas funcionan, existen, poseen redes sociales, vínculos duraderos que les dan sentido, pese a las enormes diferencias, tensiones, asimetrías y exclusiones que las caracterizan. 35 Se existe socialmente en la desigualdad ya que la igualdad sólo tiene un sentido operacional en la institucionalidad de las reglas formales de constitución; al menos en las sociedades que se reconocen como demoliberales. En la vida cotidiana, en frente de las realidades objetivas y constriñentes de existencia, el consenso se da sobre las reglas de funcionamiento, aunque muchos las acaten a regañadientes y quieran postular otras en su lugar. Las propias sociedades socialistas de inspiración marxista estuvieron muy lejos de la igualdad real. La nomenklatura burocrática de las elites gobernantes cumplió, en forma centralizada, una función de apropiación de las riquezas, los privilegios y el poder, semejante a la del señorío medieval y las formas patriarcales de dominación. Y su caída no derivó de la comprobación de la inequidad persistente, sino de la quiebra de las reglas de su institucionalidad constitutiva. De hecho, las formas sociales que las han reemplazado han acentuado enormemente las desigualdades “originarias”. En la referenciación concreta de las desigualdades y las exclusiones, la existencia misma del interés público implica por lo tanto la existencia reconocida y actuante de consensos fundantes. El orden político es, en tanto base mínima indispensable de las sociedades, un bien público. Los derechos y deberes sociales vinculantes, esenciales en un horizonte sociocultural dado, constituyen otros tantos bienes públicos; ambos del tipo procedimental. Ellos operan a la manera de medios, la posibilidad de que las sociedades se fijen consensualmente fines morales y que, a partir de estos últimos, los individuos actúen en libertad para concretar sus elecciones y preferencias. Desde luego que la configuración de consensos morales totalizantes es algo impensable, en términos absolutos, en las sociedades individualistas posesivas (C. B. MacPherson, 1962). Allí, aparecen las nociones desarrolladas por el pensador neomarxista italiano A. Gramsci (1971), las de hegemonía y bloque histórico. A través de los juegos del poder, los intereses dominantes generan una imposición de tipo valorativo que subsume a los sectores subalternos de las sociedades, mismos que se ven impelidos a aceptar los principios y metas de actuación a la manera de obligaciones morales de tipo impositivo (E. Durkheim, 1978). De tal modo la concreción históricamente dada de un conjunto de valores morales es más fruto de la acción-reacción en múltiples escalas —del proceso tenso 36 de parte de fuerzas sociales diversas y confrontacionales— que de un consenso meramente racionalista o dialogístico, como lo postula la perspectiva habermasiana (J. Habermas, 1998; C. Calhoun, 1992). No hay, en la realidad de las sociedades, espacio para un mítico Contrato Social. Existen en realidad vínculos aherrojantes, interdependencias y asimetrías en el uso de los bienes públicos y en el ejercicio real de los derechos. También ha de reconocerse que un porcentaje elevado de las luchas y conflictos sociales se resuelven con algún tipo de compromiso entre las partes involucradas, de tal manera que la ganadora no obtiene lo óptimo o ideal, sino lo posible. Sin embargo, no sugiero en modo alguno que de este tipo de compromisos diarios (entre patronos y sindicatos, entre consumidores, proveedores y productores; la sociedad civil y el Estado, y así ad infinitum) surja el interés público. En mi perspectiva, éste no es negociable. Como lo comprendió muy bien Hegel y la tradición objetivista que le siguió (Marx, Durkheim, entre otros), los intereses públicos no son el resultado ni de la suma de los intereses egoístas de los individuos ni de pactos utilitarios entre estos. De allí que sólo una parte y no la totalidad de las ejecutorías de política de un Estado cualquiera sean realizaciones del interés público. La mayoría de tales dimensiones se corresponden más con el influjo de intereses privados sobre los gobiernos, o con las propias agendas de las elites y burocracias que detentan el poder político. 3. La preeminencia del Estado sobre el interés público En este punto quisiera inicialmente despejar el equívoco que identifica el interés público con el papel del Estado. Ciertamente, muchos suponen que sólo a partir de la acción gubernamental se garantiza y configura realmente el dominio de lo público. Más adelante (en el capítulo tercero), veremos de qué forma lo público va mucho más allá de lo estatal. Tanto si concebimos lo público como macrosociabilidad o, de forma alterna, como el espacio de la publicidad. Este último nivel es el de la exposición de las cosas, las personas, las ideas, las actuaciones, etc., al libre escrutinio de los demás, sin significativas limitaciones. Obviamente en la construcción de verdaderas políticas públicas, el Estado —en tanto instancia suprema de obligación política— ha de concretar las 37 metas sociales de carácter universalista. Pero en la vida cotidiana una fuente de desestabilización del interés público que reviste gran incidencia la constituyen los propios Estados. Podría comenzar reseñando la vigencia actual de la célebre raison d’Etat. Lo que sigue siendo un típico topos hobbesiano. El Estado se reserva para sí el uso discrecional —cuando se enfrenta al peligro de su propia sobrevivencia— del antiguo derecho natural. Vale decir la regimentalidad de excepción que le permite a los gobiernos suprimir los derechos de los ciudadanos, disolver las cortes jurisdiccionales y los parlamentos. En suma, apoderarse del universalismo de las leyes y del derecho público. De estas situaciones, en América Latina tenemos suficientes experiencias, principalmente por la puesta en escena del autoritarismo militarista en el siglo pasado. Y desde luego, maximizado. Este es el fondo del totalitarismo (véase el capítulo tres). Los principales peligros de lo público en frente de la acción estatal se derivan del abuso que se hace, desde el poder político, del propio fuero gubernativo para transformar al Estado, de un medio regulatorio e institucional de las sociedades, en un fin en sí mismo. Aún los llamados neoliberales son dubitativos e inconsecuentes al momento de reducir ciertas funciones claves del gobierno, por ejemplo, en el terreno de la seguridad. Países como Inglaterra y los ee.uu. han incrementado su capacidad de observación y vigilancia sobre las conductas de sus ciudadanos, los viajeros y las propias organizaciones de la sociedad civil12. En una instancia semejante a la anterior, las formas de hegemonía y la perversión de la democracia como dictadura de las mayorías, constituyen otros grandes desafíos a lo público. En general, se exalta a la democracia procedimental como la estructura deseable o ideal de legitimación de los regímenes políticos. Sin embargo, la decisión de las mayorías no equivale siempre al interés público. Lo vimos en ee.uu. con la reelección de George Bush hijo. La democracia electoral no se escenifica en el aire; ella tiene lugar a través del influjo decisivo del dinero, el papel de las grandes corporaciones empresariales, el uso del Estado y sus recursos a favor de las reelecciones y, finalmente, por medio de la manipulación de los massmedia. La democracia expresa y refleja, además, los prejuicios y valores morales de los ciudadanos transvestidos en electores. 38 El nacionalismo y el chovinismo, las discriminaciones y violaciones de ciertos tipos de derechos cuentan a menudo con la aquiescencia de las mayorías. Es por eso que los grandes teóricos del liberalismo como J. Stuart Mill (1978) alegaban a favor de los procedimientos para proteger a las minorías. De otro lado, acechan también al interés público la ingobernabilidad y la quiebra de la función cohesionadora de los Estados. Esta patología tiene múltiples manifestaciones: burocratización, desregulación, crisis de fiscalidad, etc. De forma adicional se expresa en la carencia de políticas y estrategias que concreten la acción estatal en pro del interés público. Asunto especialmente importante para corregir y enfrentar las principales formas de exclusión y marginalización social. Pero generalmente esto obedece a la asignación por las elites del poder de otras prioridades de política en la agenda gubernativa. No dejemos de notar que en numerosas ocasiones desde los Estados se incitan e instrumentalizan la inmensa mayoría de los conflictos que desgarran y destruyen la gobernabilidad a través, principalmente, de la promoción de las guerras, la conexión de los Estados con específicas lógicas de la criminalidad globalizada, o en frente de conflictos no mediados. Es decir, a través de la inacción e indiferencia convertidas en políticas públicas de discriminación. Así que en el Estado, como en un prisma, nos encontramos tanto con las principales oportunidades de vehiculizar el sentido de lo público, como con grandes obstáculos a su concreción. 4. El interés público y sus dimensiones ecuménicas Las dimensiones que son comprobables en la literatura que podríamos llamar clásica sobre el tema son, en general, las del Estado Nación. Tal es el sentido genealógico descrito por J. Habermas (1982) en su texto fundacional. Sin embargo, la expansión mundial de los circuitos relacionales entre las sociedades, la creación —por vez primera de forma completa— multidimensional, del mercado mundial, las tendencias sociales, políticas y culturales de la globalización, han agudizado un fenómeno que por cierto no es novedoso: el de la referenciación de lo público en una dimensión ecuménica. Nuestros tiempos nos muestran en forma acusada la extensión de la conectividad global entre las diversas sociedades, tanto en sus generalida39 des incluyentes (lo público ecuménico) como en numerosísimas relaciones particulares; sobre todo en la forma de redes relacionales globales y regionales. Éstas atienden a focalizar la atención sobre intereses comunes y crean comunidades virtuales (en lo académico y científico; pero también en los usos del tiempo libre —diversiones y hobbies, pero también perversiones(!)—). Éstas son facilitadas a través de la Internet, los massmedia; la facilidad actual de circular físicamente por el planeta y comunicarse en seminarios, congresos, medios impresos y audiovisuales, etc. Una gama significativa, crucial, de problemas se interpreta cada vez más ontológicamente en su existencia universal, anclada a la noción de Humanidad. Citemos, en primer lugar, la fórmula actual de reconocimiento y demanda de los Derechos Humanos. Una de las dificultades mayores de los teóricos soberanistas es, en efecto, conjugar la defensa de las soberanías estatales como la dimensión esencial de la vida política, y mantener que esta instancia ofrece suficiente espacio de garantía a los derechos de las personas. De allí la noción alternativa, altamente debatida, del intervencionismo humanitario (E. Varela, 2005a). En efecto, ¿qué hacer cuando devastadores desastres naturales desestabilizan vastas áreas del planeta, ante la incapacidad de los aparatos estatales desvertebrados por esos desastres o con escasa capacidad de respuesta institucional? o ¿cuándo los propios gobiernos son los causantes y principales catalizadores de desastres humanitarios, guerras, hambrunas, genocidios y dantescos desplazamientos? El espacio de lo público, la responsabilidad ecuménica coloca en cuestión la propia noción de la Solidaridad. Y a quienes consideran este tipo de argumentos como justificatorios del intervencionismo a ultranza quisiera recordarles que la fundación teórica del humanitarismo, en una forma ciertamente limitada, se encuentra a mediados del siglo XIX en la proclama de K. Marx sobre el “internacionalismo proletario”. Incluso los propios portaestandartes de la Revolución Francesa pensaban en función de la emancipación de la Humanidad entera. Las utopías socialistas de los siglos XVIII y XIX estaban teñidas del ideario ecuménico. Pero fueron las revoluciones burguesas consolidadas (E. Hobsbawm, 1998) las que reemplazaron los ideales altruistas e idealistas por la conflictiva prioridad del interés nacionalista. 40 Un militantismo intervencionista es hoy claramente observable de parte de los movimientos antiglobalistas o, más exactamente, partidarios de una globalización con “cara humana”. Me refiero principalmente a la dinámica generada por el Foro Social Mundial creado en Porto Alegre, Brasil, como la contra faceta del organizado en Davos, Suiza, por las elites beneficiarias de la globalización neoliberal; pero también a las numerosas ONG de carácter global, que son las contra facetas de las grandes corporaciones multinacionales privadas. Así, uno de los efectos positivos de los procesos de globalización es la puesta en escena por vez primera en términos prácticos, alcanzables o viables, de los derechos ecuménicos. Piénsese, por ejemplo, en el ámbito económico de los derechos de propiedad. ¿A quién pertenecen los mares?, ¿los recursos naturales en su sentido lato?, ¿la continuidad y sustentabilidad de todas las formas de vida (fauna y flora) del planeta?, o, ¿la “conquista” del espacio extraterrestre? ¿Acaso al primer “ocupante”? Este tipo de debates obviamente, aunque pueden ser asuntos académicos, son en realidad cuestiones vitales —de factura política— en la opinión pública mundial. En el tema del mar, a manera de ejemplo, recordemos que la vieja discusión —que aún no termina— era los alcances de la soberanía costanera de los Estados territoriales. Esto llevó a un desarrollo complejo y doloroso del derecho internacional, para intentar zanjar conflictos que con frecuencia se han dirimido por el primado de la fuerza. Sin que este tipo de conflictos haya sido superado, ahora nos enfrentamos a otros desafíos. El agotamiento acelerado de la riqueza ictiológica requiere desde hace décadas de algún tipo de arbitraje, control o autoridad, que garantice el acceso equitativo de los países y de las grandes empresas, pero también de las comunidades ribereñas, a su disfrute continuado, al usufructo compartido sin que se precipite una catástrofe ecológica. Pero pese a los relativos avances en el campo de la diplomacia interestatal, en general domina la inacción desde el punto de vista de la construcción y puesta en marcha de políticas públicas globales. La misma situación se evidencia en el resto de la agenda ambiental (el llamado calentamiento global y las alteraciones climáticas; la sostenibilidad de la biodiversidad de la fauna y 41 la flora planetaria; la contaminación ambiental en sus diversas especificaciones; los desastres nucleares del tipo Chernobyl, etc.). La agenda global no se detiene en estos temas que son hoy recurrentes en los massmedia y en el debate público internacional. Además, con menos ruido, están en cuestión los procedimientos y los principios interpretativos del uso de la espacialidad terrestre compartida. Las migraciones, el derecho de asilo de diverso tipo y no sólo el llamado asilo político, la mercantilización ilegal de la vida humana, las redes globales de tráfico sexual, hasta los asuntos controvertidos de los derechos mundializados de propiedad intelectual. En este último terreno se creó una autoridad planetaria, la O.M.C., que dirime algunas de las cuestiones, pero cuyo arbitraje es bastante criticado. Y en el campo patrimonial, la UNESCO adelanta ciertas acciones significativas pero largamente insuficientes. Existe, por lo tanto, una profunda fractura entre la realidad de lo público-ecuménico, la idea tan reiterada hoy del “patrimonio de la Humanidad entera”, y las formas políticas estatales que no abarcan esta dimensión, salvo en la forma rudimentaria de la eufemística “comunidad internacional”. Ello se agrava, en el contexto actual del globalismo neoliberal, por la pretensión y actuación consecuente de los conglomerados y corporaciones multinacionales, que buscan llenar el vacío jurídico político, amplificando el derecho privado hacia el espacio de lo ecuménico. Como lúcidamente lo ha postulado J. Rifkin (2002), la ofensiva de las grandes empresas busca colonizar a través del reconocimiento, muchas veces unilateral —v. gr., por las autoridades judiciales de los EE.UU, de derechos de propiedad sobre la vida, sobre la vida humana misma (caso del GENOMA)— la reproductibilidad de la vida (derechos sobre semillas, genes animales y vegetales, tratamiento de enfermedades, etc.). La esquizofrenia institucional arraiga en dicha discontinuidad, por lo cual el espacio de las relaciones transestatales y transociales es, por cierto, cercana en numerosos episodios al sombrío cuadro hobbesiano de “guerra de todos contra todos”. En realidad, observamos las guerras de algunos —entre los más fuertes— contra otros a los que se sataniza en una aplicación simplista de la lógica de K. Smichtt de “amigo-enemigo”. Por ello —como nunca antes— resulta perentorio instrumentalizar de alguna manera el viejo ideal kantiano del gobierno cosmopolita. Sin embargo, pese a los avances parciales y fragmentarios hacia 42 la gobernabilidad global (D. Held, 2001; U. Beck, 2002) ésta sigue siendo una meta lejana, aún utópica. 5. La instrumentalización del interés público Los anteriores tópicos son todos ellos, en cierta medida, teoréticos. Así se refieran a cosas concretas, corresponden a lo que se denomina “comprensión”, y mantienen en ese sentido un hilo conductor de naturaleza analítica. Por lo tanto quiero cerrar este texto con un punto complementario: los grandes desafíos de lo público en el terreno propiamente instrumental de las políticas13. La pertinencia del estudio de las políticas públicas corresponde, en este caso, a la importancia del desarrollo concreto de las dinámicas de lo público y su relación con el universo de lo privado. Podríamos señalar que un marco general lo otorgan las relaciones sociales, en tanto fundamento existenciario; posteriormente el tramado institucional, como reglas de actuación para, finalmente, aterrizar en el espacio de despliegue de los grupos y factores de poder en conflicto interdependiente. El primer gran problema es la construcción de políticas. Y no tanto porque no se construyan políticas en el día a día. Por ejemplo, en los países con democracia representativa existe una abrumadora capacidad de construcción de políticas basada en la delegación constitucional de tal función al poder legislativo, a los parlamentos. En esencia esa es su tarea central, completada por el llamado control político al gobierno. Pero es notoria la tendencia mundial hacia la productividad de las políticas como una tarea del propio poder ejecutivo. En América Latina la tradición presidencialista y caudillista impele a las elites gubernativas a darle viabilidad jurídica a este tipo de alternativas. Los decretos ley, las normas administrativas que en realidad son protolegales, el control del gasto público recomendado por los organismos multilaterales, coadyuvan a afirmar esta tendencia. En general, en la productividad de las políticas por el gobierno o por los parlamentos, el fenómeno más perturbador lo constituye la práctica extendida del cabildeo o lobbying. En palabras simples, esto implica que los grupos de poder o de presión (C. Wrigth Mills, 1960), en la sombra, afectan los contenidos de las políticas redactándolas de acuerdo a sus intereses. Los grandes 43 grupos empresariales están en la primera línea de esta práctica, pero no son los únicos. Cabildean las elites de las regiones y los sindicatos; con menos suceso, ciertamente, las universidades, las comunidades profesionales y así sucesivamente. En Colombia, adicionalmente, lo hacen los grupos paramilitares —como acaba de verse en las negociaciones para su amnistía— y también las ONG de Derechos Humanos, al igual que los sindicatos con menos capacidad de convencimiento frente a los decisores . Una segunda fórmula de productividad de políticas es su privatización directa a través de outsourcings: think tanks de diverso tipo son contratados con frecuencia para redactar las normas, elaborar estudios y fijar —en borrador— las políticas. Recientemente el BID ha creado un programa para modernizar los parlamentos latinoamericanos y fortalecer su capacidad institucional. Tarea que va en la dirección correcta, pero que requiere reglamentar la función de cabildeo y fijarle marcos que eviten el ser juez y parte, los conflictos de intereses, la sustitución de la democracia deliberativa, etc.; y, de otro lado, crear condiciones técnicas para que los equipos parlamentarios, ellos mismos, estén en condiciones de estudiar los temas y elaborar las propuestas de políticas que serán después debatidas (C. Monzon, 1990). El otro asunto clave es la capacidad de los ciudadanos de participar en el debate público. La perturbación central aquí es el control férreo que los grupos económicos han alcanzado sobre los massmedia. En Francia, por ejemplo, el diario Liberation, antaño vocero del maoísmo más hirsuto, acaba de ser comprado por un poderoso conglomerado económico. Y esto ocurre todos los días en numerosos países. La triada prensa controlada, grupos empresariales y elites político-tecnocráticas —que además circulan fluidamente entre las tres funciones—, pervierte la posibilidad de un acceso abierto de la ciudadanía al debate14. La mercantilización absoluta de la información es otro factor que perturba la calidad de ésta; condición mínima de la intervención de las sociedades civiles en la discusión de sus intereses y de los problemas estratégicos de una sociedad. 44 La fragmentación de la representación política en manos, tradicionalmente de los partidos, que afrontan una crisis aguda de descomposición y, en algunos casos, de recomposición, es un fenómeno que traspasa fronteras. Dado que resulta muy difícil participar insularmente de la construcción de políticas, la desaparición de las mediaciones partidistas, la debilidad estructural de los movimientos sindicales y asociativos de base, la fragmentación y volatilidad de los intereses comunes, generan las condiciones para que los primeros dos esquemas de productividad de políticas sean largamente dominantes. No soy en este punto un pesimista absoluto, pues las sociedades, cuando tienen espacios de deliberación y decisión democráticos, por imperfectas que éstas sean, poseen una capacidad pendular de castigar electoralmente las “malas políticas” generando las alternancias en el poder, fenómeno que hemos visto en casi toda América Latina en los años recientes, cuando los neoliberales han sido prácticamente barridos del poder, salvo en Colombia, ciertamente (!). Lamentablemente estos cambios de dirección del sentido general de las políticas no anulan para nada las dinámicas cotidianas que he señalado. Pues en realidad estamos ante un problema estructural: las deficiencias de la democracia deliberativa en sociedades asimétricas, gobernadas por el mercado y la crematística. Dos breves apuntes finales: aparte de la construcción de políticas, los otros dos grandes procesos que han de tomarse en cuenta son los de: a) la ejecución de las políticas, que es el campo propiamente dicho de la tarea administrativa pública, el quehacer burocrático por excelencia; y b) la evaluación de éstas. El feed back de sus resultados o efectos para avanzar en el aprendizaje social. En estos dos terrenos se ha dado una invasión de la privatización y la mercantilización, por medio del desmonte del modelo del Estado proveedor y ejecutor. La evaluación en la actualidad le es retóricamente entregada a las comunidades según lo pregona el New Public Management. Sin embargo, —como lo veremos más adelante en el conjunto del presente libro— existen serios problemas en la puesta en marcha de estos ideales de control ciudadano. Todo el ciclo de las políticas implica la puesta en escena de las interacciones sociales que construyen el interés público. Este no es el resultado de la mera voluntariedad o de las buenas intenciones. Pero comprender sus bases teóricas y sus mecanismos de funcionamiento es un asunto clave 45 para participar activamente en términos críticos y propositivos. Los retos de política son numerosos y comprenderlos en su estructuración es un buen punto de partida para mejorar la deliberación pública y lograr consensos o acuerdos mayoritarios que plasmen, en los diferentes planos de la vida social, los intereses de lo público, el cemento de la sociedad (J. Elster, 1990). NOTAS En este texto no tengo ocasión de explicar en detalle las características conceptuales de la mundialización o globalización, como usualmente se denota el mismo conjunto de fenómenos en lengua española. Remito por lo tanto, al lector a un libro reciente donde analizo in extenso estas temáticas, titulado: “La Soberanía Transformada” (E. Varela, 2006). Por demás, la bibliografía sobre este tema alcanza hoy dimensiones absolutamente inconmensurables. 2 Estas dificultades resultan de la diversa naturaleza histórica concreta de lo público. Su contenido, límites y significación no son abstractos o atemporales. Como lo mostraré en el capítulo tres, ello depende de las características propias a las diferentes formaciones sociales, e incluso, en un mismo horizonte de temporalidad, de las diferencias civilizacionales y culturales. Véase en este sentido una penetrante compilación que incluye varios estudios sobre el multiculturalismo en la noción de lo público en Asia, el mundo árabe y el África (S. Mappa, 1996; A. Moussaui, 1996; J. Maila, 1996). 3 Toda suerte de acciones sociales, especialmente aquellas que tienen lugar en espacios de sociabilidad basados en vínculos de pertenencia fuertes —como es el caso de la adscripción a organizaciones— se ve afectada por la racionalidad limitada de parte de los individuos involucrados. Los problemas de cooperación -cómo hacer que los subordinados actúen en pos de los objetivos fijados por el propietario o por los altos directivos-, por ello, han sido y constituyen, uno de los temas centrales de análisis en las Ciencias del Management (véase: H. Simon, 1997). 4 Una muestra arquetípica de las críticas recurrentes al papel de las ONG y de la llamada sociedad civil se encuentra en: (J.F. Bayardt, 2001; B. Eme, 1 46 2001). Así mismo, una discusión matizada sobre este mismo asunto en: (J. J. Laville, 1994; J. J. Laville y R. Sansanlieu, 2002). Las críticas están básicamente centradas sobre la idea de que a este sector simplificadamente se lo presenta como el “remedio a todos los males del planeta” (Bayart). Cuando en realidad sería un mecanismo de concertación articulado funcionalmente a los intereses económicos dominantes en la globalización neoliberal (Ibíd.). 5 H. Mintzberg (1986: 400-402) ha construido un extenso y vasto repertorio sobre las ciencias de la gestión. Allí expone y critica varios modelos clasificatorios -tipologías diversas de las organizaciones contemporáneas-. Entre ellas la de Etzioni que diferencia tres tipos fundamentales de organizaciones: las coercitivas, las utilitarias y las normativas. Mintzberg considera que esta clasificación, aunque útil, se queda sólo en los niveles internos de las relaciones de poder. Blau y Scott, a su turno, propusieron un esquema tríptico: empresas comerciales, industrias de servicio y organizaciones gubernamentales. Esta propuesta le parece insuficiente a Mintzberg porque resulta ideal y no suficientemente descriptiva. Él, además, critica la alternativa diseñada por Rehnman (1973), quien diferencia: a) organizaciones marginales, b) grandes empresas, c) organizaciones anexas y d) instituciones. En este caso, considera inadecuada la primera opción; pero reconoce la utilidad de las restantes. Cito estos ejemplos para advertir que, dependiendo de las finalidades analíticas, se podrían construir numerosas taxonomías que, con base en su coherencia, serían útiles. Pero los ejemplos citados no implican una visión realmente ontológica de las organizaciones de nuestro tiempo. La propuesta de retomar el esquema tríptico formulado en este capítulo se enmarca en la referenciación esencial a la dicotomía público-privado: ésta toma en cuenta las dos cuestiones complementarias – inevitablemente entrelazadas– de los derechos de propiedad y del grado de inclusión/exclusión de éstas. 6 En un texto reciente, Stiglitz (2003), abunda en estas comprobaciones, refiriéndose en especial al papel nocivo de muchos de los altos ejecutivos (CEOs en inglés; PDGs, en francés) que se benefician de su posición de control sobre las organizaciones empresariales a su mando. Pero la lectura que, en mi opinión, se desprende de los argumentos de Stiglitz va en otra dirección: en ambas instancias existe un elevado riesgo de que los miembros -sobre todo aquellos con alto grado de control o de poder discrecional-utilicen los recursos de las organizaciones en su propio beneficio. O sea, que el lucro sí motiva al directivo, pero éste es independiente del desempeño y pertenece más al contexto de control de la organización sobre sus miembros. Desde luego que en este plano hablamos de los derechos potenciales de carácter discrecional que se convierten en beneficios o privilegios de los burócratas. Mas allá está la cuestión de la corrupción abierta, desenfrenada, recurrente en el gobierno y en las grandes corporaciones, como lo explicita claramente Stiglitz en numerosos pasajes de su libro. 7 Esta afirmación de Simon -que recoge Stiglitz- es, a mi juicio, bastante laxa. Un porcentaje creciente de las grandes empresas sí evalúa y paga los 47 salarios y otras compensaciones en función del desempeño económico de las empresas. Y esto con el NPM se ha estado extendiendo al sector público-estatal. Por supuesto que esta relación es compleja y nada mecánica. Igual que ocurre con las lógicas de gobernanza de las grandes sociedades de acciones. No sólo se pueden pagar dividendos en virtud de los resultados sino de las expectativas, lo que es el corazón de las funciones ontológicas de las bolsas de valores. De pasada, recuerdo aquí –pero no lo trato por ser ajeno al hilo expositivo trazado en este libro– que este problema (el control de las empresas por los gerentes y su grado de autonomía) es uno de los temas de frontera en las Ciencias del Management, bajo la denominación de “Gobierno Corporativo”. 8 A este argumento me refiero de forma detallada en el capítulo tercero. 9 ¿Son los bienes públicos fines o medios? A. Sen piensa que los bienes públicos son tanto lo uno como lo otro. Pues fijarlos sólo como fines sin suministrar las condiciones, las herramientas y los procedimientos para su acceso universal, es sólo retórica liberalizante, no un verdadero ejercicio de la libertad. “En esta aproximación, la expansión de la libertad es vista a la vez como 1) un fin primario y 2) el principal medio de desarrollo. Estos pueden ser llamados respectivamente el ‘rol constitutivo’ y el ‘rol instrumental’ de la libertad en procura de alcanzar el desarrollo humano?” (traduzco el texto tomándolo de la edición inglesa de Anchor Books, 2000: 36). En general, concuerdo con la tesis de Sen ya que la idea de fines o medios es relativa. Lo que para alguien es un fin central de su proyecto de vida, es, para otro, un simple medio. Y así sucesivamente. Tratar de trazar una distinción de fondo entre los dos niveles no nos lleva realmente muy lejos, tanto en la teoría como en la práctica. 10 Reconozco que existen respetables posiciones ideológicas en el terreno propiamente político, en la militancia sindical o en las franjas antiglobalización, que recusan de plano la posibilidad de que los bienes públicos sean mercantilizados. Claramente esta es la coherente postura de quienes adhieren al marxismo en sus versiones más clásicas. Estos últimos saben bien que la propuesta de Marx se encaminaba a la eliminación del mercado como institución social en donde tiene lugar la compra y venta de las fuerzas del trabajo: el régimen del asalariado. Por ello, en el largo entretanto de la vigencia de las sociedades capitalistas, pugnan por arrancar como una reivindicación temporaria el que los bienes públicos sean ofertados directamente por el Estado. Esto implica, en el campo de las políticas públicas, dos tipos de problemas. El primero: que se tendería a hacer de la provisión de los bienes públicos un monopolio estatal. De tal suerte que la educación, la salud, la justicia, la seguridad, etc., estarían bajo esa condición de un único proveedor. El segundo conflicto deriva de la condición de gratuidad o de provisión directa de estos, en el entendido de que se trata de derechos, no de relaciones contractuales de algún tipo. Sin embargo, estos alegatos dejan de lado la cuestión de los costos de la provisión, y el que muchos miembros de la sociedad civil tienen la capacidad de asegurar numerosos bienes colectivos 48 por medio de su pago. O sea que preferirían –como lo hacen en efecto– su mercantilización. Aquí entra un principio central en las políticas, cual es el de la subsidiaridad. El que el Estado deba ofrecer educación universalizada no implica que deba prohibir su provisión por empresarios privados; lo mismo vale en el campo de la salud, la seguridad, las pensiones, la recreación, etc. 11 Esta afirmación la hago en virtud de que considero que los consensos no son ni a priori teóricos ni simples ficciones jurídicas. La idea del consenso traslapado es fundacional, pero no en un sentido contractualista “clásico” como pasaje de algún “Estado de naturaleza” a la vida política organizada. Éste revela, más bien, los equilibrios sociales, las hegemonías legitimadas y las reglas de competencia en el orden social y en la apropiación de los bienes públicos y privados. En cierta medida, corresponde a una instancia de facticidad del sistema político mismo por parte de los grupos sociales que se relacionan inclusivamente con éste. Y es la base del modelo actualmente vigente de derecho internacional. Es decir, de la conceptualización de soberanía. 12 En estos tiempos de obsesión antiterrorista ello se ha acentuado. Un gobierno neoliberal como el de Álvaro Uribe Vélez en Colombia, no encuentra contradicción entre su prédica aperturista y mercantilista, y el incremento sustancial del gasto militar y en los rubros afines de “seguridad”, ni en el propósito de intervenir la vida privada, desvirtuando los principios tutelares de la presunción de inocencia (recuérdese el famoso caso de la inversión de la prueba en los delitos vinculados al narcotráfico) con la reedición de los estatutos “antiterroristas” que socavan las libertades individuales. También se evidencia tal conducta en la violación del derecho internacional para responder al secuestro terrorista con el secuestro “justo”, en términos de la más pura ley del Talión (me refiero al episodio de captura y secuestro ilegal de un dirigente de la rama internacional de las FARC en Caracas a comienzos de diciembre de 2004 revelado con todos sus episodios criminales: soborno ilegal a funcionarios del Estado venezolano, ocultamiento y manipulación de los hechos, atizamiento del choque diplomático por los EE. UU., etc.). 13 Debo también en este tema hacer la salvedad de que un abordaje en profundidad lo he hecho en mi libro La Soberanía Transformada (2006a). Remito al lector, en especial, a los capítulos uno y tres de dicho texto. Dado que no se trata de repetir lo ya escrito, aquí abordo otros nuevos aspectos del problema específicamente conectados con la dinámica público-privado. Para una relación de estudios concretos donde aplico las conceptualizaciones desarrolladas, véase además mi libro La Mercantilización de lo Público, (2005). Un buen tratamiento sistemático de la naturaleza de las políticas públicas se encuentra en A. Noel Roth (2003; 2002); Carlos Salazar Vargas (1999) y Alejo Vargas (2000), para citar sólo libros publicados en Colombia. En cuanto a revistas en español que hacen un seguimiento continuo del problema de las políticas públicas, cito, entre otras: Instituciones y Desarrollo (Barcelona, IIGOB), Reforma y Democracia (Caracas: CLAD), Gestión y Políticas Públicas (México: CIDE), Nueva Sociedad (Caracas) y la Revista de la CEPAL (Santiago de Chile). 49 14 Téngase en cuenta que los conglomerados económicos, más que obtener ganancias suplementarias —que las logran al controlar la prensa—, manejan e instrumentalizan la voluntad ciudadana, la llamada opinión pública (cfr. Una ampliación de este punto en el capítulo tercero). 50 Capítulo 2 GENEALOGÍA MODERNA DE LA FUNCIÓN PÚBLICA En este capítulo me propongo abordar parcialmente la genealogía de la moderna concepción de la burocracia estatal y de la función pública. Ciertamente, debo aclarar que la pretensión de abordar este tema en su conjunto sobrepasa los límites de este libro, por lo cual pretendo centrarme en la relación existente entre las instancias pública y privada en la vida social e institucional. En este sentido busco —precisa y delimitadamente— hacer una lectura interpretativa de las más importantes teorías filosóficas que en la Modernidad se han referido fundacionalmente a la cuestión de la naturaleza de la función pública. A pesar de la diversidad y la riqueza conceptual que podemos hallar en la reflexión política y teórica de la fase de la llamada “Modernidad”1 abarcándola en su totalidad, me limitaré solamente a realizar un análisis que tiene que ver con la formulación de una teoría política que parte del sujeto —del individuo tomado en cuanto consciencia singular— para establecer su articulación con la llamada sociedad civil o política, introduciendo en la moderna reflexión del derecho la relación entre la libertad del individuo y la seguridad colectiva, problemática que subyace en la reflexión en torno al proceso de formación de la sociedad política y el paso del llamado estado natural o de naturaleza a la comunidad o sociedad política. Para ello haré referencia, en primer término, a la génesis del sentido 51 de legitimación de la esfera política a cargo de los filósofos jusnaturalistas modernos: Hobbes, Locke, Spinoza y Kant. De este modo la temática señalada puede ser vista en correspondencia con la discusión sostenida por el pensamiento filosófico-político ilustrado del siglo XVIII acerca de la relación y contraposición entre el interés particular como ámbito por excelencia de la privaticidad, de la esfera libre y volitiva de los ciudadanos, y el interés general, ámbito de lo universal, lo público y lo estatal. Este proceso deliberativo tuvo como máximas expresiones los elementos conceptuales de los “philosophes” ilustrados franceses y su síntesis primera en la filosofía alemana, en los estudios de Kant y Fichte. Desde luego esta convergencia conceptual no se dio de forma global, y para muchos ilustrados aún la concepción de libertad política era aquella que un contemporáneo de Hegel, el pensador francés Henry Benjamín Constant, denominaría “libertad de los antiguos”, correspondiente a una tipología que podríamos quizás denominar cívico-republicana (U. Cerroni, 1972). En esa perspectiva la privaticidad —como se vería en el caso del propio Rousseau— no implicaba un desconocimiento de los ámbitos de la publicidad; es decir, del deber patriótico de servicio al Estado o comunidad de asociados. En cierta medida, puede afirmarse que la perspectiva rousseauniana limitaba o subordinaba estas esferas de privaticidad respecto a los intereses generales del Estado (valga decir de la comunidad de los asociados) y de su pulsión soberana. Su pensamiento político privilegió ciertamente el contexto del republicanismo por encima del individualismo. Posteriormente, anudaré dicha temática con el debate surgido en la filosofía política en torno de la conceptualización hegeliana de la relación entre Estado y sociedad civil. Así, más adelante, estaremos en condiciones de evaluar las contribuciones de esta polémica entre Hegel y Marx respecto de la teorización célebre y paradigmática sobre la burocracia, formulada por Max Weber (I. Fetscher, 1965). Ello nos permitirá cotejar el modelo teórico hegeliano con la conocida crítica juvenil de Marx y hacer una evaluación retrospectiva de este debate, teniendo como referente global la temática de la función pública, tal como podemos hoy evaluarla en la búsqueda de sus correlatos en la teoría dialéctica de la sociedad a partir de los señalados basamentos de estirpe he52 geliana. Así, desde estos supuestos o referentes, indagaremos la manera como vislumbra la tradición filosófica propia a la Modernidad, a partir de una lectura interpretativa del legado hegeliano, la relación entre lo privado y lo público dentro del marco más amplio de la relación o articulación entre los intereses particulares y el interés universal. Posteriormente, concluiré este capítulo con algunas referencias al estado posterior de la cuestión: en el caso del marxismo, reflexionando sobre el impacto, en su vigencia argumentativa, del desplome del socialismo de corte soviético casi en el mundo entero en las dos últimas décadas del siglo pasado. En el caso de Weber, discutiré brevemente sobre la actualidad de su paradigma para las Ciencias del Management y el campo de las políticas públicas. 1. Un discurso y una praxis individualista, secular y nacionalista La polémica sobre el lugar del interés público tuvo, en los siglos XVII y XVIII, una determinación de corte naturalista, o más exactamente, jusnaturalista, que pretendió objetivar la relación entre individuo, comunidad y Estado a partir del discurso de la ciencia moderna. Se estableció, en efecto, una estrecha relación entre el auge de las ciencias experimentales, el desarrollo de las fuerzas productivas y el contenido del trabajo filosófico de la época, en la que esta última fue jalonada por el extraordinario avance de las primeras2. El examen que de la sociedad burguesa hizo la filosofía política de los siglos XVII y XVIII, muestra cómo el método utilizado por ella (es decir, la mencionada filosofía política) fue tomado prestado de las Ciencias Naturales y de la filosofía experimental. Dicha época vivió el influjo de la llamada Revolución Copernicana (Kant). Se trató de la exaltación del método físico-matemático como instrumento para alcanzar el conocimiento no sólo de la naturaleza, sino también del hombre y la sociedad. Así, el Estado pudo ser comprendido a través de la recta aplicación de dicho método concibiendo, en consecuencia, las relaciones entre los hombres como relaciones de fuerza de orden natural, en las que intervienen la acción y la reacción. Toda pasión humana fue entendida como un movimiento mecánico que termina o comienza. Intervino, en consecuencia, el principio de causalidad como factor explicativo por excelencia y, 53 con base en éste, se captaron las determinaciones que enlazaban necesariamente la acción libre y las imposiciones propias a la naturaleza humana. Pero no se crea que existió una sola corriente de pensamiento en cuanto a la elaboración de una teoría política moderna. Desde el punto de vista que estamos considerando, existieron dos líneas disímiles. De un lado, aquellos que recogieron la herencia renacentista e insistieron en que el conocimiento habría de dirigir su atención a los problemas de la sociedad en la medida en que ésta era, en un sentido general, objeto de su propia creación; corriente que sintetiza la famosa expresión de Vico según la cual el hombre comprende la Historia porque ésta es su propia creación3. Con posterioridad al período de auge del feudalismo, teniendo como precursor al movimiento renacentista, se reivindicó el lugar central del hombre en la Historia, que lo constituyó en sujeto preeminente de la expresión estética. Tal fenómeno produjo en Europa una revolución en el campo ideológico de la cual fue punto de avanzada el singular desarrollo de las modernas Ciencias Naturales y una filosofía en estrecho vínculo con los hitos que la primera iba marcando. Hombres como Keppler, Galileo, Bacon, Leibniz, Newton, Descartes, el mismo Kant, etc., descollaron en el campo de la investigación científica, del estudio de la naturaleza y en el de la elaboración de una nueva lógica de la investigación científica del mundo. Este proceso convertiría a las ciencias en un poderoso auxiliar de la revolución terrenal que comenzaba a promover una nueva vida cultural urbana y que terminaría por destruir las fortificaciones acumuladas por el feudalismo a través de varias centurias. La vida rural teñida de romanticismo cedería el paso a una nueva épica realista que efectuaba la puesta en escena del proceso de formación de la industria manufacturera y de la expansión incontenible de la economía mercantil. Como se sabe, dichas fuerzas, en el plano político, librarían su batalla decisiva en las gestas que coronaron victoriosamente la Revolución Francesa. Desde el punto de vista filosófico, la Modernidad surgió a partir de la ruptura de las relaciones estamentarias y gregarias del orden señorial, apareciendo en su lugar el individualismo. Esta es una de las improntas características del mundo moderno: reconocer la autonomía moral y racional del sujeto. Esta ruptura entre el mundo medieval y el mundo moderno, el surgimiento del individualismo, 54 reconocido y garantizado por la teoría política en las diversas variantes del contractualismo, John Locke la definió taxativamente: “Si en el estado de naturaleza la libertad de un hombre es tan grande como hemos dicho, si él es señor absoluto de su propia persona y de sus posesiones en igual medida que puede serlo el más poderoso, y si no es súbdito de nadie, ¿por qué decide mermar su libertad?, ¿por qué renuncia a su imperio y se somete al dominio y al control de otro poder?” (J. Locke, 1990). Pregunta que ya se había hecho Hobbes. ¿Por qué el hombre acepta cambiar una libertad natural que le es inherente, a cambio de una libertad política? En la formulación del problema en Hobbes y Locke, lo político no pertenece al ámbito de lo natural; por el contrario, resulta un fruto de la voluntad, a través de un acuerdo contractual en el que los individuos pactan la reconversión de su libertad natural en libertad política mediante la conformación de una asociación artificial. En la perspectiva de Locke, se daría este paso no para garantizar el bien común, sino las libertades de los individuos; y que aquello entregado, en términos de derecho natural, pueda ser recuperado como derecho político. El liberalismo político, desde Locke, afirma el interés general como un fin societario sólo sobre la base de explicar que dicho interés garantizaría a cada individuo ejercer libremente sus derechos. Esta es la característica fundadora del liberalismo político4. En este sentido y a manera de contrapunto de la anterior definición sobre el origen del liberalismo político, en la historia de las doctrinas políticas ha sido un lugar común la tesis según la cual Hobbes es el inspirador del moderno totalitarismo. Su teoría política justificaría la obediencia ciega frente al poder establecido. El ciudadano debe entregar su libertad al Estado a cambio de su seguridad. El derecho a la rebelión está proscrito como el peor de los males que pueden afectar a la comunidad en cuanto mina las bases mismas de su existencia, etc. Como lo hemos visto, se designa con el vocablo totalitarismo la prédica a favor del régimen que recorta y limita al máximo los derechos del individuo frente al Estado policía, que lo sujeta al arbitrio del despotismo estatal, de tal modo que el blindaje de dicho Estado aparece recubierto con la sangre de 55 sus súbditos (subjects). Hobbes es acusado como propulsor de todos los abusos del Estado frente a los individuos. ¿Qué tanto de verdad hay en esta afirmación que por fuerza de repetición pasa por verdad inamovible? Hobbes fue efectivamente un hombre de partido e intervino activamente en la vida pública de su tiempo. Lejos estuvo su pensamiento de la idea de defender el derecho de las masas a la rebelión. Para Hobbes, un asunto clave en el terreno político era poder consolidar el Estado, por lo cual éste debería ser fuertemente centralista, pues de su supervivencia dependería la cohesión del cuerpo social. Este Estado impone el orden social sobre la base del ejercicio de la coacción y del terror. Hobbes, por ejemplo, explica el papel que cumple la religión en la medida en que ésta sirve como instrumento para dominar ideológicamente a las masas: el miedo a las fuerzas invisibles —señala Hobbes en “El Leviatán”— inventadas por el espíritu o imaginadas sobre la base de mentiras autorizadas por el Estado se llama religión (T. Hobbes, 1962:169). De tal modo que la religión y la superstición son simplemente instrumentos ideológicos de coerción sobre la masa, juegan el papel de freno de los estallidos sociales e impiden el desmoronamiento del Estado. El fundamento del análisis hobbesiano consiste en su comprensión de los deseos como fuerzas motrices de la voluntad humana. Existe una identificación entre poder y libertad, en la cual esta última estriba en la posibilidad —y la realización de la misma— de satisfacer los deseos. El hombre es libre en tanto pueda realizar sus deseos. “No lo es cuando sus deseos no obtienen una respuesta que los complazca”5. Como se ve, el fundamento de los actos, de la voluntad de los individuos es material. Se parte de la sensibilidad; el libre despliegue de las voluntades hace que choquen y pugnen en feroz competencia por satisfacer sus respectivos deseos. Así, la libertad y el poder de un individuo sólo pueden cristalizarse subsumiendo, anulando, la libertad y el poder de los demás. Sin embargo, para Hobbes, la comunidad no es un asunto hipotético; no parte de una consideración puramente especulativa sino, al contrario, de una realidad objetiva. La pregunta, nos dice Hobbes, no es por tanto, si es posible la vida social comunitaria, sino ¿cómo ha sido posible la constitución de la sociedad política?, ¿cuáles son sus fundamentos constitutivos, en qué elementos reposa la legalidad de su existencia? De este modo, la filosofía política inglesa colocó 56 como problema central de la temática política la cuestión del contrato, entendido como un acuerdo de las dispersas voluntades individuales para garantizar la supervivencia de la comunidad. La expresión “contrato social” no fue exclusiva de Hobbes ni de la filosofía política contemporánea; se remonta a los jurisconsultos del siglo XVI; principalmente, al teórico protestante alemán Johannes Althusius (1557-1638) y también a Grotius y Pufendorf, entre otros. Sin embargo, en esta tradición teórica, la teoría contractual aparece inficionada por concepciones medievales. Así, por ejemplo, en el caso de Althusius, el contrato no es pactado entre individuos libres sino que es un pacto entre distintas comunidades federadas. El gran mérito de Hobbes, en este punto, consiste en intentar deducir el edificio político de la conciencia singular, es decir, en colocar como fundamento constitutivo de la sociedad política al ciudadano. El punto de partida de Locke es semejante al de Hobbes. Su diferencia radica en que el primero intenta defender los derechos irrestrictos del individuo concebido en oposición al poder del Estado. El fin de dicha comunidad —que se expresa a través del poder legislativo— es, según Locke, el bien común, que consiste, ante todo, en la defensa y salvaguarda de la propiedad privada. Locke no le da, a diferencia de Hobbes, un papel singular, dominante, al Estado, sino que —al contrario— le asigna como su finalidad expresa la labor de garantizar la supervivencia y el respeto del derecho de propiedad. En el caso de Locke la función del Estado se reduce simplemente a garantizar la seguridad y la justicia. Se trata, para él, de preservar la libertad personal y la propiedad individual. El mejor de los Estados es aquel que da al individuo el máximo de libertad posible; el contrato puede, en consecuencia, asumir las más variadas formas permaneciendo los individuos con la potestad de decidir cuáles de sus derechos concederán al Estado6. Como puede observarse, Locke efectúa una transacción entre los intereses de la burguesía ascensional y la aristocracia de la tierra (Locke, en contraposición a Hobbes, escribe resumiendo desde su punto de vista los resultados inmediatos de la revolución inglesa de 1688). El derecho de propiedad en ese momento fue fundamentalmente el derecho de propiedad sobre la tierra, principal medio productivo, eje de la economía de Europa durante el siglo XVII. De otro 57 lado, sin embargo, Locke realizó un progreso importante en el desbrozamiento de una nueva teoría jurídica al concebir la fuerza de trabajo como una mercancía y definirla en función de la propiedad privada, de tal forma que, a la par con la propiedad sobre los instrumentos de producción, ha de reconocerse al trabajo como un objeto de intercambio7 mercantil y, por lo tanto —en el plano teórico— como generador de derecho (J. Locke, 1990). Pero, si Locke edifica una concepción individualista, Hobbes realiza un trabajo encaminado a un propósito diferente. Para Hobbes, el contrato entre las diversas individualidades sólo puede funcionar eficazmente sobre la base de que los miembros de la sociedad —sin excepción alguna— entreguen sus derechos al Estado, recibiendo en cambio la garantía de su seguridad, la mutua transferencia de derechos, es lo que los hombres denominan contrato (T. Hobbes, 1961). Cuando el contrato implica una cesión futura, una promesa, se denomina pacto —nos sigue diciendo Hobbes— y su incumplimiento se denomina violación de fe (violation of faith). Un contrato, para ser realmente válido, ha de ser expresado en palabras que obliguen no sólo con un tiempo, sino con el pasado, el presente y el futuro (T. Hobbes, 1961: 192). Este contrato entre personas puede ser de dos tipos en la medida en que existen, de un lado, personas naturales, y, del otro, personas artificiales. Las primeras, son creaciones de la naturaleza, y las otras, criaturas surgidas como producto de la acción social del hombre. Es importante destacar aquí cómo Hobbes introduce el concepto de persona, a renglón seguido de su explicación acerca de las características constitutivas de un contrato, para señalar cuál tipo de contrato puede garantizar la cohesión del cuerpo social. Un contrato que le entregue a un tercero los derechos originalmente correspondientes a las dos partes naturales, siendo este tercero una persona artificial; el soberano, que Hobbes identifica con la monarquía absoluta, posee —por el consentimiento de sus súbditos— el poder de determinar la guerra exterior y la paz, de definir lo que es justo e injusto, de dirimir las querellas internas, dictar las leyes, etc., sin estar atado a ninguna limitación —ni siquiera a la ley natural está sujeto el soberano—. Frente al soberano, los súbditos deben observar una obediencia total, perdiendo el derecho a recusar lo que proviene de un contrato dictado, o pactado, por la mayoría de la comunidad8. Tal es el precio que Hobbes estatuye a los ciudadanos para que se pueda 58 garantizar la posibilidad de vivir en comunidad. En este sentido, el comentarista de Hobbes, C. B. MacPherson indica que la finalidad expresa de la filosofía política hobbesiana es la de garantizar la paz y la seguridad de los miembros de la sociedad. Más adelante MacPherson expone, a un nivel específico, la naturaleza del postulado de seguridad colectiva y paz que preside la concepción hobbesiana9: sobre la base de estas consideraciones, Hobbes deduce que el soberano tiene un derecho natural a su propia perpetuación, a perseverar en su ser. MacPherson, finalmente y en relación con este corolario, alude a la recusación que núcleos de las elites empresariales partidarios de que el Estado fuera deducido a partir de una asamblea representativa hacen el Leviatán, advirtiendo que Hobbes simplemente parte en su rigurosa deducción lógica de universalizar los postulados de la competencia mercantil. De allí afirma que su falla reside en: “(...) pasar por alto dentro de su modelo la fuerza centrípeta que una burguesía cohesionadora como clase cumple dentro de la sociedad. Hobbes estuvo fuertemente impresionado por la fuerza destructiva, desgarradora, de la competencia y la lucha por el poder. Colocándola como base de su modelo de sociedad política, la falla de Hobbes consistió en no vislumbrar que este modelo generaba consecuencialmente una diferenciación de las clases sociales, de la cual se podría esperar que produjese una cohesión de las clases, o al menos, de la clase que se encaminaba firmemente al control de la sociedad” (Ibíd.: 56). A pesar de que el Estado hobbesiano aparece como la coronación omnipotente de un ser artificial, es importante subrayar que este edificio colosal se basa en la afirmación primaria de los derechos y del rol singular que le corresponde a la conciencia individual burguesa. Hobbes, al definir al hombre en términos de su igualdad natural, permitió comprender la sociedad a partir del postulado que definía sus funciones en términos de relaciones naturales10. Esta temática tendría desarrollos fecundos en la tradición teórica política jusnaturalista, en especial en la obra del filósofo judío holandés Baruch Spinoza. Aunque ciertos historiadores de la filosofía reducen su papel como pensador político al de repetidor de los postulados fundamentales de Hobbes, hemos de anotar que su reflexión posee autonomía, tiene una perspectiva enriquecedora de la teoría política. A pesar de las similitudes de los puntos 59 en que Spinoza reconoció la validez de las tesis hobbesianas11, lo cierto es que —como lo señala Stuart Hampshire— llegaron a puntos comunes a partir de principios diferentes y las conclusiones de carácter político-práctico fueron, por ende, distintas. Esta divergencia de principios encuentra su fundamento en la disimilitud de sus respectivas concepciones del mundo. Así, mientras Hobbes fue mecanicista con todas las consecuencias que de esta concepción se derivan, Spinoza destacó el papel autónomo de la individualidad frente a las pasiones. El hombre no debe ser esclavo de sus pasiones. La razón puede, verdaderamente, restringir y moderar las pasiones (B. Spinoza, 1955). La libertad estriba en la realización de la posibilidad —accesible al entendimiento finito— de comprender las causas y el orden necesario del Universo. “Para Spinoza, el ejercicio de la razón no sólo implica la búsqueda de la autoconservación y la satisfacción de los deseos, sino que constituyó en sí misma el fin supremo frente al cual las demás cosas eran simples medios; la razón no fue, pues, como en Hobbes, el cálculo empírico de probabilidades sino la reconstrucción, por medio del razonamiento lógico, del orden necesario del Universo. El criterio por el cual debe ser juzgada una organización o sistema político es si impide o hace posible el amor racional a la libertad humana y a la comprensión de la naturaleza” (S. Hampshire, 1962: 182). De este modo, la fórmula para alcanzar la libertad se realizaba por mediación del conocimiento: la libertad es la comprensión de la necesidad. Así, las líneas comunes en estos dos filósofos se separan al considerar el carácter más complejo y elaborado del método spinozista. El materialismo de Hobbes dio paso a un complejo sistema filosófico que establecía la libertad sobre el suelo de la razón con el fin de reconciliar al entendimiento finito con el infinito a través de la apelación al saber, a la ciencia; y, por este medio, brindarle al hombre su verdadera libertad, cumpliendo así la filosofía su autentico propósito. Su rechazo a los argumentos morales para juzgar las acciones de los ciudadanos en los Estados, la relación del Estado con éstos (como relación del universal hacia el singular) y de los diversos Estados entre sí (relaciones internacionales), correspondió a la firme convicción de que las leyes de la naturaleza obraban de igual forma en las relaciones sociales. Los problemas de la política podrían estudiarse de manera 60 científica comprendiendo que la cuestión esencial de la política es el poder12. Spinoza no introdujo una dicotomía de tipo cartesiano al establecer la relación entre deseos y razón; no partió de considerar los primeros como expresiones impropias del alma, como concesiones indebidas de ésta al cuerpo. Por el contrario, partió de señalar su unidad primaria en la medida en que los concibió como productos o, mejor aún, como expresiones del infinito poder de la naturaleza. Con esta precisión introdujo Spinoza su explicación acerca del carácter específico de los deseos producidos por la razón y, de otro lado, de los deseos que él ubicó como producidos por las afecciones del hombre, afecciones pasivas que son un padecimiento más que una acción propiamente dicha; y sólo con base en estas premisas, pensó el papel de la política. Así, su desdén frente a los brotes anárquicos en el seno del Estado obedeció a razones muy distintas a las de Hobbes. El campo de la política no sería, por tanto, el terreno fértil en las disensiones e intrigas cortesanas; será el campo de despliegue de las fuerzas de la razón, ya que el móvil del hombre al vivir en comunidad es la obtención de la paz y la seguridad, y salvaguardar la libertad civil; libertad que, como vimos, no expresa algo distinto a la naturaleza del hombre sino, por el contrario, la unidad del entendimiento finito con el correspondiente conocimiento (conciencia singular del mismo) de las causas y el orden necesario del Universo. Existe, en Spinoza, una creencia en el poderío incontenible de la razón y en el papel salvador que ésta juega en el manejo de las relaciones entre los hombres. Spinoza se refiere a la posibilidad de racionalizar la política asumiendo la razón, el manejo y el control de las pasiones. No se trata de negar el peso del elemento pasional en la acción política sino de establecer el primado de la conciencia racional del hombre. En este orden de ideas podemos comprender la gran importancia que la obra política de Spinoza tuvo13. 2. Las premisas rousseaunianas del interés público Rousseau en su ensayo “Sobre si el restablecimiento de las ciencias y las artes ha contribuido a depurar las costumbres”, escrito en 1750, caracterizó el paradigma del romanticismo al sostener que la moralidad, antes que significar 61 un paso en el desarrollo humano, fue una expresión de decadencia. La sociedad civil, la urbanidad, el artificio, etc., han sido generalmente elementos de la decadencia humana. De allí su acre crítica a d’Alambert, puesto que el teatro, la vida de las elites, la cortesía mundana y otras formas de sociabilidad artificiosa, son síntomas de debilidad, incluso del afeminamiento de la sociedad moderna. “Europa había recaído en la barbarie de las primeras épocas, los pueblos de esta parte del mundo, hoy en día tan esclarecidos, vivían hace pocos años en la barbarie por la decadencia del Imperio Romano. Era necesario una revolución para retornarle el sentido común y ella vino del lado de donde menos se la esperaba. Fue la estupidez musulmana la que permitió que el eterno flujo de las letras renaciera entre nosotros” (Ibíd.). Esta frase refiere que gracias a los escritores árabes los europeos logran recuperar la tradición griega. Después de que los germanos invaden Italia y el Imperio Romano se disuelve, las obras de los filósofos romanos desaparecen de Europa. No se tienen traducciones, desaparecen Aristóteles, Platón, los clásicos griegos, en un fenómeno impresionante de retroceso cultural. Fueron Averroes y otros intelectuales musulmanes los que trajeron el álgebra, a Platón, Plotino y, en general, a los escritores griegos, a los poetas Píndaro, Homero, etc. Toda esta tradición que se había perdido, se recuperaría posteriormente. La caída del trono de Constantino implicó en Italia el olvido de la vieja Grecia. Después, cuando llegaron los musulmanes, las ciencias sobreviven junto con las artes. Rousseau sintetiza este proceso: “Aparece de nuevo el arte de escribir y el arte de pensar. Se empieza a sentir el principal avance en el comercio de las musas, aquel que hace a los hombres más sociables, inspirándoles el deseo de satisfacerse los unos a los otros en obras dignas de aprobación mutua... Es un grande y bello espectáculo ver al hombre surgir de la nada por sus propias fuerzas. Anticipar por las luces de su razón las tinieblas en las cuales la naturaleza lo había envuelto, elevarse por encima de sí mismo, desarrollar su espíritu hasta las regiones celestes, recorrer a paso de gigante el sol, la vasta extensión del 62 Universo, y lo que es más importante y difícil entrar en sí, estudiar al hombre y conocer su naturaleza, sus deberes y sus finalidades. Todas estas maravillas se han ampliado con base en pocas generaciones” (Ibíd.). Rousseau explica que luego de la caída del Imperio Romano empieza de nuevo un proceso ascendente en Europa, que califica con la expresión politesse. La cual significa buenas maneras, cortesía, educación, señalando cómo el mundo europeo, desde la mitad de la Edad Media hasta los comienzos de la Modernidad, construyó un conjunto de costumbres y buenas maneras para el desarrollo de las artes y luego esto se reforzaría con el surgimiento de la ciencia. Pero paradójicamente esta característica le parece a Rousseau negativa, porque la politesse debilita al hombre moderno. Por el contrario Rousseau le canta al hombre antiguo, símbolo de virilidad y fortaleza14. Incluso sostiene que el vestuario característico del siglo XVIII, en el que los aristócratas usaban pelucas, zapatillas y medias, era ridículo. Le parecía un proceso de desnaturalización de las potencialidades humanas. Toda la erudición de las elites, en el sentido del desarrollo del arte culto, del teatro, el urbanismo, las buenas maneras, la etiqueta; todo ello implicaba su decadencia. Esta cortesía oculta el engaño: “Los hombres que forman esta multitud que se llama sociedad, situados en las mismas circunstancias hacen siempre las mismas cosas, así los motivos no sean tan importantes. No se sabrá jamás qué es lo que se hace. Es necesario para conocer a los amigos atender las grandes ocasiones, puesto que es en estas grandes ocasiones donde el hombre busca y expresa lo esencial de su conocimiento” (Ibíd.). El proceso de urbanismo, de politesse, implica, a través de la cultura, las artes y la ciencia, la decadencia del hombre en lo atinente a su fuerza y potencialidades naturales. Rousseau, en contraste, formula un elogio de los egipcios, exalta la civilización de Constantinopla y, en general, dibuja un cuadro positivo de las civilizaciones antiguas calificándolas de potentes. En el caso de Grecia, por ejemplo, resulta sintomático cómo elogia a los espartanos y critica a los atenienses: 63 “Atenas devino el refugio de la politesse y el buen gusto, se convirtió en el país de los oradores y los filósofos, de la elegancia en el vestuario, lo que correspondió igualmente en el uso del lenguaje. Se veía por todas partes el mármol y el dinero animaba las manos de los maestros más hábiles. Es de Atenas que surgieron las obras sorprendentes que han servido de modelo para todas las edades corruptas.” (Ibíd.) El análisis rousseauniano no expresa (aparentemente podríamos suponerlo así en tanto vislumbramos a Rousseau como un elemento clave de la Ilustración) que la Ilustración se vea a sí misma como el coronamiento progresivo de unas fases históricas escalonadas, que no se niegan sino que se consideran necesarias. Como lo hemos visto, Rousseau asumió una visión romántica de la Edad Moderna como una fase decadente, y por ello su programa afirma que la verdadera Ilustración debe permitir la reconciliación del hombre con su ser natural. Este es el elemento clave de su mensaje. La sociedad civil impone normas, constriñe la libertad del individuo, libertad natural propia del hombre en su estado primitivo. No se trata de que el hombre primitivo fuera bueno o malo porque dicha calificación es moral y, por lo tanto, un problema propio de la civilidad. El hombre en el estado primitivo, simplemente buscaba la satisfacción de sus necesidades. En el instinto no hay bondad o maldad, son los fines mismos de la naturaleza, pero la sociedad corrompe al hombre. Se supone en esta pedagogía que lo artificial, lo socialmente impuesto, implica una deformación de las características constitutivas del hombre premoral del estado de naturaleza. Cabe destacar cómo Rousseau no fue muy coherente en su pensamiento político con respecto al ideal moral romántico, puesto que en el Contrato Social modificó parcialmente su argumentación. Allí afirma que el hombre en el estado salvaje corría desnudo por los bosques. Esa desnudez no es elogiada por Rousseau, ya que la considera un estado de animalidad y piensa positivamente el pacto social como instrumento para transformar la libertad natural en libertad política (J. J. Rousseau, 1971). Según E. Cassirer, a pesar de su vivo interés por los problemas políticos, la Ilustración no elaboró una nueva filosofía política15. Pero esta afirmación es discutible. Una parte de los analistas considera que la teoría del Contrato Social sí 64 es nueva respecto de los contratos anteriores que se remontan hasta los filósofos católicos como Suárez16; la versión hobbesiana, donde el contrato social es el acuerdo de todos para delegarle los poderes al Leviathan; o Locke, donde en cierta medida, de la soberanía emana un poder gubernativo, poder que se puede revocar, democrático, legitimado por las mayorías del gobierno. En fin, estos modelos difieren del rousseauniano debido a su cláusula de que cada quien entregando su libertad natural recibe a cambio una cantidad proporcional o equivalente de libertad política. De tal modo que contratando entre sí todos los hombres, es como si se contratara consigo mismo. Ese contrato no pretende entregarle el poder a un tercero en discordia, a un árbitro supremo autoritario, ni le cede el poder a unos gobernantes que lo ejercen a través de la delegación. Para Rousseau, la voluntad popular, que llama general, no es delegable y, por lo tanto, cuando se constituye un contrato, la soberanía se mantiene indivisible. Este es un contrato de libertades que, al traducir en términos equivalentes lo natural en lo civil político, permitirá el explayamiento de la virtud cívica que recuperaría las características primigenias constitutivas del hombre, antes de su decadencia moral; virtud cívica republicana que se expresa en la voluntad general, en el cumplimiento de los deberes colectivos, en el amor a la patria17. Ésta constituye el contexto en el cual se desarrolla el contrato social, se cristaliza la transformación de la libertad natural en libertad política como ámbito de la soberanía popular; es el lugar en el cual la virtud cívica puede recuperar los elementos constitutivos propios al estado de naturaleza. A Rousseau corresponde un papel determinante y fundamental en la formulación de una consecuente y radical filosofía política democrática durante el período comprendido entre los siglos XVI y XVIII. Rousseau produce una ruptura profunda en relación con sus antecesores, significación que adquiere toda su nueva dimensión en su planteamiento sobre el Pacto Social que, como hemos de verlo a continuación, introduce la reflexión política en la problemática propia de la Modernidad. Al igual que sus predecesores, Rousseau inicia su trabajo de reflexión sobre lo político preguntándose por los fundamentos constitutivos de la sociedad civil. Esta línea de pensamiento podemos descubrirla en su texto de 1755: Discours sur l‘inégalité parmi les hommes, en el que da pie a una primera versión del contrato social que le servía en ese momento como 65 un simple punto de apoyo para introducir la discusión en torno a la formación de las primeras sociedades y el Estado18. De tal manera que para Rousseau no se trataba de establecer una reconstrucción histórica, sino de formular una hipótesis de orden especulativo que desde un punto de vista propio del análisis lógico-conceptual le sirviera de fundamento para explicar la necesidad de una transacción, de un acuerdo que involucrase a los miembros de la sociedad bajo las reglas comunes de un orden legal. Esta temática no es, desde luego, exclusiva de Rousseau, pero es innegable que la problemática del Contrato adquiere una nueva dimensión consistente en que éste no es un acuerdo de partes signatarias para ceder el poder y los derechos a un tercero, como ocurre en Hobbes, sino que es un nuevo tipo de contrato en el cual, alienando los ciudadanos sus derechos naturales a la colectividad, no los alienan a ningún miembro en particular. Los contratos anteriores equivalían a la cesión de los derechos de los ciudadanos a un príncipe, o bien, suponían una igualdad de contrarios entre el monarca y su pueblo (suposición en la que subyace el criterio feudal de que el monarca es el propietario de su pueblo —recordemos de pasada que esta formulación es reelaborada luego por Hegel al señalar que la verdad del pueblo reside en el príncipe—. Por lo tanto, allí, en términos jurídicos fórmales, se trataba siempre de un pacto entre personas iguales, pero en la realidad se justificaba de esta manera teóricamente la subordinación del pueblo, de la comunidad, al soberano. Rousseau no permite la legalidad de tal concesión. La única alineación posible es la que establece la comunidad consigo misma19. Rousseau intenta, a partir del establecimiento del contrato social, superar la antinomia que encierra el paso desde la libertad natural y el estado de naturaleza originario hasta la fase superior, propia de la civilización, en donde se encuentran la libertad civil y el derecho igualitario (diferente por cierto de la simple posesión por la fuerza, propia de la barbarie). Así se produce un abandono del derecho primitivo y la libertad natural en aras de una igualdad y un derecho político, frutos de un acuerdo contractual garante de la legalidad de la sociedad civil. Por lo cual el soberano que surge en este contrato, a diferencia de Hobbes y Locke, está integrado por la totalidad de los contratantes y no es un príncipe, un monarca o un gobierno delegatario. 66 En su perspectiva, dicho soberano recoge las voluntades dispersas de sus miembros a través de una figura teórica de gran importancia: la voluntad general, que representa el acuerdo de los diversos intereses particulares en un haz con una dirección que señala el bienestar de la colectividad20. Al producirse el pacto social, los miembros de la comunidad alienan sus derechos y libertades naturales y reciben a cambio derechos y libertades civiles. Pero la soberanía y su base (la voluntad general) son inalienables. A tal conclusión llega Rousseau en el famoso apartado VI del libro I del Contrato Social. Cada uno de nosotros coloca en común su persona y toda su potencia bajo la suprema dirección de la voluntad general: y cada uno de nosotros recibe en el cuerpo social a cada miembro como parte indivisible del todo (Ibíd.: 68). Las dificultades que entraña esta superación aparente del estado natural, el tránsito hasta la sociedad política y el correlato existente entre el interés particular (que surge a partir de la aparición del “amor propio” como transubstanciación del primitivo y original “amor a sí mismo”) y la voluntad general (que no equivale a la simple sumatoria de los intereses particulares de sus miembros), son aspectos que han motivado críticas respecto al carácter ahistórico, especulativo y utópico de dicha sociedad política, surgida sobre la base del contrato originario. Para allanar el camino hacia la solución de estos interrogantes conviene aclarar las relaciones que Rousseau establece entre el poder ejecutivo y el legislativo21. En este sentido, vale precisar que Rousseau supone una completa subordinación del poder ejecutivo respecto al legislativo. Este último implica una mediación entre el soberano (en cuanto potencia) y la acción a través de la cual se produce la ley. Así, aparece la ley como un mecanismo normativo de la sociedad civil con base en la voluntad general22. Surge finalmente en este punto, una pregunta que no corresponde sólo a la problemática que introduce Rousseau, y es si el pueblo puede gobernar sin equivocarse. El fundamento popular del gobierno es una constante teórica, una petición de principio en la moderna teoría política. Al mismo tiempo, el respeto y la obediencia a la ley, en tanto expresión de la voluntad general, es también parte sustancial de la teoría política moderna. Rousseau señaló que 67 el propósito de la reflexión política consistía en encontrar una forma de gobierno que coloque a la ley por encima del hombre;23 es decir, la ley, más allá del influjo de la voluntad particular y del interés egoísta de los ciudadanos considerados en su singularidad inmediata. Sin embargo, la aplicación de la ley es una prerrogativa en manos de una capa especializada de funcionarios. El hilo de la legitimidad de la representación de la voluntad general por parte de estos funcionarios es el problema que toca demostrar. Y esto no corresponde ya a la simple órbita de la reflexión teórica sino que —al contrario— corresponde a la esfera de la práctica política. En este sentido no podemos olvidar que la teoría política no puede sustituir la política misma. 3. El planteamiento kantiano sobre lo público y lo privado En 1784 Immanuel Kant, el célebre filósofo alemán, publicó un corto ensayo titulado “Respuesta a la pregunta: ¿Qué es la Ilustración? En este escrito, lejos de las complejidades de sus obras capitales se planteó, sin embargo, una cuestión vital que ha tenido profunda influencia en la tradición filosófica y política occidental. A la pregunta señalada, Kant dio una respuesta igualmente célebre: “La Ilustración es la salida del hombre de su condición de menor de edad de la cual él mismo es culpable. La minoría de edad es la incapacidad de servirse de su propio entendimiento sin la dirección de otro...” (E. Kant, 1784). Nótese, en primer lugar, cómo la fórmula “llegar a la mayoría de edad”, para nosotros una especie de metáfora, era en el siglo XVIII una expresión muy concreta que refutaba el patriarcalismo y el tutelaje. En efecto, hablar de la mayoría de edad implicaba romper el tutelaje patriarcal existente a la sazón tanto dentro de la familia como en las esferas comunitarias y en el Estado. En este último sentido, trataban los philosophes de la Ilustración de formular un nuevo paradigma respecto del estatuto ciudadano, consistente en considerar al individuo como alguien igual que establece una relación de tal naturaleza con el Estado, que en realidad es quien lo construye a partir de su asociación contractual. También Kant mira la cobardía y la pereza como causas por las cuales la ma68 yoría de los hombres, después de que la naturaleza los ha librado de la conducción ajena24, cuando logran la mayoría de edad, permanecen con gusto como menores edad, por lo cual es muy fácil para otros, los dominadores, erigirse en tutores: “Es tan cómodo ser menor de edad. Si tengo un libro que piensa por mí, un pastor que reemplaza mi conciencia, un médico que dictamina acerca de mi dieta y así sucesivamente, no necesitaré esforzarme. Si sólo puedo pagar no tengo necesidad de pensar, otro asumirá por mi tan fastidiosa tarea” (Ibíd.). Estas frases contundentes dibujan el filisteísmo burgués, la mediocridad de quienes pueden comprar y por ello dejan de pensar por sí mismos depositando en otros su responsabilidad, su autonomía moral. Hombres, como el típico ciudadano burgués, que a través de la relación de mercado pueden suplir su inanidad por medio del contrato de servicios. Y dado el carácter dominante y hegemónico que en la Modernidad han asumido las relaciones mercantiles, resulta altamente probable que tal conducta delegatoria, este mercenarismo en el cual se declinan la voluntad propia y la autonomía moral, termine por convertirse en generalidad, en norma común. Como lo sugería el profesor Estanislao Zuleta (Zuleta, 1989), ésta es quizás la principal razón por la cual Kant es muy crítico respecto a los alcances que pueda tener una revolución puramente política, pensando en varios casos de tipo histórico; pero fundamentalmente en el muy cercano paradigma de la Revolución Francesa. Una revolución puramente política que cambie un orden no va a significar una verdadera emancipación de la comunidad. Kant, en cambio, ve esta emancipación en lograr la mayoría de edad, o sea, la libertad de pensamiento plenamente ejercida desde la autonomía de cada razón individual. En la argumentación kantiana aparece en este lugar de su exposición una cuestión decisiva. Kant afirma como aspecto básico para concretar la libertad positiva del hombre, el que cada quien pueda hacer uso público de su razón y coloca, en tal sentido, varios ejemplos: “Oigo exclamar por doquier no razones. El oficial dice no razones. ‘¡Adiéstrate!’, dice el oficial del ejército al soldado; el consejero de finanzas, no razones: ‘paga’; el pastor de la iglesia, no razones: ‘cree’. Por todos los lados limitaciones de la libertad, ¿pero qué limitación impide la Ilustración y cuál por el contrario la fomenta? Respondo: 69 el uso público de la razón puede ser libre siempre, y es el único que puede producir la Ilustración de los hombres. El uso privado de la misma, en cambio, debe ser con frecuencia severamente limitado sin que obstaculice con ello particularmente el progreso de Ilustración” (Ob. Cit.). De este modo, para Kant, la salida de la condición de la minoría de edad y la posibilidad de ejercicio de la libertad política humana, la eliminación del despotismo y de la servidumbre personal, todo ello, residirá en la capacidad del hombre de hacer uso público de la razón. Pero al mismo tiempo afirma que existe un uso privado de ésta que debería ser limitado. Línea divisoria que implica definir lo privado y lo público. En este sentido, Kant entiende por uso público de la propia razón el que alguien hace de ella en “cuanto sabio ante la totalidad del público lector”. Un sujeto racional hace uso público de su razón cuando expresa opiniones fundadas, argumentadas, ante el resto de la comunidad. Los estoicos creían que un hombre podía ser libre aún cuando fuera enviado a una cárcel acuñando para ello un aforismo: “podrán encarcelar mi cuerpo pero no encarcelarán mi alma”. Kant cree que quien está detenido en una cárcel sólo será libre si puede comunicar lo que piensa. La capacidad de acceder a lo público implica como rasgo sine qua non comunicar la posición individual frente a los asuntos de la comunidad. Pero no comunicar cualquier cosa. Kant habla de un sabio, de un maestro; es decir que la comunidad sólo podrá reconocer liderazgos y construirlos en un ambiente democrático allí donde haya un libre ejercicio de lo que llamamos hoy opinión pública. Ciertamente, en el corto artículo de Kant aparece una inversión del sentido usual de la relación bajo el cual percibimos la conexión de lo público y lo privado25. Lo público sería, en esta visión kantiana, el espacio por excelencia donde el individuo, independientemente de las ataduras institucionales, puede argumentar y debatir; y lo privado, aquel ámbito de limitación en el que una persona está obligada a restringir su facultad de razonar públicamente. La idea que gobierna la distinción señalada entre lo público y lo privado se ancla en una restricción de la facultad razonadora por la inscripción del ciudadano en un contexto imperativo, institucional. Lo privado es allí, ante todo, la 70 privaticidad del Estado y las instituciones públicas, la inhibición del particular en tanto portavoz o representante de funciones públicas. En cierta medida, allí reside uno de los espacios que abordará posteriormente Hegel, el secreto de Estado, el cual subraya como uno de los rasgos esenciales del funcionarismo público. Secreto que se tipifica en el moderno Estado de derecho, v. gr., en la rama gubernativa en aspectos tales como la seguridad nacional, la defensa exterior, los tratados y convenios internacionales; y en el campo jurisdiccional bajo el nombre de reserva sumarial, etc. Lo público, alternativamente, sería el espacio de discusión de las ideas, argumentos, razones, etc., que en su mayoría de edad asumen los ciudadanos. Sin embargo, lo público no puede violentar, como ámbito de discusión o como opinión pública, la existencia y compulsionaridad de los deberes públicos, es decir, los deberes que a todos nos competen como ciudadanos miembros de la Comunidad de Estado. Así entendidas dichas acepciones, lo privado tendría entonces el sentido de privación con relación al uso activo y deliberativo de la razón por parte de los individuos vinculados a un oficio o un cargo de tipo corporativo, civil, religioso, estatal, etc., donde ellos hacen parte sustancial de una “máquina” pública. Metáfora kantiana reminiscente de la idea humana de la sociabilidad artificial en la que se conjuga el verdadero espacio de la “comonwealth” que nos dibuja TH. Hobbes en The Leviathan, como paradigma del absolutismo propio a la génesis del Estado-Nación eurooccidental. En la medida en que se esté atado a fines institucionales se mantendría un status de privacía deliberativa, puesto que si todos abandonaran (por sus creencias individuales) las responsabilidades institucionales que han asumido, la sociedad entraría en anarquía total. Así, la filosofía de Kant sobre lo público y lo privado frente al uso facultativo de la razón, va a establecer una especie de paradigma de lo que sería luego denominado opinión pública; que a partir del siglo XVIII será vista ya no como una relación de identificación del interés general, el bien común, los intereses societarios, sino como el ámbito de expresión pública de lo individual, con lo que podremos rastrear el elemento de liberalismo político en Kant, que lo enlaza con Locke y otros destacados pensadores preilustrados. 71 No bastaría, vistas las anteriores consideraciones, con tener el derecho del sufragio para que la democracia política posea una fuerza fundante, legitimadora. Debe darse además en la praxis social el derecho a la discusión pública, a la exposición de los programas, a la contratación de éstos, a los argumentos de los contrincantes que se disputan el favor de la opinión pública. Hay que recordar, como se acaba de referir, que Kant define el servicio civil como un deber público o, al menos, como aquel servicio que pretende garantizar la funcionalidad de los deberes públicos. El deber público implica un uso de la palabra público en un sentido tradicional, como bien común. De tal modo que, hilvanando la lógica de Kant, uno es el uso público de la razón y otro sería el interés público. El vínculo entre estas acepciones se dará en cuanto es del interés público (los fines societarios) que los individuos puedan expresar libremente sus opiniones sobre los asuntos inherentes a la suerte global de la comunidad. Tal paradigma ha sido, desde entonces, crecientemente mitificado, y es bajo su manto (al menos en los niveles del “deber ser” del Estado) que nos vemos discutiendo los asuntos públicos. La idea de la democracia participativa, tan insistentemente machacada en los últimos años por la teoría política moderna, sugiere que la participación ciudadana es una forma a través de la cual la Ilustración política de la que hablaba Kant se puede ejercer efectivamente. Se trata, en este modelo, de promocionar la existencia de ciudadanos inmiscuidos en los asuntos públicos, no inmersos en una actitud pasiva, donde, por pereza o cobardía, permitan que otros gobiernen y que tutores y déspotas, así sean benévolos, actúen y decidan en nombre de la colectividad. Ciertamente, cuando el pueblo discute no se necesita una concertación previa, puesto que la opinión más valiosa o válida gana consenso. Es este uno de los puntos cruciales a discutir, porque ninguna sociedad moderna, democrática o burguesa, puede asumir la prohibición del derecho a intervenir en los asuntos públicos, defendiendo el punto de vista de las diferentes individualidades y suponiendo su “buena fe”. Este consenso será objeto de crítica por el liberalismo posterior a Kant (particularmente por J. Stuart Mill), en tanto se asume una relación de ruptura de la justicia o la verdad, ontológicamente considerada, en favor de la preferencia de las mayorías; más exactamente, de la valoración utilitaria validada por ésta. 72 Kant había señalado, de otro lado, la tendencia humana doble y contradictoria hacia la sociabilidad y la insociabilidad. El hombre como ser genérico busca agregarse; pero, al mismo tiempo, busca apropiarse, por medio del egoísmo, de los bienes de los demás, explotar el trabajo ajeno, dominar a sus congéneres. Kant formula una tesis capital, también expresada en cierta medida por Locke y Rousseau, pero en éste de manera más precisa y elaborada: “sólo la competencia que se suscita entre las diferentes individualidades trae el progreso humano”. Aquellas sociedades en que los individuos no compiten ni asumen autónomamente su libertad de pensamiento, tienden a homogenizarse, a volverse mediocres y a ser tuteladas por tiranos o por déspotas benévolos. En el hombre existen la sociabilidad y la insociabilidad en una tensión dialéctica, en donde si la competencia se inhibe, el progreso humano se detiene. La teoría liberal —tanto el liberalismo político como el liberalismo económico— construye, en suma, sobre esta tradición un paradigma: sólo una sociedad competitiva será libre y sólo en una sociedad competitiva y libre serán posibles el progreso humano, los derechos individuales y el bien común. 4. Hegel: los fundamentos de la sociedad civil y la función de la burocracia estatal El contexto fundacional de la noción de burocracia arrancó mucho antes de la común referencia a Max Weber, en la teorización estupenda que hizo Hegel (1966) acerca de la burocracia prusiana —y por extensión acerca de toda la burocracia estatal del Occidente de su época—. Los aportes de Hegel a esta problemática típicamente moderna son numerosos a lo largo de su extensa obra, pero se encuentran primordialmente condensados en su análisis de la Filosofía del Derecho (1821). Precisamente, en torno a las tesis allí consignadas es que reacciona Marx en su estudio de juventud, que reseñaré un poco más adelante. Por lo tanto, en este apartado me ocuparé de sintetizar lo esencial, a mi juicio, de la aportación hegeliana sobre la función pública, en tanto que en los subsiguientes nos centraremos en la crítica de Marx a Hegel y en sus proposiciones alternativas. Mientras que hacia la parte final de este capítulo, intentaré hacer un balance general de la cuestión. 73 4.1. EL TRABAJO Y EL SISTEMA DE NECESIDADES El conjunto de los Ilustrados y, en especial, los filósofos alemanes entre los cuales se destacan Kant, Schelling, Hegel, etc., fueron, en general, estudiosos o, al menos, conocedores del desarrollo de la economía política, influencia que es muy grande sobre la filosofía clásica alemana. Incluso, podremos remontarnos más atrás: a partir de J. Locke hay cada vez más conexión entre el desarrollo de la teoría económica en la sociedad europea y la filosofía. Evidentemente, este proceso de apropiación filosófica de los asuntos económicos fue producto del desarrollo político de los Estados nacionales, del surgimiento de un contexto revolucionario de lucha contra el antiguo régimen y de la conformación de monarquías constitucionales que pactan el lugar de las nuevas clases y el reacomodamiento de las antiguas clases como ocurre en Inglaterra, Holanda y Alemania. En síntesis, el proceso político institucional de surgimiento de un orden nuevo, burgués, de una estructura de Estados nacionales, obliga igualmente a una reflexión teórica de la economía sobre la riqueza nacional. Es la característica de todo el despliegue de la economía política que fundamentalmente se da en Inglaterra y Francia. En particular, Hegel fue un asiduo lector de los economistas ingleses, Adam Smith y David Ricardo, de tal modo que captó la importancia de los elementos económicos desarrollados por el capitalismo como objeto de reflexión para la filosofía política. Ello es visible en temáticas hegelianas como su teoría de las clases sociales, el maquinismo derivado de las actividades industriales y la creciente importancia de la actividad práctico-transformativa del hombre. Hegel expresó filosóficamente algunos de los planteamientos de la economía política vigentes en su tiempo. Por ejemplo, en relación con la división y especialización del trabajo, la observación empírico-económica realizada por A. Smith en La Riqueza de las Naciones es traducida a una perspectiva filosófica por Hegel. Así, el filósofo alemán captó la importancia filosófica de los elementos económicos puestos de presente por la sociedad industrial de mercado con sus nuevas clases y ámbitos de sociabilidad tecno-productiva; distintos de los viejos estamentos y gremios del Ancien Regime, en cuyo reemplazo ocuparía un lugar preeminente la actividad humana práctico-transformativa; vista ahora no sólo en su enfrentamiento con la naturaleza sino en un marco cultural específico que reproduciría la dialéctica 74 del amo y el esclavo, del señorío y el siervo, analizadas por cierto en célebres pasajes de La Fenomenología del Espíritu. Para Hegel, la individualidad del portador de la voluntad particular expresa una determinación contingente; no posee la determinación general de lo universal y, por lo tanto, no puede ocupar el sitio (teórico) desde el cual se funda racionalmente un Estado, cualquier Estado. Hegel escribió tres textos que aluden al rol del trabajo y la producción económica en la configuración de la sociedad civil. Son ellos: los Textos de Jena (1803-1805), La fenomenología del espíritu (1807) y, finalmente, La filosofía del derecho (1821). Muchas y brillantes formulaciones pueden ser descubiertas o percibidas en dichos textos. La concepción del rol de las clases y estamentos (Stand) presentes en el sistema de las necesidades es un ejemplo vívido de la enorme riqueza filosófica del análisis hegeliano, el cual vislumbra, en particular, el nexo determinante que cumple dentro del proceso de formación del Estado moderno la categoría del trabajo, comprendiendo de forma lúcida la diferencia entre el trabajo abstracto y el concreto. Este reconocimiento pasa por la consideración del carácter del derecho político hegeliano. Hegel dice que la palabra trabajo en sí misma refleja el sentido de molestia: en latín proviene del término campo, labor, trabajo agrario, y expresa en los textos de Jena: “Trabajar significa aniquilar al mundo y/o moldearlo”. El trabajo es un proceso de aniquilación del mundo que, transformando el mundo, permite que el hombre se realice, causando efectos sensibles sobre él mismo. Algunos de los discípulos de Hegel definieron que este proceso de aniquilación podría tener una doble ontología: una de aniquilación inmediata, el acto del consumo (en que el objeto es destruido o, al menos, transformado sustancialmente) y otro acto en que el proceso de consumo es directo (proceso de trabajo que crea herramientas, a su vez, extensiones del cuerpo, de la voluntad, pero que no son consumidos inmediatamente en el proceso de trabajo). ¿Cuáles son los elementos esenciales que Hegel juzga en estos textos juveniles? En primer lugar, considera que el trabajo se puede definir como una mediación entre las necesidades y sus satisfacciones. El inevitable proceso de plasmación de esta interpretación es lo que da lugar, en Hegel, al trabajo. Incluso, el acto más elemental, como alimentarse o ser sujeto pasivo de la apropiación de la naturaleza, es en sí un proceso de trabajo. El segundo 75 elemento es descrito por Ruge como la distinción entre los útiles que superviven el proceso de trabajo como herramientas y el consumo directo. En tercer lugar, el trabajo siempre, según Hegel, implica el aprendizaje de una técnica. Cuando un animal satisface una necesidad, no está trabajando. En el animal, se trata de un mecanismo automático, repetitivo. Cualquier trabajo humano implica una técnica que debe ser aprendida para que la satisfacción de las necesidades se logre; y al necesitar de dicha técnica el trabajo adquiere un carácter vinculante. Además, en el ser humano, las necesidades y su satisfacción son procesos ascendentes en la medida en que la cultura se va acumulando y va siendo más compleja. Individualmente, el saber (la ciencia, la técnica) no se podría desarrollar —dice Hegel citando a Rousseau (cfr. el Discurso sobre la Desigualdad). Si fuera una técnica individualizada, si no existiera una cultura que implica transitividad, transmisión y acumulación, la técnica desaparecería con el creador. Por lo tanto, cuando una técnica tiene que ser aprendida, ello implica un ámbito de sociabilidad: la técnica es sociabilidad del proceso de trabajo. En especial, el trabajo en su modalidad abstracta logra su magnificación con la aparición del dinero, herramienta o el medio que permite vehicular la acumulación y el intercambio del trabajo abstracto26. Finalmente, en los Escritos de Jena, Hegel elabora una primera teoría de las clases sociales en términos de una evolución de lo sensible a lo espiritual, de lo concreto a lo abstracto. Para ello Hegel diferencia tres tipos de trabajo. En primer lugar, el trabajo de los campesinos, mineros, etc., que implica un vínculo directo con lo natural. Este es un tipo de trabajo en el que el hombre trabaja pasivamente y, en el fondo, es la naturaleza la que actúa. El hombre tutela un despliegue de la naturaleza, se trata de un trabajo plenamente dependiente de lo natural. La función activa esencial está en la naturaleza, no en el trabajo del hombre. En segundo término, aparece el trabajo de los artesanos. En éste surge una mayor función activa del hombre y una función pasiva de lo natural, pero para el artesano esta función activa del objeto producido está vinculada con su utilidad contingente, con el tipo específico de producto creado; es decir, la calidad del trabajo en términos de su concreción (valor de uso). 76 Una tercera esfera, según Hegel, la representa el comercio; ya en este ámbito se presenta claramente una superación del trabajo rural y artesanal porque el trabajo es visto en su abstracción. Al comerciante no le importa la utilidad del producto, sino cómo a través del producto puede percibir dinero y acumular riqueza. Dinero que a su vez expresa una abstracción del trabajo. El comerciante no se relaciona con la naturaleza como lo hace el productor, no la mitifica, sino que toma la sustancia pasiva de lo natural transformada como producto agrícola o artesanal, simplemente como un medio para una finalidad que se constituye en el ámbito del mercado. En la Dialéctica del amo y el esclavo, Hegel expone que el trabajo implica un vínculo indisoluble entre el amo y el siervo en el que cada cual se pone como finalidad, donde uno es dominado y el otro dominador. En cierta medida, el amo lo que busca no es el trabajo sino lo que pueda conseguir con él. De otra parte, el esclavo lo que busca no es hacer su trabajo sino obtener la subsistencia. El trabajo es visto por las dos partes: por el contratante o dueño de los medios de producción, y alternativamente, por el trabajador o siervo, como medio, no como fin. Pero se convierte en una finalidad en sí misma, porque el trabajo es el vínculo de las dos partes: el amo y el siervo. Se da una interdependencia y mutua negación. El esclavo sólo puede serlo en tanto tenga amo y el amo sólo puede serlo en tanto tenga esclavo. El capital y el trabajo, el contratante y el contratado, en todas las sociedades en las cuales actúe el trabajo bajo un esquema de disociación del productor con su producto, implicarán un vínculo de confluencia; pero, al mismo tiempo, de confrontación entre elementos opuestos. Ésta implica que ambos sean interdependientes debido a que satisfacen necesidades diferentes, por la finalidad del primero de subsistir y el segundo de acumular. Como lo acabo de mencionar, Hegel, ya maduro, dedica buena parte de uno de sus mejores libros —La filosofía del derecho— a la relación entre economía y filosofía política. En la primera parte de ese libro aborda el problema de los contratos, la posesión y la propiedad; y, posteriormente, efectúa la transición desde la familia y la sociedad hasta su síntesis en el Estado. Para definir los contratos empieza explicando la familia como primer vínculo del contrato y la posesión como la predecesora de la propiedad. 77 Hegel plantea que la sociedad civil es el ámbito de los intereses privados que están vinculados recíprocamente por la interacción de sus necesidades. Ve en el mundo del comercio el espacio para el despliegue de la sociedad civil cuya economía se desarrolla cada vez más, abstracción base del enriquecimiento; que crece en proporción contrapuesta a la singularidad del trabajo individual. La producción no está vinculada a la universalidad de la acumulación. Existe un nexo mayor del mercado que implica una conexión por medio del sistema de necesidades dando lugar al trabajo abstracto que se valora por medio del dinero y posee mayor significación social que el trabajo contingente. La vinculación de las personas con el trabajo manual los aleja de lo espiritual. Según Hegel, la clase industrial es, de suyo, más inclinada a la libertad. Ésta surge en las ciudades, mientras la clase campesina es más inclinada al sometimiento, pues el individuo no piensa en sí mismo, siendo la naturaleza el elemento activo; de ahí la inferioridad del trabajador manual. Hegel, posteriormente, elaborará una teoría, en La filosofía del derecho, en la cual el funcionarismo público aparece como la clase universal en tanto es, junto al monarca, el agente activo del Estado. La sociedad, para indemnizar a los que se dedican a lo universal, les paga un salario; son funcionarios públicos que se eligen mediante un examen de Estado. Hegel supone que todos pueden potencialmente llegar a ocupar cualquier posición. Esta teoría defiende el papel del Estado como árbitro supremo de los conflictos que se presenten en la sociedad civil; no importa si se trata del mundo de la industria, la banca, el comercio, es un mundo de intereses particulares. Además, considera que la división social del trabajo es benéfica, puesto que permite a muy diferentes intereses particulares satisfacerse recíprocamente a través del mercado. Hegel, en su explicación genealógica de la dicotomía sociedad civil-Estado, hizo una vehemente defensa del papel del funcionariado, al que consideraba portador del interés general, concreción de la sustancia ética del Estado. Por lo que, a pesar de las objeciones de sus contemporáneos, defendía y justificaba de manera firme la situación de predominio sobre la sociedad civil por parte de tal “clase universal”. Pero debemos hacer una precisión: a pesar de compartir muchos de los planteamientos de los economistas ingleses, Hegel consideró que el desarrollo del capitalismo debería darse —no mediante un esquema de laissez faire— sino bajo una fuerte dirección política del Estado. Hegel percibe, en efecto, cómo el inusitado enriqueci78 miento que el nuevo orden social industrial capitalista le depara a la sociedad civil es, al mismo, tiempo la palanca que arroja al abismo de la pauperización a masas gigantescas de la población. Por tal razón, la creación de un ejército industrial de trabajadores es concomitante con la aparición del fantasma del desempleo. Así que el trabajo, que cumple la vital función de plasmar al sujeto como un ente objetivo, resulta transmutado al implicar el extrañamiento del hombre frente a sus propias creaciones y fines de su actividad. Finalmente, esta transmutación hace que la falsa conciencia conlleve a una fetichización o cosificación de las relaciones verdaderas entre sujeto y mundo objetivo, entre lo natural y lo social. Estos aspectos, como lo acabamos de mostrar, fueron prefigurados por la propia Ilustración, para la cual el mundo del trabajo era, a la vez, la fuente de la subsistencia y la causa eminente de la desigualdad, (cfr: J. J. Rousseau, El origen de la desigualdad entre los hombres), donde apunta a definir el carácter de la riqueza social, particularmente en su forma de apropiación privada como base de la distinción entre la desigualdad natural y la desigualdad social y definiendo en términos aún clásicos, el rol de la propiedad privada en la configuración de la desigualdad social, política. Hegel es plenamente consciente de dicho proceso de elaboración conceptual a cargo de los Ilustrados románticos y repara en las limitaciones que, a su juicio, impidieron en el pasado dar cuenta de dicho problema. En la conceptualización hegeliana, una respuesta a este interrogante sólo será posible evaluando la articulación del Estado con la sociedad civil. En su perspectiva, al primero de ellos le corresponde el rol de portavoz del interés general, en tanto aquella expresa el topos donde se conjugan los diversos intereses privados. Por ello ocurre que Hegel, siguiendo paso a paso los razonamientos de Rousseau, llega a conclusiones diferentes. El Estado es, para él, la verdad de la familia y la sociedad civil en la medida, precisamente, que es de por sí la superación de la esfera egoísta del interés privado, no porque establezca un equilibrio entre las diferentes privacidades (o privaticidades); pues si así ocurriera, ello significaría —para el Estado— una inaceptable dependencia contingente —material de su derecho de existencia racional, de su legalidad objetiva—. Lo es, por el contrario, en tanto que la familia y la sociedad civil son vistas como meros momentos, singulares y particulares, de su marcha hacia lo universal e incondicionado. 79 La libertad individual ha tenido tres momentos históricos: la ciudad griega, el cristianismo y la sociedad civil. Esta última necesita instituciones de uso común que requieren la vigilancia del poder público para regular las partes. Las corporaciones constituyen un espacio de arbitraje, pues son los lugares en los que se arbitran los conflictos de productores y consumidores; los conflictos se superan por surgimiento del Estado, el cual, para defender los intereses particulares, cuenta con recursos como la policía (arbitraje mediante leyes, decretos, normas, etc.) y la justicia que posee tribunales que zanjan los conflictos de la sociedad civil. La concepción de Hegel del desarrollo humano ha sido expresada en sus Textos de Jena, en los que afirma que (antes de Cristo) el hombre no era todavía una personalidad plenamente existente por sí misma. La universalidad política se enfrenta a la personalidad privada y sus derechos; es decir, el desarrollo del hombre como ser individual y de su personalidad constituida es un fruto del cristianismo. En el fondo, el problema de la individualidad sólo se constituye con el cristianismo. Hegel —no lo olvidemos— mantiene un sistema en el cual la Historia desempeña un papel esencial. Su concepción espiritualista lo impulsa a utilizar el material fáctico, la singularidad del acontecimiento histórico, como la expresión finita, la encarnación del devenir, del sentido de finalidad (teleología) que preside y determina el despliegue de la totalidad expresiva. Hegel, siguiendo los trazos de la reflexión kantiana, se refiere a este proceso de escisión de la emergente sociedad civil: “Cuando la sociedad civil funciona sin trabas, se produce dentro de ella el progreso de la población y de la industria. Con la universalización de la conexión entre los hombres, a causa de sus necesidades y del modo de satisfacerlas, se acrecienta la acumulación de riquezas, pues de esta doble universalidad se extrae la máxima ganancia. Pero, por otro lado, tiene ello como consecuencia la singularización y limitación del trabajo particular, y, con esto, la dependencia y miseria de la clase ligada a ese trabajo, lo que provoca su incapacidad de sentir y gozar las restantes posibilidades, especialmente los beneficios espirituales, que ofrece la sociedad civil” (PH.R). 80 Algunos de estos tópicos fueron señalados por los filósofos políticos Ilustrados del siglo XVIII, por “los Iguales” y los primeros utopistas modernos. Lo realmente distintivo en Hegel, en este orden de ideas, es la articulación establecida entre la creciente división del trabajo —principalmente motivada por el desarrollo de la industria— y su efecto sobre el individuo. El trabajo mediador de la objetivación del hombre es, al mismo tiempo, el productor de su condición miserable. Al respecto conviene incluir aquí una expresión motivada del filósofo francés Eric Weil, enderezada a defender a Hegel de la acusación, muy extendida entre sus comentaristas, según la cual Hegel sólo entrevió el rasgo positivo del trabajo, dejando a un lado el “extrañamiento” del que es víctima el agente productor del trabajo concreto. “Con frecuencia se ha reconocido a Hegel el mérito de haber visto en el trabajo la esencia de la vida del hombre moderno en la sociedad. Pero casi siempre se ha seguido este elogio de la observación de que sólo habría visto el concepto abstracto del trabajo y habría ignorado sus formas concretas e históricas. Si esta crítica puede justificarse (y aún solamente en cierta medida) cuando se refiere a la fenomenología del espíritu, no tiene ningún fundamento cuando se refiere a la filosofía del derecho. No por haberlos señalado brevemente, Hegel dejó de observar correcta y totalmente los rasgos característicos del trabajo social moderno. No se trata solamente de haber distinguido los estados sociales, agricultores, funcionarios, hombres de oficios; éstos últimos divididos en artesanos, fabricantes y comerciantes, distinción clásica en la época de Hegel y ya casi superada. No se trata solamente de que considere al dinero como la mercancía universal, hecho natural en un lector de Adam Smith, J.B. Say y Ricardo. Lo que interesa comprobar es que Hegel ve y expresa claramente lo que la división moderna del trabajo significa para las condiciones de existencia del individuo” (Weil, 1985). De allí que tenga una gran importancia para el análisis la tesis consignada por Hegel en La Filosofía del Derecho referida al problema de la parcialización que el desarrollo del capitalismo introduce en el trabajo: 81 “Lo universal y objetivo reside, sin embargo, en la abstracción que ocasiona la especificación de los medios y las necesidades, y que especifica por lo tanto también la producción y produce la división del trabajo. El trabajo del individuo se vuelve así más simple y mayor la habilidad en su trabajo abstracto, así como mayor la cantidad de su producción. Al mismo tiempo, esta abstracción de la habilidad o de los medios completa y hace totalmente necesaria la dependencia y relación recíproca de los hombres para la satisfacción de sus restantes necesidades. La abstracción del producir hace además que el trabajo sea cada vez más mecánico, y permite, finalmente, que el hombre sea eliminado y ocupe su lugar la máquina” (Hegel, 1985). Ahora bien, una vez hecho este reconocimiento del proceso de extrañamiento y aún del desalojo (de la esfera del trabajo) del hombre por el maquinismo (con su secuela de desempleo), Hegel se plantea el problema de las formas primordiales —desde el punto de vista político y económico— que asume el desenvolvimiento y desarrollo de esta contradicción en el carácter del trabajo moderno, burgués. En primer lugar, en esta dirección, demuestra que es en el seno de las actividades industriales donde surge con mayor fuerza y nitidez la cuestión de la libertad del individuo; concebido no sólo bajo el rasero abstracto del citoyen, sino del hombre productor de medios de subsistencia y creador, por tanto, de riqueza. Así, acota en la Filosofía del Derecho: “En la clase industrial el papel del individuo resulta especialmente acentuado y este sentimiento de sí está muy estrechamente ligado con la existencia de una situación jurídica. Por ello, el sentido de libertad y orden surgió fundamentalmente en las ciudades...” Y haciendo una comparación con la primera clase por él mencionada, la clase sustancial —vinculada a la producción del suelo— dice un poco más adelante: “La primera clase, en cambio, tiene que pensar poco en sí misma; todo lo que adquiere es el don de un extraño, la naturaleza. Este sentimiento de dependencia es para ella lo primero, y a él se une con facilidad el aceptar pacientemente todo lo que pueda 82 sobrevenir. La primera clase está, por lo tanto, más inclinada al sometimiento y la segunda más a la libertad” (Ibíd.). Sin embargo, en el seno de esta última, se presenta la contradicción interna entre el propietario de los medios de producción y el trabajador directo, de la que Hegel da sólo una visión parcial. En efecto, para éste el representante de la clase industrial es ante todo el propietario, segmento social que hace parte de las llamadas clases medias, de tal modo que el orden y la libertad al que se haya vinculado por razón de su posición frente al mundo de lo económico es realmente el orden y la libertad de empresa, la libertad y el ordenamiento del mercado. No obstante, Hegel reconoce que la riqueza producida no resuelve el problema —para el conjunto de la sociedad civil— del mantenimiento de un nivel de vida digno y no recurre al expediente ilusorio propuesto en su momento por Rousseau de una república donde se concilien las diferencias acercando a los ricos y a los desposeídos al nivel de las clases medias. En cambio, se plantea por Hegel una cierta intervención del Estado, cuando señala, frente a la tendencia a la polarización de la feroz concurrencia en el mundo de la sociedad civil: “Surgen de este modo (necesidades que posibilitan la existencia de) procedimientos e instituciones que pueden servir para un uso común. Estas tareas generales e instituciones de utilidad común requieren la vigilancia y previsión del poder público, pues los diversos intereses de los productores y los consumidores pueden entrar en conflicto entre sí, y si bien en el todo la relación correcta se produce por sí misma, la compensación requiere una regulación consciente que esté por encima de ambas partes” (Ibíd.). Pero como la razón de ser del Estado no consiste en la simple igualación de los intereses enfrentados en la esfera económica —pues su fundamento en este punto es explicado externamente, siendo un poder impuesto desde fuera a la sociedad civil y no el fruto del acuerdo de sus distintas voluntades individuales— lo que Hegel hace es abrir el espacio para un nuevo elemento de mediación: la corporación. En efecto, este organismo cumple la función 83 de ordenar las diversas ramas de la producción, siendo un elemento de transición entre el puro interés egoísta del productor aislado, considerado en su individualidad grosera, y el bien común. Del proceso de conjunción de estos egoísmos nace la forma de un universal que viene a representar en la sociedad civil una especie de prolongación de la familia, dando lugar al desarrollo de una serie de privilegios que —a diferencia de los odiosos privilegios surgidos de la ociosidad feudal— son resultado de la comunidad de esfuerzos; entonces, el devenir de la corporación permite que la: “(...) ayuda que recibe la pobreza pierda su carácter de accidentalidad y lo que hay en ella de injustamente humillante. Al mismo tiempo la riqueza, al cumplir sus deberes con la corporación, deja de provocar el orgullo de su poseedor y la envidia de los demás. De este modo, la honestidad recibe su verdadera honra y reconocimiento” (Ibíd.). De otra parte, Hegel propugna que la sociedad logre darles pleno empleo a sus miembros para evitar la caída de una gran masa, no ya en la pobreza digna del trabajador, sino en la mendicidad, fermento de la anarquía y de la acción irresponsable de quienes no se sienten con deberes frente al Estado. De prosperar esta situación, se acrecentarían los elementos disolventes de la comunidad. Esta tendencia habrá de ser combatida con los recursos de la coerción, con el factor denominado por Hegel poder de policía, término que en su obra posee su amplio sentido original, como administración de justicia y legitimación del ordenamiento social, aunque ello no implique la imposibilidad de que se ejerza la fuerza de represión del Estado respecto a sus ciudadanos cuando estos rompan las normas y leyes del mismo. El Estado hegeliano disciplina a sus ciudadanos de tal modo, que éstos le deben obediencia; pero este deber se diferencia de la sumisión existente en el Estado despótico, en tanto existen leyes que fundan la legalidad. El camino a través del cual Hegel deduce esta legalidad conduce hasta la eticidad que preside el conjunto de las actividades del Estado, incluyendo sus momentos particulares: la familia y la sociedad civil. Las prerrogativas de las clases ciudadanas son, de este modo, limitadas. La clase industrial no es la regente de la sociedad civil y, menos aún, la conductora del Estado. Si 84 Hegel es un pensador burgués moderno, lo es combinando simultánea y contradictoriamente rasgos individualistas y totalizadores en su visión madura de los problemas de la política. Es el coronador de la arquitectónica de la razón que venían construyendo los filósofos alemanes, en particular Kant. Su expresión, por cierto, no es representativa del ala radical de la burguesía en ascenso, de aquellos sectores que llevaron sus argumentos teóricos hasta su último plano: defensa de la república democrática y de la esfera inalienable de los derechos individuales y respaldo al proyecto y a la praxis de 1789. Su visión concilia los intereses de estas clases medias con los intereses de la monarquía constitucional, llegando incluso a colocar como modelo de una forma estatal adecuada a los dictados de la razón al régimen prusiano de Federico Guillermo IV y a conceder, lo que no dejará de tener hondas repercusiones, que dicha estructura monárquica habría de ser hereditaria. 4.2. HEGEL Y LA LIBERTAD POLÍTICA Según una parte de la crítica la filosofía política de Hegel, constituyó una defensa del nacionalismo y una crítica al excesivo individualismo y al cosmopolitismo implícito en los derechos del hombre. En esa perspectiva, los individuos y sus fines no devienen fundamentales. Los individuos caen dentro de la categoría de medios. Sus deseos y satisfacciones son sacrificados para la realización de los fines más amplios de las naciones. En síntesis, Hegel creía que las realidades y causas efectivas de la Historia eran movidas por fuerzas impersonales y generales. Hegel, que recusa las propuestas —para él extremas— surgidas en el pensamiento iluminista, como una disolución de los verdaderos fundamentos éticos del Estado, no desconoce, sin embargo, la gran importancia que en la constitución del Estado moderno tiene el concepto de libertad. Al igual que su antecesor Hobbes, su defensa del papel autónomo del Estado con respecto a las dispersas voluntades individuales está basada en una firme convicción de que el Estado, al darle una forma a la estructura de la comunidad o sociedad civil, permite que el pueblo exista como tal y que los individuos, los ciudadanos del Estado, puedan gozar de libertad siempre y cuando no rompan los marcos de legalidad objetiva establecidos para la vida social comunitaria. La libertad está indisolublemente vinculada con la 85 realización de la organización racional del Estado. Sin embargo, mientras el colosal Leviatán hobbesiano deja sin derechos políticos reales a los miembros del Estado y concentra en éste las características de un poder absolutamente omnímodo, el Estado hegeliano considera el valor de los elementos constitutivos del mismo, en cuanto momentos en la marcha hacia la libertad racional, e intenta darle un espacio político a los ciudadanos ubicados en la esfera de la sociedad civil por medio de su participación en los parlamentos (# 301 de la Filosofía del Derecho) o representación de los estamentos, en el caso de la cámara baja (distinta a la cámara alta donde toma asiento la aristocracia). Podríamos resaltar aquí que Hegel, a diferencia de la filosofía absolutista, consideró al trabajo y al sistema de las necesidades; o sea, a la interdependencia mutua de los trabajos individuales, como determinantes en el proceso de constitución del sistema estatal y de la misma estructura específica de la sociedad civil. Esta concepción hegeliana de articulación entre lo económico y lo político no era, por cierto, un agregado formal dentro de su teoría del derecho. El poderoso sentido histórico de que Hegel hizo gala le impelía a considerar el proceso de desenvolvimiento de la consciencia relacionándola con el carácter determinado del trabajo, observando y dando cuenta de sus visos para explicar su contenido e historia, de tal modo que el extrañamiento de la consciencia pudiera ser explicado sin una consideración ahistórica. Así, por ejemplo, la significación básica y decisiva de la Fenomenología del Espíritu estriba en que allí se describe la marcha de la consciencia, su original distanciamiento consigo misma (la dislocación de la relación entre libertad y necesidad, los distintos momentos de este tránsito, que a su vez son los distintos períodos de la Historia Universal en consonancia con la peculiar caracterización que de ellos hace Hegel), y el reencuentro final (el devenir de la autoconciencia) bajo el marco histórico de un nuevo y superior dominio de la razón: el Estado de libertad. Si se sigue el pensamiento hegeliano desde sus comienzos —el período de juventud—, se verá que siempre vinculó el problema de la conciencia con la temática de la libertad y no exclusiva ni fundamentalmente en la libertad interior, monádica, que está presente en la moral kantiana, sino advirtiendo, como parte de su realización, la esfera de la libertad política, ciudadana. Se sabe que la órbita del Estado —en el conjunto del sistema hegeliano— no es, pese a su 86 importancia decisiva, la culminación de la arquitectónica de la razón. Más allá se encuentran las órbitas del saber absoluto. Además, la historia de la cultura no es más que un procedimiento enajenado, de exteriorización del espíritu; describe la parábola de su reencuentro y de la superación de su desdoblamiento. Si el individuo —a través de la actividad productiva y de la actividad artística, a través del mundo de las necesidades y del mundo de la cultura— supera el extrañamiento, produce en este caso particular el tránsito de la singularidad determinada a la esencia universal. De igual forma, su integración al mundo de la política por la mediación de los niveles arriba señalados, le permiten dicha concreción de lo singular con la eticidad superior del Estado. El problema que Hegel, por lo tanto, se plantea en este punto es definir, caracterizar, la forma como las diversas clases de la sociedad acceden al mundo de la política y participan en la gestión de los asuntos del Estado. O de otra forma planteado: ¿qué elementos determinan la soberanía del Estado? Para Hegel, la soberanía del Estado se deduce del contenido, de la sustancia del Estado. Hemos mostrado cómo la verdad del Estado está dada por la eticidad y que esta última no equivale a la moralidad que operaría en un plano más concreto. Hegel explica tal diferencia en los siguientes términos: “La unidad del bien subjetivo y del bien objetivo existen en y por sí en la eticidad en la que se produce la reconciliación de acuerdo con el concepto. Si la moralidad es, pues, la forma de la voluntad según el lado de la subjetividad, la eticidad no es ya meramente la forma subjetiva y la autodeterminación de la voluntad, sino el tener como contenido su propio concepto, es decir la libertad. Lo jurídico y lo moral no pueden existir por sí y deben tener lo ético como sostén y fundamento. En efecto, al derecho le falta el momento de la subjetividad que la moral tiene, en cambio, exclusivamente para sí, por lo cual ninguno de los dos momentos tiene por sí realidad. Sólo lo infinito, la idea, es efectivamente real; el derecho sólo existe como rama de una totalidad, como planta que crece en torno a un árbol firme en y por sí” (Ibíd.). Si lo ético se enraiza no en el plano de la subjetividad, que equivale a la voluntad individual y al querer (deseo y volición de la conciencia singular), sino en la objetividad racional que contiene la libertad; y, además, si 87 el derecho es “sólo la rama de un árbol firme en y por sí”, se comprende por qué Hegel considera como un asunto mal enfocado el hecho de discutir quién hace la constitución, ya que ésta no es un simple documento escrito fruto de la subjetividad; es una expresión real “anterior a toda teoría” y concordante, por lo tanto, con el carácter del pueblo organizado en comunidad. A lo sumo, se podría pensar en modificar tales o cuales aspectos de las normas llamadas “constituyentes”; pero las modificaciones sustanciales se encuentran vinculadas a la marcha del espíritu de la libertad. Sin embargo, esto no significa que en el Estado hegeliano la voluntad subjetiva de los ciudadanos sea desdeñable. Hegel precisa que al reconocer los individuos en lo universal la realización —realidad efectiva— de su libertad concreta, al mismo tiempo toma forma en el Estado moderno la idea de libertad. Su razonamiento lógico es, por tanto, perfectamente deducible de su filosofía; no por otra razón podemos comprender por qué Hegel recusa, de entrada, el supuesto procedimiento que deduce al Estado de un contrato pues ¿cómo una hipotética acción contingente va a suministrar la verdad de un universal? Inclusive, en una referencia posterior, cuando Hegel trata el carácter de la constitución del Estado, prefiere su definición en términos abstractos antes que permitir la entrada teórica de la voluntad individual como factor que, por medio de un proceso de consentimiento, asuma una postura constituyente frente a la génesis de las leyes y de los fines supremos del Estado. Los individuos no realizan un contrato entre ellos para darle el poder a un tercero, ni alienan sus derechos al Estado; el Estado es, más bien, la realidad de su derecho; este derecho es, a su vez, uno de los momentos del discurrir del concepto. Por consiguiente, la verdad del universal se da en lo particular y lo particular sólo llega a la existencia como forma históricamente dada de lo universal. Temática del devenir de las formas y las estructuras desarrollada in extenso en la ciencia de la lógica, principalmente en su parte primera. Siguiendo con la interpretación hegeliana de la constitución del Estado, vale la pena señalar cómo piensa éste la articulación del pueblo con la misma: “La expresión ‘la multitud’ designa la universalidad empírica más corrientemente que el corriente ‘todos’. Pues si se dice que es obvio que en este ‘todo’ no están comprendidos por lo menos los niños, las mujeres, etc., más obvio aún es que no debería usarse la 88 expresión totalmente indeterminada ‘todos’ cuando se trata de algo completamente indeterminado. Ha circulado una tal cantidad de falsas representaciones y expresiones acerca del pueblo, la constitución y los estamentos, que sería un esfuerzo vano querer citarlas, explicarlas y corregirlas. La opinión que suele tener la conciencia ordinaria sobre la necesidad y conveniencia de la colaboración de los estamentos consiste fundamentalmente en suponer que los diputados del pueblo, o incluso el pueblo mismo, tienen que saber mejor que nadie qué es lo que más les conviene y tienen, sin duda alguna, la mejor voluntad para este bien. Por lo que respecta al primer punto, lo que ocurre es que el pueblo, en la medida en que esta palabra designa una parte determinada de los miembros del Estado, expresa la parte que precisamente no sabe lo que quiere. Saber lo que se quiere, y más aún, saber lo que quiere la voluntad en y por sí, la razón, es el fruto de un conocimiento profundo que no es justamente asunto del pueblo” (Ibíd.). Es de este modo muy claro el sentido restringido en que Hegel se plantea el problema de la participación política del pueblo. Dado que lo universal sólo puede existir bajo la forma de un particular y que, según lo expresa con crudeza Hegel, el pueblo no interviene en los dictados del Estado en la medida en que es “aquella parte de los miembros del Estado la que no sabe lo que quiere”, ¿qué elemento del conjunto de los miembros del Estado efectúa la mediación? El problema de la relación Estado-pueblo, que está por fuera de consideración en el análisis de Hegel, resulta suplantado por el de la relación Estado-monarca. Pero allí —efectivamente— no se produce una relación real de mediación entre estos dos elementos; lo que se produce en este esfuerzo hegeliano como objetivación es la voluntad del Estado. Lo verdaderamente importante aquí es entonces la forma como Hegel define la soberanía del Estado como soberanía del monarca y, a su vez, a ésta como el lugar donde se logra concretar la soberanía de todos los miembros del Estado. Para ello, su argumentación describe, a través de un rodeo especulativo, la igualación del monarca con la soberanía nacional externa. De esta forma, en el Estado considerado como uno frente a los demás Estados en el plano de las relaciones internacionales, el ciudadano exalta su soberanía al 89 ser poseedor de un ámbito subjetivo: el de tener conciencia patriótica. Sólo en el marco de la soberanía nacional es pensado por Hegel el lugar de la soberanía del pueblo: “Se puede hablar de soberanía de un pueblo para expresar que respecto del exterior es un pueblo independiente y constituye un Estado” (Ibíd.). De allí que el patriotismo, supuestamente una de las más elevadas formas de consciencia del Estado por parte de sus ciudadanos, sea explicado de forma empírica por la simple identificación del interés particular con el universal. Pero el verdadero representante frente a los demás Estados es el monarca. Los demás miembros actúan, o bien en el ámbito económico en las relaciones internacionales comerciales y afines, o bien como guerreros en defensa del Estado frente a los enemigos exteriores. Hegel, en este punto de su exposición, hace un análisis de la función guerrera de acuerdo a la tradición del pensamiento político de una manera tal que siente el lector una reminiscencia del pensamiento político de Platón. En este orden analítico resulta altamente sintomático el rol asignado por Hegel al héroe. Este encuentra su legalidad (tanto la de su existencia empírica como la de sus propias acciones) en el marco de su lucha por constituir —más exactamente por encarnar—, en un momento histórico, el reencuentro del espíritu que se sabe y se quiere en la forma de unidad de la Nación —siendo la encarnación del proceso de conformación de un Estado—. “Cuando sólo existe el estado de naturaleza, un estado de la violencia en general, la idea (de la libertad organizada) funda en su oposición (o sea en oposición al estado de naturaleza) un derecho del héroe” (Ibíd.). Según Hegel, la guerra es el lugar del heroísmo y, efectivamente —en correspondencia con esta idea— piensa el plano de las relaciones internacionales como un ámbito gobernado por la fuerza. Por esta razón, podemos comprender la expresión hegeliana que señala que la totalidad de las expresiones de una fuerza es igual a la fuerza misma. En la filosofía de la historia, el rol del llamado gran hombre o personalidad histórica está determinado por la igualación de sus acciones y su querer subjetivo con la marcha del espíritu universal. Pero las 90 acciones de estos grandes hombres no pueden marchar, a su vez, en contravía de dicho espíritu. Hegel repite aquí, en una forma mucho más elaborada, un lugar común desde tiempos de Hobbes, y aún presente en Kant: la negación del derecho de rebelión. Si el héroe existe en el proceso de conformación de un Estado, su función no puede consistir en disolver las estructuras que trabajosamente en el decurso de la Historia se han erigido como racionales. Su negatividad no puede ser absoluta; el terror constituido en instrumento de la pura negatividad había sido rechazado por Hegel al comentar el proceso de ejercicio del poder en la Francia revolucionaria por parte de los jacobinos. En la Filosofía del Derecho es de nuevo restringido el campo en el que puede actuar la fuerza: en modo alguno, ha de resquebrajar el ordenamiento interno del Estado: su función histórica reside en ayudar a concitar la hegemonía del espíritu sobre el conjunto del devenir histórico. Napoleón es, en este sentido, un ejemplo completamente adecuado. Hegel, que se puso del lado de Napoleón frente a la aristocracia prusiana —que agitaba la bandera del nacionalismo para preservar sus privilegios—, veía en este gran hombre la encarnación del espíritu de la Historia Moderna. Las rupturas de las fronteras nacionales ocasionadas en las primeras dos décadas del siglo XIX por la expansión napoleónica, fueron saludadas por Hegel. El principio del que partía para asumir tal actitud nos muestra hasta dónde es imposible dar cuenta de manera esquemática de la riqueza de determinaciones conceptuales que este pensador produjo: “(...) Los Estados desean la independencia, y allí está su honor... Pero la independencia también debe ser considerada como un principio puramente formal... Cada vez que un Estado fue absorbido por otro, sólo perdió la independencia formal, pero no su religión, ni sus leyes, ni el contenido (concreto) de su vida... la dirección de los Estados avanza por tanto hacia su unidad” (Ibíd.). Así, el papel de la fuerza como generadora de derecho es pensado en términos históricos. Hegel, como lo han señalado muchos de sus críticos, introduce un orden racional en la Historia; ésta no es, por lo tanto, un campo fértil para el caos, la anarquía y el resquebrajamiento de las estructuras racionales del Estado. De allí que sean concebidas como válidas las acciones de coerción 91 que demuestren su legalidad y racionalidad. Sólo lo real es racional y sólo lo racional adquiere su legalidad existenciaria. En esta relación dialéctica inscribe Hegel la temática de la soberanía nacional (estatal) y popular. El rol de la fuerza no es el de ser, por sí misma, creadora de derecho. Es, más bien, un determinado instrumento de los fines universales de la razón; aunque en el caso concreto del fin hegeliano de la Historia coincida con la instauración del orden burgués y sea, por lo tanto, legal su utilización por las fuerzas que encarnen este ideal político (expresado por Hegel en la mencionada forma filosófica del Estado de libertad, con su antinómica contraposición entre el hombre real y el ciudadano abstracto), y no por parte de las fuerzas “ocultas” del proletariado, como habría de proponerlo en su momento Marx. 5. Marx y la filosofía del derecho hegeliano Aunque conocido tardíamente, ya que sólo vio la luz íntegramente en 1927, el trabajo juvenil de Marx —Crítica a la filosofía del derecho estatal de Hegel— cobró sin duda un lugar de importancia en el conjunto de la investigación y de los estudios sobre la relación entre Marx y el Estado, efectuados por filósofos, políticos, académicos, críticos de Marx y seguidores de su pensamiento. Tal interés no consiste, obviamente, en el hecho de que allí —en dicho trabajo juvenil— se encuentre la madura concepción que, sobre las relaciones entre la base económica y los diversos niveles sobreestructurales, sería definitoria, en grado sumo, de la particularidad que ha hecho tan polémica e influyente, aún en nuestros días de quiebra de los paradigmas del socialismo real, la obra de Marx y las numerosas exégesis que de ésta se han desprendido, tanto en el mundo occidental, como en sus adaptaciones orientales y tercermundistas. 5.1. LA POSTURA POLÍTICA INICIAL DE MARX En estos primeros años de lectura atenta de la obra hegeliana se produce lentamente su separación frente al pensamiento hegeliano. Marx, además, acababa de retirarse del periodismo: La Gaceta Renana había sido cerrada por las autoridades prusianas en marzo de 1843. El periodismo lo había colocado frente a una serie de complejos problemas de orden político y 92 económico que fueron la motivación inmediata para abordar el estudio de ciertos problemas teóricos; más exactamente, lo que lo compulsó a enfrascarse en cuestiones sociales y económicas, que eran el centro de discusión en las nuevas corrientes democráticas de la Alemania post-hegeliana. En suma, como lo puede demostrar un atento análisis de uno de sus textos capitales de este período, Los manuscritos económicos filosóficos de l844; en Marx, actuarían dos tensiones contrapuestas: su estudio crítico del hegelianismo y sus indagaciones en el campo de la economía política. Este proceso el propio Marx lo describiría como “un ajuste de cuentas” con su conciencia filosófica anterior. Así expresaba en su época madura al resumir retrospectivamente sus estudios juveniles: “Hace cerca de 30 años, en una época en que todavía estaba de moda aquella filosofía, tuve ya ocasión de criticar todo lo que había de mistificación en la dialéctica hegeliana”. Ciertamente, un lugar relevante en los escritos de este período corresponde al texto que nos ocupa: la crítica a la concepción hegeliana del Estado. En realidad, desde su elaboración temprana, el pensamiento de Marx centró su atención en torno a la cuestión del Estado. Marx veía como una de las insuficiencias principales dentro de la filosofía expuesta por Feuerbach en sus Tesis para una reforma de la filosofía (1843) la poca atención que allí se le concede a las cuestiones políticas y estatales. En este sentido, uno de los aspectos que más llama la atención es cómo en su trabajo teórico posterior se produce, como lo afirma H. Marcuse, su alejamiento de la filosofía. Lo que no quiere decir que la haya desechado del ámbito de sus preocupaciones, sino que no escribió trabajos filosóficos destinados a su publicación. Por varias de sus cartas nos damos cuenta de que tuvo intenciones de hacerlo: “El primer trabajo que emprendí para resolver las dudas que me asaltaron fue una revisión crítica de la filosofía hegeliana del derecho. Este trabajo me llevó a la conclusión de que tanto las relaciones jurídicas como las formas políticas no pueden ser deducidas de razones jurídicas y políticas ni exclusivamente por ellas; aún menos imposible es explicarlas e inferirlas de la llamada evolución general del espíritu humano. Tienen sus raíces exclusivamente en las relaciones materiales de vida, cuyo 93 conjunto resume Hegel siguiendo el precedente de los escritores franceses e ingleses del siglo XVIII, en la denominación de “Sociedad Civil”. Pero la anatomía de la sociedad civil hay que buscarla en la economía política.” Un análisis de los principales puntos de la “revisión crítica” de la filosofía hegeliana del derecho ayudará a la comprensión de los hitos que demarcan este proceso. Por consiguiente, trataremos ahora de ocuparnos de estos aspectos en propiedad. Lo haremos tocando tres puntos, a mi entender, primordiales: • La inversión hegeliana (una crítica materialista al idealismo de Hegel) • La relación existente entre familia, sociedad civil y Estado • Y, por último, la refutación marxista a la defensa hegeliana de la burocracia y el “funcionarismo”. Aspectos a partir de los cuales podremos desarrollar una reconstrucción del pensamiento posterior, respecto al rol del Estado. Punto que desde luego, resalta en la manera como se concretaría en el socialismo real la dominación (y tendencial anulación) que de la sociedad civil hará “el Estado político del futuro” —de acuerdo con la expresión característica del dirigente socialista alemán A. Bebel—, es decir, la “dictadura del proletariado”. 5.2. LA “INVERSIÓN” HEGELIANA El aspecto central, desde el punto de vista metodológico, en el cual se basa la crítica juvenil de Marx a la filosofía hegeliana del derecho es la utilización por parte de Hegel de un método apriorístico que combina la “vacua especulación con la imposición sublimada del empirismo”. Según el punto de vista de Marx, el análisis hegeliano parte de un error de principio, ya que invierte las relaciones entre la realidad y la idea. Ello ocurre en la medida en que mistifica su unidad de análisis filosófico- político: la relación entre sociedad civil y Estado político, como lo hemos explicado, concibe al Estado como productor sustancial de la sociedad civil y a ésta como un mero apéndice del primero, cuya funcionalidad consiste en establecer la mediación entre la esfera singular de la familia y la universalidad ética asumida como legalidad inmanente del Estado. La crítica de Marx denotaba el influjo de la llamada ala izquierda del hegelianismo, principalmente en cabeza de Ludwig Feuerbach. La obra de Feuerbach 94 fue dada a la luz en la primera mitad de la década de los 40 del siglo XIX [cfr.: La esencia del Cristianismo (1841) y Tesis provisionales para una reforma de la Filosofía (1843)]. En la primera de estas obras, Feuerbach intenta superar la filosofía hegeliana proponiendo en reemplazo del reino de la idea absoluta y la religión tras ella oculta, el “reino” del hombre. El pensamiento de Feuerbach no era un materialismo en el sentido genérico de la expresión; era en su médula, más bien, una antropología: “El hombre –éste es el secreto de la religión– objetiva su ser, y se convierte, en consecuencia, en el objeto de este ser objetivado (das Wesen), transformado en un sujeto y, respectivamente, en una persona; él se imagina que es un objeto, pero objeto de otro objeto, de otro ser. El hombre es un objeto de Dios.... el hombre en y por medio de Dios, sólo se tiene a sí mismo como última finalidad” (Ibíd.). La religión, de este modo, sólo expresa en forma invertida la verdadera potenciación, la fuerza natural del hombre, su deseo de poseer por medio de sus acciones un alcance ilimitado. La trascendencia de su finitud. Así “Es el hombre consciente de sí mismo debido al objeto: la conciencia del objeto, es para el hombre la consciencia de sí mismo. (...) el objeto es su esencia manifestada. Y esto vale no sólo para los objetos espirituales, sino también para los objetos sensibles” (Ibíd.). Si bien, a ojos de Marx, esta consideración del hombre representaba un incuestionable progreso en la filosofía social del momento, aún era insuficiente. El reconocimiento del doble rasgo general (biológico y social) del hombre, terminaba por ser una verdad a medias mientras no se introdujese el carácter esencial y, por lo tanto, concreto que —dentro de la configuración del hombre— representa el trabajo, la praxis social. Estas características de la crítica feuerbachiana no tardaron en ser vistas y señaladas por Marx. Lo que ocurre poco más adelante en las Tesis sobre Feuerbach, donde expresa: “(...) Feuerbach arranca del hecho de la autoenajenación religiosa, del desdoblamiento del mundo en un mundo religioso, imaginario, y otro real. Su cometido consiste en disolver el mundo religioso, reduciéndolo a su base terrenal. No ve que, después de realizar esta labor, falta por hacer lo principal. En efecto, el 95 hecho de que la base terrenal se separe de sí misma y se plasme en las nubes como reino independiente, sólo puede explicarse por el propio desgarramiento de esta base terrenal consigo misma. Por tanto, primero, lo que hay que hacer es comprender ésta en su contradicción y luego revolucionarla prácticamente, eliminando la contradicción. Por consiguiente, después de descubrir, v. gr., en la familia terrenal el secreto de la Sagrada Familia, hay que criticar teóricamente y revolucionar prácticamente aquella”. La inversión idealista, por ende, no sería refutada mientras no se explicite la contradicción producida por este desdoblamiento místico-ideal. Explicación a su vez mediada por el hallazgo del sustrato real en el cual se apoya su existencia. Proceso que será discutido por Marx en su conocido estudio sobre la ideología alemana y en el cual, de manera concomitante, elaboraría una primera versión de su teoría de las ideologías. Así, desde el punto de vista marxista, aunque Feuerbach intenta presentar al hombre como un ser concreto dotado de sensibilidad, su carácter genérico lo limita al dejar de lado su condicionamiento histórico-material. Para Feuerbach no existe la Historia; en consecuencia, su filosofía no puede penetrar, resulta imposibilitada para captar los nexos causales que dan cuenta del rasgo concreto-real del hombre, en la medida, precisamente, en que éste sólo puede ser comprendido como sujeto histórico, como un producto de la Historia. Marx refuta la indeterminación feuerbachiana al señalar que la llamada “esencia” del hombre, no es otra cosa que el conjunto de las relaciones sociales. Por tal razón a la hora de establecer un balance comparativo entre Hegel y Feuerbach, Marx reconoce en la filosofía hegeliana enormes potencialidades que requerirían, ante todo, desligar su núcleo dialectivo de la mistificación idealista que lo sobredetermina. En Hegel era enormemente importante su percepción histórica, la categorización del mundo como devenir, transición perenne del ser a la nada y de la nada al ser. La incompletud del punto de vista hegeliano estriba, en este punto, según Marx, en que su visión idealista le impide comprender el proceso real como autónomo, sujeto a sus propias leyes. Hegel lo vislumbra como un reflejo de la idea absoluta. En síntesis, la clave resolutiva del problema consiste en una conjunción del sentido de lo 96 histórico que preside la reflexión hegeliana, y la acerba crítica materialista que comienza a fundamentar Feuerbach. A partir de esta conjunción Marx construirá la categoría de “relaciones sociales” como los topos donde se encuentra en realidad la pretendida “esencia” humana. La verdad del hombre se encuentra en su articulación con los demás hombres y la naturaleza a través de la praxis social y el trabajo productivo. En el texto La crítica a la filosofía del derecho estatal de Hegel encontramos una clara referencia polémica contra la inversión idealista hegeliana: “La realidad no es expresada como es ella misma, sino como una realidad distinta”. 5.3. LOS LIMITES, SEGÚN MARX, DE LA DICOTOMÍA HEGELIANA (SOCIEDAD CIVIL - ESTADO) Abordaremos ahora el asunto de la relación entre Estado y sociedad civil. Como lo anotábamos líneas atrás, en la concepción hegeliana, al primero le corresponde el rol de portavoz del interés general; en tanto a la segunda, el de cobijar el círculo de los diversos intereses privados, referidos al universo de la producción y el cambio material. El pensador italiano neomarxista Umberto Cerroni explica este procedimiento señalando: “El problema de la relación entre el Estado y la sociedad civil, tan esencial en la historia del pensamiento jurídico y político moderno, es el problema de la mediación, compensación y resolución del antagonismo entre intereses particulares e interés general; y constituye, por tanto, un campo experimental importantísimo del problema más general de la mediación entre lo particular y lo universal. Bajo el aspecto que lo examina Hegel, se trata del mismo problema de la libertad concreta, que, señala Marx, consiste (para Hegel), en la identidad (necesaria y doble) del sistema del interés particular (de la familia y la sociedad) con el sistema del interés general (del Estado). (...) el problema en cuestión se relaciona, por tanto, con el tema más general de la relación entre lo finito y lo infinito, y con la estructura de la mediación en Hegel” (H. Cerroni, 1972). 97 De tal modo que el aspecto principal en esta conceptualización estriba en su intento de resolución de los antagonismos existentes entre lo privado y lo público. Ello, desde un punto de vista filosófico-político, conduce a Hegel hacia la caracterización de la soberanía popular y del Estado; a formularse el carácter inmanente de los derechos cívicos de los ciudadanos, es decir, sus relaciones mutuas y la articulación correspondiente de la esfera de la sociedad civil frente al poder público. En pocas palabras: estas oposiciones constituyen la percepción hegeliana de las contradicciones de la sociedad mercantil. Hegel, que postula la síntesis de estos contrarios, no logra explicarse la decisiva cuestión de las mediaciones que conducen a dicha identidad. Allí, según Marx, residiría el talón de Aquiles de su filosofía política, pues la prestidigitación hegeliana hace pasar por mediación real una mediación lógica; y, de este modo, retomamos como en círculo vicioso al punto de partida. En este sentido Hegel postula: “La idea efectivamente real, el espíritu, se divide en dos esferas ideales de su concepto: la familia y la sociedad civil, que constituyen su finitud, para ser, a partir de la idealidad de estas, espíritu efectivamente infinito por sí. Reparte así en estas esferas el material de su realidad finita, es decir, reparte a los individuos en forma de multitudes, de manera tal que en el individuo esta adjudicación aparece mediada por las circunstancias, el arbitrio y la propia elección de su destino” (Ob. Cit.). En su ensayo juvenil Marx glosa el anterior concepto hegeliano traduciéndolo al lenguaje “ordinario”: “La materia con la que el Estado se mediatiza en la familia y la sociedad civil está constituida por ‘las circunstancias, el arbitrio y la propia selección de su determinación’. Por esto, la razón del Estado nada tiene que ver con la repartición de las materias del Estado entre la familia y la sociedad civil. El Estado surge de ellas de una manera arbitraria e inconsciente. La familia y la sociedad civil aparecen como el oscuro fondo natural sobre el que se proyecta la luz del Estado” (Marx, Ob. Cit.). Por tanto, según Marx, la sociedad civil y la familia serían, en la concepción hegeliana, partes finitas de una totalidad expresiva y no, en modo 98 alguno, esferas materiales del Estado y por ende su fundamento objetivo. La verdadera realidad no se manifiesta en las esferas de la familia y la sociedad civil; se da en el Estado que las subsume y contiene como meros momentos. Son expresión finita de este último de donde resulta: “Un misticismo lógico panteista (...) la verdad no es expresada como ella misma sino como una realidad muy distinta. El empirismo vulgar tiene como su ley no su propio espíritu sino un espíritu extraño; por el contrario, la idea real tiene como existencia no una realidad desarrollada en ella misma, sino el empirismo vulgar” (Ídem). Según el filósofo Claude Lefort, el énfasis marxista en dicho momento de su crítica juvenil a Hegel, se basaría en argumentaciones de corte epistemológico. El artilugio hegeliano expresaría una recaída de corte empírico positivista e incluso, al formular Hegel como momento del arbitrio en la concreción de la soberanía del Estado, el rol del príncipe, Marx lo acusará de reemplazar el análisis filosófico por una “rara especie de zoología política”. De otro lado, estas críticas a la mediación establecida por Hegel entre las esferas de la familia y la sociedad civil respecto al Estado, no podrían extenderse hasta el punto de ocultar el papel activo que Hegel les concede a los ámbitos de la privaticidad, familia y del sistema de necesidades que constituye, en su teoría, el sustrato de la sociedad civil. En efecto, los extensos parágrafos (numerales 142 al 256) de la Filosofía del Derecho, e incluso la deducción de la propiedad hecha en la primera parte de su obra —todo ello consagrado al seguimiento dialéctico de la eticidad que finalmente y a la manera de síntesis expresa el Estado— son claramente demostrativos de las anteriores salvedades. Hegel, apegado a la estructura triádica de su dialéctica, exterioriza las fases que el espíritu cubre en los planos de la familia y la sociedad civil. En la primera de estas subdivide su tránsito en tres momentos, a saber: el matrimonio, el patrimonio y, finalmente, la educación de los hijos y la disolución de la familia. En la segunda esfera estos tres momentos son, respectivamente, el sistema de las necesidades, la administración de la justicia y el ámbito de la policía (die Polizei); y el poder de la corporación (reglamentación interna 99 de la sociedad civil). En tal dirección Marx concluye su comentario del parágrafo 262 anotando: “Lo real llega a ser fenoménico, pero la idea no tiene otro contenido que ese fenómeno. Tampoco la idea tiene otra finalidad que la finalidad lógica: ‘de ser para sí el espíritu real infinito’. En este parágrafo se encuentra formulado todo el misterio de la filosofía del derecho y de la filosofía hegeliana en general” (Ídem). La verdad del Estado, para Marx, en dicho momento de su crítica a Hegel, consiste en resumir en sí las esferas inmediatas de la sociedad civil y la familia. Resumir o subsumir equivale a superar, sin negar, la existencia de los momentos que integran la totalidad. El Estado político no es, por lo tanto, una simple derivación de la idea en su exterioridad histórica; es la síntesis, en un sentido materialista, de la familia y la sociedad civil, pues “el Estado político no puede existir sin la base natural de la familia y sin la base artificial de la sociedad civil” (Marx, Ob. Cit.). 5.4. LAS CONTRADICCIONES DEL FUNCIONARISMO HEGELIANO El problema de la burocracia no reviste solamente el aspecto señalado anteriormente según el cual la burocracia establece una mediación entre las clases de la sociedad civil y las elevadas esferas del poder del Estado, encarnadas en la voluntad suprema del monarca. Hegel además cree firmemente en aquello que luego el sociólogo Max Weber denominaría “racionalidad burocrática”. Veamos sus propias palabras: “La clase universal se ocupa de los intereses generales de la situación social. Debe por lo tanto ser relevada del trabajo directo para satisfacer sus necesidades, por disponer de un patrimonio privado o por medio de una indemnización del Estado, que absorba por otra parte su actividad de manera tal, que el interés privado encuentre su satisfacción en su trabajo por lo general” (Hegel, 1985). La división social del trabajo debe adaptar, conciliar, a la sociedad con los funcionarios: la elevada función que a estos les compete, encaminada a la representación y defensa del interés general, los releva del trabajo directo (intuición ésta que le sirve a Hegel para describir lo que más adelante Marx llamará trabajo productivo) y los consagra, por mediación de un pago del 100 Estado, de una retribución económica a la realización de funciones de carácter general; son responsables de la marcha concreta del proceso de dirección de la sociedad. Sin ellos la voluntad del monarca sería irrealizable. Hegel es consciente del papel que el Estado moderno debe asumir durante el largo y tortuoso tránsito de la vieja sociedad feudal con sus privilegios estamentarios, su cerrado corporativismo y su tenaz oposición a las actividades industriales, etc. —a la nueva sociedad burguesa—. En este punto podremos establecer una marcada diferenciación entre Hegel y el pensamiento liberal individualista. Mientras que para éste último, imbuido de la herencia lockeana, el Estado reviste el rasgo de un invitado forzoso cuya presencia debe evitarse en lo posible para que las fuerzas sociales y económicas de la sociedad obren con entera libertad, en el pensamiento político hegeliano se entrevé el rasgo ordenador del Estado. Sin embargo, esta misma concepción expresa un encubrimiento ideológico del verdadero rol de la burocracia, de donde resulta una mistificación del rol de los funcionarios del Estado. Así lo señala Hegel expresamente: “Las tareas de gobierno son de naturaleza objetiva; según su sustancia ya han sido decididas y deben ser realizadas por individuos. No hay entre ambas cosas ninguna unión natural inmediata, por lo que los individuos no están destinados a estas tareas por su personalidad natural o el nacimiento. Para su designación el momento objetivo lo constituye el conocimiento y la prueba de su capacidad, prueba que asegura al Estado el cumplimiento de sus necesidades y, al mismo tiempo, por ser la única condición, asegura a todo ciudadano la posibilidad de integrar la clase universal” (Ibíd.). Al referirse a esta argumentación hegeliana, Marx critica esta pretendida igualdad de todos los ciudadanos para acceder a la categoría de los funcionarios señalando: “La posibilidad ofrecida a todos de adquirir el derecho a participar en otra esfera, sólo demuestra que su propia esfera no es la realidad de este derecho...” Y agrega sobre los llamados exámenes de Estado: 101 “En un Estado racional, más bien se necesita un examen para ser zapatero, es un arte sin el cual se puede ser buen ciudadano, un hombre social; pero la ciencia administrativa necesaria es una condición sin la cual se vive en el Estado fuera del Estado, seccionado de sí mismo, separado del aire. El examen sólo es una formalidad masónica, el reconocimiento legal del saber cívico como privilegio” (Marx, Ob. Cit.). Además, se pregunta ¿qué garantiza, en efecto, que sea representado el interés general por estos funcionarios, y no, por el contrario, que estos suplanten dicho interés por su propia esfera particular? Hegel, consciente de esta objeción, intenta una solución. Comienza explicando que: “Consideradas como órgano mediador, las clases están colocadas entre el gobierno en general, por una parte, y el pueblo descompuesto en esferas particulares y en individuos, por otra parte. Su papel les impone el sentido y la mentalidad del Estado y del gobierno, así como el sentido y la mentalidad de los intereses de las esferas particulares y de los individuos. Al mismo tiempo esta posición tiene el significado de una mediación común con el poder gubernativo orgánico para que el poder soberano no parezca aislado como un extremo y por consiguiente, como una simple autoridad soberana y un arbitrio, para que los intereses particulares de las comunidades, corporaciones e individuos no están aislados ni, sobre todo, que los individuos lleguen a representar una multitud y una masa, siendo la expresión inorgánica, por consecuencia, del pensamiento y de la voluntad y un simple poder compacto frente al estado orgánico” (Hegel, ídem). Hegel supone, que las clases constituyen una indispensable mediación orgánica que cataliza los numerosos conflictos de intereses particulares. Pero en su modelo, no debe dejarse de lado este punto, la burocracia pública resulta una superior garantía de resolución de los conflictos privados. De allí la insistencia hegeliana en la profesionalización del funcionariado y su remarcada insistencia en la admisión mediante los señalados exámenes del Estado. Vista retrospectivamente la señalada objeción de Marx a la realización de exámenes de Estado, esta carece de fundamento. Desde luego, el hilo de su argumentación involucra un aspecto que hoy en día resulta una petición 102 de principio para la democracia política contemporánea. La ciudadanía es un derecho que se otorga por el sólo hecho de la naturalidad, valga decir, obedece a un procedimiento, en cierta medida, automático de acuerdo con las diferentes legislaciones nacionales. En numerosos países el simple hecho de nacer en el territorio, otorga los derechos ciudadanos como atributos inherentes a las personas. En cambio, en otras sociedades, la ciudadanía implica una relación demostrada de consanguinidad. Recientemente hemos visto que, con el despliegue de la xenofobia contra el tercer mundo y las emigraciones de los países subdesarrollados, muchas naciones europeas han restringido, de facto, la adquisición automática de la ciudadanía a través de la demostración del nacimiento en el territorio. Empero, el análisis de Marx apunta en realidad a otra dirección: se supone que ser ciudadano no requiere capacitación; mientras que, de otro lado, la funcionalidad de una profesión si lo requeriría. Sin embargo, la ciencia administrativa —que en su época era embrionaria y articulada principalmente al derecho público— tendría posteriormente un despliegue y una especialización significativa. Max Weber mostraría claramente que una de las virtudes inherentes a la burocracia es su especialización profesional, o sea, la de poseer un saber especializado que habilita al burócrata para participar de la función pública. En este orden de ideas hoy se reconoce como un lugar común la necesidad de la profesionalización para el acceso a los distintos cargos públicos y, por lo tanto, los exámenes de Estado significan un principio igualitario de acceso a la carrera pública. Por tal razón la objeción de Marx resulta endeble. La historia posterior ha demostrado la importancia de la carrera administrativa, basada en los principios de igualdad de oportunidades, demostración de habilidades y pericias, y finalmente, comprobación de la sujeción de los funcionarios a una ética pública del servicio. Marx, de otro lado, afirma que Hegel iguala e identifica Estado con gobierno, y considera como una masa iniforme al pueblo; de tal modo que la mediación que le es asignada a las clases corresponde en realidad a las corporaciones; de allí que: 103 “Las clases son la contradicción planteada del Estado, y de la sociedad civil en el Estado. Al mismo tiempo son el reclamo de la solución de esta contradicción” (Marx, Ob. Cit.). En el fondo, en este momento de la discusión, tanto Hegel como Marx parten del principio de que: “En el verdadero Estado no se trata de la posibilidad que tiene cada ciudadano para acceder a la clase general considerada como Estado particular, sino de la capacidad del Estado general de ser verdaderamente general, es decir el Estado de todo ciudadano” (Marx, Ibíd.). Pero si ambos proclaman sujetarse a principios democráticos, en Marx, esto conlleva indudablemente al ahondamiento de su postura democrática burguesa hasta afirmar, en otro lugar de su obra juvenil, que la filosofía moderna: “(...) considera al Estado como un gran organismo en el cual ha de realizarse la libertad moral, jurídica y política, y en el que el ciudadano individual, al obedecer las leyes del Estado, obedece solamente las leyes naturales de su propia razón, de la razón humana” (Ibíd.). En Hegel esta premisa resulta ilusionista, nebulosa, en la medida, precisamente en que: “(...) parte de la hipótesis del Estado pseudogeneral, del Estado ilusoriamente general, de la generalidad particular permanente...” Y de allí que: “(...) la identidad que ha construido entre la sociedad civil y el Estado, es la identidad de dos ejércitos enemigos, donde todo soldado tiene ‘la posibilidad’ de llegar a ser, mediante la ‘deserción’, un miembro del ejército enemigo” (Marx, Ibíd.). Y en otra parte del mismo texto, apunta: “Hegel da la solución al misterio hallando la definición propia de las ‘clases’ en el hecho de que el propio entendimiento y la voluntad particular de la sociedad civil encuentran en ellas su existencia en referencia al Estado. La sociedad civil es referida al Estado. Así, 104 como los burócratas son los delegados del Estado ante la sociedad civil, las clases son los delegados de la sociedad civil ante el Estado. Son siempre compromisos entre dos voluntades opuestas” (Ibíd.). Hegel, puesto en aprietos para defender la positividad del rol de los funcionarios, recurre a una garantía. Esta garantía es la jerarquía. O dicho de una manera clara, rotunda: el autocontrol de la burocracia por sí misma; y en una instancia suprema, por el monarca, poseedor del derecho de revocatoria del mandamiento de los funcionarios en sus más elevados e importantes niveles, y también, lo que no es menos importante, poseedor del derecho a modificar las estructuras legislativas y gubernativas. La presunción de incorruptibilidad de los funcionarios —problema abordado de manera “práctica” por la Revolución Francesa (recuérdense algunos de los desarrollos del terror jacobino)— posee en la estructura del pensamiento político de Hegel el carácter de una petición de principio. Lo que hay detrás es una contradicción insoluble entre la realidad de la esfera gubernativa y la idealidad dibujada, o más bien, imaginada por Hegel. El argumento de peso que nos presenta para favorecer el autocontrol de la burocracia es muy débil: “El servicio del Estado exige (...) el sacrificio de la arbitraria y personal satisfacción de fines subjetivos y da, justamente por ello, el derecho de encontrarla en el cumplimiento del deber, pero sólo en él. En esto reside, desde esta perspectiva, la unión del interés general y particular que constituye la solidez interna del Estado” (Hegel, ídem). Hablando en general de la concepción hegeliana del control jerárquico y del adicional control que se supone establecen los estamentos y las corporaciones, Karl Marx señala: “Las corporaciones son el materialismo de la burocracia y la burocracia el espiritualismo de las corporaciones. La corporación es la burocracia de la sociedad civil; la burocracia es la corporación del Estado. Allí donde la burocracia es el principio nuevo, el interés general del Estado comienza a convertirse en un interés ‘aparte’ y, por consiguiente, un interés ‘real’; ella 105 lucha contra las corporaciones, como toda consecuencia lucha contra la existencia de sus supuestos. Por el contrario, desde que despertó la vida real del Estado y la sociedad civil se libera de las corporaciones, empujada por un instinto natural propio, la burocracia trata de restablecerlas; puesto que cuando decae el ‘Estado de la sociedad civil’ decae igualmente ‘la sociedad civil del Estado’. (Marx, Ob. Cit.). La burocracia es, nos sigue diciendo Marx, un tejido de ilusiones prácticas en el cual un formalismo estatal termina por ser la realidad del Estado: “El espíritu burocrático es un espíritu totalmente jesuítico, teológico. Los burócratas son los jesuitas del Estado y los teólogos del Estado. La burocracia es la “republique préte” (Ibíd.). Resulta sintomático que se compare a la burocracia gubernamental con el jesuitismo, en especial si recordamos que dentro de las congregaciones religiosas cristianas, los jesuitas significaron el sector que lideró las sagas empresariales en los procesos de colonización en América hispana. Tal crítica al burocratismo y su comparación con el espíritu institucional jesuístico, en términos de su ideología de fe, representa la subordinación de la sociedad civil a la burocracia, en tanto ésta asume la representación del supremo interés de la sociedad. La superioridad de la burocracia se explicita, en dicha perspectiva, por la condición que posee dentro de su organización y líneas de poder el secreto de Estado, que permite la exclusión práctica de los miembros de la sociedad civil respecto del conocimiento de numerosas facetas de los asuntos públicos. Además, Marx caracteriza a la burocracia como un espiritualismo de Estado que hace del secreto su principio básico: “El espíritu general de la burocracia es el secreto, el misterio guardado en su seno por la jerarquía y hacia afuera, por su carácter de corporación cerrada” (Ibíd.). De otro lado, es necesario recabar el citado sentido contradictorio que posee la jerarquía estatal; si Hegel ofrece una imprecación moral como condición necesaria y suficiente de la probidad inmanente al cuerpo de 106 funcionarios, de otro lado llega a los umbrales mismos de la solución democrática del asunto, a saber: la vigilancia sobre los funcionarios no sólo desde arriba sino desde abajo. En este sentido agrega en el texto citado anteriormente: “(...) la seguridad del Estado y de los gobernados contra el abuso del poder por parte de las autoridades y de los funcionarios radica, por una parte, inmediatamente en su jerarquía y responsabilidad; y por otra parte, en la legitimación de las comunas y corporaciones que refrena la intromisión del arbitrio subjetivo en el poder confiado a los funcionarios y completa desde abajo el insuficiente control que se ejerce desde arriba sobre la conducta individual” (Ibíd.). Cuando Hegel postula su teoría acerca del rol mediador de las clases, afirma enfáticamente que tal función catalizadora la realiza el conjunto de las corporaciones. De este modo éstas aparecerían como una sumatoria de instituciones a medio camino entre la organización gremial típica del feudalismo y el papel moderno, propio a la sociedad industrial, de los gremios y las asociaciones de empleados (sindicatos). Como se sabe, Rousseau había anticipado radicalmente su rechazo a la función mediadora del interés general por parte de este tipo de agrupaciones. Por ello justificó lógicamente su postulado de democracia directa, que tuvo hondas repercusiones en la fase final de la Revolución Francesa. La experiencia de 1789 —de la cual fue testigo y analista Hegel—, del mismo modo que las frecuentes revoluciones y sublevaciones populares en Europa continental e Inglaterra, corroboraron a la teoría política del siglo XIX las consecuencias prácticas de la puesta en escena de dicho proceso de control ciudadano. Hegel, como ya lo hemos advertido, es, en este orden de ideas, enemigo declarado del derecho a la rebelión. Por lo tanto, lo que aparece supuesto en su teoría, sería aquello que los constitucionalistas de la unión americana habían definido con claridad, a saber, que el control ciudadano se expresa fundamentalmente mediante institucionalidades que generan diáfanamente frenos y contrapesos del poder público. Ahora bien, según Marx, el lado flaco del control desde abajo que reclama Hegel para complementar el insuficiente control ejercido por 107 la institución de la jerarquía, estriba en que los funcionarios, en modo alguno, obedecen al mandato imperativo; pieza angular de la democracia revolucionaria implementada por los más radicales luchadores de la revolución burguesa en Francia y reencontrada por Marx en la comuna de París de 1871. Esta ambivalencia remite a otra de mayor complejidad: ¿cómo reconciliar —en un sistema que se proclama la armónica expresión de la razón— la monarquía con un control democrático de las instituciones y los funcionarios a ellas vinculados? Tal encrucijada motivaba a Marx a proclamar que la democracia es la verdad de la monarquía; pero, por el contrario, la monarquía no es la verdad de la democracia. En dicho momento de su elaboración teórica, para los izquierdistas hegelianos era una petición de principio el postular la oposición entre democracia y monarquía. En este sentido no resulta gratuita, como a veces muchos de sus analistas lo proponen, la acerba crítica que Marx tuvo frente al rol histórico de Bolívar, al cual comparó a una caricatura de Napoleón. Es la perspectiva eurocéntrica de lo que luego sería el epítome del dictador latinoamericano. ¿Qué era en realidad aquello que alejaba a Marx de Bolívar? En relación con el anterior tópico, corroboro la hipótesis sobre la imposibilidad, o insuficiencia, de los postulados democráticos e institucionales en el discurso y la praxis bolivariana propia de las sagas independentistas del siglo XIX. De otro lado hemos de reconocer que, ya en su tiempo y como los filósofos políticos franceses más clarividentes tales como Montesquieu, Voltaire y otros lo habían precisado, existía dentro de la teoría política europea una clara diferenciación entre el despotismo monárquico y los modelos monarquistas de tipo constitucional, con base principalmente en el referente de la Inglaterra postrevolucionaria. Sin embargo esta es una diferenciación muy importante que Marx en su crítica juvenil a Hegel no toma en consideración. Sin embargo, cabe recordar que el análisis hegeliano toma en consideración el eco que existe en la filosofía política ilustrada francesa del modelo inglés, en el cual se daba una articulación funcional entre la monarquía constitucional y la democracia censitaria que expresaba el bicameralismo. Por ello puede afirmarse que el democratismo de los textos juveniles de Marx es bastante radical27. Téngase en cuenta que en la primera mitad del 108 siglo XIX, el republicanismo liberal dominante hacía suyo el modelo monárquico constitucional en la mayor parte de los países europeos. Para esa misma época, en el contexto latinoamericano, las revoluciones de independencia, pese a contar con una participación decisoria del pueblo (principalmente del ejército libertador como pueblo en armas), se resistían a generar instituciones republicanas de corte democrático. No es gratuito por ello que haya tenido tanto influjo el presidencialismo bolivariano, al cual complementaba el senado vitalicio y hereditario, elementos simbólicos del peso del caudillismo estamentario (cfr. E. Varela: 1990. Véase además el capítulo IV de la presente obra). Ciertamente Marx en ninguna parte dejó una definición expresa de sociedad civil; sólo anotó que la anatomía de la sociedad civil se encontraba en la economía política. Así, en su perspectiva, los temas de la política, el problema del Estado, la vida partidista y la lucha de clases encuentran su explicación a partir del análisis económico y, por lo tanto, en el estudio de la cultura material; objetividad que dota de sentido a una sociedad, incluyendo, no faltaba más, los campos del arte. Marx tuvo una concepción de la sociedad civil muy cercana a la de Hegel en el sentido de entenderla como ámbito de la economía política. Pero a diferencia del Hegel, no consideró que fuera el Estado quien crease la sociedad civil, sino la sociedad civil quien le otorgaría contenido al Estado. La idea que subyace a los ilustrados románticos y al propio Hegel, establece un supuesto implícito: vislumbrar la familia como parte o principio de la sociedad civil, la cual constituye realmente el crecimiento de la misma como pueblo, como nación y, por consiguiente, tiene un origen natural común. El concepto de comunidad involucra un vínculo de pertenencia religioso, étnico, cultural y consanguíneo. El individuo no se concibe en términos de comunidad, como alguien separado de los demás; por el contrario, el individuo sólo existe en la medida en que tenga un sentido de pertenencia claro con un grupo étnico, religioso, con una territorialidad. Sin embargo, el mercado rompe con las pretensiones del jusnaturalismo puesto que abre, libera la propiedad y, de este modo, desata los vínculos y las solidaridades culturales. Por ello es claro que a la sociedad tradicional la disuelve el mercado, teniendo en cuenta que allí reside la diferenciación entre comunidad y sociedad civil. En síntesis, la teoría marxista sobre el Estado y el funcionarismo público tuvo dos períodos claramente diferenciados. Su fase juvenil, que corresponde las 109 consideraciones críticas sobre la filosofía política hegeliana, ya reseñadas; y el período posterior a 1847, cuando arranca su formalización madura. Su actitud frente a la democracia se modificó, por consiguiente, de forma profunda y significativa. En esencia, el aporte relevante de su crítica juvenil a Hegel consistió en su reivindicación del ámbito de la denominada sociedad civil. No se necesita mayor perspicacia para correlacionar dicha visión con los predicamentos del liberalismo radical, v gr. J. Stuart Mill y A. de Tocqueville entre otros. Desde luego, no pretendo equivaler el democratismo juvenil marxista con las versiones clásicas del liberalismo individualista. Más bien, de todo lo hasta aquí analizado, se desprende su concepción de lo político como derivada de la soberanía popular, siendo visibles en dichos postulados ecos del republicanismo de estirpe rousseauniana. 6. La concepción marxista del Estado y de la burocracia pública Las experiencias del auge revolucionario del medio siglo y el surgimiento de la clase trabajadora urbana como una fuerza política que le disputaba el poder, tanto a la aristocracia del suelo como a la burguesía industrial y mercantil, catalizó la revisión del paradigma juvenil marxista sobre la democracia. Le permitió, a partir de la teorización funcional de Hegel sobre las clases sociales, con base en la apropiación crítica del discurso de la economía política, entender el carácter conflictivo de la democracia. En tal dirección Marx no opondrá más la sociedad civil al Estado sino que destacará las profundas fisuras de las relaciones de propiedad y acumulación de riquezas, y el carácter instrumental del poder público. El Estado no es, pues, un árbitro imparcial, sino la expresión de un dominio social que le permite al bloque de clases hegemónico sustentar, consolidar, la subordinación material y espiritual del resto de la sociedad. En la perspectiva marxista, por lo tanto, antes de analizar el Estado, de ocuparse de la teorización de las esferas de lo político como un campo autónomo, debe estudiarse la economía y el proceso de producción y reproducción material y espiritual que en el fondo, ineluctablemente, en la sociedad industrial constituyen el suelo del que emerge la llamada sociedad civil. 110 En función de tal cosmovisión Marx no se preocupó por construir argumentos para debatir los temas del contractualismo, porque no quiso ocuparse de la filosofía, ya que sólo escribió textos filosóficos en su fase juvenil, puesto que creía resolver el reto de estudiar científicamente la sociedad sobre la base del análisis socioeconómico de la formación capitalista, para luego dar cuenta del universo de lo político. 6 . 1 . LA CRÍTICA MARXISTA AL ARBITRAJE ESTATAL DEL INTERÉSGENERAL La idea común que nos hacemos de las tesis del marxismo acerca del Estado se desprende de la lectura de la obra de Engels: El origen de la familia, la propiedad privada y el Estado, escrita con base en unos borradores y cuadernos de lectura de Marx sobre las investigaciones del antropólogo norteamericano Lewis Morgan. Según las teorías de estirpe marxista, el Estado es un instrumento de dominación de clase. Esto significa que existe como componente esencial del Estado, de un lado, un aparato ideológico de dominación y, de otro lado, un aparato coercitivo de dominación. Al mismo tiempo el marxismo plantea que la principal característica del Estado es la prevalencia dentro de él de este segundo aspecto. Lo anterior no significa que se nieguen otros de sus componentes como la burocracia y el funcionarismo, o su ingerencia creciente en la actividad productiva de la sociedad en la época del capitalismo monopolista. Sin reducir el Estado solamente a este cuerpo coercitivo, la teoría marxista lo señala como su aspecto principal. Esta sería la madura conceptualización funcional o instrumental que el marxismo opondría a las teorías calificadas como liberales, donde se ocultaría el estrecho vínculo entre Estado y dominación de clase. Anotemos adicionalmente que K. Marx, frente al argumento contractualista como base de la sociedad política, apunta su crítica en otra dirección. Según Marx, fundar el Estado político sobre el individuo “libre” es la expresión sublimada, característica del pensamiento del siglo XVIII, según la cual el principio del análisis no reside en la relación social de los individuos, sino en su “aislamiento” artificial, en su consideración como sujetos “naturales” con rasgos que eternizan las condiciones específicas de producción social propias del capitalismo. Al igual que sucedía en el caso de Hobbes, este in111 dividualismo tipificó una forma representativa para su época de expresión de la sociedad civil. De tal forma, que la manera como Rousseau concibió al individuo y como presentó el origen de la sociedad política, fue no sólo especulativa sino además irreal. La verdadera respuesta a estos interrogantes, según Marx, estaba en manos de la economía política y no en la filosofía social del siglo XVIII28. Este camino llevaría a la conocida elaboración marxista en relación con el origen de la familia y la sociedad. El paso de la barbarie y el salvajismo a las primeras formas de civilización, las grandes formas históricas de división social del trabajo, el surgimiento de las primeras formas de excedente social y la correspondiente esfera de distribución a él vinculadas etc., son temas abordados no en una perspectiva filosófica sino histórico-antropológica. Lo que Marx reconocería primordialmente en la obra rousseauniana es su anticipación en el terreno político de la problemática de la desigualdad. Punto que tiene que ver, precisamente, con la cuestión señalada por Lenin en torno a las fuentes constitutivas del marxismo y, en particular, a la manera como Rousseau prefiguraría el “socialismo utópico”; lo cual —según la citada interpretación leninista— constituiría una de las vertientes del proceso formativo del pensamiento marxista. Galvano Della Volpe, una de las figuras teóricas del moderno pensamiento socialdemócrata, intentó una lectura del enlace existente entre Marx y la tradición política de los siglos XVII y XVIII; principalmente haciendo referencia al legado teórico-político de Rousseau. Según esta interpretación la tradicional visión que se ha tenido de Rousseau, como filósofo que no supera los límites de la utopía en la concepción de las estructuras formales de la sociedad civil que postula una irrealizable comunidad de hombres iguales acordados en torno a un pacto contractual, es completamente equívoca. En Rousseau se encuentra mucho más: se trata de concebir el contenido de la obra rousseauniana en relación con la posibilidad de extender la validez histórica, de defender la realidad y la vigencia de la libertad política, de darle un contenido actual a sus postulados igualitarios29: de tal manera que la interpretación de la tesis de Marx que plantea como corolario del socialismo inferior “la satisfacción del consumo y las necesidades según el trabajo”, y la premisa de su etapa superior (el comunismo entendido como sistema social de producción): 112 “Dé cada quien según su trabajo, a cada quien según sus necesidades”, desarrollada por Marx en la Crítica al programa de Gotha y por Lenin en el Estado y la Revolución, está destinada a representar la satisfacción histórica de la instancia rousseauniana del mérito personal” (Ibíd.). Della Volpe (1975), plantea una enumeración de problemas que se desprenden de la tesis arriba anotada, a saber: los alcances teóricos y político-históricos de la obra rousseauniana; la renovación de la herencia democrático-burguesa en la legalidad democrática socialista; la correspondiente estructura del Estado socialista; la “destrucción” del Estado; la naturaleza de la sociedad comunista, etc. Ello implica una recusación vigorosa de dos interpretaciones de Rousseau. La primera, la lectura socialdemócrata que enfatiza el aspecto de la igualdad política (la libertad civil) en la medida en que se inscribe dentro del afán rousseauniano por incorporar como sustancia legítima de la comunidad el derecho inalienable de cada uno de sus miembros. Allí, sólo se trataría de tomar como punto de partida la libertad individual, la igualdad formal de los ciudadanos frente a la ley. La segunda lectura tiene —según Della Volpe— su paternidad en el marxismo ortodoxo. Rousseau aparece ante tal vertiente como un pensador de la pequeña burguesía, alegrándola con el ilusorio ideal de una república de pequeños propietarios que tendría como divisa la eliminación de las distancias en el ingreso. Para Della Volpe, la importancia de Rousseau sobrepasa la ubicación de su aporte en los límites de una formulación política y teórica que sirvió de guía en las luchas del ala radical de la burguesía media francesa en el siglo XVIII. Es verdad que su influjo directo fue muy grande en el proceso de 1789, principalmente en la etapa anterior al Termidor, cuando los jacobinos, desde el gobierno, trataron de llevar a la realidad inmediata del momento algunos de los más importantes postulados planteados en el Contrato (así ocurrió con la idea expresada por Rousseau de conformar una religión civil, con el mecanismo de la democracia directa, el papel de la soberanía, etc.). Pero su verdadera significación estuvo por encima del alcance y la comprensión de sus contemporáneos. Rousseau —insiste Della Volpe— es el padre de la democracia moderna. Y ello porque la consecuencia de su posición supera el liberalismo clásico de Locke. Mientras éste se limita a defender al indivi113 duo como eje de la moderna sociedad civil haciéndolo partícipe de una cesión de poderes al gobernante, es decir, justificando la monarquía constitucional y estableciendo la separación entre el citoyen y el burgués, Rousseau supera esta antinomia y proclama el carácter igualitario de los hombres. Su argumento esencial se apoya en el deber de legitimar la libertad. El soberano garantiza la seguridad de los asociados sin desmedro de su libertad; y, alternativamente, la libertad civil encuentra su límite en los derechos y libertades recíprocas de los miembros de la comunidad. En el fondo de esa argumentación subyace la discusión acerca de las relaciones entre la libertad individual y la obediencia a las normas colectivas que prescribe la ley. Locke representaría una forma de liberalismo hipostasiado, en la cual el Estado no juega un rol moral de dirección ética de la sociedad, sino un papel de garante colectivo del inalienable derecho de propiedad, lo que en este punto representa la antítesis del postulado rousseauniano. En efecto, la concepción rousseauniana, en la perspectiva del análisis de Della Volpe, garantiza a un mismo tiempo la seguridad colectiva y el respeto a los derechos inalienables de los ciudadanos. El derecho de vida y muerte es competencia del Estado (es decir, de los ciudadanos organizados como comunidad) y está subordinado a la supervivencia del mismo. Por tal razón, es loable perecer en defensa de la patria, dar la vida por la defensa y la salvaguarda de la libertad30. De otro lado, según Della Volpe, el democratismo rousseauniano no ha perecido en la ambivalente relación entre la realidad efectiva con sus privilegios y la persistencia del antagonismo social y el ilusorio terreno del derecho “igual”. El contenido democrático del mensaje rousseauniano va más allá. El mérito personal, subrayado enérgicamente en su obra —y del cual su propia vida fue un ejemplo dramático— debe traducirse en la búsqueda del “desarrollo libre de las potencialidades del individuo”, tal como lo señaló Marx en los Manuscritos económico-filosóficos del 44 (Ibíd.: 38). De esta manera se postula un enlace entre Marx y el legado de Rousseau31. Satisfacer la igualdad política (o libertad civil) con el complemento indispensable de la libertad “igualitaria” (o igualdad económica) en un proceso que subsume y supera dialécticamente a la primera en la segunda, es, desde el punto de vista de Della Volpe, el mérito del verdadero socialismo (democracia) frente a la 114 pretensión ideal del liberalismo. Y en este postulado reside, a su juicio, el valor intrínseco del marxismo: su realización efectiva de la libertad. Ciertamente, según Marx, la transformación de las condiciones económicas en las cuales la sociedad produce y reproduce las condiciones materiales para su existencia traen consigo, de manera “fatal”, un cambio cualitativo en el contenido y en la forma jurídica que asume la igualdad económica y política. Al respecto de las relaciones entre lo económico y lo político, recordemos la tesis de un teórico ruso de la década del veinte del siglo pasado, Isaac Rubin, quien en un interesante estudio sobre la teoría marxista del valor señalaba que la diferencia capital entre la economía política marxista y la burguesa residía en que la primera desdeña su igualación con el simple nivel de análisis correspondiente al desarrollo técnico de las fuerzas productivas. En tal dirección subrayaba que la economía política se caracteriza por esclarecer la naturaleza de las relaciones sociales de producción y de cambio. El problema de la burocracia, en este orden de ideas, es por tanto, sólo parcialmente un problema técnico32. Aunque la tesis de Rubin resulta sugerente en la medida en que subraya el carácter político de la división social del trabajo, de la cual la burocracia pública es un subsistema, no comparto su pretensión de subsumir la genealogía de lo político en el campo de las meras relaciones económicas. Como se sabe, dicha formulación ha sido muchas veces criticada como una de las insuficiencias cruciales de la teoría social de estirpe marxista. Además, la propia esfera de la política tiene un conjunto de estructuras funcionales basadas en una tecnología del poder (M. Foucault, 2004a) que se expresa parcialmente a través de formas de burocratización. Éstas se dan en la acción partidista, la lucha política y en la arena parlamentaria, y se sienten fuertemente sus influjos en las dinámicas asociativas de las sociedades civiles, sindicatos y demás aparatos subalternos de expresión y movilización social incluidos. En un texto juvenil, Marx destacaba ya el carácter contradictorio, irrealizable, de este “deber ser” expuesto por la filosofía política. Así, ésta ve en las relaciones sociales, antes que un elemento positivo, un factor de restricción de la libertad del individuo aislado. Dicotomía que erige a las necesidades y las articulaciones del hombre en su inserción social como obstáculos de la libertad: “La libertad –tal como la conciben los derechos del hombre de 115 1791– consiste en que cada persona tiene el derecho de hacer lo que quiera en tanto no se afecten los derechos de los demás. Los límites bajo los cuales cada quien puede moverse sin ir en detrimento de los derechos de otro son marcados por la ley, de igual forma a como los límites entre dos terrenos están marcados por un mojón. Se trata, en consecuencia, de la libertad del hombre considerado como una mónada aislada, replegado sobre sí mismo. (...) pero el derecho del hombre, su libertad, no reposa de este modo en las relaciones del hombre con el hombre, sino ante todo, en la separación del hombre con respecto al hombre. Es así la libertad el derecho a esta separación, el derecho del individuo limitado sobre sí mismo. (...) es esta libertad individual con sus expresiones prácticas la base de la sociedad burguesa. Ella le hace ver a cada hombre en los demás hombres no la realización sino la limitación de su propia libertad”. Más adelante en la Introducción a la crítica de la economía política (Grundrisse), Marx señala cómo este pensamiento que concibe al hombre como mónada aislada va en contravía de las condiciones objetivas de la época; se halla desfasado en sus propias circunstancias históricas: “La época que engendra este punto de vista es precisamente la época de las relaciones sociales más desarrolladas hasta el momento (y desde este punto de vista, generales). El ser humano es, en el sentido más literal del término, un animal político; no sólo un animal social, sino además un animal que sólo se puede aislar en sociedad”. Esta expresión es una constante en el pensamiento de Marx poseyendo incluso una elaboración nítida en su crítica de 1844 a La filosofía del derecho estatal de Hegel: el hombre no es algo abstracto; quien dice ‘el hombre’, dice el mundo del hombre: Estado, sociedad”. La abstracción del término hombre ha de fijar, por lo tanto, no un semblante nebuloso o la mistificación de lo humano que constituye un rasgo tan característico de la religiosidad, como lo acababa de explicitar Ludwig Feuerbach. Al contrario, con esta expresión habría de subrayarse que el hombre se realiza, como tal, en su relación con los demás hombres. Por ello, en relación a su concepción del Estado afirmó, 116 que “los asuntos estatales, etc., no son más que los modos de existencia y de actividad de las cualidades sociales de los hombres” (Ibíd). 6.2. LA UTOPÍA MARXISTA DEL ESTADO REVOLUCIONARIO DE TRANSICIÓN Y DE LA SOCIEDAD SIN CLASES A pesar de no establecer aún las determinaciones específicas del Estado y de la sociedad civil de la forma que lo haría en su madurez, Marx, en su crítica a Hegel, dio claras muestras de su visión peculiar acerca de la relación existente entre la individualidad y el marco de las relaciones sociales, de manera opuesta a la “libertad individual, base de la sociedad burguesa”. Marx señala que la distinción entre el hombre aislado ‘como un todo perfecto y solitario’ que trastoca esta condición original, de orden natural, por su inserción al todo social para así devenir en una existencia parcial y moral —lo que antes era existencia física independiente (Rousseau, Contrat Social)— es una expresión general, brillante para el siglo XVIII, de la abstracción del hombre político, o citoyen, que caracteriza a dicho pensamiento político. Para luego señalar a renglón seguido que en la distinción entre el ciudadano y el hombre burgués, la segunda expresión designa la verdadera sustancialidad de la sociedad civil: expresa la negación de la realidad efectiva de la primera. De allí que Marx postule su famosa aserción: “La emancipación del judío es la emancipación de la sociedad del judaísmo”. Marx insistió en la necesidad de dotar a la corriente revolucionaria obrera de un programa no limitado al alcance inmediato de reformas susceptibles de ser logradas en los marcos del capitalismo. Y vislumbra un anticipo de este programa de transformación radical en la experiencia en la Comuna de París (1871). Marx enumera las principales medidas adoptadas por los comuneros, destacando, en primer término, la relacionada con el principal componente del Estado: las instituciones coercitivas, el ejército y la policía. Esta última fue despojada por la Comuna de sus atributos políticos y convertida en un organismo dependiente, al que se le dio un carácter revocable. La milicia popular no era una invención de la Comuna de París. En numerosas revoluciones burguesas estas fuerzas, que representan a la población levantada en armas, jugaron un papel de primer orden, un rol protagónico. En América Latina, por ejemplo, durante el proceso independentista contra la Corona Española, la base de las fuerzas patrióticas 117 fueron estas milicias populares que equivalían a lo descrito por Marx en el París de 1871: la población armada espontáneamente, en condiciones en las que el control ‘desde arriba’ de dichas fuerzas es dificultoso, por no decir imposible. Marx, al hacer referencia a la disolución del ejército y la constitución —en su reemplazo— de una milicia que sea el mismo pueblo armado, parte de la base de que estas nuevas clases populares están sometidas a la dirección del proletariado. La Comuna sería una nueva forma política de estructuración del aparato coercitivo encaminada a devolver las esferas del Estado del cielo inalcanzable y prepotente del funcionarismo, hasta el firme terreno del control, la revocabilidad y la vigilancia por parte de los ciudadanos. Marx considera al Estado un ‘mal necesario’, enormemente costoso para la sociedad y, en especial, para sus capas productivas, cuya desaparición, por consiguiente, aliviaría el pesado fardo de la burocracia y el militarismo consustanciales al Estado ‘parásito’. Sin embargo no compartió el anarquismo que imprecaba por la lucha contra toda forma de autoridad y proclamaba como el summum del revolucionarismo el oponerse a las diversas formas de ejercicio del poder político; de allí el lema anarquista: labor revolucionaria es aquella encaminada a la destrucción del Estado. Marx pensó en la posibilidad de que el Estado de dictadura del proletariado eliminara en lo fundamental el rasgo nocivo que le corresponde en todo Estado a la burocracia y al ejército. Esto lo podemos ver con claridad en el curso de la exposición citada: “La Comuna convirtió en una realidad este tópico de todas las revoluciones burguesas, que es ‘un gobierno barato’ al destruir las dos grandes fuentes de gastos: el ejército permanente y la burocracia del Estado. (...) El régimen de la Comuna habría devuelto al organismo social todas las fuerzas que hasta entonces venían absorbiendo el Estado parásito que se nutre a expensas de la sociedad y entorpece su libre movimiento. Con este sólo hecho habría iniciado la regeneración de Francia” (Ob. Cit.). En segundo término, la Comuna no había de ser un organismo parlamentario, sino una corporación del trabajo ejecutiva y legislativa al mismo tiempo (C. Marx, Ibíd.: 499). Sería, por lo tanto, sustancialmente diferente en su 118 funcionamiento de las tradicionales repúblicas democrático-representativas. La separación entre los tres órganos del poder público sería reemplazada por una articulación, —en el seno del nuevo Estado— de las funciones ejecutivas y legislativas de tal modo que los representantes del pueblo legislasen y gobernasen al unísono. Marx proclamó, como norma básica de conformación de los órganos del poder público, la sujeción en su proceso de conformación al mandato imperativo (la delegación popular revocable como la llama en el texto citado), lo que diferenciaría esencialmente este sistema de representación de la democracia burguesa donde el elegido no responde de sus actos ante el elector. Iring Fetscher, en su texto Marxismo y burocracia (1965), dice que Marx sometió a una crítica demoledora a la burocracia (entendiendo dentro de la burocracia, principalmente a los funcionarios ejecutivos del Estado, v. gr., los ministros y el cuerpo parlamentario y demás miembros de las instituciones legislativas, judiciales y gubernativas). Kautsky afirmaba: “Hoy sólo puede afirmar un ciego político que el sistema representativo asegura, bajo el imperio del derecho electoral general, el poder de la burguesía; y que para derribar a ésta, habría que eliminar primeramente el sistema representativo. Ya ahora empieza a hacerse evidente que un verdadero régimen parlamentario puede ser igualmente el instrumento de la dictadura al proletariado, así como es en la actualidad un instrumento de la burguesía. No la eliminación del sistema representativo sino la destrucción del poder del gobierno frente al parlamento, la facilitación de una vía amplia hacia éste al proletariado, mediante la igualdad del derecho electoral, la homogénea división de los distritos electorales, el respeto al voto secreto, parlamentos cortos, total libertad de prensa, de y asamblea y sobre todo, la extensión del derecho electoral a todos los ciudadanos que han cumplido 20 años, es una de las tareas principales de la clase trabajadora en su lucha por el poder político”33. Como se sabe, la socialdemocracia alemana, bajo el influjo ideológico de Eduard Bernstein y de Karl Kautsky, estableció notables diferencias frente a la concepción marxista de la democracia. En efecto, los socialdemócratas alemanes concibieron el parlamentarismo desde una perspectiva sustancial 119 que superaba el utilitarismo político de los socialistas radicales. Según tal utilitarismo, habría que utilizar los canales democráticos para acumular fuerza, a la manera de tribunas de opinión, pero sin reconocer su carácter sustancial. Empero, certificando tales ideas, la burocracia pública no podría ser encasillada en un corpus único, sino que deberían considerarse las funcionalidades institucionales divergentes y/o paralelas entre las distintas ramas del poder público. La función de representación política, adscrita a las colegiaturas de elección popular dotaba de una legitimidad superior a esta modalidad de funcionarismo público. El tercer campo de experiencia de la Comuna de París relacionado además de manera directa con la cuestión del centralismo democrático como principio organizativo del Estado, fue la relación entre el poder central y los poderes locales34. La idea que Marx traza de la nueva forma de la unidad nacional en Francia —sobre la base de la organización comunera extendida a todo el país— y su afirmación acerca de las limitadas funciones que le corresponderían en este orden de cosas al Estado central, fue, por decir lo menos, audaz. Obsérvese que Marx contrapone la realización de la unidad de la nación a la existencia del Estado. La afirmación de su realidad se asegura en la medida en que se destruye el poder estatal. Marx efectivamente pensó en la posibilidad de cristalizar en un futuro inmediato la disminución de las funciones de las autoridades centrales y propulsar la ampliación de la autonomía de las formas de gestión popular. La Comuna, en este sentido, le pareció la forma más adecuada para llevar a la realidad estos propósitos. Por esta razón nos habla de que el poder central poseería, bajo este sistema político, ‘pocas pero importantes funciones’. Marx rechazó desde un comienzo las acusaciones contra él y contra la experiencia de la Comuna por parte de aquellos que veían en dicho experimento un paso hacia el federalismo y una copia que hacía tabulares de las comunas revolucionarias que Francia había conocido en el pasado, en el curso de las revoluciones burguesas35. Marx —como ya lo dijimos— finca sus esperanzas en el hecho de que la Comuna pueda desembarazarse del ominoso centralismo característico de los Estados políticos conocidos hasta entonces por la Historia. 120 Esta conceptualización muestra que la visión propiamente marxista sobre el Estado socialista del futuro distó muchísimo de las reformulaciones hechas por Lenin y sus continuadores. Marx pensó que muy rápidamente podría disolverse el papel centralizador y omnímodo del Estado mediante mecanismos de cesión al pueblo (autogestión, descentralización, democracia directa, mandato imperativo, etc.). Paralelamente, el otro punto que reviste capital importancia en la conceptualización fue la reivindicación programática elaborada en Gotha con respecto a la prédica a favor del “Estado libre”. Marx señala que dicha reivindicación es, para comenzar, una contradicción en los términos. Así, el Estado popular libre se ha convertido en el Estado libre. Gramaticalmente hablando, Estado libre es un Estado que es libre respecto a sus ciudadanos, es decir, un Estado con un gobierno despótico (Ob. Cit.). Marx se refiere al Estado al tratar lo que él mismo denomina “la parte democrática” del programa de Gotha. Comienza por indicar qué se entiende por ‘Estado libre’. Su respuesta es célebre y ha sido causa de más de una polémica en virtud de su peculiar posición acerca de la libertad. En efecto, Marx estipula: “La libertad consiste en convertir al Estado de órgano, que está por encima de la sociedad, en un órgano completamente subordinado a ella, y las formas del Estado siguen siendo hoy más o menos libres en la medida en que limitan la ‘libertad del Estado’ ”. La primera parte la tesis contrapone la libertad del Estado a la libertad de la sociedad civil. Entre más libre sea un Estado, mayor será el despotismo del gobierno. Un vivo ejemplo de lo que Marx comprendía por libertad del Estado lo observamos en su análisis del llamado bonapartismo, forma europea clásica del despotismo estatal. Igualmente, como lo analizaremos en el tercer capítulo de este libro, Marx da cuenta de la variante oriental de libertad del Estado que caracteriza el modo de producción asiático donde el déspota, identificado con el Estado, es propietario de la tierra, único o principal medio de producción y donde la ley equivale a su voluntad. La disminución de la libertad del Estado significa, por consiguiente, en toda circunstancia, y principalmente en el marco de la sociedad burguesa, una 121 disminución de las prerrogativas del Estado en favor de la esfera de la sociedad civil. 6.3. EVALUACIÓN RETROSPECTIVA DE LA CONCEPCIÓN MARXISTA SOBRE LA BUROCRACIA Lo primero que quiero subrayar —como lo anoté de pasada— es que vale la pena tomar consciencia de que el debate Marx-Hegel sobre la burocracia pública es realmente el punto de partida de la teorización sobre la burocracia, tanto en la perspectiva que legitima su función de mediación y arbitraje del interés general (Hegel) como en su crítica y desmitificación (Marx, en su escritos juveniles). De tal manera que en la primera mitad del siglo XIX, tras los ecos de la ilustración alemana, surgen dos tempranos paradigmas. El primero de ellos, altamente crítico respecto de la naturaleza negativa de la burocracia. Mientras que el segundo, que podríamos calificar de ‘realista’, consideraba a las organizaciones burocráticas como una emanación institucional inevitable, fruto de la gran organización industrial del desarrollo del Estado corporativo, y producto adicional de sus evidentes conexiones y ramificaciones, más allá de los límites de la soberanía estatal nacional. Marx fue, en sus escritos de juventud, uno de los primeros críticos de la burocracia y de sus pretensiones de funcionalidad universalista. Y ello pese al enorme despliegue que el modelo burocrático tendría en el siglo XX en los países, encabezados por Rusia y China, que desarrollaron poderosas revoluciones socialistas inspiradas en su ideario. Hegel es por lo anterior fácilmente reconocible como el gran autor fundacional de la positividad de la burocracia estatal en términos de que ésta encarna la racionalidad teleológica del Estado moderno constitucional. Como todos lo sabemos, Marx, pese a su enorme influjo sobre el proceso político de la Modernidad, no fue en sí mismo un activista político. Sus participaciones en la llamada acción directa fueron significativas mas episódicas y operaba, a la larga, más como una suerte de gran consejero del movimiento revolucionario de su época. Así, la posibilidad de concretar la utopía de la sociedad sin clases y de la extinción del Estado, a la que acabamos de referirnos, fue cuestión práctica para las siguientes generaciones de teóricos del socialismo radical. En especial, para Lenin y sus contemporáneos, Plejanov, Kautsky, R. 122 Luxemburgo, Kamenev, Martov, y Trostky; y para Stalin, que consolidó los trazos totalitarios o nacionalistas de modelo. Y, en fin, para una serie de grandes líderes revolucionarios de los países que después se denominarían eufemísticamente como el “Tercer Mundo”, líderes estos de los que el más destacable sin duda sea Mao Tse tung. En este apartado, de una forma concisa pero espero que adecuada, dentro del hilo argumentativo de este capítulo y teniendo en cuenta los objetivos que me he trazado al escribir este libro sobre lo público y lo privado, quisiera hacer un balance de los principales problemas que presenta la teorización marxista sobre la burocracia y la función pública. Obviamente colocando las tesis del marxismo en relación con Hegel y con los desarrollos posteriores sobre la burocratización y el Estado. En general, he mostrado que Marx en sus escritos de juventud le prestó gran atención al problema de la burocracia y sometió a una aguda crítica a los exaltatorios postulados hegelianos. Más tarde Marx precisa y afina su concepción del Estado y asume como hilo vertebral de su análisis la relación de dependencia de la variable política respecto de la economía. En esa misma vena, Lenin afirmaría taxativamente que “la política es la expresión concentrada de la economía”. Las clases sociales harían uso del Estado como de un instrumento, el Principal o central como fuente del poder político, de tal forma que la acción política sería vista, a la larga, en un sesgo semejante al de Maquiavelo: la lucha por adquirir el poder que no se tiene, y por conservar y ampliar el poder que ya se tiene. La burocracia, en tal perspectiva, sería entendida como el corazón mismo del aparato estatal. Usualmente, los teóricos de corte marxista ortodoxo diferenciarían a ésta en dos grandes ramas. La burocracia coercitiva: ejércitos y fuerzas de policía, pero también el sistema judicial, las cárceles, etc. Y una burocracia paralela de corte civil (en el sentido de no militar) encargada de innúmeras actividades, desde la administración del propio poder político, hasta las tareas que hoy se denominan de tecnoestructura, tales como recaudar impuestos, las relaciones exteriores, la función legislativa, etc. La crítica marxista condenaría, en general, la actividad de las burocracias gubernamentales en las sociedades capitalistas, llegando al extremo de considerar que éstas eran agentes mecánicos de los centros de poder del 123 capital. Recuérdese aquí la idea marxista, que no se puede rechazar sin más, el Estado como un comité que administra los negocios de la burguesía. Pensadores franceses como Louis Althusser o Maurice Godelier, hoy caídos en el olvido, fueron célebres hace unas décadas por radicalizar las tesis de Marx y ver al Estado en su totalidad (burocracia de todos los rangos incluida) en tanto aparatos ideológicos de dominación. La crítica marxista al papel mediador del Estado respecto de las clases sociales, desde su primaria etapa juvenil de elaboración conceptual, evidencia su rechazo a la concepción funcional, basada únicamente en la división del trabajo, de la teoría hegeliana de las clases, para postular un modelo que, como se sabe, se basó en el reconocimiento del antagonismo inmanente a las clases sociales, principalmente aquellas propietarias, frente a las masas asalariadas. En suma: el meollo de la reacusación marxista de la burocracia pública seguía siendo el mismo de sus textos de juventud: la imposibilidad de la burocracia de representar y ejecutar, esto último menos aún, el llamado interés general o público. De allí que la solución del problema burocrático fuera su eliminación o extinción progresiva. Marx, observando el contexto social de su época, supuso la inminencia de la confrontación antagónica entre el capital y el trabajo. La alternativa mitificada de la Comuna de París, suponía un camino más o menos fácil y viable de desaparición de la burocracia gubernamental, incluyendo su aparato civil y las propias fuerzas militares. Las clases sociales sin el sustento ontológico de las relaciones asimétricas de propiedad tenderían hacia una lógica evanescente, etc. Pero, en este segundo aspecto, la continuidad del pensamiento neomarxista recusaría las ilusiones sobre la pronta desaparición de ésta. La noción de dictadura del proletariado fue acuñada por el propio Marx. No así la teoría y la práctica del partido único y del llamado ‘centralismo democrático’. Lenin, Stalin y Mao serán, en el siglo XX, los grandes arquitectos del socialismo real. La centralización política de los proyectos en cuestión reforzaría la aparición y el dominio de una nueva burocracia gubernamental en las sociedades ‘sin clases’. Por ello, al considerar la concepción de Marx sobre el Estado, sus críticos hacen una observación casi unánime. No es cierto —afirman— que el Estado socialista, en sus distintas variantes, tanto aquellas desaparecidas como en las 124 actualmente existentes, haya constituido, en la experiencia vivida a lo largo del siglo XX, gobiernos que eliminasen el burocratismo y fueran más “baratos” que sus similares capitalistas. Allí, se dio burocracia y funcionarismo en gran escala. En este sentido, cabe subrayar que los postulados marxistas tuvieron profundas transformaciones en la práctica del proceso político revolucionario a lo largo del siglo XX, particularmente en aquellas revoluciones políticas triunfantes que, bajo el ideario marxista-leninista, se llevaron a cabo en regiones periféricas o excéntricas al desarrollo clásico del capitalismo eurooccidental. Es sabido que las expectativas tempranas de Marx sobre la transición del capitalismo al socialismo en sociedades de un industrialismo complejo y avanzado no tuvieron lugar. En estos países las corrientes socialdemócratas y “revisionistas” coparon el apoyo del movimiento sindical en los términos planteados por el teórico socialdemócrata alemán Eduard Bernstein: “el movimiento lo es todo, el objetivo final nada”. En esta conceptualización, lo importante para la causa obrera serían las reivindaciones económicas y materiales. Efectivamente, dicha política posibilitó procesos tanto de confrontación como de concertación que dieron lugar a una ampliación de la democracia social y económica, cobijando el despliegue de los derechos humanos denominados de ‘segunda generación’. Esta posibilidad de concertar reivindicaciones sociales debilitó, a la larga, la fuerza de las facciones más radicales y revolucionarias del socialismo de estirpe marxista en los países avanzados. En su lugar, las revoluciones políticas marxistas ocurrieron en países del llamado ‘tercer mundo’. En especial los casos paradigmáticos de Rusia, China y otros países del continente asiático. La paradoja fue que el cambio revolucionario tuvo lugar en sociedades que no tenían apego a las tradiciones explícitas de la democracia liberal, el parlamentarismo, el respeto a los derechos ciudadanos, la separación de las ramas del poder público, la existencia de una opinión pública deliberativa, etc. La tradición centralista, despótica y arrogante propia de estas culturas, hizo que se articulara el funcionarismo renovado de los Estados socialistas con el burocratismo prerrevolucionario. El régimen leninista del partido único, desde luego imputable como teorización a Lenin, no a Marx, se basó en reminiscencias de la Revolución Francesa que incluyen la jerga, la estructura jerárquica y vertical de los aparatos revolucionarios, y la idea de que al triunfar 125 la revolución sólo existe una opción: el ideario socialista revolucionario era la verdad, y ésta la expresaba funcionalmente el partido. Por lo tanto, todas aquellas facciones y organizaciones políticas diferentes al partido obrero, fueron vistas como rivales y enemigas. Esta construcción intelectual leninista permitió el desarrollo hegemónico del aparato socialista del poder público. De este modo, el modelo socialista ortodoxo se diseñó bajo el esquema de la democracia totalitaria (H. Arendt, 1956). Hoy, en la primera década del siglo XXI, tenemos una adecuada distancia para formular un balance de la experiencia del llamado “socialismo real” a partir de su implosión en el mundo entero. Uso deliberadamente esta expresión que implica un estallido hacia adentro, por la física incapacidad de continuar el régimen comunista en ejercicio del poder político para sacar avante su proyecto estratégico. En la Unión Soviética de los 80 del siglo XX, la audaz apuesta renovadora de Mijail Gorbachov, la perestroika, fue una mera anécdota histórica. El proceso rápidamente desbordó a quienes no tenían en sus manos el decurso de los acontecimientos. No quiero emplear aquí con todas sus letras la noción de ‘restauración’ del capitalismo porque de hecho, en gran medida, la URSS fue un férreo capitalismo de Estado, cuyos trazos iniciales fueron idea del propio Lenin a través de la Nueva Política Económica (HEP, por su sigla en ruso) en los años 20. Stalin después apeló, en gran medida, al nacionalismo, o más exactamente, al chovinismo de la gran Rusia para consolidar el ‘socialismo en un sólo país’. Los comunistas rusos fueron, en esas épocas, admiradores del taylorismo que copiaron sin reato (recuérdense el stajanovismo y las concepciones antisindicales y autoritarias de Trostky al proponer su alternativa de industrialización). Y asumieron una competencia desenfrenada por superar a las potencias occidentales a partir de una versión totalitaria de la acumulación originaria del capital plasmada en los llamados planes quinquenales. Aún los países que en la actualidad se consideran formalmente socialistas como la China continental, Cuba o el Vietnam, le han abierto las puertas al capitalismo de mercado, a la inversión extranjera, creando zonas económicas especiales y un curioso ‘partenariat’ de lo público y lo privado, donde por el lado de lo público actúa el capitalismo de Estado comunista, y por el lado de 126 lo privado, incursionan —en febril competencia entre sí— las más grandes multinacionales que fueron expulsadas —éstas y sus semejantes de la época revolucionaria— tras el radicalismo de la eliminación de la gran propiedad privada y la explotación del hombre por el hombre, según rezaba la jerga marxista ortodoxa hoy olvidada. No se sabe si el sepultamiento del socialismo marxista ortodoxo sea definitivo. Como historiador, sé que los ciclos históricos de larga duración pueden expresarse en siglos. El propio capitalismo sólo advino a la democracia liberal después de centurias de incubación en el absolutismo de las monarquías europeas; se apoyó durante muchísimo tiempo en la esclavitud, que reinventó y en las peores formas de coloniaje y afectamiento de la dignidad humana. Ciertamente, esa democracia de la que tanto se ufanan los teóricos de la neo-derecha, es un fenómeno reciente y no necesariamente consolidado. La ecuación capitalismo y democracia es sólo una de las tantas opciones que la Historia nos ofrece y no tiene, en modo alguno, un rasgo ineluctable. Pero lo que sí resulta claro es el fracaso de una forma de socialismo que pretendió eliminar de un plumazo el mercado reemplazándolo por la asignación planificada de los recursos productivos y sociales. Como resultado de esta dinámica, el aparato burocrático alcanzó cotas inimaginables bajo el socialismo real —de las presenciadas en los países capitalistas— a causa, precisamente, de la férrea centralización política del régimen leninista del partido único. La paradoja histórica advino ineluctable: la aplicación de las teorías del primer gran crítico de la burocracia condujo a un tipo de sociedad donde el ‘fenómeno burocrático’, en su patología, alcanzó su paroxismo. 7. A manera de conclusión: vigencia del rol del funcionarismo público Finalizaré este capítulo —de corte principalmente filosófico— sobre la base del balance ya hecho atrás de las principales aportaciones de las teorías clásicas sobre la función pública y la burocratización, destacando en especial la cuestión crucial de su pertinencia para la teoría política contemporánea. Y la forma como se reinterpretan los problemas mayores de la legitimación del Estado, la institucionalización social y política; y la dialéctica de lo público y lo privado. 127 7.1. ¿CUÁL ES —ENTONCES— LA PERTINENCIA DE LA VISIÓN HEGELIANA SOBRE LA BUROCRACIA? F. J. Guinle (1974) remarca la actualidad del pensamiento del Hegel maduro en relación con el Estado moderno. Hegel, según Guinle, capta de forma magistral la relación entre la libertad de los individuos y su necesaria articulación con el Estado, de tal manera que nos ofrece una salida que supera la antinomia: anarquía o totalitarismo. En este sentido, la idea central que resulta indispensable recuperar en el pensamiento hegeliano, es su firme convicción del papel civilizador (el sustrato ético) llamado a cumplir por el Estado moderno. Función social que se adecúa con las reivindicaciones en torno a la defensa de los Derechos Humanos o genéricos, en cuanto plasman la relación histórico-natural del hombre socializado. En tal lectura neohegeliana, el Estado como totalidad aparece, en consecuencia, como el lugar de encuentro, el espacio de la resolución, de los intereses, ideales o utopías sociales. Esto es lo que Hegel denomina “un Estado racional”. Así, más allá de la acerba crítica formulada desde distintos ángulos a la filosofía del derecho hegeliano, encontramos en su conceptualización una actitud filosófica extremadamente valiosa y de gran importancia práctica. La crítica al funcionarismo deviene una aportación positiva que posibilita el ajuste (la reforma) del desfase existente entre el Estado y la sociedad civil. El señalado desfase de la gestión estatal, expresada en la inoperancia de su estructura administrativa, reclama de parte de la sociedad civil una vigilancia hacia ésta que busque el cumplimiento de las funciones prescritas. Sobre este tópico he señalado en otro texto (E. Varela, 1995) cómo el Estado civilizatorio se entiende a partir de la concepción de ser éste un Estado democrático basado en la legitimación social de sus funciones. En este orden de ideas, el Estado —como órgano de control y mediación social— ejecuta un sinnúmero de tareas públicas v. gr. elevar la sociedad desde formas jurídicas y políticas de convivencia inferiores para superar conflictos; monopolizar y dar cumplimiento efectivo a las funciones de tipo coercitivo justicialista; generar mecanismos redistributivos de los bienes sociales tangibles e intangibles con base en los mecanismos de consentimiento establecidos, etc., donde el factor clave, piedra de toque de la democratización o el autoritarismo, estriba en el tipo de relaciones que establezca dicho Estado con su correspondiente sociedad civil. 128 Resultó claro que la justificación del arbitraje del Estado consistía precisamente, y en esto tuvo clarividencia Hegel, en que las clases y corporaciones cumplían funciones limitadas de mediación, puesto que, en realidad, aparecían como portavoces de intereses particulares buscando, a través de medios de presión, convertir sus prerrogativas en políticas de Estado, en políticas públicas. Por lo cual, un análisis en profundidad de esta discusión permite afirmar que la razón parcial la tenía Hegel; aceptada, desde luego, la crítica a la mistificación de las clases y las corporaciones que formulan los textos juveniles de Marx, puesto que dicha crítica no debilita sino, al contrario, fortalece la teoría que supone al Estado moderno garantista, defensor de los derechos civiles y ciudadanos, árbitro, portavoz del interés público y, por ende, actor institucional colocado por encima de la sociedad civil, esfera de los antagonismos, egoísmos y conflictivos intereses personales. El punto de vista hegeliano parte de reconocer la inevitabilidad, la necesariedad, de la función de representación del interés general que asume la burocracia gubernamental. Consciente de las limitaciones que supone el autocontrol ya referido, propugna entonces por una solución “democrática”: el control gubernativo realizado desde las esferas de las corporaciones, los estamentos y, en general, la sociedad civil. El democratismo hegeliano es limitado ya que su teoría sobre el control ciudadano no retoma elementos que, en su tiempo, habían sido abordados en la teoría y la práctica, principalmente mediante el proceso de la construcción intelectual y social de la opinión pública. En realidad, una revisión de los principales textos hegelianos nos corrobora en la anterior afirmación. Aunque debo reconocer que Hegel no fue consciente del papel de la llamada opinión pública que ya en sus tiempos, como Chartrier y otros autores lo han analizado, se expresaba en la forma de un vasto escenario de confrontación de los intereses privados y de consolidación del debate público (Habermas, 1978). Esto último, expresado en la profusión de medios impresos —libelos, folletos, prensa periódica— y de espacios de confrontación política y social v. gr., la plaza pública, los mítines y manifestaciones callejeras, los espectáculos públicos tales como la ópera, el teatro, etc. (Al respecto, véanse los capítulos tercero y cuarto de este libro). 129 Un factor relevante dentro del análisis aquí plasmado que quisiera subrayar, es la originalidad de la aportación hegeliana sobre la burocracia. Ello es más destacable en cuanto que en su época el Estado era relativamente pequeño, simple. Lejos de la enorme masificación de la burocracia contemporánea. Lejos también de la capacidad fiscal exacerbada que sería definida por los modelos europeos de tributación después de la segunda mitad del siglo XIX. Y lejos, finalmente, del ‘Estado providencia’, que inaugurarían los regímenes liberales en Inglaterra y el autoritarismo prusiano en Alemania. Por su contexto de época, Hegel no tuvo en cuenta la enorme tarea administrativa que asumirá posteriormente la burocracia pública. Su análisis es, por ello, primordialmente un discurso de mediación de conflictos en el marco de la sociedad civil. Es la imposición del Estado árbitro sobre los particularismos sociales y económicos. De cierta manera, Hegel postula una dicotomía central al análisis social del siglo XX: la antinomia sociedad civil vs. Estado, que podría ser leída igualmente como una contraposición asimétrica entre la esfera de lo político y la de lo económico. Y en plano analítico que evidencia la superioridad de la esfera de lo público-estatal, en cuanto encarnación de la universalidad de los fines societarios. Empero, si quisiéramos relievar algunos de los rasgos positivos de la teoría de Hegel sobre la burocracia que serían vistos luego en el modelo de Weber36, nos encontramos con el principio de universalidad para el acceso a la carrera burocrática. Éste se basa en la crítica de Hegel al ideal antiguo de los gobernantes por nacimiento o linaje, o por la compra del cargo. Como lo vimos atrás, el mecanismo sugerido es el examen de Estado, una prueba imparcial de conocimientos y aptitudes, abierta a todos los que quieran aspirar. Hoy en día, éste es un lugar común en la profesionalización de la gestión pública y expresa jurídicamente el arraigo del garantismo democonstitucional. Y para evitar la corrupción, el empleado público, según Hegel, tendría un control con base en estrictos sistemas de lo que hoy se denominan ‘inhabilidades e incompatibilidades’. Finalmente, de manera embrionaria, aparece la idea de que, al menos en parte, los funcionarios públicos deben ser asalariados por el Estado. La perspectiva hegeliana serviría indirectamente al influjo apoteósico del nacionalismo estatal a lo largo de los siglos XIX y XX. La totalidad elevada a concepto integrador de la sociedad civil, pautará una lógica argumentativa de 130 subordinación de los individuos a los supremos intereses de la nacionalidad o de la soberanía encarnadas en el aparato político administrativo. De esta forma, las ideas de Hegel resultaron anticipatorias e influyeron ciertamente en la evolución posterior de los aparatos burocráticos gubernativos; obviamente en un tipo de discurso, el hegeliano, con un elevado nivel de abstracción. Aunque muchos de sus hitos sobre la justicia, la familia, la institución de la ‘policía’, eran esencialmente síntesis conceptuales de las realidades administrativas de los principados y reinos alemanes de su tiempo. Estas lógicas de gobernabilidad política serían llevadas a sus extremos por los modelos nacionalistas y autoritarios de extrema derecha e izquierda, nazismo y sovietismo —valga la aclaración— que marcaron la crisis de la democracia liberal en el medio siglo pasado37. 7.2. LA CONTINUIDAD DEL ‘FENÓMENO BUROCRÁTICO’ En un plano genérico, podemos comprobar que el Estado moderno, en sus diversas gamas —como arquetipo matricial del poder centralizado— se ha continuado desarrollando, ocupando de manera cada vez más prominente, un lugar de dominio preponderante sobre las la sociedades contemporáneas. La centralización no ha cedido el paso a las formas de gestión directa por las masas que Marx predicó como una de las enseñanzas fundamentales de la Comuna. Por el contrario, cada día es más notoria la separación entre sociedad política y sociedad civil. El antiguo dilema rousseauniano, ‘soberanía del pueblo versus soberanía del Estado’, se ha resuelto finalmente en favor de este último. Una forma concreta de expresión de esta ideología política que considera inevitable la superioridad de la burocracia en su variante contemporánea la encontramos en el sociólogo Max Weber38. Generalmente las reflexiones sobre la burocracia se remiten fundacionalmente a su conocido paradigma del llamado ‘fenómeno burocrático’ (Crozier, 1964). En esencia, Weber (1976, 1985) considera que la burocracia, como fenómeno social, era el resultado inevitable de la modernización. Recordemos su poética expresión sobre el desencantamiento del mundo. Pero las fuentes y motivaciones principales de la concepción de Weber acerca del desarrollo del sistema capitalista de producción, se pueden entrever en su oculta pero evidente polémica con Marx. Y, en tal sentido, la 131 axiología weberiana nos coloca en el centro de la discusión al problema de la motivación y la cuestión de la acción valorativa. Ello, además, determinaría el enorme influjo de Weber sobre la sociología americana (Parsons, Merton), las teorías sobre la desviación social y, parcialmente, sobre las propias Ciencias del Management en los EE.UU (Perrow, 1991). Weber, en este sentido, afirma la existencia de una legalidad racional para el comportamiento y la función social de la burocracia: “El fundamento decisivo para el proceder de la organización burocrática ha sido (...) siempre su superioridad puramente técnica sobre cualquier otra forma: un mecanismo burocrático, plenamente desarrollado está, con respecto a cualquier otra forma, en la misma relación que una máquina con respecto a los medios no mecánicos de producción de bienes” (M. Weber, 1.985.). Así,pues, Weber reconocía, con algún grado o dosis de pesimismo, la inevitabilidad y la superioridad del fenómeno burocrático. En su perspectiva, la modernización social del orden capitalista implicaba de suyo la configuración, el surgimiento de instituciones y organizaciones fuertes, basadas en la centralización y en la disciplina normativa; lo cual producía inexorablemte el carácter impersonal de la acción de los miembros de dichas organizaciones, de lo cual derivaría posteriormente la distinción de la sociología funcionalista americana entre status y rol (Berger y Luckmann, 1968); y, ante todo, el carácter profesional y tecnocientífico o de racionalidad instrumental, como se le conocería después, de la imposición del poder de la burocracia sobre la sociedad. En esta perspectiva, el propio Weber veía con notable desconfianza este tránsito hacia la socialización de la producción adjetivándole con denuestos que describen la inevitable corrupción consustancial al preceder del gobierno socializado aun cuando se trata de la gestión de los asuntos públicos bajo el régimen capitalista. Se pregunta, por tanto ¿qué podemos oponer a este sistema para salvaguardar una porción del alma? (M. Weber, 1966). La diferencia más importante entre la visión weberiana de la burocratización y las teorías precedentes es su análisis del proceso activo de las interacciones sociales, basado en el modelo denominado individualismo 132 metodológico39. La inmensa mayoría de las recensiones hechas sobre la teoría de la burocratización weberiana se centran en la descripción de las tipologías, los célebres tipos ideales, carismático, tradicional y racional-legal; pero dejan de lado el sustrato epistemológico y ontológico —a mi juicio bastante problemático— del cual parte Weber. En el fondo, el cálculo utilitario del individuo transmutado en funcionario, o del alterno usuario de las relaciones estatales, será la base de la construcción de un poderoso aparato de dominación impersonal. Las motivaciones de los llamados “actores”, como lo reiterara la sociología postweberiana, serán las claves explicativas de los vínculos societales. Aunque la burocracia es un fenómeno social complejo, la interpretación weberiana lo devela, en términos de una teoría de la acción social, enfrentando acremente el determinismo40. En este caso, primordialmente el economicismo propio a la visión marxista del desarrollo social del capitalismo, y, de otro lado, a la visión teleológica de la Historia de parte del hegelianismo. Ciertamente, en el modelo de análisis hegeliano no hay lugar —en modo alguno— para un sujeto individual. La burocracia resulta para Hegel, como ya lo vimos, una encarnación activa del principio de la totalidad que está como pulsión original en la voluntad del monarca. La autonomía en las interacciones del aparato burocrático con la sociedad civil no era un asunto central a ojos de Hegel. Sí lo fue, en cambio, y de forma decisiva en la teorización weberiana. La otra cuestión decisiva que separa la visión de Weber de la de Hegel es, por supuesto, el plano de análisis filosófico y especulativo —en el recto sentido de la expresión— en Hegel; sociológico o, más exactamente, socio-histórico en Weber. Los puntos de confluencia, aunque son menores respecto a las diferencias, resultan de percibir a la burocracia como un fenómeno que encarna —como el que más— la Modernidad; específicamente en términos de la racionalidad dominante en las sociedades capitalistas mercantiles centradas en una gobernabilidad de las relaciones sociales que dimana del Estado-Nación. La incomprensión del rol progresivo y civilizador que puede eventualmente cumplir el Estado obedece en, primer lugar, al punto de vista de la conceptualización, a la gran fuerza de las ideologías que lo recusan y la simplificación antinómica del debate sobre el papel del Estado (debo advertir que sobre este punto ahondaré en los capítulos cinco y seis del presente libro). 133 Pensemos adicionalmente la referida crítica al funcionarismo público de las sociedades contemporáneas. Tal como lo vimos en la crítica juvenil de Marx a Hegel, la recusación del funcionarismo ha insistido en su falsa representación del interés general. Teóricamente se supone que el funcionario o burócrata enquistado en el aparato del Estado asume la defensa de los intereses colectivos de los ciudadanos; no es el funcionario de éste o de aquel grupo social. Está, además, sometido a una rigurosa comprobación de su neutralidad. Para ello, el Estado se vigila a sí mismo a través de instituciones como las procuradurías, las contralorías y los ombudsman (veedores), y mediante la implementación de modelos de control interinstitucional, en desarrollo de la conocida teoría constitucional de los frenos y contrapesos del poder público. Igualmente las democracias avanzadas posibilitan que esta función fiscalizadora recaiga en miembros de la sociedad civil. Para lo cual se introduce con frecuencia en los cuerpos directivos de los entes públicos la presencia de representantes de la comunidad: los consumidores, los gremios productivos, los intelectuales, etc.; o, en su defecto, se acude a la delegación del jefe del Estado en una persona de su confianza, que al representarlo, asume la dignidad de la Nación. Pero el problema estriba en que la tosca realidad no marcha de la mano con los bellos ensueños y los ideales republicanos. El funcionario es corruptible y corrupto en más ocasiones de las que el público sospecha; está sometido al vaivén de los poderosos intereses privados e, incluso, de los intereses de las potencias extranjeras, cayendo en dolos antipatrióticos. Por esta razón existe una desconfianza acentuada en el rol del funcionarismo. La crítica a la burocracia y al funcionarismo se engloba así en un marco mayor que insiste en la recusación del propio Estado. No siempre en la forma del clamor ácrata contra sus bases constitutivas, pero sí en un tono que exige depurar el manejo del Estado frente a los abusos y privilegios de la burocracia. Esta ha sido la motivación ideológica en ciertos discursos antiestatales que acuden a la descripción del Estado como si fuera aquel terrible y voraz Leviatán del que hablara Hobbes. No se trata sólo del discurso anarquista ni tampoco exclusivamente de la prédica antiautoritaria desarrollada en la crítica sociológica y en la psicología social moderna. El discurso antiestatal proviene de dos fuentes plenamente identifica134 das. De un lado, de la enorme influencia ideológica del liberalismo libertario, y, del otro, de una interpretación socialista anarquista que señala el problema de la desigualdad y la injusticia en la existencia de un cuerpo público-estatal coercitivo, base de la apropiación patrimonial, clasista del interés general. En este sentido el anarquismo coincide con la utopía marxista de “un futuro mundo feliz, sin clases, sin política, sin Estado”. Si bien es cierto, tales tesis han demostrado su carácter deleznable al probarse que el Estado nacional está lejos de ser un gran aparato sobredimensionado con respecto a la sociedad civil, han logrado permear la opinión al acusar al Estado como inoperante y adjudicar dicho inmovilismo a la fuerza de la burocracia oficial: al funcionarismo. Evidentemente, que tal situación exista no es un secreto para nadie. Lo difícil de explicar y de aceptar es la subsistencia del fenómeno, pese a las prédicas de las altas esferas gubernamentales contra la ominosa y anónima maquinaria. Parecería, en efecto, que nos hallamos frente a un mal no sólo inevitable sino, además, invencible. En este sentido, la experiencia indica que se debe propugnar por una crítica positiva del funcionarismo. El atraso de la gestión estatal —expresada en la inoperancia de su estructura administrativa— reclama de parte de la sociedad civil una vigilancia hacia ésta que busque el cumplimiento de las funciones prescritas. La impunidad absoluta de los funcionarios frente al incumplimiento de sus deberes es una de las fuentes corrosivas de la credibilidad en las formas de gestión propias de los altos niveles de socialización y civilidad. La manifestación de este fenómeno opera, en forma directa, en los criterios sociales de valoración del Estado; si a éste, por ejemplo, se le vislumbra como socializador de las pérdidas y quiebras, y privatizador (redistribuidor) de las utilidades. En este sentido, véase el caso tan frecuente del subsidio estatal a la actividad privada de acumulación de capital, el cual magnifica la separación entre el Estado, como praxis del ejercicio del poder por parte de grupos dominantes, y la constitución política como postulado o perfectibilidad puesta. Y por consiguiente expresa la ruptura entre ese mismo Estado y el conjunto de la sociedad civil. Esto es particularmente importante cuando contemplamos la realidad efectiva de los acuerdos contractualistas por los que la sociedad civil se rige y los cuales se encuentran sobredeterminados por la estructura y la dinámica del Estado. En 135 este sentido, el Estado parece (o semeja) una función de mediación entre los distintos intereses particulares. Por ejemplo, entre los industriales, los sindicatos, los productores de materias primas, etc., de donde se colige la importancia de historizar positivamente la función de mediación del Estado y la validez operacional de los acuerdos entre las clases y demás grupos de interés que se enfrentan en el seno de la sociedad civil. El Estado resulta catalizador de los cambios sociales; este es el caso de los procesos que cumplen las revoluciones institucionalizadas a partir de la transformación de la fuerza —victoriosa— en una nueva legalidad, en un nuevo corpus del derecho positivo. O asume tal rol por la vía pacífica de la reforma social y política. Los derechos civiles de los miembros de la comunidad de Estado —en tanto que individuos—, las esferas de organicidad de lo político, de la cultura, (en el sentido lato de la expresión), las formas de punición, castigo y socialización de los transgresores, etc. Todo lo anterior, hace parte sustancial de la temática civilista, de la búsqueda de formas más avanzadas de organización social, adecuadas al “espíritu de los tiempos”. De tal modo que la democracia social, la consciencia del ciudadano para asumir sus deberes, etc., no es, en modo alguno factor que surja gratuitamente como efecto de la mera transformación de los patrones culturales de los “de abajo”. Pero ello deviene, lo concreta, tal como las experiencias históricas lo indican, la articulación global de los distintos actores sociales; y en su plasmación intervienen el señalado programa en pos del Estado civilizatorio, así como las fuerzas de la sociedad civil, en sus diferentes y complementarios campos: (economía, política y cultura) y, en particular, aquellas que fungen un rol de liderazgo histórico. El moderno Estado industrial y postindustrial de nuestros tiempos está muy lejos del viejo veilleur de nuit en el que se pensaba en el siglo XIX. Es hoy, y desde hace mucho, un inmenso aparato productivo, de control y subordinación de la sociedad civil, cuyas funciones sociales son cada vez más indispensables. Al contrario de la progresiva desaparición o extinción del Estado pronosticada por Marx, las sociedades contemporáneas enfrentan, desde hace 136 tiempo, el problema de repensar, en términos histórico-concretos, la vigencia de este órgano de dominio que aparece colocado por encima de la sociedad, sobredeterminándola. La sociedad industrial y postindustrial, arraigada con un carácter transnacional y globalizador, con posterioridad al final de la segunda guerra mundial, ha mostrado un grado de complejidad no comparable con los modestos niveles de organicidad pública y privada, propios a la Nación-Estado que se coronó históricamente al final del siglo XIX. Los ideales del ultraliberalismo —como lo veremos en los siguientes capítulos—, pese a su radical retórica antiestatal en la vida diaria, significan modelos y alternativas que conducen a su predominio sobre las sociedades civiles. La función pública es —por lo tanto— un factor estructural de la Modernidad que necesita ser pensado críticamente para debatir sobre la oportunidad y la justeza de las políticas públicas, como eje de la acción estatal y de las interacciones sociales, en un época como la actual que se ha abierto a dimensiones ecuménicas y globales como nunca antes en la Historia. NOTAS 1 Sobre el concepto “Modernidad” y la inflexión verbal “modernización”, véase en este libro el capítulo quinto. 2 Al respecto señaló Hegel con perspicacia en alguna ocasión: “Para decir algo más sobre la pretensión de enseñar cómo debe ser el mundo, la filosofía, en todo caso, llega siempre demasiado tarde. Como pensamiento del mundo, aparece solamente cuando la realidad ha consumado su proceso de formación y se ha realizado... Cuando el filósofo pinta gris sobre gris, una forma de la vida ha envejecido y no se deja rejuvenecer sino solamente reconocer; el búho de Minerva sólo inicia su vuelo a la caída de la noche” (Hegel, 1966.) 3 “Este mundo civil ha sido ciertamente hecho por los hombres, por lo cual se puede y se debe hallar sus principios en las modificaciones de nuestra propia mente. Debe causar asombro a todo el que reflexione sobre esto el que todos los filósofos intentaron alcanzar la ciencia del mundo natural, ciencia que sólo puede tener Dios que lo hizo; y que descuidaron pensar el mundo de las naciones, o sea, el mundo civil, del cual por haber sido hecho por los hombres, los hombres podían tener ciencia”. G. Vico. (1964: 200). 137 4 En este contexto entenderemos la respuesta de Locke: “aunque en el estado de naturaleza tiene el hombre todos esos derechos, está sin embargo expuesto constantemente a la incertidumbre y a la amenaza de ser invadido por otros, pues como en el estado de naturaleza todos son reyes lo mismo que él, cada hombre es igual a los demás. Como la mayor parte de ellos no observa estrictamente la equidad de la justicia, el disfrute de la propiedad que tiene en un Estado así es sumamente inseguro” (Ob. Cit., J. Locke). 5 Para esta filosofía las acciones de los hombres obedecen a móviles de orden natural y no son distintas en cuanto a su naturaleza del conjunto de las acciones que existen en el mundo. Así, su postura psicológica es mecanicista: “Ese movimiento en el que estriban el placer y el dolor es una solicitación o una atracción que conduce hacia el objeto que gusta o que nos hace alejar del que nos disgusta. Tal movimiento se llama apetito o deseo cuando el objeto es agradable; aversión cuando el objeto desagrada; tenor, relativo al no placer que espera”. (T. Hobbes, 1962:119). 6 “La única manera –señala Locke en su ensayo sobre el gobierno civil (1690)– de que alguien se despoje de su natural libertad y la coloque en la sociedad civil es por convenio con otros hombres con objeto de formar una comunidad y vivir en seguridad y en paz unos con otros... cuando un cierto número de hombres ha convenido así en formar una comunidad o gobierno, quedan incorporados y constituyen un cuerpo político en el cual la mayoría tiene derecho a obrar y obliga al resto a obrar”. 7 Locke, al caracterizar el papel que el Estado debe cumplir en el seno de la sociedad civil, precisa: “El principal fin por el cual los hombres se reúnen en repúblicas y se ponen bajo su gobierno es la preservación de su propiedad, para lo cual, en el estado de naturaleza faltan muchas cosas: ante todo una ley establecida, fijada, conocida que sea aceptada y permitida por el consenso general como patrón de lo justo y de lo injusto para medir las controversias, pues aunque la ley de la naturaleza sea clara para todas las criaturas racionales, como los hombres sufren la parcialidad de sus intereses y son ignorantes por no estudiarla, propenden por no considerarla como ley que los obligue en casos particulares”. 8 Aquí hallamos las famosas restricciones que Hobbes le prescribe a los ciudadanos: 1- los súbditos no pueden cambiar la forma del gobierno. 2- el poder soberano no puede ser alienado (forfeited). 3- ningún hombre, sin caer en injusticia, puede reclamar contra la institución soberana constituida por la voluntad manifiesta de la mayoría. 4- las acciones del soberano no pueden ser justamente juzgadas por sus súbditos. 5- el soberano, cualquiera sean las acciones que haga o ejecute, no puede ser castigado o sancionado por sus súbditos”. 9 “Su Estado soberano no fue diseñado para negarle a los hombre una vida de competencia y adquisiciones, sino al contrario, para asegurarles que ella podría tener lugar: el Soberano está ideado para proveer las condiciones en las cuales los hombres pudieran llevar una vida segura, sin que peligrara la paz civil. El precio a pagar no le parece a Hobbes demasiado 138 alto: deben reconocer, como si lo hubieran convenido así, por intermedio de un contrato, una obligación de obedecer las leyes del soberano en la medida, precisamente, en que el soberano les garantice protección” (MacPherson; 1962). 10 Aunque ciertamente, según Hobbes, el método que debe utilizarse no debe ser el simple método geométrico euclidiano, porque en la Ciencia Política no es posible partir de axiomas tan sencillos que sean absolutamente demostrativos, por lo cual, debe utilizarse un método compuesto-resolutivo (resolutive-compositive method), que Hobbes toma prestado del pensamiento científico de Galileo: “La parte resolutiva brindaba la manera de alcanzar las proposiciones básicas simples requeridas; la parte compuesta permitía construir las proposiciones complejas a partir de las simples” (MacPherson, Ob. Cit.). ¿Por qué, en la segunda mitad del siglo XX, aún leemos a Hobbes, que escribió tres siglos atrás? ¿Por qué la permanente fascinación con sus escritos? Cuando estas preguntas se ponen en una misma dirección, una respuesta común surge: nuestro mundo está obsedido por la problemática del poder, y Hobbes fue un analista del poder. Nosotros queremos, en la actualidad, no sólo comprender y por lo tanto controlar el poder, sino, además captar la cuestión de su derecho. Al igual que Hobbes. Éste, en efecto, expuso los lineamientos del poder más claramente que ninguno otro desde Maquiavelo, de una manera tan sistemática como nadie lo había expuesto hasta el momento y como muy pocos lo harían desde entonces. Pero, además, Hobbes afirmó la naturaleza igualitaria de los derechos del hombre y trató, a partir de estos dos postulados, de construir una teoría del derecho y de la obligación, al mismo tiempo que intentó construir una teoría del poder. Sin embargo, su problema central fue la paz, tal como nos sucede a nosotros. En verdad, Hobbes pensó poco en la guerra entre las naciones. Su atención se centró en la guerra civil; cómo evitarla fue el propósito central de su reflexión política. “La utilidad de la filosofía política y de la moral- y por lo cual son estimadas estas ciencias -no es tanto porque conocerlas nos suministre la posibilidad de muchas comodidades, sino porque al conocerlas nos evitamos muchas calamidades”. 11 En ambos pensadores se encuentra la teoría de un soberano que subsume los intereses individuales de los miembros de la sociedad para que ésta pueda existir cumpliendo la finalidad primordial de salvaguardar la paz y la seguridad de los asociados en comunidad; es común a ambos también la creencia de que la Ciencia Política no tiene que ver con nociones de orden moral y religioso, poseyendo –al contrario– su propio específico que puede ser comprendido partiendo de concebir al hombre y a sus actos como fenómenos de carácter natural, estableciendo una identidad entre el ser natural y el hombre; también comparten el principio según el cual todos los seres se mueven de acuerdo al propósito de perseverar en su ser (conatus). Lo anterior para señalar sólo algunas de las más importantes similitudes que hallamos al comparar los sistemas políticos de Hobbes y Spinoza. 12 Resumiendo el punto de vista de Spinoza, en relación con este asunto, 139 señala Stuart Hampshire: “Todas las actuaciones políticas pueden ser juzgadas por sus efectos en la distribución del poder dentro del Estado, y además, por el efecto que tal distribución del poder tiene para evitar la anarquía, que es para los hombres el más grande de los males” (Ibíd., pág. 179). 13 En estos análisis spinozistas se cree ver, por parte de muchos de los estudiosos de su obra, una anticipación de la temática contemporánea de las “ideologías”. Y hemos de señalar que, en efecto, en la Ética se presenta el problema de los niveles de conocimiento de tal forma que la imaginación no es negada en su verdad, relativa a una imagen que el hombre recibe como afección externa, sino colocada en situación de menor comprensión, o de incomprensión, de las causas y el orden necesario que sólo se dan con propiedad en el segundo y tercer niveles de entendimiento. Confrontar al respecto el texto del tempranamente fallido profesor Augusto Díaz S., de quien mucho recibí en cuanto a mi formación académica : “El concepto hegeliano de la historia universal”, en Estudios de historia de la filosofía, Cali, Universidad del Valle, 1997. 14 Rousseau hace un elogio del hombre primitivo, de lo que se llamaba, en la época, el buen salvaje: “El hombre de bien es un atleta que se complace en combatir desnudo, despreciando todos estos viles ornamentos que debilitan el uso de sus fuerzas, de los cuales, la mayoría ha sido inventada más que todo para ocultar sus deformidades” (Rousseau, Ob. Cit.). 15 “Al estudiar las obras de los autores más famosos nos sorprende descubrir que no tienen ninguna teoría totalmente nueva. Son las mismas ideas repetidas, ideas no creadas por el siglo XVIII. Rousseau, a quien le gusta cultivar la paradoja, emplea un tono muy distinto y muy serio cuando se ocupa de la política. En su concepción del objeto y el método de la filosofía política, en su doctrina de los derechos del hombre, no hay nada que ya no tenga su paralelo y modelo en Locke, Groccio o Pufendorff” E. Cassirer (1945). 16 Inspirados en Tomás de Aquino, bajo la idea de que todo gobierno es un pacto entre el príncipe y el pueblo. Y bajo una concepción de subordinación del príncipe a los designios del derecho natural, llegando incluso a la defensa del tiranicidio. 17 Este republicanismo ha sido reivindicado por los propios líderes americanos del XIX. Recuérdese que en los escritos de Simón Bolívar, v. gr., la carta de Jamaica, sus discursos en la convenciones de Angostura y Cúcuta, en los documentos que escribió para la de Ocaña, en donde no estuvo presente, hay numerosas referencias directas a Rousseau. Sus biógrafos cuentan que Bolívar siempre portaba un ejemplar de las obras de Rousseau. Se puede constatar fácilmente la gran importancia simbólica otorgada por los republicanos latinoamericanos del siglo XIX a la obra rousseauniana, porque piensan que el problema de la virtud y la moralidad republicana sólo se puede resolver sobre la base de la identificación del individuo con el ideal patrio. 18 Es casi innecesario señalar que este trabajo no es un intento de recon140 strucción histórico-antropológico, sino una brillante hipótesis de carácter lógico. “Comencemos por descartar los hechos, porque ellos no vienen al caso. No es necesario tomar las investigaciones por verdades históricas, sino solamente por razonamientos hipotéticos y condicionales; más apropiados para establecer la naturaleza de las cosas que para mostrar su verdadero origen”. Rousseau, Discours sur lòrigine et les fondements de la inégalité parmi les hommes (Ed.de La Pléaiade: 132). 19 Así lo expresa en el Contrato Social: “Encontrar una forma de asociación que defienda y proteja la fuerza común, la persona y los bienes de cada asociado, y por la cual cada quien, uniéndose a todos, no obedezca más que a sí mismo, y permanezca tan libre como antes, tal es el problema principal al cual da el pacto solución. Las cláusulas de este contrato están determinadas de tal forma por la naturaleza del acto constitutivo del mismo, que la menor modificación las haría vanas y sin ningún efecto; de tal suerte que, en tanto que ellas no han sido formalmente enunciadas, son, sin embargo, las mismas (cláusulas), tácitamente admitidas y reconocidas, justo hasta que, siendo el propio pacto social violado, cada quien se reencuentre con sus derechos originarios y, retomando su libertad natural, pierda la libertad convencional que había adquirido. Estas cláusulas, bien entendidas, se reducen en el fondo a una sola, a saber: la alienación total de cada asociado a la comunidad” (Rousseau, 1971: 66). 20 Precisando, en este sentido, el ejercicio de la soberanía y la articulación de ésta con la voluntad general, expresa Rousseau: “La soberanía, no siendo cosa distinta al ejercicio de la voluntad general, no puede jamás alienarse; y el soberano, que no es más que un ser colectivo, no puede ser representado por nadie distinto a él mismo: el poder puede ser transmitido, pero en manera alguna puede transmitirse la voluntad general” (Ibíd.: 78). 21 En efecto Rousseau señala al hacer referencia a las causas que promueven la aparición de la voluntad general: “Toda acción libre tiene dos causas que concurren a producirla. Una moral, a saber: la voluntad que determina el acto; otra física, a saber: la potencia que la ejecuta... (en el cuerpo político) se distingue igualmente la fuerza y la voluntad; ésta recibe el nombre de potencia legislativa (puissance législative); la otra recibe el nombre de potencia ejecutiva (puissance exécutive). Hemos visto que la potencia legislativa pertenece al pueblo y no puede pertenecer más que a él. Es fácil de ver, al contrario, por los principios arriba anotados, que la potencia ejecutiva no puede pertenecer a la generalidad en tanto se la conciba como legisladora o soberana, porque esta potencia no consiste más que en actos particulares que no son del exhorte de la ley, ni por consiguiente, del resorte o esfera del soberano, puesto que todos sus actos (los del soberano), no pueden ser cosa distinta que leyes” (Ibíd..: 115). 22 “Cuando digo que el objeto de la ley es siempre general, entiendo que la ley concibe a los ciudadanos y a las acciones que estos realizan como abstractas; jamás a un hombre como individuo ni a una acción en cuanto particular: así, la ley puede estatuir que habrá privilegiados, pero no puede 141 asignar tales privilegios a ninguna persona; la ley puede prescribir que habrá muchas clases de ciudadanos, definir de igual manera las características que conceden tales derechos a determinadas clases de ciudadanos, definir de igual manera las características que conceden tales derechos a determinadas clases, pero no puede asignar a tales o a cuales para que los disfruten; puede establecer un gobierno monárquico y una sucesión hereditaria, pero no puede elegir un rey, ni nombrar una familia real. En una palabra, toda función que se refiera a un objeto individual o singular no pertenece a la esfera de la potencia legislativa” (Ibíd.: 94). Como se ve, esa diferenciación rousseauniana introduce una polémica importantísima. Se trata de definir el lugar que ocupa, dentro del contrato, la acción singular del poder ejecutivo. ¿En manos de quiénes está la potestad de aplicar la ley? Evidentemente, la teoría rousseauniana busca deliberadamente defender la intervención de la comunidad en el manejo y la aplicación de la ley. Pero el mismo Rousseau reconoce que siendo la democracia la mejor de las formas de gobierno, es aplicable sólo en casos excepcionales para pequeños Estados o ciudades. Si la democracia no puede operar, si es inaplicable, ha de intervenir un elemento político práctico que Rousseau había recusado anteriormente: la delegación de poder, la representación de los ciudadanos. De aquí se desprenden las asociaciones partidarias, los grupos políticos que intervienen en la vida social buscando el consenso mayoritario para monopolizar la interpretación y la aplicación de la ley. 23 Rousseau, Lettre au Marquis de Mirabeau (1767) citado por J.L. Lecercle en “Introducción al Contrato Social” en Du Contrat Socialde. cit., pág. 32. 24 Kant recoge la argumentación clásica en Locke y en Rousseau, que negaba la posibilidad de derivar el poder del Estado del poder familiar, visto como pulsión natural. La obediencia paterna, nos decía Locke es, en la edad temprana, fruto de la debilidad, la dependencia; y en el hombre adulto, una inclinación de la voluntad, en forma de gratitud, pero no algo imperativo. 25 Así lo reconoce el filósofo norteamericano John Rawls cuando afirma de manera taxativa la conexidad entre la razón pública y la democracia política. Lo público –en la perspectiva rawlsiana– necesariamente implica la institucionalidad del Estado, sus basamentos contractuales y consensuales. Ello sólo sería posible en un contexto validado por la aceptación de la razón pública (que Rawls, para el caso norteamericano, supone reside institucionalmente en la Suprema Corte) como razón ciudadana, además de darse en torno de la vigencia del principio del bien público como base de la justicia, procesos que a su vez demandan la existencia de una comunidad de individuos que persiga el interés general. Cfr. al respecto, J. Rawls, “Political liberalism” (1992: 229ss). 26 Hemos de recordar que estas tesis habían sido esbozadas por Locke. 27 “La monarquía es necesariamente democracia en tanto que es inconsecuencia con respecto a sí misma; el momento monárquico no es una inconsecuencia en la democracia. La monarquía no puede ser comprendida 142 en sí misma, pero la democracia sí. En la democracia ninguno de sus momentos adquiere otra significación que la que le corresponde. Cada uno de ellos no es más que un momento del gran demos. En la monarquía una parte determina el carácter del todo. Toda la constitución debe modificarse según un punto fijo. La democracia es el género de la constitución. La monarquía es una especie, y a decir verdad, una mala especie. La democracia es ‘el contenido y la forma’. La monarquía no debe ser más que forma, pero altera el contenido. En la monarquía, la totalidad del pueblo es clasificado en una de sus maneras de existir: la constitución política. En la democracia, la constitución misma aparece simplemente como una determinación única, la autodeterminación del pueblo.La democracia es el enigma descifrado de todas las constituciones” (Marx, Ob. Cit.). 28 Afirma K. Marx en los Grundrisse: “El punto de partida, naturalmente, está constituido por los individuos que producen en sociedad, es decir, por la producción de los individuos socialmente determinada. El pescador y el cazador individual y aislado, con el que comienzan Smith y Ricardo, pertenecen a las imaginaciones carentes de fantasía de las robinsonadas del siglo XVIII que, en modo alguno, expresan como se imaginan los historiadores de la civilización, simplemente una reacción frente al superrefinamiento y un retorno a una vida natural mal comprendida. De la misma forma que tampoco descansa sobre tal naturalismo el contrato social de Rousseau que pone en conexión y relación mediante el contrato a sujetos independientes por naturaleza. Esta es la apariencia, y la apariencia exclusivamente estética, de las pequeñas y grandes robinsonadas. Se trata, más bien, de la anticipación de la moderna “sociedad civil” que se preparaba desde el siglo XVI y que en el XVIII marchaba a pasos de gigante hacia su madurez. En esta sociedad de libre competencia el individuo se presenta desprendido de los lazos naturales, etc., que lo convertían en épocas anteriores en un elemento accesorio de un conglomerado humano limitado, determinado. A los profetas del siglo XVIII, sobre cuyos hombros se apoyan todavía por completo Smith y Ricardo, este individuo del siglo XVIII –producto, por una parte, de la disolución de las formas sociales feudales y, por otra parte, de las nuevas fuerzas productivas desarrolladas desde el siglo XVI– se les presenta como un ideal cuya existencia pertenece al pasado. Se les presenta no como un resultado histórico sino como un punto de partida de la Historia”. 29 “La esencia fecunda del mensaje rousseauniano de la libertad” (igualitaria) debe verse en la instancia universal (democrática) del mérito personal, o sea, en la exigencia del reconocimiento (social) de todo individuo humano, con sus aptitudes y necesidades peculiares. 30 Rousseau, con plena conciencia de la dificultad, intenta dirimirla en los siguientes términos atinentes a la polémica en torno a los derechos que el Estado puede abrogarse frente a los individuos: “Todo hombre posee el derecho de arriesgar su propia vida buscando conservarla. ¿Se ha dicho alguna vez que aquel que se arroja por una ventana para escapar de un incendio sea culpable de suicidio? ¿Se ha imputado igualmente este cri143 men a quien perece en una tempestad en la que se ha visto involucrado no ignorando el peligro? El tratado social tiene por finalidad la conservación de los contratantes. Quien quiere el fin quiere los medios, y estos medios son inseparables de algunos peligros e igualmente de pérdidas. Quien quiere conservar su vida a expensas de los demás, debe darla también cuando sea necesario. Ahora, el ciudadano no es juez del peligro al cual la ley quiere que se exponga; y cuando el príncipe (le prince) le dice: “Es necesario que tú mueras”, él debe morir puesto que es bajo esta condición que ha vivido en seguridad y su vida no es solamente una bienandanza de la naturaleza sino también un don condicional del Estado. Rousseau, Du Contrat Social, pág. 90. 31 Della Volpe expresa estas ideas –no sin antes ubicar el rasgo de investidura a-histórica e intemporal del “individuo valor o persona” que consigna las limitaciones reales del Contrato tal como lo concibe Rousseau– en los términos perentorios que son propios a su vehemente lectura del mensaje del filósofo ginebrino: “...la originaria formulación contractualista –jusnaturalista– para resolver tal sociedad no consentía más que la garantía de los méritos y de los derechos de una parte de los individuos integrantes del cuerpo social, de una clase: la burguesa (...) y en ella resultaban contradichos los conceptos principales de voluntad general del cuerpo social y de soberanía popular. Pero también es cierto que la original problemática democrática de Rousseau, con sus mismos conceptos radicales de un cuerpo unitario soberano y del decisivo carácter racional –artificial– histórico de tal cuerpo, amenaza ya con romper el esquema racional-natural a priori del jusnaturalismo: hecho que, lo sabemos, no se produce en Kant. Así, el problema descubierto por Rousseau, el del reconocimiento social del individuo mismo, o sea de la susodicha proporcionalidad universal de valores sociales y méritos individuales, el de una libertad igualitaria, sigue siendo un problema real aún después de la revolución burguesa; y aguarda su plena solución. Debemos ver cómo, en el complicado contexto del desarrollo histórico ideal de la democracia moderna, tal problema se hace íntegramente soluble, en virtud de un método muy distinto del racionalista-voluntarista-abstracto- espiritualista-humanitario de Rousseau. Tal método, enunciado anteriormente, como el de un racionalismo concreto, materialista, es el método del socialismo científico”. 32 Es decir que a los funcionarios no se les asigna una categoría especial dentro de la sociedad sólo por que sean técnicos de los distintos niveles del aparato del Estado (desde un punto de vista formal, como lo aclaró Gramsci, todos los hombres integrados al proceso productivo pueden ser considerados como técnicos con distinto grado de calificación de su fuerza de trabajo) sino, lo que es una cuestión de naturaleza totalmente distinta, porque - y en la medida en que - cumplen una función política. De allí, dice Rubin, a la economía política corresponde la labor de explicar la génesis de la burocracia y de manera correlativa a lo anterior, la vigencia histórica de ésta. 144 K. Kautsky. Parlamentarismo y Democracia. Marx, al analizar este aspecto de la experiencia comunera, anotaba: “La Comuna de París habría de servir de modelo a todos los grandes centros industriales de Francia. Una vez establecido en París y en los centros secundarios el régimen de la Comuna, el antiguo gobierno centralizado tendría que dejar el paso también en provincias al gobierno de los productores por los productores. En el breve esbozo de organización nacional que la Comuna no tuvo tiempo de desarrollar, se dice claramente que la Comuna habría de ser la forma política que revistiese hasta la aldea más pequeña del país y que en los distritos rurales el ejército permanente habría de ser reemplazado por una milicia popular, con un plazo de servicio extraordinariamente corto. Las comunas rurales de cada distrito administrarían sus asuntos colectivos por medio de una asamblea de delegados en la capital del distrito correspondiente y estas asambleas, a su vez, enviarían diputados a la Asamblea Nacional de Delegados de París, entendiéndose que todos los delegados serían revocables en cualquier momento y se hallarían obligados por el mandato imperativo (instrucciones) de sus electores. Las pocas pero importantes funciones que aún quedarían para un gobierno central no se suprimirían, como se había dicho, sino que serían desempeñadas por agentes comunales y, por tanto, estrictamente responsables. No se trataba de destruir la unidad de la Nación, sino, por el contrario, de organizarla mediante un régimen comunal convirtiéndola en una realidad al destruir el poder del Estado” (Ob. Cit.). 35 K. Marx, La guerra civil en Francia (Ob. Cit.: 501). Allí, Marx precisa la diferencia capital entre la Comuna y las antiguas formas federativas del feudalismo: “Generalmente, las creaciones históricas completamente nuevas están destinadas a que se las tome por una reproducción de formas viejas e incluso difuntas de la vida social, con las cuales pueden presentar cierta semejanza. Así, esta nueva Comuna, que viene a destruir el poder estatal moderno, se ha confundido con una reproducción de las comunas medievales, que primero precedieron a ese mismo Estado y luego le sirvieron de base. El régimen de la Comuna se ha tomado erróneamente por un intento de fraccionar en una federación de pequeños Estados, como lo soñaban Mostesquieu y los girondinos, esa unidad de las grandes naciones que, si bien en sus orígenes fue instaurada por la violencia, hoy se ha convertido en un factor poderoso de la producción social. El antagonismo entre la Comuna y el poder del Estado se ha presentado equivocadamente como una forma exagerada de la vieja lucha contra el excesivo centralismo”. 36 En mi revisión de la extensa literatura weberiana que he podido consultar primordialmente en inglés, francés y español, ya que no leo alemán, no he encontrado referencias explícitas a esta discusión de Marx sobre el argumento hegeliano ni a la propia visión de Hegel sobre la burocracia. La primera omisión es claramente explicable, aunque Marx en la Ideología alemana menciona el tema, su principal textoreferido, sólo fue publicado después de la muerte de Weber. De otro lado, las biografías que he podido 33 34 145 abordar no mencionan alguna conexión o influencia de las teorías hegelianas en la fase formativa de Weber. Aunque la Filosofía del Derecho de Hegel era un texto de común recibo en los estudios jurídicos y no sólo en el ámbito filosófico, en el medio académico alemán de finales del siglo XIX. 37 Filósofos liberales como K. Popper (1974) anotaron acremente que la genealogía del totalitarismo estaba en la exaltación hegeliana del Estado y la Nación, y... en la tradición utopista que en Occidente inauguró Platón en la «República». En ambos casos, el individuo resultaba aherrojado por la totalidad de lo público-estatal. 38 Sobre Weber una salvedad. Como su teorización sobre la burocracia es suficientemente conocida y referida en numerosísima bibliografía, me eximo de hacer aquí una presentación in extenso de sus cuadros analíticos (Sobre el tema véanse: S. Kalberg, 2002; J. Freund, 1991; P. De Gay, 2000; N. Mouzelis, 1970). 39 Véase, por ejemplo, la peculiar lectura de M. Crozier (1964) sobre el paradigma weberiano. Tales críticas a Weber llevan el “individualismo metodológico” a extremos de formalismo –la nebulosa de las negociaciones intencionales de los “actores”– que deslindan la teoría de la burocracia del rico y fino análisis contextual e histórico del que hizo gala Weber en numerosos textos y, en especial, en su obra mayor, Economía y Sociedad. Empero, el problema no está sólo en la interpretación de Crozier, sino en el punto de partida weberiano que Crozier radicaliza. 40 Esta tendencia weberiana le aleja notoriamente de la tradición dialéctica que encarnan Hegel y Marx. Como lo afirma S. Kalberg (2000: 57-58), “En la sociología de Weber, las personas son verdaderos actores, capacitados para interpretar las realidades sociales en las que se inscriben y para asumir iniciativas de acciones creadoras”. Kalberg subraya, en ese sentido, que no existe en la visión weberiana ni una teoría de las significaciones objetivas de la realidad, ni ningún espíritu que encarne la Historia Universal. Y nos recuerda una conocida expresión de Max Weber (1978): “La acción, entendida como una orientación significativamente comprensible a partir de un comportamiento propio, no existe más que bajo las formas del comportamiento de muchas personas singulares”. 146 Capítulo 3 CONCEPCIONES DE LO PÚBLICO 1. Campos temáticos de lo público La conceptualización en torno de lo público y su correlación con lo privado —en Colombia y América Latina— no ha sido objeto de análisis riguroso en los ámbitos de la ciencia política, la sociología ni de la teoría de la administración pública; para citar sólo campos académicos que tienen pertinencia cercana al tema. Pero, en general, se percibe una suposición básica respecto a que lo público es equivalente a lo estatal y que Administración del Estado es igual a Administración Pública1. Este es, desde luego, uno de los focos centrales de la investigación aquí asumida, de la cual por ahora abordaremos el aspecto atinente a la relación espacial entre lo público y lo estatal, señalando que la visión tradicional asimila estos dos ámbitos, al punto que vemos ciertas reformas de Estado como procesos de privatización simplemente porque se considera el ámbito de lo público como correspondiente a las funciones del Estado. Para mirar de cerca esa identidad de lo estatal y lo público debemos rastrear históricamente cómo se formó el concepto de lo público. Considerado lo público desde una perspectiva distinta a esta simple sinonimia con lo estatal, pudiéramos señalar tres campos analíticos o líneas de trabajo que han venido escudriñando sobre lo público y su relación con lo privado. La primera de estas líneas ha sido la historiografía moderna; particularmente la llamada historia de las “mentali147 dades”, básicamente desarrollada en Francia y que —con algunas variantes— ha logrado avances en Inglaterra, Estados Unidos e incluso en América Latina. En particular, desde el punto de vista conceptual de lo que significan lo público y lo privado, señalemos los trabajos que fueron publicados con el título de Historia de la vida privada por dos grandes historiadores franceses de la escuela de los Annales, Philippe Ariès y Georges Duby (1990). En efecto, en una compilación de varios volúmenes, Ariès y Duby impulsaron un análisis monográfico muy exhaustivo de la vida privada, desde la Antigüedad hasta el siglo XX2. En la discusión de esta temática no podemos referirnos de manera extensa a la citada obra, porque desbordaríamos los límites y el perfil analítico de este texto3. Sin embargo, un par de trabajos allí incluidos son interesantes para nuestros propósitos. El primero de ellos, el propio prólogo de Philippe Ariès, al tomo sexto de dicha compilación. En segundo lugar, en ese mismo volumen, el ensayo de Jaques Revel sobre Los usos de la civilidad. En general, esta monumental compilación de Ariès y Duby es sólo una muestra del creciente interés de la ciencia histórica por rastrear la vida privada, la cotidianidad, ocultas tras la predominancia de la historia diplomática, institucional, política, económica, etc. Esta línea historiográfica destaca cómo el sentido moderno del concepto de lo público nace hacia finales del siglo XVI y comienzos del siglo XVII. Según Philippe Ariès: “El estudio sobre la vida privada se centra en el cambio de sociabilidad, en la sustitución de una sociabilidad anónima de la calle, el patio del palacio, la plaza, la comunidad; por una sociabilidad restringida que se confunde con la familia o también con el propio individuo. Por lo tanto, el problema está en saber cómo se pasa de un tipo de sociabilidad en la que lo privado y lo público se confunden, a una sociabilidad en la que lo privado se halla separado de lo público” (Ariès, 1990). De este modo, la privacidad se conforma en los orígenes de la Modernidad al establecerse una separación entre los espacios público y privado. Es decir que la separación del ámbito privado respecto a lo público no es una regularidad de las sociedades humanas sino una característica que aparece sólo en ciertos períodos específicos. Un segundo aspecto en la relación público-privado, estriba en que lo público aparece como el Estado o el servicio del Estado y, por otra parte, lo privado, como 148 lo particular, lo que se sustrae del Estado. Relación de implicación que deviene de un contexto medieval signado por una comunidad fuerte y un Estado débil; marco donde la individualidad incipiente no se pudo afirmar. En este sentido, el individualismo moderno sólo se constituyó con base en un arbitraje público-estatal que garantizara los derechos individuales, tales como el derecho a la seguridad de la vida, el derecho a la propiedad, la libertad de locomoción por el territorio y el derecho a presumir la inocencia del acusado y su juicio imparcial por tribunales del Estado; o sea, la eliminación de la justicia privada y de la venganza de sangre, característica de las sociedades en las que el Estado no existe o es débil. Ariès coloca ejemplos que evidencian cómo en la Edad Media, y en muchas otras sociedades en las que el Estado es débil o simbólico, la vida de cada particular depende de las solidaridades colectivas o de dominios que desempeñan una función de protección. No se tiene nada, ni siquiera el propio cuerpo, que llegado el caso no se halle en peligro y cuya supervivencia no esté supeditada a un vínculo de dependencia. Bajo tales condiciones, el Estado —débil y fragmentado— no puede garantizar los derechos ni la seguridad de los individuos. La figura del individuo está puesta o subdimensionada por su relación gregaria y comunitaria. En suma, la argumentación de Philippe Ariès lleva a dos puntos. Uno es el que acabo de mencionar, donde lo público aparece como estatal y lo privado como opuesto a lo estatal, fenómeno que ocurre en el tránsito de la sociedad feudal europea a la Nación-Estado moderna. Proceso típico de la Modernidad, y por lo tanto, condición sine qua non para la existencia del individualismo burgués moderno. El otro tópico consiste en mirar lo público como un campo de sociabilidad, proceso que va a permitir la separación de lo privado como un campo desligado y opuesto a dicha sociabilidad. Se trata de dos fenómenos diferentes: el primero implica una mediación política; y el segundo ubica el proceso de conformación de las formas comunitarias de sociabilidad. No es que en lo privado no existan ciertos niveles de sociabilidad sino que existiría una especie de sociabilidad macro colectiva, común, y otras menores que se segmentan y se oponen a lo público. Se comprueba así el despliegue de una nueva corriente enfrascada en la elucidación de lo social que dentro de este campo enfatiza el estudio de la vida cotidiana (bajo muchas formas: historia familiar, de género, etc.), con lo cual se constituye una investigación histórica muy prolija sobre la vida privada. 149 De otro lado, la historiografía tradicional de tipo institucional también ha trabajado la sinonimia de lo público con lo estatal; es decir, que allí la información historiza la administración del Estado como administración pública, lo cual es, desde nuestra perspectiva, una limitación. Un segundo campo de estudio sobre lo público corresponde a la sociología. En particular me refiero a la gran Sociología Clásica de final del siglo XIX y comienzos del siglo XX, cuya obra arranca con los aportes de destacados teóricos (v. gr. Tonnies, Spencer, Durkheim) y se corona en los trabajos de Max Weber, particularmente con sus estudios histórico-sociológicos sobre el carácter de la burocracia, el origen del funcionarismo y el papel del Estado. Como se sabe, en el pensamiento sociológico existe una teoría sobre el funcionarismo cuyo influjo ha sido dominante a lo largo del actual siglo. Según este punto de vista comparado con el funcionarismo de la sociedad antigua, el moderno es mucho más complejo. Weber enumera una serie de características que tipifican el funcionarismo típico contemporáneo. Son básicamente las siguientes: empleo asalariado; carrera burocrática de ascensos; subordinación y ordenamiento jerárquico; funciones claramente prescritas en códigos y manuales; formalismo documental, órdenes por escrito; constancias de que cada uno de los pasos se realice y de que todo ello se pueda revisar por los superiores jerárquicos. Además, un rasgo esencial del moderno funcionario estriba en que éste es un trabajador intelectual; alguien que desempeña su papel en función de sus destrezas y saberes, constituyendo dicha condición una de las grandes diferencias frente al anterior trabajador manual de la sociedad premoderna. Quienes han discutido el problema de la burocracia, necesariamente repiten o retoman estos elementos. Weber hizo, en este sentido, una síntesis muy lograda en los finales del siglo XIX y comienzos del siglo XX de los rasgos del funcionarismo moderno. Además Weber no solamente fue sociólogo sino un profundo conocedor del campo histórico destacándose como especialista en sociedades antiguas. En sus trabajos indaga sobre la naturaleza de la administración pública de los viejos imperios primigenios (la civilización china, hindú, egipcia, etc., donde existió un gran aparato estatal que controlaba el regadío, la construcción de 150 obras monumentales, templos, la canalización, el sembradío, los ejércitos, etc.) con el empleo de una compleja burocracia pública. De otro lado, en fechas más recientes, podremos destacar el brillante aporte del sociólogo e historiador alemán Karl Wittfogel que, en su texto clásico El despotismo oriental (1966), desarrolló una explicación teórica —sobre la que volveré más adelante— de la subordinación de la administración pública al poder despótico en las sociedades antiguas. Es decir, no solamente retomó los elementos tradicionales del análisis weberiano, sino que los extendió a una consideración de carácter genérico donde incluso los casos de las civilizaciones precolombinas (incaica y mesoamericanas: maya y mexica), fueron analizados proponiendo una serie de invariantes estructurales. Estas teorías sobre el despotismo antiguo fueron discutidas de manera muy intensa hacia los años cuarenta y cincuenta en Europa y los Estados Unidos, ya que muchos investigadores encontraron que en el modelo llamado por Marx “modo asiático de producción”4 estaba prefigurado el totalitarismo del siglo XX. Es decir, que el nacional-socialismo, el fascismo italiano y los propios Estados comunistas poseían rasgos comunes a las sociedades antiguas despóticas: una concentración de la propiedad pública en el Estado, una disminución del papel de la propiedad privada hasta llegar —en algunos casos— a una anulación de la misma, una centralización vertical, autocrática de la red de funcionarios públicos y del poder político-militar, etc. En el caso de las sociedades modernas, ya no se basaba el control político sólo en el manejo centralizado de los recursos hídricos, sino en el monopolio de los principales recursos tecnoproductivos como la industria moderna —que llamamos estratégica—, el monopolio sobre la producción del conocimiento sofisticado, la física nuclear, etc. Todo ello en manos del Estado despótico. Recordemos el conocido aforismo leninista según el cual el socialismo era igual a “el poder de los soviets, más la electrificación”. Así, el totalitarismo pudo ser analizado con herramientas conceptuales que lo parangonaban con los antiguos despotismos de los Estados prístinos. Hannah Arendt, filósofa y socióloga alemana emigrada a los Estados Unidos, fue tal vez quien primero teorizó esta relación y acuñó 151 el concepto totalitarismo, que no equivalía simplemente a la descripción del fascismo. Precisamente, uno de los rasgos que también igualaba al totalitarismo contemporáneo con los antiguos despotismos era la identificación de lo público con lo estatal y, por lo tanto, la desaparición de la sociedad civil (H. Arendt, 1976). Arendt caracteriza el totalitarismo ante todo porque anula y destruye la sociedad civil; es decir, la esfera de lo privado. Bajo este modelo de Estado policial, los asuntos que supuestamente son de la esfera privada resultan reglados, intervenidos y sancionados por el poder supremo. De este modo, se constituye una dominación policial sin ningún resquicio democrático donde los ciudadanos no pueden ejercer sus derechos, ya que éstos son permanentemente vulnerados por el Estado. En suma se conforma un régimen de Estado carente de controles ciudadanos, que utiliza sistemáticamente el terror. La tercera línea de trabajo que ha investigado el problema de lo público ha sido la Ciencia Política, particularmente en los campos de la teoría del Estado y en la teoría de la administración pública. Podemos afirmar que el origen de la administración pública se dio en el marco de estos Estados despóticos antiguos. Y esta aclaración tiene sentido en tanto los manuales de administración pública y otros textos académicos, fundamentalmente escritos en los Estados Unidos, afirman que fue el expresidente norteamericano W. Wilson, en los años ochenta del siglo XIX, en un pequeño artículo, quien fundó la disciplina académica de la administración pública. Posteriormente, a comienzos del siglo XX, en los Estados Unidos, Taylor, y en Francia, Fayol, un ingeniero, elaboraron teorías sobre el problema de la serialización de la producción, la racionalización del trabajo, la estructura de delegación y jerarquización, etc., que se asumieron como elementos constitutivos de la moderna doctrina administrativa. Parecería en estas visiones como si el Estado, durante miles de años, no hubiera tenido una disciplina, una racionalidad para estructurar la compleja división del trabajo y la administración de los asuntos generales de la sociedad para beneficio de una minoría. Lo que sucede es que la administración pública, como disciplina académica, con una autonomía y especificidad relativa en su objeto de estudio, es un tanto reciente, pero como práctica 152 social, es muy antigua. Incluso, la tradición de reflexión sobre la organización del aparato público del Estado se remonta a los propios orígenes de la civilización, y por ello encontramos elaboraciones muy sofisticadas en los propios filósofos de la Grecia clásica. Obviamente, la administración pública como fenómeno social y como ámbito académico ha tenido vaivenes, se ha mezclado con procesos de privatización de lo público, etc. En general, las estructuras básicas de identificación de la administración pública con la función del Estado le corresponden a estas sociedades despóticas, pues ha sido históricamente hablando la forma de sociedad en la que la relación de identidad entre lo público y lo estatal ha significado de forma acusada una anulación del universo de lo privado. Quizás el influjo mayor en este último énfasis lo hallemos en dos clásicos estudios sobre la administración pública del Estado moderno: en los trabajos de Marshall Dimock, La administración pública; y de Fritz Morstein Marx, Las funciones sociales de la administración pública. Textos inmersos en una reflexión hecha por la academia norteamericana a mediados del pasado siglo encaminada a elucidar la praxis del gobierno en la sociedad post-industrial y cuya edad de oro se situó en la segunda postguerra. Estos estudios fueron pioneros en la conformación de la administración pública como disciplina académica en las universidades norteamericanas y, en esa misma medida, influyeron en América latina marcando el paradigma de dicha disciplina académica. Adicionalmente, en el contexto latinoamericano las escuelas dominantes en las pasadas décadas —la cepalina, la teoría del desarrollo y la protomarxista de la teoría de la dependencia— han producido un importante corpus analítico que avanza en el campo que pudiéramos denominar situacional, de lo cual son ejemplos los estudios de Faleto, Jaguaribe, Cardoso, Medina Echavarría, Ozlack, O’Donnell y otros. De otro lado, sobre lo público propiamente dicho, hay también trabajos sociológicos que anudan el tema en una perspectiva diferente a la propiamente política; es quizás uno de los más importantes el producido en Norteamérica hacia los años cincuenta por Irving Goffman (1979). El estudio de Goffman sobre lo público escapa a la restricción temática que equivale lo público con lo estatal y asume lo público en la perspectiva más genérica de la vida pública, 153 del espacio público y la representación social. Goffman en particular estudia los problemas de interpelación y construcción de la personalidad a partir de la relación comunitaria. Lo público sería entonces el espacio en el cual las personalidades se configuran en su interacción, en su negación y en su afirmación. Estas son las dos grandes tendencias en la definición de lo público. En la última dirección anotada, lo privado sería un ámbito protegido de vida, que no tiene por qué exponerse a la vista de todos. Público y privado poseen un amplio campo semántico correlativo. Público: lo permitido al acceso de la comunidad o, de otro lado, lo público como bien común. Privado: lo que está oculto o resguardado; de lo cual se desprende, entre otros aspectos, la postulación de la intimidad como derecho o, de otro lado, lo privado como espacio legítimo de las posesiones de los individuos, igualmente consagrado como fundamento legal del orden social5. 2. Génesis y desarrollo histórico de la contraposición público-privado En cualquiera de los modelos y perspectivas señaladas que se pueden enfatizar en mayor o menor grado, existe una conexión necesaria entre lo público y lo estatal, la cual ha sido una constante en la articulación entre una comunidad y su correspondiente aparato de gobierno. En efecto, no existe sociedad donde el Estado no tenga un predominio, mayor o menor, sobre los asuntos públicos. Desde luego, vale la pena resaltar que en dicho contexto tomamos lo público en sus diferentes especificidades: en cuanto ámbito de discusión, de concertación de intereses societarios que expresan una determinada visión que en una época tiene una comunidad de sus puntos de confluencia; o, incluso, en un sentido más moderno: en cuanto se constituye la administración pública como algo que es asumido por el Estado. La segunda idea que podría destacarse en este orden expositivo es que no siempre ello ha significado que lo público equivalga a lo estatal. Es decir, podríamos hablar de una relación —como lo hace la teoría de conjuntos— donde un ente se intersecciona con otro pero no lo comprende en su totalidad, no existiendo una relación simple y pura de identidad. Es necesario que exista la intersección entre lo público y lo estatal para que lo estatal se constituya como tal, pero esto no implica que lo público quede cobijado bajo lo estatal. 154 En realidad, existen dos grandes concepciones de lo público. En primer lugar, lo público como bien común, donde entraría el concepto latino de res publica. En este sentido, lo público, de una forma u otra, remitiría a un cuerpo político. Esta es la base histórico-social que ha permitido en diferentes momentos de la Historia identificar lo público con estatal. En esta perspectiva podríamos afirmar que esta relación del bien común con lo público no corresponde a la Modernidad; aparece ya en las formas sociales comunitarias en los antiguos imperios despóticos, en Persia, Egipto, el Imperio Romano; es decir, en las matrices de la civilización occidental, y posteriormente, en Grecia y Roma. En este sentido la obra de los estoicos, por ejemplo, Séneca y Cicerón o los jurisconsultos romanos de la época del Imperio, sería altamente ilustrativa. 2.1 ORÍGENES DE LA HEGEMONÍA DE LO PÚBLICO-ESTATAL Sabemos muy bien por la Historia de las demás civilizaciones que Europa advino tardíamente al escenario. Como claramente lo anotó el historiador alemán ya mencionado, Karl Wittfogel (1966), las siete grandes civilizaciones originales o prístinas fueron configurándose en —diferentes temporalidades— a través de complejos procesos evolutivos —desde las comunidades primitivas, tribales y nómades— en asentamientos no europeos. Wittfogel lista siete: China, la India, el Egipto antiguo, la civilización Sumerio-babilónica; dos americanas: el Imperio Inca y los imperios mexico-toltecas; y, finalmente, describe una que en tiempos de la Modernidad (siglos XVI-XIX), estaba en trance de configurarse en el África ecuatorial, y que, al igual que las americanas, fue descuajada por el colonialismo europeo: la civilización Zulú. Tales sociedades milenarias poseyeron complejas burocracias adminstrativas, generalmente bajo la forma de teocracias altamente centralizadas (despotismos). Fueron capaces de realizar obras y redes de infraesctructura de diverso tipo que rendían cuenta de necesidades económicas (regadíos, embalses, caminos y sistemas de comunicación, puertos, etc.), administrativas (ciudades como sedes de megajefaturas administrativas públicas), religiosas y simbólicas (pirámides, templos, etc.). Pueden señalarse evidentemente ciertas fases de la Historia donde la asimilación del Estado frente a lo público se ha acrecentado, y otros períodos en los que se ha debilitado mucho el control del aparato estatal sobre el ámbito público. Incluso ello ha ocurrido en períodos de forma concomitante al fenómeno de debilitamiento del propio espacio social de lo público. 155 En la sociedad antigua llamada por K. Marx asiática y que autores posteriores, críticos de la formulación teórica de Marx, como Karl Wittfogel, denominaron despotismo oriental o hidráulico, en los momentos en los cuales se comenzó a configurar el Estado, se evidenció el rasgo de absorción de lo estatal sobre lo público. No hay una simultaneidad en estos casos, porque el Imperio egipcio es muy temprano, sus bases se sientan en el milenio tercero antes de la era cristiana, mientras que los toltecas-mexicas y los incas, apenas estaban configurándose como Estados a la llegada de los españoles. Es decir, no existe una época antigua común, en la perspectiva de un tiempo lineal, sino que los procesos de estructuración social pueden ser vistos a través del método de sociología comparativa como expresando irregularidades comunes. Las características esenciales de estos llamados despotismos antiguos básicamente son las siguientes. En primer lugar, un débil o nulo peso de la propiedad privada; en segundo lugar, como resultado o complemento de lo anterior, la propiedad dominante es estatal; el Estado es dueño de la tierra, las aguas, los diferentes recursos. En tercer lugar, la estructuración del Estado obedece a diversos modelos o grados de gobierno despótico6. Pero lo que caracteriza al despotismo, desde el punto de vista político, es el carácter omnímodo, es decir absoluto, del dominio de quienes detentan el poder sobre la comunidad. Hay una cuarta característica que describe Karl Wittfogel cuando señala que los Estados antiguos son sociedades hidráulicas. La clave de que estos imperios se hayan constituido y consolidado de manera estable y en fases de “larga duración” consistió en que una minoría poseída del poder militar o religioso pudo establecer una extracción de trabajo o de tributos en especie a partir del control de las aguas y los recursos fluviales y marítimos. No es casual que la civilización babilónica se haya estructurado sobre el enorme delta formado por los ríos Tigris y Éufrates; o que la civilizaciones hindúes se hayan establecido básicamente sobre el río Ganges, o que en China se haya dado lo mismo en el río Yan-tsé Kyang o Amarillo; y diríamos algo semejante en el caso de los incas y los mexicas7. En todo caso, en cada una de estas civilizaciones antiguas, el uso central de las aguas se daba a la comunidad por la mediación y bajo el control de una minoría despótica. 156 ¿Cómo fue posible este dominio, estamental, despótico, sobre la comunidad? En realidad, no era posible estructurar en dichas comunidades pequeñas unidades de producción autosuficientes y autónomas. Era hasta cierto punto, inevitable y necesario, por la geotropía, por la relación entre el hombre y los recursos naturales, el establecimiento de lazos de cooperación social para controlar, por ejemplo, las crecientes e inundaciones de los ríos, lo que implicaba construir canales, taludes, exclusas. En épocas de sequías muy fuertes, de manera contraria, habría que construir sistemas de riego, etc. Todo ello sólo se podía dar en ese tipo de sociedades sobre la base de una gran centralización política. Esta férrea centralización sirvió para que una minoría de sabios, que podían utilizar conocimientos de agrimensura, de matemáticas y de astronomía (para poder hacer el cálculo de las lluvias y los períodos en que se pudiera desarrollar los cultivos o las crías de ganado, o pastar a través de diferentes circuitos de propiedad) generaron una forma comunal de propiedad que muy pronto se convirtió en una forma estatal despótica. Esta estructura despótica creó el funcionarismo. El déspota, o los déspotas, necesitaron un cuerpo de funcionarios que eran, por ejemplo, personas insertas en diferentes oficios: escribas, contadores, jueces, sacerdotes, cortesanos, cobradores de impuestos; en fin, sabios, elites intelectuales, literatos. La característica de todo este cuerpo de personas es que eran empleados del Estado, funcionarios vinculados a la corte y dependientes verticalmente de la voluntad del grupo dominante y del déspota. El terror es una clave esencial del despotismo. El terror en las sociedades antiguas lo describe muy bien Wittfogel desde el sacrificio de las vírgenes y las núbiles en las ceremonias religiosas, desde el sacrificio de los prisioneros de guerra o de los rivales políticos, todo aquél que estuviera en contra del Estado era victimado. Y había un terror también sobre la población civil por razones económicas. Para organizar la gran estructura productiva de los despotismos, se requiere la utilización de medios coactivos sobre el común de la población. El ejército y los funcionarios militares ejercían un control coactivo que en los despotismos antiguos fundamentalmente se estructuró en torno a la esclavitud, fuerza de trabajo cautiva para el desarrollo de las labores productivas. 157 2.2. “OCCIDENTE”: FUENTE EMANATORIA DE LA DICOTOMÍA PÚBLICO-PRIVADO Siempre que nos interrogamos acerca de la etiqueta que aún los propios latinoamericanos portamos como miembros de la civilización Occidental, la pregunta inmediata que salta a la vista es: ¿cuál es la ontología que define a esta civilización, hoy dominante y avasalladora? La civilización eurooccidental fue ciertamente tardía, y probablemente una comparación de estas dimensiones de la vida social no muestre ninguna superioridad del mundo griego respecto a sus pares. De acuerdo a lo planteado por autores como Gusdorf, Jaeger, De Konik y otros, la clave de la diferencia del Occidente emergente estuvo en la forma singular como apareció en la vida griega, durante unos pocos siglos (-IV a I (-)) el pensamiento libre. Es decir, la posibilidad de sobrepasar el peso dominante del universo mítico y religioso, que en estas sociedades prístinas estuvo siempre instrumentalizado como un conjunto de discursos del poder. Es cierto, salvo excepciones singulares, las ciudades-Estado del Mediterráneo no fueron teocracias. Y el peso de la religión obedeció al ámbito de la vida privada, al fuero interior de los individuos. Agregaría a lo dicho por estos autores, la interpretación que al respecto elaboró F. Hayek (1990). Según el economista austríaco, la otra pieza del enigma fue la aparición en esa época y en ese espacio de la institución social de la propiedad privada. Ciertamente, hoy es un anacronismo teórico la explicación rousseauniano-marxista sobre el origen del Estado. Como se recordará, tanto Rousseau como Marx identificaron al Estado como un aparato que surge para defender la acumulación asimétrica de la propiedad privada en un polo privilegiado de las sociedades. La antropología y la historiografía han comprobado y enriquecido las ideas de Wittfogel. El Estado surgió precisamente atado, anudado a la existencia misma de la propiedad pública, como factor que permitía la vida pública, en un esquema tal que la propia vida privada carecía de sentido. La originalidad griega no es su papel pionero en la constitución del Estado y la civilización. Su punto de quiebre fue el hecho de que en Grecia, por vez primera, aparece la dimensión de la privacidad, que será el rasgo central del mundo occidental. Privacidad económica: derecho privado garantizado, propiedad fundiaria, herencias, acumulación de riquezas, comercio (el factor 158 catalizador según Hayek). Propiedad intelectual: autores (no relatos anónimos), debates (el paradigma del sofista); universo ético (el ámbito de la decisión libre); limitaciones severas de los poderes religiosos, políticos y militares sobre la propia sociedad. La misma idea de polis refleja muy bien esa relación que Platón describía como dike o justicia: la interdependencia y satisfacción recíproca de necesidades. Una justicia que no pertenecía al reino de la igualdad (la justicia con equidad de J. Rawls y el pensamiento liberal contemporáneo), sino que era una justicia retributiva. Cada cual aporta desde ‘su lugar en el mundo’ y recibe de acuerdo a su propia ontología; los esclavos procuran medios materiales a la ciudad y reciben de ésta su proteccion y la garantía de su sobrevivencia. La vida política es el coronamiento de la especificidad de Occidente. En los antiguos imperios había poder —pleno, absoluto—; había vida pública, comercio, circulación de las lenguas, intercambios sociales, vida urbana, etc.; pero no había vida política. Ya que ésta implica la interrogación sobre el poder, su cuestionamiento, la necesidad de fundarlo discursivamente para rendirlo legítimo. La tradición occidental siempre, desde sus orígenes griegos hasta nuestros dias, ha estado marcada ontológicamente por la dinámica tensa y contradictoria entre lo privado y lo público, entre el individuo y la sociedad (H. Arendt, 1958). Pero no pretendo sugerir una contraposición simplista entre Occidente y Oriente (véase E. Said: 1978) en los términos de Huntington (1996). Más bien, podría vislumbrarse esta relación como un proceso comunicacional, político, económico, etc., donde la cultura occidental le asignó recurrente importancia al individuo, al sujeto —incluso en el plano gnoseológico— respecto del horizonte de vida más gregario de las restantes civilizaciones. También, desde la misma genealogía occidental, podremos comprender la otra gran dimensión de la vida pública. En efecto, una segunda categoría de lo público consiste en definirlo como aquello expuesto o manifiesto a la observación general, abierto a la indagación de todos. Sentido usual cuando se habla de documentos públicos por oposición a los privados (por ejemplo, las actas de las sociedades del Estado o las cuentas de las empresas); todo esto se supone debe ser del dominio público o potencialmente público. La relación específica de la vida política con la vida pública fundó 159 igualmente el nacimiento de la Filosofía como el campo discursivo por excelencia de la busqueda de la verdad. Como lo vimos en las discusiones y lectura de la primera parte, esta instalación de Occidente en referencia al saber como verdad discursiva, requirió la ruptura de la representacion mítica. Los tiempos heróicos en Grecia, dan cuenta de un pasado nebuloso; pero no referencian la propia civilización helénica. De otro lado, esta construye el campo epistémico de la verdad no como el objeto de una reflexión intelectual individual, autocentrada, sino que los diferentes individuos (intelectuales o filósofos) debían acudir a las pruebas de su demostración. Sólo con la aceptación social es posible la verdad. Ella es, por lo tanto, una de las bases esenciales de la sociabilidad. El propio reconocimiento al que se refería Aristóteles se definía así puesto que la identidad de cada individuo existe en tanto producida como mecanismo de asignación social. El juicio y condena de Sócrates revela ciertamente en forma emblemática el paso fundacional de la verdad controlada por la totalidad, en términos de una cierta hegemonía moral de la ciudad frente al desarrollo posterior del pensamiento libre. La dialéctica y la mayéutica como herramientas de construcción de la verdad son, en ese sentido, en el mundo griego, marcantes de la condicionalidad social de su productividad. Desde luego, no creo que esto suprima los nexos existentes entre el saber y el poder descritos en nuestra época por Foucault (2004). Debemos precisar que esta relación se da en un contexto abierto, relacional, dado que existen numerosas vinculaciones entre el conocimiento y el poder y que, además, el poder político no es el único tipo de poder existente en las sociedades de tradición occidental, pese a la persistencia en el tiempo de los órdenes monárquicos, dictatoriales o a las propias restricciones de la democracia constitucional, paradigma de la vida pública hoy en día. 2.3. DOMINIO DE LA PRIVATICIDAD EXTENDIDA EN EL MUNDO MEDIEVAL Posteriormente, devino un período muy largo, de siglos, en el que decayó el poder de los viejos despotismos; y es esta fase lo que llamamos Edad Media o Medioevo, en Europa occidental, el norte de África y otros lugares. Ocurrió entonces una fragmentación del Estado; y una especie de vuelta hacia lo privado. 160 Hay que tener mucho cuidado con el uso de la expresión privacidad o privado en la Edad Media, porque lo privado moderno no es equivalente a la privacía medieval, puesto que la característica de la privacidad moderna es el individualismo. No se puede hablar de un mundo de lo privado en la sociedad medieval como opuesto a un mundo de lo público. Alternativamente existió un mundo de la privacidad, en el sentido de que el Estado no intervenía en una serie de asuntos que eran de competencia de las comunidades y de la familia, lo que determinó estructuralmente a la privacía medieval por oposición a lo estatal. La característica de la privacía medieval fueron las prerrogativas que se establecieron entre el mundo comunitario, el mundo familiar y el régimen eclesiástico. La llamada familia total es una red en la que no está la figura de la pareja que se aísla, forma sus hijos, tiene su propia esfera, sino que en una misma casa, hacienda, estancia o explotación, coexiste un tronco común vinculado en torno a la autoridad patriarcal. Se trataba de familias numerosas que involucraban no sólo a los familiares consanguíneos o afines, sino a los criados. En esa familia no existió privacidad individual porque el gobierno de la misma estaba en manos del patriarca familiar. Los asuntos de la pareja, o los asuntos del individuo, el paso a la mayoría de edad, todo ello no logró zafarse del control de la familia patriarcal o de la comunidad; esta última vista como un agregado de familias, establecida sobre la base de la jerarquía de los patriarcas. Lo que pudiéramos llamar privacidad, desde el punto de vista de nuestro análisis, básicamente sería una limitación de tipo negativo; fue el hecho de que no se dio un control estatal total desde arriba porque el conjunto de la sociedad se guiaba por el derecho natural. Cuando se lee a los teóricos cristianos del derecho natural, los grandes clásicos de la patrística: Santo Tomás, San Agustín, San Buenaventura y San Bernardo; o a los teóricos jesuitas de la fase final de la Edad Media: el padre Suárez, Juan de Mairena, etc., todos ellos suponían una relación contractual que vinculaba a los soberanos con el pueblo. Las teorías contractualistas no las inventaron los liberales del siglo XVII. Son teorías que se anclan en la Edad Media. Lo que pasa es que el contrato social de la Edad Media no fue un contrato entre individuos para cederle el poder al Estado; fue visto como un contrato entre seres creados por Dios, y en tanto creados por Dios, todos poseían derechos naturales y le delegaban al monarca el ejercicio de una autoridad que se asemejaba a la paterna. 161 La iglesia románica, desde antes del cisma, defendió como argumento central el libre albedrío. Éste fue una piedra angular del catolicismo que empezó a ser construida con San Agustín. Desde el punto de vista filosófico, una de las características esenciales por las que se señala una diferencia en los orígenes de la polémica entre las corrientes del protestantismo y el catolicismo, fue la defensa que hizo el catolicismo del libre albedrío. Según esta doctrina, cada cual construye su propia salvación y su perdición. En eso se puede ver un embrión de individualismo. Cada individuo define si se hunde en el pecado, no por otro, sino por sí mismo ( Huizinga, 1979). ¿Quién ocuparía entonces el espacio de lo público en la sociedad medieval? El espacio de lo público se construyó fundamentalmente a partir de las instancias de la corporación. Oficiantes, artesanos, constructores de caminos; no sólo la corporación como estamento. De un lado, estaba el mundo feudal: la corte, las diferentes jerarquías o gradaciones de la nobleza. De otro lado, el clero, también con sus distintas gradaciones. Y, finalmente, un estamento que no alcanzaba el nivel de corporación: la plebe, los siervos. En estos dos niveles, la corporación clerical y la nobleza, se puede encontrar un cierto embrión de lo que llamaban república, en el sentido de la discusión y la formación de una voluntad colectiva del Estado, del poder público. Valga aclarar que la Iglesia ejerció un monopolio sobre lo público, fundamentalmente en el campo de la ideología y en el tutelaje de la moralidad. La ideología y la moral fueron controladas no por el Estado sino por la Iglesia, a la manera de una suerte de límite a las posibilidades de la propia comunidad, de los vínculos de familia, del patriarcalismo y, posteriormente, de los estamentos que trataron de construir un espacio distinto, como ocurrió, por ejemplo, en la época del Renacimiento. Se puede aún percibir la tensión de los renacentistas en función de toda la preeminencia, el auge y, a veces, la decadencia del papado romano. La elite intelectual de la Edad Media, en todo el mundo occidental cristiano, fueron los sacerdotes, ya que prácticamente toda persona que quisiera acceder a una universidad, ejercer la medicina u ocuparse de una labor intelectual como la escribanía, etc., tenía necesariamente que convertirse en clérigo. Por lo tanto, la Edad Media, que estuvo lejos de ser una edad de las oscuridades y la ignorancia, tuvo elites cultas que deliberaban a través de medios de intercambio con una 162 lengua erudita, el latín, y a través de medios de intercambio por excelencia: las epístolas que se cruzaban los sabios, el intercambio de libros copiados a mano entre los bibliotecarios de esta elite, etc.; todo ello representó un embrión de configuración de lo que llamaríamos opinión pública. Hubo, por encima de estas prácticas, otras instancias más regladas como las órdenes religiosas, en donde también se desarrollaron discusiones entre tendencias y matices, enfrentamientos y luchas por el poder; y estaba el gran centro o el vórtice en función de estas fuerzas: los sínodos papales, en los cuales se definía la doctrina, se orientaba políticamente la gran superestructura eclesiástica que tutelaba los Estados y los señoríos. El otro nivel, la corte, fundamentalmente sirvió para estructurar progresivamente lo público en relación con la aparición de la Nación-Estado. El caso más conocido fue Inglaterra, alrededor de la época en que Juan sin Tierra gobierna Inglaterra (año 1250 en adelante), fenómeno que describe muy bien Shakespeare en sus obras sobre el proceso de formación de la Nación-Estado en Inglaterra, escenificándose dicha lucha en la configuración de las dos cámaras mediante un doloroso proceso de disputa entre los nobles y la Corona. En otros escenarios alternativos, valga decir en aquellos correspondientes a las luchas sociales, el estamento de la nobleza desarrolló el drama de la pérdida de un poder que fue centralizándose en el Estado, abatiendo los poderes dispersos de las localidades. En este orden de ideas, se presentaron arduas discusiones sobre los impuestos, la lógica de las guerras civiles, los enconados conflictos entre las iglesias; en fin, todo ese universo sociopolítico fue debatido en términos de igualdad estamentaria por una cúpula que dirigía la sociedad feudal y la nobleza cortesana y que asistía al drama de su propia decadencia. No es que lo privado surja en una relación independiente de lo público. Por el contrario, el sentido moderno de lo privado y de lo público se formó en una dinámica muy lenta que finalmente se aceleró a partir del colapso del universo medieval, crisis revolucionaria que implicó la aparición de la Modernidad y que confluyó finalmente en la conformación de los Estados absolutistas tales como Francia, Inglaterra, España, Portugal, los señoríos germánicos, el dominio del papado, las ciudades-Estado de Holanda, Bélgica, Italia, etc. Adicionalmente, Inglaterra fue pionera de un proceso que permitió, a partir 163 del bicameralismo, el surgimiento de una opinión pública ciertamente en un contexto de nuevas elites que actuaron en la vida social urbana más allá de los ámbitos tradicionales, es decir los círculos cortesanos concentrados políticamente en las cámaras de los lores y los comunes. Desde luego, no fue éste un proceso espontáneo sino que tuvo profundas vicisitudes. Así, muy de pasada, haciendo un sobrevuelo del mundo medieval, podemos señalar que el Medioevo implicó una disminución muy grande del papel del Estado respecto a las sociedades antiguas. Significó igualmente una disminución de la dicotomía entre lo público y lo privado mediante una identificación de estos elementos con la corporación, la comunidad y la familia. 2.4. PUBLICIDAD Y VIDA MODERNA Abordaremos ahora la relación de lo público con la Modernidad, es decir, con el mundo posterior al Medioevo, en el cual surgió un sentido altamente valorativo del individualismo. Trataremos, en consecuencia, de precisar en sus grandes trazos los rasgos más relevantes de esta interrelación. En la Modernidad hubo una diferenciación creciente entre la esfera privada y la esfera pública. Se dieron tres factores que permitieron a la Modernidad un proceso de concentración de lo público hacia lo estatal. En primer lugar, el desarrollo técnico-productivo; en segundo lugar, la acumulación y reproducción del capital; y en tercer término, la preeminencia del Estado-Nación y el surgimiento del mercado mundial. El desarrollo tecno-productivo llevó al Estado a asumir funciones que no existían o que, en la sociedad premoderna, desarrollaban las comunidades y los individuos particulares. Por ejemplo, en las sociedades mercantiles en España en los siglos XVI y XVII, la empresa de la conquista y la colonización de América fue una empresa impensable sin el Estado imperial centralizado español. La colonización de vastos territorios, la exportación de esclavos africanos, el proceso de saqueo y colonización no pudo darse en los términos en que se dio, sin un papel creciente por parte del Estado. En mayor o menor grado (unos Estados intervinieron mucho más que otros), ello tuvo efectos sobre las sociedades de la nueva América. Me refiero particularmente a que Inglaterra no tuvo el mismo grado de centralización política del proceso de 164 emigración y colonización que sí tuvo España y eso marcó diferentes modelos entre Norteamérica y la América hispánica. Básicamente, en la primera fase, la actividad de colonización fue la que signó la función de intervención del Estado y de concentración de funciones públicas. En segundo lugar, el crecimiento de los ejércitos nacionales, con miles y centenares de miles de hombres disciplinados, centralizados, bajo estrictas jerarquías, con instrucción militar relativamente sofisticada fue una hechura de la Modernidad. Ciertamente existieron en las antiguas sociedades despóticas los ejércitos de los faraones, el gran ejército de Alejandro Magno, los grandes ejércitos de los imperios chinos, japoneses, etc., pero luego virtualmente desaparecieron ya que en la Edad Media predominaron los caballeros, que fueron grupos muy pequeños de aristócratas y nobles cuya sola armadura podía valer más que todo el sobreproducto de una aldea completa en la Edad Media. En realidad, el campesinado y los vasallos no participaban en las guerras; las guerras eran entre elites que intervinieron directamente en el sentido de combate militar. De allí el alto valor que ha tenido el honor militar, porque cuando éste se construyó, luchar militarmente fue una excepción. La canalla no participaba en las guerras, y en las luchas y contiendas bélicas actuaban aristócratas que se entrenaban y se educaban como caballeros y que portaban la función real, puesto que manejaban un sable, ostentaban la instrucción militar, y esto era algo que correspondía a ese estamento de los cortesanos. Todo esto cambió con la sociedad moderna puesto que la lógica de los conflictos en los nacientes Estados-Nación impulsó a construir grandes ejércitos que ya no van a ser ejércitos privados de un señorío sino ejércitos de la Nación-Estado que se estructura en función de dichos ejércitos nacionales. En síntesis: la colonización implicó un primer grado de intervención económica. El segundo fenómeno fue el surgimiento de los ejércitos nacionales y el tercero un régimen estatal de tributos y, sobre la base del mismo, la consolidación de una burocracia o funcionarismo público que dependería del Estado. El sociólogo francés Maurice Duverger hace, en relación a este punto, una descripción de la forma en que los reyes logran progresivamente arrebatarle a los nobles la función voluntaria del tributo estableciéndolo como coactivo. Describe cómo en los siglos XII, XIII y XIV había asambleas, que en Francia se llamaron Estados generales, en las que la nobleza se reunía 165 con el Rey y se ponían de acuerdo sobre el grado de donaciones que le harían los nobles a la Corona para sostener una empresa común, que podía ser de la conquista o la defensa de Francia o de Inglaterra, o hacer una cruzada contra los árabes; pero siempre estaba en función esa contribución de la voluntad de los nobles y de una especie de programa o mandato programático contratado entre estos y el Rey. En caso de no ser así los nobles podrían sustraerse de pagar los tributos. Pero luego se suscitó un proceso muy complejo, que no puedo narrar aquí, en donde el tributo se volvió coactivo. Entonces el Estado pudo ejercer una actividad que sería decisiva para su imposición sobre los señoríos y sobre el universo de lo privado: el control de la hacienda pública. De este modo, si algo caracteriza la identificación entre administración pública y Estado, ha el sido manejo de la hacienda pública. Podemos resumir lo dicho hasta aquí señalando que el ejército nacional y la hacienda pública fueron los dos pivotes sobre los cuales se estructuró el Estado moderno y en los cuales podemos ver la identificación entre lo público y lo estatal. Esta lógica social de centralización política nos permite comprender por qué los problemas de turbulencia social más caracterizados a lo largo de los siglos XVI, XVII y XVIII fueron las rebeliones populares frente al cobro de los impuestos y, adicionalmente, las contradicciones de las diferentes facciones dominantes sobre el problema de los tributos. El Estado sólo puede desarrollar planes, ejecutar gestiones y emprender guerras ampliando la base de tributación, exaccionando del plusvalor al máximo nivel posible a la comunidad, al sector productivo. En la medida en que el Estado se fortalece requerirá indudablemente un incremento de los impuestos y estructuras para ello, una función coactiva. Hay otros dos campos de la administración pública que el Estado fue asimilando progresivamente. El primero de ellos, la Administración de justicia, que no llegó muy pronto a convertirse en un monopolio de la Nación-Estado, porque la tradición de la Edad Media y del derecho natural era arreglar los conflictos de acuerdo a las costumbres. Las comunidades tenían una serie de mecanismos muy viejos, de carácter milenario incluso, que permitían zanjar las diferencias (cf. Foucault, 1978). El Estado despojó a la comunidad del ejercicio de la justicia en un proceso bastante dificultoso. Nunca el Estado-Nación pudo hacer que el sentido de justicia privatizado, propio de una sociedad premoderna, lo pudiera asumir el 166 sector público estatal. La vendetta, la venganza de sangre, el automatismo en la condena de quien transgrede una norma social, todos estos fenómenos premodernos aquí se han mantenido. En Europa, éste fue un proceso también muy difícil, muy lento, pero se puede señalar que después del siglo XVIII los Estados nacionales más importantes de Europa habían logrado ya concentrar las funciones de la administración de justicia en manos del Estado central. Un cuarto proceso de relación de lo público con lo estatal fue el proceso, por parte del gobierno central, de sometimiento de las autoridades locales y regionales. Es decir, la estructuración de formas gubernativas en que la centralización política pudo ser plenamente válida porque podía existir un Estado con una concentración de funciones en los aspectos básicos como hacienda pública y manejo del ejército nacional; pero que no tocaba las autonomías de las regiones para el manejo de otro tipo de asuntos como el mismo manejo de la justicia o las relaciones con la Iglesia, (por ejemplo a nivel de las comarcas, de los arzobispados, de las feligresías, las funciones que tendrían los alcaldes o las autoridades locales, formas estas que fueron las últimas que el capitalismo pudo reducir, someter, para estructurar el concepto de soberanía nacional-popular). En el libro de Karl Smitch, Teoría de la constitución, puede vislumbrarse de qué manera la soberanía nacional-popular, que tanto se reclama hoy, sólo pudo construirse abatiendo las autonomías locales nacionales. Por eso resulta tan difícil devolver el proceso, lograr que haya un reconocimiento de parte de la Nación-Estado de la tesis según la cual los municipios son soberanos y autónomos, las regiones. Un punto decisivo, por ejemplo, es el de la soberanía fiscal. Es decir, que una vez que el Estado-Nación ha logrado capturar para sí el derecho a tutelar y controlar los impuestos y la hacienda pública, no va a ceder fácilmente estas prerrogativas a los municipios y a las regiones. La Nación-Estado le arranca y cercena prerrogativas de manejo de administración que les correspondían a las comunidades locales, a los gobiernos regionales, a las provincias. Todo ello se centraliza en el Estado- Nación. Respecto de la conceptualización de lo público y en relación con las anteriores tendencias históricas, en la últimas décadas ha sido conocido el punto de vista de Habermas a través de su opúsculo Historia y crítica de 167 la opinión pública (1962). En esta obra el autor de la escuela de Frankfurt critica, particularmente, la conceptualización kantiana, arriba descrita, sobre lo público. En especial, basa su argumentación en que Kant asume de forma idealizada el intercambio mercantil, propio de los circuitos económicos que Hegel denominaría luego sistema de necesidades, punto de vista que le sirve a Kant, y en general a los ilustrados, como constructo para justificar la ficción de universalidad en el ámbito de lo público; y equivaler así este campo con el espacio natural de los sabios, lectores; círculo en donde mediante la deliberación se lograría la razón ilustrada. Habermas señala que tal visión resulta en realidad restringida sólo a una parte de los ciudadanos activos, puesto que, en su conjunto, la opinión pública aparecería en una forma más difusa, vaga, que la presunción del circuito relativamente estrecho de los “sabios deliberantes”. Quizás podría interpretarse este tipo de crítica como si en el siglo XVIII, la Ilustración, incluso en sus versiones más democráticas, no fuera capaz de romper la camisa de fuerza del denominado despotismo ilustrado. En Europa, el nuevo sentido de lo público apareció acusadamente en el siglo XVIII, en la época de la Ilustración, básicamente en la exhibición pública de las personas que iban a ciertas diversiones y a través de prácticas sociales como las salas de teatro. En París, Londres, Roma y en general en las grandes urbes era signo de distinción asistir a ese tipo de veladas, tales como exposiciones teatrales, pictóricas, audiciones musicales, el café espectáculo, etc., donde se producía no sólo la representación de una estética del drama, de la relación actor-público, sino la cotidianidad de la relación entre las elites8. Estas fueron formas tempranas de la Modernidad en las que se empezó a usar lo público, en el sentido etimológico de el público; entendiendo por tal el espectador que asiste a ciertas funciones y representaciones, y se separa del resto de la comunidad; prácticas aristocráticas que la burguesía copia y que fungen ahora como prácticas de minorías ilustradas. Lo público apareció en dicho contexto como un espacio de sociabilidad diferenciado de manera nítida; pero la sociedad industrial posterior a la Revolución Francesa y a la Ilustración disolvieron esa posibilidad y la volvieron sutil, ya que los escenarios de lo público se convirtieron en accesibles a todos. 168 Adicionalmente, la crítica habermasiana implica que la identificación de lo político con lo moral (supuesto kantiano clásico) desconoce la naturaleza conflictiva y terrenal del citado espacio político, precisamente el real lugar de despliegue de la deliberación pública, que no sería en este sentido una catalizadora del consenso ilustrado sino el escenario de conflicto de intereses de clase. Igualmente, en tal medida, Kant exaltaría mitificadamente el rol de la competencia. En este orden de ideas, Habermas repite las conocidas críticas de las escuelas neomarxistas frente al rol del liberalismo afirmando que éste es sólo un momento de la historia moderna, posteriormente liquidado por el ‘imperialismo y la centralización del poder político’. Ello nos conduce a una discusión de mayor importancia que he pretendido destacar a lo largo de las anteriores páginas y espero se refleje en el conjunto de este texto: se trata de la posibilidad de conciliar una postura que quizás caracterizaríamos con certeza como posthegeliana (desconociendo el lugar de encuentro entre lo ético y lo político que le corresponde al Estado de derecho, y en especial a sus versiones sociales y garantista, respecto a la necesaria construcción de un ámbito de la institucionalidad pública en términos del garantismo de los derechos individuales, en lo cual se conduce la discusión al terreno del liberalismo clásico. El proceso civilizatorio de los últimos siglos —desde el XV— fue, en suma, con todas sus complejidades, un proceso de imposición de una sociabilidad: la europeo-occidental sobre el resto del mundo. Ésta responde a una combinatoria efectuada en primer lugar —históricamente— por el cristianismo que sincretizó los ejes medulares del pensamiento, la vida pública y la estética griega. Y la revolución científico-técnica de la Modernidad, desde los grandes descubrimientos y conquistas de los siglos XV y XVI, creó las condiones de posibilidad de su expansion hegemónica, como la empresa del mundo euroccidental, al tiempo que se movilizaba a causa de los intereses materiales de los mercaderes emergentes y de las grandes Naciones-Estado. La dinámica de dominación fue compleja dado que se basó en el uso de la fuerza (guerras de conquista y sistemas coloniales) y en el agenciamiento de la cultura y del ethos civilizatorio occidental para ganar las mentes de los colonizados (véase: E. Said, Ibíd.). 169 En América latina, desde finales del siglo XIX, a los partidarios de esta mimetización se les ocurrió justificarla con un término cargado de sentidos: cosmopolitismo. Enfrentando las reacciones del señorío patriarcal y del catolicismo hispanista que preferían la continuidad de las sociedades cerradas, autores como Sarmiento, Rodó, Vasconcelos, José Ingenieros, Bulnes y otros se arropaban bajo las banderas que ellos mismos nominaban tomando prestado otro término: positivismo. El liberalismo hizo posible la confluencia de esta adhesión con la fórmula preeminente del individuo como centro de los derechos y la juridificación. 4. Lo público y los desafíos de la modernidad En este sentido, la Modernidad percibe a lo privado como un fin en sí mismo. Es decir, la sociedad de derechos liberal individualista, propia del siglo XX, como coronación de un proceso de cinco siglos, tiende a postular como su finalidad, como su telos, garantizar un ejercicio libre por parte de los individuos en la esfera privada. Pudiéramos parangonar esta constatación señalando que así como Kant habla del hombre como un fin en sí mismo, observaríamos la evolución del sentido de finalidad humana en el logro de la privacidad. La construcción de dicho ámbito resguardado sería la plasmación, auténtica, de la finalidad humana moderna. El hombre se consolida en la sociedad moderna de los siglos XIX y XX a partir de una afirmación de su individualidad generando salvaguardas con respecto a lo público. Igualmente, se manifiesta y exalta una privacidad del hombre público, la cual se evidencia, por ejemplo, en la lucha por separar la representación social del político, del hombre de empresas, del hombre de los medios, etc., con respecto a su vida privada. Véase el caso tan discutido desde Max Weber sobre el carisma de los hombres públicos, particularmente aguzado en la segunda mitad del siglo XX por la puesta en escena de la imagen política a través de los massmedia: como el hombre público sabe que lo juzgan por lo privado, crea una privacidad artificial para poder mostrar esa privacidad artificial que le exigen los acuciosos ojos de sus electores y dirigidos, escondiendo detrás de tal privacidad artificial su verdadera privacidad. En dicho proceso se configura una suerte de doble privacidad: la que se exhibe deliberadamente, de 170 forma intencionada, y la real. Dicotomía a través de la cual el hombre público moderno se escinde. Podríamos indagar, con la ayuda de la psicología social, no solamente este caso —como una línea de exploración investigativa— de los hombres públicos que ejercen liderazgos; sino en el ciudadano anónimo, en el “hombre de la calle”, cómo existe esa doble privacidad: una privacidad construida como parte de su representación, es decir, una falsa vida privada, en tanto esa privacidad se convierte en parte de su representación pública (en los hombres de negocios, los profesores universitarios, en la gente frente a su barrio, frente a su mundo laboral, etc,). Es decir que muchas de las facetas de la vida privada: el auto que se posee, la casa donde se habita, el lugar donde estudian los hijos, dónde y cuándo se vacaciona, el tipo de esposa(o) que se tiene, etc., son, en suma, una serie de elementos aparentemente privados pero que, en el fondo, corresponden a un componente de la vida pública, porque son parte de la representación social. Hacen parte esencial, como lo diría Hegel, de la puesta en escena del reconocimiento, de la valoración y la afirmación de la personalidad en el contexto público. Richard Sennett (1979), quien hace una profunda reconstrucción de lo público usando elementos conceptuales de la psicología social con el fin de discutir problemas tan específicos como los del narcisismo y el carisma en su relación con lo público y lo privado, sugiere, como lo veremos luego, que en últimas el predominio moderno del intimismo produce la “muerte” de lo público. Por ello resulta conveniente que discutamos si en verdad hay una caída del espacio público como resultado del dominio del intimismo, o si existen límites y condicionantes en relación con la anterior tesis de Sennett. Así, cada vez con fuerza mayor, la comunidad se encierra en la privacidad y le da la espalda a los asuntos públicos; al punto que se evidencia un fenómeno: la huida de los formalismos de la red pública, presentándose una suerte de vuelta hacia lo privado y abandono de lo público. Sennett condensa su estudio en una tesis, en cierta medida discutible, según la cual el aislamiento aparece como condición previa de la privacidad y, por lo tanto, el dominio de la privacidad es la muerte del espacio público. El intimismo, dice Sennett, surge como resultado de un proceso de modernización que implicó la destrucción masiva del dominio público. Por ejemplo, las tecnologías electrónicas en 171 auge durante las últimas décadas (la televisión especialmente, pero cuéntense además las videocaseteras, la Internet y muchas otras nuevas tecnologías) ofrecen a las personas una alternativa de escape, pues ahora pueden disfrutar, desde su casa, su propia visión o relación con el mundo sin contacto físico, lingüístico, significativo. Las nuevas tecnologías permiten una selectividad y jerarquización del gusto a partir de un mismo medio electrónico; y establecen, incluso, la posibilidad de desconectarse de la programación masiva. Recordemos, a manera de norma general, que cuando se habla de lo público como espacio de exposición existe un nivel concreto y altamente valorativo (en términos de intercambio social) cual es el reconocimiento9. Tal situación implica una gradación en términos de calificación y jerarquización social. La referencia a la conexión entre civilidad y sociabilidad expresa un sentido renovado de lo público, tanto en la configuración sociohistórica de jerarquías y distancias como en sus implicaciones dimanantes hacia el terreno político. Sin embargo hoy —y desde hace unas cuantas décadas— precisamente con el advenimiento del postindustrialismo, el postmodernismo y la globalización, se ha cobrado consciencia en el debate intelectual, político y científico contemporáneo de la relevancia de la postura multicultural. Ello significa una crítica acerba al imperialismo cultural —y no sólo al político o al económico—, igual el reclamo del respeto a las diferencias, a los otros ethos contruidos en evoluciones sociales de larga duración, etc. La tensión profunda de la actual mundialización está allí: ¿la innegable universalización del capitalismo por el orbe es correlativa al avasallamiento definitivo de las alteridades? Una sola verdad: la tecnociencia; una sola moral: el utilitarismo; una praxeología: la eficacia racional instrumental; una sola ontología: el homo oeconomicus? Nuestro tiempo es, en esta perspectiva, un gran escenario de negociación-confrontación. En el campo de estudio que corresponde a las llamadas Ciencias del Management, se instala con sus especificidades este dilema. Aunque la emergencia de paradigmas alernativos al dominante está muy lejos de los logros en otros terrenos de la vida cultural (la política, el arte, las ciencias humanas, etc.). A despecho de la consideración optimista de los ilustrados sobre las bondades de la deliberación pública, tanto en el proceso formativo de los ciudadanos como en la legitimación del orden político, a lo largo del siglo XX se ha visto 172 tergiversado dicho ideal por el papel enorme de los massmedia, a su vez manipuladores de la opinión pública y manipulados por los intereses de los grupos o elites dominantes, tanto públicas como privadas. De otro lado, el ideal del ciudadano deliberante, que reflejaba la utopía griega, e incluso, el republicanismo rousseauniano —tal como lo mostré en el capítulo anterior— no se ha plasmado en la sociedad contemporánea. De esta manera la investigación implica delimitar los contenidos específicos del espacio público, que sólo los podemos aprehender a través del estudio de las maneras concretas en que se desarrollan dichas representaciones. Desde luego, existen contratendencias respecto a este problema de representación social; mecanismos de diferenciación en un momento, son de homogeneización en otro. Para lo cual habría que construir una semiótica de la vida pública, es decir, un estudio de los signos, significados y códigos que constituyen los ámbitos de la vida pública como escenarios de representación. No es lo mismo la representación de una sociedad cortesana que está plenamente jerarquizada y donde el solo vestuario establece un claro lenguaje de separación respecto a una sociedad liberal democrática donde, en teoría, todos nos podemos vestir igual. Desde luego, allí hay lecturas; por ejemplo, la diferenciación de las marcas (éstas son más sutiles, se aprende a leer la calidad de la ropa de la persona o incluso la gente es muy aguda para leer hasta el corte del cabello) de tal modo que muchas señas resultan fácilmente indicativas de segmentación y diferenciación social; pero son lenguajes que cada vez se han convertido en marcas de civilidad sofisticadas. Así, se adquiere a través de la resocialización grupal una aguda capacidad de lectura, relativamente rápida, de estos símbolos de representación. Existe evidentemente toda una compleja mecánica expresiva del cuerpo, v. gr., las formas de cortesía moderna, los grados sutiles de denotación que marcan niveles de significación, tanto en el lenguaje discursivo como en el lenguaje corporal o en el vestuario, parte sustancial de la representación social. El médico con su indumentaria blanca, el ingeniero con su indumentaria semejante a la del operario; el antropólogo, en nuestro caso, disfrazado de hippie de los años sesenta, el yupi o joven ejecutivo; la típica ama de casa, el joven de clase alta, los adolescentes rebeldes, los punks, etc., en fin, cada persona inserta en grupos sociales, representa un papel que se incorpora dentro del 173 lenguaje corporal, del lenguaje discursivo y del vestuario. Obviamente existen vestuarios totalmente segregados del resto de la sociedad como el militar o el eclesiástico. Pero incluso este último tuvo que reducirse; fue derrotado en el siglo XX y realmente solamente acuden a él las altas jerarquías de la Iglesia o, en ciertos casos, de ritualidad muy significativos (matrimonios, bautizos, inauguraciones, etc.). En el caso de los militares, sí tienen claramente una separación en su vestuario con respecto al resto de la comunidad, por obvias razones de Estado. Hoy en día existe una enorme complejidad en los mecanismos de representación. Así, la ciudad y su espacio público son escenarios de representación y de exigencia de derechos. Un lugar de múltiples encuentros, de sociabilidades. Igualmente existen espacios urbanos modernos que toman como modelo el esquema actual de la ciudad, vista como ámbito mundial por excelencia de sociabilidad, de civilidad, cuyos espacios de intercambio están restringidos o son semipúblicos. La ciudad actual en general está pensada (y habría que investigar más este tópico con arquitectos, viviendistas y psicólogos sociales) en términos de un intimismo; allí, el lugar de una vida pública no está delimitado. Cierto: el paseo por la ciudad popular no es una costumbre cultural; aunque todavía existen algunos lugares donde la gente se encuentra que son muy pequeños con respecto al conjunto urbano. En el plano de la vida pública urbana que dimana hacia la expresión política, vemos que el modelo de la plaza pública es un paradigma ficticio, un deber ser artificial. Incluso movilizar multitudes hacia la plaza pública es muy difícil. Así, vemos cómo la representación política que acude al espacio arquetípico de la ciudad ha desaparecido al menos en sus formas primigenias. ¿Qué la ha reemplazado? Ya lo hemos visto: los massmedia. Quienes actúan en la vida política, o la analizan, saben muy bien que la clave del éxito está cada vez más condicionada a las posibilidades de acceder a los medios masivos de comunicación social. “Salir” en televisión, en los grandes periódicos y cadenas radiales. Quienes puedan acceder a estos grandes medios logran una representación y una puesta escena pública inmejorable. Adicionalmente, la manera como la imagen sea construida para el público que consume los medios es también algo decisivo. Una de las trazas singulares de nuestra época consiste en el predicamento según el cual el capitalismo liberal le dice a la 174 comunidad que los bienes están al alcance de todos, y en las posibilidades de todos el libre acceso a estos. ¿Cómo las tendencias hacia la frustración social se corresponden con el utilitarismo puesto por los medios o por la misma Modernidad, según el cual todos tenemos el mismo derecho de acceso a los bienes de la sociedad de consumo? Este consumismo es la forma específica que adopta el paradigma universalista de la igualdad ciudadana. Es indudablemente el espacio por excelencia de lo público, en la vía preeminente, por supuesto, de la publicidad; es decir, la construcción deliberada, intencional, de objetos de gusto, vitales para que a través de su apropiación el individuo se reconcilie con el ser colectivo, lo que nos lleva al paradigma del uniformismo. Además, esta pulsión consumista no es problemática sólo porque le diga al consumidor que tiene que ser como las elites (o al menos imitarlas, asumiéndolas como un deber ser), sino porque genera en muchos espacios de la vida un paradigma inalcanzable para todos. Ello crea una esquizofrenia del espacio público: lugar de todos y ámbito de realización de muchos o de pocos, pero en donde, en todo caso, se presenta de forma dramática el fenómeno de la exclusión social, de la marginalidad. Vinculado a esto no debe olvidarse que existen poderosos mecanismos democráticos que vehiculan la opinión masiva, que reproducen el autoritarismo y la discriminación (por ejemplo exaltando el nacionalismo, la subvaloración racial, etc.). El espacio público mantiene su postulado democrático en tanto se apoya en el punto de vista de ser un espacio de todos. Pero el problema de los contenidos del espacio público y la significación del espacio público es algo que no se resuelve únicamente por la cobertura o la extensibilidad del espacio. Puesto que surge también la posibilidad de que la representación no le corresponda exclusivamente a la elite y que el espacio público sea usado por las demás clases sociales. Incluso se pueden dar casos de sincretismo cuando clases y sectores de elite se apropian del lenguaje popular y de los gustos populares. En todos estos fenómenos de la vida pública se presentan una vasta gama de entrecruzamientos y canales de circulación, sentidos, etc., que merecen una indagación semiótica. Así, pudiéramos sintetizar este punto: hoy en día, el gran espacio de representación de lo público son los massmedia. Y estos evidencian 175 en su funcionalidad diferencias sustanciales frente a las distinciones jerárquicas de la plaza, el café, la sala de teatro, etc., propias de la sociedad ilustrada que le tocó en suerte a Kant. En el pasado, en tales escenarios la puesta en escena de la diferenciación de las elites era muy clara. Piénsese en eventos contemporáneos propios del star system tales como la premiación de los Óscares, donde los artistas de la elite artística se diferencian del resto en la puesta en escena del jet set de la cinematografía; lo mismo puede verse en el caso de las exposiciones de pintura, donde una elite (la elite de los críticos, la de los compradores y los bon vivants) se diferencia del resto de los consumidores de arte; porque a diferencia del pasado, en la sociedad aristocrática, no son ellos ya los únicos consumidores: no son el único público. Mientras que en el siglo XVIII, el único público eran las minorías ilustradas ricas y aristocráticas que iban a ver teatro o escuchar música de cámara. En el siglo XIX y más aún en el siglo XX (una sociedad masificada), aparecen gradaciones y separaciones de los públicos. Ya no hay un solo público, sino muchos públicos y probablemente existen todos los públicos. En síntesis, desde los tiempos de la Ilustración, ha cambiado el sentido político y cívico de lo público, para ser éste cada vez más el ámbito por excelencia del mercado, del intercambio y la sociabilidad reglada en un contexto de permanente diferenciación y, a la vez, aunque ello resulte paradójico, de homogeneización. La crisis de la Edad Moderna —las señales de su agotamiento— es en tan pocos ámbitos evidente como en la quiebra del universo de lo público. Las llamadas teorías “neoliberales” o, expresadas de manera más exacta, los “liberalismos libertarios” encuentran puntos de confluencia con las llamadas tendencias postmodernas precisamente en su acerba crítica al papel del Estado —y junto a él, otras instituciones del ámbito privado— que surgen como instancias que se abrogan para sí la representación cuasitotal, la hegemonía de lo público. El retorno a una teoría filosófica que reivindica el papel de la subjetividad no esquematizada en torno de una razón fundante o de un sujeto racional sino que alternativamente involucra múltiples sensibilidades, potencialidades del gusto, redefiniciones del universo de la política y de las antiguas fundamentaciones contractualistas, etc., recoge indudablemente la experiencia del final del siglo y 176 redefine el papel de lo público y lo privado en términos tales que la distinción kantiana recobra, ciertamente recontextualizada, utilidad analítica. La función pública del hombre de Estado, de las burocracias impersonales, es sin duda el espacio de la privacía, donde se manifiesta hasta ahora la imposibilidad de ocupar simultáneamente el doble rol de sujeto público y privado. ¿Cómo distinguir la opinión personal de una oficial del ejército o de un ministro sobre los temas de orden público, frente a su rol institucional? Sin embargo, se ha venido avanzado en la recuperación del espacio de la deliberación pública para el enorme grupo social que constituye el funcionariado. Véase, por ejemplo, en el caso colombiano la permisibilidad para que los funcionarios de nivel medio y bajo puedan ejercer sin restricciones significativas la militancia política, algo vedado en el pasado dentro de nuestra ritual doble moralidad (ya que en verdad todos los funcionarios públicos han sido, sin excepción, hombres políticos o que ejercen las formas clientilazadas de la política). O tómese en cuenta el papel de las distintas ramas del poder público que deliberan entre sí, bajo el principio de la construcción de poderosos frenos y contrapesos, al absolutismo del poder de Estado. Fenómeno que por cierto, gracias al predominio de un hegemonismo herededado del Frente Nacional, los colombianos, después de la promulgación de la Constitución de 1991, hemos entrevisto, equivocamente, como una situación indeseable o caótica, cuando se trata precisamente de lo contrario: crear el orden social e institucional como un proceso de equilibrio funcional entre intereses y puntos de vista diversos. Así, incluso en el ejercicio de la función institucional, donde Kant veía como algo normal la restricción a las pulsiones deliberativas de los individuos, hoy se avanza en la conformación de un renovado garantismo. De otro lado, el discurso político predominante en nuestro contexto insiste en la limitación, en las insuficiencias de la democracia representativa. Y en la necesidad de un recambio o renovación del ejercicio de los derechos deliberativos de los ciudadanos en términos de la participación de la comunidad en el proceso mismo de implementación de las políticas públicas. Con ello se busca romper la separación entre la sociedad civil y el Estado, entre los niveles de lo público y lo privado. Hasta dónde este supuesto constituya una utopía es, desde luego, una temática objeto de acres dimensiones que no pretendo plantear aquí. Pero, 177 en la medida en que la teoría política no se limita a describir una realidad sino que pretende ofrecer alternativas de mejorabilidad (es decir, estatuir un deber ser) aparece el problema de la configuración de una opinión pública madura, libre hasta donde sea posible de las censuras e imposiciones de los poderes establecidos, como un mecanismo civilizatorio, creador de equilibrios, de fuerzas que canalizan los conflictos de intereses, nunca originados en la ignorancia —como ingenuamente se cree— sino en el saber interesado. En dicho programa político veo la funcionalidad y la actualidad de la reflexión kantiana sobre los usos de la razón deliberativa. 5. La configuración del espacio público Como lo advertíamos líneas atrás, la ciudad como espacio de reconocimiento o exposición pública, fue característica del siglo XVIII. Confróntese, en este sentido, la mayoría de las crónicas históricas en las cuales se ha podido constatar este fenómeno en América Latina. Desde muy temprano los españoles se cuidaron de construir ciudades a imagen y semejanza de las europeas. A quienes vivimos en siglo XX, en medio de urbes anarquizadas, nos cuesta trabajo creer que la ciudad latinoamericana fue planeada y diseñada con una semiótica pública deliberada. Los virreyes y, en general, las autoridades coloniales, tenían que someter a la aprobación de la corona española los planos y trazados de las ciudades del Nuevo Mundo. La ciudad española construida en América hablaba un lenguaje arquitectónico: estaba diseñada en términos de una plaza, una iglesia con su sacristía, la sede de los poderes jurisdiccionales correspondientes; y en el resto del cuadrado vivía la aristocracia citadina compuesta por los vecinos de mayor prestancia. Luego, la ciudad se extendía en cuadrículas en las que a mayor distancia era menor la prestancia social de los habitantes. En la plaza principal había posibilidades de resguardo militar, puesto que también las ciudades eran construidas como fortalezas militares. Además, la plaza fue el lugar por excelencia que constituía el ámbito de lo público, donde se realizaban las festividades religiosas, la entrega del mando de un virrey a otro, la posesión de las autoridades locales en las provincias; donde se realizaban en ocasiones solemnes misas campales, se hacían los jolgorios y ocurrían las revoluciones. Es decir, en la historia latinoamericana no se puede pensar la vida pública sin la 178 plaza. Hoy en día es un arcaísmo hablar de la plaza pública; pero desde mediados del siglo XX, hasta remontarnos a la Colonia, la plaza pública fue el espacio de sociabilidad más importante. En síntesis, la ciudad ha sido un espacio de exposición en que se ha representado y escenificado el drama cotidiano de la vida pública, lo que explica la función del paseo citadino. Pensemos en la Colombia, en la América Latina de los años sesenta hacia atrás; ¿qué hacía la gente al acabar la jornada de trabajo?, pasearse por la plaza pública. Este era el lugar de encuentro, la conquista amorosa, el lugar del intercambio social, reconocimiento de jerarquías; mientras que el café de la plaza pública fungía como el lugar de los negocios. Han sido nuestras ciudades vividas, históricamente hablando, como espacios de exposición y diferenciación. ¿Cuándo puede señalarse que este fenómeno de exhibición y marca civilizadora en las ciudades se disuelve? Ello ocurre en nuestros tiempos y una causa evidente ha sido el excesivo crecimiento de la urbe. En el siglo XIX, en Europa la urbe se volvió tan grande, que las calles de la ciudad ya no fueron el espacio del reconocimiento público, comparadas con las ciudades más pequeñas y manejables de los siglos XVII y XVIII. En nuestro caso, las pequeñas ciudades aún son espacios de reconocimiento, pero otras razones hacen que en ellas no se repita la situación anteriormente descrita. En parte, porque los medios de locomoción varían: el paseo de la exhibición ya no es directo, a través de una caminata, sino que está mediado por el automóvil como artefacto de comunicación y representación. En las grandes ciudades de la sociedad industrial y mercantil, el espacio público se vuelve anónimo. Es muy difícil diferenciar y crear jerarquías a partir de la representación en los espacios públicos. Al menos en nuestro contexto específico, mientras la ciudad aparece como un espacio de reconocimiento de lo público en la época colonial y la primera época republicana, ya no lo es en la sociedad modernizante o en tránsito hacia la Modernidad. Pero esto no es, desde luego, algo absoluto. Indudablemente, existe una diferente valoración de los públicos: quienes van a la Media Torta de Bogotá para escuchar a un cantante popular, a través de este acto de exhibición pública se ubican en un segmento de la sociedad; en tanto que si se asiste a una presentación del mismo cantante en un hotel o centro nocturno de lujo, ello determina la pertenencia de un público de otro segmento social. Esto es así por cuanto el ámbito público 179 es un escenario de representación, donde de manera cada vez más sutil o fina se construyen diferencias sociales. Esta puesta en escena que se evidenció a partir de la Ilustración posteriormente se profundizó con la masificación del espectáculo público. Sin embargo, en América Latina no ha desaparecido del todo la ruralización de las ciudades y las costumbres (Veliz, 1972); el populismo en América Latina ha tenido tanto arraigo social, pese a la modernización y al despoblamiento del campo, sobre la base del discurso carismático del caudillo que le habla al campesino, y aunque éste ya no lo es porque vive en la ciudad, posee una cultura política campesina. Entonces coexisten una ruralidad y prácticas sociales modernas en relaciones a veces muy sórdidas, principalmente en las barriadas, en la medida que el obrero, el emigrante en su segunda o tercera generación va integrándose hacia la civilidad urbana y cambian tanto sus gesticulaciones como su cultura (cfr. Varela, 1995a). Así, incluso la violencia, en el pasado directo, se convierte en una violencia parcialmente simbólica. El campesino hace uso de la violencia del cuerpo, de la agresión, mientras el hombre de la ciudad en su vida privada tiende a usar (claro que en nuestro país no es tan exacto esto) la violencia simbólica del lenguaje; es decir la agresión verbal y la capacidad sofisticada de subordinar al otro a partir del lenguaje (cfr, Basil Bernstein, 1978). Veamos ahora el lugar, en este fenómeno, de los medios masivos de comunicación social. El lugar del espectáculo aparece primordialmente en la televisión y, en segundo lugar, en los demás medios públicos de difusión: radio, prensa, revistas, folletines, etc. Sin embargo, los massmedia no logran reproducir de una manera adecuada las distinciones que correspondían a los escenarios ilustrados (la sala de concierto o de teatro en la sociedad en transición hacia el industrialismo), porque los massmedia están dominados por una lógica hiperactiva muy fuerte: la del mercado, y mercado en la sociedad de masas es básicamente mercado popular. Jesús Martín Barbero (1995) muestra cómo en la cultura popular las clases populares hacen una lectura y una reelaboración de la televisión y los medios masivos, una relectura a nivel de los sectores populares e incluso muchas veces se logra la utilización de los medios para fines totalmente diferentes. Un caso muy interesante que sugiere Barbero se dio en el Perú con la utilización del radio parlante. En particular, las mujeres en las plazas de mercado utilizaron la radio 180 para solucionar sus problemas, además, lo popular (como el mercado popular es el mayoritario) se vuelve un objeto deseable de representación social. Ahora bien, también existe un movimiento de reconocimiento de lo provincial, por ejemplo, con el descubrimiento de que los mejores literarios nuestros están en las provincias. Se empieza a ver una tendencia global en América Latina, los brasileños comienzan también a elaborar y presentar en sus telenovelas elementos que permiten la expresión de sentidos regionales, costumbres y formas de ser de lo popular que empiezan a admitirse como válidas. Es la misma gente viéndose representada. O, en ese mismo sentido, piénsese en la extensibilidad del impacto de la salsa, que hace 30 ó 40 años era una música marginal y sórdida; o tómese en cuenta la extensibilidad de los gustos en Bogotá sobre los vallenatos y los merengues, originalmente música popular de provincia. Incluso, vemos fenómenos bien difíciles de comprender como el hecho de que ciertos programas pudieran sostenerse, debido a su público selecto, a pesar de que su rating fuese muy bajo. Siendo sectores pudientes o intelectuales quienes los sintonizan, resultan excluidos de la programación en la TV y la radio, porque se mide homogéneamente el público y por lo tanto se toma el rasero de la ama de casa que ve telenovelas y programas sórdidos, o el gusto del habitante de barriada (que es quien más ve televisión). Así, a la televisión o a la radio —noticias o la radio que entretiene— les queda difícil ser un escenario adecuado para la jerarquización social, en el sentido elitista antes descrito. Aún más , si se toma en cuenta otro fenómeno: quien que se presenta en la televisión, no se puede ufanar de pertenecer a elites; tiene por el contrario que construir un discurso democrático liberal enfatizando un interés general en su acción. Por ello, los capitalistas, en la publicidad masiva, aparecen como benefactores; las empresas y sociedades anónimas tienen un estilo de publicidad que pudiéramos llamar subliminal, consistente en anunciar indirectamente sus productos a través de un mensaje que enfatiza en torno de bienes comunes; vemos a diario a bancos y corporaciones financieras anunciando la defensa de los bosques; compañías que se lucran con los niños, pues son ellos quienes consumen sus bienes, patrocinando campañas para la defensa de los menores; a grandes firmas papeleras o madereras defendiendo en su publicidad los bosques y las aguas que ellos depredan, etc. El mensaje no puede ser directo, porque hay 181 un inhibitorio en relación con la capacidad que tiene el medio de masas cual es la capital suposición de actuar con el fin de lograr el consenso del gran público. Es decir que los medios masivos de comunicación social no permiten establecer una gradación tan tajante entre el público segmentado que reproduzca jerarquías sociales. Y esta es, a su pesar —o sin que sea algo consciente— una labor democratizadora de los medios. Desde luego, este no es un fenómeno absoluto y existen salvedades: por ejemplo, la publicidad que establece como móvil el adquirir un estatus diferenciado. Como contrapunto a lo anterior, diríamos que el medio masivo posee un límite inmanente, ya que establece un rasero uniforme y bajo. Y en ese sentido, la televisión y los otros medios masivos generan grandes dificultades para elevar las características cualitativas de la cultura popular. Pero esto es objeto de otro análisis y simplemente lo menciono para mostrar cómo este mecanismo no actúa de manera uniforme (B. Berger y T. Luckmann, 1968). En otros campos operan mecanismos de publicidad en los que es reconocible un hito inalcanzable para la mayoría, pero que no se pone como tal. Un ejemplo en tal sentido es el paradigma racial, la abominación establecida en términos de que la “sociedad” es de los rubios. Un artículo de la revista Time, publicado en 1993, hace una reseña de la propaganda de la televisión mexicana y se pregunta “¿México rubio?”. Cualquiera que llegue a la ciudad de México y vea un canal de televisión creería que está en Suecia. Todos los modelos, todas las personas que salen en la pantalla son rubias, pero en la calle el observador encuentra una sociedad abrumadoramente compuesta de indios y mestizos, en la cual sólo el 10% de la población es blanca y de ese 10%, sólo el 3 ó 4% es realmente rubia. Eso constituye un paradigma incesante en la publicidad latinoamericana. En realidad, no hay nada más indignante que ésta afrenta a nuestra identidad constitutiva, desde el punto de vista de la pluralidad étnica y racial. 6. El sentido político de lo público Quiero señalar ahora otro punto frente a lo público y es el verlo como espacio para el ejercicio de derechos sociales y políticos. En este orden de ideas, una primera conceptualización de lo público propone que este es un ámbito universal. Véase el caso citado de la circulación por las calles como parte integral del espectáculo público de la representación. Ser reconocido y 182 reconocer son potencialmente elementos constitutivos de todos los miembros de la comunidad. Por lo cual nos explicamos el hecho de que muy pronto la libertad de locomoción se convirtió en libertad de expresión de derechos sociales y políticos. Hay trabajos muy interesantes en este sentido de la historiografía moderna sobre lo que llaman revuelta de la turba popular; particularmente se refieren al hecho de que muchos de los grandes estallidos de cólera popular en Europa en los siglos XVIII y XIX se pudieron presentar teniendo como escenario las grandes ciudades. Hasta el siglo XVI, en efecto, las revueltas populares eran fundamentalmente revueltas campesinas, estallidos de cólera rural, a veces patrocinados por intereses feudales, otras veces espontáneos y supremamente destructivos por cuanto el campesinado ha sido una clase social que por sus condiciones de vida, como diría Hegel, como rebelde se expresa en su pura negatividad, porque carece de capacidad para construir una alternativa social. El campesinado, en general, opera como una fuerza conservadora en el sentido de que es salvaguarda principal de tradiciones y costumbres. Los campesinos, a diferencia de lo que el marxismo ha creído en el último siglo, no tanto actúan como una fuerza revolucionaria sino como una fuerza que defiende un cierto orden. Veamos ejemplos concretos: el papel de las masas rurales en el caso de Irán e Iraq, en el respaldo al gobierno de los musulmanes rechazando la modernización industrial occidental; lo podemos ver también en el caso de Rusia y de los países de la antigua Unión Soviética, donde las fuerzas que respaldaron al Partido Comunista dentro de su base de adherencia eran las fuerzas rurales, mientras que en las ciudades se gestó todo el movimiento para hacer un trastorno del orden social. La ciudad y el espacio público urbano han sido —en los últimos tres siglos— ámbitos de la expresión revolucionaria popular en la forma de la rebeldía espontánea o como un movimiento político dirigido. Casos clásicos: la Revolución Francesa, una revolución en París; la Revolución Bolchevique en Moscú y Leningrado, como la llamaron después. El modelo de la insurrección militar y la toma de la capital de una ciudad o un imperio era una demostración de que el espacio de lo político se definía fundamentalmente en el ámbito de vida pública de las ciudades. Nuestra historia, sin embargo, muestra variantes frente a este modelo: los latinoamericanos hicimos una revolución en el 183 siglo XIX que fue predominantemente rural, pues el desenlace de la revuelta promovida por Bolívar y los líderes patriotas no se definió en las ciudades. Y a lo largo de la vida republicana del siglo XIX las luchas armadas, las guerras civiles, no se definían en las capitales. Éstas fueron lugares donde se cimentaba y reagrupaba el nuevo orden, el orden revolucionario, o, en algunas ocasiones, el orden restauracionista. Pero un indicio muy interesante es que sí se inicia el estallido a través de revueltas urbanas. Son las ciudades quienes protagonizan el grito de independencia: Mariquita, Ambalema, Cali, Buga, Popayán, Cartagena, el propio Socorro, Cúcuta; hay unas trece ciudades en la Nueva Granada que están al frente de la emancipación. Pero, por las características de la sociedad colonial y republicana del siglo XIX, el escenario de la confrontación no fue tanto la ciudad, sino el campo. Es sintomática, sin embargo (ya entrados en el final del siglo XX), la medida débil en la que la protesta popular urbana se manifiesta. Nuestra sociedad posee, en términos comparativos, muy poca protesta popular urbana. ¿Qué pasa entonces con los movimientos sociales que surgen a veces por algún tipo de confrontación, es decir, que, en un momento dado, surgen como expresión, así sea aislada, de conflictos? El problema es que se disuelven. Los movimientos sociales han atravesado en Colombia por una etapa de guerra muy violenta. La década de los sesenta fue la década, en cierta medida, marcada por movilizaciones populares; un periodo donde había posibilidad de protestar, pero después —en la década de los setenta— se generan enfren tamientos fuertes y se da un fuerte retroceso de los escenarios de la calle y la plaza, del desfile, como topoi de protesta civilizada. Dicho de otro modo, estamos viviendo la misma etapa que sufrió la Argentina y otros países del cono sur, durante los regímenes militares, donde muchas pulsiones sociales se prohíben y autocensuran porque hay un gran miedo social y una introyección del ámbito de lo político. Es en nuestro caso, una consecuencia larvada por la suciedad de un conflicto político, gangsteril, lumpenizado. ¿Cómo pudiéramos manifestar esto en términos conceptuales? Es cierto: existe sin duda intolerancia estatal, la cual se apoya en la inexistencia en nuestra tradición de una cultura política que legitime la protesta urbana pacífica. Y, al mismo tiempo, vivenciamos intolerancia de las fuerzas paraestatales y de sectores de la propia sociedad civil que entran a 184 intervenir de manera violenta contra quien proteste o se oponga a sus intereses y designios. Así, la expresión civilista —utilizando los rituales propios a la civilidad— se anula pues ¿qué es una protesta urbana? El típico mitin callejero es una exposición de simbolismos: desfile, escenificación, teatralización de la vida política a cargo de actores que son los marchantes ante un público, la comunidad, que los observa. Pero tenemos rasgos autoritarios expresados en la recurrencia con la cual la protesta ciudadana es inhibida o reprimida. O, cuando esto no ocurre, la gente no sale porque les preocupa mucho la seguridad por las mil cosas que pasan. Ahora bien, el que la represión debilite la protesta popular no es una explicación suficiente del fenómeno, porque en otras sociedades logra aflorar la protesta popular urbana aún en condiciones de represión. Este es un tema muy complejo que no podremos en estas líneas elucidar. Pero lo que quiero mostrar aquí es como la ciudad y su espacio público son escenarios de representación y de exigencia de derechos. Un lugar de múltiples encuentros, de sociabilidades. Veamos ejemplos que muestran una cultura popular divergente a ese modelo: cuando se recorren los barrios populares de las ciudades latinoamericanas, las barriadas de invasión, no se encuentran plazas o lugares diseñados arquitectónicamente como espacios de sociabilidad, no hay lugares de encuentro de la comunidad. Obviamente en las zonas antiguas se manifiesta la pervivencia de una ciudad vieja hispánica, donde sí existen tales ámbitos. Además, hay una ciudad republicana que tiene una identidad y que corresponde también al trazado español, con similares rasgos. Finalmente, existen espacios urbanos modernos que toman como modelo el esquema actual de la ciudad mundial, cuyos espacios de sociabilidad son restringidos o semipúblicos. La ciudad actual, en general, está pensada (y habría que investigar más este tópico con arquitectos, viviendistas y psicólogos sociales), en términos de un intimismo; allí el lugar de una vida pública no está delimitado. Cierto: el paseo por la ciudad popular no es una costumbre cultural, aunque todavía existen algunos lugares donde la gente se encuentra, pero son muy pequeños con respecto al conjunto urbano. Al respecto, apoyémosnos en una disciplina relativamente reciente: la criminología y, particularmente en las variantes de ésta, que se han ocupado del control social. La idea básica de la nueva criminología es que una sociedad se puede leer desde la Historia o se puede ver a sí misma desde el presente 185 en función de lo que resulta criminalizando. Es decir, qué considera válido y legítimo y qué considera una transgresión; y al mismo tiempo mirar cómo las sociedades, al estructurar este horizonte —la separación entre lo legítimo y lo punible— construyen el orden social a través de mecanismos de control. Un mecanismo de control, por ejemplo, típico de las sociedades autoritarias tradicionales es el patriarcalismo, el poder de un padre de familia que no sólo maneja su familia pequeña, ya que el patriarca es una especie de gobernante de un clan o semiclan. Quien elabora un modelo de jerarquía y orden basados en elementos de respeto y subordinación que no se pueden poner en discusión con elementos morales muy claramente definidos, que aparecen como el modelo desde el cual se reproduce la misma sociedad. Se postula que el Estado sería una forma de estructuración del poder, correspondiente o equivalente al que el padre tiene sobre los hijos10. En realidad, ocurrió lo contrario. El Estado organizó a la sociedad, la familia, y delimitó claramente el lugar del patriarca y la estructura de dominio propia de la sociedad rural y campesina. Ello ocurrió en España, en América en los siglos XVI, XVII, XVIII; en Europa desde el siglo XVI, más o menos, hasta final del siglo XIX. Es decir que el Estado sometió como ámbito de reglamentación y de control social a las esferas anteriormente tuteladas desde una perspectiva patriarcal (privatizadas) de la familia y la comunidad rural. De donde surgió un nuevo modelo según el cual el orden social pasaría por el orden público. Hay que referir aquí igualmente las conocidas y ya citadas historizaciones de la rebelión urbana. Estas acciones fueron elementos que obligaron a darle un tinte político al control del espacio público. Cuando se vive en las grandes concentraciones urbanas, para todo Estado es indispensable la creación de una estructura de tipo preventivo, y las policías corresponden a ese período y a tales tareas. Tal desagregación entre el ejército como defensor de la Nación (Estado vigilante de las fronteras) y un cuerpo armado destinado al orden interno, claramente provino de las contradicciones y tensiones que vivieron las sociedades en el proceso de delimitación de las esferas de sociabilidad públicas y privadas. Hay además elementos, no de tipo político sino del orden social y sociocultural, que llevan al Estado a intervenir el espacio público. Por esta razón no hay ciudad importante que no tenga normas sobre uso del mismo. Precisamente, nuestra discusión en una ciudad como Cali se basa en que estas normas no sirven 186 puesto que no hay espacio público respetado, puesto que se manifiesta como rasgo dominante su privatización. La gente del común, los empresarios y los propios organismos estatales, se apoderan de calles, parques y demás ámbitos del espacio público. He observado, por ejemplo, contratos en los cuales vendedores ambulantes (buhoneros) “venden” el andén sobre el que ejercen posesión y tratan de registrar dicho traspaso en notarías. Confróntense igualmente las demandas judiciales a personas que usufructúan un parqueadero usurpando un espacio público, no para retornarlo a la comunidad, sino alegando pertenencia privada legítima (J. Romero, 1976). En Colombia, como en muchos otros países, existen reglamentos sobre el uso de antejardines y zonas de uso común que, ubicadas en el contexto de propiedades privadas, son parte del espacio público; pero buena parte de las viviendas y locales comerciales asumen dicho espacio como privado estableciendo cerramientos que impiden su visualización y disfrute comunitario. Incluso multinacionales productoras de bienes de consumo colonizan las áreas públicas de las ciudades con vallas de publicidad y casetas de ventas que no pagan impuestos y que nadie les ha asignado. En fin, se intenta validar un derecho privatizado sobre espacios que son de todos. Si se asume que esos espacios se han privatizado lo que hay es una colonización incesante de lo público hacia la esfera privada. Pero espacio público no solamente son las calles; es igualmente el ámbito de los medios de comunicación, que sin ser propiedad de todos, sí son por excelencia hoy en día lugares de articulación de lo público y lo privado, paradigmas de las señaladas sociabilidades macro. En ese sentido, es un derecho público el derecho a la libre información, al punto que se lo plantea constitucionalmente, lo que implica, por ejemplo, que el dueño de un periódico o cadena radial, o el Estado —si es el dueño de la televisión que hace concesiones de la misma a las productoras privadas— debe establecer, en condiciones de igualdad, el acceso de los particulares a la propaganda de los artículos y los servicios que se consideren legalmente establecidos en la sociedad. La libre información es, por lo tanto y desde un punto de vista ontológico, una potencialidad. Se tiene un derecho no cuando se lo ejerce necesariamente, sino cuando existe la posibilidad de ejercerlo. No todo el mundo concurre a gozar de los mismos derechos. Todos tenemos derecho, por ejemplo, a ir 187 los domingos a los sitios de recreación públicos: ríos, parques, playas, etc. Sin embargo, si toda la población fuera a los mismos sitios en forma simultánea habría que limitar este derecho, porque llegaría un momento en que no se podría permitir el acceso, produciéndose una limitación del derecho. De allí que el derecho siempre esté visto en términos relativos. Puedo tener derecho al ingreso a la educación superior pero no podemos esperar que la totalidad de los bachilleres ingresen. Ello sería una utopía. No se plantea como frustrante para un régimen social que no permita a todos los egresados de la secundaria ingresar automáticamente a la universidad. ¿Dónde estaría la validez del derecho igualitario al acceso? En los exámenes de Estado realizados sobre la base de condiciones equilibradas que midan aptitudes y conocimientos. La igualdad, en realidad, es siempre relativa. Si se tiene el mismo derecho a aspirar a que los niños ingresen a un ballet, tal condición estará limitada por las restricciones físicas, v.gr., el biotipo exigido por el propio arte del baile: el tener determinada estatura, los pies de una forma dada, etc., lo cual se concreta en pruebas de admisión a las que acceden favorablemente quienes posean las destrezas requeridas. Así, el derecho consiste en la posibilidad igualitaria respecto a un patrón universal; y se ve, al contrario, un privilegio, allí donde no existan patrones de universalidad en la posibilidad del disfrute del mismo. Y en tal sentido lingüístico decimos que un servicio es público, aludiendo a que la totalidad de los ciudadanos pueden disfrutar del mismo. El servicio de telefonía es público porque no se le prohíbe a ningún ciudadano que pueda acceder a él. El derecho a la información es igualmente un derecho público, porque todo miembro de la comunidad puede reclamarlo; pero esto no significa que no tenga límites. Se evidencian si injurio al informar o puedo inhibir una información porque viola mi privacidad. Igualmente en otros terrenos se manifiestan también salvaguardas frente a lo público. En el derecho norteamericano se prohíbe en muchos Estados tomar fotos de las audiencias. De los juicios norteamericanos publican los periódicos y la TV retratos artísticos y no fotografías, lo que constituye un arcaísmo, porque no se podía reproducir en público la audiencia que, a pesar de ser pública, está circunscrita respecto a su publicidad. Otro ámbito de restricción ocurre cuando las informaciones violan secretos de Estado. Y en este último sentido existen muy 188 precisas normas, tales como el manejo de presupuesto de defensa, la política de seguridad interior, que en la mayoría de los Estados modernos está exceptuada del derecho de la información. Ningún Estado-Nación discute públicamente cómo gasta los recursos de defensa porque se supone una cláusula de supervivencia del Estado. Pero ello no implica que el derecho, como generalidad, a la información, no se pueda ejercer. Llegados a este punto vemos a lo público no como generalidad sino en el terreno específico problema de la publicidad. La publicidad puede ser vista como un ámbito de extrañamiento. ¿En qué sentido? El extrañamiento es la condena de la comunidad sobre el individuo. La pena más fuerte que las sociedades primitivas hacían de quien transgredía una ley o una norma colectiva era la discriminación, que pudiera ser pasiva (el rechazo total que llaman Freud y los antropólogos de comienzo de siglo, tabúes). El extrañamiento, como expulsión, era la salida coactiva del individuo respecto a la comunidad y en todo caso eran eventos públicos. En otro escrito (Varela, 1995), he referido el orígen de la expresión “bandidismo”, que posee una doble acepción en inglés (band), el grupo como banda y el bando como enseña o notificación pública. Dichas bandas recorrían las villas denunciando delincuentes, leyendo proclamas, notificando tributos. Este punto es muy importante recordarlo porque en la justicia que realiza el Estado se mantiene en buena medida el elemento de lo público. Es decir, la notificación universal que incorpora en la publicidad dos formas de condena: la condena justicial directa, el tamaño de la afrenta visto desde el Estado y sancionado como multa, cárcel, destierro, pena de muerte, etc., y el tamaño de la condena visto desde la afrenta moral que la comunidad hace. ¿Cuál es la lógica que gobierna todo ello? La reacción de sobrevivencia del Estado y la sociedad sobre el factor discriminado. Es la historia de la persecución a los herejes, las minorías étnicas, los enfermos, los extranjeros, etc. Siempre hay una especie de reacción, no desde el punto de vista moderno o liberal individualista, sino desde el punto de vista del derecho de la comunidad como un ente gregario a defenderse a sí misma de amenazas potenciales: la enfermedad, la locura, las desviaciones sociales. Aún en la sociedad moderna subsiste una tendencia a discriminar, romper y vulnerar lo privado, porque cuando hay una separación se disuelve la privacidad en el sentido utilizado en este texto. Surge otra privacidad: privacía, carencia o ausencia de derechos. 189 Igualmente en otros campos se aplica una suerte de secreto que semeja la reserva sumarial. Por ejemplo en la medicina, a los pacientes se les atiende llevando registros, historiales clínicos que tienen un carácter confidencial. ¿Hasta qué punto una historia clínica puede ser privada?¿En qué momento o bajo qué argumentos las autoridades clínicas pueden develar esta privacidad? ¿Qué pasa con los pacientes enfermos de sida vistos como una amenaza al orden social? ¿Cuáles en estas circunstancias serían los límites de la privacidad? En las sociedades antiguas e incluso en la Modernidad, las cárceles, hospitales, asilos, batallones, etc., son en su conjunto instituciones de reclusión discriminatoria, como si sus “internos” portaran enfermedades sociales. La privacidad, es en dichas instancias, una privacía, una segregación donde la sociedad avecindada actúa bajo la compulsa colectiva del extrañamiento. En este sentido, vale la pena subrayar la tesis de Michel Foucault quien en su ensayo La verdad y las formas jurídicas sintetizó una transición desde la justicia retributiva, señorial, hacia la justicia moderna. “Las prácticas judiciales –la manera en que, entre los hombres, se arbitran los daños y las responsabilidades, el modo en que en la historia de Occidente se concibió y definió la manera en que podían ser juzgados los hombres en función de los errores que habían cometido; la manera en que se impone a determinados individuos la reparación de algunas de sus acciones y el castigo de otras, todas esas reglas, si se quiere, modificadas sin cesar a lo largo de la Historia– son algunas de las formas empleadas por nuestra sociedad para definir tipos de subjetividad, formas de saber, y en consecuencia relaciones entre el hombre y la verdad que merecen ser estudiadas” (M. Foucault, 1991: 17). Resulta interesante además su análisis por cuanto anuda el problema del juicio como institución social con la búsqueda de la verdad, la cual se obtiene por medio de la confesión, la indagación (que por supuesto en muchas ocasiones ha incluido la tortura) y la prueba, procedimiento que legitima el fallo sancionatorio o exculpatorio de la institución social, privada o pública que se encargue de la administración de justicia. Foucault, de otra parte, sintetiza de forma penetrante la naturaleza de la justicia premoderna, señalando que, en esencia, esta es una justicia de tipo retributivo, fundada sobre 190 la reparación automática, valga decir, sobre el automatismo de la pena (Ob. Cit.: 71). Los procesos de juzgamiento premodernos fueron básicamente rituales donde el juez no atestiguaba sobre la verdad, sino sobre la regularidad del procedimiento. En el sistema feudal aparecería una doble tendencia justicial. La concentración de la capacidad coactiva en manos de los más poderosos, los señores, los nobles, los caballeros, con lo cual apareció una forma de privatización de la justicia que consistía precisamente en que los nobles privaron al pueblo llano de su capacidad de ejercer la violencia y en segundo término concentraron el poder sobre los litigios judiciales, impidiendo que éstos se desenvolviesen espontáneamente entre los individuos. Recordemos que Foucault caracterizó la moderna sociedad burguesa como una sociedad disciplinaria, constituida por una red paralela de poderes reclusorios: cárceles, escuelas, hospitales, asilos, panópticos, fábricas, cuarteles militares, etc, y que el pensador francés le otorgó la paternidad de dicho sistema opresivo moderno a las teorías del jurista y publicista inglés Jeremy Bentham11. El criminal es, visto por la teoría liberal, como aquel que perturba la sociedad, que se convierte en virtual enemigo de la patria, de la soberanía popular. Por lo tanto, su emanación política, el poder público, asume de suyo la iniciativa de castigar al criminal y no sólo a quien se rebela políticamente sino, en general, a todo elemento antisocial que se aparte de los imperativos legales del orden público constriñendo de este modo el universo de lo privado en campos tan vastos como la violación del derecho a la propiedad, las normas de los contratos, los imperativos legales que plasman una moralidad dada por ejemplo la bigamia, la infidelidad, el estupro, etc. En este sentido, resulta claro que al asumir el Estado moderno el cuidado de los individuos, su garantía y protección, se estatuye para éste una función moral. En realidad Foucault precisa que el poder no es una totalidad, algo centralizado que se ejerce sobre el resto de los individuos o la sociedad, sino que se constituye a través del conjunto, del tejido de comunicaciones entre las personas. Por tal razón, no comparte la concepción de Hobbes según la cual el poder se ejerce por la fuerza, porque éste es, en suma, toda relación de vínculo entre los individuos y marcara la pulsión ambivalente entre lo público y lo privado a lo largo y ancho de la tradición occidental. 191 NOTAS 1 La crítica a esta identificación de dichos ámbitos aparecería como contradictoria, pero la investigación aquí emprendida permite aseverar, como lo mostraremos en detalle luego, que lo público apareció, que fue una cuestión concomitante con la aparición del Estado. 2 Debe aclararse no obstante que su utilidad para los latinoamericanos es ciertamente relativa ya que focalizan su atención en las experiencias históricas de Francia y tangencialmente del mundo euroccidental. 3 En cuanto concierne a la investigación dirigida por Aries y Duby, ésta da una información que debemos procesar y refuncionalizar para nuestros fines analíticos; porque el objetivo de estos textos no está concentrado tanto en discutir lo público sino en mirar lo privado, y además no lo hacen desde una perspectiva filosófico-política. Obviamente, existe en ellos una dialéctica entre lo público y lo privado que lleva a que un análisis de la vida privada tienda necesariamente a referir y construir un contexto analítico de lo público. 4 Marx anota en los Grundrisse, respecto a las caracteristicas del asiatismo: “...dentro del mismo despotismo oriental y de la ausencia de propiedad que parece existir jurídicamente en él, existe en realidad, como fundamento del 192 mismo, la propiedad de la comunidad o tribu engendrada en la mayor parte de los casos por una combinación de manufactura y agricultura dentro de las pequeñas comunidades, que de esta forma devienen autosuficientes (...) una parte de su plustrabajo pertenece a la comunidad suprema, que existe en último extremo como persona, y este plustrabajo se manifiesta tanto en tributos como en trabajos comunes para la glorificación de la unidad; en parte para glorificación del déspota real, en parte para la del sistema atribal ideal, es decir, Dios.” (De. ob.cit. pág. 429; y cfr El Capital (III, PCE México, 1966 pág 733). Posteriormente escribió que en este sistema los productores directos: “no se hallan frente al terrateniente privado, sino, más bien como en Asia, están en subordinación directa al Estado que se halla por encima de ellos; a la vez como terrateniente y como soberano. El Estado es aquí el supremo terrateniente. La soberanía consiste en la propiedad de la tierra concentrada a escala nacional. Pero, por otra parte, no existe propiedad privada de la tierra, aunque existen la posesión y el uso de la misma, tanto privado como común.” 5 Desde una perspectiva conceptual diferente, la economía distingue los dos ámbitos en virtud de la exclusión o externalidad de los efectos del disfrute de los bienes y servicios. Sin embargo, no pretendo abordar las consideraciones, numerosas y complejas, que desde esta perspectiva existen en relación con la naturaleza de los llamados “bienes públicos”, puesto que tal discusión nos alejaría de los propósitos analíticos del presente capítulo (véase más adelante, el capítulo 6). 6 Uso la expresión despotismo en el sentido clásico de la Ciencia Política tal como aparece en Montesquieu, o sea la idea del despotismo como gobierno absoluto, que no tiene control por parte de sus subordinados y que puede ser ejercido por un solo hombre, por un monarca o por una teocracia, una nobleza o una aristocracia. 7 Esto es más difícil de probar como han tenido ocasión de formularlo los estudiosos de historiografía andina (cfr Murra: 1968,) aunque se puede afirmar que hubo un control sobre los recursos hídricos por parte de los habitantes de la parte alta de los Andes respecto a quienes dependían del agua en la parte costera, o sobre los recursos de una fuente de agua tan fundamental como el Lago Titicaca, en Bolivia. 8 Un buen ejemplo de este problema lo suministra la novelística de Flaubert. Madame Bovary aparece en dicho relato ficcional como una provinciana frustrada casada con un médico que carece de “vuelo”, mientras que sus expectativas hacen parte sustancial del nuevo espíritu pequeño burgués, el arribismo social. Por ello el drama se presenta cuando el personaje anhelando la gran ciudad y desdeñando su vida gris con un médico de provincia, establece su utopía eficiente: ir a los escenarios, adquirir y exhibir un vestuario elegante, etc.; síntomas de las damas de alta sociedad. Y ella ansía, al máximo, hasta su final desgraciado, representar un nuevo rol. Cfr. sobre esta novela y la temática aquí sugerida, a Mario Vargas Llosa Flaubert- La orgía perpetua, Barcelona: Seix Barral 1975. 193 9 Cabe subrayar que un apunte brillante, aunque no original, sobre el rol del reconocimiento dentro de la teoría liberal individualista, que además enlaza lógicamente el pensamiento hegeliano con la citada tradición liberal, lo formuló explícitamente hace unos años el polémico filósofo norteamericano Francis Fukuyama (1992). 10 Locke discute el argumento patriarcal como fuente del poder del Estado y lo rechaza. El poder paternal sólo se extiende sobre la minoría de edad o la incapacidad de la familia, de los niños; pero cualquier hijo, una vez alcanzada la condición de autosuficiencia, ya no puede ser sometido por el derecho natural a la tutela paterna. Tienen que intervenir, dice Locke, elementos de consenso, compromiso, convenio o contrato. Pero tal idea gobierna a muchos de los análisis que ven el origen del gobierno en la sociedad patriarcal: el patriarca es un modelo que luego se reproduciría a nivel del Estado. 11 Debe recordarse en este sentido el concentrado interés de Bentham en el diseño, en el modelamiento institucional, de las repúblicas americanas en trance emancipatorio frente a la corona española. Con base en los desarrollos de la teoría justicial utilitaria de Bentham y otros, surgieron en Europa y en Norteamérica tanto nuevas instituciones disciplinarias, nuevos corpus doctrinarios de derecho público –entre ellos destaquemos el novísimo derecho penal, que desecha los argumentos retributivos– como nuevas ciencias sociales: la sociología, la criminología, la psicología, etc. 194 Capítulo 4 LA CONTRAPOSICIÓN PÚBLICO-PRIVADO EN SU PERSPECTIVA LATINOAMERICANA Como lo hemos visto en el capítulo anterior, se afirma con frecuencia que el patriarcalismo constituye la estructura de poder que replica, fundamenta y legitima el poder soberano del monarca. En tal dirección, arguyeron muchos de los tempranos teóricos del jusnaturalismo al vislumbrar a éste como un tutor de la comunidad, la cual se encontraba en condición de “minoría” de edad, necesitada de protección y consejo por parte de sus gobernantes. En el caso de América Latina, se reforzó el dominio del Estado a partir de una cierta versión jusnaturalista: la idea del pater familiae según la cual el Estado aparece como un vicario del padre de familia y la autoridad paterna. Desde este punto de vista, la legitimación, la acción inmanente sustancial del Estado, supuso el reconocimiento de una productividad del derecho natural que implicó al poder del Estado como si fuese de la misma sustancia del poder paterno. Por ello, en cierta medida, el rey era un padre de la comunidad que la protegía y la tutelaba. Las instancias de privaticidad donde se ejerció un dominio político, sino delegado, al menos virtual, que significaron una red de relaciones de dominio paralelas a las líneas de poder del Estado colonial, fueron los ámbitos del señorío, el cual se transmutaría posteriormente hacia el “cacicazgo” blanco-mestizo dependiente de la voluntad del amo o de la discrecionalidad de la administración política en la respectiva territorialidad. Podremos en este 195 sentido discutir sumariamente el rol de la administración estatal respecto al proceso de las políticas públicas1 en América Latina; temática abordada con base en una visión retrospectiva, de corte histórico-político, del devenir de lo público y lo privado en la vida social del continente. 1. Privatización, patrimonialismo y regalismo ibérico Ello significa tomar en cuenta las grandes fases de nuestra evolución, incluyendo el universo que constituyó el punto de partida de la dinámica estatal, el dominio colonial ibérico, que como se sabe abarcó un ciclo sumamente amplio de más de tres siglos. Asumo este criterio para colocar a prueba la tesis que sostiene cómo buena parte de las características de la administración pública latinoamericana corresponde a una tradición centralista y patrimonial que se desarrolló y tuvo su despliegue a partir del momento mismo del descubrimiento de América con la imposición del paradigma social europeo, en particular su versión castellana. Tesis que alude al menos a dos elementos iniciales: en primer lugar, la estructura política propia de las sociedades precolombinas más complejas, fundamentalmente los grandes imperios andinos y mesoamericanos que poseían al arribo de los primeros conquistadores una forma estatal, la cual cobijaba en su seno millones de súbditos y etnias tributarias; y en segundo término, las especificidades del Estado español, que de cierta forma se reprodujeron en el Nuevo Mundo. La tradición precolombina en ciertas regiones de América y en toda ella la dominación hispánica, afirmaron rasgos singulares a la administración colonial y a la relación dimanante dentro de dicho contexto histórico entre lo privado y lo público. Veamos en primer lugar una imagen vívida de la organización política del Estado español en la época del descubrimiento y conquista de América, recordando que dicho Estado operó en la época del absolutismo. El profesor mexicano Ómar Guerrero afirma sobre los rasgos dominantes en dicho período: “Tan pronto como la estructura económica feudal fue puesta en crisis y el Estado absoluto demandó el monopolio de los medios de administración pública, recientemente integrada por comisarios, 196 inició la época de expropiación de aquello que estaba en manos de propietarios privados. Fue en Francia donde este proceso se manifestó en su forma más pura. El conflicto latente entre el rey y la aristocracia terrateniente que tiene lugar principalmente en los grandes cuerpos colegiados y las localidades también incluyó la administración central del monarca, el ejército, las finanzas y otras ramas que pasan a ser monopolio del soberano. Y, a su vez, el control de la administración central permite al monarca subordinar gradualmente a las localidades implicadas en el control de la administración pública. Esta es pues la centralización gradual a la que es sometida toda la organización del Estado. El monarca absoluto reconoció el contenido estamental de los oficios como criterio de distinción entre comisión y oficio” (Guerrero, 1986)2. El profesor Guerrero describe, a renglón seguido, una serie de rasgos presentes en el Estado español, entre los cuales podremos destacar: a) La distinción entre comisarios y oficiales. Es decir, entre un cuerpo substituible, revocable, de funcionarios directamente vinculados al poder del monarca y, de otro lado, una estructura administrativa más estable, fundamentalmente desarrollada en los sistemas de justicia y en la hacienda pública. Estos tuvieron más acusadamente un carácter de carrera, mientras que el poder político, el ejecutivo, el ejército y otros afines manifestaron una mayor dependencia con respecto al poder del monarca. Así, ocurrió una rápida centralización del poder político, antes disperso en la organización premoderna feudal. b) España al iniciar la conquista de América venía de cumplir un doble proceso: una guerra de siglos por emancipar el territorio del dominio moro y, en segundo lugar, el establecimiento de un equilibrio, hasta entonces relativamente precario, entre las distintas regiones. Los reyes católicos representaban la unión de dos grandes territorios imperiales, Aragón y Castilla. Las legislaciones aragonesa, castellana, catalana, vasca, navarra y, en fin, de otras etnias constituidas como protonaciones poseyeron de todos modos sus especificidades. Sin embargo, no podemos asumir plenamente esta estructura como válida para las Américas porque en España se dio una mayor diversidad de los ámbitos regionales. Pero sí podre197 mos recabar el que a lo largo y ancho de sus dominios la legislación, y en especial el derecho público, estuvieron regidos por la casuística, es decir, aún cuando tuvieron la pretensión de formularse como universales, en el fondo, la estructura legislativa resultó abigarrada3. Adicionalmente, ha de recordarse que España mantuvo una fuerza mayor o un espacio social mayor para las municipalidades. Éstas tenían una cierta independencia y trataron de preservar prerrogativas, aunque éstas eran ya de orden secundario en la época del descubrimiento. La conformación plena del Estado imperial las limita en aspectos decisivos de política fiscal y tributaria, en el manejo de la legislación sobre tierras, en la estructura del poder político; en esos campos se cimentó, en el siglo XVI, un claro predominio y una hegemonía regalista”. c) Otro elemento que caracteriza al régimen español fue un cierto nivel de privatización o delegación del Estado de funciones públicas hacia el ámbito de lo privado, lo cual constituyó una característica dominante del proceso de la “empresa” de la conquista y la colonización en América. Los adelantados, quienes tuvieron a su cargo el descubrimiento y la conquista del territorio, eran en cierta medida agentes del rey, pero al mismo tiempo sus socios. La empresa de conquista4 y colonización de América fue en gran medida, y ello ante todo en sus fases iniciales, una empresa privada contratada por particulares bajo una delegación del Estado central. En los primeros hitos de la Conquista se les dio a los conquistadores (incluso a Colón) títulos de tierras a perpetuidad, revocados posteriormente. Germán Colmenares, al sintetizar este problema de la privatización de las empresas de la Conquista, afirmó: “El carácter privado de las empresas de la Conquista convirtió a los primeros gobernadores que habían capitulado con la Corona en una mezcla de empresarios y aventureros, de caudillos militares y de dispensadores de la justicia. La atribución contractual de funciones públicas y el ejercicio simultáneo de negocios privados dieron un carácter peculiar a los caudillos de las huestes de conquista. Aunque en teoría estaban sujetos a restricciones legales que poco a poco se iban elaborando por parte de una burocracia imperial, de hecho se beneficiaban con la dispensación que proporcionaban la absoluta ausencia de antecedentes de sus acciones” (cfr. Colmenares, 1988). 198 De allí que para un Estado como el español, basado en la exacción tributaria, la catástrofe demográfica que implicó este proceso de enriquecimiento del empresariado del terror, quienes en su afán de enriquecimiento inmediato exterminaron etnias completas de aborígenes, fue terrible. En ese sentido, puede ser entendida la conjunción entre los intereses de la Iglesia y el Estado para limitar el poder de los conquistadores, cuyo epítome fue el padre Fray Bartolomé de las Casas. Además, el Estado español legisló para evitar que los nativos fueran esclavizados. Por ejemplo, desde muy temprano se le prohibió a los pioneros de la Conquista que esclavizaran indígenas. Así, fue propósito del Estado español constituir garantías para que las comunidades indígenas sobrevivieran puesto que constituían la base de la transferencia tributaria. Para expoliar una población tributaria hay que saber cuántos tributarios existen, a qué tareas se dedican, pues sólo con esta base hacendística y tributaria5 resulta posible asignar desde el Estado la división de trabajos de acuerdo con principios estamentarios, establecer los límites en los cuales los empresarios privados (encomenderos, dueños de mitas, mineros) podrían explotar la población aborigen sin que ella fuera exterminada; reglamentar el proceso de importación de los esclavos africanos; la circulación transoceánica de los barcos, y, en fin, la acción de las compañías mercantiles que se encargaban de estas tareas. Recuérdese que allí se daba una instancia de administración local y regional a diferencia del simple poder del padre, que no requería medios administrativos porque la relación cotidiana le daba los elementos de poder absoluto sobre la casa. Además, en la relación social patriarcal cabe recordar cómo las familias eran extendidas; no solamente actuaba el jefe sobre su círculo consanguíneo inmediato, sino que su poder se extendía a otros parientes por afinidad, v. gr. las esposas de los hijos, sirvientes, esclavos, etc., eran integrantes de la instancia privada sometida a su férula, pues el familismo implicaba una vasto universo de relaciones patrimoniales. Sin embargo, definir un vasto campo de patrimonialismo y patriarcalismo en la vida social colonial conllevaría una paradoja respecto a la funcionalidad típica de estas categorías. Ello, en especial por la coexistencia de dicho régimen de dominio privado, al lado de los ámbitos de sometimiento expreso al 199 poder público real. La organización del sistema colonial hispánico implicó la superación de los rasgos simplemente patrimoniales y la configuración de un sistema burocrático señorial. Como lo anotaba Ots-Capdequí, la burocracia imperial fue decisiva para la imposición del dominio ibérico en el Nuevo Mundo, por la fuerza subterránea de las formas privadas de dominio, en la hacienda, mitas, resguardos y, en general, en el conjunto anónimo de la vida civil. 2. Vida familiar y sobredeterminación de lo público Uno de los campos donde, de manera más diáfana, se planteó la sobredeterminación del Estado sobre lo privado ha sido la vida familiar6. Del mismo modo, lo que estaba en juego era el complejo problema de la valoración que hasta entonces se hacía del honor en el seno de la sociedad colonial. Recuérdese aquí que la violación o el sexo impuesto, el rapto, podrían conducir fácilmente a la sacralización de estos vínculos impuestos o fortuitos por medio del proceso de reparación por la vía de contraer matrimonio. Investigaciones recientes han adelantado caminos en la resolución de esta tesis. En particular, me refiero al trabajo de Susan Sokolov, Cónyuges aceptables en la Argentina colonial 1776-1810, y a los ensayos de Ann Twinn, María Isabel Queiroz, y otras estudiosas de las relaciones coloniales entre lo público y lo privado. Resulta útil, en este sentido, analizar brevemente un ejemplo ilustrativo de las pretensiones de la Corona de avasallar los del ejercicio hermenéutico sobre el contenido de la pragmática, ámbitos de la vida privada. Para ello, ilustramos el caso de la real pragmática de 1776 que intervenía en las decisiones de los esponsales sobre el vínculo matrimonial. Como la recopilación de documentos sobre la historia colonial hispanoamericana de Richard Konetzke lo evidencia, una serie de decisiones anteriores iban en la misma dirección. El argumento central de la política del Consejo de Indias tenía que ver con la posibilidad de separar las jurisdicciones eclesiásticas y civiles, y quebrar así los efectos civiles del matrimonio eclesiástico, particularmente en una serie de cuestiones decisivas: la dote. La herencia, la posibilidad de interdictar como ilegítimos a los hijos de tales uniones, etc. Los consejeros del rey se vieron obligados a discutir los argumentos del papado y, 200 en este sentido, se refieren al Papa Benedicto XIV quien había legislado sobre dichos matrimonios ocultos. Aunque el propósito del Papa era limitar dichos matrimonios, en parte porque facilitaban la bigamia, la bula papal mantenía el espacio legítimo de dichas uniones ocultas, sin que dejara de ser deseable el que todos los matrimonios se celebraran públicamente. El papado prescribió, en efecto, las condiciones bajo las cuales estos podrían celebrarse de forma secreta. En real cédula del Consejo se concluye que debe separarse el ámbito de la república civil respecto de la potestad eclesiástica. Allí se afirma: “Que ésta sólo atiende directamente al fuero interno y al bien espiritual de las almas”. La pragmática reconocía la imposibilidad de que matrimonios entre desiguales puedan ser sancionados desde una perspectiva eclesiástica. Las dos instituciones públicas de control del orden social, el Estado y la Iglesia, reconocían tradicionalmente para esta última no solamente un ámbito espiritual sino una serie de prerrogativas de carácter legislativo: el control y la posesión de numerosos bienes y de importantes funciones de control administrativo que muchas veces desempeñaba en nombre del Estado7. El argumento que se señaló para subrayar la urgencia de tal política de control social, fueron los desórdenes que causaba la mezcla de castas y razas. Se reconocía por parte de la Corona el papel de la Iglesia en el intento de arreglar esta situación, pero se afirma la necesidad —según consulta con el Consejo de Indias— de que esta pragmática sea extendida a América con algunas ampliaciones, modificaciones y restricciones, con el fin de darle una respuesta de fondo a tal situación8. Todos aquellos exceptuados, aconseja la pragmática, deben acatar el criterio de respetar y obedecer a los padres. En este sentido, la adaptación de la pragmática en las Indias se fundamenta sobre un punto decisivo: la preocupación esencial sería la conservación del orden social a partir de la estricta separación y delimitación del espacio social de los correspondientes estamentos. El anhelo de los legisladores es evitar, precautelativamente, la quiebra del orden público. Este es, en el fondo, el gran peligro del matrimonio entre desiguales: producir el mestizaje y la mixtura de las castas, que han de permanecer separadas. El orden social dependería del mantenimiento de las separaciones estamentarias, y ello es vital en tanto el orden social. El énfasis que plantea Patricia 201 Seed respecto a una fase diferenciada en la que la Iglesia decae o se repliega, me parece más ajustado al análisis histórico. Hay una disputa entre laIglesia católica y la Corona española que podemos definir básicamente en términos de la resolución de una ambigüedad entre el derecho canónico y el derecho civil. Es decir: el derecho público. En un segundo plano, existe una contradicción entre la ley y la práctica social efectiva. En particular, por el complejo discernimiento que, desde el punto de vista jurídico, significaba el proceso de mestizaje (G. Colmenares, 1991). En la construcción constante del orden social, la mediación eclesiástica fue siempre fundamental, pero desde los últimos decenios de la dinastía de los Habsburgo, el Estado se sobrepuso a los intereses de la Iglesia. Patricia Seed cita el caso de los reclamos de la Iglesia para que las fuerzas policiales intervinieran cuando los padres se opusieran a las bodas de sus hijos. Luego de un tiempo, las autoridades se negaron a ejercer este papel y más bien actuaron en el sentido contrario, o sea, impidiendo los matrimonios no aprobados por los padres9. Pero fue indudablemente cierta la importancia política de la familia como factor que constituyó el humus cultural para el dominio de la hacienda. En suma, tal política pública permitió el surgimiento de formas de organización social estratificadas, constituyéndose un sistema de estamentos, un sentido de pertenencias, una red de lealtades y solidaridades basadas en valores de la hispanidad, de tipo medieval. Cabe destacar, en este sentido, que la forma fundamental de estructuración socioproductiva fueron las comunidades de trabajo, sobre las cuales se asentó la acumulación de riquezas típicas del Estado tributario10. La imagen que podemos tratar de reconstruir hoy es la de un Estado español que pese a la inmensidad de los nuevos dominios intentó controlarlo todo; aunque delegaba ciertos aspectos en el sector privado, otorgándole, por ejemplo, funciones públicas a la Iglesia en campos vitales como la educación y los sistemas de justicia (ver, por ejemplo, la Inquisición); e incluso permitiendo su participación en instancias muy diversas del control político, pues muchos gobernantes de las Indias fueron prelados. Los españoles desvertebraron, en suma, la estructura de poblamiento de América, transfiriendo a su voluntad masas enormes de indígenas, creando nuevas ciudades, poblados y planteándose como criterio básico de la organización social el vivir bajo policía (es decir, bajo orden civil) en comunidad con jerarquías, jurisdicciones delimitadas, bajo un cierto modelo de contrato social11. En este sentido, la forma dominante en los inicios de nuestra vida colonial fue la organización de los indios y de las poblaciones negras en unidades productivas diversas que se basaron en dos fuentes fundamentales de la riqueza: las minas, muy importantes como acumuladoras de capital-dinero, y en segundo lugar, la organización tributaria agrogerencial sobre la base de las “repúblicas de indios” en agencias productivas que evolucionaron lentamente desvertebrando el esquema estamentario formulado desde “arriba” por la Corona (G. Colmenares, 1990). Es este el caso de la progresiva extensión de la propiedad privada, del derecho sucesorial, en unidades agrarias tales como las estancias ganaderas, las propiedades fundiarias de la Iglesia, las haciendas, la pequeña propiedad rural, etc. En cuanto a la hacienda, ésta surgió fundamentalmente a tono con el declive de las encomiendas. En algunos trabajos historiográficos pioneros, por ejemplo en los escritos de Luís Eduardo Nieto Arteta (1942), se manifiesta una afirmación que vale la pena retener. Según dicho autor, el régimen municipal (con el poder de los cabildos) y, en cierta medida, las formas como se tomaban las decisiones en la sociedad colonial americana, implicaban una privatización de las relaciones públicas. Instituciones tan antiguas como la encomienda o las mitas (prestación de servicios de trabajo personal) eran instituciones en que el poder público del soberano se delegaba. Prácticamente desde los comienzos mismos del proceso de conquista y expoliación, se establecieron formas de delegación de lo público hacia lo privado. Desde un comienzo, el proceso del domino hispánico sobre el Nuevo Mundo fue un proceso de privatización. Nunca quedó plenamente claro qué era lo público y qué lo privado. Muchas funciones, típicamente privadas en nuestra mentalidad contemporánea, eran en ese entonces públicas; la separación de los cónyuges, raptos, disputas familiares, etc. Estos asuntos no podían ser zanjados de manera privada entre los partícipes o los protagonistas, sino que sufrían la intervención forzosa del Estado y de la administración eclesiástica. En el terreno de la vida económica fue también muy difuso el horizonte que podría separar lo privado y lo público. 3. El centralismo borbónico en el nuevo mundo Los elementos antes anotados, el reforzamiento del control central por parte de los borbones y, de otro lado, la redefinición institucional de los ámbitos de lo público y de lo privado, no bastaron por sí solos la pacífica sujeción del Nuevo Mundo a los mandatos y dictados de la Corona. En la segunda mitad del siglo XVII, comienza un proceso aparentemente esporádico y discontínuo de protesta social y política, tanto proveniente de algunos grupos indígenas (recuérdese el célebre episodio de la rebelión inca del auto nombrado Tupac Amarú), de los artesanos comerciantes y súbditos plebeyos de las áreas más desarrolladas de la Nueva Granada (cfr. la rebelión comunera en Pamplona, Charalá, San Gil, etc.) o, finalmente, los peligrosos brotes de rebelión y descontento cada día más evidentes entre las ricas elites de los criollos o americanos españoles. Las autoridades coloniales, cada vez más receptivas de estas situaciones, escribían a la Corona solicitando reformas sociales, políticas e institucionales que echaran para atrás las principales directrices de los borbones. Así, por ejemplo, la Real Audiencia de Santafé de Bogotá le escribía al rey unas “precauciones en consulta a Su Majestad dirigidas a conservar la tranquilidad pública del Nuevo Reino de Granada”, en marzo de 1796. El virrey Ezpeleta comenzaba su documento afirmando: “El respeto a la autoridad pública mantiene la conservación de los imperios. La veneración del supremo poder hace felices los reinos. El temor a la potestad superior conserva la paz entre los individuos de la sociedad. Estas sagradas obligaciones precisas a los hombres, no siempre se consiguen por la razón y justicia. Una justa violencia hace efectivos semejantes deberes. De estos respetos nace la indispensable necesidad de las armas... No obedecerían los hombres si el temor de la fuerza no los contuviese” (J. de Ezpeleta, 1875: 275). Con base en la anterior premisa, la audiencia de Santafé solicitaba al rey mejorar sustancialmente el pie de fuerza militar en el territorio. Así mismo, evitar el contrabando de armas y controlar el crecido número de armas de fuego existentes en el reino, reconociendo que, aunque acertada providencia 204 sería el recogerlas, “(...) las presentes circunstancias no permiten abrasar este partido porque entonces se crearía desconfianza en los naturales”. Adicionalmente, proponen morigerar las prohibiciones al libre comercio, puesto que el trato, la comunicación y el comercio son necesarios y benéficos de acuerdo al derecho de gentes (E. Kant, 1987). El comercio bien llevado pacífica las relaciones entre los semejantes, pese a las diferencias políticas y morales entre los distintos pueblos. Es necesario delegar las leyes municipales que han prohibido el comercio, la comunicación y el ingreso de extranjeros en estos dominios. En particular, sugieren que se permita el ingreso de médicos y cirujanos. En tercer lugar, la Audiencia percibe azorada los influjos incontenibles de la ilustración europea: “La variación de los tiempos ha corrompido en gran manera las costumbres: es verdad que en todas las edades se han experimentado estos excesos; también lo es que nunca en tanto grado como en nuestro días. En ellos se han propagado aquí sus libertinas proposiciones, escandalosa doctrina, opiniones detestables. Buen testimonio de esta verdad es lo que consta de los expedientes seguidos por la Audiencia”(Ibíd.: 277). Las autoridades del virreinato consideraron fundamental acrecentar la vigilancia y el cuidado en las capitales mucho más que en los pueblos. Propusieron que el mantenimiento del buen orden público se hiciera premiando virtudes y castigando vicios. Para ellos resultaba fundamental reformar la administración de justicia, enfatizando en los negocios criminales más que en los asuntos civiles. Se quejaban respecto a los métodos de investigación y a la excesiva impunidad. Reclamaban por la falta de oficiales de sala para los sumarios, escribanos, preceptores, etc. Sería necesario crear una sala del crimen compuesta de gobernador, tres alcaldes y fiscal. De igual modo, enfatizó la audiencia en torno al papel de las autoridades eclesiásticas como instrumentos de control del orden público. “La conservación de los dominios americanos se ha de medir irregular por la dependencia de los naturales en todas las clases, sexos y estados; ya sean seculares ya regulares. Lo que no alcanzan muchas veces las armas políticas, vencen las armas eclesiásticas. El púlpito y confesionario son los tribunales que 205 proporcionan a la felicidad pública y particular innumerables beneficios” (Ibíd.). En el documento se exalta el papel de los “padres de la patria”, aquellos ciudadanos que poseen la representación y la voz del pueblo, los honrados vecinos encargados de los empleos de alcaldes y regidores. Para que su oficio se desempeñe con imparcialidad, se sugiere que se dividan por mitad entre europeos y americanos. El documento se extiende largamente en el análisis de la educación en el Nuevo Reino de Granada. Sobre la base de que es conveniente evitar que la corrupción se introduzca en los jóvenes, se sugieren reformas educativas en los establecimientos de Santafé de Bogotá. Pero ello no implica descuidar el control político sobre la comunidad. En tal dirección, se quejan del estado de las cárceles12, que impiden que los reos estén custodiados por tiempo indefinido. Igualmente, reconocen el efecto positivo de los juzgados de vagos para combatir la ociosidad, y señalan la importancia de mantener los padrones o matrículas por parte de los alcaldes para que sólo vivan en las ciudades los que tengan motivo justificado. En general, las autoridades virreinales perciben soterradamente un clima de inconformidad. En una nota reservada del virrey de Santafé, se expresa que en América se ha extendido el espíritu de partido, que divide a los naturales en dos fracciones llamadas vulgarmente “criollos” y “españoles europeos”, respectivamente. Estos últimos en minoría. De allí la propuesta de que los batallones del rey asuman progresivamente puestos civiles tales como cabos del resguardo, guardas de rentas reales. Y sugiere que de España de promuevan la emigración de cadetes jóvenes que deseen ascender por sus méritos con lo cual estas tropas influirían en el mejoramiento de la disciplina para generar el dominio de los vasallos sin generar desconfianza13. Otros acontecimientos como la pérdida de la isla de Trinidad, arrebatada por los ingleses en 1797, las frecuentes fugas de presos tanto en España como en América, entre ellas la muy connotada evasión de Antonio Nariño; el descubrimiento de tentativas de rebelión en Venezuela; la difusión subrepticia de libelos y panfletos revolucionarios que circulaban entre las elites criollas, etc., generaban un clima de incertidumbre que impelía a 206 buscar caminos para satisfacer de la mejor manera el ideario borbónico de reforzar los lazos del control social, tanto sobre la vida pública como sobre las conductas privadas de los ciudadanos. 4. El rol del militarismo en la vida colonial tardía El peso heredado del patriarcalismo colonial será decisivo en la fase republicana. Del criollaje a la oligarquía republicana habrá una línea de continuidad en la conformación de las elites que van a apropiarse del discurso liberal durante el siglo XIX, manteniendo las diferencias y las discriminaciones estamentarias. Los borbones asumieron una política de recelo y desconfianza con respecto a las elites criollas, a los americanos-españoles. Ello se evidenció, por ejemplo, en el manejo de la institución militar, donde cambios tales como reconocer en ciertas circunstancias el fuero militar a pardos y mestizos se asimilaron como ofensas a las antiguas separaciones y privilegios estamentarios. O, en un plano más general, téngase en cuenta la mayor concentración del poder en las autoridades de ultramar respecto a las tradiciones observadas por la Casa de los Habsburgo. Al respecto, recordemos el verdadero rol de las elites criollas: “(éstas) también conservaron el control de los cabildos. Tanto la forma jurídica de sus peticiones como la “exposición de méritos”, descansaban en los valores de un linaje “limpio” y de un comportamiento previo de “vasallo” confiable. Hasta el advenimiento de la Independencia, muchas actitudes de los criollos obedecían a la noción de cortesanos: tanto su existencia social como su situación económica dependían del prestigio y consideración que merecían a los ojos del rey, su virrey y su Audiencia. La obtención de oficios tenía el valor simbólico de reconocimiento de méritos; no en vano Santa Fe era llamada “la corte”. El capital simbólico atesorado por el grupo se vio fuertemente afectado por la creciente subasta de puestos” (M. Garrido, 1993: 34-35). Tómese en cuenta esto en los propios descargos de Nariño durante su proceso, o además recuérdese el carácter vacilante, con lealtad formal al 207 rey, propio de las declaraciones políticas del criollaje en la fase 1801-1811. Al comienzo de este texto señalábamos de qué manera la doctrina liberal recusaba la homologación entre el poder del patriarca y el poder del rey. Como lo ha señalado el profesor chileno Claudio Veliz, la Colonia, exceptuando su fase temprana, no se caracterizó por la presencia del militarismo. Fue la compleja construcción de un orden social basado en estas distancias sociales, la clave para la larga continuidad del dominio español. Veliz apunta que la paradoja de la emancipación criolla republicana fue el surgimiento del militarismo que sintomáticamente obtuvo su base cultural autoritaria en el patriarcalismo que la Corona reforzó poco antes de su desplome. La lucha de clases sofocada a lo largo de la Colonia estallaría con la emancipación republicana, y el patriarcalismo no sería ya garantía alguna de mantenimiento del orden social. Ahora bien, la inexistencia de un férreo poder militar numeroso y bien dotado técnicamente, por parte de la Corona española, durante las tres largas centurias que duró el dominio colonial sobre el Nuevo Mundo, obliga a plantearse un interrogante: ¿Qué tan efectivo en realidad fue el rol de las tropas de ocupación? Una reciente monografía del investigador español Juan Marchena Fernández arroja luz sobre ese tópico. Allí se describe cómo a lo largo del siglo XVIII el ejército español en América fue creciendo en tres grandes unidades, a saber: En primer lugar, el ejército de dotación, con unidades fijas a la manera de guarniciones, en las principales ciudades americanas, con tareas eminentemente defensivas; para precaver a las ciudades, puertos y centros comerciales de ataques por parte de las potencias europeas rivales, incluyendo la piratería. Su estructura era idéntica a las unidades peninsulares, pero “su composición a nivel humano lo caracterizó como un ejército netamente americano; era el núcleo fundamental del ejército de América” (J. Marchena, 1992: 110). En segundo lugar, aparecía el ejército de refuerzo, o ejército de operaciones en las Indias. Este consistía en guarniciones enviadas temporalmente desde España para fortalecer puntos débiles en las líneas defensivas del Imperio, rechazar la amenaza de invasión, efectuar alguna campaña de exterminio, de tal modo que al finalizar su tarea regresaba a España. Y, en tercer lugar, las milicias, unidades de reserva con jurisdicción territorial que englobaban 208 teóricamente la totalidad de la población masculina entre los 15 y los 45 años. El análisis de Marchena muestra cómo desde la Conquista hasta el período de las reformas borbónicas los ejércitos del rey modificaron progresivamente su composición y estructura. En un comienzo eran simples huestes de conquista al servicio de los señores de la guerra y los encomenderos. Posteriormente, en el siglo XVII, fueron tropas destinadas a consolidar la ocupación del nuevo territorio creando principalmente plazas fuertes y realizando la construcción de caminos, puertos y otras obras de infraestructura. En ese tiempo se desarrollaron levas de reclutas, vagos y desterrados, como una fuente para surtir las necesidades del servicio. Marchena recuerda el poco aprecio del español por la vida militar y la impopularidad de dicho sistema de reclutamiento forzoso que dio pie para que el lumpen se cobijara bajo las banderas reales. Desde comienzos del siglo XVIII, los errores e imprevisiones cometidos por la Corona que habían permitido, por ejemplo, el saqueo reiterado de ciudades como Cartagena de Indias y Guayaquil —el desprestigio social de la carrera militar—, hicieron imperiosa la reforma del ejército. Se dignificó el rol social de la oficialidad con una mayor profesionalización del servicio, elitizando el reclutamiento de los mandos. De esta manera, la carrera de las armas fue uno de los mecanismos de movilidad social más importantes que configuraron una nueva oficialidad criolla, “aumentando los naturales de la misma plaza donde estaban en guarnición, por lo que no se desvincularon de sus intereses comerciales o productivos, e incorporaron su grado militar a su más que notable esfera de poder”. Respecto a las milicias, afirma Marchena cómo: “Normalmente en torno a estos núcleos de población y con la participación de las autoridades municipales o del ‘capitán a guerra’ –si en la zona en cuestión no se había aplicado aún el reglamento para las milicias de Cuba–, se establecían una serie de unidades conforme lo permitía la población masculina: piquetes, compañías, batallones o regimientos, según el número de aptos para el servicio (...) las elites locales dirigieron y controlaron este enorme y heterogéneo conjunto social que fueron las milicias americanas (...) buena parte de estas unidades quedó adscrita bien a clanes familiares - en los cuales el patriarca era coronel; sus hijos los capitanes; los sobrinos, los tenientes; los nietos, los cadetes, más 209 los mayordomos y gentes de confianza de la casa que figuraban como sargentos, por ejemplo los famosos regimientos del comercio, en casi todas las grandes ciudades americanas” (Ob. Cit.). En el caso de la Nueva Granada, las milicias fueron desmanteladas después de la insurrección comunera, puesto que ellas fueron la base para crear el ejército rebelde. Por ello, como se recordará, la rebelión sólo pudo ser aplastada con el envío de tropas regulares de la guarnición de Cartagena de Indias. En general, las milicias fueron miradas con desconfianza por los españoles por su ineficacia y potencial peligrosidad. Tal previsión se corroboraría durante los gritos de independencia de las principales ciudades americanas entre 1810-1811. 5. Lo público y lo privado bajo el proyecto liberal republicano Una segunda fase en la relación de lo público y lo privado la constituyó el proceso de conformación de las repúblicas independientes en América Latina; primordialmente bajo el discurso del republicanismo liberal. Durante este período comenzó a utilizarse la expresión liberal, acepción usual en la época, tomada de las cortes de Cádiz que, a grosso modo, significaba para los independentistas, principalmente, la recusación de la monarquía en cualquiera de sus manifestaciones; la soberanía popular como fundamento del Estado; y, finalmente, un régimen de libertades públicas donde el Estado garantizaría el orden social y la realización de los fines privados de los individuos (J. Lynch, 1987). Este trípode se estructuraba en torno de la idea —común en la época— de felicidad; expresión tomada de la Revolución Francesa. Desde luego, no existía un pleno consenso respecto al significado de dicho término. Cabría distinguir, en este sentido, dos matices. De un lado, aquellos que consideraban prioritaria la felicidad pública, tendencia que compartían la mayoría de los ideólogos y políticos revolucionarios. Para ellos la felicidad no constituía el ámbito de los fines privados, de las mezquinas metas individuales, sino que era el terreno por excelencia del despliegue de las llamadas virtudes cívicas. Y, de otro lado, se alinderaron aquellos que privilegiaban las libertades individuales y que en consonancia con tal supuesto construyeron un discurso 210 que condenaba lo español fundamentalmente en virtud de su régimen de monopolios y prohibiciones apuntando a la emancipación del mercado, al despliegue del liberalismo económico. Las constituciones de las ciudades y provincias durante la primera fase de la Independencia, en 1810, tomaron su modelo, con algunas variaciones, de derechos de la convención francesa, traducidos por Nariño. En estos documentos fundacionales se estipularon (tomemos como ejemplo la Constitución de Cundinamarca) que los derechos del hombre en sociedad consistían en “la igualdad y libertad legales, la seguridad y la propiedad”. En este sentido, la libertad implicaba escoger el bien mediante el libre albedrío; toma de opción por un principio católico, opuesto al determinismo calvinista. Igualmente, se definía la libertad como una facultad humana de “hacer todo lo que no sea en daño de tercero o en perjuicio de la sociedad”. Principio, decían los redactores de dicha constitución, “inspirado por la naturaleza, sancionado por la ley y consagrado por la religión: no hagas a otro lo que no quieres que se haga contigo”. Un aspecto fundamental de dichos principios era equivaler la ley a la voluntad general, “explicada libremente por los votos del pueblo en su mayor número o por medio de sus representantes legítimamente constituidos”. En este sentido, se asumía la idea del contrato social, no en términos de un absoluto consenso de todos, como expresión de la mayoría, a través de sistemas de representación. Era un rechazo a cualquier posibilidad de ejercer la democracia directa. Así mismo, se definió el derecho a la propiedad como la facultad ciudadana “de gozar y disponer libremente de sus bienes y rentas, y del fruto de su ingenio, trabajo e industria”. De tal modo que: “Ninguno puede ser privado de la menor porción de sus bienes sin su consentimiento, sino en el caso de que la necesidad pública, legítimamente acreditada, así lo exija; pero aún entonces, es bajo la implícita condición de una justa y precisa indemnización”. Tampoco nadie podría ser privado o perturbado, según los constituyentes cundinamarqueses, del derecho de manifestar sus opiniones por “medio de la imprenta o de cualquier otro modo que no le sea prohibido, en uso de su libertad y propiedad legal” (Ibíd.). 211 Respecto al tema espinoso de los impuestos, que había sido uno de los focos de acres enfrentamientos entre el imperio español y los criollos ricos e incluso con vastas masas de plebeyos en las décadas anteriores al estallido insurreccional, se afirmaba que “ninguna contribución puede establecerse sino para la utilidad general, y por lo mismo, todo ciudadano tiene derecho de concurrir a su establecimiento y a que se le dé noticia de su inversión” (Ibíd.). Otro de los principios liberales que claramente se estatuyó fue el de la participación ciudadana afirmándose que “todos los ciudadanos tienen igual derecho a concurrir directa o indirectamente a la formación de la ley y al nombramiento de sus representantes (...) La soberanía reside esencialmente en la universalidad de los ciudadanos”. Empero, se limitó el rango de los ciudadanos en textos que serían repetidos y reformulados en parecidos términos a lo largo de todo el continente. “No son ciudadanos ni gozan de estas consideraciones: los vagos ni los que por sentencia dada con las formalidades necesarias hayan sido arrojados del seno de la sociedad, ni los que siendo llamados al servicio de la Patria, se excusaran sin legítimo impedimento” (Ibíd.). Frente a la arbitrariedad de la justicia colonial se postuló el principio liberal según el cual: “Ninguno puede ser llamado a juicio, acusado, preso, arrestado, arraigado ni confinado, sino en los casos y bajo las formas prescritas por la Constitución o la ley” (Ibíd.). En lo que respecta al sentido de justicia, cabría diferenciar igualmente dos tendencias. Como corriente dominante una concepción escolástica, neotomista de la justicia, la cual incluso corresponde a un tópico ya citado al comienzo de este texto, a la libertad civil de los antiguos, al decir del ilustrado francés Benjamín Constant. La justicia en la sociedad consistiría en el adecuado diseño de un orden social, retributivo y, hasta cierto punto, estamentario. De tal manera que en su cosmovisión del orden político del Nuevo Mundo “justicia” y “libertad” constituían expresiones equivalentes. Así, sólo habría justicia bajo un régimen expreso de libertades civiles, y alternativamente, un sistema político constitucional para América habría de ser garante de la justicia social. Bolívar, como lo veremos más adelante, trataba de descalificar las acusaciones que, en torno a la configuración de una nueva aristocracia, le criticaban sus rivales al referirse a su recurrente proyecto constitucional, que cristalizaría a la postre en su frustrada idea de una Constitución Boliviana. Este es el caso fundamental, en 212 la arquitectura política del libertador, de su distinción entre ciudadanos activos y pasivos. Tal distinción de dos clases diferentes de ciudadanos la justificó señalando que de esta manera se lograría el propósito de excitar la prosperidad nacional movilizando las dos grandes palancas de la industria: el trabajo y ‘el saber’. Aquí no deja de observarse, de forma reminiscente, un influjo platónico: la antigua diferenciación griega entre el trabajo manual y el intelectual. El libertador señaló que estableciendo restricciones justas y prudentes en las asambleas primarias y electorales se le colocaría un dique a la licencia popular “evitando la concurrencia tumultuaria y ciega que en todos los tiempos ha imprimido el desacierto en las elecciones y ha ligado por consiguiente el desacierto en los magistrados”. Por tal razón, posteriormente se afirmó en el sentido de establecer un senado vitalicio y hereditario. Su pregunta fundamental fue: ¿cómo educar a los descendientes? Este interrogante lo resolvió propugnando que el gobierno republicano ejerciese una tutoría sobre las nuevas generaciones, los descendientes de la elite ilustrada, a través de la creación de un ‘colegio especial’ advirtiendo que “no crearíamos una nobleza sino un oficio” (Simón Bolívar, 1951). La argumentación de fondo que presidió la propuesta tan combatida por los historiadores y por sus propios contemporáneos se basó en un diagnóstico pesimista, pero, a su juicio, inobjetable: la debilidad de la democracia. Bolívar, en efecto, apeló a argumentos históricos. Recordó la contraposición entre los paradigmas atenienses y espartanos subrayando la debilidad del modelo democrático y la fortaleza institucional del primigenio despotismo; luego, retomó los imperios milenarios, China, India, Egipto y Roma. En especial, respecto al mundo romano subrayó que era débil cuando gobernó el senado electo, en cambio fue fuerte y duradero durante la égida de sus libertadores usurpadores. En tal sentido, citaba a Volney, quien decía que los errores del mundo antiguo le enseñan sabiduría y felicidad al mundo nuevo. Bolívar y muchos de los más destacados líderes independentistas concibieron el orden político subsumido dentro del orden moral. Ahora bien, este orden moral no era un elemento estable, un dato fijo, inscrito en la naturaleza genérica del hombre. Bajo el influjo de Montesquieu, los jefes políticos de la Independencia, barruntaban, de cierta forma intuitivamente, una teoría acerca del carácter peculiar del hombre americano. 213 En este sentido, Bolívar apareció como una curiosa combinación entre el soñador y el pragmático. Una de las principales preocupaciones de Simón Bolívar a lo largo del período durante el cual se ocupaba de dirigir militarmente la guerra de independencia fue lograr la necesaria centralización del mando. En particular porque, de un lado, los jefes militares en una dilatada área de operaciones tendían a autonomizarse, a subordinar el poder civil y a ejercer una virtual dictadura de guerra en los territorios controlados. Bolívar afirmaba, como contrario al nuevo orden institucional, la confusión o supresión de las ramas del poder público14. Por lo cual solicita del Consejo de Estado la convocatoria del congreso constituyente para “restaurar nuestras instituciones republicanas”. Bolívar hizo suyos los argumentos de Montesquieu acerca de la separación de los poderes públicos, como una manera de establecer lo que la teoría constitucional denomina frenos y contrapesos. En general, las elites intelectuales militares y políticas del nuevo mundo tenían claro que la revolución de independencia buscaba afirmar la Nación construyéndola autónomamente en términos republicanos. Es decir, en una lucha frontal contra los privilegios estamentarios de la Corona. Por ello, de manera correlativa, habría de ser un pivote del nuevo orden político en reconocimiento de los derechos ciudadanos; “la libertad de obrar, de pensar, de hablar, de escribir, son actos eminentemente liberales15”. Un magistrado republicano es un individuo aislado en medio de una sociedad, encargado de contener el ímpetu del pueblo hacia la licencia; la propensión de los jueces y administradores hacia el abuso de las leyes. Está sujeto inmediatamente al cuerpo legislativo, al senado, al pueblo: es un hombre solo resistiendo el ataque combinado de las opiniones, de los intereses y de las pasiones del Estado social que, como dice Carnot, no hace más que luchar continuamente entre el deseo de dominar y el deseo de sustraerse a la dominación (Ibíd.). Fue reiterativa la insistencia bolivariana a favor de la estabilidad política que sólo podía lograrse mediante un gobierno fuerte capaz de contener el tumulto y la anarquía. Bolívar criticó acerbamente la democracia absoluta calificándola como el régimen de “la libertad indefinida”; y recordando los conceptos sociológicos de Montesquieu, hizo un balance respecto a la composición del pueblo americano: “la mayor parte del indígena se ha aniquilado, el europeo se 214 ha mezclado con el americano y con el africano, y éste se ha mezclado con el índio y el europeo” (Ibíd.). Esta peculiaridad de no ser ni indígenas ni africanos ni europeos representó, en cierta medida, un reto inédito para los legisladores que habrían de guiarse, según Bolívar, por los siguientes principios: “El sistema de gobierno más perfecto es aquel que produce mayor suma de felicidad posible, mayor suma de seguridad social y mayor suma de estabilidad política. Ellos han de tomarse en cuenta para no copiar indiscriminadamente modelos políticos ajenos a nuestras realidades16. Según Bolívar, la felicidad consistía en la práctica de la virtud, lo cual implicaba políticamente reconocer la equivalencia entre justicia y libertad. No podía darse la libertad sino sobre la base del reconocimiento de las leyes objetivas, y por lo tanto impersonales. Al respecto, recordó una frase rousseauniana, “la libertad es un alimento suculento, pero de difícil digestión”. Por tal razón, era una tarea necesaria echar abajo la vieja legislación española, “enciclopedia judiciaria, monstruo de diez mil cabezas, suplicio refinado”. En su lugar, convendría designar un cuerpo estable de jueces, jurados; elaborar un nuevo código civil y penal, en fin, corregir radicalmente los vicios de la judicatura. La meta que propuso Bolívar consistió en que una legislación civil y criminal, análoga a la constitución, dominase imperiosamente sobre el poder judiciario. Una de las utopías más discutidas y conocidas que Bolívar elaboró fue su señalada propuesta del poder moral. Éste tendría como meta regenerar el carácter y las costumbres de los pueblos que, afectados por la tiranía y la guerra, se encuentran en el fondo de una oscura antigüedad, para lo cual se “requiere recuperar aquellas olvidadas leyes que mantuvieron, durante ciertas épocas, la virtud ante los griegos y romanos”. En fin, la antinomia que ha de resolver el congreso constituyente estriba en, a la vez, “moderar la voluntad general y limitar la autoridad pública”, con el propósito de garantizar la menor confrontación posible entre dicha voluntad soberana y el poder legítimo. Bolívar, haciendo un balance en cierta medida pesimista, expresa que “casi toda la tierra ha sido, y aún es, víctima de sus gobiernos”. Todas las variadas formas de gobierno existentes son otras tantas formas de opresión. Y ello gracias a que: “La naturaleza nos dota al nacer del incentivo de la libertad; más sea pereza, sea propensión inherente a la humanidad, lo 215 cierto es que ella reposa tranquila aunque ligada con las trabas que le imponen” (Ibíd.). Los constituyentes de Cúcuta adscribieron el llamado poder moral como un apéndice de la Constitución, aunque en realidad nunca tuvo efectos prácticos. Sin embargo ,su articulado resulta bastante sintomático. Se supone que dicho poder moral lo conformarían un presidente y cuarenta miembros, denominado areópago, con una autoridad plena e independiente sobre las costumbres públicas y la primera educación. Tendría dos cámaras; una de moral y otra de educación. Sus miembros los escogería el Congreso entre los mejores padres de familia quienes, además de ser ciudadanos destacados en el ejercicio de las funciones públicas, serían Padres de la Patria, personas sagradas, objeto de respeto filial por las autoridades republicanas, los tribunales y las corporaciones. Se instalaría así mismo mediante: “Una celebridad extraordinaria con ceremonias y demostraciones propias para inspirar la más alta y religiosa idea de su institución, y con fiestas en toda la república (...) El congreso reglará por un acta especial los honores que deben hacerse al areópago en las fiestas y actos públicos: su traje, sus insignias y cuanto concierne al esplendor de que debe estar revestido este poder moral”17. ¿Cuáles serían las tareas específicas de dicha institución? Premiar y otorgar corona cívica a los ciudadanos virtuosos, nombrar los grandes héroes y hombres, así como ignominiar los viciosos, los corruptos e indecentes. En especial, la cámara moral dirigiría la opinión moral de la república, no ocupándose de los actos singulares de los individuos sino de aquellos repetidos “que constituyen hábito o costumbre”. Tendría igualmente a cargo la censura posterior de los libros, periódicos y otros impresos; pero esta censura no sería política sino moral. De su parte, la cámara educativa tendría a su cargo la educación física y moral de los niños hasta los 12 años de edad. Los curas y los agentes departamentales serían los instrumentos de que se valdría para esparcir estas instrucciones. Tal cámara establecería, organizaría y dirigiría las escuelas primarias especificando la educación que respectivamente convendría a los niños para que adquiriesen, desde su niñez, ideas útiles y exactas, nociones más adaptadas a su estado y 216 fortuna, sentimientos nobles y morales, principios de sociabilidad y patriotismo... la cámara educación, finalmente, convendría a la opinión pública las materias literarias mientras se establece el instituto filosófico18. Este proceso de discusión y la no aplicabilidad del denominado poder moral nos plantea la siguiente reflexión: en primer término, la propuesta bolivariana recogía un espíritu de época, propio a las elites criollas, acerca del papel que le correspondería al Estado, constituyendo una cuarta rama del poder público, articulada desde luego con sectores en capacidad de liderar la sociedad, para que, con base en la ejemplarización pedagógica y en la respectiva sanción moral, en los casos y conductas que lo ameritasen, se procediese a direccionar, elevar y cualificar las virtudes morales sin las cuales la civilidad no tendría posibilidades de concreción. Muy pocas voces en los debates de los congresos de Angostura y Cúcuta se levantaron en su contra. Sin embargo, como lo hemos anotado, el poder moral fue una utopía ya que en modo alguno se implementaron sus normas y procederes. En este orden de ideas, quisiera destacar cómo en el proceso de las revoluciones de Independencia tuvo un arraigo más eficaz el sentido liberal del “dejar hacer, dejar pasar”. Es decir, los sectores liberales que propulsaron la tolerancia, el respeto a la libre expresión de las ideas, ámbitos por excelencia de la libertad individual, se sobrepusieron a quienes pretendieron, desde la órbita de lo público estatal o civil, gobernar y regir las conductas privadas. Formulo como hipótesis explicativa de este resultado la señalada reacción contra el universo de la Colonia, pues ésta, como lo anotara Germán Colmenares (1991), utilizaba como uno de los mecanismos más sólidos de control social el intervencionismo en el ámbito de la moralidad “pública” que, como se sabe, se expresaba en la intromisión en el gobierno desde el Estado y/o desde la Iglesia, las comunidades sociopolíticas respecto a la vida familiar y privada. La reacción consecuente, desde el punto de vista liberal, cerró la posibilidad que bajo argumentos republicanos tal control desde lo público sobre lo privado continuase. Era, desde una perspectiva moralizante, de tal tipo que el Estado, la Iglesia y la propia comunidad avecindada se inmiscuían y sancionaban las conductas privadas. Ahora, en cuanto a lo que pudiéramos considerar como vida privada, su enfoque se dio a través de la exaltación del pequeño suceso, de la 217 anécdota biográfica. Germán Colmenares hablaba, en este sentido, del papel central de la categoría escándalo para marcar las articulaciones existentes desde la tradición colonial entre los niveles público y privado: “El trasfondo ideológico que buscaba mediante la represión la conservación del orden social colonial nos revela los patrones más esenciales de este orden social. Si quisiera buscarse una categoría específica en el lenguaje de la sociedad colonial, capaz de resumir el conjunto de actitudes con respecto a dichos patrones, la mera frecuencia de la palabra escándalo la señala como un buen candidato. El escándalo poseía la virtualidad capaz de convertir en hechos sociales conductas privadas, aún las más íntimas. (...) obedecía a la noción de que la sociedad reposaba en un frágil equilibrio en el que dominaban las apariencias. La estabilidad política y social exigía la aceptación de que ningún acto podía violar las obligaciones morales impuestas por un orden jerárquico. Esta categoría ayudaría a comprender las consecuencias de un localismo de sociedades encerradas en sí mismas, en las cuales el control de la conducta individual se ejercía como una tarea colectiva y en las que el chisme, la comidilla y la conseja aparecían como correctivos sociales, sino como auxiliares de la justicia” (G. Colmenares, 1990). Desde el punto de vista del discurso dominante, que cabría calificar de estirpe santanderista, la función de mejoramiento de las virtudes cívicas se le asignó a la educación. Es sabido el enorme interés que en torno de la pedagogía se tuvo en la etapa inicial de nuestra vida republicana. Según diversos estudios (Aline Helg y Pilar Moreno de Angel, 1989 y David Bushnell, 1954), se pretendió desarrollar un sistema de cobertura nacional que, promovido desde el Estado con una característica predominantemente laica, resolviese el problema del analfabetismo, la superstición y el predominio de los curas en la vida cultural y en las creencias de las comunidades. Santander, en efecto, evidenció un denodado interés por la ampliación del sistema educativo, tanto de la oferta a cargo del gobierno como de la educación en manos de los particulares. Así, reseñaba en su informe al congreso, en 1836, algunos de los avances logrados: 218 “La educación primaria y la instrucción pública han continuado recibiendo del poder ejecutivo, de las corporaciones locales, de los agentes del gobierno y de los buenos ciudadanos, la protección y fomento que demandan su importancia y las peculiares exigencias del gobierno republicano. Reorganizado el colegio de Pamplona, se ha abierto nuevamente a la juventud por los esfuerzos de su gobernador y de varios hijos de la provincia; se ha fundado el patriotismo de los vecinos de Girón; se ha dado una nueva planta al de Panamá, y se han creado nuevas cátedras en los colegios de Antioquia, Chiquinquirá, Vélez, Cali, Tunja, Guanentá, Popayán, Cartagena, Socorro, Santa Marta y Bogotá”. Santander no veía la reforma de la educación únicamente como un problema fiscal, de ampliación de la oferta mediante la creación de nuevos centros de estudio. Una de sus convicciones íntimas era la necesidad de crear una educación alejada del fanatismo religioso19. Pero esto, desde luego, fracasa en parte por las carencias, las incapacidades fiscales del novel régimen y porque el predominio de la Iglesia como institución socializadora seguía siendo formidable20. 6. Centralización política y búsqueda de la unidad nacional Al fracasar la centralización supranacional, tanto en el temprano experimento del congreso anfictiónico de Panamá, como en un plano histórico concreto; al disolverse la Gran Colombia, separarse Bolivia del Perú, Uruguay de la Argentina; al tomar los países centroamericanos instancias autónomas de gobierno, etc., no puede suponerse el fracaso del centralismo, sino el específico declive del ideal o la utopía bolivariana de un orden social estable que trastocara las viejas separaciones territoriales —y por supuesto los arraigados intereses regionales— que sobre este orden se habían consolidado a lo largo de más de tres centurias. En realidad, se consolidó una cierta forma de centralización nacional. Los Estados nacionales se convierten en el proyecto viable durante el siglo XIX y la discusión de política se trasladó al terreno de si se constituirían como repúblicas centralizadas o como repúblicas federales. En América Latina se 219 tomó un camino diverso, ya que países como Brasil, que no tuvo una guerra de independencia, mantuvo una vocación monárquica que dura varias décadas. Incluso el emperador de Portugal en algún momento se instala en América; eludiendo los conflictos europeos, coloca su corte en Río de Janeiro y en Bahía. De otro lado, México y Argentina también, en ciertas fases del siglo XIX, logran estructurar un sistema federativo. En algunas partes de América, en el siglo pasado, el federalismo logró un triunfo relativo, mientras que en otras partes lo obtuvo el centralismo. ¿Qué ocurrió en el caso colombiano? En un primer momento se asume el modelo centralista bolivariano. Pero, a pesar de que se quiera colocar el debate entre Santander y Bolívar en términos de centralismo y federalismo, no se puede afirmar que Santander fuera un federalista. Cuando estuvo en el poder, fundamentalmente en la segunda administración que es muy larga, desde 1832 hasta su muerte, aún cuando no siempre fue el presidente, era quien manejaba el país. El modelo político santanderista fue también un modelo centralista. Santander había luchado como letrado y constitucionalista para crear un sistema político nacional que incluyese un sistema de hacienda pública nacional, un sistema de justicia nacional, un ejército nacional; todo ello bajo el gobierno de una elite política que, a diferencia del modelo bolivariano, no concentraba excesivamente el poder en el presidente sino que establecía un equilibrio de poderes entre el presidente (ejecutivo) y el congreso (legislativo). En realidad, la diferencia entre Bolívar y Santander reside más en que Santander tiende hacia el parlamentarismo, se muestra proclive hacia un congreso que tenga una función relativamente importante, que represente las regiones, elabore las leyes y mantenga una relación de tutela sobre el poder ejecutivo. En estos puntos programáticos residen las célebres y mistificadas contradicciones entre el proyecto de Bolívar y el de Santander (J. Jaramillo, 1982). Alejandro Vélez, secretario del interior, en su mensaje al congreso de 1831, hacía un balance global de la situación de la república, poco después de la separación de Venezuela y cuando la disolución de la Gran Colombia era un hecho. Un aspecto central en sus memorias fue la colocación de una frontera divisoria de la sociedad neogranadina entre civilistas y militaristas. 220 Según Vélez, la situación en que estaba el gobierno no podía ser más crítica. Existía un verdadero desbarajuste del modelo de Estado que la primera administración Santander había trabajosamente construido ya que la anarquía reinaba por doquier21. Las autodenominadas corrientes liberales, es decir, los partidarios de Santander, consideraban decisivo continuar la senda originaria de libertad inscrita en el congreso de Cúcuta en 1821. Por tal motivo, las magistraturas constituidas en 1831 se veían como una continuación de dicho proceso fundacional, aunque su precio fuera el despedazamiento de Colombia a causa del alzamiento de los republicanos de Venezuela y la subsiguiente separación de los quiteños bajo el influjo separatista del general Flórez. Los neogranadinos percibieron en un tono más dramático la separación con Venezuela, en virtud del enorme influjo del recién desparecido libertador presidente, y porque las gestas decisivas de la guerra de Independencia habían ensamblado para la Historia a cundinamarqueses y venezolanos. Según los partidarios de Santander, ello significaba redimir a la Nueva Granada y a la propia Venezuela del yugo monárquico ya preparado. La continuidad de la república dependería de la capacidad de sus líderes para aislar del poder a los antiguos partidarios y sostenedores de la dictadura22. El periodo de la dictadura de Bolívar y sus seguidores fue descrito como un proceso destructor del gobierno legítimo constitucional de la República establecido por los representantes de los pueblos, colocando en su lugar un despotismo militarista. 7. Orden social y libertades republicanas: un difícil balance Ciertamente, los rasgos de dicho proyecto liberal se asemejaron tanto en sus realizaciones como en sus fracasos en la mayor parte del continente. En efecto, se presentó una tensión social enorme que desgarró bajo luchas políticas, guerras civiles, desórdenes, levantamientos, inestabilidad institucional, etc., los intentos en pos de lograr un nuevo orden social. Y ello en virtud de que los hitos centrales y la razón de ser del liberalismo, destruir el orden colonial y construir una nueva sociedad, no se cristalizaron con éxito y el ámbito del 221 interés público se difuminó con frecuencia bajo el predominio del liberalismo decimonónico. Así, ninguna de estas tareas se pudo cumplir plenamente. La Colonia, manifiesta en muchas realidades sociales, culturales e institucionales, sobrevivió en buena parte del XIX. Y los cambios reformistas y revolucionarios, aunque muchas veces radicales y dramáticos, no aparecieron consolidados. Nosotros, al comienzo del siglo XXI, tendemos a aceptar como un lugar común el que se asumiera el modelo de república liberal. Pero hay que imaginarse lo que era la América que se independizaba para pensar que el proyecto político liberal tenía muy poco que ver con la sociedad de ese entonces. Los nuestros eran países mayoritariamente rurales donde el 90% de la población vivía en el campo, en los cuales una parte significativa de la fuerza de trabajo eran esclavos cautivos ocupados en las haciendas, plantaciones, minas y, además, en la servidumbre doméstica. Eran países con una situación de inferioridad para las masas de índios y donde cada vez más aumentaba el peso numérico e importancia cualitativa de los mestizos. En realidad, lo que me parece fundamental es esclarecer el verdadero alcance y contenido (a la manera de una semiótica de la dominación) del discurso criollo liberal que se proponía el establecimiento de un modelo de república con ciudadanía y separación de las ramas del poder público y, en este último campo, que se colocaba como meta la construcción de un juspublicismo laico, igualitario y, por ende, contrapuesto a la heredada construcción jusnaturalista del coloniaje hispánico. De otro lado, al mismo tiempo, el proceso de formación de los nuevos ámbitos políticos positivos (el referido marco institucional del nuevo orden social) implicó por la fuerza de los propios acontecimientos, el desatamiento de los lazos señoriales y devino el predominio de un sentido de libertad jusnaturalista. Pruebas de ello: la preeminencia del derecho natural y la apropiación, desde un puro sentido de privaticidad natural, de un sistema de derechos por parte de los actores sociales, lo cual debilitó en grado sumo la libertad negativa. Preeminencia, agreguemos, de una democracia vista más como libertad que como equidad. El clientelismo, el patrimonialismo, etc., en esta perspectiva, anudaron en la fase republicana dentro de un contexto prepolítico. Por ello, desde entonces una de las tradiciones más arraigadas en nuestra cultura ha sido el culto 222 ritual a la juridicidad, y muchos de los conflictos de intereses y confrontaciones políticas han resultado mediados o definidos a favor de algunas de las facciones en pugna a través de la expedición de sucesivas reformas constitucionales, de nuevas leyes, en fin, de la reforma legal como paradigma o mito viviente. Además, la cultura política de los países latinoamericanos comportó un contenido de lo civil demarcado por la antigua subordinación de las esferas privadas frente a lo corporativo, a las instituciones regalistas y al poder cultural de la Iglesia. Esto ha de ser cabalmente apreciado en su enorme importancia cuando tomamos en cuenta que la civilidad, el orden republicano, se opone conceptualmente a la idea de derecho natural (E. Varela, 1995). Pocos estudios han indagado la compleja estructuración de los sentidos justicialistas inherentes a los ámbitos de la vida pública y la vida privada en la sociedad republicana surgida luego del desplome del vasto aparato administrativo (jurídico, tributario, político e ideológico) cimentado por los españoles a lo largo de tres centurias. Dicha estructuración operó en un doble terreno. En primer lugar, en el deber ser institucional a través de la creación de instituciones estatales de administración de justicia y de control social, entre ellas el propio ejército neogranadino y la articulación entre el novel Estado y la Iglesia, o al menos una fracción de ésta; y, de otro lado, en el discurso liberal republicano, que incidió mediante procesos sociales determinados de apropiación en la moralidad privada y pública de los autodeclarados ciudadanos, las elites criollas blancas mestizas, y de quienes aún permanecieron en la aherrojada condición estamentaria: esclavos, libertos negros y las masas indígenas organizadas en comunidades o subsistiendo como “salvajes” más allá de las márgenes de la civilización. El señorío se transmutaría posteriormente hacia el “cacicazgo” blanco-mestizo dependiente de la voluntad del caudillo, del hacendado o de la discrecionalidad de la administración política en la respectiva territorialidad. Este familismo implicó un vasto universo de relaciones patrimoniales. Obviamente discriminatorias hacia las mesnadas que dieron su sangre en las guerras de independencia; la pervivencia del régimen esclavista hasta más allá del medio siglo en la mayoría de las nuevas naciones del continente; los frustrados procesos de acumulación originaria pensados para dotarnos de una base industrial; la traumática relación entre Estado e Iglesia en virtud del mantenimiento del patronazgo, etc. 223 A manera de ejemplo, recordemos cómo el gobierno santanderista en 1836 expresaba su preocupación por las quejas que de manera numerosa habían presentado los líderes indígenas de diferentes resguardos a lo largo y ancho del país contra la práctica de los repartimientos. El gobierno reconocía la justeza de dichos reclamos, pues muchas veces se abusaba de su “sencillez e ignorancia para despojarlos de una parte de sus propiedades en ilegal provecho de los encargados de repartirlas”. En tal dirección, el régimen liberal criticaba acerbamente la institución colonial-indígena de los cabildos. En efecto: “Se conserva todavía, en muchas poblaciones de indígenas en que no se han repartido los resguardos, el pequeño cabildo autorizado por el artículo noveno de la ley del 4 de octubre de 1821; y además de que la renovación de tales cabildos presenta embarazos, por dudarse con razón cómo deba hacerse, ellos, lejos de proporcionar ventajas a los indígenas, les producen el mal de que se les impongan cargas arbitrarias y pesadas en clase de miembros de dichas corporaciones y de que se mantenga establecida entre sus familias una cierta aristocracia que hace su condición más miserable. Piensa el ejecutivo que deberían mandarse suprimir los expresados pequeños cabildos de indígenas, quedando la administración política y judicial de sus poblaciones bajo el mismo pie de la que se halla establecida en todos los distritos parroquiales de la República” (L. Pombo, 1836: Vol. V, 77). En dicho contexto podemos ubicar el rol de textos pioneros, como los de Juan Friede, que discuten la condición del índio en Colombia. Y su oposición respecto a puntos vista más “oficiales”, como los de historiadores tradicionales (v. gr., Eduardo Santa, Lievano y otros), que exaltan como uno de los logros fundamentales de la novel república liberal, la extensión de la condición ciudadana, hecha en el congreso de Cúcuta, hacia los grupos étnicos nativos. Ello evidencia hasta qué punto la polémica sobre el grado efectivo de integración de los anteriores grupos estamentarios a la condición republicana, ciudadana, ha estado presente también en Colombia23. Como tuve ocasión de explicarlo en otro texto (E. Varela, 1990), lo que ha tenido un débil arraigo puramente formal, ha sido el justicialismo demo224 constitucional que se supone a sí mismo dentro de un ámbito universalista. Por ello, las propias elites republicanas procedieron a definir un sistema igualmente arbitrario y privatizador en la aplicación y calificación de los delitos comunes y políticos. Es pertinente, por lo tanto, analizar las formas específicas de ejercicio de este justicialismo privado en el proceso mismo en el que se afirma como un espacio institucionalizado de control social, en tanto jusnaturalismo redivivo. Las elites parcialmente herederas de las prerrogativas de los peninsulares instauraron un sistema de imposición del sentido de justicia donde valores tales como el honor, la virilidad, la palabra empeñada, etc., de estirpe hispánica, fueron el eje justificador de la desigualdad liberal. Igualmente, pretendieron conciliar lo anterior con un cuerpo doctrinal utilitarista que significó la rotura con el anterior espíritu jusnaturalista. Se trataba de naciones que intentaron olvidar ese pasado y lo consideraron enterrado, vinculado a la nefasta experiencia colonial. La república liberal, particularmente en el campo agrario, intentó desvertebrar, y de hecho lo logró en buena medida, en toda América la estructura de las antiguas comunidades indígenas. En Colombia hay ejemplos muy claros como la disolución de los resguardos y las leyes que buscaron descongelar propiedades rurales que podían integrarse a los círculos de compra y venta mercantiles. Con esta política se buscó que las propiedades comunales, unas de origen indígena y las otras de la gran superestructura institucional de la Iglesia, se descongelaran e ingresasen al mercado (Colmenares, 1978). La disolución de las comunidades indígenas estuvo vinculada a la ampliación de los horizontes del mercado, no sólo del mercado de la tierra, sino de la fuerza de trabajo, la circulación dineraria y la producción para el consumo. Integrar al índio al consumo para que produjese más allá de las estrechas esferas de la economía natural sería una de las tareas constitutivas de la creación de un mercado interior para las nuevas repúblicas. No se trató de repúblicas liberales que intentaron volver hacia lo indígena sino que partieron de cero, negando radicalmente a España y al mundo de la Colonia. ¿Era el liberalismo un discurso con arraigo en las anónimas masas indígenas, negras, mulatas y mestizas?, y ¿de qué forma afectaba este discurso la 225 apropiación popular del sentido de justicia? Por ahora quiero subrayar el hecho de que el concepto de libertad manejado por el liberalismo en el siglo XIX fue dúplice. Recordemos la conocida expresión de Benjamín Constant de libertad de los antíguos opuesta a libertad de los modernos. Puesto que esta conceptualización nos da la clave para percibir la conexión entre lo público y lo privado, entre sociedad civil y sociedad política. Hasta donde he estudiado el fenómeno, y esto es una hipótesis, creo percibir que en los orígenes de las repúblicas latinoamericanas estuvo presente tal dicotomía. Y ello resulta particularmente ilustrativo si analizamos el contenido del discurso republicano de Simón Bolívar: “Para cualquier republicano el recurso a la imagen mitológica de imperios índios era aún más inquietante que el recurso a las antiguas constituciones feudales. Todos los intentos por revivir el simbolismo político del mundo indígena habían sido confinados, por razones obvias, a México y Perú, de lo cual estaba plenamente consciente Bolívar. Pero también estaba consciente de que podían ser generalizables esos intentos (...) por medio de una confusa y popular exigencia de libertad promovida no en nombre del pueblo sino en representación de los indígenas oprimidos y de las masas mestizas. El pueblo, para Bolívar, significaba algo muy parecido a lo que significaba para Montesquieu, a saber, los “ciudadanos”, esto es, aquellos que tuviesen una condición y educación suficientes para poder participar de manera plena en los procesos políticos” (A. Padgen, 1991). Si asumimos la tesis de Padgen, podríamos indagar en torno a la existencia de dos sentidos justicialistas en la república neogranadina, uno público, político; y otro privado, civil. En este orden de ideas, la investigación debería evidenciar cómo opera el entrecruce entre lo público y lo privado a nivel del funcionamiento (y apropiación de éste por los diversos grupos sociales) del aparato estatal. Ello expresaría igualmente el topos del enfrentamiento entre un proyecto de modernización y la defensa de las tradiciones. El conocido historiador argentino Abelardo Ramos afirma que a lo largo y ancho de América Latina se intentó la construcción de la nación, ante todo, buscando liquidar el pasado colonial. De allí la recurrencia de los enfrentamientos en torno a los bienes eclesiales y cofradiles de manos muertas, la política de disolución de los resguardos y demás bienes 226 de las comunidades indígenas, y la adopción de un criterio de alienabilidad del conjunto de la propiedad donde incluso la propia perspectiva de lo ejidal comunal tendría fuertes variaciones. Uno era el ejido colonial y otro el ejido municipal republicano. En síntesis, la ideología cardinal de dicho proceso se encuadró en la propuesta dicotómica de civilización contra barbarie qu,e al final del siglo, propugnarían los más lúcidos ensayistas del continente. El liberalismo, como espíritu de estas elites criollas formadas por la ilustración europea e influidas por la revolución norteamericana, con toda su problemática de derechos ciudadanos, construye un paradigma que se sobreimpone y que, desligando el modelo monárquico y la antigua estructura colonial, deja un camino, una única alternativa histórica: la república liberal. Entonces tenemos el caudillo clientelista, el caudillo militar, etc., originados o basados en torno a la hacienda y, en términos más vastos, al señorío; paradigma de organización y racionalización de la vida económica y social que desplazaría el papel del Estado. Este retroceso del rol estatal en la época del liberalismo republicano no es sólo fruto de las anteriores tendencias centrífugas. De otro lado, el propio discurso liberal dominante en el siglo XIX, cuando discute el problema de la educación, subraya que el Estado no promocionaría campañas de alfabetización ni construiría escuelas, colegios y universidades puesto que tales actividades deberían hacerse en función de las necesidades de la demanda y serían, por lo tanto, proveídas por empresarios particulares bajo el paradigma de la libertad de empresa. Se ve cómo el manejo educativo —al igual que muchos otros campos de la política pública— son, en cierto sentido, reprivatizados; dejan de ser jurisdicción del Estado y pasan a ser un campo proveído por la propia comunidad a través de la iniciativa particular. Recuérdese que incluso la llamada seguridad y el monopolio del uso de la fuerza, garantía esencial de lo que constituye la legitimidad de un Estado, se le delega a los particulares puesto que se defendía el armamento general de la población bajo una filosofía individualista según la cual los ciudadanos tienen el derecho a portar armas y a su legítima defensa. En suma, educación que no cumplía el Estado, sistemas de justicia que la tradición patriarcal bajo los códigos de honor y la venganza de sangre adscribían al ámbito de lo privado, armamento general de la población, débil 227 poder del Estado central con respecto a los caudillos locales y regionales, y una débil base tributaria fueron las características que, en cierta medida, hicieron imposible que el proyecto de Estado liberal republicano se pudiera construir en el medio siglo posterior a la Independencia. Por ello, el proyecto de Estado colombiano, al igual que en muchos otros países de América Latina, se realizó primordialmente en la esfera ideal, en el deber ser institucional y/o constitucional. Por tal razón, el estudio de la formación de la Nación-Estado ha sido obra primordialmente de la historiografía institucional y del derecho público, visto desde una perspectiva genealógica (cfr. Restrepo Piedrahita, 1979). De otro lado, el análisis de este proceso de conformación del tejido social ha estado a cargo de la historiografía económica. En este sentido, se han construido diversos modelos interpretativos en relación primordialmente con los ciclos de expansión económica, migración, cambios demográficos y, posteriormente, con los procesos de industrialización y modernización social24. Los liberales decimonónicos propusieron por tal circunstancia un modelo que a la larga resultó utópico intentando crear desde arriba una identidad nacional. En realidad, hemos de reconocer que la Nación y el Estado se encuentran aún en proceso de creación. En América Latina se construyó idealmente, arquetípicamente, al Estado; pero, incluso hoy, estamos buscando configurarnos como Nación. Quizás el ejemplo más claro de la creación artificial de naciones fuese el caso de Bolivia. Como se recordará, el libertador Simón Bolívar le encargó la tarea de crear una institucionalidad para esta región del denominado Alto Perú a su protegido Sucre. Pero en dicho proceso no reconocieron que allí existía ya una cultura nacional: la indígena. Por verla tan indigente, situación lógica después de tres siglos de dominación, la dieron por perdida y en condición de servilismo; sintiéndose los criollos en el deber de volverlos ciudadanos, introdujeron como principio integrador el del mercado con la pretensión de aniquilar las primigenias comunidades, forma por excelencia de la organicidad colonial (M. Carmagnani, 1984). En síntesis, muchos de los rasgos premodernos de la organización del Estado y de las relaciones políticas en América ´Latina, aún manifiestos a lo largo del siglo XX, corresponden a tradiciones que han logrado sobrevivir, en cierta medida, refuncionalizadas. Y aunque tales marcas han tenido adaptaciones en 228 la manera como se articulan con la actual sociedad (puesto que es diferente la sociedad capitalista o con elementos predominantes de capitalismo propia del siglo XX, respecto a las sociedades coloniales, o a las republicanas del siglo XIX), estos elementos se han mantenido, sino incólumes, al menos como importantes rezagos o trabas en las relaciones políticas y en la configuración del Estado nacional (J. O. Melo, 1991). Pudiera decirse que el clientelismo de hoy corresponde a la adaptación “modernizada” de la tradición patrimonialista; y el centralismo de hoy ancla igualmente en una lógica del poder político que se remonta a los orígenes mismo de la constitución del Nuevo Mundo. Muchos estudiosos señalan la importancia de esta tradición que arraiga nuestra cultura de la violencia, particularmente en la tipología de las bandas armadas, en sus métodos, secuestro de rehenes (político o económico), boleteo, homicidio fuera de combate, tortura, exterminio de pobladores civiles, etc. Estas continuidades nos afirman en la convicción de que las elites y las masas, pese a su distancia, han compartido un humus uniforme en el que ambas, de forma dialéctica, avivan mutuamente los rasgos de sedición, clientelismo, depredación e ilegalidad. Actúan de consumo en el plano de una doble moralidad. En este sentido, resulta conveniente subrayar cómo en los conflictos sociales existen puntos de contacto entre la cultura política de los de arriba y los de abajo. La lógica del caudillismo y la cultura patriarcal son evidencia palpable de lo anterior. Y tras ellas aparece como figura tutelar la privatización de las relaciones públicas, que ha implicado, desde el siglo XIX, el arraigo de la justicia privada. 8. La preeminencia del intervencionismo estatal sobre lo público Posteriormente, hacia el inicio del siglo XX, se impulsó el paradigma centralista del Estado intervencionista (J. M. Keynes, 1988), lo que se dio con mayor vigor en algunos países que lograron afirmar una base industrial sólida implicando procesos de acumulación de capital significativos, como ocurrió en Argentina y Brasil. Naciones que, de otro lado, promovieron la inmigración masiva de millones de europeos conformando una suerte de “islotes” o archipiélagos de lo “occidental”, los cuales sirvieron como base social para la 229 destrucción violenta del pasado. La épica de José Hernández, en su conocida saga gaucha Martín Fierro, retrató muy bien este programa de reconversión de las anteriores realidades sociales. En el caso colombiano, se pretendió un programa positivista de industrialización y de libre cambio. Pero este proyecto, como lo acabamos de anotar, igualmente fracasó y al final del siglo XIX, fue el cesarismo de la Regeneración el que tuvo la posibilidad de darle un sentido realista al proyecto de configuración de la Nación-Estado. Para ello se restablecieron los viejos pactos y ligazones entre Estado e Iglesia y se asumió un modelo exclusivista, elitista y oligárquico como alternativa de cohesión, primordialmente entre las elites que lideraban el señalado proceso. En realidad, desde las tres últimas décadas del siglo XIX, se comenzó a vislumbrar un proceso de desarrollo capitalista, particularmente bajo el marco de la integración del continente a la división internacional del trabajo. Con base en la funcionalidad de las llamadas por Adam Smith ventajas competitivas, estos países se convirtieron en proveedores de alimentos, productos agrícolas y materias primas diversas con destino al mercado de la metrópolis desarrollada, fundamentalmente de Norteamérica y Europa. En el caso colombiano, ciertos productos de exportación cumplieron ciclos efímeros. En efecto, la quina, el añil, el tabaco durante ciclos relativamente cortos tuvieron un auge en el mercado internacional; pero sólo con la economía cafetera, se estableció una base de acumulación originaria de capital que permitió la constitución de un mercado interior y la dinamización de procesos demográficos y de urbanización, que actuaron como un contexto más sólido para la conformación de la Nación-Estado, al menos desde un punto de vista socioeconómico (S. Kalmanovitz, 1988). En lo atinente al proceso propiamente político, este paso se comienza a concretar con la derrota del liberalismo radical en 1885 y con la reforma constitucional y el modelo político que establece Rafael Núñez en sus alternados períodos de gobierno, durante los tres últimos lustros del final de siglo. Desde luego, estas reformas no operaron inmediatamente, y básicamente su concreción se logra sólo en los primeros decenios del siglo XX. A lo largo de la hegemonía radical, la idea base de la organización unita230 ria de la república, la consolidación del Estado nacional, no fue un objetivo expreso de la programática del devenir político. Sin embargo, la crisis económica, social y político-militar que sacudió a la república federativa generó el fermento para una escisión del liberalismo, que sería, a la larga, decisiva para la conformación del esquema estatal y político que dominaría la vida de la Nación a lo largo del siglo XX. El sector liberal encabezado por Rafael Núñez, en el cercano pasado muy vinculado con el radicalismo, se separó y estableció una coalición con la mayoría del conservatismo. El alejamiento doctrinario de esta facción liberal respecto del ideario del llamado “Olimpo radical” se basó en dos puntos capitales: la denuncia del federalismo como forma de gobierno, acusándolo del caos, la crisis económica y la parálisis de cualquier plan de modernización y desarrollo. Recuérdense aquí los grandes problemas que, en razón de la debilidad del Estado central y de las numerosas contiendas armadas, enfrentaron proyectos tales como la navegación a vapor por el río Magdalena, la construcción de los ramales más importantes de la red ferroviaria; la suerte de algunos programas de industrialización como las ferreterías de Pacho, etc. En segundo lugar, en el plano ideológico, esta coalición intentó recuperar el rol de la iglesia católica como factor de cohesión del orden social. La Constitución Política de 1886 marcó, un punto de viraje. El programa radical cedió el turno al ejercicio de una compleja coalición en la que los sectores señoriales, las elites políticas y los círculos empresariales, aunados, propugnaron por la concreción de la unidad nacional. Si se quiere hacer una comparación histórica, diríamos que así como la burguesía europea logra la unificación nacional aliándose a las monarquías para combatir al disgregacionismo feudal, en Colombia, la unidad nacional es fruto de un plan sui generis de grupos de terratenientes, la Iglesia y un sector de comerciantes saturados del caos promovido por el radicalismo. Éstos, frente al alegre utopismo ideológico de aquéllos, pactan con la institución que —desde el punto de vista de su fuerza institucional— podría actuar en pos de dicha finalidad. De este modo, con la complacencia de una fracción importante del liberalismo, la Iglesia recupera buena parte de las prerrogativas de las que gozó a lo largo del período colonial (G. Molina, 1981). Por esta razón al acto de reformas de la Constitución se 231 le suma, un año después, la suscripción de un concordato con la Santa Sede. El corolario de este proceso fue la derrota de los últimos ejércitos y guerrillas del liberalismo en la Guerra de los Mil Días. Al producirse este hecho, este puntillazo al utopismo militarista, un plan hegemónico del conservatismo y el sector neonacionalista permitieron la irrupción, en las décadas iniciales del siglo XX, del mencionado programa positivista. Esto alteraría de manera incesante el cuadro del país decimonónico. Con los inicios de la industrialización surgirían clases modernas y el universo cultural se renovaría considerablemente. Los preceptos constitucionales de esta reforma y el nuevo balance de las fuerzas políticas desvertebraron el poder de los caudillos. El Estado central consolidó un ejército que reemplazó las milicias privadas enseñoreadas de la república a lo largo del siglo XIX. El caudillismo clientelista, desde luego, no desapareció en el curso del citado proceso de nacionalización de la vida pública. Pero los rasgos del caudillaje refuncionalizado fueron desde entonces diferentes. Estos nuevos caudillos no se basaron exclusiva ni prioritariamente en el poder privado de la hacienda y no constituyeron una fuerza de la misma naturaleza del caudillismo patriarcal del siglo XIX. Efectivamente, siguieron manteniendo elementos de autoritarismo y patriarcalismo; pero la lógica funcional en la que se inscribieron fue distinta. Esta es una característica de civilidad que el país logró establecer desde comienzos de la presente centuria. Desde el punto de vista de la administración pública, se dio un proceso en el que el Estado empezó a recuperar funciones que el viejo Estado colonial tuvo y que en la primera fase republicana había perdido. En primer término, se logró por el Estado central el control sobre las regiones y la subordinación de los departamentos y provincias al poder presidencial. En segundo lugar, la señalada disolución de las milicias y su sometimiento por un ejército nacional permitió el monopolio estatal del uso de la violencia. En tercer lugar, se configuró una legislación nacional, particularmente en el campo de la justicia. Este fue un proceso complejo porque Colombia no tuvo códigos penales efectivos de ese tipo en el siglo XIX, puesto que en esa fase la legislación penal era casuística, en parte heredada de la Colonia y adaptada con ciertos criterios tomados del ideario de derechos y deberes propios del liberalismo. 232 En realidad, un código penal que abarque y sistematice, desde el punto liberal republicano, el monopolio de la justicia del Estado solamente se plasmará en la administración López Pumarejo con la reforma del Código Penal del año 1936. Desde el punto de vista tributario también se evidenció una recuperación del poder del Estado central. Las finanzas que habían sido delegadas a las provincias y municipios fueron de nuevo recapturadas por el Estado. Empezó a desarrollarse un proceso de tributación sobre la propiedad predial y volvieron a tomar importancia como factor del financiamiento del Estado los recursos provenientes de aranceles y aduanas, puesto que desde el gobierno de Núñez, en cierta medida, se reinstaló el esquema proteccionista dejando de lado el esquema que caracterizó al liberalismo radical. Con ello se fortalecieron los fiscos, particularmente a través de los impuestos sobre las exportaciones cafeteras. Así, la economía agroexportadora cafetera, en buena medida, contribuiría al fortalecimiento del Estado central. El Estado trató de comprometerse en el desarrollo económico, particularmente con la inversión de recursos en el campo de la infraestructura y las obras públicas, v. gr. la constitución de una red férrea nacional, puertos marítimos, etc., complementarios del tradicional sistema de transporte fluvial (J. A. Ocampo, 1992). En síntesis, este fue un esfuerzo que acentuó el predominio, la ascendente hegemonía de la burocracia pública, que lentamente asumió funciones nuevas; burocracia intelectual de abogados, ingenieros, servidores del Estado, que reemplazó poco a poco al paradigma decimonónico del burócrata letrado, incompetente en la gestión técnica. Con ello se establecieron mediaciones mucho más sofisticadas de carácter clientelista entre el poder político y la administración del Estado. Fenómeno que no existió de manera sólida en el siglo XIX, probablemente por la simple razón de que el Estado era muy pequeño y el tipo de gajes y prebendas que podían obtenerse eran ínfimas respecto a las que suministraban la hacienda, el señorío y la explotación privada de las actividades económicas. De otro lado, el positivismo económico fue una ideología asumida con gran vigor en América Latina y cuyo máximo mentor fue el porfiriato mexicano. Sus propugnadores argumentaron que las guerras civiles —en las que se batallaba por ideas, las más de las veces abstractas y/o utópicas, carentes de sentido concreto— sólo 233 dejaban empobrecimiento y destrucción de la infraestructura material. Por ello, como condición indispensable del progreso, sería necesario suspender este tipo de conflicto entre las elites y colocar, fungiendo la administración estatal como árbitro, a estas facciones de acuerdo con la necesidad de que el progreso, visto en términos de desarrollo económico, llegara a América Latina. Modernidad, ansiada quimera plasmada en la plantación cafetera, la producción agrícola para la exportación, el desarrollo de carreteras, puentes, ferrocarriles, infraestructura básica de las ciudades, urbanismo, etc. Igualmente, se propugnó la creación de universidades que permitiesen formar elites tecnocráticas buscando reformar la universidad colonial, que todavía sobrevivía en el siglo XIX, para que nuevas carreras y disciplinas surgieran. Es decir que la elite empresarial supondría su consolidación con base en un cierto desarrollo de procesos de industrialización y de conformación del sector de servicios públicos. En este último campo, resultó titánico, por ejemplo, el esfuerzo de las últimas administraciones neoconservadoras y republicanas conservadoras inspiradas en Núñez para crear una banca central. En efecto, uno de los grandes problemas del siglo XIX, prerrequisito para la modernización económica en los países de América Latina, fue la falta del sector financiero y de la banca central. En Colombia, la actividad bancaria se desarrollaba en las regiones, v. gr. Santander, el Viejo Cauca, Antioquia, etc., donde una banca incipiente e inestable, por autorización de los gobiernos provinciales, emitía dinero; aunque muchas veces ese dinero tenía dudoso respaldo o se utilizaba para operaciones especulativas existiendo una gran desconfianza del público, por lo cual se seguían utilizando los metales preciosos, las transacciones en oro y plata. Estos son aspectos en los cuales se plasmó el ideario positivista del desarrollo económico promovido desde el Estado. Éste, además, impulsó la iniciativa privada pero entendiendo que ella no sería fructífera sin el respaldo del sector público. Proceso que implicó además la asimilación creciente por éste de funciones diversas, v. gr., la educación para formar elites, la salubridad, el aseo de los espacios públicos urbanos, etc. De ello resultó que bajo el ideario del positivismo, en lo concerniente al caso colombiano, se reinstalara el esquema centralista diseñado inicialmente por el imperio español, incluso manteniendo prerrogativas eclesiásticas que databan de ese periodo. Este régimen político con su estructura bipartidista y 234 este modelo constitucional, con sus innumerables reformas de pequeño calado, constituyó durante buena parte de la actual centuria una suerte de síntesis de los rasgos predominantes de la cultura política nacional. El culto ritual al cuerpo normativo, el parlamentarismo clientelista y la permanente dependencia de los partidos respecto al tutelaje del aparato del Estado, etc., han conformado una trama de democracia limitada, recortada, sustento de las manifestaciones de violencia social y política, del autoritarismo en sus diferentes manifestaciones. No puede dejarse a un lado, como factor determinante de las formas políticas y jurídicas del Estado nacional, la imposibilidad histórica de plasmar el conjunto ambicioso del programa modernizador. La sustitución, o mejor, el tránsito del señorío colonial al señorío republicano; la débil piel de unas instituciones republicanas transplantadas de la experiencia europea y norteamericana, produjeron ese peculiar tramado de nuestra historia, donde de todas formas resultó un rasgo innegable, por lo menos durante la primera mitad del siglo XX: la preeminencia del poder político laico y civil frente al militarismo. Así, la reconstitución del ejército nacional en los comienzos del régimen de Reyes e incluso los reacomodos que se efectuaron durante los años 10, estuvieron encaminados a otorgarle al brazo armado del Estado una posición subordinada de frente a la esfera política; vale decir, a los grupos partidistas que tienen acceso al control del gobierno. De allí la profesión de fe en los principios de neutralidad, la no deliberación y la supresión del reclutamiento forzoso como pilares constitutivos de las Fuerzas Armadas. Ahora bien, esto en modo alguno quiere decir que se esfumen las prerrogativas en el manejo de los asuntos militares. Este tópico puede sintetizarse afirmando que la administración pública liberal republicana es tardía; la programática general de la Independencia logra configurarse parcialmente sólo en los últimos tres decenios del siglo XIX y en el siglo XX. Los dos últimos puntos, de alguna manera ya vistos, describen cómo el Estado intervencionista configura una relación patrón-cliente que es diferente a la del siglo XIX. El patrón-cliente para la administración pública del siglo XX cambia, ya que el Estado es ahora, y de forma significativa, el mediador retributivo por excelencia. De otro lado, aunque la administración estatal empieza a cimentarse sobre el pivote de una burocracia moderna, en el sentido weberiano, subsisten la 235 gestión moderna y el patrimonialismo, fenómeno que también ocurre hoy en día. El patronazgo-cliente fue una relación conservadora en tanto que la conformación de una burocracia moderna aparece como lo progresivo, no en el sentido de progreso sino en el de dominante. Este elemento también puede vislumbrarse en el Estado de beneficencia o benefactor impulsado particularmente por las reformas liberales. También tiene que ver con procesos de urbanización y con el crecimiento de la administración pública, por ejemplo, en el número de funciones asumidas, no sólo a nivel central, sino en las regiones y los municipios. 9. La crisis del interés público en el contexto colombiano Finalmente, sin pretender una elaboración exhaustiva del tema, quiero plantear algunos elementos medulares que expresan la quiebra, práctica e institucional, del predominio del interés público encarnado en la arriba referida estatización y hegemonización gubernamental de las formas de control social propias del discurso constitucionalista demo-liberal. En primer término, quisiera destacar una comprobación fácilmente demostrable. En efecto, una característica primordial de la crisis de gobernabilidad en Colombia y algunos otros países latinoamericanos —no me atrevo a globalizar en demasía— consiste en que el interés público se encuentra, desde hace varias décadas, en un proceso de destrucción en tanto categoría legitimadora del orden social. En realidad, la noción de bien común, la vinculación societal sobre la base del acatamiento de una serie de normas que constituyen la moralidad social ha progresivamente perdido vigencia. Por supuesto, como salida a dicha crisis, las personas buscan de alguna manera trascender, desde sus finalidades privadas, hacia fines colectivos superiores, pero no encuentran valores que merezcan calificarse como tales y que sean garantes de la cohesión social. Aunque nos hemos modernizado bastante, aún persiste en la sociedad colombiana una violencia recurrente, prácticamente única en América Latina en cuanto a su intensidad y duración. Adicionalmente, constituimos una sociedad con un moderno sistema bancario, millones de tarjetas de crédito y dinero plástico; se ha extendido el uso social de los computadores personales, se accede a un consumo sofisticado por parte de vastas capas sociales que 236 hacen parte sustancial de nuestra cotidianidad laboral, al igual que ocurre en el universo de la vida privada. Esta paradoja es quizás uno de los rasgos más sorprendentes, puesto que muchos analistas suponían que bastaba con que el país se volviera más rico y moderno, con grandes ciudades, para que la violencia fuera un fenómeno del pasado. Como si éste fuera un problema de la ruralidad, de la tradición y no de la Modernidad. Entre tanto, lo que ha ocurrido es que hemos sofisticado la violencia, los mecanismos de muerte. Así, por ejemplo, en los últimos decenios, el principal factor de muerte lo constituyen las armas de fuego, de alcance corto y medio. ¿Qué ha pasado? ¿Por qué coexiste la violencia con la Modernidad? En mi opinión, ello ha ocurrido porque hemos construido una modernidad sin civilización en el sentido estricto de la palabra. No hemos logrado crear una red de relaciones sociales que limite, que claramente diferencie, el ámbito público respecto de lo privado. Sin embargo, numerosas sociedades contemporáneas, y aún vastas parcelas de la nuestra, presentan un fenómeno inverso: la predominancia de la disciplina y el autoritarismo en desmedro de la libertad natural de los individuos. Recuérdese, como lo hemos subrayado en capítulos anteriores, la tesis que en esta dirección formuló el lúcido filósofo francés Michel Foucault (1978), quien nos hablaba de la sociedad disciplinaria evidenciando como signos inequívocos de la Modernidad, el agobio totalizador de innúmeras instituciones de reclusión tales como la cárcel, la escuela, la clínica, los hospitales, los colegios, etc. En el caso colombiano, evidenciamos parcialmente rasgos de tal sociedad disciplinaria, pero al mismo tiempo “gozamos” de un elemento que nos diferencia de las sociedades de capitalismo avanzado, puesto que adoptamos un esquema patrimonialista de apropiación de los ámbitos y espacios de lo público. Precisamente afrontamos una crisis de la civilidad porque dichos valores —tales sistemas de respeto mutuo— de convivencia ciudadana que asumen diversas formas, están en quiebra. En el interés general los valores sociales colectivos no son reconocidos ni interiorizados por las personas, lo cual expresa el desacato a las normas generales como una suerte de recuperación de la guerra de todos contra todos de la cual hablaba Thomas Hobbes (1962). Es algo que se ve muy claro en nuestras sociedades dada la predominancia 237 del esquema de la sociedad competitiva, de la sociedad de mercado. En efecto, el capitalismo moderno ha hecho que no tengan función relevante, como en el pasado, los valores de aquello llamado bien común. Ni la religión ni la patria ni los partidos ni tampoco otros valores generan identidades fusionadoras de la sociedad. Anteriormente he subrayado la pertinencia en nuestro contexto de la reflexión política hobbesiana en razón de que la sociedad colombiana posee paralelos con la Inglaterra del siglo XVII, desgarrada por las guerras civiles, para la cual propuso Hobbes su modelo de Estado totalizador y omnímodo como instrumento artificial para superar la condición de guerra de todos contra todos, propia al predominio del derecho natural. Nosotros, como lo subrayaran diversos analistas —entre ellos el desaparecido líder liberal Luís Carlos Galán—, tenemos más territorio que Nación, e incluso más Nación que Estado, lo que en cierta medida resulta paradójico dadas las mencionadas correlaciones entre estas dos instancias. El liberalismo político ha tenido peculiares desenvolvimientos en nuestra historia. En realidad, primordialmente ha significado el desatamiento de los lazos que articulan las comunidades y los individuos a un orden institucional, a un proyecto dado de civilidad. La libertad, para la mayor parte de los colombianos, y ello visto en un balance retrospectivo, implica fundamentalmente la libertad del hacer y del querer, en donde los trazos de su ejercicio se articulan con el derecho natural. Recordemos que Hobbes había anotado que en dicha condición la libertad es simplemente el equivalente del grado de fuerza y la voluntad que ésta plasme. El proyecto inconcluso de la revolución democrática del siglo XIX coexistió con la esclavitud; se fundó adicionalmente sobre el poder señorial y oligárquico de la hacienda, y posteriormente se transmutó en un modelo dado de civilidad segmentada que permitió, por ejemplo, el desarrollo temprano de los denominados enclaves del terror (cfr. E. Varela B, 1997). Así, desde los espacios propios a la sociedad civilizada, hemos comprobado la recurrencia de este fenómeno. En tal sentido, se evidencian igualmente procesos de colonización que generalmente han implicado el alejamiento de grupos y comunidades enteras del ámbito del justicialismo público estatal. De otro lado, en las últimas décadas, dos manifestaciones de primordial 238 importancia que expresan la funcionalidad utilitaria de la violencia se han apoyado en dicha tradición. Me refiero, por supuesto, al narcotráfico en las vastas regiones de la Orinoquía, la Amazonía y algunos enclaves del terror en regiones del interior del país como la Costa Atlántica, el Magdalena Medio, Urabá, etc. Y, alternativamente, al despliegue del dominio móvil, volátil, si se quiere, pero arraigado en diversas áreas de la sociedad colombiana de las diversas facciones de la insurgencia extremo-izquierdista. Hoy, se reconoce que las relaciones entre narcotráfico y guerrilla han transitado desde la colusión temporal hasta los conflictos más radicales y abiertos que, entre otras cosas, han producido el despliegue del paramilitarismo de las denominadas Autodefensas. En realidad, la situación colombiana en los últimos años se acerca vertiginosamente a la condición de una guerra irregular civil, abierta y desenfrenada, semejante en muchos de sus rasgos a los peores momentos de la violencia del medio siglo. En los espacios sociales no relacionados específicamente con la violencia social y política ha operado un proceso de fragmentación, donde distintas fuerzas centrífugas se han sobrepuesto a las tendencias cohesionadoras que pretenden como proyecto político el fortalecimiento de la institucionalidad democrática, civil y republicana. Un ejemplo claro ha sido la recurrente abstención ciudadana pese a la modernización institucional y a la mayor depuración del proceso electoral25. Otra prueba palmaria la encontramos al comprobar que el grado de sindicalización de los trabajadores colombianos ha descendido vertiginosamente y probablemente alcance una cifra del 4 ó 5 %, uno de los más bajos del continente. Alternativamente, la representación gremial sigue siendo, desde el punto de vista empresarial, corporativa y concentrada en carteles y oligopolios. Pero quizás el fenómeno que de manera meridiana comprueba la grave fragmentación de nuestra sociedad es la profunda crisis de la representación política. La corrupción, el dolo, el fraude como herramientas y a la vez finalidades de la acción política, han terminado por deslegitimar los partidos, desvertebrados, fragmentados en señoríos electorales, en baronías unipersonales. Mientras, entre tanto la oposición democrática, llámese M-19, Unión Patriótica u otras fuerzas cívicas y regionales, impedida para consolidarse se desvanece. De tal modo que el espacio real, sórdido pero funcional de la oposición al stablishment, lo representan los carteles de las mafias y las cúpulas de las guerrillas ex239 tremo-izquierdistas. Otro rasgo de esta crisis de participación en las instancias de mediación del interés público, se expresa en la defensa corporativa de sus intereses por parte de muchos miembros de las altas esferas burguesas y empresariales. La costumbre política en Colombia ha permitido pactos entre los gremios y el gobierno de turno concretados en la sistemática y reiterada presencia de delegados del Estado (ministros, funcionarios de los departamentos centrales, etc.) en sus eventos y congresos, y la fácil intercambiabilidad de los cuadros de las clases dirigentes entre los ámbitos públicos y los privados (fundacionales). De otro lado, en los últimos años este fenómeno de alejamiento de las comunidades de los ámbitos de la política y la gestión pública se ha visto acentuado por el efecto corruptor del narcotráfico que se ha imbricado con formas antiguas, tradicionales, de privatización patrimonialista del interés público, sea bajo el discurso nacionalista de los otrora extraditables o simplemente por el ascenso social del propio capitalismo criminal, como se ve en la reciente crisis política que ha hecho tambalear al stablishment colombiano. Esta crisis del interés público manifiesta, en el fondo, la privatización de los espacios de lo público y se comprueba con la crítica acerba en los últimos años a las financiaciones provenientes de los grupos empresariales en las campañas electorales y el funcionamiento de los partidos. Por ello mismo, no se trata sólo de los fondos provenientes de sectores ilegales o criminales, sino del problema más vasto de lograr transparencia, gobernabilidad e imparcialidad en la gestión pública frente a los diversos y cotidianos conflictos de intereses que éste debe mediar. La corrupción derivada de estas y otras prácticas, el dolo como conducta normal por parte de muchos funcionarios públicos, su apatía y falta de compromiso frente al cumplimiento de sus deberes, son —de otro lado— algunas de las fuentes corrosivas de la credibilidad en las formas de gestión propias de altos niveles de socialización y civilidad. En síntesis, la crisis de nuestra civilidad resulta de la contraposición entre lo individual y lo colectivo estando imposibilitados tanto la sociedad civil como el aparato público del Estado para generar un consenso ético que 240 implique una alternativa viable a la denominada reforma moral e intelectual de la sociedad, como lúcidamente lo planteara, hace más de medio siglo, el pensador italiano Antonio Gramsci (1971) (E. Varela, 1995). Lo anterior sólo tendrá efectividad si se establecen claros rumbos para la funcionalidad complementaria de los actores públicos y privados bajo un eje central: la recusación del patrimonialismo clientelista, del autoritarismo social y estatal y la promoción, en su sentido lato, de las libertades públicas. Ello no implica, desde luego, ni la desaparición del Estado ni su debilitamiento en cuanto instancia suprema de arbitraje y control social. NOTAS En este texto focalizamos el vasto campo de las políticas públicas, dentro del ámbito del control social y del ejercicio de las relaciones de dominación que dimanan desde el Estado hacia la comunidad de vida, la sociedad civil. En particular, en este ensayo pretendo destacar la larga continuidad y los procesos adaptativos sufridos por dos semblanzas medulares de nuestra cultura política: el patrimonialismo y el clientelismo. 2 Este aspecto es destacado en el estudio clásico del pensador francés Bodino, Los seis libros de la srepública, Barcelona: Orbis, 1987. 3 En tal dirección, aún mantienen relevancia los conocidos análisis del profesor José María Ots-Capdequi: “el derecho propiamente indiano presentó pronto como rasgos más característicos los siguientes: 1. un casuismo acentuado y en consecuencia una gran profusión. No se intentaron, salvo en contadas ocasiones, amplias construcciones jurídicas (...) se legisló por el contrario, sobre cada caso concreto (...) 2. una tendencia asimiladora y uniformista. Se pretendió, desde la metrópoli, estructurar la vida jurídica de estos territorios con visión uniformadora y tratando de asimilarlos a las viejas concepciones peninsulares. No hay que exagerar, sin embargo, los 1 241 resultados, ni aún los propósitos de dichas políticas. 3. una gran minuciosidad reglamentarista. Los monarcas españoles quisieron tener en sus manos todos los hilos de gobierno de un mundo tan vasto, tan complejo, tan lejano?”(cfr. J. M. Ots-Capdequi, 1950: 124). 4 En el siglo XVI, se escenificó una lucha entre estos pioneros de la dominación sobre el mundo indígena y el propio Estado español. La iglesia católica y el Estado español muy pronto comprendieron que el mantenimiento de sus estructuras de dominio solamente podría darse en la medida en que lo privado estuviera subordinado a lo estatal y en que la propiedad de los territorios fuera simplemente una concesión temporal del rey. Ots-Capdequi, alude igualmente al carácter privado de la empresa de la conquista: “ La pugna entre los intereses privados de los conquistadores y el interés político de la Corona había de estallar, con violencia mayor o menor, tan pronto como el Estado español quisiera acentuar su presencia en estos territorios. Los reyes, ayudados por sus fiscales, trataron de reivindicar los atributos esenciales de la soberanía que, en buena parte, por vías de hecho o de derecho, habían caído en manos de los grandes descubridores y de sus descendientes (...) los monarcas españoles sólo pudieron alcanzar la victoria porque contaron con el apoyo de una doctrina y con los servicios eficaces de una burocracia que había logrado sólida madurez (...) América tuvo que ser reconquistada cuando apenas había sido descubierta; y fueron principalmente las huestes burocráticas, los oidores y los fiscales de las audiencias, los relatores y los escribanos, y los oficiales de la real hacienda, los verdaderos animadores de esta empresa reconquistadora más política que militar.” 5 También podemos, en ese sentido, recordar que la administración española se caracterizó por levantar censos y estadísticas prolijas, cual era el propósito de las visitas que han constituido una fuente invaluable de información historiográfica. 6 En especial, refiriéndonos al poder de los padres para limitar la decisión de los hijos de asumir libremente su conyugalidad, es particularmente claro en la pragmática de 1776 promulgada por los reyes españoles, la cual fue primero aplicada en España, y después extendida al Nuevo Mundo. Ella estatuía el control de tutela de los padres sobre los hijos para la celebración de esponsales y casamientos. Vale decir, la necesaria aquiescencia paterna a la aceptabilidad de los cónyuges. Los efectos que esta medida buscaba eran muy claros, y aparecían señalados en su propio cuerpo normativo. Para muchos críticos y analistas de la historia social colonial fueron parte de una deliberada restauración del orden y el control social. Resulta significativo cómo estas políticas implicaron un enfrentamiento del Estado con el poder eclesiástico. La Iglesia, a partir de las deliberaciones del Concilio de Trento, había difundido e impuesto el concepto del libre albedrío. De éste se desprendía que la aceptabilidad del matrimonio era decisión personal de los respectivos contrayentes. Se trataba, entonces, de un desafío al poder clerical y de un intento por sobornar el poder espiritual al poder temporal del Estado, a partir de la propia estructura del padrinazgo. Pero también constituía 242 un ejercicio del poder del Estado sobre las comunidades y los individuos. No tanto porque no fungiera la nueva política como un reforzamiento de la autoridad natural en las familias (el pater familiae), sino porque el ámbito de jurisdicción lo asumía ahora el Estado. 7 La pragmática precisa que su atención se concentre fundamentalmente en solidificar los efectos de prohibición de este tipo de matrimonios en el campo del contrato civil. En la reglamentación propiamente dicha se estipula en síntesis lo siguiente: “La limitación se aplica a los hijos menores de 25 años tanto para el contrato de esponsales como en el matrimonio propiamente dicho; estos deberán pedir y obtener consentimiento de sus padres”. Afirma, además, que todas las clases sociales del reino quedan cubiertas: “En todas ellas debe lugar la indispensable y natural delegación”. El argumento de fondo en este punto se plantea explícitamente. Se trata de corregir una práctica social que afecta como perjudicial “al público y a las familias”. Hay una clara defensa de la honorabilidad en virtud fundamentalmente de que estos conflictos, como lo han demostrado Ann Twinn y Susana Sokolov en los casos de México y Argentina, afectaban a las elites aristocráticas y a los estamentos superiores de la sociedad colonial. 8 A renglón seguido se señalan las mismas que, en síntesis, son las siguientes: “Que de la pragmática se exceptúen a los mulatos, negros, foyetes e individuos de castas y razas semejantes, reconocidos o reputados públicamente por tales, excepto a quienes sirvan de oficiales en la milicias o que, distinguiéndose de los demás por su reputación, buenas operaciones y servicios, deberán comprenderse en ella”. 9 Cfr. la tesis de Couture sobre la decadencia de la dote en las uniones matrimoniales, ya que muestra —con base en estudios de archivo— de qué forma va decayendo la costumbre de exigir la dote. Su hipótesis sugiere que una nueva elite en trance de autonomía se estaría conformando muy probablemente entre las burocracias criollas e hispánicas. Pero, hasta donde conozco el tema, no se evidencia hasta ahora una correlación entre esta decadencia de la dote y las normas de la pragmática. 10 Esta es una variante del Estado precapitalista que censa a sus pobladores y les establece tributos, individuales y/o colectivos, en las formas de trabajo en especie, en dinero, de tributos sobre la propiedad de la tierra o cuando la tierra está simplemente dada en capitulación o en concesión. 11 En América se establecieron “repúblicas” de índios y “repúblicas” de españoles. Cabe aclarar que los colonizadores usaron la palabra república en un sentido antiguo, como instauración de comunidades, asentamientos, congregaciones. 12 Los miembros de la Real Audiencia de Santafé evalúan, en dicho sentido, el uso del presidio urbano como correccional: “Para el castigo de los delitos leves es bien proporcionado y útil el presidio urbano porque en él al mismo tiempo que se corrige el exceso, se saca utilidad del trabajo de los presidiarios. Por ellos se puede proporcionar la limpieza de las calles, la composición de los caminos y paseos públicos. Por lo que conviene que 243 se fomente por todos los medios buscando arbitrios para su subsistencia” (Ibíd: 281). 13 Correspondencia del virrey a Vuestra Majestad los Reyes de España. Santafé, mayo 13 de 1796. En: Documentos para la historia de la vida pública del libertador, ed cit., tomo 1, pág. 282-283. 14 Bolívar afirmaba, en efecto, la “imposibilidad que hay para que un general sea a un mismo tiempo el magistrado y el creador de un nuevo gobierno”. Correo del Orinoco No.12, 10 de octubre de 1818. 15 Bolívar coloca como ejemplo clásico de dicho equilibrio entre los poderes de la monarquía inglesa: el poder ejecutivo británico está revestido de toda la autoridad soberana a que él pertenece, pero también está circunvalado de una línea de diques, barreras y estacadas. Es eje del gobierno, pero sus ministros y subalternos dependen más de las leyes que de su autoridad, porque son personalmente responsables, y ni aún las mismas órdenes de la autoridad real los eximen de esta responsabilidad real. 16 Así, criticando a los federalistas, Bolívar se extiende en demostrar por qué el modelo federativo funciona en la Unión Americana y cómo representaría grandes inconvenientes al intentar implantarse en América del Sur. Uno es el inglés americano y otro el español americano. En el caso de estos últimos dice Bolívar: “No somos europeos, no somos índios, sino una especie media entre los aborígenes y los españoles”. Los criollos le han arrebatado a los naturales sus títulos de posesión, pero a lo largo de la Colonia han tenido una existencia política nula, dado que estábamos excluidos del curso de los negocios públicos; abstraídos, ausentes del universo en cuanto era relativo a la ciencia del gobierno”. En: Correo del Orinoco, No 20, 27 de febrero 1819. 17 Actas del congreso de Angostura 1819-1820. Bogotá: Biblioteca Presidencia de la República, 1988. Roberto Cortázar y Luís Augusto Cuervo, compiladores. Segunda edición, página 175. 18 Ibíd., 174-181. 19 Por eso afirmaba en la citada memoria: “Persuadido el poder ejecutivo de que los sacrificios hechos por los granadinos en favor de su independencia y de sus libertades no han podido limitarse a sacudir la natural dominación de España conservando el sistema de educación e instrucción que el régimen colonial había establecido, está sosteniendo una lucha contra feroz fanatismo y las pasiones políticas bajo el nombre de la religión. No retrocederé una línea del puesto en que se ha colocado el gobierno, mientras que la ley, la cooperación del cuerpo legislativo y la sana opinión pública le presten su ayuda y apoyo”. 20 Los estudios sobre historia de la vida privada y la correspondiente relación con las instancias públicas son relativamente recientes en Colombia. El esfuerzo más destacable en tiempos recientes se encuentra en Historia de la vida cotidiana en Colombia (B. Castro, 1996). 21 Señalaba Vélez en su informe: “Dislocada la máquina política hasta en sus elementos primitivos, él se hallaba entregado a la merced de los movimientos populares y de las invasiones de la fuerza armada, sin medios 244 de reprimir los agentes destructores que amenazaban al Estado con una subversión inevitable. Necesario es decirlo, la Nación estaba dividida en dos grandes bandos o partidos. Componíase el uno de la parte sana y de los hombres ilustrados del país, entre quienes se distinguía una juventud ardiente, impetuosa y llena de patriotismo y de entusiasmo, y de la masa general del pueblo, que exigían una administración republicana, firme, estable y que asegurase a todos el tranquilo goce de sus derechos. El otro era compuesto de la mayor parte de los jefes y oficiales del ejército, que con algunos pocos ciudadanos incautos, ilusos o interesados, estaban prontos a luchar por la continuación del gobierno militar que, desde 1827, había regido la República sin sujeción a las leyes ni a pacto alguno escrito”. En: Administraciones de Santander, Vol. III, 12-13. 22 Alejandro Vélez: “Exposición que el ministro secretario del Estado en el despacho del interior y justicia del gobierno de la República, presenta a la convención granadina de 1831 sobre los negocios de su departamento”. En: Administraciones de Santander 1831-1833, Bogotá: Presidencia de la República, 1990 vol. III, 12 y 13. 23 En un plano más general, Germán Colmenares, en Las convenciones contra la cultura, encabezando el nuevo revisionismo historiográfico contra las “ historias patrias” (por ejemplo, Henao y Arrubla, J. M. Restrepo, etc.) abrió de forma renovada la discusión sobre la verdadera significación de la prédica liberal en el siglo XIX como discurso de legitimación del nuevo poder en manos de las oligarquías criollas. En este orden de ideas, un aspecto que resulta interesante en esta cotejación fue la inconsistencia del predicamento ciudadano al ser mantenida la esclavitud durante casi tres decenios; remanente sólo resuelto a través de una guerra civil, la de 1851. 24 Sin embargo, el siglo XIX evidenció una ruralización de la vida ciudadana en el nuevo continente. Las antiguas ciudades centralizadas construidas por los españoles bajo preceptos establecidos por la Corona cederían su lugar, y el mundo urbano se plebeyizó, perdió la estructura ordenada que caracterizó el mundo colonial, y, al mismo tiempo, en la medida en que los poderes políticos y sociales se ampliaron, este viejo universo cerrado tuvo una menor capacidad de dominio sobre la vida rural. Mientras la ciudad colonial era un centro de dominio, la ciudad republicana del siglo XIX pierde ese carácter, y el mundo rural adquiere preeminencia y autonomía dándose una fragmentación del poder político; aparecen los caudillos locales y regionales y la cuidad es una especie de punto de encuentro en donde los líderes, las elites, hacen negocios, conciertan, pero la disputa por el dominio se ejerce fundamentalmente en la ruralidad. 25 Incluso hoy, después de la reforma política de 1991, el fenómeno de apatía hacia la política –este declive del “hombre público” en el plano de la contienda por el poder público– persiste con rasgos significativos. Si tomamos en cuenta los cambios otorgados por el nuevo orden institucional para hacer más transparente y abierta la contienda en el terreno de la democracia representativa, evidenciamos cómo los ciudadanos mantienen su alejamien245 to de los asuntos públicos. En el pasado, bajo la premisa de los auxilios parlamentarios, se sabía claramente que la retribución se encontraba en la repartición del festín en fundaciones anónimas y piratas, en el manejo de los contratos de obras públicas, los puestos y prebendas, etc. Sin embargo, la participación, en términos relativos, era mayor en el pasado. Es cierto que hay un electorado independiente que vota por consideraciones no clientelistas, movido por el carisma de los candidatos, por sus ideas, por la representación emblemática de sus intereses, etc. Pero este electorado, llamado por algunos politólogos “franja”, no es un fenómeno nuevo, imputable a la institucionalidad derivada de la Constitución de 1991. Como se recordará, esta “franja” intervino decisivamente en las elecciones presidenciales de 1970, 1984 y 1990. Y fue igualmente la base de apoyo que permitió la consolidación del liderazgo de Luís Carlos Galán en los años 80. Este apoliticismo y el odio hacia los políticos profesionales o de “carrera”, se expresa en el desprestigio del parlamento y de los partidos como instituciones confiables y serias, sin que se descuente en razón a este severo juicio la utilización de sus beneficios potenciales. 246 Capítulo 5 MODERNIZACIÓN INSTITUCIONAL Y NUEVOS ROLES PÚBLICOS Y PRIVADOS El debate en el contexto de la globalización respecto de las nuevas funciones de los actores públicos y privados, dentro del proceso de modernización de las organizaciones público-estatales, ha sido en nuestro medio un ejercicio académico y político apasionado. Ciertamente agudo, pero con un fondo teórico y analítico superficial, carente de una visión de conjunto y de una adecuada fundamentación conceptual. Una relectura de las discusiones recurrentes sobre la privatización de las empresas del Estado demuestra una confusión conceptual acerca de las nociones de lo público y lo privado (R. Moe, 1999). Desde luego, este no es un problema meramente epistemológico. Existen fuertes intereses particularistas y patrimonialistas que, de manera sesgada, tienden a obnubilar el asunto para simplificar demagógicamente las opciones en juego. De tal modo que, con base en la aclaración necesaria de estos conceptos en un terreno sociopolítico, creo que podremos establecer un adecuado marco analítico de la reforma del Estado tanto en América ´latina como en sus generalidades globales. 1. Los términos del debate No pretendo, desde luego, hacer de las anteriores temáticas una exposición exhaustiva. Quiero sí, situar un conjunto de referentes para un eficaz proceso de conceptualización de estas complejas realidades, en particular, 247 con el fin de elaborar líneas de seguimiento para detectar las nuevas fronteras entre lo público y lo privado, el surgimiento de nuevos actores sociales, el carácter recesivo de otros, realidades institucionales y conflictuales que marcan nuestra época. Estas categorías —lo público y lo privado— aunque corresponden a una larga tradición de la teoría política y gestionaria, han sido reformuladas, en cierta medida de manera inconsciente, en el contexto de la reforma del modelo prevaleciente de Nación-Estado en América Latina. Y no se trata de un asunto de poca monta. En realidad, este es el problema capital para definir políticas públicas que permitan un desarrollo sostenible de la región basado en principios de eficiencia, transparencia, equidad y favorabilidad (E. Albi, 2000). En particular, visto dicho proceso en dos dimensiones. Primero, como una rectificación del esquema intervencionista y paternalista asumido por las oligarquías republicanas durante la mayor parte del siglo pasado. Resultado del predominio de este intervencionismo fue la hipertrofia burocrática del aparato estatal, incidiendo en la pérdida de la autonomía de lo local y lo regional, avasallados por el presidencialismo y el autoritarismo centralista, confirmando las tendencias a eliminar y oprimir lo privado. En segundo lugar, para determinar el tipo de políticas y las fuerzas sociales que las encarnarían, capaces de salir del dilema estrecho entre estatismo o neoliberalismo. Además, afirmo la cuestión de las fuerzas sociales dado que en años recientes las rectificaciones de políticas que se han concretado, a través de las evaluaciones electorales, están por doquier en América Latina (Brasil, Argentina, Venezuela, Ecuador, Uruguay, etc.). Salvo en Colombia; aunque en nuestro país también la elección de alcaldes y gobernadores alternativos marca la misma dirección. Sin un adecuado diagnóstico, no hay garantía alguna de que en ciertos años el péndulo sea retomado por las corrientes neoliberales. Lo importante, en este sentido, es la capacidad de construcción de consensos de largo plazo sobre las políticas públicas y el papel complementario de lo público y lo privado. La cuestión no es sólo ni principalmente teórica, aunque estoy convencido de que la claridad en las políticas resulta de forma parcial de nuestra capacidad de aprendizaje social y de la evaluación de los caminos y experiencias recorridos. Además, está de por medio la llamada inteligencia interesada. Recuérdese que en las sociedades capitalistas de América Latina, cuando fracciones o grupos de intereses correspondientes a la sociedad civil cuestionan al Estado, éste con frecuencia asume dichos desafíos 248 como una amenaza a su régimen político y procede en consecuencia instaurando el terror del Estado1. Un balance de estas décadas nos muestra que a la par con la violencia directa, igualmente se disminuyeron los derechos individuales básicos y el valor real, para la retórica asistencialista, de los derechos sociales en virtud de la agobiante marginalidad y exclusión que dichos modelos de Estado ha implicado para quienes poco importan en términos de una pragmática de gobernabilidad. Aunque inicialmente el modelo de desarrollo de las organizaciones empresariales resultó afectado por el excesivo reglamentarismo y proteccionismo implícitos en dicha política pública, lo que a la larga significó un retroceso de la región latinoamericana —Colombia incluida— respecto al proceso de apertura y globalización en curso desde hace algunos lustros. Posteriormente, el aperturismo no ha traído beneficios tangibles para las mayorías; ni siquiera para los sectores de la llamada economía real. Como se sabe, se percibe generalmente dicho debate en torno a dos posturas opuestas. En primer término, la de quienes impulsan el desmonte, tan amplio como sea posible, del antiguo Estado de “bienestar” o patrimonialista, identificados en Latinoamérica como neoliberales. En segundo, la de los grupos y sectores interesados en el mantenimiento e incluso en el reforzamiento de las funciones tradicionales del Estado latinoamericano, particularmente de los roles que a éste le asignó el modelo desarrollista imperante en la mayoría de nuestros países después de la última postguerra. En lo que concierne al neoliberalismo, éste ha sido percibido frecuentemente en términos esquemáticos como si fuera exclusivamente resultado de una imposición de los organismos multilaterales de crédito (Banco Mundial, F.M.I.) (cfr. World Development Report, World Bank, 1991, Section VI), entidades que ciertamente representaran la doble moral del mundo desarrollado que decreta la apertura de los mercados del tercer mundo al tiempo que se practica un feroz proteccionismo. Alternativamente, tales fuerzas liberalizantes se apoyaron en los excesos del anterior modelo intervencionista. Su verdadera semántica se escudó en el alegato a favor del desmonte del asistencialismo y de los privilegios estamentarios que caracterizaron al “Estado de bienestar” promovido por las oligarquías republicanas en Latinoamérica. 249 Todo ello basado en un principio central que toma en cuenta un esquema de acción de los actores públicos y privados con base en este principio: la necesidad de un ambiente de competencia que tendría su complemento en la liberalización política para, de este modo, sacudir nuestras sociedades del paternalismo y el patrimonialismo2. Lo que conlleva una evaluación del rol del Estado frente a los procesos de apertura, privatización y reconversión, sin que ello implique per se el crecimiento o la disminución de las organizaciones público-estatales. De todas formas, estas distinciones, si bien funcionales en términos de captación del debate para el amplio público, resultan excesivamente esquemáticas. En efecto, los primeros, es decir los “neoliberales”, no llegan hasta el punto que correspondería a un postulado anarquista, de pretender la desaparición total de las funciones del Estado para que, en su lugar, opere una suerte de autorregulación social, terreno soñado por las diversas utopías libertarias. Por el contrario, incluso en sus versiones radicales, los enemigos del rol intervencionista del Estado, reconociendo implícita o explícitamente el postulado durkheimiano acerca de la inevitabilidad del control social, de la coerción y del papel compulsatorio de las instituciones públicas y privadas (como producto de la división del trabajo y de la jerarquización social en tanto elementos cohesionadores y vinculantes de la vida en sociedad), expresan la tesis de reducir el Estado hasta un límite: el denominado “Estado mínimo”. Pero en modo alguno llegan al nivel de la utopía anarquista: la desaparición del Estado político (R. Nozick, 1974). La discusión sobre el mínimo tamaño del Estado posee, en ese sentido, dos facetas. En primer lugar, el espacio de la disquisición teórica donde el modelo radical de Nozick (cfr. Nozick:1974) y otros liberales libertarios expresa que el Estado, no fundándose en contrato social alguno, sólo puede hacer legítimamente aquello autorizado expresamente por los ciudadanos. Ese límite mínimo es simplemente el de la garantía de las propiedades de los particulares y sus libertades básicas, incluyendo el respeto a la transparencia de los contratos, donde encuentra Nozick la única fundamentación válida de cualquier reclamo justicialista en la labor del gobierno3. En un plano diferente al filosófico, la teoría económica liberal aperturista v. gr. Arrow, Walras y otros, en su definición de los ámbitos de la competencia y 250 el libre mercado, afirma un lugar necesario para la acción estatal, por ejemplo, en el terreno de la subsidariedad (G. Tullock, 1974). Esto implica discutir y evaluar en términos concretos qué es una funcionalidad “mínima” del Estado. En la mayoría de las visiones liberales, el problema se expresa en términos de asignarle al Estado las tareas que la sociedad civil, bajo el móvil de la ganancia privada, no asumiría y dejaría de lado. En este sentido, el liberalismo aperturista actual no se opone al esquema del Estado Social de Derecho, que aparece consignado en la promulgada constitución política colombiana de 1991. Quienes, a su turno, propugnan por la continuidad y modernización del Estado intervencionista no dejan de reconocer las graves deficiencias que este tipo de modelo ha representado, desde el punto de vista de su implementación en términos de equidad, justicia e igualdad —para no hablar de la corrupción, inoperancia, tramitomanía, burocratización, etc., rasgos premodernos de la gestión pública inherentes a dicho modelo—. En realidad, muy pocos podrían negar que dicho Estado ha servido de instrumento para reforzar las desigualdades sociales, la inequidad y el usufructo del poder público por parte de las oligarquías estamentarias y republicanas. En fin, esta discusión se hace más urgente si sopesamos la necesidad de establecer una correlación entre la liberalización política —al parecer una situación consensual, donde los contradictores manifiestan confluencia a tono con los vientos que hoy soplan a favor de la democracia participativa— y la liberalización económica y social, donde surge inmediatamente el disenso. 2. Modernidad, racionalismo y mercantilización de la vida social En conexión a lo ya dicho, puede establecerse una articulación entre Modernidad y racionalidad, donde la última opera como un principio de causalidad, pretendiendo estatuir un discurso de tipo filosófico y científico que reemplazaría la fe, la intuición, el escepticismo. Camino que establece una relación entre sujeto y objeto, de tal modo que el sujeto racional y autónomo puede conocer el objeto (el mundo) siempre y cuando acuda a un método, al método científico, con base en el cual se da la articulación sobredeterminante entre los principios generales de una ciencia y su aplicación a contextos empíricos. 251 La Modernidad se pone en una suerte de pragmatismo con respecto al horizonte de lo que es conocible y articula funcionalmente la racionalidad de la ciencia con su utilización social. De este modo, se produce una ruptura radical respecto a la tradición ilustrada filosófica, uno de los paradigmas más claros de la Modernidad, no tanto en su versión inicial de los siglos XVI y XVII, sino en su acabamiento dentro de la sociedad industrial de los siglos XVII y XVIII y, en especial, del XIX. Estos son elementos que retomarán tanto Rousseau como Kant y Hegel, y que serán, en fin, patrimonio común de la Ilustración y el Romanticismo. Y que incluso dominan la visión de intelectuales como Comte, que funda las ciencias positivas de lo social a través de una argumentación recusatoria de cualquier argumento trascendental o especulativo abstracto, urgiendo en su lugar la aplicación del método que lograba éxitos portentosos en el campo de las ciencias naturales y las matemáticas. Esta ciencia positiva entraría a reemplazar las anteriores explicaciones de la religión y la metafísica, permitiendo una explicación racional de lo que antes era de carácter religioso4. Una segunda característica de la Modernidad, dejando de lado el campo filosófico y entrando en el terreno sociológico, fue el surgimiento de las Naciones-Estado. Coronamiento del decurso de una tendencia de largo aliento, cual fue el surgimiento de dichas formas políticas en los principales países europeos: España, Portugal, Francia, Inglaterra, los Países Bajos, etc. Además, un fenómeno de carácter histórico que se anudó con este proceso y que resultó de suma importancia fue el descubrimiento de América y la colonización de África. De tal modo que la Modernidad inauguró una época que, entre los siglos XV y XVI, significó para los europeos su colocación como cultura dominante respecto a esferas nuevas, a mundos vastos y desconocidos cuales eran los mundos de la colonización y las empresas de conquista (M. Kaplan, 1969). Se estableció una relación de apertura de la mirada del mundo distinta a la de la sociedad cerrada del medioevo. El surgimiento de la nacionalidad en dicho contexto pudo darse a partir de portentosos cambios en la cultura. Como lo subraya el profesor James Kurth: “Antes de la Era Moderna había tribus, grupos étnicos, pueblos y aún nativos. Sin embargo, todos ellos se convirtieron en naciones con el arribo de la alfabetización, su difusión entre las diferentes clases con un idioma común y sobre todo con una literatura com252 partida. El proceso se inició en el siglo XV con la invención de la imprenta, su propagación inicial estuvo a cargo de la reforma protestante y más tarde la Ilustración; y culminó con la normalización de las escuelas y la educación para las masas en el siglo XIX y a principios del siglo XX” (J. Kurth, 1992). Este rasgo es un apunte conceptual que muy pocas veces se subraya con la suficiente importancia, quizás exceptuando el trabajo de Marshall McLuhan, La Galaxia de Guttemberg (1986), donde se explora el impacto enorme de la imprenta y el cambio paradigmático que significó la alfabetización y el surgimiento de las lenguas nacionales con escritura para el desarrollo de la Nación-Estado5. En efecto, en todos los países europeos surgen simultáneamente lenguas escritas, e incluso los grandes clásicos de la literatura son todos contemporáneos entre sí, pertenecen a una época específica de la literatura y de las artes: el Barroco. Con base en tales cambios, se dio inicio a la moderna cultura de masas, que empezó a romper el divorcio entre una elite culta, señorial y aristocrática, y una plebe que carecía de todo tipo de cultura alfabética y que incluso no tenía un consenso mínimo en su cultura lingüística. Además de lo anterior, constituyó un rasgo central de la Modernidad la configuración de la Nación-Estado en términos de una vasta institucionalidad pública, mediante la conformación de una burocracia relativamente sólida y compleja que incluyó, desde luego, al ejército6 y que se colocaría, de forma impersonal, sobredeterminando la naciente sociedad civil. Resurgió esta burocracia de Estado principalmente en el poder ejecutivo (bajo la distinción de oficiales y comisarios o ministros), en la hacienda pública, la administración de justicia y el parlamento (fuente de la productividad legal-normativa)7. Generalmente, tales burocracias estuvieron bajo una égida común ya que el Estado se unificó desde arriba mediante un proceso de centralización que concretó el Estado absoluto. Otro rasgo importante para el tipo de Estado propio a la Modernidad fue su progresiva separación respecto de la Iglesia. La Reforma y la Contrarreforma —ideologizada por los sacerdotes e ideólogos españoles jesuitas— significaron una separación entre Estado e Iglesia. En el caso protestante y anglicano, el proceso fue tal vez mucho más radical que en nuestro contexto hispano253 portugués. Pero incluso en la larga alianza entre la corona española y el papado, el Imperio logró prerrogativas: el nombramiento de los obispos, el pago de los sacerdotes, el financiamiento mismo de la Iglesia por parte del Estado, y un cierto poder de veto estatal sobre las autoridades eclesiásticas, todo ello bajo la figura tipificada del patronato. De otro lado, uno de los puntos básicos de la Modernidad ha sido su creencia en el progreso, con la subyacente imagen de una temporalidad lineal y ascendente. En efecto, si se revisan autores representativos de las más disímiles corrientes de pensamiento, v gr., aquellos que en su momento fueron irreconciliables rivales y presentaban opiniones totalmente antagónicas en los siglos XVIII y XIX —con la distancia que nos da el tiempo— se pueden comprobar las enormes coincidencias existentes entre estos. Véase el modelo del progreso plasmado por Vico, Kant y Comte y recuperado por las ideologías más disímiles del siglo XIX a la manera de un consenso de época. Hay, de este modo, una creencia relativamente uniforme en un progreso que probablemente tiene retrocesos, ciclos; pero que siempre marcha de forma ascendente. Este discurso lo elaboraron prácticamente todos los hombres de pensamiento del siglo XIX. Véase el caso de Marx, para quien el proletariado era el sector que encarnaba de mejor manera el progreso, la Modernidad; y, por lo tanto, habría de ser el beneficiario de esa marcha hacia lo mejor8. Por su parte, los nacionalistas conservadores consideraron depositaria del progreso a la Nación, encarnada en el monarca, o en un proyecto expansionista. En el siglo XIX, después de la Revolución Francesa y la gran transformación industrial, se interpretó al progreso sólo dentro del marco de despliegue de la Nación-Estado9. Desde el punto de vista económico fue distintivo de la Modernidad su contexto mercantil capitalista. En este sentido, cabe subrayar que, en general las teorías liberales nunca se han opuesto a ciertas restricciones en las libertades mercantiles. Es decir, ningún doctrinario de este corte ha postulado la inexistencia de un límite para el ejercicio del derecho a la propiedad, el papel del mercado o la competencia. Por el contrario, se han hecho frecuentes salvedades respecto a que la propiedad pueda ser limitada en el caso de los monopolios o por el uso indebido del derecho propio; por ejemplo, a través de la especulación o mediante prácticas que se consideran desleales como quebrar a la competencia vendiendo 254 por debajo de los costos de la producción (dumping)10. En todo caso, el Estado a través de los sistemas jurisdiccionales ha intervenido la competencia durante todo el largo periplo del predominio de las escuelas y prácticas mercantilistas. De tal modo, no puede pensarse que la sociedad y los instrumentos de arbitraje existentes dejen hacer y dejen pasar en forma absoluta. Los marcos de la Nación incorporan reglas de concurrencia y un papel de arbitraje del Estado que tutela el desenvolvimiento económico. Solamente en los manuales de economía que utilizan modelos puramente teóricos aparece la estructura del dejar hacer y del dejar pasar bajo la forma de la llamada competencia perfecta, donde se supone que el Estado no interviene, los precios no son alterados desde fuera; se establecen unos marcos ideales, en los cuales vendedores y compradores, oferentes y demandantes, actúan racional y libremente, por lo que a través de ellos se logra un equilibrio (A. Buchanan, 1985). Es cierto que Adam Smith habla de la mano invisible del mercado, una de esas frases como la de Maquiavelo: “el fin justifica los medios” que se ha retenido como la esencia de su conceptualización. Esta idea está muy en la tónica de la concepción de Kant a la que aludimos en otro lugar de estos ensayos, pero en Smith actúa en un plano económico. Recuérdese que la obra de Kant es contemporánea a la de Adam Smith, quien publica su libro en 1776. Kant y Hegel fueron atentos lectores de los economistas ingleses; su concepción filosófica no estuvo desvinculada de un conocimiento general de los debates económicos en las escuelas y tendencias que la economía política estaba cimentando en el siglo de la Ilustración y en los comienzos del siglo XIX. Bajo este “espíritu de época”, entenderemos la afirmación de Smith de que la sociedad se estructura sobre la base del mercado y la competencia. Smith parte de la división del trabajo, con lo cual asume un doble problema: ¿cómo operan los procesos de producción y cómo se distribuyen los resultados del trabajo acumulado?11 El primer punto sobre el que insiste Smith es que existe dentro de los seres humanos (organizados en comunidad), una tendencia natural a dividir y especializar las funciones productivas: “los mayores adelantamientos en la facultades o principios productivos del trabajo, pericia, acierto con que éste se aplica y dirige a la sociedad no parecen efectos de otra causa que la 255 división del trabajo mismo. (...) esta división del trabajo, que tantas ventajas le trae a la sociedad, no es en su origen efecto de una premeditación humana que prevé y se proponga como fin intencional, aquella opulencia que la división ocasiona. Es una consecuencia necesaria que lenta y gradual de cierta propensión genial (en el sentido de ingenio, de inclinación del hombre) que tienen por objeto una utilidad menos extensiva... la propensión humana es negociar, cambiar o permutar una cosa por otra” (Smith). Las limitaciones al mercado deben permitir una mayor eficiencia del desarrollo de la producción y de la propia división del trabajo. De allí que, por regla general, el Estado no intervenga frente a la división del trabajo ni a la estructuración de la concurrencia mercantil. No se trata, desde luego, de un capitalismo mundial que carezca de fronteras nacionales, sino que se piensa allí a cada orden económico desde una perspectiva nacional. El proteccionismo era en ese entonces una política económica discutida por los librecambistas, en tanto afirmaban que solamente sobre la base de eliminar las restricciones al comercio podría establecerse una prosperidad general de las naciones. Smith plantea que la ruptura de esa especie de choque insoluble entre el libre cambio y proteccionismo estriba en su teoría de las ventajas comparativas de la división del trabajo. Como se ve, este argumento es simplemente una aplicación a un marco diverso, nacional o internacional, del mismo principio: así como cada individuo se beneficia con el mayor número de productores y con la especialización, de la misma manera, lo hace como productor de un mercadería o de un fragmento de la misma pudiendo obtener en el mercado el resto de bienes que demande. La teoría sobre la funcionalidad del mercado y su carácter libre, abierto, es una de las tesis mayores que arranca con A. Smith (K. Galbraith, 1989: Cap. 6), aunque en el propio Smith con mayores matices y consideraciones contextuales que en la lectura simplificada que hicieron las partes interesadas. Así caracteriza irónicamente Galbraith a los empresarios que se enorgullecen de su egoísmo al ser la base misma del interés general como en la corriente desde entonces dominante de la economía como disciplina. La acumulación de las riquezas en el caso de Smith se basa en el factor productivo por excelencia: el trabajo. Pues éste constituye la mercancía por excelencia, con lo cual Smith hizo una significativa contribución teórica. La corriente dominante en la economía defenderá la naturalidad y legitimidad del atesoramiento (la que denomina el profesor O. Aktouf (2002) “crematística” —retomando un concepto aristotélico—); criticará el proteccionismo y limitará el papel del Estado a su condición de instrumento garante de los fines privados. Pero aún Smith sabía que el Estado cumple funciones que van más allá de la mera garantía de la existencia duradera de los mercados. Recuérdese que él era filósofo de formación y sus preocupaciones por las motivaciones de la acción humana lo condujeron a la economía, un poco, aunque en un sentido ideológico diferente, a la manera que ocurriría con Marx en el siglo XIX. De acuerdo al punto anterior, la función estatal sería subsidiaria. Es decir, el Estado comenzaría su funcionalidad social allí donde se agota la eficiencia del mercado. Esto tiene dos componentes. El primero, el carácter de arbitraje o mediación del Estado, necesario para que el mercado funcione. Así, las normas jurídicas y administrativas de tipo público crean un marco institucional que evita el robo, el fraude, garantiza el libre acceso al mercado, promueve la competencia, etc. El segundo supuesto de la teoría económica dominante es más estrictamente el de la señalada subsidiaridad. Desde la teoría de los bienes públicos, se justifica una intervención económica del Estado sólo para proveer cierto tipo de bienes cuya nula o poca rentabilidad impiden que los empresarios los provean en un esquema de mercado abierto. Esta es la base teórica de la actual tendencia a la mercantilización de los bienes públicos (J. Stiglitz, 2002b). Se parte del principio de que siempre que sea posible debe permitirse al mercado proveer aquellos bienes públicos que puedan ser pagados por la esfera de los consumidores, respetando los principios de acumulación de capital o de obtención de tasas medias de beneficio. Obviamente, el principio que defiende Smith nunca ha tenido plena o total aplicación y las sociedades de Nación-Estado han limitado la competencia internacional. Siempre han existido limitaciones a la competencia. Así que lo planteado como modelo está lejos de ser igual en la práctica. Smith, continuando su argumentación, señala límites a la división del trabajo según la extensión del mercado público. Vale la pena resaltar esta acepción: el mercado público, aunque Smith no hace una disquisición teórica sobre lo que significado. En la medida que el capitalismo esta diseñado para vender más, y correr así las fronteras del mercado homogeneizando tanto la producción como el consumo, no se trata sólo de que siempre se divida el trabajo pues al mismo tiempo que hay más suboficios, hay una homogeneización. Unos ejemplos son los médicos, que tienen tantas especializaciones; pero la especialización no implica que el nivel de homogeneidad se pierda12. Es decir que la especialización no es un vector que surge de un sólo punto común, sino que se produce en líneas paralelas. El mercado no está siempre dividido, ya que hay tendencias a su homogeneización. Después de Adam Smith, en la teoría económica, se ha planteado una discusión sobre los mercados segmentados; se subraya como no hay un solo mercado sino que, en el fondo, existen una serie de mercados, de acuerdo con reglas especificas del juego de oferta y demanda, y la vez sobre la base de las limitaciones que establecen las legislaciones nacionales, las relaciones sociales, las diferencias regionales, etc. Puede hablarse hipotéticamente de un mercado público que se segmenta. Sin embargo, tendría que discutirse en qué medida ese mercado público se pudiera segmentar hasta llegar a niveles de privatización del mercado o a una privatización de la competencia. Este es el caso de los “carteles” (oligopolios), muy discutido por la teoría económica, los que ya pueden ser vistos como espacios organizacionales de privatización del mercado público en tanto limitan la libre oferta y la demanda, ya sea en la forma de acuerdos de precios, prefijando estructuras de producción o por medio de la creciente intervención del Estado. En Colombia, ha sido usual en las últimas décadas, en el mercado una economía dirigista y reglamentada desde arriba, que el Estado defina multitud de variables como los precios de los transportes urbanos, municipal, intermunicipal; se tiene igualmente un control de tutela sobre los precios de los productos de primera necesidad y existen numerosos mecanismos de carácter coactivo para incidir los precios, las tasas de interés a través del aumento de la captación en los niveles de encaje o incluso herramientas y facultades legales para intervenir directamente el mercado. De este modo, sobre el mercado, en teoría libre, ámbito de la interacción 258 pública entre demandantes y oferentes, existen una serie de factores que perturban esta libre competencia no sólo del orden internacional (proteccionismo), sino que corresponden a reglas del juego y equilibrios políticos y sociales. No hay que olvidar, además, que uno de los factores más importantes del mercado es la propia fuerza de trabajo y que ella también se segmenta en mercados diferenciados: trabajadores manuales, intelectuales; mujeres, el trabajo del campo y la ciudad, etc. Y recordemos cómo sobre este suelo de divisiones socio técnicas del trabajo surge luego un proceso dialéctico, contradictorio, entre la especialización y la homogeneización del trabajo. Éste, en ciertos ciclos, tiende a homogeneizarse de acuerdo a estándares logrados desde el punto de vista de la tecnoproducción; pero, al mismo tiempo, en la medida en que la sociedad se vuelve más compleja, tiende de nuevo a dividirse (E. De la Garza, 2000). Es decir, con base en estos procesos contradictorios, se plasmó el predominio de un nuevo orden basado en la primacía de una sociedad posesiva y abierta que estructuró un discurso de tipo institucional y jurídico basado en el reconocimiento de los derechos individuales. Surgió la teoría moderna de los derechos ciudadanos o civiles, que básicamente fue una teoría para definir las relaciones entre el Estado y los propietarios privados. Por ello, el derecho fundamental —dentro del orden capitalista— ha sido el derecho a la propiedad, al cual se le añadió el derecho a la libertad de expresión, configurando los pilares fundamentales del derecho moderno. Estos son dos pivotes fundamentales del derecho moderno, que sirven para expresar la autonomía de las personas al perseguir sus propios fines privados sin la intervención de los actores públicos más que allí donde se enfrenten conflictivamente los derechos de estos sujetos. En tal dirección, se reconoce explícitamente que la gente tenga la libertad de expresión social sin que sea censurada por razones religiosas o políticas, a no ser que ataque los fundamentos mismos del contrato social (R. Dworkin, 1995). Esta fue la gran batalla que le correspondió a la Ilustración en Europa, y cuyo eco, en cierta medida, también influyó en España y en la América colonial. Ello significó la consolidación de los principios fundamentales del derecho civil moderno, la separación entre lo privado y lo público, la creciente diferenciación entre Estado y sociedad civil (entre política y economía), 259 otorgándole al gobierno fundamentalmente funciones de arbitraje de los conflictos propios a la sociedad de mercado. La democracia, tal como la conocemos y la practicamos en el siglo XX, es un fenómeno supremamente reciente. Ningún país del mundo en el siglo XIX tenía sufragio universal, pues éste en todos los países era censatario, basado en criterios tales como la propiedad privada, la renta, o un nivel dado de instrucción cultural, prerrequisitos para ingresar a los listados de elegibles y electores (R. Dahl, 1992). Una de las mitificaciones que se tiene con respecto a la Modernidad es que ésta implica necesariamente la democracia y el ejercicio de todo un programa político que solamente se dio en el coronamiento del proceso. Sin embargo, antes que de democracia, se habló de soberanía popular, lo que correspondió al influjo dominante de las teorías contractuales. En efecto, la burguesía, el empresariado y las elites políticas modernas que se enfrentaron al poder de la Iglesia y de los señoríos utilizaron, como uno de sus argumentos esenciales de legitimación, el que todo poder político era fruto de un convenio; y ese convenio residía en la capacidad potestativa del pueblo de generar poder. Por lo tanto, la soberanía estaba en el pueblo o en la Nación, que significaba para ellos lo mismo: Nación como conjunto o suma de habitantes de un territorio. En tal perspectiva, se supuso que todo gobierno respondía, en últimas, al constituyente primario. En ese sentido, se puede hablar de democracia; pero esta democracia así vista fue siempre directa. Por ejemplo, la que un dictador usaba al convocar plebiscitos para legitimarse a sí mismo, fenómeno muy frecuente en la historia contemporánea. O, alternativamente, democracia en la que el pueblo mediante un alzamiento, a través de la violencia revolucionaria, legitimaba un nuevo orden. De tal modo que la democracia estuvo pensada, durante los siglos XVII, XVIII y XIX, en términos de la legitimación de hecho, fáctica, de un ejercicio dado del poder político. Es decir, lograr tener el respaldo directo de la población antes que de una apelación consciente mediada por los partidos, por instituciones como la prensa, los medios masivos, la discusión pública de una programática de gobierno, elecciones, etc. En síntesis: la Modernidad, fenómeno múltiple, cultural, político, sociológico, económico e ideológico se originó en Europa y se fue explayando hacia el resto del mundo. En Estados Unidos con su proceso de independencia en el siglo XVII; 260 en América Latina con la revolución criolla durante la primera mitad del XIX, y de manera más tardía esta Modernidad llegó bajo la forma del cosmopolitismo y el colonialismo al resto del mundo. 3. La crisis de la modernidad El reconocimiento de que existe una crisis de los valores en torno de los cuales se ha constituido nuestra época, puede cotejarse confrontado uno de los grandes mitos de la Edad Moderna. El espíritu que entra en crisis en el siglo XX, como uno de los primeros elementos de la Modernidad, es la creencia en el progreso. Esta crisis es de doble tipo. En primer lugar, con base en una pregunta de cálculo, surge la dubitación: ¿Es posible aún llegar al progreso?, ¿es posible seguir progresando? En esa perspectiva, podría pensarse que —cambiando ciertas líneas y esquemas de causamiento— podrían evitarse dichos males, además de lograr que las gentes encuentren un sentido de pertenencia. Discusión sobre las vías equivocadas o correctas para alcanzar un ideal por todos compartido. Probablemente haya quienes estén de acuerdo con el progreso en tanto meta deseable; pero ahora, desde una perspectiva pesimista, aparece la crítica a la imposibilidad de continuar o mantener el crecimiento ascendente en el cual se plasma dicho progreso, principalmente visto en realizaciones tangibles de la cultura material. Discusión que palpa los enormes problemas actuales en campos como la calidad de vida, la crítica ambientalista y ecológica a la relación costo-beneficio dentro de la sociedad industrial y postindustrial. Este tipo de crítica no cuestiona tanto el progreso, como sus límites y su viabilidad (M. Albert, 1991). Una segunda crítica de carácter mucho más profundo apunta contra la noción misma de progreso, desenmarcada al colocarse como meta la idea de un crecimiento acumulativo, continuo, que es también por su propia naturaleza inalcanzable, a la manera del mito de Tántalo. Esta segunda crítica la formulan principalmente los filósofos de las llamadas escuelas postmodernas y es también punto de confluencia de estos con ciertos pensadores liberales anarquistas o cuasi anarquistas, o incluso de reaccionarios y nostálgicos conservadores. Así, se ataca la funcionalidad de los grandes paradigmas conceptuales de la Edad Moderna como las ideas de libertad, justicia e igualdad; el esquema 261 de la democracia política representativa, la idea de soberanía, etc., valores hasta hoy estatuidos como dominantes. En especial, se señala que dichos conceptos pretendieron, de manera abusiva, utilizar la argumentación hegemónica del discurso de la ciencia, con base en un principio de autoridad y necesidad insostenible. Discursos que, en opinión de los postmodernos, no son más que otra variante de los llamados metarrelatos. La característica de un relato, por ejemplo, una fábula o un mito es que su continuidad, un grado dado de coherencia lógica-discursiva que los hace inteligibles, aunque sean totalmente arbitrarios, son invenciones. Toda discursiva es narrativa y por lo tanto tiene lógica, coherencia entre sus partes. Un mito indígena igualmente posee coherencia, y siendo la ciencia también discursiva, asume un grado de coherencia semejante al de éste. Todas estas estructuras, las de un relato infantil, un mito indígena, la ciencia, o un discurso político, son todas ellas de tipo convencional. De tal forma, se recusa la pretensión de la ciencia de convertirse en un supradiscurso por encima del relato normal, del mito o cualquier otra forma de discurso oral o político. La legitimación por la ciencia es una característica de la Modernidad que ha entrado en crisis, incluso dado el propio de la ciencia y la crítica a ésta surgida desde la propia opinión pública. Todos estos argumentos de recurso racional y científico son utilizados por el Estado para autolegitimarse y perpetuarse. La crítica opera en muchos planos pero desde la perspectiva que más nos interesa, se esboza como una crítica a la legitimidad del Estado, anclada a su vez sobre la superioridad de la ciencia y de la técnica que aparecen como ideologías. De otro lado, se reclama para el individuo un espacio autónomo que ese individuo ha venido perdiendo a manos de la Nación-Estado13. En opinión de los pensadores post-modernos, mientras el mundo pre-político es plural y diverso, el mundo de la racionalidad política es ordenado. Por lo tanto, es un mundo basado en un principio totalizador. Así, la democracia no es un principio de diversidad sino de uniformidad. Sobre la base de uno de los grandes entes de la sociedad moderna: el partido, la nación, el Estado, etc., principios y valores que recogidos y apoyados por la comunidad, la decencia, las buenas costumbres, la moral, la virtud, resultan uniformadores de un orden 262 social, de forma tal que bajo este orden se “buscaría que no existieran zonas de incognocibilidad” (Oñate, Ob. Cit.). Es decir, se plantearía que los límites del conocimiento son solamente de tipo relativo. No existiendo limitación absoluta a la capacidad que el hombre tiene de conocer el mundo y de dominarlo. De tal suerte que existiría una sinonimia entre conocer y dominar14. En tercer lugar, se señala como presupuesto central de quienes sostienen esta operación de largo alcance, la interpretación lineal del tiempo, un esquema de lineabilidad ascendente de pasado, presente y futuro. Obviamente, bajo el supuesto de que todo presente es mejor que el pasado, y de que todo futuro será mejor que el presente. En pocas palabras, tales críticas pretenden preguntarse por la razón de ser del proyecto moderno. Este es el foco central de dicha crítica y no el debate en torno de si el proyecto moderno está inacabado. Estas son las bases fácticas que le dan importancia a la opinión elaborada de manera muy heteróclita por los filósofos postmodernos, y que le otorgan pertinencia, más allá de las alternativas que presenten o de los vacíos de dichos discursos críticos de la sociedad contemporánea. Se puede constatar igualmente un hecho, que los valores correspondientes a toda una época de cuatro o cinco siglos, desde el punto de vista de su consenso y su operatividad —y aún de su fundamentación— han entrado en crisis. Probablemente, lo que caracterizará a nuestra sociedad como sociedad global es la transición a un orden del cual aún no se sabe cuáles serán sus características. Sobre este tópico dice el filósofo alemán Arthur Kauffman: “Mi primera tesis que afirma que la Edad Moderna ha terminado, casi no requiere ser controvertida. Hace ya cuarenta años, Romano Guarddini, un genio de la Universidad de Munich, describió la Edad Moderna en un libro llamado “La Edad Moderna”, y mi primer maestro en filosofía, el célebre filósofo alemán Karl Jaspers, aclaró los signos que vinculan nuestro tiempo con anteriores cambios de época: la amechania (en griego: el extravío, el abandono, la angustia). En síntesis, las raíces de donde emergen todas esas épocas de la filosofía de la existencia: en el cambio de la edad arcaica a la antigüedad, acaso con Heráclito; en el cambio de la antigüedad a la Edad Media, acaso con San Agustín; en el cambio de la Edad Media a la Modernidad, acaso con Pascal; y ahora en el cambio de la Edad Moderna, una cuarta 263 edad aún innominada para la que muy especialmente es apropiado Kierkegaard, gran pensador danés del siglo XIX, que fue el antecesor de toda la filosofía existencialista” (Ibid). Ocurre un cambio de paradigma de lo moderno a lo postmoderno. La postmodernidad no es, en dicha línea analítica, la negación de todos los valores y los sistemas de sociabilidad de la Edad Moderna, es decir, que a diferencia de un discurso como el del marxismo o como los discursos fascistas que hablaban de crear un orden nuevo, una nueva era, los postmodernos no tienen un programa político, no pretenden transformar su crítica a la sociedad actual en una alternativa, receta y programa, de lo que debería ser la sociedad. En mi opinión el postmodernismo ha sido básicamente una estética15 de carácter intuicionista, que carece de un discurso, de un argumento de tipo político. Incluso la propia palabra postmodernidad ha sido muy debatida porque post es un prefijo lingüístico que significa después y así, cuando se habla de postmoderno, se alude a algo que va después de lo moderno, frontera temporal donde lo moderno ha llegado ya y sigue un mundo posterior. De otro lado, y mostrando las réplicas de los modernistas a dicha perspectiva, una de las objeciones que más se le han hecho a los postmodernos, es que critican la sociedad moderna enfatizando los graves síntomas de su agotamiento. Pero dicha Edad Moderna siempre ha vivido en crisis. Así, cuando surgió la Edad Moderna, hubo guerras, portentosas reformas y cismas religiosos, esclavitud masiva como fermento, se dio el proceso de la colonización de América, supremamente sangrienta; disputas entre las diferentes facciones que se disputaban el poder en Europa, etc. Es decir, la guerra, la inestabilidad política, la pobreza, las crisis, la hambruna, las desigualdades sociales, los cambios muy veloces de modelos de vida que al mismo tiempo han desubicado a la gente con relación a su sentido de pertenencia, a sus ideales; todos estos rasgos son característicos de la Edad Moderna siendo una equivocación fundamental ver estos fenómenos como algo nuevo, inédito, que implica el desplome del capitalismo, ya que este orden social ha demostrado una enorme capacidad de recomposición. El capitalismo surge como un modelo de sociedad capaz de ajustarse a todos sus vicios y salir triunfante de los mismos. No se ve de qué manera, por más postmodernidad de la que se hable, se 264 producirá un orden económico nuevo que difiera sustancialmente del modelo económico propio del capitalismo. Es probable que existan alternativas: capitalismo de bienestar o capitalismo salvaje, donde el papel del Estado sea mínimo, sólo el de ser un guardián de los propietarios; o que se llegue a un capitalismo de tipo intermedio; pero, en suma, siempre serán todas ellas variantes del capitalismo. 4. Redefinición del ámbito propio al interés público Ahora nos ocuparemos, tomando lo ya dicho como premisa, de la cuestión tan debatida de la crisis del interés público o el declive de lo público, según la caracterización citada en otros textos de este libro. En particular, planteándonos la pregunta de si esta crisis opera por determinaciones sociales o si, de otro lado, la actitud de los individuos, la relación anómica de desadaptación o desajuste valorativo de estos, constituye el fondo del problema. ¿Es la crisis de los intereses públicos un mero problema moral o anuda en un plano ético-individual? Como se sabe, durante el siglo XIX, una serie de pensadores, particularmente ingleses (Bentham, J. Mill, J. Stuart Mill y otros) reconstruyeron la teoría utilitaria y la plantearon en términos liberales y democráticos. Este replanteamiento se basó en la suposición de que existe un objetivo común en la sociedad que, en cierta manera se convierte en un imperativo u obligación, cual es lograr la felicidad de la mayoría. El liberalismo económico se basaba en una premisa, una especie de tautología, que consistía en que cada cual buscaba el bienestar individual a partir de su acción libre, natural; y como resultado de esto, se producía un bienestar colectivo que no estaba en la mente del productor o del agente inmediato, pero que significaba un resultado tangible y objetivo. He aquí el ideal moderno por excelencia, y esto lleva inmediatamente a la formulación de la pregunta ¿qué es felicidad? En la visión liberal del siglo XIX la respuesta está en manos de cada cual, apareciendo una separación muy abrupta entre los ámbitos de lo privado y de lo público. Así, lo privado, es decir, mi propia deliberación autónoma —no impuesta por otro, ni por una iglesia, partido o etnia— determinaría mi sentido de costo 265 beneficio con base en el cual calculo e identifico mi felicidad. Y, adicionalmente, otorga una guía respecto a lo que es útil y beneficioso, desde el punto de vista de un proyecto individual. En realidad el liberalismo económico resultó de una curiosa mezcla entre socialismo y el utilitarismo; y surgió bajo el predominio de un individualismo que exaltaba la norma. Subrayo esto porque aparentemente contradice las visiones normales del liberalismo económico, pues se afirma a menudo que éste desconoce al Estado y nada tiene que ver con el socialismo. Se supone que el individualismo es enemigo de todas las cortapisas a la acción del individuo y que bajo dicha óptica las normas legales, institucionales, las políticas públicas, etc., serían una limitación de la libertad de los individuos. Sin embargo, la adhesión a una cierta normatividad se entiende en tanto ella garantice la propiedad privada. En tal sentido el liberalismo tiene unos presupuestos básicos que corresponden al garantismo del Estado. En el modelo de Locke, como lo anotábamos al comienzo de este texto, el Estado tiene un lugar: es el garante de los derechos naturales a la propiedad, que se convierten en derechos positivos para evitar el estado de naturaleza. El individualismo liberal supone el acatamiento de una norma fundamental, la de convertir al Estado en garante del derecho de propiedad privada. Pero aparte del problema estricto del derecho de propiedad, con todo lo que involucra la existencia de la justicia, seguridad, policía, etc., un segundo problema de normatividad, fundamental para el liberalismo, consiste en que el Estado debe regular la competencia, no en el sentido de intervenir los precios y afectar con su favoritismo a ciertos agentes productivos, sino ejerciendo la tarea de garante equitativo para los participantes en el mercado, haciendo prevalecer la transparencia de los contratos. Por lo tanto, este modelo individualista liberal se opone al proteccionismo porque significa ponerle trabas y barreras a la libre concurrencia. La economía política clásica retoma el elemento rousseauniano de soberanía nacional y de la voluntad general; pero se aparta de Rousseau en la crítica que éste hace al individualismo. Rousseau es muy claro al afirmar que la suma de las voluntades particulares no hace la voluntad general. Sólo parcialmente se retoma a Rousseau, pero reinterpretado en un contexto utilitarista en la medida que el esquema de Bentham y de otros utilitaristas consiste en que la economía de mercado, la libre competencia, tiene por finalidad el logro del bienestar. 266 En esa perspectiva se sustenta la tesis democrática-liberal según la cual es necesario que la propiedad y la acción individual se traduzcan socialmente en un bienestar colectivo que consiste, a diferencia de Rousseau, en la suma de los bienestares individuales. Pero no se trata de un liberalismo que desatienda los problemas de las mayorías. Por el contrario este liberalismo supone que el mercado libre —sobre la base de estas normas mínimas de intervención del Estado, de carácter regulatorio y de garantía de los beneficios de la acción individual de una manera en cierta medida automática— dará lugar al beneficio común (S. Lidblom, 1999). Según sea el punto de enfoque, este liberalismo sería rapaz; mientras que existe un modelo político que plantea el principio de la justicia y la equidad social como una función legítima del Estado (F. Vallespin, 1985). Los ideólogos del siglo XIX no veían en esta dicotomía una contradicción, puesto que creían posible conciliar algunos de los elementos fundamentales del socialismo con el esquema de la sociedad liberal, donde se lograba una armonía sólo perturbada por el mantenimiento de privilegios de carácter sectorial y estamental, más propios del antiguo régimen que de la Edad Moderna. Muchas de las teorías liberales de este periodo son de una u otra forma darwinistas y están basadas en la selección natural que propicia que los menos hábiles queden eliminados y la sociedad progrese, teniendo mejores recursos bajo un esquema claro de competencia. Esto generó un rechazo de los socialistas al discurso liberal del mercado como igualador natural y como asignador óptimo de los recursos. La sociedad capitalista ha vivido sucesivas crisis en el contexto del libre mercado. Son los ciclos depresivos de la economía que se presentan con regularidad cada cierto número de años. Este modelo impulsado por la teoría socialista se concilió con buena parte de las vertientes liberales en tanto propugnaba que el individualismo pudiera ser satisfecho por un mecanismo planeado de intervención de las instituciones del Estado en la asignación de recursos. Ya que en la formulación del modelo liberal está supuesto el bienestar colectivo, y dentro de él se interpreta que el Estado cumple una función, se percibe el eco del socialismo en el problema del bienestar para la mayoría, el papel necesario del Estado para arbitrar, modificar y compensar estas irregularidades. Es evidente que la sociedad moderna se encuentra en una crisis de valores. Una serie de elementos que consideramos —al menos desde un discurso que 267 tiene su origen en el cristianismo— fundantes de las relaciones sociales se que han deteriorado. Piénsese, por ejemplo, en valores tales como la solidaridad, la cooperación, el desinterés, la ayuda a los otros, la caballerosidad, etc. De estos valores muchos correspondían a un modelo de sociedad tradicional —el modelo del honor— donde las conductas estuvieron sujetas a un reglamento no escrito; pero del cual todos los miembros de la comunidad jerarquizada eran responsables, desde luego, bajo un esquema de asignaciones específicas. Este modelo, basado en vínculos sociales dados por la religión, las costumbres nacionales, etc., se rompió. En su lugar existe hoy en día una apertura muy grande en la cual la gente puede hacer uso —sin ninguna dependencia o trascendencia de un valor superior— de su propia opción para configurar su conducta privada y para reflexionar sobre la conducta pública del resto de los miembros de la sociedad. En ese sentido hay una crisis y una quiebra de los valores. No porque la gente carezca de ellos, ya que en general quien actúa en la vida privada o pública tiene un punto de vista valorativo, posee la capacidad de juzgar qué es bueno, qué es malo, justo o injusto, sino porque los términos de referencia, desde los cuales se produce este juicio, se han transformado y relativizado. Ha ocurrido que lo privado y lo público se han disociado, siendo hoy contradictorios, antinómicos. Recuérdese que lo privado es al mismo tiempo privacía, es decir, limitación. Así cuando alguien habla de lo privado está señalando un límite frente a otras personas: este es el sentido ontológico de la propiedad privada (N. Bobbio, 1986). Derecho individualizado en el cual los demás individuos no pueden intervenir, ni menos aún el Estado, a no ser que existan las limitaciones del monopolio o el desmedro evidente del interés de los demás. Adicionalmente, lo privado aparecería de manera positiva como el ámbito de vida donde realmente la comunidad de ciudadanos construye sus finalidades, es decir, su felicidad personal. De tal modo que la vida privada es un bien que cuidamos y resguardamos celosamente. Esta privacidad es una positividad, es aquello que nos pertenece y donde construimos nuestra ética, establecemos nuestro sistema de valores y jerarquizamos cuáles son las metas de nuestra vida. Tal como lo expliqué en anteriores capítulos, lo público equivale al bien común, en su etimología original o como 268 ámbito de la llamada publicidad. Es aquello dispuesto a la observación de todos: espacio público, opinión pública, etc., donde quien tiene unos fines privados se ve obligado a entrar en relación con los demás para intercambiar y así plasmar sus propias finalidades privadas, ya que la mayoría de los fines privados no pueden realizarse al margen de las relaciones sociales. Resulta altamente complicado colocar una barrera infranqueable entre los fines privados y el medio público en el cual logramos relacionarnos con los demás y así poder en ese intercambio del que hablaba Platón, en su idea de la justicia como armonía común. Esta doble configuración de lo público y lo privado con sus contraposiciones es, a mi juicio, lo que ha puesto en una grave crisis la eticidad de nuestro tiempo. Vivimos en una sociedad donde valores ordenadores del pasado como la religión, la patria, los partidos políticos, las ideologías, están perdiendo terreno. De allí la gran apatía por lo político, no de ahora, sino de hace mucho tiempo. Existe, en cambio, el estigma de que lo político es lo corrupto, ámbito necesario pero negativo, y de que la vida privada debe estar en cierta manera opuesta al denominado bien común o interés público (N. Cunill Graw, 1997). El Estado de bienestar, el Estado intervencionista diseñado en los últimos dos siglos, asumió como deber de Estado una serie de funciones y de recursos que antes eran propios de la comunidad, de la familia o del ámbito de lo privado. Al respecto caben dos ejemplos. En primer término, la educación que hasta el siglo XIX era básicamente un asunto familiar. Las personas eran educadas por sus padres; los campesinos recibían de sus padres la instrucción que los capacitaba para continuar la labor milenaria de las cosechas y la domesticación de animales. Las clases ricas contrataban preceptores privados para la instrucción en el ámbito de la propia casa. Solamente en un nivel superior las elites se educaban en seminarios y en algunas universidades. Sin embargo, con el despegue del capitalismo industrial apareció la educación masiva o pública, característica de nuestra sociedad. El Estado empezó a reglamentar y a establecer códigos en esta materia y a controlar, desde muchas vías, la actividad educativa privada. Otro campo donde de lo privado se pasó al control público, fueron las relaciones matrimoniales y de familia. Una serie de cuestiones —como el castigo de los padres a los hijos, la cuasipropiedad del jefe de familia sobre su parentela (potestad marital y patria potestad)— empezó a ser limitada por 269 el Estado hasta llegar a un grado muy grande de intervención de lo público sobre lo privado. Podemos señalar que un ámbito antes privado, la familia, se convirtió en un ámbito público, reglamentado y tutelado por el Estado (F. Weffort, 1989). Así podríamos enumerar otros aspectos de la vida social, como los derivados de los llamados derechos sociales y asistenciales, a saber, la salud y la vivienda, que de ser ámbitos y responsabilidades puramente privadas se convirtieron en cuestiones de tipo público. En todas las ciudades modernas existen una serie de oficinas públicas ante las cuales se debe pedir permiso para buena parte de las actividades de la vida cotidiana. Quien quiere instalar un negocio, debe obtener una licencia; si se quiere construir una edificación, debe someterse el proyecto privado a las normas y ordenamientos públicos tales como patrones de seguridad sismorresistente, normas sanitarias, etc., las cuales constituyen tipologías externas a la propia voluntad de cada persona. Hay un desarrollo muy grande en todo este proceso de conformación del Estado de bienestar, pero ello implica una expansión de lo público a expensas de lo privado. Lo que ha venido ocurriendo en estos dos siglos, jalonado por la programática de la Modernidad, ha sido la estatización o nacionalización de buena parte de las esferas privadas tradicionales, bien sea porque el Estado las asume directamente o les impone límites y controles. Ello ocurrió porque incluso para pensadores como J. Stuart Mill (1971), quien se autocalificaba como “socialista”, desde mediados del siglo XIX entra en crisis la idea de satisfacer los intereses colectivos por la vía de la defensa de los intereses particulares desde el punto de vista de la asignación de recursos. Pero en la formulación del liberalismo económico existe un supuesto que no es estrictamente el de que cada cual haga lo que quiera y reciba lo que puede en una especie de individualismo rapaz, donde no tiene ningún sentido moral la dinámica de consolidación de un orden social; sino que en la formulación del modelo liberal está supuesto el bienestar colectivo y dentro de él se interpreta que el Estado cumple una función. Desde esa perspectiva se percibe un eco socializante en el discurso que propugna el bienestar para las mayorías, cual es el destacado papel del Estado para arbitrar, modificar y compensar las desigualdades. La contradicción en la que nos hallamos inmersos, derivó desde el mismo momento en que se le planteó a los ciudadanos que podrían hacer su propia individual y que el Estado no intervendría en los ámbitos de la vida privada y ello se planteó 270 como un principio del Estado de derecho moderno. En ese mismo momento empezó a ocurrir un proceso donde, por la naturaleza de las funciones públicas del Estado, éste avasalló los ámbitos específicos de deliberación y vida que correspondían a la privaticidad. El fuero interior resultó restringido y así se ha creado una esquizofrenia social: la contraposición entre un postulado ideal, la autonomía ética de cada persona para definir su vida y obligaciones; y, de otro lado, la existencia de una barda de reglamentaciones impuestas en nombre del interés público. Probablemente el argumento a favor de estas restricciones es el mismo del utilitarismo. Se trata de garantizar la felicidad del mayor número y el Estado moderno —no importa si es un Estado socialista, fascista, demócrata liberal o autoritario— siempre proclama el interés común como su fundamento de legitimidad. No conozco un sólo Estado moderno que haya planteado como su norma central algo diferente al bien común. Véase la nueva Constitución colombiana que hace una definición taxativa al definir a Colombia como un Estado Social de Derecho. Lo cual quiere decir que el Estado debe garantizarles a las personas, bienes y medios para que construyan su propia vida privada. Pero al ser éste un campo del derecho público, el Estado arbitra y reglamenta los conflictos de la llamada sociedad civil. Éste, en general, ha sido el mecanismo que ha operado en los últimos dos siglos y por medio del cual se ha legitimado el discurso sobre la tutela de lo público estatal sobre los espacios de la privaticidad. Este problema se expresa en el plano teórico en términos de la articulación resultante entre lo objetivo y lo subjetivo, entre lo individual y lo social. Un célebre sociólogo norteamericano, Robert K. Merton, trató de hacer una síntesis desde una perspectiva postweberiana del debate frente al modelo durkheimiano, cuyos elementos coactivos, inconscientes y objetivos tenderían a articularse con los elementos valorativos planteados por Weber. No se trataba entonces sólo de valorar cómo las tendencias gobernaban la sociedad, sino cómo éstas se construyen desde las subjetividades valorativas inherentes a la acción de los individuos. Toma de partido entre lo objetivo e inevitable, y lo valorativo y susceptible de cambio en tanto la sociedad genera un consenso de valores. Merton, hablando de la sociedad norteamericana, dice que su valor fundamental, la sustancia social cohesionadora, es 271 la riqueza, paradigma de la cosificación; de tal forma que los individuos sólo se reconocen en términos del consumo y la acumulación de riquezas. El concepto de desviación social y los estudios sobre la desviación individual o patología del desviado, surgieron con fuerza en la tercera década de este siglo, particularmente en una discusión entre dos tendencias. La primera de ellas estaba conformada por las concepciones del psicoanálisis freudiano, en particular desarrolladas por discípulos de Freud que se esforzaron por integrar su pensamiento con un punto de vista marxista. En particular, me refiero a la corriente de pensadores alemanes-judíos que en el transcurso de los treinta emigraron desde Alemania hacia los Estados Unidos: Erich Fromm, Marcuse, Horkheimer, Adorno y otros; pensadores miembros de la “Escuela de Frankfurt”. Freud propuso la teoría de la naturaleza humana según la cual en el hombre existe una ambivalencia u oposición innata entre los instintos naturales y la cultura, entre Eros y Tánatos. Oposición fundada en la inclinación hacia la reproducción social, el altruismo, etc., y de otro lado, en el arraigo de una actitud individual egoísta de tipo natural, base de los hondos conflictos humanos. Como se recordará, para Durkheim la conducta no dependía de un simple acto de voluntad, puesto que el hombre no actúa mediante o movido por una deliberación anterior a cualquier acción social. Al contrario, los individuos se encuentran sobredeterminados por fuerzas, tendencias e intereses que éstos no controlan. Las normas, como lo diría Durkheim (o el imperativo moral tal como lo formuló Kant), resultan, en términos explícitos (o son fruto de la fuerza del grupo) de un gregarismo colectivo que limita y maniata la autonomía propia y supuesta en la voluntad individual. Efectivamente, los individuos disponen de libertad natural, pero su margen efectivo de ejercicio se halla controlado por un sistema de prohibiciones. Alternativamente, desde el punto de vista marxista, la naturaleza humana es una abstracción, en cierta medida percibida como inocua, o una cuestión retórica, en tanto el marxismo —como lo vimos en el capítulo primero— es un historicismo, ya que define al hombre gobernado en cada fase histórica según las leyes económicas y sociales. En esa medida, las críticas de Durkheim al final de siglo XIX, al ideario socialista no podían ocultar el hecho de que tanto Durkheim como los marxistas compartían la concepción de la prioridad de las normas sociales. Pero para el marxismo éstas resultan históricamente condicionadas y 272 explicadas en una perspectiva genética producida desde las categorías inherentes a la economía política. En segundo lugar, se desarrolló una explicación alternativa del problema de la desviación que también tiene su origen en la teoría popularizada en los Estados Unidos durante los años 30 por el citado sociólogo Robert Merton, quien publicó en ese periodo sus conocidos estudios sobre el conflicto social y la desviación. Para Merton, la anomia equivale literalmente a la carencia de normas, siendo, sin embargo, un estado normal dentro de la sociedad. Pues bien, Merton amplió los conceptos de anomia definidos por Durkheim operando mediante un desdoblamiento del concepto de anomia desde el individuo, visto como un ente autónomo que rechaza las normas sociales, afirmando que también existe anomia cuando la sociedad entra en crisis de sus sistemas de control social y se produce una situación en la cual los individuos no saben qué pautas de conducta adoptar. En otras palabras, en ese segundo campo, se presentan graves fallas en los procesos de socialización. Esto es para Merton un fenómeno doble que va desde la colectividad al individuo; y no sólo al individuo aislado sino a la suma de los individuos. De otro lado, la desviación también podría analizarse como un fenómeno manifiesto en un plano individual presente en una sociedad plenamente vigente y funcional. De esta manera pueden existir diferentes niveles de conducta. Y ello en el contexto de una sociedad bien ordenada, con una relativa armonía donde las causas de anomia o de desviación son muy pocas, y en la que prácticamente existiría un proceso automático de criminalización de la conducta desviada. La sociedad paradigmática de este tipo de orden, para Durkheim fue la sociedad tradicional, rural o precapitalista. En tal sociedad, la estructura de socialización, el patriarcalismo, el patrimonialismo, la división en estamentos, etc., haría que los diferentes individuos y grupos, desde su infancia, supiesen ubicar su lugar, sus sistemas de normas, de permisibilidad y/o prohibición (E. Durkheim, 1976). Ello explica por qué muchas de estas sociedades tuviesen una larga duración sin que produjesen transformaciones fundamentales. El feudalismo, la esclavitud y, en nuestro caso latinoamericano, la colonia hispánica son ejemplos palmarios de lo anterior. En realidad, el aporte más importante de Merton, a su vez respuesta a la perspectiva neofreudiana, fue la construcción de un sistema analítico que 273 estableció una correlación entre fines y medios. Una sociedad inscrita dentro de un sistema de normas, socializa un conjunto de fines culturales que guían la acción del individuo; pero para la persecución de los fines se requieren medios. De este modo existirían medios sociales para la consecución de fines culturales, de tal manera que en su esquema aparecería una estructura de fines (valores) paralela a la estructura de medios, que instrumentalizan, estos últimos, los canales legítimos o ilegítimos para la movilización social. Como lo afirma con frecuencia Omar Aktouf (2002), el pensamiento administrativo ha sido de manera dominante un “brazo armado” de la economía. Tanto de la disciplina económica, que ha influido muchísimo sus postulados, métodos de análisis y problemas de investigación, como de las prácticas económicas y los intereses de política prevalecientes —en el contexto del sistema global hoy dominante— de capitalismo post-industrial. La economía de mercado ha marcado los contenidos de la enseñanza managerial. Y evidentemente el Management, como campo académico, es un resultado “natural”, en primer término, de la aparición de las modernas formas organizativas, complejas, que han caracterizado al mercado a partir de finales del siglo XIX (A. Chandler, 1962); y de otro lado, del desarrollo gigantesco del aparato administrativo de los Estados centralizados bajo el modelo genérico del intervencionismo y el Welfare State. Si hoy en día releemos a los clásicos de la economía política, por ejemplo la referencias del profesor M. Patry (2004) sobre A. Smith, o las de O. Aktouf (2002) sobre Marx, Ricardo, Hobson, Say, etc., notaremos a primera vista que el fenómeno organizacional no estaba en el centro de sus preocupaciones.Desde luego, sí vislumbramos parcialmente muchos de sus elementos constitutivos. En el caso de Smith, para éste una noción central fue la cooperación, coordinación y división del trabajo, temática que luego sería enriquecida por Marx y Durkheim. En el análisis de Marx —que sigue claramente los trazos de Smith y Ricardo—, otro aporte decisivo fue su rica y matizada descripción de la ganancia y el sistema del trabajo asalariado (I. Rubin, 1977). El propio Say hizo una primera definición del espíritu del entrepreneur. Empero, las dinámicas económicas que dieron nacimiento a la gran corporación industrial, al sistema organizacional bancario, a la aparición de las 274 bolsas de valores, las grandes redes de distribución mercantil del capitalismo de consumo y producción en masa, en fin, fueron los catalizadores para que surgiese el pensamiento administrativo; de forma independiente al desarrollo de la ciencia económica. En una primera etapa fundacional, uno de los retos del capital al campo académico fue la racionalización del proceso de trabajo en el interior de la planta industrial, lo que sería el epítome del Management científico en el marco del modo de producción fordista. Posteriormente otras demandas sobre la productividad llevaron a la aparición de las versiones más sofisticadas de la escuela de relaciones humanas. Incluso, hasta en nuestros días, la reingeniería, el downsizing, el TQM y otras de las llamadas “modas administrativas” son en general, respuestas académicas, tanto interpretativas como instrumentales, de las formas racionalizadas de composición de los distintos mecanismos operadores de la lógica del capital. De cierta manera, a lo largo del siglo pasado, el Management y la economía como campos disciplinarios (o transdisciplinarios en el caso del Management) han discurrido de manera paralela, con puntos de confluencia en la interpretación del sistema económico del capitalismo mercantil. Cabe anotar que estos discursos se entrecruzan, especialmente, en la utilidad normativa formalizada de sus marcos analíticos y en la supuesta capacidad predictiva sobre las actuaciones de los agentes económicos y las decisiones empresariales. No obstante estas confluencias topológicas, el problema más importante —desde el punto de vista epistemológico— han sido los intentos por lograr efectivamente la autonomización del pensamiento administrativo. En este sentido resulta evidente hoy, que desde el punto de vista de las metodologías empleadas, las diferencias en cuanto al estatuto de cientificidad son enormes respecto del pensamiento económico. En efecto, la economía, de la mano de la matematización y la formalización, ha seguido “exitosamente” al pie de la letra la programática del positivismo del siglo XIX, o sea asumir como objetivo de la investigación social, en este caso, la naturalización de las realidades económicas. A partir, por supuesto, de la instauración hegemónica del individualismo metodológico. Así vemos que en la economía éste domina la escena como sujeto individual transpuesto, transvertido de muchas maneras: en firma, en consumidor, en productor, en trabajador asalariado, en 275 inversionista, etc. Y a partir de una optimalidad paretiana —simple—, refleja que permite la estandardización de las leyes naturales de la economía micro. Esta disposición de la ciencia económica está en la base de su pretensión de tipo imperialista. En efecto, si leemos a un economista Nobel, como Gary Becker, encontramos que en su análisis aplica la lógica mencionada al matrimonio. Nos casamos, divorciamos y/o tenemos hijos de la misma manera que actúa un decisor económico; al fin de cuentas un matrimonio es un contrato. En el campo de la ciencia política o de la administración pública, esta misma lógica (G. Tullok; J. Buchanan) pretende dar cuenta de la corrupción —como cálculo de tipo paretiano de los agentes en términos de incentivos e inhibiciones o de los resultados electorales— las justas electorales son un mercado con vendedores (los candidatos) y compradores (los electores) y así ad infinitum. En realidad la corriente más innovadora en el marco del propio Management made in Usa ha debido romper parcialmente con el economicismo. Esta es quizás la principal aportación de la escuela de mediados de los 50 que cristalizó el trabajo de Herbert Simon, al reducir los límites de la racionalidad de los agentes organizacionales y criticar la inviabilidad del análisis microeconómico transpuesto a las empresas y demás organizaciones modernas. Aunque debemos matizar que igual Simon y sus sucesores (March, Cyert, Homans, y otros) siguieron fieles al individualismo metodológico. En el fondo, el management no ha salido en su versión principal de la “aula de hierro” —como decía Weber— del racionalismo instrumental. Del lado del Management, en cuanto teorización, tal tipo de consenso metodológico no existe en modo alguno. Como muy bien lo describe Dery, coexisten una multitud de paradigmas simultáneos en las ciencias del Management. Éste atraviesa desde siempre —perdidas tempranamente las ilusiones del cientifismo tipo Taylor, Fayol y otros— la imposible —según Dery— búsqueda de su propio campo epistemológico. Quizás la imposibilidad estriba en la intención de ser un corpus monodisciplinario. Pero en una perspectiva trans o multidisciplinaria como lo aclara agudamente Gusdorf, existirán alternativas epistemológicas. Al fin de cuentas el Management se teoriza de la forma que precisaba Hegel: la lechuza de Minerva alza su vuelo sólo al anochecer. Únicamente con la plena consolidación de los fenómenos organizacional y managerial estaremos en condiciones de elaborar 276 teorías maduras sobre los mismos. Las prácticas sociales más significativas (producir medios materiales de vida, reproducir y rendir sustentable la vida misma, generar simbologías y discursos en los planos lingüístico y simbólico, etc.) se constituyeron como campos disciplinarios en los siglos xviii y xix, aunque se ejercían desde milenios. Las demandas del mundo de los negocios hacia el Management —como campo académico— lo impelen hacia el esquema normativo-prescriptivo. Se espera que las investigaciones de los centros de estudios de más alto nivel (grandes universidades, institutos y escuelas) digan, indiquen taxativamente, cómo mejorar la gestión de la organizaciones, incrementar sus eficacias, disminuir sus costos, posicionarse en términos de la competitividad global.Y por supuesto que esto se refleje en los diseños y contenidos de los principales programas para la formación de los futuros administradores de negocios. Es aquí donde podremos ubicar la contribución al debate sobre el Management y su enseñanza que ha suscitado el reciente libro de H. Mintzberg (2004). Éste alude a la necesidad de rectificar a fondo estos programas, alejarlos del mecanicismo, de la excesiva formalización, de la esterilidad agobiante del método de “caso”. Mintzberg defiende una perspectiva teórica según la cual el Management no se aprende fuera de las organizaciones y de las propias prácticas administrativas. Así, que sólo una fuerte interacción teoría-práctica será una real alternativa de resolución de la crisis epistemológica del mismo. Quizás debamos avanzar más allá de las críticas fundadas de Minztberg hacia las profundas dimensiones ontológicas que sólo aparecen de manera tácita en su texto. Si el Management es un arsenal más o menos estructurado de herramientas para manejar a las personas y a las cosas (según la conocida metáfora del conde Saint-Simon), el problema de fondo reside en la función social de las tecnologías, del saber instrumental. Creo que en este punto es donde se encuentra el mérito del reciente texto de Celine Lafontaine (2004). La cibernética de mediados del siglo XX (Wiener, Bateson, M. Mead, Lévi-Strauss y otros) es en esencia, la puesta en marcha de la aritmética moral de los ilustrados del siglo XVIII (Condorcet, Buffon, Holbach). Una nueva mecánica funcionalista y comunicacional basada en la desaparición del sujeto consciente y razonador, eje del modernismo filosófico. La sofisticación de las técnicas de autocontrol social y organizacional con 277 el advenimiento del ideal sombrío del cyborg, durante mucho tiempo sólo premonición mediática ficcional y ahora figura central del nuevo modo de gestión en la época de la tecnociencia informacional. Como lo explica Lafontaine, el nombre de la actividad denota la intención, una construcción artificial de la gobernanza humana. La base del control de los flujos relacionales en una perspectiva sistémica. Las alternativas de salida al círculo vicioso estarían en la ruptura del modelo instrumental, del tecnicismo unidimensional, en la recuperación del papel social de los sujetos, en la humanización del Management. Se trataría de volver a colocar en el centro de las reflexiones las llamadas dimensiones olvidadas del Management. Este es uno de los puntos fuertes de la proposición teórica de R. Bedard y de R. Dery en un plano epistemológico; y de Chanlat y Aktouf en un terreno propiamente ontológico (la re- situación del lugar ontológico del Management). Si el Management da cuenta del inevitable proceso de mediaciones organizacionales en las actividades humanas de las edades modernas y postmodernas (reemplazando los espacios de la vida familiar extendida y de las comunidades naturales), su esencia no se agota en la lógica del capital, y menos aún en su variante financiera-especulativa. Prueba de ello es el Management público, que se ocupa del Estado y su función administrativa —no reductible a la relación economicista (J. P. Noche, 1996)—. Y el Management de las organizaciones del llamado tercer sector tales como el benevolat, la acción solidaria, el sindicalismo, las cooperativas, las iglesias, los centros de educación, cultura, pensamiento, recreación. Todo ello nos pone en evidencia que en la vida práctica de las sociedades existe un amplio espacio para el ejercicio de la gestión pluralista, que asume como eje un tejido de valores adheridos a la cultura con sus diferencias nacionales, étnicas, de género, generacionales, etc. Ahora el reto es ofrecer alternativas de interpretación de los fenómenos gestionarios que rompan el paradigma dominante y reconozcan las especificidades sociales y culturales. El universalismo del Management en cuanto discurso objetivador ha sido por ello acremente cuestionado en escuelas críticas del mundo de la francofonía; pero en América Latina, primordialmente. 278 Sin embargo, esta es una tarea académica que no se agota ni mucho menos sólo en la deconstrucción del discurso dominante. Requiere también ofrecer alternativas en el campo de la actuación managerial con útiles que permitan reconocer la diversidad, la participación, la diferencia de intereses, las relaciones complejas entre lo político, lo sociológico, lo individual y lo managerial. En esta dirección, esquemas de análisis como los suministrados por la profesora Renée Bedard (1995) y el profesor Alain Chanlat (1993) son pertinentes y merecen ser afinados, socializados y contrastados. Esperemos que en nuestros trabajos de investigación ello se pueda escenificar para contribuir a determinar valores, ontologías o modos del ser burocrático; así como al estudio específico de la relación entre las realidades nacionales, la mundialización y las formas más abstractas de la teoría. 5. El predicamento intervencionista Debemos fijar ahora nuestra atención en los contenidos esenciales de la reforma del Estado. Lo primero a saber es que éste es un debate antiguo. Recuérdese el estado del mismo durante los años veinte y treinta en los EE. UU. y Europa, cuando el lalsseferismo hizo crisis y surgieron triunfantes teorías intervencionistas de estirpe keynesiana (J. M. Keynes, 1943). Un síntoma de época fue el ensayo La orientación de la producción en un Estado socialista; escrito por Fred M. Taylor (1928), quien era en ese entonces presidente de la Sociedad de Economistas de los Estados Unidos. En dicha coyuntura estaban los EE.UU. lejos de la oleada de furibundo anticomunismo, característica del macartismo posterior a la Segunda Guerra Mundial (guerra fría). Al contrario, en muchos círculos intelectuales se profesaba simpatía respecto al experimento soviético. La Unión Soviética llevaba escasamente una década de haber empezado el intento de construcción del socialismo en un solo país. En el pasado reciente en la URSS por parte del propio Lenin se habían implementado políticas económicas heterodoxas como la llamada N.E.P., un régimen de capitalismo de Estado. Posteriormente se daría inicio a la implantación de los llamados planes quinquenales, con los cuales se pretendió dejar de lado el rol del mercado, el valor de cambio, la asignación de recursos por la vía de los precios, etc., 279 estableciendo, en su lugar, la planificación centralizada. La Unión Soviética iniciaba la exaltación ortodoxa socialista de la planificación estatal16. En el mundo entero, América Latina, el resto de Europa, los EE.UU, etc., se miraba con gran simpatía la propuesta socialista de resolver los problemas de inequidad, desigualdad social, pauperización de las masas obreras, desempleo y también las guerras y la corrupción inherentes a la “decadente” sociedad capitalista17. En tal contexto se puede entender esta lectura de Taylor, donde hace una defensa conceptual de la posibilidad del socialismo como un sistema económicamente viable, racional. Y además, dentro de dicho marco ha de valorarse la elaboración mucho más sofisticada del economista polaco Oskar Lange sobre el problema de la planificación económica (O. Lange y F. Taylor, 1970). Alternativamente a esta justificación de la viabilidad del socialismo, surgió con fuerza en los años treinta la crítica al liberalismo manchesteriano, incapacitado según sus opositores para garantizar el desarrollo económico y combatir la desigualdad y las marginalidades sociales. Ha de recordarse en esta dirección a Keynes quien tuvo una gran influencia, no sólo como académico, pues fue consultor del gobierno inglés, y luego se convirtió en un asesor del gobierno norteamericano, particularmente a raíz de la crisis muy profunda del año 1929. Según Keynes, había llegado el tiempo del final del laissez-faire, pues fue derrotada la idea de que automáticamente se daría una asignación óptima de recursos por la vía del mercado libre. Según tal criterio, el revés de este viejo modelo consistió en generar una distorsión de la economía, un desempleo a gran escala, e incluso, lo que es más grave, incitó directamente la crisis del veintinueve con la suspensión de circuitos completos de la actividad económica, la bolsa, la actividad de los bancos, etc., que según tal teoría eran claros y racionales en un esquema de mercado abierto. Su propuesta, por lo tanto, estuvo encaminada a asignarle de nuevo un papel económico fundamental al Estado. Básicamente la concepción de Keynes estribó en tomar las riendas de la política económica y monetaria, logrando un control estricto sobre los fenómenos monetarios macro, tales como la devaluación, inflación, etc. En segundo lugar, propuso Keynes promover desde el Estado un apalancamiento de la inversión productiva. En términos populares lo que Keynes sugirió fue una política en que el Estado debería dedicarse a la inversión en 280 las obras de infraestructura y al mismo wazzu al tiempo fomentar el empleo. Efectivamente, el keynesianismo permitió superar la crisis de los 30 en los EE.UU y en alguna medida en el resto de Europa, puesto que el proceso se vio interrumpido por la Segunda Guerra. Pero incluso la guerra, negocio del Estado, cimentó el intervencionismo del Estado. Desde la época en que las anteriores ideas gozaban de un gran consenso a nuestros tiempos actuales existen desde luego hondas diferencias. En general no motivadas tanto por elucubraciones teóricas sobre los modelos de desarrollo y gestión social, sino como fruto del balance de la experiencia histórica al final del siglo. Es así como se presentan tendencias mundiales muy significativas; la más importante de ellas el derrumbe del llamado “socialismo real” en los países de Europa oriental y la disolución de la antigua Unión Soviética. Y de ellas colegimos sin mayor disenso que hoy en día se palpa un fracaso estrepitoso del modelo socialista leninista. Preciso es que se trata de un modelo específico puesto que es muy temprano aún para hablar de un fracaso total o más exactamente de la imposibilidad del socialismo, pues esta es una ideología política que tiene muchos matices18. Recuérdese, además, el rol nada desdeñable en las más importantes vertientes políticas burguesas del compromiso programático con lo social. Precisemos entonces qué tipo de gestión socialista ha hecho crisis: aquel socialismo que limita en forma severa el papel de la propiedad privada y restringe al máximo posible el rol del mercado, colocando, en su lugar, como preeminente, la propiedad colectiva y estatal en función de una utilidad común, desarrollando tendencialmente una apropiación colectiva sobre medios y relaciones de producción. Esta definición implica el reemplazo del rol distributivo del mercado por la asignación de recursos arbitrados desde la planeación estatal. En cuanto concierne al socialismo leninista, éste incluye un rasgo adicional en su funcionalidad política: la simbiosis del partido único y el aparato del Estado. Aspecto en el cual, no por razones puramente aleatorias, coincidieron el fascismo y el autoritarismo marxista leninista (R. Bharo, 1979). Situados en las anteriores precisiones no se puede hablar de que los procesos de socialización, asignación de una función preeminente a la propiedad pública, hayan fracasado en el mundo entero, pues muchas ramas de la vida económica y social permanecen con dicho carácter de propiedad pública o colectiva y no hay 281 una posibilidad inmediata de restitución total del viejo esquema laisseferista en el cual desaparezca en muchas ramas de la vida social la apropiación colectiva (cooperativa, solidaria, autogestionaria, estatal, etc.) sobre los medios de producción y trabajo. 6. La modernización del estado Cuando se discute el problema de la modernización del Estado se nota la escasa utilización de una conceptualización rigurosa sobre lo que significa la Modernidad19 y su corolario, la modernización de la vida pública, de la cual el aparato del Estado constituye una parte sustancial. Existen, muy al contrario, supuestos comúnmente aceptados —principalmente por las élites depositarias del poder, por los partidos y facciones políticas vinculados al stablishment— según los cuales hay ya Modernidad, uniforme, global. A partir de dicha presunción, frente a esta realidad, no tendríamos tarea distinta que acelerar nuestros propios pasos para alcanzarla. Se trata de la conocida dicotomía entre desarrollo y subdesarrollo, parte sustancial del discurso que apuntala los organismos multilaterales internacionales en forma repetitiva desde la lejana época de los años 50. En dicha perspectiva, por la inflexión modernización, aunque esto se vea sólo como una cuestión de tipo lingüístico, se entiende la puesta en marcha de una programática de Modernidad. Por ello la primera discusión presente, desde una perspectiva analítica, consiste en definir, tanto la especificidad de la llamada Modernidad, como —correlativamente— indagar en qué consisten realmente los procesos de modernización. Piénsese en que nunca se habla de Modernidad del Estado, sino de su modernización. Hay en esta distinción una suposición racionalista en el sentido de que los procesos de modernización pueden ser inducidos o dirigidos por actores sociales, trátese de partidos, dirigentes o proyectos políticos que arriban a la dirección de un Estado. Por el contrario, podrían concebirse simplemente como procesos inconscientes a cargo de fuerzas sociales subyacentes en una sociedad, tanto en el plano económico como en la cultura civil que, sin la intencionalidad de ningún actor, propulsarían una alternativa de cambio. En cierta medida se trata de tomar en cuenta dos visiones del problema: a) Una alternativa dirigista, racional-teleológica, donde las elites dirigen el proceso, lo impulsan y lo encausan como parte de un proyecto político. Este 282 nivel implica el estudio del tema de una manera básicamente descriptiva, por lo cual se lo encasilla con relativa facilidad en los marcos de la ciencia, del saber positivo. b) Otra perspectiva reconoce la modernización como un proceso, pero más bien como expresión de la fuerza de determinadas tendencias sociales. Aquí, se parte desde un nivel filosófico donde se plantearía el problema de los fundamentos o causalidades profundas de los fenómenos sociales descritos desde la causalidad eficiente del primer nivel. Podrían caracterizarse la mayoría de estas interpretaciones como anudadas en una concepción de tipo hegeliano en la que la Historia evoluciona en el sentido del progreso, actuando portavoces del mismo con un mayor o un menor grado de conciencia. La Modernidad, de otro lado —es difícil definirla con una sola línea de mirada—, podríamos captarla destacando los rasgos globales que implican su arribo y su diferencia frente a las sociedades tradicionales. Adicionalmente en muchos de los presupuestos conceptuales vigentes en esta temática, se establece un derrotero: mientras de un lado, la Humanidad, al iniciar el medioevo se aprestó a una ruralización de la vida pública, proceso que duró unos doce siglos, el surgimiento de la Modernidad implicó de nuevo la colocación en primer plano de la ciudad o de lo urbano como ámbito de dirección social y el relegamiento de lo rural colocado como un elemento al servicio de lo urbano. Dicotomía tantas veces expresada como oposición entre civilidad y barbarie, entre lo artificial y lo natural. Hay, de este modo, una creencia relativamente uniforme en un progreso que probablemente tiene retrocesos, ciclos, pero que siempre marcha de forma ascendente. Desde el punto de vista económico, fue distintivo de la Modernidad su contexto mercantil capitalista. Así, con base en estos procesos contradictorios, se plasmó el predominio de un nuevo orden basado en la primacía de una sociedad posesiva y abierta, que estructuró un discurso de tipo institucional y jurídico basado en el reconocimiento de los derechos individuales. Surgió la teoría moderna de los derechos ciudadanos o civiles, básicamente encaminada a definir las relaciones entre el Estado y los propietarios privados. Evidentemente de lo dicho a lo largo de los anteriores capítulos, se desprende que cuando se alude como programa a la modernización del Estado, 283 se parte del supuesto del progreso y se asume una temporalidad de carácter ascendente. Así la modernización resulta vista como un esfuerzo por modificar, recambiar al Estado como institución pública que constituye en realidad un sistema de organizaciones con responsabilidades en el manejo de los bienes y servicios públicos a través de la implementación de una política de ingeniería gradual (K. Popper: 1974). Por supuesto que dicha aserción corresponde a la vigencia de un modelo relativamente consolidado de la economía capital-mercantil. En otros órdenes o modos de organización social, v gr., en las sociedades tradicionales, en el feudalismo clásico o en sociedades autoritarias, las fuerzas de la denominada sociedad civil, o simplemente no existen, o se encuentran maniatadas; aunque ello no impida su influjo en un largo plazo. Pero bajo el supuesto señalado y en la medida en que resulten más rápidos los cambios de la órbita del mercado, la reforma política pretende ajustar la marcha del Estado a las crecientes demandas sociales. Desde esa perspectiva, la modernización se vislumbra como algo fundamentalmente instrumental. Como se trata de que el Estado cumpla funciones nuevas o refuncionalice su estructura para lograr resolver nuevas metas, esto sólo puede darse cambiando o mejorando los instrumentos y el abigarrado tejido de las organizaciones que lo constituyen. La instrumentalización racional se logrará a través del impulso de políticas —las que suponen un conjunto de insumos materiales y epistemológicos—. Hoy en día el discurso de la modernización del Estado implica, a lo largo y ancho de los países insertos en dicha programática, una serie de tópicos comunes, de invariantes estructurales del proceso que, en orden de relevancia, podremos enumerar así: a. El manejo de las organizaciones empresariales públicas de forma semejante a las privadas, con base en el criterio racional del costo-beneficio y atendiendo a las señales del mercado. Es el predominio del llamado modelo de Estado empresarial, sobre el cual se ha suscitado una copiosa literatura en los últimos años; particularmente en la versión anglosajona que propone la “reinvención del Gobierno” (Osborne y Gaebler: 1992). b. Impulso a políticas de descentralización, desconcentración y autonomía en la gestión de las entidades públicas, otorgándoles a sus mandos 284 una responsabilidad grande, y disminuyendo el peso, paternal o autoritario, de los niveles superiores. Esta tesis de la ruptura de una unidad vertical, jerárquica en las estructuras de la Nación Estado, es en realidad el corolario de la propuesta de democratización de la gestión de los asuntos públicos a partir de la puesta en marcha de la autogestión, la participación ciudadana, con base en un discurso que reconoce las dimensiones creativas de las comunidades y la importancia de la reconstrucción de la sociedad civil (E. Varela: 2005b). c. La supremacía de la razón ciencia-técnica moderna en la conducción de los procesos administrativos y de tipo productivo de bienes y servicios. d. Finalmente, a la modernización se la vislumbra como un continuum de agregados y cambios tanto en el hardware como en el software gestionario. Todo este conjunto de elementos se ensambla coherentemente a nivel del diseño, formulación y evaluación de las políticas públicas (Public Policies). En este sentido es muy importante delimitar hasta qué punto los procesos de política pública tienen que ver con contenidos específicos. Las políticas públicas se rigen por diversos modelos; principalmente oscilando entre el centralizado o principesco, donde un centro, una elite decide desde arriba; y el democrático o participante, donde el proceso de decisiones es abierto, se construye a diario mediante la práctica de consultas reconociendo los intereses afectados, etc. La modernización del Estado supone que la racionalidad burocrática no ha llegado a su límite y por lo tanto se trata de maximizar dicha racionalidad como leitmotiv del proceso de modernización (N. Crozier, 1992). Esa es, en esencia, la idea tan recurrente en las escuelas de administración de articulación entre la eficacia y la eficiencia de la acción pública como verdadero paradigma o norte de un programa de modernización estatal. Recuérdese el lema que la administración Gaviria en Colombia acuñó hace unos años para darle impulso a su programa de liberalización y reforma de las organizaciones públicas: “Entidad del Estado que no se modernice, está condenada a desaparecer”. Así aparece el ajuste a un modelo modernizante de gestión, no como una opción, sino como un imperativo categórico. De otro lado, cada vez más en este impulso del programa de modernización le fue dado un papel dominante a la planificación, bien sea desde 285 la cúpula del Estado, o resultado de procesos de concertación con diversos grupos de interés. La planeación corresponde al paradigma de la racionalidad instrumental, que ha sido debatido por filósofos y sociólogos a lo largo de los últimos dos siglos, particularmente en torno a la obra de Max Weber (1969). La racionalidad instrumental proviene de la relación que en Occidente se ha construido desde Descartes, dándole un papel preeminente a la persona, a la capacidad que el hombre tiene de gobernarse a sí mismo. La articulación entre fines y medios expresa el contenido filosófico conceptual de la planeación. Se asume en tal perspectiva que los agentes sociales podrían darle una dirección intencional a la suerte, al destino de los individuos y grupos que constituyen una sociedad. Con base en esta suposición filosófica, se constituyeron una serie de herramientas y metodologías técnicas para volver realidad este supuesto. Generalmente las políticas públicas de los últimos decenios han estado inspiradas en la concepción utilitarista del bienestar general, particularmente en términos cuantitativos y cualitativos (A. Hirschman: 1958). Bienestar que no se deja a la deriva, al albedrío automático de las diferentes individualidades. Al contrario, un papel preeminente del Estado como responsable del interés común consiste en encausar y planear mediante etapas, caminos que lleven a la sociedad hacia una meta determinada. ¿Qué discurso hay detrás de la instrumentalidad? El discurso racional que el propio Habermas (Habermas y otros: 1991) ha calificado como racional instrumental, basado en una relación costo/beneficio consistente en el logro de un óptimo: minimizar costos, aumentar los beneficios. El Estado, en esta perspectiva, existe para generar medios con el fin de que los miembros de la sociedad se doten a sí mismos, de manera libre, de sus bienes (como finalidades) privados. Las finalidades son individuales en cuanto a sustancia, proyecto de vida. En esta elegibilidad se basa el esquema de la sociedad liberal con toda su parafernalia de los derechos individuales y sociales. Por ello los programas de Estado son fundamentalmente programas de tipo cuantitativo o cualitativo para ampliar o mejorar servicios: la cobertura de la escuelas, recoger más toneladas de basura, reciclarla, prestar un servicio más eficiente de energía eléctrica, guardar de mejor manera el orden público, disminuir la tasa de homicidios, etc. Así se muestra la correlación entre la eficacia y la eficiencia en términos 286 tales, que cuando un gobernante entrega resultados de su gestión, los expresa en términos mensurables. Frente a la lucha contra la criminalidad se informa: antes mataban dos personas por semana y ahora sólo una, de tantos heridos por accidente de tránsito hemos bajado a la mitad, etc. Con ello se evidencia una cuantificación de los objetivos y metas del Estado. La racionalidad instrumental es, de otro lado, una racionalidad formal porque el Estado no puede dotarse a sí mismo de contenidos éticos, o al menos cuando intenta dotarse contenidos éticos, supone que estos son concesiones que la propia sociedad civil le otorga a la manera de un consenso de autodefensa. Véase el debate en Colombia y otros países sobre dos temas cruciales: el aborto y el uso de la dosis privada o personal de droga. Estos ámbitos, desde el punto de vista teórico liberal, son asuntos privados y no públicos mientras que los intereses de terceros no se vean afectados. Adicionalmente otro aspecto destacable en la praxis de la modernización del Estado es el proceso de reformas políticas e institucionales. Su importancia deriva de que el Estado, siendo una red de organizaciones estructuradas bajo normas, principios, jerarquías y procedimientos, puede funcionar en una dirección determinada si se adecua o modifica su diseño, su estructura normativa. De allí la fetichización que tiene en nuestros tiempos la reforma de la Constitución, ya que se la supone como clave o matriz de la acción del Estado, instancia de modernización donde se dota al Estado de sus principios, objetivos y valores (E. Varela: 2005a). Ello explica por qué las sociedades en transición, en crisis, lo primero que hacen es cambiar el fetiche, transformar y colocar una serie de principios de articulación entre el deber ser y la realidad imperfecta. En este sentido se entienden dichos valores; los diseños cibernéticos estructurales de la normatividad legal y supraconstitucional como la base misma que determina la naturaleza y estructura del Estado. No entendiendo que poseen solamente una verdad parcial y formal; no la sustancialidad real del Estado. 287 NOTAS 1 Evidentemente pueden darse medidas típicamente autoritarias que significan la subordinación del poder jurisdiccional al poder ejecutivo a través de un uso excesivo de los poderes discrecionales que otorgan las constituciones cuando se estipulan los estados de excepción. Existen tanto el autoritarismo social como el autoritarismo estatal. Pero el despotismo o el totalitarismo se diferencian del autoritarismo en su carácter absoluto. Es la dominación policial sin ningún resquicio democrático, y sin ninguna posibilidad de que los ciudadanos puedan ejercer derechos, que no sean vulnerados por el Estado. Es un régimen de Estado 288 carente de controles ciudadanos que utiliza sistemáticamente el terror. 2 Recuérdese, en este orden de ideas, cómo el sentido conceptual de la organización burocrática fue precisado en los trabajos teóricos de Max Weber (1985) como una administración racional con arreglo a fines impersonales. Alternativamente la dominación patrimonial puede ser vista como la manera política o pública de extensión del principio autoritario propio del patriarcado. 3 La segunda parte del libro de Robert Nozick, titulada “Más allá del Estado mínimo”, se centra en el problema de la justicia distributiva y desarrolla “la teoría del Estado mínimo”, la cual plantea de la siguiente forma: “El Estado mínimo es el Estado más extenso que puede ser justificado; cualquier otro Estado más extenso viola los derechos de los individuos”. Para demostrarlo, Nozick no tiene más que aplicar rigurosamente la teoría de la mónada moral y su despliegue en el minimal State, de modo tal que cualquier intento de justicia distributiva implicaría una violación de la “igualdad” moral individual que expresan los derechos lockeanos. El verdadero soporte de Nozick es su absolutización del derecho de propiedad como soporte de los demás derechos del individuo; e incluso los más mínimos impuestos redistributivos aparecen como atentados contra la inviolabilidad personal. En consecuencia, el rol del Estado se agota en el papel de “Estado sereno”: hacer respetar los derechos individuales mediante su monopolio legal-moral de la violencia. Respecto de la adquisición original, Nozick se basa en la tesis de Locke según la cual la adquisición original de una cosa es justa si cumple dos condiciones: a) si nadie posee un título de propiedad sobre la misma, b) si el que se la apropia “mezcla” con ella su trabajo, y a esto Locke añade su “proviso”. En éste se establece que pese a cumplir las condiciones señaladas , una apropiación no será legal si no “deja suficiente, y tan bueno, en común para los demás”, de modo que no “empeore” la posición de los otros. 4 Comte avala la argumentación basada en la experimentación lejos del anterior carácter especulativo abstracto. Ésta reemplazará las anteriores explicaciones de la religión y la metafísica. Así en la Edad Moderna aparece una ciencia positiva que no pretende explicar los fenómenos últimos del mundo, que renuncia a toda explicación trascendental. En lugar de esto, pretende explicar los procesos con base en el principio de causalidad aplicándolo a contextos empíricos. De este modo el argumento comteano, recogiendo esta herencia, significa la percepción del progreso en una perspectiva utilitaria. 5 “Antes de la Era Moderna había también ciudad, reinos e imperios y hasta demarcaciones políticas; sin embargo, sólo llegaron a ser Estados estas demarcaciones anteriores con la llegada de burocracias sistemáticas y profesionales que hablaban también el mismo idioma”. cfr. J Kurth, Ob. Cit. 6 Uno de los rasgos más significativos de la obra de Maquiavelo reside precisamente en su planteamiento crítico frente a los mercenarios. El Príncipe alerta contra dichas fuerzas y plantea como una de las tareas fundamentales a cargo del gobernante, la de conformar un ejército o una milicia nacional que tenga un sentido patriótico, que esté anudado al principio de la virtud, que no reciba una paga simplemente por la acción bélica, sino que sea parte sustancial del Estado. 7 En un destacado texto sobre el estudio del origen del parlamento B. De 289 Jouvenel explica cómo el Parlamento surgió como un instrumento de mediación entre los señores feudales, los jefes de las áreas regionales, respecto al monarca, mediante un procedimiento en el cual acordaban un programa de gastos y un compromiso político voluntario del rey con los caudillos señoriales y la nobleza. 8 El desajuste de las relaciones sociales de producción frente al incesante crecimiento de las fuerzas productivas llevó a Marx a pronosticar que el mundo tendería a ser cada vez más socializado en términos del proceso productivo. Esto demandaría, por consiguiente, una correspondiente socialización del poder político y la eliminación de la propiedad privada. Contraposición entre lo público y lo privado sobre la base del predominio de lo público. 9 Los anarquistas igualmente consideran que el progreso, la marcha hacia lo mejor, se puede dar pero tratando de eliminar el obstáculo fundamental para el despotismo, cual es el Estado, la dictadura política; la escisión entre gobernantes y gobernados. Por lo tanto aquí se inscriben todas las utopías acerca de comunas y demás formas de convivencia no estatales o prepolíticas que serían alternativas a la relación política de la sociedad moderna. 10 Obviamente esta salvedad no es meramente teórica, pues resume la experiencia mercantilista de las potencias europeas. Recordemos cómo, en el marco del mercantilismo, se desarrolló una vasta reglamentación estatal, la cual estableció parámetros y reglas del juego a la concurrencia comercial y a los mismos procesos productivos. Véase en dicha dirección el caso de las grandes compañías comerciales. Éstas, en general, usufructuaban franquicias para ejercer derechos, traficar con esclavos, comprar especias, exportar oro y, en fin, realizar mil y una actividades pagando tributos, gravámenes y participaciones a los Estados nacionales. Y de otro lado éstos expidieron, normaron la competencia a través de la fijación de criterios para evaluar los niveles de aceptabilidad de compromisos crediticios, sanciones de carácter judicial, etc. 11 Tómese en cuenta cómo Marx creía que la economía clásica – particularmente en David Ricardo, que desarrolla muchos de los temas de Smith y hace una elaboración más sofisticada– se había equivocado al plantear que el problema estaba en la distribución. Al contrario Marx afirma (lo que es tan conocido por todos como una suerte de moneda común) que el problema está en la producción. Las relaciones de explotación a partir de la plusvalía o del sobretrabajo son la clave de la inequidad de la distribución de los ingresos o de la riqueza en la perspectiva marxista. Marx además criticaba a Adam Smith, y en general a los denominados economistas “clásicos”, por postular como naturales características meramente socio históricas la división del trabajo, la existencia del mercado, el régimen de competencia; todos esos elementos que Smith concebía como naturales, según Marx, son expresión de una ideología mistificatoria que eterniza el carácter transitorio del capitalismo. Se trata, en suma, de un doble ataque: una crítica a la naturalización y otra a la universalización de relaciones de corte histórico o específico, como las propias al régimen capitalista. 12 Pongamos un ejemplo muy simple: las actividades de servicios, tales como llevar cuentas, inventariar facturas, desarrollar ciertos trámites, se especializan de acuerdo con cada oficio o subdivisión, pero al mismo tiempo, se homogenizan. 290 Con el mismo programa de computador o las mismas bases de datos puede una persona indistintamente pasar de un oficio al otro porque el oficio simplemente está expresando una formalidad, pero en el fondo mantiene una misma sustancialidad. Incluso el consumidor especializado por la lógica misma del mercado, tiende a universalizarse. 13 Veamos lo que sobre este tema dice la profesora española Teresa Oñate: “El programa de la metafísica, ciencia técnica, no es otro que el de dominar el mundo”, expresión redundante ya que tal apropiarse de la realidad se lleva a cabo, en este caso, instituyendo un cosmos y ella recuerda que la palabra Kosmos significa orden. El surgimiento del cosmos, ustedes recuerdan, según las leyendas de la mitología griega que le enseñan a uno en el colegio, se hace porque según las leyendas había otro mundo preexistente al cosmos, que era el caos. El caos es reemplazado por el cosmos, el principio de la diversidad, y la heteronomía es un principio de ordenamiento de una situación preexistente. 14 Según M. Foucault (2004) el saber es un instrumento por excelencia de dominio; se conoce para dominar. Expresión que destacaba ya la crítica célebre de Foucault sobre la relación entre saber y poder. Este autor fue muy perspicaz al entender que toda relación discursiva es una relación de dominio, y que la diferencia entre el poder social y el poder político es una diferencia meramente esquemática y externa a la naturaleza misma del poder; ya que el poder en sí mismo tiene un carácter discursivo. 15 Recuérdese adicionalmente que la citada expresión tuvo su origen hace más de medio siglo, en el contexto de las artes plásticas y la arquitectura. 16 Precisamente en momentos en que Taylor pronuncia su discurso, está en marcha el primer gran plan quinquenal de Stalin, quien había llegando al poder cuando Lenin enfermó en 1923 y pudo derrotar a sus rivales internos, particularmente a Trotsky, Bujarin y Kamenev, e imponer su esquema de socialismo de Estado. 17 Este episodio del plan quinquenal tuvo un profundo impacto en el mundo entero. Incluso fue celebrado por la literatura rusa en una serie de obras; la más conocida de ellas bajo la autoría del escritor Mijail Sholojov. Vease su relato épico El don apacible que describe la rebelión de los cosacos y las minorías no rusas contra la imposición de la planificación estatal. 18 Incluso los propios teóricos marxistas reconocían la existencia de variantes socialistas muy disímiles, entre ellas feudales y cristianas (Marx, 1955, PP 39 SS). 19 Desde luego no pretendo negar los intentos por resolver este problema en formulaciones diversas de analistas latinoamericanos. Cfr la compilación Colombia el despertar de la Modernidad, los ensayos de Garcia Canclini, y otros. 291 PÁGINA EN BLANCO EN LA EDICIÓN IMPRESA Capítulo 6 REFORMA DEL ESTADO Y REDEFINICIÓN DEL CAMPO DE LAS POLÍTICAS PÚBLICAS. La experiencia reciente bajo el influjo del Globalismo Neoliberal 1. Management y gestión pública En el ámbito de las ciencias administrativas resurge con cierta frecuencia el debate acerca de la diferente naturaleza de la gestión pública respecto de la gestión privada. En particular se ha reavivado la polémica y se ha puesto en cuestión la legitimidad del campo de las políticas públicas en la medida en que algunos autores consideran dominante la tendencia hacia la integralidad de la gestión de las organizaciones, independientemente de su naturaleza pública o privada (B. Bozeman, 1998). Fenómeno acentuado por la adopción en la gestión pública de los procedimientos y teorías provenientes del campo privado. Este es el caso de la reingeniería, la calidad total, el justo a tiempo, etc., conceptualizaciones extendidas desde el espacio “natural” de las organizaciones empresariales privadas hacia la gestión de las organizaciones sin ánimo de lucro y las instituciones públicas (Donahue: 1.991; Osborne y Gaebler: 1.992). En el caso del Management, este nivel reviste una gran complejidad por el carácter multi y transdisciplinario de la gestión. En general, en mi pers- 293 pectiva, esta cuestión se presenta bajo la forma de un gran debate referido a la naturaleza humana y la discusión sobre las articulaciones entre sociedad y organizaciones. En el primero de los puntos señalados, el discurso dominante intenta la hegemonía del homo economicus, con sus corolarios: el individualismo metodológico y el utilitarismo. Autores centrales y caros a la tradición humanística como Weber, Freud, Veblen, Simon, etc., refuerzan con sus aportaciones teóricas esta tradición hoy hegemónica. Aunque alternativamente las visiones provenientes del hegelianismo y el marxismo, de la teoría social de Durkheim, de la teoría crítica de la sociedad (Escuela de Frankurt) y de ciertas vertientes del estructuralismo, entre otras, significan un polo alternativo que enfatiza las determinaciones sociales, las megatendencias objetivas que no resultan de la mera suma de las acciones individuales. La discusión ontológica se referencia por lo tanto en función de los principios alternativos de preeminencia genérica, el individuo o la sociedad. La teoría organizacional, en cierto sentido, aparece aquí como intentando una mediación, la unidad artificial de la vida organizacional, que sería una suerte de prisma integrador de los elementos dicotómicos anotados. Sin embargo, la adscripción de las ciencias manageriales al ámbito ontológico de lo organizacional, pese a las tendencias clásicas del mismo (Barnard, Homans, Selznick y otros) no es afirmada consensualmente y éste se ve prioritariamente como una rama o un brazo del economicismo utilitario de racionalidad limitada. Además este discurso sobre la unificación del método y del objeto se recubre con argumentos cientificistas. En efecto, al leer la crítica de Genereux (2001), a mi juicio fundamental, ésta nos recuerda las expresiones que Marx formuló en su momento contra las pretensiones de la propia economía política clásica (Smith, Ricardo, Say, Bastiat et al) de construir leyes de tipo ahistórico válidas en todas las condiciones y situaciones, sobre la base de una hipostatización de la llamada condición o naturaleza humana. En efecto, en la nueva versión de la economía neoliberal, aparece de nuevo esta mistificación (Friedman, 1983). Quizás cabe agregar que en la teoría económica actual se presenta un uso —metodológico y expositivo— abusivo del citado naturalismo. Este pretendido cientificismo usa como cortina de humo el lenguaje y los modelos 294 primordialmente provenientes de las ciencias físicas. De otro lado, adicionalmente, se introduce un cientificismo formalista basado en la matematización del discurso. Lo que hoy en día es dominante en las escuelas de Management (cfr. los MBA y las críticas de Mintzberg y otros). Sin embargo, la pretensión cientifista u objetivista de las Ciencias del Management es en general eso: sólo una pretensión. Véase al respecto la aguda crítica del profesor Aktouf al llamado método de Casos, base de la enseñanza en la mayoría de las escuelas de gestión y que, como lo ha retomado recientemente el propio H. Mintzberg (2004), no contribuyen en nada ni a formular teorías rigurosas o a mejorar la capacidad de gestión de los egresados. En este sentido resulta de enorme importancia establecer las correlaciones fundamentales entre ética y Management. El cientifismo y el formalismo positivista de los últimos dos siglos, rompieron precisamente las conexiones entre las grandes interrogaciones éticas y las prácticas sociales humanas, vistas estas últimas en términos puramente utilitarios e instrumentales. Pero la propia crisis del campo managerial ha obligado de cierta manera a intentar restablecer esta conexión. Fijémonos en el auge repentino del business ethics en las escuelas americanas y por reflejo en su enorme esfera de influencia mundial —una moda más que ha llegado con fuerza en América Latina—. Sin embargo, no se trata sólo de algo cosmético o superficial para fijar la coartada de que se enseña éticamente, lo que después no se practica salvando la responsabilidad de las grandes escuelas de negocios respecto de los desastrosos resultados de la gestión en buena parte de las grandes corporaciones y megaorganizaciones de nuestros tiempos. Desde luego que muchas de tales escuelas de negocios salvan su responsabilidad en la ejecutoría de sus egresados, cuando éstos dejan de lado cualquier consideración, salvo la maximización del lucro. Pero la reemergencia del debate ético deviene principalmente de la consciencia crítica de las sociedades civiles y de las fuerzas sociales afectadas por la ruptura de los principios éticos:a) los trabajadores, empleados y cuadros manageriales; b) los consumidores y usuarios de los bienes y servicios; y c) los ciudadanos particularmente enfrentados a las acciones de la administración pública, etc. En nuestro mundo globalizado, estas fuerzas se unifican transnacionalmente ofreciendo su accionar, la otra cara de la mundialización. Las redes 295 contra las guerras, el hambre y la pobreza extrema y contra los abusos al medio natural se crean hoy igualmente en el terreno político como críticas y acciones contestatarias frente a las formas asimétricas y salvajes de la gestión de los negocios del capital mundial (Hardt y Negri, 2004). De tal manera que más allá de la cosmética managerial del business ethics, se despliegan una serie de reflexiones críticas (Aktouf, 2002; Lafontaine 2004; Minztberg, 2004). Esto hace necesario delimitar de forma precisa y rigurosa el campo propio de las políticas públicas frente a la teoría general de la gestión. En el caso del Management, el carácter altamente utilitario tanto de la profesión como de la disciplina, expresa una separación que Dery describe acertadamente entre lo que podríamos llamar el espacio múltiple del mundo de la vida (Husserl) y el mundo del trabajo (Habermas). Dicotomía que se agudiza en tanto que a un sector creciente de nuestras sociedades se le arrebata el acceso a este último. Esto nos conduce, por supuesto, al cuarto de los grandes marcos de referencia descritos en el seminario: el espacio de la praxeologia. El conocimiento aún en sus niveles más abstractos, el saber filosófico, la lógica y las matemáticas, tienen la virtud de servir de marco de actuación, de clivaje referencial y constitutivo de las diversas praxis sociales. En sus distintos grados de complejidad, el pensamiento abstracto se articula funcionalmente con la actividad creativa de los seres humanos, por su esfuerzo por la sobrevivencia, el dominio de la naturaleza, y además teje los conceptos necesarios para la operacionalidad del gobierno de las sociedades. Es el lugar del uso del poder entre los seres humanos que abre el mundo de nuestras experiencias al campo de la política, el impulso vital del zoon politikon, el mundo artificial de las relaciones comunicacionales y simbólicas que tan bien anticipó Aristóteles. Esa distinción del ámbito de la política, los griegos la fundaron sobre la dicotomía entre lo público y lo privado asumiendo la superioridad del primer espacio, la condición preeminente de la ciudadanía. Y dejando en las sombras de lo privado precisamente a la praxis económica, al oikos nomos, que fue en el mundo antiguo lo más parecido a las ideas y a las prácticas manageriales en tanto éstas expresan hoy el lugar ontológico de lo privado. Lo público es hoy el Estado, como instancia separada del mercado, en tanto en el mundo griego a lo público se lo identificaba con la polis, con la preeminencia de la deliberación de los ciudadanos. Allí la po296 lítica expresaba claramente su superioridad ética y ontológica sobre la economía, lo que no representa la tendencia dominante en nuestras sociedades postmodernas, exaltatorias de lo privado, del individualismo, del éxito material y económico. 2. Los avatares y retrocesos en el cambio de modelos de gestión pública Una vez precisadas sumariamente estas consideraciones previas sobre el Management y su articulación al campo de las políticas, tomemos en consideración los acelerados cambios presentados desde la segunda mitad del siglo XX, que han replanteado los roles tradicionales de las organizaciones en la región latinoamericana, en los espacios de lo público y lo privado. La reforma del Estado que sacudió y aún impacta a América Latina, ha sido frecuentemente percibida en forma negativa y crítica por parte de vastos sectores de la opinión pública que la consideraron, o bien una moda pasajera como tantas otras que a menudo aparecen en el campo de la gerencia (reingeniería, calidad total, justo a tiempo, benchmarking, etc.), o alternativamente, como resultado o fruto exclusivo de la imposición de los organismos multilaterales de crédito y de los países industrializados (Medellín, 1991). Efectivamente sectores de la izquierda, el sindicalismo y muchos de los partidos y ciertos sectores de las elites vinculadas a los poderes tradicionales, apuntan a que este proceso es sólo una manera de ajustar las economías regionales para pagar la deuda externa y satisfacer los requerimientos de la apertura y la globalización. En fin de cuentas, estos enfoques críticos coinciden en que las reformas de los aparatos públicos latinoamericanos no serían necesarias. Su razón de ser no consultaría los legítimos intereses de los pueblos latinoamericanos. Pero si se observa el panorama político latinoamericano, en los últimos lustros, se encuentra que bajo unos patrones muy parecidos, todos los países de América Latina han enfrentado procesos semejantes de reevaluación de las funciones de los sectores público y privado1. El fenómeno es global y ello constituye un primer dato para comprobar que las anteriores explicaciones, aunque sean ciertas, son insuficientes. El repensamiento del papel del Estado obedece a causas más profundas, de tipo estructural; se explica por la necesidad de encontrar correctivos 297 que rectifiquen, que ajusten los fallos de un modelo de administración pública desgastado y en crisis. En realidad las reformas liberales fueron propulsadas y catalizadas por el fracaso estrepitoso que, en el final del siglo pasado, le correspondió a aquella política pública que pretendió reemplazar los indicadores de eficacia y eficiencia que da el mercado a través de los sistemas abiertos de formación de precios y asignación de recursos (humanos, de capital, etc.). Al contrario, se impuso una solución centralista y autoritaria a problemas endémicos de las sociedades de mercado como el desempleo o las inequidades extremas, adoptando medidas simplemente políticas y voluntaristas, propias de los modelos más radicales del socialismo de estirpe leninista o a las variantes extremas del Estado de Bienestar. Generalmente la patología de tales modelos la dio el fenómeno de desbordada corrupción compartido por las elites políticas, militares y una parte no desdeñable del gran empresariado articulado a las multinacionales. El desgaste del modelo, sin embargo, visto el proceso retrospectivamente, no dio las alternativas de depuración de la política ni implicó las mejorías de la gestión pública que muchos esperamos. Aunque es cierto que en los años recientes a nivel mundial prácticamente ya no se discute acerca del desmoronamiento de ese viejo modelo de Estado asistencialista y benefactor, sino la medida en la que ese desmonte debe darse; y qué tipo de políticas deben agenciar la filosofía del interés público. La discusión se convierte así en procedimental: cómo desmontar partes del Estado y devolverle funciones al ámbito privado; definir qué ramas o ámbitos de la vida social y económica del Estado contemporáneo deben dejar de corresponderle. Qué actividades deben ser conservadas en manos del Estado, incluyendo una parte de las funciones empresariales del mismo, etc. Obviamente, como ya lo vimos, no hay, al menos exceptuando algunas teorías bastante radicales, una ideología política que cimente la destrucción total de las funciones públicas por parte del Estado. En el fondo hoy sabemos que necesitamos tanto del Estado como del mercado. Y el asunto es cómo combinar estos dos componentes en la política. El definir qué funciones de regulación son propiedad inherente del aparato del Estado y cuál es el grado deseable o conveniente de mercantilización de lo público (E. Varela, 2005), son los grandes interrogantes de las políticas públicas a comienzos del siglo XXI. 298 Respecto al conjunto de campos en los cuales el Estado ha venido a lo largo de los decenios anteriores asumiendo un rol de control, tutelaje, o ha asumido directamente funciones tecnoproductivas, la discusión sobre su reconversión ha sido muy aguda. Al punto que ella ha reemplazado en gran medida la vieja oposición entre izquierda revolucionaria y defensores del status quo. En este sentido puede observarse el caso colombiano, particularmente el proceso de la reforma del Estado iniciado desde finales de la administración Barco en los años 80 del siglo pasado y que con zigzag y avatares, continuó en las demás administraciones colombianas, acentuándose en el actual gobierno de corte abiertamente neoliberal de Álvaro Uribe Vélez. En Colombia el debate ha escindido, por vez primera en muchos años, al propio partido tradicionalmente mayoritario, el liberalismo, como lo hemos visto en recientes foros ideológicos, y se vislumbra en las contiendas políticas en el terreno electoral. La posible reelección del actual presidente Uribe Vélez ha concitado por vez primera en muchos una polarización de clara diferenciación ideológica entre los partidarios de la profundización del recetario neoliberal y los que buscan una opción alternativa, aunque esta última luzca confusa o falta de consensos significativos en términos de opciones alternativas y no sólo simplemente contestarias. Ciertamente el proceso de modernización del Estado en América Latina ha implicado, en términos relativos, una emancipación de la sociedad civil respecto al centenario yugo del autoritarismo centralista que caracterizó nuestro modelo de desarrollo; además se vislumbra en dicha óptica cómo éste, en buena parte, constituyó una mixtura entre una tradición patrimonial y centralista de origen colonial. En segundo lugar, es conveniente analizar la manera como devino el específico proceso de conformación de las repúblicas liberales en el siglo XIX. Una de las características del proceso de democratización en América Latina ha implicado reforzar los sistemas de control social de la comunidad hacia el Estado y afianzar el tradicional modelo de frenos y contrapesos entre las distintas ramas del poder público (F. Leal, 1989). De otro lado, en contraposición a las anteriores tendencias, igualmente es comprobable en el contexto latinoamericano que el papel del presidencialismo o el cesarismo providencial no ha desaparecido. Desde luego ésta no es una tendencia exclusivamente latinoamericana: el rol del Estado y de 299 las relaciones del gobierno con la sociedad civil y con el mundo empresarial, los cambios en el papel del Estado, etc., son fenómenos que revisten hoy en día el carácter de una tendencia mundial. Son principios que se reflejan como fundamentos del cambio. En la base de las presentes trasformaciones y desafíos de la política pública se halla un nuevo espíritu de recuperación del papel del gobierno tanto en Estados Unidos como en Europa. En este último caso, incluso la socialdemocracia, ha jugado un papel relevante en la implementación del cambio valorando adecuadamente el rol del mercado, la integración social y económica, las privatizaciones y, por último, intentando combinarlas con la preservación de los principios básicos del Estado social de derecho. Indudablemente nos encontramos aquí ante una profunda transición en los esquemas de gestión de la relación entre lo privado y lo público. En tal dirección los socialistas europeos se encuentran ideológicamente distantes de sus antecesores de hace dos o tres generaciones. Muchos de los viejos principios y ortodoxias se han dejado a un lado. Se reconoce abiertamente que un gobierno socialista que no atienda las señales del mercado va a generar no sólo ineficiencia sino además corrupción (cfr. Collomb y otros: 1994). Pero al mismo tiempo, el balance de la última década, ha morigerado las ilusiones iniciales sobre una retirada general del Estado, dado que las fuerzas del mercado libradas a su suerte son fuente de grandes exclusiones y enormes inequidades. La tendencia dentro de la reforma del Estado principalmente enfatizó un mayor grado de autonomía del empresariado público. Empero el balance de tal autonomía ha sido altamente negativo (Stigliz, 2003). En el modelo colombiano, se ha venido dando un grado relativamente alto de competencia regulada con controles desde arriba tanto a nivel de la planificación, la fijación de los objetivos y metas sectoriales, como en cuanto a los criterios técnicos para la evaluación del desempeño y análisis de los determinantes de los comportamientos de ejecutivos de alto nivel, lo cual se vislumbra en la creación de una multitud de superintendencias o agencias gubernamentales que tienen a su cargo el control de dicha competencia regulada, principalmente para el sector público, pero también para los prestatarios privados. Sin embargo, las dinámicas patrimoniales y la corrupción no se han visto significativamente desplazadas por la liberalización. Los indicadores de equidad tampoco han mejorado. Ciertamente el aparato de 300 frenos y controles funciona formalmente, pero en una gran medida lo hace de manera tritualista e inefectiva. En el caso de los entes territoriales descentralizados, lo que se ha percibido en los últimos tiempos es un predominio de lo político sobre lo gestionario y ésto se ha visto generalmente reforzado. Estas expresiones poseen en general, sin embargo, un alto grado, comparado con el pasado, de especialización profesional y gerencial, principalmente en las áreas más desarrolladas del país, en los sectores denominados de punta. Podríamos sintetizar lo dicho enfatizando que la reforma del Estado es un proceso mundial, hace parte de la globalización, expresa una poderosa tendencia de cambios en las relaciones sociales, económicas y culturales del mundo contemporáneo. Los procesos de modernización de los aparatos públicos y privados, en especial si nos referimos a los nuevos paradigmas en el terreno de las teorías organizacionales, no implican, desde luego, la disolución de las fronteras entre lo público y lo privado. Por tanto, vale la pena precisar, en este orden de ideas, las diferencias básicas entre estas dos instancias en un terreno que precise las diversas lógicas de la gestión (R. Dombois y N. Pries, 2000). 3. Principales diferencias entre lo público y lo privado En primer lugar, debe señalarse la diferente naturaleza del régimen de propiedad público respecto del privado. En el último caso los propietarios privados son responsables ante sí mismos, mientras que en la propiedad pública los agentes organizacionales, funcionarios, gobernantes, actúan en una relación de medios-fines; le deben responder a la comunidad, en términos de su legitimidad en relación con los objetivos de política pública que establecen el norte, la misión, la visión, de las organizaciones públicas en donde actúan como responsables y ejecutores (recuérdese que la política pública se refiere al Estado de acción). Una segunda brecha sustancial, se desprende de la característica diferenciada que en ambas esferas asume la apropiación de los beneficios. Resulta meridianamente claro el propósito de la empresa privada: la obtención legítima de ganancias, la maximización del lucro, bien sea primordialmente material, dinerario o, alternativamente, como lo había expresado ya hace 301 muchos años Max Weber, algún beneficio intangible en términos de status y prestigio por parte de los propietarios de las organizaciones y empresas de tipo privado. La empresa privada en su acepción más extendida, supone como condición de posibilidad el mercado público (cfr. Adam Smith, El Contexto Mercantilista, ob cit). La diferencia entre lo público y lo privado no puede equivaler al dilema Estado versus mercado, como usualmente se cree. El espacio constitutivo, como ya lo vimos en capítulos anteriores, de la llamada sociedad civil o el sistema hegeliano de las necesidades —compuesto de intercambios recíprocos, interdependencias y mutua satisfacción entre los miembros individual o particularmente considerados de la comunidad de vida de una sociedad determinada— es, al menos en la época moderna, el del mercado. La división del trabajo, la especialización, los referentes nacionales e internacionales, la segmentación, fragmentación y especialización de dicho mercado, habían sido estudiados con clarividencia por los llamados economistas clásicos o políticos incluyendo, desde luego, a Marx. Lo privado en este sentido, no se opone a lo público; entendiendo por este término la aserción de publicidad, sociabilidad macro; se opone a lo público articulando primordialmente con su referente gubernamental. Es decir, entendiendo las mediaciones políticas que provienen de las determinaciones del Estado, los sistemas legales y papel de arbitraje (incluso del favorecimiento, restricciones, protección, concesiones) que el Estado define como reglas del juego de un mercado determinado. Igualmente, cuando el Estado actúa como empresario se desdobla, en un ámbito mercantil propiamente dicho, la provisión de bienes y servicios, en su inicial y ya señalada función de mediación y arbitraje. En este orden de ideas diferenciaría a lo público de lo privado, como lo han aclarado los economistas a partir de Allen Buchanan (1985), la función de provisión de bienes públicos por parte de este último sector y/o su intervención para garantizar la naturaleza de la distribución de dichos bienes públicos, sean ellos de consumo o apropiación exclusiva o universal. Lo que resaltaría de dicha intervención, sería la contradicción (aparentemente oponiendo privado contra público) entre la competencia “natural”, en donde los individuos, propietarios de sí mismos —según el discurso liberal posesivo— participan en la conformación de la sociedad civil a través de una mutua reciprocidad egoísta e involuntaria; campo este por excelencia del derecho natural, de la libertad natural, frente a un universo, el de la sociedad política, donde predominan las restricciones artificiales impuestas por el Estado (sea éste un Leviatán omnímodo o un Estado liberal demogarantista). En la literatura política permanece, no obstante, una gran confusión respecto a qué se considera función pública. Y esto primordialmente cuando aquella no es asumida directamente por el Estado (este es el caso sinuoso de la dificultad para delimitar cuándo se permiten huelgas, reconocidas en diversas constituciones colombianas incluyendo la actual, pero restringidas al campo de los servicios privados). ¿Cuáles de los servicios que se otorgan a través del mercado con base en la relación oferta-demanda, serían públicos? Si se analizan las actividades financieras, el suministro de alimentos, el transporte en sus diversas modalidades, los servicios domiciliarios (energía, acueducto, alcantarillado, telefonía, televisión, etc.) en una dimensión —la del acceso y la teórica universalidad— todos son, o al menos deberían ser, servicios y bienes públicos. El transporte, por ejemplo, así sea operado empresarialmente, sigue siendo parte del ejercicio de un derecho público esencial: la libertad de locomoción por los territorios, componente esencial de la autonomía irrenunciable de los seres humanos. La información, a su turno, es una garantía y a la vez una función pública que se suministra mayoritariamente en las sociedades contemporáneas casi de forma absoluta por organizaciones empresariales privadas. La privatización se sitúa, por regla general, en tres grandes campos: Primero, la venta de los activos públicos de empresas y organizaciones gubernamentales a compradores, consorcios y grupos empresariales nacionales y extranjeros. Este es estrictamente el significado etimológico de la expresión privatización. En un segundo plano, caracterizaríamos como procesos de privatización, la apertura a la competencia donde existe monopolio público, de tal modo que otros grupos empresariales y organizaciones que proveen bienes y servicios participen en el mercado. Se trataría, en cierta manera, de una desnacionalización o desestatización de determinadas actividades para que la ley de la oferta y la demanda asigne racionalmente los recursos y, de esta manera, adicionalmente el Estado pueda ubicar su inversión en otros campos en donde exista prioridad, desde el punto de vista de la política pública (inversión social, gasto educativo, inversiones en infraestructura etc.). En tercer lugar, mediante mecanismos de descentralización territorial, aunque no necesariamente debe ser entendida, como usualmente se lo hace en Colombia, como una relación entre el Estado central y los territorios y regiones (departamentos, municipios, distritos, etc.). La palabra descentralización implica pérdida o redistribución del poder de decisión desde un centro, hacia una pluralidad, multitud de ámbitos y lugares donde se toman decisiones sin necesidad de un control de tutela (M. Urrutia, 1991). En esta dirección la competencia, como muy bien lo explican los economistas contemporáneos, es sin duda la máxima forma en que se expresa la descentralización. Contrario a ello aparecen los carteles, oligopolios y los monopolios públicos y privados. Y, finalmente, una cuarta modalidad de la privatización ha sido la autogestión, proceso mediante el cual se ha producido la devolución de una serie de atributos que el Estado, como ya lo mencioné en otros capítulos de este libro, se ha apropiado para sí; campos de acción que serían delegados o devueltos al sector empresarial privado o a las llamadas organizaciones sin ánimo de lucro, no gubernamentales y a las comunidades autogestionarias. De otro lado, pese a la amplia gama de opciones que ofrece la privatización, es indudable que dicho concepto ha sido satanizado por parte de un sector mayoritario de los académicos en América Latina. En el otro polo de la opinión, quienes son sus partidarios la suponen como un sinónimo de eficiencia, eficacia y una panacea. Pero ambos extremos simplifican en demasía la naturaleza compleja del problema. Ciertamente un sentido positivo de la privatización o la reconversión de las relaciones entre la comunidad y el Estado, es la devolución, la recuperación de esferas que pertenecían a ésta, a sus organizaciones civiles, culturales y empresariales, a la propia acción de los individuos, en campos tan diversos como la cultura, las artes, la vida familiar, la actividad económica, etc., y que le fueron arrebatados por el intervencionismo del Estado patrimonial (Cárdenas, 1991). En realidad, el fondo central del debate reside en el carácter acusadamente político en la gestión de las organizaciones públicas. Es imposible, en este sentido, que pueda pensarse en una transformación de las funciones públicas 304 que las lleve a un nivel de despolitización equivalente al proceso administrativo privado, en donde, aunque existen consideraciones sobre el entorno político, la legislación, principalmente en relación con definiciones, restricciones y/o favorecimientos por parte de los agentes públicos gubernamentales; en últimas, allí el resorte de las decisiones corresponde a la lógica, libremente definida, de la empresas u organización privadas. Valga decir, la política es prioritariamente definida por las instancias del staff respectivo: la junta de accionistas, las atribuciones que correspondan a las juntas directivas y a los equipos de Management o en virtud de la diferente naturaleza de la organización jurídica de la empresa (sociedad accionaria, comanditaria, propiedad familiar, etc.) (A. Przeworski, 1991). En este sentido la disposición de las ganancias privadas aparece como decisión autónoma de los propietarios referida a criterios de racionalidad económica, primordialmente en términos de la reposición del capital, las tasas de rentabilidad esperada, las percepciones de propietarios y administradores sobre las condiciones del negocio en el mercado nacional e internacional. En cuanto a las organizaciones públicas debe precisarse que no existe una definición que les impida, en ciertos casos, tener lucro o ganancias. Un buen gerente o directivo de las empresas comerciales e industriales que pertenecen total o parcialmente al Estado posee deberes inherentes a su función, y ellos incluyen tanto el servicio al público —velar por los intereses generales de la sociedad— como, desde luego, maximizar la eficacia y la eficiencia de la organización pública a su cargo. Sin embargo, hay serias reservas por parte de los teóricos de la gerencia respecto a los alcances de la racionalidad administrativa y gerencial en la toma de decisiones, particularmente en lo concerniente a las macro-organizaciones. En este sentido, no se trata de un caso específico del aparato público, sino que también se suponen dificultades semejantes en el manejo de grandes empresas privadas, tales como corporaciones multinacionales, en las cuales una serie de factores de distinto tipo, limita la capacidad de acción por parte de los planeadores de política. En realidad difícilmente éstos gozan del poder que generalmente se les atribuye en el proceso efectivo de decisiones y de implementación de sus políticas. 305 4. Dinámicas recientes de la relación público y privado En nuestro continente, después de la Segunda Guerra Mundial, se privilegió un modelo de desarrollo con base en el reforzamiento del rol del Estado. En esa edad de oro del intervencionismo, al cual aludíamos líneas arriba, ello se concretó en varias áreas, principalmente: a) La política de sustitución de importaciones buscando crear una base tecno-productiva propia, proceso gradual donde primero se arraigaría la industria de bienes de consumo, luego el sector de servicios y finalmente el de bienes de capital. b) La conformación de una vasta infraestructura para lo cual el Banco Mundial y otras entidades de tipo multilateral, a partir del decenio de los años 50, enfatizaron los megaproyectos de comunicaciones, la producción de energía, la duración básica en salud y educación técnica. c) Finalmente la ampliación de los mercados interiores aplicando políticas de ingresos y salarios. Ello estuvo vinculado, en mayor o menor medida, al populismo y a las negociaciones corporativas de la propia política pública concertada con sus directos beneficiarios. En el caso particular de Colombia se continuó, de este modo, una política comenzada desde finales del siglo XIX por la oligarquía republicana, la cual pretendía —con visos claramente positivistas— impulsar la configuración de un mercado interno. Cabe destacar cómo muchas de las actividades productivas y de servicios estuvieron a cargo del sector privado (por ejemplo, la prestación de los servicios de energía eléctrica, acueducto y alcantarillados, los puestos, etc.). Posteriormente, cuando se empezó a cimentar de manera nítida la constitución progresiva de una base industrial, ésta se encaminó fundamentalmente hacia los bienes de consumo. Fue al mismo tiempo un período en el que se vivió una significativa exportación de capitales hacia América Latina, provenientes de las metrópolis más desarrolladas de Europa y Norteamérica. En Colombia, en particular, este modelo Cepalino tuvo un antecedente cercano en la experiencia de gobierno por parte de la república liberal (1930-1946). Había en las reformas que se emprendieron en el primer gobierno de la administración 306 de López Pumarejo (1934-1938), una línea que anticipó el modelo Cepalino de los años 50, en el sentido de fortalecer el papel del Estado en la actividad económica y esbozar su rol en la planificación del desarrollo. Más adelante, en el decenio de los años cincuenta, cuando surgió un clima de violencia y confrontación armada anarquizada, fue también, y ello es una paradoja, un período de gran inversión extranjera. La modernización del Estado en Colombia se cimienta entonces en un proceso doble: de un lado, en el esfuerzo por dotar al Estado de instrumentos gerenciales de corte racional-burocrático, principalmente introduciendo el principio de la planeación, ya que ésta es una fase en la que se sentaron las premisas de la planificación o la intervención racional e intencional del Estado como agente de cambio, de lo cual son claros ejemplos trabajos pioneros como el informe de la Misión Lebret (Lebret: 1985) y la Operación Colombia del Doctor Lauchin Currie (Currie y otros: 1993). Efectivamente, en los años cincuenta, y luego hacia el final de la década con el comienzo del Frente Nacional, se incrementó en nuestro país la planeación institucional y los planes de desarrollo. De otro lado, se acentuó el rol empresarial del propio Estado. Este es el caso durante dicha década de las telecomunicaciones, el petróleo, los puertos, parcialmente la banca, las empresas de servicios públicos domiciliarios. Nuevas ramas de la actividad empresarial público-estatal añadidas al manejo directo que venía dándose de la gestión ferroviaria y de la infraestructura de las llamadas obras públicas. Con relación al fortalecimiento global del papel del Estado, se trataba de aplicar el modelo del llamado Welfare State o Estado de Bienestar. Es así como el Estado asume una serie de funciones públicas, particularmente en la atención de los servicios educativos, salubridad, vivienda y, posteriormente, en el suministro de otros servicios sociales —básicamente los denominados servicios del Estado empresarial (el Estado propietario de organizaciones empresariales en el sector industrial, de servicio y mercantil)—. Esta tendencia —manifiesta en la política de sustitución de importaciones y la creación de una base industrial nativa— junto con las obras de infraestructura encaminadas a estructurar el mercado interno; y el creciente papel del Estado en sus funciones asistencialistas a través de la extensión en la cobertura de los servicios públicos, de la redistribución de ingresos por 307 vías tributarias con el destino a la inversión social, etc., fue característica de un modelo recortado y limitado del Estado de Bienestar. Mucho se ha discutido respecto a la existencia en Colombia y en otros países de América Latina, del tal Estado de Bienestar. En mi opinión éste si se ha dado, pero de forma diferente al paradigma clásico europeo o al norteamericano y de otros países capitalistas. Baso mi afirmación en el hecho de que el Estado en Colombia y en el resto del continente ha cumplido su papel de redistribución utilizando mecanismos muy característicos de nuestra cultura política, como los esquemas clientelistas y patrimoniales de vinculación con la comunidad. El problema deriva ahora hacia la gradualidad, la velocidad del proceso de cambio, existiendo un relativo consenso sobre los modelos (Hafsi y Jorgensen: 1990) organizacionales que reemplazaron el anterior esquema centralista, burocratizado, y, desde luego, altamente formalizado (Kliskberg: 1989). De otro lado, tomemos en cuenta que vastas esferas de la sociedad civil, a través de sus múltiples actores, han escapado del tutelaje estatal, expresando dicha emancipación, en la tozuda realidad, en las formas de la economía marginal (De Soto, 1989), la socialización del delito, las prácticas de la violencia, etc. Además, ello se ha producido en las formas más civilizadas de desobediencia consciente respecto al Estado, apartándose de sus normas, de su capacidad de control sobre la comunidad de vida. La introducción del nuevo paradigma en la gestión pública colombiana, se dio primordialmente en la parte final del gobierno de Virgilio Barco (19861990) implicando tanto la prevalencia de la apertura económica y la internacionalización de la economía, como —adicionalmente— enfatizando los procesos de cambio de la administración pública en torno de los principios de eficiencia, eficacia y competitividad. Para lo cual profundizó la política aperturista y, hasta cierto punto, desreguladora adoptando el conocido aforismo: “el gobierno no rema, dirige el timón”, y además se suscitó la derogación de la centenaria constitución de 1886 dando lugar a un formidable proceso constituyente que culminó con la redacción de una nueva carta política en 1991. En ésta se acentuaron las tareas y deberes asistenciales y garantistas del Estado social de derecho; mientras que, de otro lado, se concibió la función administrativa pública bajo un esquema de eficiencia, ahorro, celeridad e igualdad, 308 apoyándose en la delegación tanto en el sector no gubernamental como descentralizando funciones y responsabilidades en las regiones y autoridades locales2. Como lo ha anotado Peter Drucker (1993), el nuevo modelo de gestión de los aparatos públicos y privados se basa en la percepción de cambios dramáticos del contexto: el predominio de la sociedad del conocimiento. Y ello le otorga un sentido dialéctico a la decisión de qué dejar en manos del Estado y qué funciones devolver al ámbito privado. Y este no es un debate o una decisión simplista, fácil, pues incluso las funciones fundamentales del Estado, las que predicó el liberalismo clásico —justicia y seguridad— implican, hoy en día, un grado de delegación hacia el sector privado. Las dimensiones del cambio, del rediseño entre los roles público y privado, son principalmente las siguientes: a. Política pública de infraestructura. Se ha definido que la tarea debe ser compartida entre el nivel central y los entes territoriales (municipios y regiones). No obstante, la ejecución de las obras públicas se hace por el sector privado a través de ofertas de licitación, concesión y contratación directa, incluyendo tanto los megaproyectos como la construcción y reparación de las redes vial, ferroviaria, porturaria, aeroportuaria, etc. De otro lado, en el campo de las telecomunicaciones, una infraestructura de tecnología sofisticada se ha avanzado audazmente para recuperar el terreno perdido en la región y en el contexto global. La telefonía celular, la televisión por cable, la eventual privatización de los canales nacionales expresan la tendencia de que la oferta pública en esta materia será en lo fundamental provista por las organizaciones privadas, con ánimo de lucro. b. Políticas públicas de seguridad social. Se está hoy en día en el proceso de implementación de la reforma al sistema de seguridad social y pensional que ha permitido la creación de empresas privadas prestatarias de estos servicios en un esquema de competencia regulada. En el futuro, es previsible que en particular los fondos pensionales adquieran la importancia económica y social que hoy poseen en otras latitudes. En este campo se ha desmonopolizado la función pública. c. Política social. Tanto en la educación como en la salud pública, la vivienda, los servicios culturales, etc. se ha profundizado una tendencia 309 que venía de tiempo atrás: la oferta de dichos bienes y servicios satisfecha por organizaciones particulares en competencia regulada con las organizaciones públicas, y la reubicación del rol del Estado como agente regulador y fiscalizador del interés general. d. Descentralización. Como lo anoté de pasada, éste es un punto conflictivo donde se han presentado zigzag y retrocesos. En particular por la preocupación del gobierno central respecto del control de la eficiencia y calidad del gasto y la inversión pública ejecutado en las regiones y municipios, pero transferido desde el nivel central. En iguales términos, podríamos referirnos frente a la ley de competencia para las entidades territoriales, a la ley sobre regalías, a las normas de control sobre la gestión de los alcaldes y los gobernadores, etc. Tal acierto se demuestra, por ejemplo, en el tema de la paz. En especial si analizamos la estrategia del gobierno Gaviria al señalar que la negociación regional con los grupos alzados en armas era caer en un caballo de Troya, porque la guerrilla tenía un mando altamente centralizado que negociaría en cada una de las regiones, en cada municipio y vereda a cambio de que no hicieran secuestros, no desarrollaran ciertas actividades propias de su naturaleza; los gobiernos locales firmarían una especie de contrato social en donde, a cambio de guardar las armas debajo de la mesa, se entregarían obras, se haría inversión pública y se propiciaría empleo en una política de concertación que generaría simplemente que el Estado le entregase a la guerrilla recursos y la guerrilla los cobraría políticamente como beneficio suyo, pero sin un real proceso de desarme. Evidentemente la estrategia de las organizaciones insurgentes en Colombia busca prolongar el conflicto en un contexto regional, diversificado y frente a autoridades civiles relativamente inermes, como son la mayoría de alcaldes y los gobernadores. Pero la posición del stablishment, al comienzo clara en la década de los noventa, finalmente terminó por agotarse porque el problema en política no es que se pueda decir: “tengo esta posición y me quedo con ella quince años”, puesto que la política es muy dinámica, cambiante, y se debe tener la capacidad de reaccionar frente a nuevas situaciones e interpretar la naturaleza de la sociedad en la que se viven los dramas. Tal es el caso actual del juego de ajedrez entre las partes en torno al tema del llamado “intercambio humanitario” de soldados, policías, civiles y guerrilleros, que se ha prolongado por años, mientras los cautivos 310 sufren las consecuencias. Este caso, el tema de la paz, igualmente se puede asumir en otros campos, las políticas de vivienda, de empleo, de salud, educativa. En Colombia, frecuentemente, cuando se han impulsado procesos de descentralización, los propios alcaldes no han querido “recibir” la educación y rehúsan recibir la salud. Confróntense las denuncias que hizo la Procuraduría respecto al hecho de que más de doscientos alcaldes de los mil que tiene el país habían hecho mal uso de los recursos y las transferencias fiscales, llegando a casos escandalosos como las gobernaciones de Arauca y Casanare, donde los gobernantes han saqueado el erario público, las transferencias fiscales, las regalías. Las comunidades en general, en dichas regiones son poco críticas con sus autoridades locales. Y tal permisividad con sus autoridades locales, acarrea el que se siga manejando el esquema de que el Estado central es el que da los recursos (Estado providencia). Persibe una cultura que tradicionalmente —desde que se implantó el régimen colonial hispánico— supone la pretensión de que los problemas no se resuelven en la esfera local o regional, sino que la solución se define en la esfera nacional a partir de ejecutorías paternalistas y asistencialistas (tal es la base sociocultural del neoclientelismo reeleccionista de los consejos comunitarios del presidente Uribe Vélez). Al respecto recuerdo un relato del escritor peruano José María Arguedas, donde describe cómo una comunidad indígena del Perú en los años 30 decidió —porque el gobierno no tenía la decisión ni la voluntad política— construir por sí misma una carretera para conectar a su pueblo y sacar al mercado sus productos como efectivamente lo hicieron. Una comunidad vecina se dio cuenta del fenómeno y lo empezó a emular; como producto de la competencia, finalmente las dos comunidades accedieron a construir sus carreteras por sí mismas. Poco después llegó el Estado a la región y empezó a construir otra carretera con fondos nacionales. En el momento en que el Estado llegó el espíritu comunitario de resolver los problemas por su propia fuerza desapareció y la gente quedó “reeducada” en el sentido de sólo esperar la atención del Estado o la inversión pública como factor de decisión y resolución de sus problemas. En el fondo de todos estos asuntos aparece una cuestión de actitud. La necesidad de superar la crisis de gobernabilidad y afirmar la eficiencia de la política pública consiste principalmente en lograr que la transformación 311 del papel del Estado se concrete en procesos de verdadera descentralización y autonomía para los diferentes entes y organizaciones estatales y sociales (recuérdese que el Estado es mucho más que el gobierno). Desde luego se corren muchos riesgos, y la plata que maneja un municipio se la pueden robar; pero para ello deben promoverse procesos de evaluación de la inversión pública y de la gestión; utilizar mecanismos del tipo mandato programático que aparece en la Constitución colombiana y en otras de la región, y hacer efectivos una serie de elementos a los cuales las sociedades latinoamericanas no están acostumbradas, pero que realmente son el único camino para que se entienda que el Estado no sólo es el Estado central, sino también el Estado de las regiones y los municipios. Y por supuesto la colaboración entre esos tres niveles y los órganos de control y fiscalización. Esto expresa de qué manera no ceja en su empeño el antiguo modelo de Estado providente, que “civiliza” y “educa” a la sociedad civil desde “arriba” suponiendo, en cierta medida, la incapacidad de la comunidad y de sus organizaciones para regir sus asuntos. La baja participación electoral en las elecciones, muestra las grandes dificultades de acceder a tipos cualitativamente fuertes de democracia de participación. El fenómeno aparentemente contradictorio de una baja participación en los sucesivos comicios realizados luego de la implantación de reformas políticas, pese a la financiación estatal (parcial) de las campañas, al acceso formal a los grandes medios de comunicación y a la aparición en la arena política de nuevas fuerzas sociales y políticas (guerrillas desmovilizadas, movimientos suprapartidistas, minorías étnicas y religiosas) indica una diferencia entre los patrones de la cultura civil y política de la población y los predicamentos “participacionistas” de las nuevas elites políticas, beneficiarias inmediatas de los procesos de modernización. En Colombia, los desarrollos legislativos y la puesta en práctica de la reforma constitucional de 1991 no han sido tanto dados en el sentido de fortalecer la democracia participativa, la autogestión o la descentralización; sino, por el contrario, en el de aunar —desde un punto de vista pragmatista— la funcionalidad del Estado en términos del papel preeminente del ejecutivo. Sin embargo, los avances políticos y sociales de la descentralización son hoy en día algo irreversible. 312 e. El Estado empresario. No se trata de que el Estado abandone totalmente su función empresarial, ya que muchas de las actividades empresariales que éste tiene a su cargo son altamente rentables y, manejadas en un esquema de competencia, transparencia y rechazo a los modelos de gestión clientelistas y patrimoniales, resulta importante la presencia estatal para proveer recursos adicionales al fisco público y establecer criterios racionales respecto de la competencia privada (Osborne y Gaebler, 1992). El punto decisivo aquí es la lucha contra todo tipo de monopolios, sean público-estatales o privados. f. Los mecanismos de control de gestión. Para combatir la corrupción se ha considerado vital fortalecer la gestión pública a través de la puesta en escena del conocido mecanismo de los “frenos y contrapesos entre las propias agencias estatales”. Adicionalmente la Constitución colombiana y el sistema legal que de ella se ha desprendido, han dotado a los ciudadanos y a las comunidades de herramientas eficaces para ejercitar el control popular de los asuntos públicos (R. Klitgaard, 1994). Igualmente la reforma constitucional de 1991, estableció una serie de instancias nuevas y reforzó algunas viejas que pueden ser vistas en el sentido de una amplificación de la triada modélica de las ramas del poder público: ejecutivo, legislativo y jurisdiccional. Es evidente que instituciones como la Procuraduría, el Consejo Nacional Electoral, la Comisión Nacional de Televisión, el Sistema Estatal de Universidades Públicas, la Veeduría del Tesoro —para mencionar sólo algunas instituciones nuevas o reformuladas— implican, de hecho, una capacidad mayor de establecer un equilibrio de los poderes públicos constituyendo sistemas a veces sofisticados de controles que, en mi opinión, están lejos de expresar una anarquía o una atonía de la ciudadanía y de los propios actores políticos con respecto a la dirección misma del Estado. En uno y otro caso, el modelo que en la práctica se ha venido implementando, expresa una recusación al presidencialismo y a la identificación equívoca entre poder público y poder ejecutivo. g. La política criminal. En este campo se ha reforzado, sin duda, la política pública que concentra y centraliza el llamado orden público en el gobierno central. Tal es el sentido de la ley de estados de excepción recientemente presentada por el gobierno al congreso y el hecho mismo de que durante 313 los años pasados, a pesar de las promesas de levantamiento del Estado de excepción —antiguo Estado de sitio— el país ha vivido bajo la férula de los ucases y decretos presidencialistas. Pero, de otro lado, la nueva Constitución promulgada en 1991, permite mecanismos de justicia popular, como los llamados jueces de paz. En la medida que actúen, el Estado delegaría una serie de funciones justiciales con el fin de que sean asumidas por la comunidad. Además, muchos de los conflictos económicos los zanjan organizaciones privadas como las cámaras de comercio, etc. Fenómenos, de otro lado, que significan una delegación de la seguridad son la prestación de vigilancia privada y las organizaciones civiles de defensa propia. Se supone que la policía, como cuerpo armado de orden interno, cumple una función monopólica del tutelaje; pero no solamente en Colombia, sino en la mayoría de los países de América Latina, la institución policial coexiste con mecanismos coactivos y preventivos de carácter privado. Así en el propio terreno de la justicia y la seguridad no se asume hoy un monopolio total de la prestación directa de las funciones estatales. Y ello en tanto ha cambiado, se ha relativizado, el principio decimonónico de la soberanía nacional. De otro lado, en el país se avanzó hacia una mayor autonomía del poder judicial respecto del ejecutivo, como los sonados episodios de juzgamiento de altos funcionarios públicos por parte de la Fiscalía General de la Nación lo atestiguan. h. La política ambiental. En este campo se ha diseñado un esquema institucional que ha recogido la legislación más avanzada sobre política ambiental existente hoy en día. Pero este es un proceso que, desde el punto de vista de la gestión pública, es aún incipiente. La política pública no ha precisado si adopta el esquema teórico de la ecología como administradora del capital ambiental, visión antropocéntrica del problema, o si los bienes naturales poseen determinaciones externas a las necesidades humanas. Cabe anotar, de manera positiva, cómo en el nuevo esquema institucional en este campo se le ha dado un reconocimiento expreso a las comunidades étnicas (indígenas y negritudes) y a las numerosas organizaciones ecológicas de tipo no gubernamental. De tal modo que la política pública colombiana le asigna claras responsabilidades a los actores públicos y privados. i. El esquema de competencia política. Este es quizás uno de los aspectos críticos en la revaloración de lo público y lo privado. Los partidos 314 y fuerzas políticas representan el canal de mediación por excelencia entre estos dos espacios sociales. Son los agentes que concentran las demandas sociales y las transforman en temas de política pública, pero su crisis es hoy bastante aguda. Desarticulados, afectados por la corrupción y la pérdida de su credibilidad, no son funcionales, y lo peor, no se vislumbra su reemplazo. Por ello, se dan con frecuencia opciones organizacionales difusas de corte alternativo que resultan sólo efímeras y coyunturales. Este es uno de los puntos críticos en un programa de modernización de la vida pública colombiana (E. Varela: 1995). 5. Surgimiento de nuevas fuerzas sociales y perspectivas de afirmación del interés público El análisis de los procesos de modernización del Estado no puede hacerse sin tomar en consideración los cambios dramáticos en cuanto a los actores sociales públicos y privados. Nuestro peculiar proceso de incorporación a la Modernidad, tanto en los niveles de identificación y adscripción a las tendencias globales, como en cuanto a las especificidades del fenómeno, muestra cambios acelerados en las décadas. Entre ellos podemos destacar: En primer lugar la modernización parcial de sectores de la vieja clase empresarial que han transformado sus estructuras organizacionales y de propiedad, desde la unidad de tipo familiar hasta llegar a niveles tales de organicidad como la sociedad corporativa anónima. Esto ha ido de la mano con el afianzamiento de algunos emporios transnacionales cuya base son las propias economías latinoamericanas. En buena medida, la profesionalización de la gerencia se ha apoyado en la asimilación, muchas veces acrítica, de los modelos norteamericanos y japoneses. Pero en la práctica, en las relaciones disciplinarias, la actitud frente al mercado, a las regulaciones estatales, en fin, en los rasgos más relevantes de la cultura empresarial estas elites no expresan una ruptura radical con los esquemas patrimoniales históricamente prevalecientes (González y E. Ogliastri, 1995). Adicionalmente, en el caso colombiano, vale la pena subrayar el importantísimo peso que, dentro de la acumulación de capitales, le ha correspondido, en este mismo período, a sectores criminales, ilegales y/o “informales” 315 que podemos englobar bajo la denominación de “lumpenburguesía”. Como actores más significativos de dicho pathos mencionemos: • La delincuencia política, en primer lugar, en la funcionalidad social de la tradicional “clase política”, usufructuaria sempiterna de los bienes públicos o en la nueva delincuencia política que, desprovista de su justificación ideológica marxista o nacionalista, ha transmutado su abierto desafío al stablishment por la depredación bandidesca de las esferas públicas y privadas mediante el uso de la violencia y el terror. • La narcoburguesía, cuyo afloramiento evidencia de la mejor manera el rumbo específico de nuestro tránsito hacia la Modernidad. En particular, la conformación de una cultura de capitalismo salvaje, de anulación de las inhibiciones morales que subyacen a la adecuación entre capitalismo y democracia. En efecto, un rasgo dominante en la praxis social, empresarial, política y delincuencial del narcotráfico es la impronta de la antidemocracia, de la intolerancia llevada hasta sus últimas consecuencias. No se trata en el contexto latinoamericano de un simple episodio de surgimiento de un sector neocapitalista que se proponga como móvil el rasgo central de la actividad empresarial privada: la maximización de la ganancia y el lucro. Al contrario, lo peculiar de la burguesía mafiosa es su imperativo impulso de control sobre las esferas públicas: en los massmedia, en el mundo de la política a través de su financiación fraudulenta e interesada y en la conformación de una cultura de enclave, rural y urbana, que reemplaza el poder del Estado por la voluntad omnímoda del “padrino”. Desde una perspectiva analítica, quisiera destacar el enorme papel que estos grupos, generalmente de origen plebeyo, han jugado en la privatización patrimonial de los asuntos públicos. Otro fenómeno característico de la modernización es el menor peso específico otorgado a la institución que, en buena medida, coadyuvó a la cohesión social en el primer siglo y medio de vida republicana en nuestro continente: la iglesia católica. Es sintomático que dicha institución, basada en el discurso de la justicia social, esté mayoritariamente en oposición a los procesos de reforma y modernización de la vida pública y económica, expresando su complacencia con el esquema patrimonialista de administración de 316 lo público. En esta perspectiva podemos entender la tensión existente entre las esferas de los altos gobiernos y la Iglesia en los últimos años; contienda centrada en la recuperación por parte del Estado de un estatuto laico y de separación respecto a los roles público-estatales que, en virtud de la vieja constitucionalidad y el mantenimiento del concordato aún le correspondían a la iglesia católica. Es, en síntesis, un conflicto derivado de un tardío proceso de secularización del Estado, de definición institucional y política del ámbito religioso hacia el terreno de la privaticidad (el fuero interno de los creyentes y la libertad para el ejercicio de los diversos cultos o la indiferencia social en materia religiosa). Alternativamente el papel recesivo de las fuerzas armadas en América Latina durante las últimas décadas, expresa un arraigo cada vez mayor del liberalismo político, de la conjunción, no siempre fácil de lograr, entre capitalismo y democracia. En el caso colombiano, la modernización del Estado no ha modificado la relación de subordinación de los estamentos castrenses, que se remonta incluso al final de la última de las grandes guerras civiles del XIX. La reforma del Estado ha propulsado —tras la búsqueda del monopolio estatal de la justicia y la coerción— un incremento del tamaño de las FF.AA (aumento del pie de fuerza, mayor participación en el presupuesto estatal, etc.) debido a que, como lo he señalado en otras ocasiones (E. Varela, 1995), nuestra Modernidad no nos ha alejado de la barbarie; y aún persiste, lejos de su finalización, un conflicto sordo y sucio con base en el uso de la violencia y el terror entre las distintas facciones, políticas y prepolíticas que desgarran al Estado y la sociedad civil colombianos. El primer fenómeno que se destaca en este sentido es el carácter recesivo, decadente, de las formas paradigmáticas de confrontación social, política y reivindicativa de estos actores sociales. No se trata simplemente de constatar cómo en América Latina, al igual que en el mundo entero, se ha presentado un desplome de las tendencias políticas de estirpe marxista-leninista. Ello es así. Pero, además, se presentan rasgos que van mucho más allá de la acción política que proviene de dichas vertientes, como ocurre en el declive del movimiento sindical, tozudo en la repetición ritual y ahistórica de esquemas de confrontación desfasados con las nuevas realidades. Es éste tal vez dentro 317 de los grupos sociales que se han manifestado como un férreo crítico de los procesos de modernización del Estado en América Latina. En términos globales, a mi juicio, el quid del problema se deriva de obtener una plena comprensión del rol que han jugado aspectos determinantes del decurso de la historia moderna del país. En primer término se destaca la debilidad de las transformaciones democrático-liberales en Colombia, proceso histórico inconcluso y que, por lo tanto, no despejó el desfase existente entre los procesos socioeconómicos y los cambios culturales y político institucionales: meollo de la cara bifronte de la democracia formal y la doble moralidad colombiana. Ello nos conduce a un tema crucial: el carácter apenas formativo o embrionario de la llamada identidad nacional, lo que podemos resumir señalando que el Estado-Nación apenas está en su fase constitutiva. Los desequilibrios de la mayoría de las naciones de América Latina, las guerras civiles, las soluciones de fuerza, la violencia social, etc. se explican parcialmente en razón de los límites alcanzados por la fuerza cohesionadora de los Estados y, por ende, se explican debido a la débil percepción de este nivel en la conciencia nacional popular: la lumpenización de la política o el neobandidaje politizado se erigen desde este sustrato; el deslizamiento del universo político hacia la corrupción, el mercenarismo y el caudillismo resultan, en consecuencia, de fácil génesis. Igualmente la ruptura de los lazos orgánicos que unen a los jefes, los partidos, las instituciones con sus propias o correspondientes bases sociales es un fenómeno reiterativo. Los anteriores fenómenos responden a la debilidad de las clases y núcleos sociales que habrían (teóricamente) de liderar el tránsito del “en sí” al “para sí”. Es este el caso de la burguesía, y en general del empresariado colombiano, infeccionada por la vinculación con el latifundio, el gamonalismo, el rentismo, la especulación, el dolo, etc., lo que la ha llevado a privilegiar un modelo deforme de sociedad burguesa. Lo que se profundiza más cuando se piensa en la hibridación que durante las décadas de los ochenta y noventa, se ha venido dando con respecto a las llamadas clases emergentes. De esta fusión saldrá cualquier cosa, menos una elite con afanes de profundización de la democracia política y modernización sustancial del Estado. En realidad el entrabe del tejido social en la región corresponde al hecho 318 de que las fuerzas sociales en pugna aparecen incapacitadas para promover un programa cabal de tipo orgánico —estructural que implique una valoración del rol público-estatal desde una visión moderna—. Es evidente que la crisis de la Modernidad se ha expresado en el agotamiento de los procesos de intervención política de la sociedad civil. Es este fenómeno evidente en la crisis de los partidos, en el enorme poder de los massmedia, en la desideologización de los conflictos sociales y políticos, en el surgimiento de reivindicaciones centrales para el interés común antes desdeñadas: derechos individuales, desarrollo sostenible con profunda conciencia ecológica, frenos y contrapesos al leviatán estatal, convergencia en la necesidad imperiosa de eliminar los particularismos armados y desarmados, que imposibilitan el ejercicio pleno de la soberanía en el marco de la Nación-Estado, etc. La manifestación de la crisis estructural opera, en forma directa, en los criterios sociales de valoración del Estado. Si a éste, como en el caso de Colombia y otros países latinoamericanos, sólo se le vislumbra como socializador de las pérdidas y quiebras; y privatizador (redistribuidor) de las utilidades. Véase el caso tan frecuente del subsidio estatal a la actividad privada monopólica de acumulación de capital, el cual magnifica la separación entre el Estado, como praxis del ejercicio del poder por parte de las elites dominantes, y la Constitución Política como postulado o perfectibilidad puesta. Y por consiguiente expresa la ruptura entre ese mismo Estado y el conjunto de la sociedad civil. Esto es particularmente importante cuando contemplamos la realidad efectiva de los acuerdos contractualistas por los que ésta se rige y los cuales se encuentran sobredeterminados por la estructura y la dinámica del Estado. En este sentido el Estado semeja una función de mediación entre los distintos intereses particulares, donde inclinándose de facto a favor de un grupo, impone y legitima una relación de fuerzas al conjunto de la sociedad. Así, el arbitraje pierde su función consensual, democrática, y se anuda al viejo corporativismo y a formas sublimadas de patrimonialismo. El problema, desde una perspectiva concreta, al menos para el caso colombiano, reside en nuestra incapacidad de historizar positivamente la función de mediación de Estado y/o reclamar la validez operacional de los acuerdos entre dichos intereses conflictuales. Tal es el sentido en el que ha resultado efectivamente funcional la vieja consigna de democratización del Estado que 319 generalmente conduce a la práctica de un corporativismo que puede ser capitalizado por el amarillismo sindical, los gremios y el burocratismo liquidacionista. Sólo abriendo la sociedad civil, desatando sus energías constreñidas, podrá recuperarse el terreno perdido. La politización de la sociedad habrá entonces de provenir de la esfera privada, lo cual ciertamente —y aunque luzca paradójico— es el punto de partida para el fortalecimiento del hombre público. (Sennett, 1979). No se trata aquí de establecer un calificativo moral (bueno o malo) respecto del papel que el Estado cumplió a lo largo del periodo señalado. Una evaluación retrospectiva del Welfare State debe reconocer su fundamental función como catalizador en la acumulación de capital en sociedades débiles con un enorme peso de la tradición señorial, carentes siquiera de la mínima infraestructura básica necesaria para configurar una sociedad industrial. El Estado, efectivamente, construyó los puertos, las vías férreas, las centrales eléctricas, los sistemas de banca central y, en general, jugó un papel determinante en la prestación de servicios sociales, en el garantismo asistencial que no fue, en modo alguno, una dádiva sino un resultado de acelerados procesos de cambio que involucraron, y aún lo hacen, dolorosos y violentos conflictos sociales. De una forma u otra, además, el sector público propició la sindicalización de vastos sectores asalariados y el reconocimiento de muchas comunidades tradicionales como poseedoras de derechos legítimos ante los cuales el Estado debía responder. Se validó así que los derechos humanos no serían sólo retórica sino elementos que condicionarían las políticas públicas (López, 1992). Incluso ello se dio en procesos de democratización perversa propulsados por la corrupción y el llamado clientelismo. Debe retenerse aquí que la palabra modernización o reforma del Estado no posee un sentido per se ahistórico, sino que depende de la evolución social, de procesos económicos, políticos; del contexto internacional, etc., que le dan significados cambiantes. La modernización del Estado al final del siglo XIX y primera mitad del siglo XX es muy distinta a la modernización del Estado de nuestro tiempo. Modernizar al Estado hoy implica rectificar la hipertrofia del aparato público que resultó uno de los efectos negativos del anterior proceso. Lo que antes era necesario frente a los retos de una sociedad tradicional hoy puede decirse corresponde a tareas cum320 plidas. El efecto final ha resultado altamente negativo distorsionando la filosofía inicial que le dio fundamento. En efecto, no hay gobierno contemporáneo que no tienda, de suyo, a aumentar el gasto público, al reglamentarismo, a intervenir en la vida privada, a moralizar a sus ciudadanos, etc.; así, muchos de sus líderes e ideólogos se pregonen como neoliberales. Se parte del principio hoy de que sólo es posible asumir nuevos retos públicos con el aumento de impuestos y la hegemonía gubernamental. Esta hipertrofia llevó a la consolidación de burocracias públicas anudadas con modalidades clientelistas de gestión que terminaron por convertir los derechos sociales en privilegios de la burocracia pública y de los grupos de interés que la controlan (Peters, 1999). De ser el Estado una emanación de la sociedad civil, se convirtió, a menudo, en su verdugo. Y a pesar de que en las democracias contemporáneas se parte del principio de la soberanía popular, que cada tantos años los pueblos eligen a sus gobernantes, no se ha eliminado el que la población resulte virtualmente “explotada por sus gobernantes”. Incluso, en términos de decisiones, quien quiera colocar una empresa, fundar una escuela, desarrollar cualquier actividad, está sometido al arbitrio de las reglamentaciones del anónimo poder público estatal. Son demasiado conocidos entre nosotros los excesos, por ejemplo, de la tramitología kafkiana para obtener un permiso o una licencia. En buena parte del mundo, durante este último periodo, una parte significativa de las actividades que asumió el Estado empresarial las tomaron bajo su rienda productores, empresarios y grupos sociales en un ambiente de competencia regulada. Ello por dos vías: la venta de los activos del Estado o sucedáneamente por la ruptura de los monopolios público-estatales. Incluso en este último caso, manteniendo la función empresarial del Estado, pero bajo claras reglas de competencia con los prestatarios o productores privados. Se intentó entonces generar estructuras sociales en donde los agentes públicos y privados concurriesen en la satisfacción de los principios de equidad y de provisión de los servicios sociales que justificaron durante los últimos ciento cincuenta años el modelo del Welfare State. Se trataba de buscar una definición del contrato social, una negociabilidad en un contexto de apertura y respeto a los derechos ciudadanos y a la libre iniciativa de los particulares. Sólo sobre esta base, sociedades como la 321 colombiana lograrían un grado importante de gobernabilidad. No obstante, las prácticas de modernización de las instituciones públicas que sólo tomen en cuenta los aspectos procedimentales, el eficientismo, la rentabilidad en el corto plazo, etc., no alcanzarán dicho propósito. La sociedad moderna expresa cada vez más una separación tajante entre la simple función de propiedad (vista en sí misma como algo pasivo) y la función del Management, donde tanto el Estado como las demás organizaciones sociales actúan como satisfactores de la demanda en el mercado público. Además la civilidad, en tanto potencialidad —o puesta en marcha— de una nueva cultura, se apoya de manera indispensable en el sesgo o en el contenido de las funciones sociales (direccionales, promocionales y redistributivas) que se le asignen al Estado. Pero no se trata de implantar un modelo de relación entre Estado y sociedad civil que desatienda los problemas de las mayorías. Por el contrario, es un modelo demoliberal que supone al mercado libre, sobre la base de estas normas básicas de intervención del Estado de carácter regulatorio, capaz de satisfacer tanto los beneficios de la acción individual como, de forma en cierta medida automática, el beneficio común. Lo anterior sólo tendrá efectividad si se establecen claros rumbos para la funcionalidad complementaria de los actores públicos y privados bajo un eje central: la recusación del patrimonialismo clientelista, del autoritarismo social y estatal y la promoción, en su sentido lato, de las libertades públicas. En síntesis, los procesos de modernización en su peculiar decurso dentro de la historia latinoamericana en general y colombiana, en particular, han cambiado el sentido político y cívico de lo público, definiéndose éste a sí mismo en una tensión dinámica con la órbita estatal que lo cobija, y pugnando por aflorar en términos del logro de una real democratización de todos los ámbitos, públicos y privados, de vida e institucionalidad. 322 NOTAS 1 Este es el caso de México, donde la implementación de las reformas liberales ha generado enormes conflictos sociales, propiciado la aparición de guerrillas antimodernistas, y se ha imbricado con la propia crisis del PRI (cfr. C. Fuentes, 1995). Alternativamente, en el caso Argentino, el proceso ha tenido un éxito reconocido aún por contradictores de la señalada tendencia “neoliberal”. Estas reformas han implicado profundos cambios en Chile, Bolivia (donde se redujo la hiperinflación recurrente a menos de dos dígitos) y el Perú, donde bajo un modelo autoritario y “antipolítico” se han logrado notorios éxitos de política pública. Esto para sólo señalar algunos casos relevantes con resultados disímiles (B. Levine y otros, 1992). 2 Cfr. los artículos 209-210 de la Constitución colombiana de 1991 que consagran los citados postulados (cfr., Manrique: 1991). 323 PÁGINA EN BLANCO EN LA EDICIÓN IMPRESA Capítulo 7 ¿RENOVAR O DESTRUIR LAS BUROCRACIAS ESTATALES? EL IMPACTO DEL NEW PUBLIC MANAGEMENT En los últimos años un conjunto más o menos coherente de teorías han reclamado una transformación sustancial de la administración pública en una escala mundial. Estas nuevas teorias, a diferencia de muchos debates precedentes que se enmarcaban desde el origen de sus proposiciones en disciplinas tales como la sociología, la ciencia política o el derecho, han provenido primordialmente de las Ciencias del Management (Barzelay, 2001; Evans, 2004; Peters, 2003), tradicionalmente centradas en el desempeño de las organizaciones económicas empresariales privadas movidas tras la obtención del lucro. Bajo la influencia de concepciones de carácter neoliberal, y con base en indicadores de eficiencia y eficacia, la teoría dominante dentro del Management público propugna, desde hace unas dos décadas, una transformación sustancial de la función tradicional de las burocracias clásicas gubernamentales. En general, su reclamo pretende corregir los abusos del intervencionismo del Estado; y para ello se apoya en la mala imagen social de las burocracias gubernamentales, a las que acusa primordialmente de clientelismo e ineficiencia. Quizás a nivel de la literatura gerencial del tipo best sellers, el epítome de esta corriente fue el conocido libro, vendido por centenares de miles de ejemplares en el mundo entero, titulado La reinvención del 325 Gobierno (Osborne y Gaebler, 1992). Uno de sus autores (Gaebler) fue un consultor cercano a los círculos de poder de la adminstracion Clinton, y sus páginas ilustran frecuentemente las realizaciones del modelo de “reinvención” en diversos niveles de la administración pública de los Estados Unidos en el citado periodo. En los países anglosajones, Europa Occidental y América Latina, este tipo de literatura ha gozado de una enorme influencia, no sólo en los círculos académicos, sino primordialmente a nivel de los policy makers. De tal suerte que muchas de sus recetas y lecciones fueron asumidas con alborozo e irreflexión al diseñar modalidades de cambio sustancial en la manera de adminstrar al Estado, y definir sus relaciones tanto con la sociedad civil como con la esfera del mercado. El modelo de gestión denominado genéricamente New Public Management, NPM (Spink y otros: 2001; A. Nickson, 2002), resultó de una combinación de dos ideas: la primera, aplicar los criterios y los modelos dominantes para la gestión empresarial privada al sector gubernamental a través de mecanismos de rendición de cuentas, evaluación del desempeño, intentando desvirtuar el modelo burocrático clásico (N. Mouzelis, 1991). En segundo lugar, cuando ciertos de sus teóricos reconocen la imposibilidad de llevar esta meta hasta sus últimas consecuencias, han erigido un discurso alternativo (o complementario) centrado en una peculiar profesionalización de la gestión pública (Ospina, 2001). El supuesto básico de esta segunda estrategia reconoce que en numerosos países, entre ellos los latinoamericanos, el proceso de consolidación de la burocracia de estirpe clásica nunca se llevó a cabo, al menos en un estadio maduro y, por eso, antes de pensar en destruir totalmente la burocracia, habría que emprender la tarea de su radical transformación. Sin embargo, ambos matices concuerdan en que allí donde la mercantilización libra a su suerte al sector de la burocracia pública empresarial, serán los ajustes del mercado los ejes de esta transformación. Mientras ese mecanismo no sea posible, se requerirá una reingeniería gradual con el objetivo de lograr el desmonte de los privilegios de la burocracia clásica. En este texto, por lo tanto, pretendo describir críticamente los principales ejes que han guiado la transformación emprendida sobre las estructuras y 326 funcionalidades de las burocracias públicas. E igualmente mostrar de qué manera estos cambios se articulan con la programática de mercantilización de la provisión de la mayoría de los bienes y servicios públicos. Sobre este esquema comprensivo, finalmente, me interesa destacar en qué terrenos percibo se mueve actualmente la ejecución de estas políticas, así como las resistencias y oposiciones que éstas han suscitado, afectando su viabilidad. 1. El ideario del New Public Management El tono general, entusiasta, del New Public Managment fue ilustrado centralmente con la tesis de que con sus predicamentos saldríamos al fin de la caja de hierro de la racionalidad burocrática, según rezaba la conocida expresion weberiana (Powell y DiMaggio, 1989). Las tesis o axiomas de los que partían los reformadores eran del siguiente tenor: una sociedad “postburocrática” (C. Heckscher y A. Donellon, 1994) habría emergido tras el desplome del viejo orden industrial de tipo taylorista-fordista (H. Jorda, 1999). La postmodernidad social, cultural, tecno-científica, etc., nos demandaba poner las instituciones de los gobiernos a tono con la nueva epoca. A la caída de la sociedad industrial, debería acompañarla el final del Estado burocrático, etc. Cabe recordar aquí sus orígenes históricos: las ejecutorías reformistas del aparato público y las privatizaciones en la administración Tatcher en Inglaterra en los ochenta; el reaganismo en EE.UU en ese mismo periodo; y los extraordinarios cambios que se ejecutaron, durante los 90, en la función pública en países de la órbita de la Commonwealth. En efecto, esta corriente —a diferencia del Management estratégico americano diseñado para el sector privado— ha surgido casi simultáneamente en tres esquinas del mundo anglosajón: Norteamérica (Canadá y los EE.UU.), Inglaterra y Australia. El modelo por excelencia dentro de este subconjunto, han sido las ejecutorías adoptadas en Nueva Zelanda. Este país, relativamente poco poblado y con un nivel elevado de riquezas, ha sido origen y ejemplo de este ideario en dos órdenes de cosas: el manejo del presupuesto público, asumiendo una solución de continuidad de los esquemas tradicionales de la hacienda pública; y la gestión del personal. Todos los tópicos e ideales que abajo consignaremos han sido adoptados en dicho país. 327 1.1. LA PROGRAMÁTICA DEL NEW PUBLIC MANAGEMENT Precisamente un académico angloparlante del Canadá ha hecho una sistematización de los principales aspectos de la doctrina contrastándolos, con los modelos llamados tradicionales (Kernaghan, 2000: 92), en los siguientes términos que resultan ciertamente paradigmáticos. Las organizaciones postburocráticas son: 1. Centradas en la calidad de los servicios ofrecidos a los ciudadanos. 2. Con liderazgo participativo, valores compartidos y procesos decisionales democráticos. 3. Con capacidad de consulta a los ciudadanos y con real poder de los empleados. 4. Orientadas al cambio. 5. Orientadas a los resultados. 6. Descentralizadas, tanto en la esfera de las autoridades, como en lo atinente a su control. 7. Son funcionales con base en estructuras no departamentalizadas. 8. Pretenden buscan ser autosustentables financieramente y depender lo menos posible de asignaciones presupuestales inerciales de tipo estatal. 9. Competitivas tanto con otras agencias públicas como con el sector privado. Por supuesto en este modelo todos los aspectos contrarios o antagónicos de cada ítem corresponden a las características de las organizaciones burocráticas tradicionales. Vale subrayar que en este esquema el autor citado supone a los primeros seis puntos como referidos a la cultura administrativa; el siete y ocho a la estructura; y, finalmente, los últimos dos, a los mecanismos y principios de financiación. En realidad, estos tres grandes ejes temáticos ilustran bien las dimensiones del cambio. Veámoslos con mayor detalle. En lo concerniente a la cuestión presupuestal y a su pregunta central ¿cómo financiar el funcionamiento de las burocracias del Estado? y al asunto más general del gasto y la inversión pública, el gran cambio predicado por el NPM es el desmonte de las asignaciones presupuestarias permanentes para los distintos programas y agencias del Estado. Estos usualmente se han con328 certado políticamente en los parlamentos, o quizás, siendo más precisos: han sido decididos en la arena parlamentaria, pero con una creciente intervención e influencia del poder ejecutivo; aunque ciertamente con debates inevitables para los centros del poder, con esferas crecientes de la sociedad civil, los grupos de presión y cabildeo, los expertos y los medios académicos, etc. Los ideólogos neoliberales piensan que este tipo de asignación es indebido, dado que se somete el gasto público a demasiadas presiones políticas y no se toman en cuenta los criterios de racionalidad administrativa (eficacia, eficiencia, oportunidad, evaluación). Se produce así —nos dicen— un despilfarro estructural y se abren las puertas a la corrupción. La receta o antídoto implica entonces hacer que en la medida de lo posible cada agencia y organización pública sea responsable por la ejecución del presupuesto y no que simplemente mantengan su carácter de ejecutoras. Así, al medir la calidad del gasto público, con base en distintos indicadores técnicos, se asume un modelo de premios y castigos. Esto, desde luego, no se puede hacer en la forma tradicional de asignación del gasto que, como ya expliqué, es de naturaleza política. La tecnocracia de ciertos ministerios, hacienda pública, planificación, etc., asumen progresivamente estas competencias. De otro lado, y ello fue aplicado en buena medida en los países señalados arriba, especialmente en Nueva Zelanda, se adopta el modelo de presupuesto cero. Vale decir que no habrá más incrementalismo en uno de los sentidos que las teorías clásicas determinaban como rasgo de la política pública (Lindblom, 1963, 1999). Con la adopción del nuevo ideario no se le garantiza a nadie la continuidad inercial de las asignaciones presupuestales y todo será decidido evaluando. Los presupuestos por programas y proyectos son, en este sentido, un intento de viabilizar tal opción de políticas. El efecto más dramático de este cambio es la flexibilización presupuestal1. De tal suerte que ningún programa, ninguna política, tendría estabilidad en el largo plazo y las razones fiscales pesarían sobre otras consideraciones, sean éstas éticas, de justicia social, redistribución de las riquezas, etc. La segunda gran “innovación” presupuestaria es el origen de los recursos fiscales. En general se tiende a disminuir los impuestos al capital alegando que éste crea empleo y si se lo grava excesivamente se alejaría. La globalización mercantil también ha implicado el desmonte de buena parte de los gravá329 menes arancelarios. El centro de tales gravámenes está, más que nunca, ahora en los ingresos salariales y afines, y en los impuestos y tasas de tipo indirecto que penalizan el consumo. Otro hito significativo en materia presupuestaria es la financiación del gasto —o del déficit en la mayoría de los casos relevantes— a través de la venta de los activos gubernamentales. En América Latina, países como la Argentina y el Perú, alcanzaron cotas —de subasta de empresas y entidades de propiedad estatal— inimaginables. Pero el fenómeno fue generalizado y continúa su aplicación en diversas regiones del mundo2. Aunque existe una creciente conciencia sobre la cortedad e ineficacia estructural de esta política para un financiamiento duradero del gasto gubernamental. La descentralización territorial igualmente ha sido reintepretada. Cada territorialidad debe, en adelante, ser responsable de un porcentaje cada vez mayor de sus ingresos. La autonomía fiscal —un viejo anhelo de muchas regiones frente a los avasalladores controles del centro— se constituye en una clave nueva: no más dependencia fiscal respecto del Estado central3. Ese modelo, además, se ha aplicado radicalmente en todos aquellos aspectos de la provisión de bienes públicos que han sido mercantilizados (E. Varela, 2005a). Los hospitales públicos colombianos, por ejemplo, ya no reciben asignaciones directas, inerciales, del presupuesto general del Estado. Facturan y a cambio del número y costo de los servicios, por ejemplo, para los grupos más pobres, el Estado les devuelve una parte de sus ingresos. Por supuesto, el dinero facturado nunca equivale a las necesidades de los mismos y la inmensa mayoría vive al borde del cierre, cuando no son simplemente clausurados. Sobre la base de estas políticas fiscales y presupuestarias, se ha dado un gran cambio en el nivel propiamente político de la gestión de la burocracia. En el lenguaje managerial se habla de reestructuración, reingeniería y redefinición de la formulación y aprobacion de las políticas (Kernaghan, 2000: 97). Así mismo de las formas de control y rendición de cuentas. En la jerga del NPM: responsabilización o accountability. La teoría del NPM supone que uno de los obstáculos a remover es precisamente el excesivo poder de las burocracias públicas. Su condición de monopolio sobre la gestión de muchos de los asuntos de la sociedad; la centralización de las decisiones y la forma igualmente centralizada del 330 desempeño del Estado son, por ello, los focos de las reformas (O. Ozlack, 1984). 1.2. LAS ESTRATEGIAS DE DEBILITAMIENTO DE LAS BUROCRACIAS PÚBLICAS Como lo expliqué en otro texto (E Varela, 2001), existen al menos 5 componentes de la estrategia de debilitamiento de la burocracia gubernamental: (a) la descentralización terrritorial y funcional, (b) la desmonopolización de las funciones del Estado, (c) las privatizaciones o venta de los activos y las empresas del Estado a los entrepreneurs (d) la mercantilización de los bienes públicos y sociales, y (e) la competitividad como criterio para asignar recursos. Por ahora quiero fijar la atención en los aspectos propios al diseño organizacional, uno de los trazos actuales del oficio managerial (en tanto saber disciplinario aplicado o práctico). Los cambios que se supone introduce el nuevo modelo son: a. Una estructura organizacional menos jerárquica, más aplanada. Se dice, en favor de la nueva modalidad, que existen demasiados niveles jerárquicos, que son muy rígidos los mecanismos de comunicación entre ellos y difícil el tránsito del personal entre los distintos niveles. Por lo tanto, desde el punto de vista del diseño organizacional, ahora se intenta una flexibilización de la estructura. En cierto modo, su parcial desregulación. Éste es quizás uno de los puntos que más se presta a la propaganda que defiende ante el vasto público al nuevo modelo frente al burocrático. En el caso alternativo que se nos vende habría una mayor democratización de la gestión, mayor profesionalismo —el verdadero reino de la experticia— y, además, mayor nivel de control y autocontrol. b. Una mayor separación entre la competencia profesional o el saber especializado respecto de las interferencias de naturaleza externa (léase: políticas). Aquí el reto implica alejar a las burocracias profesionales del debate político atendiéndose en sus procesos de gestión, decisión y evaluación a criterios primordialmente técnicos. Esto se haría rompiendo la carrera del servicio civil, que, en buena parte del aparato gubernamental en el mundo entero es la principal garantía de la estabilidad, la permanencia en los cargos, la posibilidad de ascensos basados en el sistema de concurso de méritos. Estos logros de la profesionalización de la función pública son denotados y, en su lugar, se 331 defiende ahora la movilidad, la fluidez de un sistema que permita la subcontratación de la experticia fuera de las líneas organizacionales standards. c. Una mayor movilidad de las burocracias. Esto último se desdobla en la polivalencia, o sea, en el desempeño flexible de funciones sobre la base de equipos y tareas asignadas facultativamente por los niveles superiores, o por las leyes. De otro lado, la flexibilidad también se refiere a la discrecionalidad en el uso de las tareas que cumplen las burocracias. Allí encuentra su lugar el outsourcing o subcontratación, y todas las teorías del adelgazamiento organizacional (Drucker, 1999), ahora visto en términos de racionalidad económica. Las políticas de despidos masivos, jubilaciones anticipadas, contratos a término fijo, nóminas suplementarias, etc., hacen parte de una visión precarizadora de las relaciones laborales en el sector. d. Esto se ha complementado con un debilitamiento cuantitativo y cualitativo del sindicalismo en el sector (De la Garza, 2000). En el fondo, en extremo se trata ni más ni menos que de la supresión, la destrucción de las burocracias gubernamentales. Para ello son frecuentes argumentos del tipo: “el empleo público es improductivo”, “destruir empleo público es generar recursos para crear empleo privado”; “éste sí es productivo”, etc. (Los argumentos que cito aquí son frecuentes en el discurso del actual presidente colombiano, Álvaro Uribe Vélez, y su equipo de gobierno; pero no es éste un discurso aislado. Lo repiten diversos gobiernos en todas las latitudes). f. Formas flexibles de remuneración, compensación, promoción etc. Para lograr este objetivo, se debe previamente alcanzar la desregulación laboral a nivel macro modificando las leyes laborales, tanto del sector público como del privado, dado que en esquemas de mercantilización los costos laborales, en virtud del rápido isomorfismo en las formas de gestión y en las tecnologías, resultan un factor decisivo (Genereux, 2001: 120-130). En América Latina en numerosos países como en Chile, Ecuador, Colombia, se han cambiado las leyes laborales para desmontar la estabilidad de los contratos de trabajo con el argumento de que estas leyes de hierro afectan la creación de nuevos empleos y disminuyen la competitividad. 332 1.3. ESTADO, SOCIEDAD, MERCADO Otra rama del cambio la constituyen las relaciones del aparato público con la sociedad y con el mercado. En este terreno, que podríamos llamar externo o relacional, la idea central del NPM es la rendición de cuentas a la sociedad en términos de la eficiencia-eficacia de los servicios a cargo de las burocracias, en todos los terrenos. En lo esencial se transita desde la abstracta y politizada noción de ciudadano a la despolitizada y acéptica condición de cliente. El público usuario consume unos servicios y puede por ello reclamar en relacion con la calidad, costo, oportunidad, diversidad, etc., de los mismos. El instrumento tranformador es —como lo mencioné brevemente— la mercantilización de las acciones del gobierno. Desde la seguridad, las tareas tecnoadministrativas, hasta la provisión directa de bienes y servicios. Realmente no habría en esta perspectiva algún tipo de límite o demarcación que separe lo que es mercantilizable de aquello que no. Incluso, frente al recurrente debate de cómo evitar o combatir la corrupción, este tipo de discurso nos habla de los costos, los incentivos, como factores claves para inhibir este tipo de conductas. Antes, en la era de la ciudadanía, al menos todos eran potenciales ciudadanos en el plano formal o jurídico. La clientelización de la sociedad marca un turning point de la exclusión silenciosa de los marginados respecto de la sociedad de consumo. En suma, como lo describe reiteradamente Omar Aktouf, el Management (en este caso la variante aplicada al sector público) es una suerte de brazo “armado” del economicismo neoliberal; y de los intereses de las grandes corporaciones y centros de poder del capitalismo financiero. Como lo subraya Aktouf: “(...) los programas de formación en Management a la americana están en el proceso de usurpar sobre el terreno de la comprensión, de la explicación y de la propia acción, los términos de las cuestiones económicas e igualmente de las políticas. Se trata - en suma - de un deslizamiento de primera importancia, ya que se hace pasar por ‘económico’, o peor aún por ‘político-económico’, un discurso que no es a la larga cosa distinta que un mensaje de estirpe financiera y managerial, es decir basado en los negocios y la rentabilidad” (O. Aktouf, 2002). 333 Esta dinámica de dotación de contenidos y significados de la acción managerial se da en un contexto de privilegio de la mercantilización. De tal suerte se busca que las experiencias de vida queden inevitablemente inscritas como experiencias de venta-compra-consumo. La idea de un consumo no mercantil tendería a ser descartada. He allí una nueva utopía. Más aún cuando el propio consumo resulta subsumido en un cuadro dominante macro: el de la actividad especulativa del capital y su rentabilidad no productiva. En esta lógica el funcionariado gubernamental sería prescindible. La delegación o devolución de la gestión de los asuntos públicos al mercado (eufemísticamente nominado aquí como “sociedad civil”) desconoce que una parte significativa de los bienes públicos no tiene una naturaleza mercantil. Y que esta esfera mercantil requiere siempre, para que ésta exista aún en la peor de las utopías liberal-libertarias del tipo imaginado por Nozick (1974), de un organismo macro, dueño de un poder coercitivo y normativo no reductible a la rentabilidad o al costeo económico. Y por supuesto el Estado sin un aparato estable de funcionarios de carrera es impensable —como la experiencia contemporánea lo demuestra—. Allí donde las reformas han tenido “éxito” o donde los efectos de la globalización han sido terribles desplomando el aparato del Estado (en aquellas regiones denominadas cínicamente “Estados fallidos”), las sociedades se han retrotraído hacia formas de confrontacion e incivilidad no vistas ni siquiera en el Medioevo. 2. Globalismo y desmonte de las burocracias públicas El influjo mundial del NPM en el periodo señalado es fácilmente comprobable con una somera revisión de la literatura sobre el tema. Además, otro indicio claro es el enorme esfuerzo que a su promoción dedican las agencias globales multilaterales, en especial el World Bank y, en Latinoamerica, la OEA, la Cepal, el BID, entre otros. El Banco Mundial expresamente destina con regularidad fondos para financiar el downsizing del sector público en numerosos países. Con esto facilitan o viabilizan los despidos masivos, las jubilaciones anticipadas, igual el desmonte radical de ciertas agencias públicas (World Bank, 2004). Como lo afirma Chossudosky: “Los préstamos de apoyo a las reformas (policy based loan) se han establecido en el marco de las instituciones de Breton Woods. Se suministra dinero a los países para ayudarlos en los ‘ajustes’. Pero los fondos sólo son entregados cuando los países recipientes aceptan las determinaciones de políticas de los países prestamistas” (Chossudosky, 2004: 69). De manera adicional, los planes de ajuste del FMI incorporan elementos de tipo financiero-fiscal. Agreguemos las políticas del FMI sobre el déficit presupuestario cero, la lucha contra la inflación, la moneda sana, la liberalización de los mercados, que son ejes complementarios que a la larga convierten al NPM en un modelo fuera del cual no hay opción. Ello inscrito dentro de un proceso profundo como el de la globalización de las políticas, la unificación de los mercados y el control de los mismos por parte de las grandes potencias capitalistas (Amin, 1991; Negri y Hardt, 2000). Como lo afirmó recientemente el profesor Andre-Noel Roth (2003): “Si el desarrollo tradicional del Estado y de las políticas públicas desde los años 50 tiene su origen en la socialdemocracia, los principios económicos de Keynes y la administración pública weberiana, la renovación administrativa propuesta se inspira directamente en la escuela neoliberal y en la gestión empresarial privada”. En efecto, el grueso de las fórmulas adoptadas por el NPM es una adaptación, o mejor, una exportación de los principios del Management hoy dominante en los EE.UU —al menos en su formulación retórica más edulcorada— al sector gubernamental. Esta expansión comenzó en los ochenta en los principales países anglófonos; luego, tal ideología fue adoptada por los organismos multilaterales, hizo parte sustancial del Washington consensus y constituye en la actualidad una de las políticas públicas globales determinantes, con sus componentes: los planes de ajuste y rediseño institucional de los Estados. Una de las formas más corrientes de justificación de este préstamo o transferencia consiste simple y llanamente en negar las diferencias entre los ámbitos público y privado (véase el ilustrativo texto de Bozeman, 1998). Para éste todas las organizaciones son públicas, en el sentido de que actúan socialmente4. En realidad, al negar la diferenciación de los dos términos, se justifica todo. Desde la provisión mercantil de la totalidad de los bienes públicos (incluso el gobierno), hasta la utilización de los recursos de todos (la accion del Estado y el financiamiento estatal) para socializar las pérdidas de las grandes corporaciones privadas o subsidiar la inversión empresarial. Así, la relación de fusión o de mezcla público-privado —lo que llaman en francés partenariat— se utiliza corrientemente para procurar fondos públicos, subsidios y compensaciones a las grandes corporaciones. Lo acabamos de ver en Canadá cuando a la principal compañía de fabricación aeronáutica, Bombardier, le fue avalada una megaventa de aviones a Air Canada. Si esta última tuviese dificultades para el pago, el Estado lo asumirá. Este tipo de acciones requieren nuevas regulaciones, desregulaciones y políticas públicas más flexibles que rompan las fronteras entre lo privado y lo público. Ello mismo facilita que los empresarios hagan un rápido tránsito a la política y al gobierno, y viceversa. 2.1. UNA RADICAL TRASNFORMACIÓN DE LOS PRINCIPIOS DEL GASTO PÚBLICO Otra gran causalidad de la transformación de la burocracia pública tiene relación con los aspectos fiscales del problema. Como lo destaca Genereux (2001: 135-149), esta política se desprende de que ahora existe un cambio signicativo en la comprensión del rol y la legitimidad del gasto público, dado primordialmente que las tendencias neoliberales intentan generalmente reducir el porcentaje del gasto público frente a los ingresos generales de la sociedad. Hoy en día se aboga por las exenciones fiscales como mecanismo preferente para incentivar la creación de riqueza privada; y se busca atenuar la dimensión del gasto público recortando los subsidios y compensaciones redistributivas. Como lo mencioné arriba, la teoría neoliberal supone que el crecimiento se induce si los empresarios no son “desmotivados” por altos impuestos; y que éstos disminuirán si se reducen sustancialmente las burocracias gubernamentales. Así que la principal prioridad al recortar gasto público es la destrucción y/o el debilitamiento de las burocracias estatales. Los gobernantes con frecuencia se presentan en las rendiciones de cuentas periódicas ante la opinión pública con los partes de “victoria” que muestran con orgullo, dando cuenta de la cantidad de empleos públicos suprimidos. O se construyen golpes 336 de opinión reduciendo demagógicamente los salarios de los altos ejecutivos, como si en esquemas de sociedad basadas en una cultura de la corrupción, estos dependieran y usufructuaran exclusivamente sus salarios. Y no de su enorme poder discrecional de contratación y asignación de recursos y cargos. 2.2. LA LIBERALIZACIÓN ECONÓMICA Y LA COMPETITIVIDAD DE LAS FUERZAS DEL TRABAJO En tercer lugar, la competitividad global, en condiciones de creciente y rápida estandarización de las formas tecnológicas de producción y gestión (por la vía de la mercantilización, sobre la base de los derechos de propiedad intelectual de la tecnología de diverso tipo), hace que la verdadera diferenciación para atraer la inversión de capitales la suministre el precio de la fuerza de trabajo. La deslocalización de las plantas de producción de bienes y servicios, de los medios de producción de las nuevas industrias del consumo masivo (J. Rifkin, 2000), centros financieros, etc., depende de la caída de los salarios. Ello en parte porque a nivel mundial los salarios de las burocracias públicas han constituido un parámetro para determinar el precio “justo” de la fuerza del trabajo al que aludía Adam Smith (Galbraith, 1989:92). Pues bien, la precarización laboral, condición sine qua non de la competitivad global, requiere romper esta columna vertebral del garantismo social. Del mismo modo, al romper la espina dorsal de las burocracias se le asesta un golpe mortal a los derechos de asociación sindical, los cuales los ejercen prioritariamente los funcionarios públicos. En América Latina, este ha sido un fenómeno dramático que ha alcanzado cotas inimaginables en países como la Argentina, Ecuador, Perú y Colombia, entre otros (OIT, 2002; Carrillo e Iranzo, 2000). La tasa de sindicalización en la región ha caído dramáticamente, aunque nunca fue siquiera comparable con el nivel promedio del capitalismo occidental avanzado (Norteamérica y Europa occidental). En 2002 en Chile se alcanzaba sólo el 11% del PEC (población económicamente activa); mientras que en Colombia era del 6%, en el Salvador del 5,6% (Cepal, 2003:198). El salario mínimo del sector formal de la economía tuvo una leve alza en los noventa en Chile y el Brasil, pero empeoró en el resto de la región, siendo los casos más dramáticos Ecuador, México, Argentina y Uruguay (Cepal, 2003: 199). 337 Las reformas tienen un impacto directo sobre el empleo. En efecto, el desempleo abierto o disfrazado es, según los analistas, el principal problema laboral de América Latina y se produce en parte como resultado de la recesión económica de la región en los 15 años recientes. El factor central explicativo de este fenómeno son los términos inequitativos de intercambio con el capitalismo globalizado. En realidad, durante los últimos años se han producido tasas de crecimiento económico muy bajas y con frecuencia negativas. Las políticas de flexiblizacion laboral ya señaladas, no han producido los efectos proclamados por los reformadores, que justifican la destrucción del empleo estatal con el argumento de propiciar el crecimiento del empleo en el sector privado al disminuir el gasto público en nómina y liberarse recursos para el desarrollo económico (inversión social e infraestructura). La Cepal, en el estudio citado, reconoce las dificultades de política en la región para “generar empleo productivo de buena calidad”; en parte derivada dicha situación de la baja productividad de las economías latinoamericanas, las condiciones relativamente precarias de la infraestructura básica, la baja calidad y escasa pertinencia de la educación, etc. La única alternativa “exitosa” ha sido la autogeneración de empleo, un eufemismo de la jerga económica para calificar el “rebusque” o trabajo informal, que implica entre otros impactos negativos el elevado procentaje de trabajo infantil (J. P. Pérez, 2003). 3. tendencias en la transformación de las burocracias públicas: trabajo alienado, precarización, pauperización Veamos los principales cambios y las tendencias actuales en relación con los cambios producidos por esta megapolítica pública transnacional en las burocracias estatales. 3.1. LA ALIENACIÓN LABORAL EN EL SECTOR PÚBLICO A algunos lectores podrá sorprenderles que aborde este tema. En general se supone que la alienación del trabajo es resultado de la apropiación del excedente en los sectores económicos productivos, que agregan valor. Este, por supuesto, en términos estrictamente económicos no es el caso de las 338 burocracias gubernamentales; salvo aquellas que actúan en las dimensiones empresariales del Estado. La dimensión económica del trabajo se encuentra en la dicotomía que diferencia en las sociedades mercantiles al valor de uso y al valor de cambio. A este respecto, el profesor Aktouf (2002) recuerda la diferenciación moral introducida por Aristóteles entre oikosnomos y crematística. La segunda no se interesa por el carácter útil-material de la producción humana (poiesis y autopoiesis), sino por la acumulación de riquezas, principalmente dinerarias. El propio Locke, que reconocía como fuente de riqueza al trabajo humano, legitimaría la diferenciación social, la polarización entre propietarios de riqueza y gentes del común que carecían de ésta, pese a trabajar duramente en la función legítima y equitativa de la compra de la fuerza de trabajo. Con la progresiva imposición de la filosofía económica del neoliberalismo, la robótica y la automatización, la apertura de los mercados, la deslocalizacion del capital, las maquilas, el retroceso del Estado de bienestar, etc., empezó a ser cada vez más creciente el fenómeno del desempleo estructural, que no tenía que ver con los ciclos económicos, entre prosperidad y depresión. Autores como Rifkin (1994) han advertido sombríamente la desaparición del mundo del trabajo. 0tros como Beck (2002) describen el mundo laboral actual como una serie de sociedades transestatales, divergentes en niveles de complejidad, pero unidas por el riesgo. La propia imagen pública del burócrata, desacreditada por el discurso neomanagerial, afecta hondamente la autoestima de los funcionarios; destruye su sentido de pertenencia y el arraigo del interés público. Es la imagen clisé del holgazán e inoperante; cuando no del manipulador corrupto, en su propio beneficio. Por supuesto, no niego que existan funcionarios holgazanes y corruptos. Pero el clisé hace de esta situación una generalización abusiva, una etiquetación que desvaloriza a la totalidad de los funcionarios públicos ante la sociedad y deslegitima sus reclamos y sus derechos. La alienación se expresa más profundamente en el problema del poder. El Estado burócratico constitucional moderno ha construido un tejido sofisticado de reglas de disciplinamiento organizacional. Este proceso lo podemos describir como un largo camino de institucionalización del deber ser de las burocracias públicas. Hoy —a la par con los avances tecnologicos— se 339 avanza correlativamente en el control sobre los empleados con base en sistemas de información sofisticados, bases de datos, mecanismos de evaluación del desempeño, interceptación de comunicaciones y violaciones recurrentes de la privacidad. La razón de Estado ha ganado nuevos ímpetus con la ideología unidimensional (Marcuse,1968) de la sociedad managerial, postpolítica. Además, si se lee con cierto detalle el discurso del NPM, se notará fácilmente que éste es externo. Es un conjunto ideológico normativo sobre la burocracia, asumiéndola como un objeto no sólo en el plano epistemológico, lo cual es evidente, sino —lo que es mucho más revelador— es la intencionalidad de manipulación, supresión, intervención, etc. Los salarios y el régimen laboral —por ejemplo— en el modelo del NPM no se deciden en negociaciones internas, ni siquiera en la relación de la burocracia con el parlamento y el poder ejecutivo. Son decididos unilateralmente sin su participación. La forma como se diseña el trabajo administrativo y se lo ejecuta tiene en la actualidad la misma connotación de absoluta externalidad. 3.2. LA PRECARIZACIÓN LABORAL DEL FUNCIONARIADO La incertidumbre y el mundo del riesgo (Beck: 2002) son las percepciones dominantes para los funcionarios públicos en esta época de reformas neoliberales y reestructuraciones. La reingeniería, los ajustes organizacionales, las reformas administrativas, etc., son pensados y ejecutados —por regla general— para reducir las plazas de empleo y aumentar la precarización del trabajo. La primera forma de precarización es la ruptura de la estabilidad laboral, el quiebre del contrato a término indefinido. La segunda, el desvertebramiento de las asociaciones sindicales, cambiando las reglas laborales, facilitando los despidos “sin justa causa” o la persecución abierta. En Colombia se cuentan por centenares los sindicalistas asesinados y desaparecidos en los últimos años; pero un fenómeno semejante —aunque en menor escala— se da hoy en México, Guatemala, Perú, en el África Ecuatorial y en ciertas áreas de Asia. Anotemos que el espectro de la tendencia a la disminución del porcentaje de la población laboral sindicalizada ha caído en los países desarrollados tales como el Canadá, los EE.UU, Australia, Europa occidental (Informe OIT, 2003). Aunque la persecución directa contra los sindicatos no es 340 allí la causa central, sí lo es la flexibilización de los contratos, el número creciente de funcionarios en régimen de outsourcing y las privatizaciones. La precarización tiene relación también con el suministro y la calidad de los medios de trabajo. Piénsese —al menos en América Latina— en los juzgados donde los empleados carecen de papelería y computadores; en los hospitales donde no hay adecuados suministros y equipos médicos; o en las escuelas y colegios públicos, con infraestructuras inadecuadas; obsolescencia cuando no nulidad tecnológica, etc. Incluso en muchos países del “tercer mundo” las propias fuerzas de policía están en inferioridad de medios para asumir la protección ciudadana y el combate a la criminalidad (Carrillo e Iranzo, 2000). 3.3. LA PAUPERIZACIÓN DE LAS BUROCRACIAS Como resulta evidente de acuerdo a lo ya descrito, las vías para la pauperización son múltiples. En primer lugar, se manifiesta en la caída de los ingresos reales de los funcionarios públicos. Éstos son las primeras y preferenciales víctimas de los ajustes fiscales. En Latinoamérica, por ejemplo, ya es costumbre la congelación salarial o los ajustes de los pagos por debajo de la inflación, en medio de la retórica de la austeridad y la lucha contra el despilfarro. Se le dice a la opinión pública y se la manipula con la complicidad de los massmedia ligados al gran capital, que esos recursos “liberados” de la inutilidad del gasto burocrático se destinarán a la “inversión social” (Evans, 2002). Los cambios en los regímenes de pensiones y de cesantías están en la misma línea de actuación. Más tiempo para pensionarse (edad mínima aumentada y tiempo maximizado de cotización), se rebaja el monto real de las pensiones, además de la pérdida en toda América Latina de la retroactividad de las cesantías. Tanto éstas como las pensiones se han privatizado en su gran mayoría, generándole enormes rentas a las grandes corporaciones financieras y drenando —allí sí— cuantiosos recursos estatales, al reconocerse los estudios actuariales de las cotizaciones acumuladas, las cuales los ministerios de hacienda y finanzas giran prontamente a los fondos, aunque sea todavía lejana e incierta la fecha de jubilación de los cotizantes. La cuestión pensional y las prestaciones laborales son hoy igualmente parte de los objetivos de la modernización del Estado, ya sea en el Quebec, 341 Francia, Alemania, los países de la antigua órbita soviética. Del mismo modo, los conflictos con los sindicatos del sector público se han incrementado en América Latina, Europa occidental, el Canadá y ciertas áreas de Europa oriental. Otro tópico de pauperización son los servicios médico-asistenciales que, al mercantilizarse, han reducido simultáneamente la cobertura, calidad, amparo a las familias. Han encarecido las cotizaciones y cobran servicios que antes eran incluidos integralmente. La precarización también implica pauperización; los trabajadores contratistas, adicionales, etc., no tendiendo estabilidad, “entran” y “salen” del sistema; cotizan, pero rara vez se pensionan y pierden muchos de los derechos de que gozan todavía los trabajadores más antiguos. Muchas veces la lógica perversa de las reformas administrativas, que deben respetar los llamados derechos adquiridos —protegidos por convenios del tipo OIT (Organización Internacional del Trabajo)— supone negociar el sacrificio de la siguiente generación de empleados. Estos perderán todos los rubros y normas que se conservan a los actuales, para comprar la aquiesencia de los últimos. Aquí, igualmente mencionemos que el grado de cobertura y eficiencia de los sistemas médico-asistenciales en los propios EE.UU está muy lejos de los estándares europeos o del Japón, y una porción significativa de su población laboral carece de aseguramiento (Sennett, 2000). 4. Para las teorías manageriales dominantes, ¿es viable la burocracia estatal? Parecería claro el contenido arrasador de la política neoliberal contra las burocracias públicas. Sin embargo, casi 20 años después de sus ejecutorías, el Estado a nivel mundial —sin excepción— sigue siendo la principal fuerza económica de las sociedades, el primer empleador; y se habla y actúa —vista la “lucha antiterrorista”— en pro de su fortalecimiento, al menos en sus componentes policiales y coercitivos (Negri y Hardt, 2004: 211-214). ¿Hasta dónde llegan las metas desvertebradoras del New Public Management? Una revisión rápida pero representativa de la literatura managerial en boga nos revela pistas de sus actuales proposiciones y rectificaciones. B. Guy Peters, profesor de la Universidad de Pittsburgh y una de las voces más calificadas del Management público en los EE.UU, afirma: 342 “(...) las reformas señaladas han logrado un notable éxito mejorando la eficiencia de la administración cotidiana. Pero también han creado una serie de problemas significativos, cuando analizamos la función gubernamental en una perspectiva más amplia (...) fragmentando las administraciones públicas y generando un estilo altamente incoherente de gobierno. Ahora podemos observar las reacciones suscitadas por la excesiva descentralización y el intento por recuperar una más centralizada misión de la gobernanza hacia el centro. Requerimos una visión más coherente del futuro del Estado y de la sociedad” (B. G. Peters: 2003). Peters insiste sobre la base del anterior argumento en la necesidad de construir una agenda de prioridades estratégicas para los gobiernos. Sería necesario aplicar las metodologías del Management estratégico en los niveles decisores más elevados de la administración pública. La idea principal de este autor es que se necesita reequilibar la balanza colocando al Estado en el centro. De allí la opción del “justo medio”: ni mucho Estado, ni tampoco el desmonte total del Estado. Para ello deberían tenerse en cuenta los niveles diversos de complejidad de las administraciones públicas, situándolas en espacialidades y temporalidades precisas. No dándole prioridad a recetas genéricas. “Las concepciones de buena gobernanza son cultural e históricamente contingentes” (Ibíd). Otra autoridad de gran renombre —como quiera que se le reconozca su rol como fundador del neoinstitucionalismo—, el profesor sueco Johan P. Olsen, señaló recientemente que existen buenas razones para “redescubrir la burocracia”: “(...) una razón central es que el escenario de la burocracia como un dinosaurio (...) es equivocado e insuficiente (...) La burocracia tiene un rol como custodia institucional de los principios constitutivos de la democracia; los principios precisamente de la racionalidad procedimental. Pero debe actuar bajo esquemas concurrentes con otras agencias implicándose en criterios de competitividad y éxito. La burocracia posee además un rol como eficaz herramienta para los legisladores y la democracia representativa. Está positivamente implicada con los aportes sustantivos que concretan los valores en las democracias con343 temporáneas, en unos casos más que en otros. La juridificación de muchas esferas de la sociedad, el desarrollo de los derechos humanos, el incremento de la diversidad, la carencia de bases comunes para definir metas sociales y renovar las demandas de rendición de cuentas, todos estos factores - en suma - han contribuido al emergente interés en los aspectos burocrático-legales de la administración y del gobierno” (J. P. Olsen: 2004). Un par de breves menciones adicionales. En un tono semejante, el científico político Peters Evans propugnaba por la triple “hibridación” como la estrategia administrativa más adecuada. Ello quiere decir combinar: 1) las fuerzas y capacidades del Estado, 2) la democracia deliberativa, y 3) las fuerzas o señales del mercado. La democracia deliberativa requiere, en la visión de Evans, el complemento del mercado. Y sobre estas dos bases puede actuar adecuadamente la burocracia de estirpe weberiana. Al respecto, formula un balance crítico: “Los esfuerzos exagerados para lograr que la administración imite los mercados no sólo amenazan con minar la capacidad burocrática sino que, además, traen consigo sus propias e intrínsecas irracionalidades e ineficiencias (...) el peso creciente de los mercados financieros internacionales en la determinación de decisiones económicas nacionales y el surgimiento del neoliberalismo global como la doctrina que define la política económica internacional dio como resultado una serie de poderosas limitaciones a la trayectoria de la reforma del Estado, especialmente en América Latina. Los administradores que deseen experimentar con reformas que enfaticen cualquier otro factor que no sea la confianza creciente en el mercado, temen ser castigados precisamente por los mercados, e indudablemente muchas veces lo son. Lo que denominé monocultivo institucional es un potente impedimento para lograr la hibridación” (Evans, 2002). Finalmente, cierro con un tema conexo este recorrido rectificatorio. Los politólogos Peter De León y Mark T. Green (2000) de la Universidad de Colorado, cuestionaron la teoría del New Public Management dado que al permitir explícitas licencias para que los funcionarios gubernamentales 344 operen de manera relativamente independiente de una supervisión directa, más bien como agentes libres (free lance agents), semejan una suerte de cowboys en la periferia del control burocrático de tipo weberiano. Aunque se gana en agilidad empresarial, ellos cuestionan la capacidad del modelo del NPM para prevenir la corrupción política y administrativa, o siendo más realistas, para minimizar su impacto. Si además reconocemos los niveles inferiores de remuneración dados a trabajos de tipo comparable —entre los sectores público y privado— encontramos que la discrecionalidad en sus funciones y poderes, abre las compuertas al ajuste ilegal de los ingresos de funcionarios en los diferentes niveles, principalmente los decisorios. Obviamente, estos críticos identifican un punto neurálgico del modelo. Al alejar la política, han con ello mismo, alejado de la gestión pública la capacidad de control ciudadano. Para no hablar de la función difusa hoy en día de los organismos de control (contralorías, tribunales de cuentas, de ética, jueces de la función pública, etc.), dado que muchas inversiones y activos del sector estatal se han transpuesto a los términos del derecho privado para ganar “agilidad”. Una somera revisión a las reseñas de los medios de comunicación y reportes periodísticos de los últimos15 años, en diversas latitudes del mundo, nos evidencia la impresionante tendencia al incremento de la corrupción como fruto del maridaje abierto entre los negocios y las administraciones públicas (Stiglitz: 2003). En suma: lejos estamos en la actualidad del optimismo que respiraban los promotores del NPM hace 15 ó 20 años. El tiempo transcurrido permite ciertamente hacer un balance y la mayor parte de sus ejecutorías son desastrosas. Aunque algunos de los autores mencionados reconocen mejorías en la llamada competitividad, inmediatamente matizan este punto enfatizando las consecuencias inopinadas, las “externalidades” negativas de la aplicación de estas políticas. En general, el Estado se ha desvertebrado —allí donde se siguieron sus doctrinas al pie de la letra— (léase: Argentina, Ecuador, África Ecuatorial, Bolivia, Colombia, Indonesia, Filipinas, entre otros; véase sobre este tema: Chossudosky, 2004; Stiglitz, 2003) generándose impactos graves para los propios programas de desarrollo económico y la inserción en la globalización. Los países más exitosos en esta línea de acción son precisamente aquellos como Corea del Sur, Japón, los países escandinavos, 345 Alemania, Taiwán y los otros tigres asiáticos, en los que el Estado sigue siendo un actor económico de primer nivel. El ascenso meteórico de China continental es, por supuesto, un caso mucho más claro, si se quiere, de capitalismo de Estado, con la puesta en escena de un plan estratégico —soportado en la inversión pública a gran escala— para lograr un puesto de primera línea en la economía mundial. Estos logros, en el marco de una exacerbada competencia entre los países y las corporaciones multinacionales, han dejado en claro la simplificación excesiva del discurso del NPM. Autores como Bob Jessop (1999) y Poulantzas (1973) advertían sobre la necesidad del Estado en las sociedades capitalistas desarrolladas para acometer, en el largo término, las tareas de la acumulación y la reproducción ampliada del capital. Ello se corrobora en la actualidad y tal fenómeno —en mi opinión— está en el centro de las rectificaciones por parte de los principales ideólogos del neomanagerialismo público. De manera adicional, a sus efectos discutibles en la esfera económica, mencionemos sus impactos sobre la gobernabilidad, la estabilidad institucional, el aumento ya señalado de la corrupción, el desborde de la criminalidad del globalismo en la sombra (P. Lock, 2003). Estos factores se aúnan a las crisis sociales y al aumento inconmensurable de la pobreza y la exclusión de clases, regiones y países. Aunque estos tópicos no entran en las sensibilidades del neomanagerialismo, salvo por sus efectos indirectos sobre la estabilidad institucional. 5. Entonces, ¿cuál es la funcionalidad actual de las burocracias públicas? La teorización contemporánea sobre la burocracia no es simplemente un dato histórico en la evolución de las Ciencias Sociales. Cada vez resalta la pertinencia de su completo y matizado análisis paradigmático, pues necesitamos capturar los trazos esenciales de tal fenómeno organizacional e institucional, imbuído de la racionalidad del capitalismo industrial y postindustrial, moderno y posmoderno. No es este el lugar para rastrear y describir in extenso las consecuencias de 346 estas proposiciones y sus conexiones con la teoría weberiana5. Empero, quisiera retomar recientes interpretaciones del llamado fenómeno burocrático que, a mi juicio, resultan decisivas para responder la pregunta arriba señalada. La funcionalidad de las burocracias públicas —más allá de los topos clásicos del weberianismo— consiste primordialmente en un esquema de actuación tríptico. MEDIACIÓN Gestión burocrática NEUTRALIDAD IMPARCIALIDAD 5.1. NEUTRALIDAD E IMPARCIALIDAD En primer término, ella posee, como ninguna otra forma organizacional, la capacidad de rendir de manera eficiente la tarea de mediación entre el Estado y la sociedad (incluído el mercado como realidad socioespacial de intercambio e interdependencias). En tal sentido, vale la pena diferenciar claramente estos tres términos del tríptico en mención. Veamos a continuación las características de la mediación. Literalmente dicho, mediar significa ocupar un espacio de conectividad entre dos o más fuerzas sociales. Clásico caso: las oficinas e inspectores de asuntos laborales cuando se presentan conflictos entre patronos y sindicatos. O las agencias de protección del consumidor en otros ámbitos de conflicto. Estos son ejemplos que escojo de forma deliberada porque en ambos casos se supone la asimetría ontológica de la relación. Tienen mucho más poder en las sociedades capitalistas —por regla general— los patronos y productores que los trabajadores o los simples consumidores. Por lo tanto, cuando el Estado garantista de derechos media, generalmente lo hace para proteger los derechos potencialmente más vulnerables de la parte débil de la relación. Es decir: no se trata de acciones basadas en el principio de neutralidad, sino en el de imparcialidad. En este principio, se encuentra la base legítima de la función pública. En tanto que la neutralidad puede equivaler a la indiferencia, y por supuesto, al laissez faire a la liberalización de las políticas. Las ideologías en boga hoy sobre la gobernanza, se apoyan con frecuencia en la neutralidad gubernativa dándole al Estado principalmente una asig347 nación de tareas, en tanto máxima instancia social de coordinación. Pero esta tarea —aunque posible y conveniente— no define la legitimidad de la existencia del Estado social de derecho. La esencia de la burocracia gubernativa, su permanencia exitosa por más de un siglo desde que surgió el Estado moderno constitucional, estriba en esa función de mediación sin la cual las reglas de la concurrencia, la lógica misma del mercado como institución, se debilitan seriamente. Esa fue la clave del éxito del modelo de Estado empresarial. Este asumía la provisión de ciertos bienes y servicios a la sociedad sobre la base de una toma de posición política. Se suponía así la primacía del interés nacional-estatal a través de la monopolización de ciertas actividades productivas en recursos energéticos estratégicos, o en el campo de las comunicaciones, los servicios públicos domiciliarios, la industria de armamentos, etc. Los fallos de las privatizaciones —extendidas más allá de las áreas típicamente empresariales del Estado— configuran una situación de crisis estructural en la medida en que el aparato estatal pierde su relacionalidad de arbitraje y mediación universal. Además, en el modelo hoy en auge del partenariat se hace del Estado “juez y parte”, quebrantándose así los principios universalistas y la propia equidad. Del mismo modo se abren las compuertas de la patrimonialización de lo público-estatal y, por ende, de la proliferación del dolo y la corrupción administrativa. Es comprobable, ciertamente, la capacidad de las burocracias para tranformar los fines sociales o públicos en finalidades corporativas. Pero, precisamente, como Weber lo senaló, el contrapeso a esta tendencia son los numerosos elementos de la juridización de la función gubernativa, que limitan —en el modelo clásico bien entendido— la discrecionalidad que sí da el NPM. Así que, en fin de cuentas, con el neomanagerialismo se facilita mucho más la tendencia nociva al corporatismo. Además, porque el público-privado, al romper la función de arbitraje neutro, permite la captura del interés social por parte de los grupos económicos que establecen “alianzas estratégicas” con los gobiernos. Recuérdese aquí que la contratación administrativa es el principal dispositivo discrecional de corrupción, corporativismo burocrático 348 y, también, de favorabilidad asimétrica para los grupos de presión. La juridización del Estado social de derecho se expresa en la creación de un espacio por el ejercicio eficaz, y no sólo retórico, de los derechos ciudadanos (Habermas,1998). No es fácil, por supuesto, lograr que el Estado sea un árbitro neutral, que sus decisiones se den sin que las cartas del juego estén marcadas, teñidas de favoritismo. La sociología crítica americana (Wrigths Mills y otros), desde mediados del siglo pasado, abundó en demostraciones del poder de las elites y los grupos de presión sobre las decisiones de política y la asignación de recursos y beneficios. En este sentido el análisis de la lógica burocrática en términos de políticas públicas subraya que, más allá de la autonomía interna y la capacidad de fijar iniciativas de política, los Estados —inevitablemente— deben responder a las demandas y presiones que provienen de la sociedad o de otros actores estatales o transestatales. Como en un prisma el Estado a través de sus decisiones y actuaciones refleja y expresa, al mismo tiempo, las correlacciones de fuerzas entre los distintos grupos sociales y los factores de poder que interactúan —en esquemas sistémicos— entre sí. 5.2. DINÁMICAS INTERNAS EN LOS PROCESOS DE BUROCRA TIZACIÓN Comienzo este punto recordando que la burocracia pública no es una totalidad uniforme, homogénea. Ella posee niveles o jerarquías claramente delimitadas. En general, una elite burocrática preside el proceso controlando las principales decisiones del aparato burocrático y actuando como un núcleo de enganche con el aparato propiamente político del Estado (presidencia y ministerios, gobiernos locales y regionales, parlamentos, etc.). Dado que la burocracia constitucional asume sus funciones en una maraña de reglas imbricadas, el poder decisional autónomo no le pertenece. Ella es una ejecutora de las decisiones asumidas en la esfera política propiamente dicha. Su ontología, su razón de ser, la constituye el proceder técnico especializado de la implementación de las decisiones. La encuentra en su aterrizaje en el aparato ejecutor de tipo procedimental, anónimo, prefijado, etc. 349 En tiempos recientes, un fenómeno altamente comprobable en el mundo entero es la circularidad de la capa superior, entre el dominio managerial y el público. Así, un cuadro de alta responsabilidad político-administrativa está un tiempo en la empresa privada y luego transita a la dirección administrativa de lo público y viceversa. En parte porque el sistema de nominación es, a ese nivel, cada vez más discrecional y sometido al lobbying. O se transita desde la esfera de lo político a la dirección adiminstrativa: el político profesional “aprende” a administrar en la práctica, de manera empírica, sólo para obtener beneficios de diverso tipo (asignación de recursos, manejos de las nóminas burocráticas, favorecimientos de causas electorales, regionales, etc.) Y por supuesto el homo politicus retorna a la política electoral con un plus que reivindica ante el electorado: es el advenimiento y puesta en escena de la “política gerencialista”. En un segundo nivel del aparato burocrático encontramos la capa de funcionarios con saberes especializados, que configuran el corazón mismo del sistema burócratico. Son la clave de la permanencia y durabilidad de la memoria institucional, generalmente basada en los principios meritocráticos como ingreso por concurso universal (pruebas abiertas a quien cumpla los requisitos, en absoluto pie de igualdad formal); ascensos por méritos y antigüedad; relativa democratización en la toma de decisiones de tipo técnico (relación directa entre saber y poder). Este es un nivel que tiende a fortalecerse en la medida en que las sociedades de comienzos de siglo son cada vez más complejas y abiertas. Inevitablemente ello obliga, demanda, la fuerza técnica del aparato. Este es un escenario donde las rectificaciones se presentan de manifiesto. Por ejemplo, en los EE.UU antes del 11 de septiembre de 2001 se había privatizado la seguridad aeroportuaria. Y en muchos países el colapso de servicios públicos claves que se han privatizado implica, ahora, una gran presión social —igual proveniente de las elites empresariales que no son directas beneficiarias de estos negociados— para su retorno al sector público. Finalmente, en la base del sistema burocrático, se encuentran los empleados que desempeñan tareas mecánicas, operacionales, por llamarlos del mismo modo que a sus pares en el mundo laboral industrial y privado. Esta última capa tiende a ser cada vez menos importante por el doble cambio derivado de las reformas 350 neoliberales de adelgazamiento y reingeniería de los gobiernos; y, en segundo lugar, por el enorme impacto de las tecnologías: sistematización informática, estandarización de procederes, siendo reemplazados los burócratas de bajo rango por medios robóticos y analógicos. La destrucción de las burocracias en este escalón es un fenómeno mundial prácticamente arrasador. Por supuesto, impactando duramente la capacidad gubernamental de absorber parcialmente las demandas de empleo de quienes llegan por vez primera al mundo laboral (lejos estamos de los buenos tiempos del keynesianismo que, al menos en este punto, no volverán). Las relaciones entre estos tres niveles son uno de los aspectos que merece amplio estudio en el sector público, para determinar las tendencias actuales en la interrelación de los mismos, y los grados de la posible democratización del proceso decisional (Aktouf, 1991; Arellano y otros, 2003) en las acciones del gobierno. 5.3. EL FUTURO DE LAS BUROCRACIAS PÚBLICAS En estos momentos nos encontramos en la encrucijada de un gran debate sobre la suerte de las burocracias, donde el péndulo oscila entre grandes alternativas. El sentido de la acción política frente a las burocracias será: ¿fortalecerlas?, ¿destruirlas?, ¿transformarlas? Cuestiones derivadas son preguntas como: ¿qué tipo de trabajo han de desempeñar?, ¿bajo qué modalidades de control político y social se espera éstas rindan la función de concreción de los intereses sociales? Esta discusión —aparte de sus componentes técnico-científicos— revela las preferencias ideológicas y morales de sus participantes. No hay, por ello, lugar para consensos sociales. Las tendencias neoliberales más radicales continúan su pregón en pro de la destrucción —hasta donde les sea posible— de las burocracias gubernamentales. Los sindicatos, buena parte de la izquierda política y del propio centro, consideran válida y necesaria la funcionalidad de las burocracias y su preservación a ultranza. Empero, difícilmente encontramos defensores incondicionales del statu quo. Y ello porque las burocracias siempre han generado resistencias sociales. En cierta forma, los ciudadanos corrientes nos sometemos a su férula, a sus imperativos, muchos de los cuales se perciben como inadecuados, injustos, excesivos, etc. Además, no olvidemos que las burocracias —aún en las de- 351 mocracias liberales— no resultan de una emanación democrática. No elegimos a los funcionarios que las constituyen, a nuestras autoridades policiales y militares, a los cuerpos jurisdiccionales, etc. De otro lado, la complejidad de las alternativas tiene igualmente relación con las áreas donde éstas se desempeñan, con los contextos culturales, históricos y sociales en los que han arraigado unos u otros modelos de gestión burócratica. Y, finalmente, están los conflictos de intereses, que son, en el fondo, los grandes propulsores de las políticas públicas respecto del fenómeno burocrático. La política se condensa a la manera de un cristal de luz prismática— en las luchas por el poder burocrático y en su instrumentalización relacional o contextual. En lo que concierne al campo concreto de los discursos manageriales, podríamos sintetizar el estado actual del mismo señalando que el NPM ha perdido su capacidad de convocatoria y unificación dentro del universo de las comunidades académicas y científicas. Algunas evidencias he citado de ello en la sección anterior. Pero sigue siendo dominante a nivel de las agencias multilaterales y preside aún —ideológicamente hablando— los planes de ajuste del sector público. Como lo podemos comprobar al mirar los contenidos de estas políticas en el Canadá, México, Colombia y, parcialmente, en la propia Unión Europea. Obsérvense al respecto ciertas de las reglas de su constitucionalidad elaboradas con el influjo de la third way (Blair y Giddens y de V. Giscard D-Estaing). El stablishment se divide en la actualidad, revelando con sus posturas al respecto, la capacidad de los diversos sectores para actuar de manera estratégica. Un sector reconoce la necesidad de pensar en el largo plazo, puesto que los proyectos sociales y económicos requieren una definición coherente sobre el fortalecimiento del aparato del Estado y la burocracia a éste ligada. En ciertos contextos —primordialmente financieros y especulativos—, predominan las miradas de corto plazo, el rentismo inmediatista, que resulta generalmente ajeno a las preocupaciones ecológicas, a la sustentabilidad de un modelo de sociedad, a la calidad del trabajo de las burocracias —considerando a sus miembros como ciudadanos con legítimos derechos en su específico campo de actuación—. La suerte de las burocracias, desde luego, no depende de un sólo factor, el del poder establecido neoliberal. También concurren a la resultante tendencial los factores objetivos, tecnológicos, las megatendencias no decisionales de las 352 transformaciones sociales. En lo que concierne a la transformación de la burocracia administrativa (Kallinikos, 2004), la experiencia que he analizado y vivido como consultor y agente de cambio organizacional en diversas organizaciones estatales en Colombia, así como una revisión sucinta de las experiencias documentadas y la literatura sobre el tema, me afirman en la convicción de que, en buena medida, la burocracia contemporánea mantiene rasgos comunes a los definidos por la teoría clásica. En este orden de ideas, cabe resaltar que un Estado regulador como el que hoy en día se promueve implica, en forma cada vez más creciente, la adopción de numerosas reglas y principios de arbitraje de los conflictos entre los particulares y entre las diversas instancias del Estado, de tal forma que una hermenéutica legal cada vez más hipertrofiada implica la configuración de poderes burocráticos en campos tales como la justicia, las relaciones comerciales, los derechos de propiedad, etc. Un fenómeno burocrático adicional consiste en que con los procesos de globalización y de interdependencia, ha surgido —con renovados bríos— una nueva burocracia internacional que se apoya estructuralmente en el declive de las soberanías nacionales estatales. Por supuesto me refiero al papel que juegan los organismos multilaterales tales como el Banco Mundial, FMI, BID, BIRF, OEA, el Parlamento Andino, etc. Igualmente, en el sector público sigue siendo vigente la jerarquización y subordinación en líneas de mando y la división entre el ámbito decisional o político y las esferas propias a la tecnocracia. Ésta, a su turno, ha ganado espacios de autonomía relativos derivados de lo que el gurú del Management, Peter Drucker, ha denominado “la sociedad del conocimiento”. Muchas de las actividades de la burocracia contemporánea tienen un alto nivel de sofisticación, el uso de instrumentales completos, de metarrelatos científicos, de lenguajes codificados y esotéricos, así como el manejo de una enmarañada red de intercambios entre las organizaciones en el espacio y en el tiempo (E. Varela, 2005a). Finalmente —pero no por ello lo menos importante—, las dinámicas de la acción política, los procesos decisionales de las políticas públicas, son, en gran medida, influidas por las movilizaciones ciudadanas, el debate público, el papel del sindicalismo estatal, la visión influyente de la academia, etc. 353 Estos procesos, aunque discurren por vías específicas en cada sociedad, se encuentran interconectados en red, por así decirlo. La suerte de las burocracias es por ello una de las cuestiones decisivas en las actuales derivas de la globalización de políticas. La defensa del papel institucional de las burocracias, como necesarias mediadoras en la concreción de las demandas sociales y del interés público, ha vuelto a ganar —lentamente todavía— un lugar en los medios académicos (Du Gay, 2000; J. Kallinikos, 2004). Ello se refleja igualmente en el tono rectificatorio y matizado del actual discurso del New Public Management. Empero, el debate público y las ejecutorías reformatorias sobre las burocracias gubernamentales constituyen hoy en día uno de los temas más sensibles para el futuro del garantismo social del Estado de derecho en el proceso actual de mundialización. El eje direccional de las políticas dominantes sigue vigente y tiende a su debilitamiento y/o desmonte. Allí no hay consenso alguno y la suerte de las burocracias públicas, su sobrevivencia dependerá: 1) de su funcionalidad estructural al sistema, 2) de su capacidad de defensa, y 3) de la legitimidad de la burocracia como una fuerza social reconocida y reclamada. Este punto se desprende de que la propia sociedad civil le otorgue a ésta un rol central como vehículo de democratización y freno a las arbitrariedades patrimoniales del rentismo financiero. Este último aspecto —sin desdeñar los primeros— es, en mi percepción, el eje de la continuidad de las burocracias gubernamentales en los tiempos por venir. 354 NOTAS Así, por ejemplo, en Colombia la principal crítica de los financistas de esta última ola tiene como blanco a la Constitución política redactada por una Asamblea Constituyente en 1991. En dicha legislación, el garantismo de los derechos sociales elevó al rango constitucional el gasto público social y las transferencias a los territorios subnacionales. En años recientes el Congreso, bajo el influjo del ideario neoliberal y de los organismos multilaterales, ha flexibilizado (es decir, recortado) la mayoría de estos rubros. 2 En una crítica descarnada de tales excesos, afirma Stiglitz (2002a) que en numerosos países el efecto simple y directo fue el privilegiar “el desmantelamiento de los activos y no la creación de riqueza (implicando) el exagerado enriquecimiento de unos pocos individuos en tanto que las masas se veían empujadas a la pobreza atroz.” 3 Esta política, aunque aligera la “carga” del Estado central, que, en general, bajo el FMI es un garante de la deuda externa, implica una profundización de las diferencias entre las regiones ricas y las pobres. Se acentúan, por tanto, las tendencias inherentes al capitalismo de su desarrollo desigual (S. Amin, 1990). 4 Las puntualizaciones aquí formuladas suponen que el lector ha abordado los análisis sobre la relación público-privado que he desarrollado en los capítulos primero, tercero y sexto del presente libro. 5 Sobre Weber y la burocratización véase un análisis en un plano teórico en el capítulo 2, y en su dimensión socio-histórica en el apartado 3 del capítulo 3 del presente libro. 1 355 PÁGINA EN BLANCO EN LA EDICIÓN IMPRESA BIBLIOGRAFÍA ADORNO, Theodor. “Dialéctica negativa”. Madrid: Taurus. (l989) AKTOUF, Omar. “La stratégie de l’Autruche”. Montréal: ECOSOCIETE. (2002) —————. “Adhésion et pouvoir partagé”. Dans Gérer et Comprendre, Annales Des Mines. (1991) ALBERT, Michel. “Capitalisme contre capitalisme”. Paris : Ed. du Seuil. (1991) ALBI, Emilio. “Público y Privado - un acuerdo necesario”. Barcelona: Ariel. (2000) ALTHUSSER, L. “Para leer El Capital”. México: Siglo XXI. (1982) ALLAIRE, Yvan y FIRSIROTU, Mihaela E. “Les fondements stratégiques du phénomène de mondialisation”. En : Revue internationale de Gestión, Montréal, Vol 15, No 3, Sept. (1990) ARELLANO, David y Otros. “Retos de la Profesionalización de la Función Pública”. Caracas-Venezuela, CLAD. (2003). ARENDT, Hannah. “Los orígenes del totalitarismo”. Madrid: Alianza. (1976) —————. “La condición Humana”. Barcelona: Paidós. (1993) ARIES, Philippe y DUBY Georges (compiladores). “Historia de la vida privada”. Argentina: Editorial Taurus. (1990) AUBERT, Nicole y GAULEJAC, Vincent. “El coste de la excelencia ¿Del caos a la lógica o de la lógica al caos?”. Barcelona: Ediciones Paidós Ibérica. (1991) BAKUNIN, Mijail A. “Escritos de filosofía política”. Barcelona, Ediciones Altaza. (1994) BANCO MUNDIAL. “World development report”. New York. (1991) —————. “Poverty”. New Cork. (1990) BARBERO, Martín Jesús. “Pre-textos -Conversaciones sobre la comu357 nicación y sus contextos”. Cali: Ediciones Universidad del Valle. (1995) BARZELAY, Michael. “La nueva gerencia pública. Un ensayo bibliográfico para estudiosos latinoamericanos (y otros)”. En Revista del CLAD -Reforma y Democracia - Caracas, pps. 7-66. (2001). http://www.clad.org. ve/reforma.html BAYARDT, J. F. “Le Pidgin de la Societé civile”. Paris: ALTERNATIVES ECONOMIQUES, No 190, Marzo. (2001) BECK, Ulrich. “Pouvoir et Contre - Pouvoir a la ère de la Mondialisation”. Paris: Flammarion. (2003) BERTRAND, Badie. “Un mundo sin soberanía. Estados entre artificio y responsabilidad”. Colombia: Tercer Mundo Editores. (2000) BELTRAME, Pierre. “Los sistemas fiscales”. Barcelona: Ediciones Oikos - tau. (1977) BERGER, Peter L. y LUCKMANN Thomas. “La construcción social de la realidad”. Buenos Aires: Amorrortu Editores. (1968) BETTELHEIM, Charles. “The great leap backward”. New York: Monthly Review Press. (1978) BHARO, Rudolf. “La Alternativa- Crítica del Socialismo realmente existente”. Barcelona: Ediciones Materiales. (1979) BOBBIO, Norberto. “Estado Gobierno y Sociedad -la gran dicotomía de lo público y lo privado”. México. FCE. (1998) —————. “El futuro de la democracia. México: F.C.E. (1994) BODINO, Jean “Los seis libros de la república”. Barcelona: Orbis. (1987) BOLIVAR, Simón. “Obras completas”. Caracas: LECUNA V. (compilador). (1951) BOURGOIS, Bernard. “La pensée politique de Hegel”. París: P.U.F. (1969) BOZEMAN, Barry. “La gestión pública. Su situación actual”. México: Fondo de Cultura Económica y Colegio Nacional de Ciencias Políticas y Administración Pública. (1998a) —————. “Todas las organizaciones son públicas”. México: Fondo de Cultura Económica y Colegio Nacional de Ciencias Políticas y Administración Pública. (1998b) BRADING, David. “De la monarquía católica a la república criolla, 1492 - 1867”. Mexico: F.C.E. (1991) BRAUDEL, Ferdinand. “The Mediterranean and the Mediterranean World in the Age of Philip II”. Berkeley: Univ. of Cal. Press. (1966) BUCHANAN, Allen. “Efficiency arguments for and against the market” In: “Ethics efficiency and the market”. New York: Rowman and Allan Held. (1985) CALHOUN, Craig. “Habermas and the public sphere”. Cambridge: MIT Press. (1992) CÁRDENAS, Jorge Hernán. “Alcances y razón de ser de la privatización”. Bogotá: Economía Colombiana. No. 236, pag. 33-42, Julio-Agosto (1991) 358 CARLYLE, A. J. “La libertad política- Historia de su concepto en la edad media y en los tiempos modernos”. México: F.C.E. (1942) CARMAGNANI, Marcello. “Estado y mercado: la economía pública del liberalismo mexicano 1850-1911”, México: FCE, 1994. —————. “Estado y sociedad en América latina 1850-1930”. Barcelona: crítica. (1984) CARRACEDO RUBIO, José. “Paradigmas de la política - Del Estado justo al Estado legítimo”. España: Editorial Anthropos. (1990) CARRILLO, Jorge e IRANZO Consuelo. “Calificación y competencias laborales en América Latina”. En: “Tratado Latinoamericano de Sociología del Trabajo”, DE LA GARZA T. Enrique (comp.). México: Fondo de Cultura Económica. (2000) CASSIRER, Ernest. “El mito del Estado”. México: F.C.E. (1945) CASTRO CARVAJAL, Beatriz (editora). “Historia de la vida cotidiana en Colombia”. Bogota: Norma. (1996) CEPAL (Comisión económica para la América Latina-ONU). “Panorama social de América Latina”. Santiago de Chile: Agosto (2003) CERRONI, U. “La libertad de los modernos”. Ediciones Martínez Roca. Barcelona. (1972) —————. MILIBAND. R. et al. “El Derecho y el Estado”. Barcelona: Ediciones Oikus Tau. (1969) COLMENARES, Germán. “Región-Nación. Problemas de poblamiento en la época colonial”. Medellín: Revista de la Universidad Nacional. (1991) —————. “La ley y el orden, fundamento profano y fundamento divino”, Bogotá: B.B.B.R., No22, 1990. —————. “Historia económica y social de Colombia”. Bogotá: La carreta. (1978) CORTINA Adela y otros. “Ética de la Empresa”. Madrid: Editorial Trotta. (1994) CROZIER, Michel. “Estado modesto, Estado moderno. Estrategia para el cambio”. México: Fondo de Cultura Mixto, Segunda edición. (1992) —————. “Le phénomène bureaucratique”. Paris: ed. de seuil. (1964) CUNILL GRAW, Nuria. “Repensando lo público a través de la sociedad - Nuevas formas de gestión pública y representación social”. Caracas: Centro Latinoamericano de Administración para el Desarrollo, y Nueva sociedad. (1997) CURRIE, Lauchlin. “Los objetivos del desarrollo la política urbana en un marco macroeconómico”. Bogotá: Cuadernos de economía, Universidad Nacional de Colombia. (1993) CHENERY, Hollis y MONTEK Ahluwalia S. “Redistribución con crecimiento”. Madrid: Editorial Tecnos. (1976) CHILD, Jorge, RESTREPO Darío, VARGAS Alejo et al. “Rompiendo la corriente: un debate al neoliberalismo”. Bogotá: CEIS. (1992) DAHL, Robert. “Un prefacio a la teoría democrática”. Bogotá: cerec. (1988) 359 —————. “La democracia y sus críticos”. Barcelona: Paidós. (1992) DAVILA, Carlos. “Teorías organizacionales y Administración. Enfoque crítico”. Bogota: Editorial interamericana. (1985) DE EZPELETA, José; INCLAN, Joachín y otros. “Comunicación de la real audiencia de Santafé para conservar la tranquilidad del Virreynato”. 30 de Marzo de 1776. En: “documentos para la historia de la vida pública del libertador” (José Félix Blanco - Compilador). Caracas: imprenta de la opinión nacional 1875, Tomo I, Pág. 275. DE LA GARZA T, Enrique. “La flexibilidad del trabajo en América Latina”. En la compilación: “Tratado Latinoamericano de Sociología del Trabajo”. México: Fondo de Cultura Económica. (2000) DE LEON, Peter y GREEN T., Mark. “Cowboys ant the New Public Management- Political Corruption as a Harbinger”. Denver: U de Colorado. (2000) DE POMBO, Lino. “Memorias del secretario del interior al Congreso de 1836”. En: “Administraciones de Santander”, Vol. V, Pág. 77. DELLA VOLPE, G. “Rousseau y Marx”. Barcelona: Ediciones Martínez Roca. (1975) DEWEY, John. “El público y sus problemas”. Buenos Aires: ágora. (1958) D’HONDT, Jacques. “Hegel filósofo de la historia viviente”. Buenos Aires: Amorrortu, s.f. (ed. original francesa, 1966) DILTHEY, Wilthey. “Hegel y el idealismo”. México: F.C.E. (1978) DIMOCK Marshall Eduard y DIMOCK, Gladis. “La administración pública”. Madrid: Editorial Hispanoamericana. (1967) DOMBOIS ,Rainer y PRIES, Ludger. “Relaciones laborales entre Estado y Mercado- Sendas de transformación en América Latina”. Caracas: Nueva Sociedad. (2000) DONAHUE, John D. “La decisión de privatizar - Fines públicos y medios privados”. Barcelona: Ediciones Paidós Ibérica. (1991) DOUGLAS, Mary y BARON, Isherwood. “El mundo de los bienes - Hacia una antropología del consumo”. México: Grijalbo. (1990) DRUCKER, Peter F. “La sociedad postcapitalista”. Bogotá: Grupo Editorial Norma. (1994) —————. “Gerencia para el futuro: El decenio de los 90 y más allá”. Bogotá: Editorial Norma. (1993) DU GAY, Peter. “In praise of Bureaucracy- Weber-Organizations- Ethics”. London: Sage publications. (2000) DURKHEIM, Emile. “La división del trabajo”. DUVERGER, Maurice. “Les parties polítiques”. Paris: Libraire Armand Colin. (1951) DWORKIN, Ronald. “Los derechos en serio”. Barcelona: Ariel. (1995) ECHAVARRIA, OLÓZAGA, Hernán. “Macroeconomía y partido liberal”. Bogotá: Fondo Editorial Legis. (1994) Edición inglesa: “The Human Condition”. Chicago: U of Chicago Press. (1958) 360 ELIAS, Norbert. “La sociedad cortesana”. Madrid: F.C.E. (1993) ELSTER, Jon. “Justicia local: De que modo las instituciones distribuyen bienes escasos y cargas necesarias”. Barcelona: Editorial Gedisa. (1994) EME, Bernard. “Les associations ou les tourmentes de l ambivalence”. En: Association, démocratie et société civile. Paris : La découverte- Mauss. (2001) EVANS, Peter. “Hybridity as an administrative strategy: combining bureaucratic capacity with market signals and deliberative democracy”. Lisboa: VII Congreso del CLAD sobre la reforma del Estado y la Administración Pública. (2002) FETSCHER, Iring. “Marxismo y burocracia”. Bogotá: Revista ECO, tomo X/5. (1965) FEUERBACH, L. “La esencia del Cristianismo”. La Habana: Editorial de Ciencias Sociales. (1976) FINDLAY, J. N. “Hegel - A re-examination”. London: George Alien Unwin, Ltd. (1958) FISK, Milton. “El papel de los intelectuales en el cambio social”. Cali: Praxis filosófica, No 2, pag. 87-99, Marzo. (1992) —————. “The state and the justice”. Cambridge: U. P. (1989) FOUCAULT, Michel. “Sécurité, territoire, Population”. Paris: Gallimard-Seuil. (2004a) —————. “Naissance de la Biopolitique”. Paris : Gallimard Seuil (2004b) —————. “La verdad y las formas jurídicas”. Barcelona: Paidós. (1991) —————. “Tecnologías del yo - Y otros textos afines”. Barcelona: Paidós. (1990) —————. “Vigilar y castigar”. México: F.C.E. (1978) —————. “Historie de la folie a la age clasique”. Paris. (1962) FREUND, Julien. “L’essence de la politique”. Paris: Editions Dalloz. (2004). Ed original (1965). —————. “Etudes sur Max Weber”. Paris: Librairie Droz. (1991) FRIEDE, Juan. “El indio en la lucha por la tierra”. Bogotá: s.e. (1974) FRIEDMAN, Milton y Rose. “Libertad de elegir”. Barcelona: Ediciones Orbis. (1983) —————. “Essays in Positive Economics”. Chicago: The University of Chicago Press. (1966) FUKUYAMA, Francis. “El fin de la historia y el último hombre”. Bogotá: Planeta. (1992) —————. “Respuesta a mis críticos”. Washington: Revista Facetas, No 4. (1990) GALBRAITH, John Kenneth. “L’économie en perspective- Une histoire critique”. Paris: Ed Seuil. (1989) GARRIDO, Margarita. “Reclamos y representaciones. Variaciones sobre 361 la política en el Nuevo Reino de Granada. 1770-1815”. Bogotá: Banco de la República. Vol. III, Pág. 34-35. (1993) GENEREUX, Jacques. “Les vraies lois de l’Economie”. Paris: Editions du Senil. (2001) GIDDENS, Anthony. “El capitalismo y la moderna teoría social”. Barcelona: Labor. (1985) GOFFMAN, Edwin. “Relaciones en público-Microestudios del orden público”. Madrid: Alianza Editorial. (1969) GONZALEZ, Gustavo y OGLIASTRI Enrique (compiladores). “La gerencia publica: ¿asunto privado? gerencia privada, asunto público”. Bogota: Tercer mundo editores. (1995) GRAMSCI, A. “Obras de A. Gramsci”. México: Juan Pablos Editor. (1975) Gran problema social de fin de siglo. Buenos Aires: Editorial Sudamericana. (1994) GUERRERO, Omar. “Del Estado gerencial al Estado cívico”. México: Miguel Angel Porrua, Universidad Autónoma del Estado de México. (1999) —————. “La teoría de la administración pública”. México: Harla. (1986a) —————. “La administración pública del Estado capitalista”. México: Halar. (1986b) GUINLE, F. J. “Reflexions sur l’tat hegelienne”. En: Etudes Philosophiques. Bruselas. (1979) HABERMAS, Jurgen. “Historia y critica de la opinión pública”. Barcelona: G. GILI. (1982) (Ed original, 1962) —————. “The Structural transformation of the Public sphere- An inquiry into a category of bourgois society”. Cambridge: MIT Press. (1995) HAFSI, Taieb y JORGENSEN, Jan J., “ L’état en mutation: désiventir pour mieux agir”. Montreal: Revue internationale de Gestión, Vol 15, No 3, Sept. (1990) HALL, Richard. “Organizaciones estructura y proceso”. Londres: Prentice-Hall International. (1983) HANSEN, Alvin H. “Guía de Keynes”. New York: Editorial MacGraw Hill. (1974) HARING, C. H. “El imperio hispánico en América”. Buenos Aires: Peuser. (1958) HAYEK, Friedrich V. “La fatal arrogancia, los errores del socialismo”. Madrid: Editorial Elnion. (1990) —————. “Camino de servidumbre”. Madrid: Revista de derecho privado. (1950) HECKSCHER, C. y DONNELLON, A. “The Post-bureaucratic Organization: New perspectives on Organizational change”. Thousand Oaks, CA: SAGE. (1994) HEGEL, G. W. F. “Lecciones de Filosofía de la historia”. Barcelona: SARPE. (1986) 362 —————. “Fenomenología del espíritu”. México: F.C.E. (1977a) —————. “Le droit, le moralité et la politique” (esseais choises). Paris: PUF. (1977b) —————. “La filosofía del derecho”. Buenos Aires: Claridad. (1966) HELD, David, MC GREW, Anthony, GOLDBLATT, David, PERRATON, Jonathan. “Global Transformations - Politics, Economics and Culture”. Stanford University Press. (1999) —————. “La democracia y el orden global - Del Estado moderno al gobierno cosmopolita”. Barcelona: Paidós. (1995) HILFGERDING, Rudolf. “El capital financiero”. Madrid: Tecnos. (1985) HIRSCHMAN, Albert. “The strategy of economic development”. London: Yale University Press. (1958) HOBBES, Thomas. “The Leviathan”. London: Penguins B. (1962) HOBSBAWM, Eric. “Las revoluciones burguesas”. Barcelona: Guadarrama. (1980) HUILZINGA, Johan. “El otoño de la Edad Media”. Madrid: Alianza. (1979) HYPPOLITE, Jean. “Génesis y estructura de la Fenomenología del Espíritu de Hegel”. Barcelona: Herder. (1996a) —————. “Lógica”. Barcelona: Herder. (1996b) JARAMILLO URIBE, Jaime. “Ensayos de historia social”. Bogotá: Tercer mundo. (1989) —————. “Nación y región en los orígenes del Estado Nacional en Colombia”. En: “Problemas de la formación del Estado y la Nación en Hispanoamérica”. Bonn: Internationes. (1984) —————. “El pensamiento colombiano en el siglo XIX”. Bogotá: Temis. (1982) JARAMILLO VELEZ, Rubén. “Ideas para una Historia Universal en clave cosmopolita y otros escritos sobre Filosofía de la Historia”. Madrid: Editorial Tecnos. (1987) —————. Diciembre de l784. Traducción. Bogota: Argumentos, Nos. 14-17. (1986) JELLINEK, G. “Teoría general del Estado”. Buenos Aires: Albatros. (1970) JESSOP, Bob. “Desarrollos recientes en la teoría del Estado- Enfoques, problemas, agendas”. Bogota: Siglo del Hombre- Universidad Nacional. (1999) JORDA, Henri. “Travail et discipline. De la manufacture á l’entreprise intelligente”. Paris: L’HARMATTAN. (1999) KALMANOVITZ, Salomón. “Economía y Nación - Una breve historia de Colombia”. Bogotá: Siglo XXI, U. Nacional, Cinep. (1988) KALLINIKOS, Jannis. “The Social Foundations of the Bureaucratic Order”. London: ORGANIZATION, Sage Publications, Volume 11 (1), pag. 13-36. (2004) KAMERMAN, B. Sheila y KAHN, Alfred (compiladores). “La privati363 zación y el Estado benefactor”. México: F.C.E. (1993) Kant Inmanuel. “Respuesta a la pregunta ¿Qué es la Ilustración?”. Berlinische Monatsschrift, KAPLAN, Marcos. “La formación del Estado nacional en América latina”. Buenos Aires: Amorrortu. (1969) KERNAGHAN, Kenneth. “The post-bureaucratic organization and public service values”. London: INTERNATIONAL REVIEW OF MANAGEMENT SCIENCES, Sage Publications, Vol. 66, Pág. 91-104. (2000) KEYNES, John M. “Ensayos sobre intervención y liberalismo”. Barcelona: Orbis. (1988) —————. “Teoría general de ocupación el interés y el dinero”. México: F.C.E. (1943) LAÏDI, Zaky. “La Grande Perturbation”. Paris: Flammarion. (2004) LANGE, Oscar y TAILOR Fred M. “Sobre la teoría del socialismo”. Barcelona: Ediciones Ariel. (1970) LARSON, Brooke. “Explotación y economía moral en los Andes del sur- hacia una reconsideración crítica”. Bogotá: Historia crítica No. 6. (1992) LAVAL, Christian. “L’ambition sociologique”. Paris: Éditions la Découverte/ M.A.U.S.S. (2002) LEAL, Francisco. “Estado y política en Colombia”. Colombia: Editorial Siglo XXI. (1989) LEBRET, J. C. “Misión economía y humanismo sobre las condiciones del desarrollo de Colombia”. Bogotá: Editores Acolita. (1958) LEVINE, Barry (compilador). “El desafió neoliberal”. Bogotá: Norma. (1992) LIEVANO AGUIRRE, Indalecio. “BOLIVAR”. Bogotá: Oveja Negra. (1983) LINDBLOM, E. C. “Democracia y sistema de mercado”. México: Fondo de Cultura Económica, Colegio Nacional de Ciencias Políticas y Administración Pública, A.C. (1999) LOCK, Peter. “Transformaciones de la Guerra- Hacia la dominación de la violencia reguladora”. Cartagena: Foro social mundial. (2003) LOCKE, John. “Segundo tratado sobre el gobierno civil”. Madrid: Alianza Editorial. (1990) LOPEZ CALERA, Nicolás. “Yo el Estado - Bases para una teoría sustancializadora (no sustancialista) del Estado”. Madrid: Editorial Trotta. (1992) LOW, MURTRA, Enrique y GOMEZ RICARDO, Jorge. “Teoría fiscal”. Bogotá: ediciones Universidad Externado de Colombia. (1996) LOWITH, Kart. “De Hegel a NIEZSTCHE”. Buenos Aires: Sudamericana. (1974) LUHMANN, Niklas. “Poder”. Madrid: Anthropos. (1995) LYNCH, John. “Hispanoamérica 1750- 1850- ensayos sobre la sociedad y el Estado”. Bogotá: U. Nacional. (1987) MACPHERSON, C. B. “La teoría política del individualismo posesivo”. Barcelona: Editorial Fontanella. (1979) 364 MAESTRE ALFONSO, Juan. “El tratamiento del componente indígena en el derecho constitucional americano”. Barcelona: Boletín americanista No. 32. (1982) MAILA, Joseph. “La notion de interet general dans la penseé musulmane”. En la compilación: “Essai historique sur l interet general”. Paris: Ed Karthala. (1996) MAJONE, Giandomenico. “Evidencia, argumentación y persuasión en la formulación de las políticas”. México: Fondo de Cultura Económica y Colegio Nacional de Ciencias Políticas y Administración Pública. (1997) MANCUR, Olson. “La lógica de la acción colectiva- Bienes públicos y la teoría de grupos”. México: Editorial Limosa. (1992) MAPPA, Sophia (comp). “Essai historique sur l intérêt général”. Paris: Ed Karthala. (1996) MARCUSE, Herbert. “L’homme unidimensionnel: essai sur l’idéologie de la société industrielle avancée”. Paris: Les Éditions de Minuit. (1968) —————. “Razón y revolución Hegel y el surgimiento de la teoría social”. MARCHENA FERNÁNDEZ, Juan. “Ejército y milicias en el mundo colonial americano”. Madrid: Editorial MAPFRE. (1992) MARX, Kart. “Critica a la filosofía del derecho estatal de Hegel”. En: Escritos de Juventud. México: Fondo de Cultura Económica. (1971) —————. y ENGELS Federico. “Obras fundamentales”. México: F.C.E. —————. “El 18 brumario de Luis Bonaparte”. Pekín: Ed. en lenguas extranjeras. (1978) MASUR, Gerhard. “Simón BOLIVAR”. Bogotá: Círculo. (1984) McLUHAN, Marshall. “La galaxia Gutemberg”. Barcelona: Planeta. (1986) MEDELLIN TORRES, Pedro. “La reforma del Estado en América latina”. Bogotá: Editorial Presencia. (1991) MILIBAND, R. “Marxism and Politics”. London: Oxford Press. (1977) MILL STUART, John. “Sobre la libertad, el utilitarismo”. Barcelona: Editorial Orbis. (1971) MINTZBERG, Henry. “Le pouvoir dans les organisations”. Paris: éditions d’organisations (1986) MOE, Ronald. “Explorando los límites de la privatización”. En la compilación: Clásicos de la Administración Pública, SHAFRITZ Jay y HYDE Albert (comp). México: Fondo de Cultura Económica. (1999) MOLINA, Gerardo. “Las ideas liberales en Colombia”, Bogotá: Tercer Mundo. (1981) MONTHLY REVIEW, “Publicación socialista independiente”. New York. MONZON ARRIBAS, Cándido. “La opinión pública: teorías, concepto y métodos”. Madrid: Editorial Tecnos. (1990) MORGAN, Gareth. “Imágenes de la organización”. México: Ediciones Alfaomega. (1991) 365 MORSTEIN MARX, Fritz. “Las funciones sociales de la administración pública”. Barcelona: Sarpe. (1984) MOUSSAOUI, A. “La notion de l intérêt général en Algérie”. En la compilación: “Essai historique sur l’intérêt général”. Paris: Ed. Karthala. (1996) MOUZELIS, Nikos. “Organización y Burocracia”. Madrid: Ediciones 62. (1991) NEGRI Tony y HARDT, Michael. “Multitude- Guerre et démocratie a l’age d’Empire”. Montréal: Boreal. (2004) NIETO ARTETA ,Luis E. “Economía y cultura en la historia de Colombia”. Bogotá: Librería Siglo XX. (1942) ————. “Ensayos históricos y sociológicos”. Bogotá: Colcultura. (1978) NIOCHE, Jean Pierre. “Ciencia Administrativa, Management Público y análisis de las políticas públicas”. En la compilación: “Cuatro lecturas clave sobre políticas públicas”. Bogotá: Pontificia Universidad Javeriana. Septiembre. (1996) NOZICK, Robert. “La naturaleza de la nacionalidad”. Barcelona: Paidós. (1995) —————. “Anarchie, State and Utopia”. New York: Basic Bokks. (1974a) —————. “State, Anarchy utopia”. New York: Basic Books. (1974b) OCAMPO, José Antonio. “Reforma del Estado y desarrollo económico y social en Colombia”. Bogotá: Análisis político, U. Nacional, No 17, Pág. 5-40, Septiembre. (1992) OCAMPO LÓPEZ, Javier. “El proceso ideológico de la emancipación en Colombia”. Bogotá: Ediciones Presencia. (1991) OLSEN, P. Johan. “Maybe it is time to rediscover bureaucracy”. Madrid: IX Congreso del CLAD sobre la reforma del Estado y la Administración Pública. (2004) OLSON, Mancure. “Poder y prosperidad. La superación de las dictaduras comunistas y capitalistas”. Siglo Veintiuno de Argentina Editores. (2001) OSBORNE, D. y GAEBLER, T. “Reinventing Government”. New York: Addison-Wesley. (1992) OSLACK, Oscar (Compilador). “Teoría de la burocracia estatal”. Buenos Aires: Paidós. (1984) OSPINA BOZZI, Sonia. “Evaluación de la gestión pública: Conceptos y aplicaciones en el caso latinoamericano”. Caracas: Revista del CLAD Reforma y Democracia, No. 19, Pág. 89-122, Febrero. (2001) OTS-CAPDEQUI, José M. “Instituciones de Gobierno del Nuevo reino de Granada durante el Siglo XVIII”. Bogotá: U. Nacional. (1950a) —————. “El Estado español en las indias”. México: F.C.E. (1950b) —————. “Nuevos aspectos del siglo XVIII español en América”. Bogotá: U. Nacional. (1948) OVEJERO, Lucas Félix. “La política de la desconfianza -límites de la 366 concepción liberal de las políticas públicas”. En la revista: “AGENDA: Ética, economía y políticas públicas”, No. 2. Bogotá: Tercer Mundo Editores, octubre 1999 - mayo 2000 PADGEN, Anthony. “El final del imperio: Simón BOLIVAR y la república liberal”. En: “El liberalismo como problema”. (comp.). Caracas: Monte Avila eds. (1991) PAZ, Octavio. “El ogro filantrópico”. Barcelona: Editorial Seix Barral. (1983) PEEMANS, Jean Philippe. “Revoluciones industriales, modernización y desarrollo”. Bogotá: Historia critica, U de los Andes, No 6, Pág. 15-34, Enero- Junio. (1992) PEREZ, Juan Pablo. “Globalización, riesgo y empleabilidad”. Caracas: NUEVA SOCIEDAD, No 184, Marzo-Abril (2003) PETERS, B. Guy. “Institutional theory in political science- The New Institutionalism”. New York: Continuum. (2001) —————. “La política de la burocracia”. México: Fondo de Cultura Económica y Colegio Nacional de Ciencias Políticas y Administración Pública. (1999) PINSON, Jean Claude. “Hegel, le Droit et le libéralisme”. Paris: PUF. (1989) POPPER, K. “The Open Society and it´s enemies”. London: Oxford Press. (1974) PORTELLI, H. “Gramsci y el Bloque histórico”. México: Siglo XXI. (1979) POULANTZAS, Nikos. “Political Power and Social Class”. New York: The Free Press. (1973) POWELL, Walter y DIMAGGIO, Paul. “El nuevo institucionalismo en el análisis organizacional”. PRICE, Vicent. “La opinión pública - Esfera pública y la comunicación”. Barcelona: Ediciones Paidós. (1991) PRZEWORSKI, Adam. “Democracia y mercado”. Cambridge: Ediciones organización Editorial de la universidad de Cambridge. (1991) RAWLS, John. “A Theory of Justice”. Oxford: University Press. (1971) RESCHER, Nicholas. “Welfare the social issues in philosophical perspective”. U. Pittsburg. (1972) RIFKIN, Jeremy. “The Age of Access”. New York: Penguins Books. (2000) —————. “El Fin del Trabajo. Nuevas tecnologías contra puestos de trabajo: el nacimiento de una nueva era”. Barcelona: Editorial Paidós. (1994) ROMERO ,José Luis. “Latinoamérica: las ciudades y las ideas”. México: Siglo XXI. (1976) ROTH A, Noel. “Introducción para el análisis de Políticas Públicas”. Cali: Cuadernos de Administración, Universidad del Valle, No 30, Págs. 113-128, Diciembre. (2003) 367 ————. “Políticas públicas. Formulación, interpretación y evaluación”. Bogotá: Ediciones Aurora. (2002) ROUSSEAU, J. J. “Discurso sobre la economía política”. Madrid: Editorial Tecnos. (1985) —————. “Du contrat social, Editions Sociales”. Paris. (1978) —————. “Discours sur l´origine et les fondements de l´inégalité parmi les hommes”. Paris: Editions sociales. (1971) —————. “Ouvres”. Paris: Garnier. (1958) RUBIN, Isaias. “Ensayos sobre la teoría marxista del Valor”. Buenos Aires: Cuadernos de Pasado y Presente. (1977) RUBIO CARRACEDO, José. “Paradigmas de la política- Del Estado justo al Estado mínimo”. Barcelona: Antrhopos. (1990) RYAN, Alan; TUGENDHAT, Ernesto y otros. “El liberalismo como problema”. Caracas: Editores Monte Ávila. (1991) SALAZAR VARGAS, Carlos. “Políticas Públicas”. Bogotá: Pontificia Universidad Javeriana. (1999) SANTANA, L. Rabell y NEGRON PORTILLO, M. “Reinventing government: Nueva retórica, viejos problemas”. Caracas: Revista del CLAD No. 6. (1996) SARMIENTO ANZOLA, Libardo. “El paquete social del neoliberalismo”. Bogotá: FORO, No 20, Pág. 4-19, Mayo (1993) SAVATER, Fernando. “Para la anarquía y otros enfrentamientos”. Barcelona: Ediciones Orbis. (1984) SCHMITT, Carl. “Teoría de la Constitución”. Madrid: Editorial Alianza. Universidad textos. (1988) —————. “La dictadura”. Madrid: Alianza. (1985) SENNETT, Richard. “La corrosión del carácter”. Barcelona: Anagrama. (2000) —————. “El declive del hombre público”. Barcelona: Península. (1979) SHAW, G. K. “Introducción a la teoría de la política macroeconómica”. Madrid: Ediciones icf. (1974) SIMEON, J. P. “La démocratie selon Rousseau”. Paris: Editions du Seuil. (1976) SIMON, A. Herbert. “Administrative Behavoir- A study of decision making processes in administrative organizations”. New York: The Free Press. (1997) SPINOZA, B. “The chief Works of Spinoza”. New York: Dover Publications. (1959) STIGLITZ, J. “Quand le capitalisme perdre sa tête”. Paris: Fayard. (2003) —————. “El malestar en la globalización”. Bogotá: Editorial Taurus, S.A. (2002a) —————. “Mejorando la eficiencia y la capacidad de respuesta del sector público: lecciones de la experiencia reciente”. Caracas: En Revista del CLAD: Reforma y Democracia, No. 22, Pág. 7-68. (2002b) STUART, Hampshire. “Spinoza”. London: Penguins Books. (1962) SUNKEL, Osvaldo. “Auge, crisis y renovación del Estado”. Bogotá: 368 FORO, No 20. Pág. 77-94, Mayo. (1993) TALMON, Y. “Los orígenes de la democracia totalitaria”. TENZER, Nicolas. “La sociedad despolitizada”. Barcelona: Paidós. (1991) TEXIER, J. “Gramsci”. Barcelona: Grijalbo. (1973) THOMPSON, E. P. “Tradición, revuelta y consciencia de clase”. Barcelona: Crítica. (1979) TULLOCK, Gordon. “Necesidades privadas y medios públicos. Análisis económico del nivel deseable de intervención pública”. Madrid: Ediciones Aguilar. (1974) URRUTIA MONTOYA, Miguel. “Posibilidades de privatización en Colombia”. Bogotá: Economía Colombiana, No 236, Pág. 18-32, Julio-Agosto. (1991) VALLESPIN, Fernando. “Nuevas teorías del contrato social: Rawls, Nozick y Buchanan”. Madrid: Alianza. (1985) VARELA B, Edgar. “La soberanía transformada”. En proceso editorial. (2005a) —————. “La Mercantilización de lo Público y otros ensayos sobre Gestión y Políticas Públicas”. Cali: Facultad de Administración, Universidad del Valle. (2005b) -—————. “Mercantilización de lo Público”. Barcelona: INSTITUCIONES Y DESARROLLO, Nos: 13 y 14, Diciembre (2003) —————. “La Transformación de la burocracia pública: Una perspectiva crítica del proceso de reforma del Estado en América Latina 19802001”. Barranquilla: Universidad del Norte. Pensamiento & Gestión No. 12, Junio. (2002) —————. “Empresarios del Terror”. (1997) —————. “La crisis de la civilidad en Colombia”. Cali: Centro editorial U del Valle. (1995) —————. “La cultura de la violencia en Colombia el siglo XIX”. Cali: Imprenta departamental. (1990) VARGA, E. “Essais sur l´economie politique du capitalisme”. Moscú: Editorial Progreso. (1962) VICO, G. “Una nueva ciencia sobre la naturaleza común de las naciones”. México: Aguilar. (1964) VIVIESCAS, Fernando y GIRALDO Fabio (compiladores). “Colombia: el despertar de la modernidad”. Bogotá: Carvajal. (1991) WEBER, Max. “Economía y sociedad” (dos Vols.). Bogotá: F.C.E. (1985) —————. “La ética protestante y el espíritu del capitalismo”. Barcelona: Ediciones Orbis S.A. (1985) WEFFORT, Francisco. “Los dilemas de la legitimidad política”. Bogotá: FORO, No 10, Sept. (1989) 369 WEIL, Eric. “Hegel et L’état”. Paris: J. VRIN. (1985) —————. “Problemes Kantiens”. Paris: J. VRIN. (1982) WEINTRAUB, Jeff. “The Theory and Politics of the Public/Private Distinction”. En el libro: “Public and Private in Thought and Practice Perspectives on a Grand Dichotomy”. Chicago: The University of Chicago Press. (1997) WIEACKER, F. “Historia del derecho privado de la Edad Moderna”. Madrid: Aguilar. (1957) WILGUS, A.C. “Colonial Hispanic America” (compilador). New York: Russell and Russell. (1963) WITTFOGEL, Karl. “El despotismo oriental - Estudio comparativo del poder totalitario”. Madrid: Guadarrama. (1966) WOLFE, Alan. “Public and Private in Theory and Predicament of Private Life at the End of the Twentieth Century”. En el libro: “Public and Private in Thought and Practice - Perspectives on a Grand Dichotomy”. Chicago: The University of Chicago Press. (1997) WORLD BANK GROUP. “Governance and Public sector Reform sites”. www.worldbank.org public sector 2-52 ( Web consultada en Noviembre del 2004) ZULETA, Estanislao. “Para una concepción positiva de la democracia”. En la compilación: “Derechos humanos y modernidad”, MARTINEZ Carlos A. y VARELA Edgar. Cali. (1989) 370 Ciudad Universitaria, Meléndez Cali, Colombia Teléfonos: (+57) 2 321 2227 321 2100 ext. 7687 http://programaeditorial.univalle.edu.co programa.editorial@correounivalle.edu.co
https://openalex.org/W4383052003	https://zenodo.org/records/8112418/files/100525.pdf	English	null	Hardsness Testing Of MIG Welding Process For Similar And Dissimilar Materials	Zenodo (CERN European Organization for Nuclear Research)	2,023	cc-by	5,210	Abstract: MIG Welding is massively used in industrial process. The major advantage of the MIG welding process is that not only its versatility but also it causes almost no loss of metals. MIG can be operated as a semiautomatic or fully automatic welding. During literature review we found that there is very less study and research work done on welding of two different metals joining together. Then we have been decided and selected material of Mild steel and Aluminium for the study and investigation of Mechanical properties. Initially, with the help of trial-and-error method, we successfully welded aluminium and mild still together using MIG Welding Process. MIG Welding setting Parameters we keep constant like voltage and current for different specimen for getting same result. We have also completed hardness testing for specimen. As a result, this study has been successfully completed with the help of MIG welding experimentation and welding joint hardness test for specimens. welding, and plasma arc welding are all various types of welding techniques used in the metalworking industry. A variety of metals, such as stainless steel, aluminium, nickel and copper alloys, cobalt, and titanium, can be welded using these techniques, which usually require the use of a filler material. Different sectors, including oil and gas, automotive, aerospace, power, and others, rely heavily on arc welding methods. Gas metal arc International Journal of Advance and Applied Research www.ijaar.co.in ISSN – 2347-7075 Impact Factor – 7.328 Peer Reviewed Bi-Monthly Vol.10 No.5 May – June 2023 HARDSNESS TESTING OF MIG WELDING PROCESS FOR SIMILAR AND DISSIMILAR MATERIALS Prof. Vaibhav V. Kulkarni1, Vaishnavi Ghatge2, Rutuja Kadam3 & Vaishnavi Karhale4 1Assistant Professor, School of Mechanical Engineering, Dr. Vishwanath Karad MIT World Peace University, Pune. 2,3&4School of Mechanical, Engg., Dr. Vishwanath Karad MIT World Peace University, Pune. Corresponding Author - Prof. Vaibhav V. Kulkarni Email - Shridharshana@gmail.com DOI - 10.5281/zenodo.8112418 International Journal of Advance and Applied Research www.ijaar.co.in ISSN – 2347-7075 Impact Factor – 7.328 Peer Reviewed Bi-Monthly Vol.10 No.5 May – June 2023 HARDSNESS TESTING OF MIG WELDING PROCESS FOR SIMILAR AND DISSIMILAR MATERIALS Prof. Vaibhav V. Kulkarni1, Vaishnavi Ghatge2, Rutuja Kadam3 & Vaishnavi Karhale4 1Assistant Professor, School of Mechanical Engineering, Dr. Vishwanath Karad MIT World Peace University, Pune. 2,3&4School of Mechanical, Engg., Dr. Vishwanath Karad MIT World Peace University, Pune. Corresponding Author - Prof. Vaibhav V. Kulkarni Email - Shridharshana@gmail.com DOI - 10.5281/zenodo.8112418 Lap joint: A lap joint is a type of joint in which two overlapping pieces are joined together at an angle ranging from 0° to 5° in the area surrounding the weld. Fig.1. MIG Welding Fig.1. MIG Welding ISSN – 2347-7075 IJAAR welding (GMAW) is an alternative name for Metal Inert Gas (MIG) welding, which employs a consumable wire electrode to act as the filler material, a semi-automatic or completely automatic arc to form the weld, and a shielding gas to protect the weld, encourage weld penetration, and lessen weld bead porosity. how quickly the electrode is supplied into the weld. Introduction: Many common automatic, semi- automated, and manual processes fall under this group. Stick welding, gas welding, metal inert gas welding (MIG), tungsten inert gas welding (TIG) (also referred to as gas tungsten arc welding or GTAW), metal active gas (MAG) welding, flux cored arc welding (FCAW), gas metal arc welding (GMAW), submerged arc welding (SAW), shielded metal arc 149 Vol.10 No.5 ISSN – 2347-7075 ISSN – 2347-7075 IJAAR portion of the base metal surrounding the weld that has undergone metallurgical changes due to the heat of welding or thermal cutting, but has not melted. Fusion Line: The boundary between the heat affected zone and the weld metal in a fusion weld. This term is sometimes used to refer to the weld joint itself. Weld Zone: The area including the heat affected zone and the weld metal. Weld Face: The visible surface of the weld bead on the side where welding was performed. Weld Root: The point on the opposite side of the weld from where the first weld run began. electric shock, visual damage, radiation exposure, or inhalation of toxic welding fumes and gases. portion of the base metal surrounding the weld that has undergone metallurgical changes due to the heat of welding or thermal cutting, but has not melted. Fusion Line: The boundary between the heat affected zone and the weld metal in a fusion weld. This term is sometimes used to refer to the weld joint itself. Weld Zone: The area including the heat affected zone and the weld metal. Weld Face: The visible surface of the weld bead on the side where welding was performed. Weld Root: The point on the opposite side of the weld from where the first weld run began. Weld Toe: The point where the weld face or run intersects with the base metal. This is a critical area for stress concentration and crack initiation, so it is essential to ensure a smooth transition from the weld to the base metal. Excess Weld Metal: The weld metal that extends beyond the surface of the base metal, often used to fill gaps or reinforce joints. This is sometimes referred to as overfill or reinforcement. Run (pass): (Weld run pass) Metal that has been melted or deposited after passing via an electrode, torch, or blowpipe. Layer: In welding, a layer is a stratum of weld metal consisting of one or more weld runs. Energy Sources: The energy source used affects various processes, and there are numerous diverse approaches available. These procedures must be used with ti b th lt i b Problem Statement and Objective: To investigate and analyze MIG Welding for Dissimilar Materials. Selected materials are Aluminum and Mild Steel and type of joint is Lap joint and Butt joint. Various types of tests can be performed on selected specimens in welding. Current and voltage to weld such material is determined. Analysis of different parameters of aluminium alloy and mild steel weld by MIG welding technique. To determine mechanical properties and behaviour of different material of welding specimen. Weld Toe: The point where the weld face or run intersects with the base metal. This is a critical area for stress concentration and crack initiation, so it is essential to ensure a smooth transition from the weld to the base metal. Excess Weld Metal: The weld metal that extends beyond the surface of the base metal, often used to fill gaps or reinforce joints. This is sometimes referred to as overfill or reinforcement. Run (pass): (Weld run pass) Metal that has been melted or deposited after passing via an electrode, torch, or blowpipe. Layer: In welding, a layer is a stratum of weld metal consisting of one or more weld runs. Energy Sources: The energy source used affects various processes, and there are numerous diverse approaches available. These procedures must be used with caution because they can result in burns, Features of Completed Weld: Features of Completed Weld: Fig.2.Lap joint Fig. 3. Butt Weld Fig.2.Lap joint Fig.1. MIG Welding Fig.2.Lap joint The welding gun is fed with both the electrode and the shielding gas (or torch). Other mixtures may also be employed, depending on the materials being welded and other factors, but the standard shielding gas composition is 75% argon and 25% carbon dioxide. Additionally, a welding gun is fed a constant, consumable wire electrode and a shielding gas through a lead (sometimes called a torch). The types of metal being bonded, the thickness of the part, and the design of the connection all affect the composition and diameter of the wire electrode. The wire feed speed (WFS) settings, which required to supply adequate weld metal for the join, control Fig. 3. Butt Weld Fig. 3. Butt Weld Fig. 3. Butt Weld Parent Metal: The metal pieces that will be bonded, brazed, braze-welded, or surfaced together are typically referred to as the parent metals. Filler Metal: The metal that is added during brazing, braze- welding, welding, or surfacing to bond or fill the base metals is called the filler metal. Weld Metal: Any metal that is melted during welding and remains in the welded joint. Heat Affected Zone: The Prof. Vaibhav V. Kulkarni, Vaishnavi Ghatge, Rutuja Kadam & Vaishnavi Karhale Prof. Vaibhav V. Kulkarni, Vaishnavi Ghatge, Rutuja Kadam & Vaishnavi Karhale 150 Vol.10 No.5 ISSN – 2347-7075 IJAAR welding. Welding defects like as porosity, fractures, and other issues are not visible in the welding. The bend test also produced the greatest results for this work. The ultimate result was satisfactory and gave clear guidance on welding under bend force. Since relatively little Cr is lost from stainless steel during TIG welding, the material is resistant to field corrosion. Additionally, welding is carried out in an inert atmosphere to shield the joint from environmental hazards like hydrogen. According to reports, TIG welding produces excellently strong joints between different types of metal. Hardness rating at the filler metal's point. The Effect of Gas Metal Arc Welding (GMAW) Processes on different Welding Parameters. Izzatul Aini Ibrahim, Syarul Asraf Mohmat, Amalina Amir, Abdul Ghalib, 2012[9] conducted an experiment to investigate the effects of arc voltage, welding current, and welding speed on the depth of penetration and hardness of the weld bead. The results showed that increasing the welding current from 90 A to 210 A resulted in an increase in the depth of penetration. However, the penetration was also influenced by the welding speed and arc voltage. The highest value of penetration was achieved at 210 A and 22 V for all three welding speeds. At a welding speed of 60 cm/min, the best rates of 14 l/min and 18 l/min, respectively, Samples 2 and 5 were determined to be the best samples among the others, demonstrating how mechanical qualities depend on certain welding parameter values. Study the effects of Welding Parameters on TIG welding of Aluminum plate. It can be possible to weld aluminium plates uniformly using an automated welding system, according to Umesh Sharma and Viswas Yadav in 2020[7]. The tensile strength of a weld joint can be significantly affected by various welding parameters such as welding speed and welding current. The tensile strength of the weld joint increases with an increase in current. Due to changes in microstructure, the weld zone’s hardness value varies with distance from the weld center. Strength is greater at lower welding rates because the current is more intense. Tensile strength on both sides of the welding is found to be about equal to the base material's strength. The tensile strength of the weld joint is unaffected by welding speed when both sides are welded at a high current (180 A). Experimental Investigation of TIG Welding on Stainless steel and Mild Steel Plates. Literature Review: dam & Vaishnavi Karhale To determine effects of gas metal arc welding (GMAW) parameters on mechanical properties of aluminum alloys. Lochan Sharma and Chandan Kaushal, 2015 On 6 mm thick aluminum alloy plates, explained free welds may be produced utilizing a gas metal arc welding method with a current of 150 to 170 A and gas flow rate being 14- 18 L/Min. Due to incorrect welding values, some welding samples, such as sample 4 and sample 6, failed the Tensile tests in the welding zone. The tiny grain structure of silicide led to an increase in Hardness. With welding parameters of 160 A and gas flow Prof. Vaibhav V. Kulkarni, Vaishnavi Ghatge, Rutuja Kadam & Vaishnavi Karhale to the base metal. Excess Weld Metal: The weld metal that extends beyond the surface of the base metal, often used to fill gaps or reinforce joints. This is sometimes referred to as overfill or reinforcement. Run (pass): (Weld run pass) Metal that has been melted or deposited after passing via an electrode, torch, or blowpipe. Layer: In welding, a layer is a stratum of weld metal consisting of one or more weld runs. Energy Sources: The energy source used affects various processes, and there are numerous diverse approaches available. These procedures must be used with caution because they can result in burns, To determine effects of gas metal arc welding (GMAW) parameters on mechanical properties of aluminum alloys. Lochan Sharma and Chandan Kaushal, 2015 On 6 mm thick aluminum alloy plates, explained free welds may be produced utilizing a gas metal arc welding method with a current of 150 to 170 A and gas flow rate being 14- 18 L/Min. Due to incorrect welding values, some welding samples, such as sample 4 and sample 6, failed the Tensile tests in the welding zone. The tiny grain structure of silicide led to an increase in Hardness. With welding parameters of 160 A and gas flow 151 Vol.10 No.5 Process Selection: dam & Vaishnavi Karhale The development of the high- productivity, high-quality arc welding technique is being led by the GMAW method. A Metal Inert Gas (MIG), is also called as Gas Metal Arc Welding (GMAW), is a technique which heated, melted and solidified the filler materials and base metals in a constrained fusion zone. Due to this a joint is created between parent metals. The internal resistive power and heat from the welding arc melt the continuous wire electrode that is fed through the contact tip of the welding torch from an automatic wire feeder. Heat generated by the molten metal that transferred to the weld pools and from the melting electrode's end to the weld pools. The quality, productivity, and cost of the welding joint are influenced by the GMAW welding parameters. If all welding variables are in compliance, the ideal arc will be produced. Arc Voltage, arc welding current, welding speed, free wire length, nozzle distance, welding location and direction, torch length and gas flow IJAAR ISSN – 2347-7075 bead was found to be the highest at 90 A and gradually decreased with increasing welding current up to 150 A. At 210 A, the hardness slightly increased compared to 150 A. The highest hardness value was achieved at 90 A and 26 V at a welding speed of 60 cm/min. The microstructure's grain boundaries changed from a larger size to a smaller size as the welding parameters were changed. Fig.5.MS Specimen Sample Fig.6. Al Specimen Sample Fig.6. Al Specimen Sample Vol.10 No.5 Vol.10 No.5 ISSN – 2347-7075 Keyur Panchal [8] (2016) Investigations were conducted on the tensile strength dilution of welded joints rates of 14 l/min and 18 l/min, respectively, Samples 2 and 5 were determined to be the best samples among the others, demonstrating how mechanical qualities depend on certain welding parameter values. Study the effects of Welding Parameters on TIG welding of Aluminum plate. It can be possible to weld aluminium plates uniformly using an automated welding system, according to Umesh Sharma and Viswas Yadav in 2020[7]. The tensile strength of a weld joint can be significantly affected by various welding parameters such as welding speed and welding current. The tensile strength of the weld joint increases with an increase in current. Due to changes in microstructure, the weld zone’s hardness value varies with distance from the weld center. Strength is greater at lower welding rates because the current is more intense. Tensile strength on both sides of the welding is found to be about equal to the base material's strength. The tensile strength of the weld joint is unaffected by welding speed when both sides are welded at a high current (180 A). Experimental Investigation of TIG Welding on Stainless steel and Mild Steel Plates. Keyur Panchal [8] (2016) Investigations were conducted on the tensile strength, dilution of welded joints, Experimental Investigation of TIG Welding on Stainless steel and Mild Steel Plates. Keyur Panchal [8] (2016) Investigations were conducted on the tensile strength, dilution of welded joints, hardness, and bend test. For the best S.S. and MS plate joint, use tungsten inert gas Vaibhav V. Kulkarni, Vaishnavi Ghatge, Rutuja Kadam & Vaishnavi Karhale 152 ISSN – 2347-7075 Fig.5.MS Specimen Sample Fig.6. Al Specimen Sample Vol.10 No.5 The findings of an investigation into the weld deposit area were provided, and they showed how the weld deposit area was affected by the electrode polarity, extension, and diameter, arc voltage, welding current, power source settings, flux and travel speed .These are some of the key factors that need to be considered for optimal results. The MIG welding process offers several benefits such as high deposition rates, fast welding speeds, deep penetration, consistent wire diameter, and the ability to fully automate the process. It is commonly used for unalloyed and low- alloy steels (MAG), but is also increasingly used with CrNi steels (MAG) and aluminum materials (MIG). Some common applications of MIG welding include steel construction, shipbuilding, rail vehicle construction, and container construction. Vol.10 No.5 Vol.10 No.5 IJAAR IJAAR Vol.1 rate are among the various parameters that affect the quality, productivity, and cost of a welding joint in the GMAW process. In 1987, researchers utilized regression analysis to establish a correlation between the process factors and the bead geometry in welding. By considering factors such as the type of base metal, the shape of welded components, and the welding process, it is possible to create optimal welding conditions. A previous study that employed the MIG or GMAW welding process found that by increasing the welding current, the depth of penetration increased, while decreasing the voltage resulted in a reduction in penetration depth. Furthermore, by reducing the arc travel rate, the penetration depth increased again, with a minimum value being dependent on the arc power. The analysis of the GMAW process and its relationship with bead geometry indicates that the arc current has the most significant impact among all process factors. The findings of an investigation into the weld deposit area were provided, and they showed how the weld deposit area was affected by the electrode polarity, extension, and diameter, arc voltage, welding current, power source settings, flux and travel speed .These are some of the key factors that need to be considered for optimal rate are among the various parameters that affect the quality, productivity, and cost of a welding joint in the GMAW process. In 1987, researchers utilized regression analysis to establish a correlation between the process factors and the bead geometry in welding. By considering factors such as the type of base metal, the shape of welded components, and the welding process, it is possible to create optimal welding conditions. A previous study that employed the MIG or GMAW welding process found that by increasing the welding current, the depth of penetration increased, while decreasing the voltage resulted in a reduction in penetration depth. Furthermore, by reducing the arc travel rate, the penetration depth increased again, with a minimum value being dependent on the arc power. The analysis of the GMAW process and its relationship with bead geometry indicates that the arc current has the most significant impact among all process factors. Methodology: Aluminum is a particularly intriguing metal in the world of engineering because of its special qualities of light weight and corrosion resistance. Because of their light weight and great strength to weight ratio, aluminium alloys are widely used in the aerospace, automotive, rail, bridge, and offshore, and high-speed ship industries. Mild steels have great weldability and don't require any additional alloying elements. Because it is affordable, weldable, and readily available, mild steel is the optimum material for welding. Mild steel is available in any shape or thickness throughout the world and is of exceptional quality. Mild steel and Aluminum were selected for the experimental study because of their many benefits and straight forward accessibility to the market. Prof. Vaibhav V. Kulkarni, Vaishnavi Ghatge, Rutuja Kadam & Vaishnavi Karhale have great weldability and don't require any additional alloying elements. Because it is affordable, weldable, and readily available, mild steel is the optimum material for welding. Mild steel is available in any shape or thickness throughout the world and is of exceptional quality. Mild steel and Aluminum were selected for the experimental study because of their many benefits and straight forward accessibility to the market. through the contact tip of the welding torch from an automatic wire feeder. Heat generated by the molten metal that transferred to the weld pools and from the melting electrode's end to the weld pools. The quality, productivity, and cost of the welding joint are influenced by the GMAW welding parameters. If all welding variables are in compliance, the ideal arc will be produced. Arc Voltage, arc welding current, welding speed, free wire length, nozzle distance, welding location and direction, torch length and gas flow 153 ISSN – 2347-7075 ISSN – 2347-7075 ISSN – 2347-7075 Vol.10 No.5 ISSN – 234 Table.1 Chemical Composition Table.1 Chemical Composition Chemical Composition of Specimen: The ratios of atoms in a chemical substance's molecule are referred to as chemical makeup. In a nutshell, it is the proportion of a specific ingredient in a compound as a percentage. The substance's physical and chemical qualities are determined by its chemical composition. We require the chemical makeup of aluminium and mild steel materials for our study. Specifications for this samples are as given below: Mild Steel - 50mm x 50mm x 6mm, Aluminum - 50mm x 50mmm x 6mm. Prof. Vaibhav V. Kulkarni, Vaishnavi Ghatge, Rutuja Kadam & Vaishnavi Karhale 154 Vol.10 No.5 ISSN – 2347- Table.1 Chemical Composition Vol.10 No.5 Vol.10 No.5 Vol.10 No.5 IJAAR IJAAR Final samples (Al Butt joint): 120 mm x 20 mm x 5 mm Given samples (Al): 120 mm x 50 mm x 6 mm Final samples (Al Butt joint): 120 mm x 20 mm x 5 mm Given samples (Al): 120 mm x 50 mm x 6 mm Final samples (Al Lap joint): 120 mm x 20 mm x 5 mm Fig. 7. Al Lap and Butt Welding Specimen. Fig. 7. Al Lap and Butt Welding Specimen. Given samples (MS): 120 mm x 20 mm x 6 mm Given samples (MS): 120 mm x 20 mm x 6 mm Final samples (MS Lap joint):120 mm x 20 mm x 5 mm Final samples (MS Butt joint): 120 mm x 10 mm x 5 mm Vol.10 No.5 Vol.10 No.5 Mechanical Properties of Specimen: several other mechanical characteristics, including elastic modulus, yield strength, and tensile strength. The ability of a material to resist indentation under a given load and its ability to recover after the load is removed are both affected by these characteristics. The most used mechanical characteristics to evaluate a material’s properties are yield stress, elongation, hardness, and toughness. Yield stress and elongation are measured in a tensile test, where a sample is subjected to a load until it begins to experience plastic strain, and the amount of strain that is not recovered when the load is removed is measured. Hardness, on the other hand, is a quasi- mechanical property that is influenced by Specifications for this samples are as given below: 1. Mild Steel: 15mm x150mm x 5mm 2. Aluminium: 15mm x150mm x 5mm Table.2. Mechanical Properties Table.2. Mechanical Properties Prof. Vaibhav V. Kulkarni, Vaishnavi Ghatge, Rutuja Kadam & Vaishnavi Karhale 155 Table.2. Mechanical Properties Specifications and Preprocessing of Aluminum Sample: Samples were collected and expected dimensions of the samples were achieved by performing various manual processes like cutting by hacksaw, filing etc. Table.2. Mechanical Properties Specifications and Preprocessing of Aluminum Sample: Specifications and Preprocessing of Aluminum Sample: Prof. Vaibhav V. Kulkarni, Vaishnavi Ghatge, Rutuja Kadam & Vaishnavi Karhale 155 Samples were collected and expected dimensions of the samples were achieved by performing various manual processes like cutting by hacksaw, filing etc. processes like cutting by hacksaw, filing etc. 155 Specifications and Preprocessing of Mild Steel Sample: Samples were collected and expected dimensions of the samples were achieved by performing various processes like metal cutting by machine, filing etc. Fig. 8. M.S. Lap and Butt Welding Specimen Fig. 8. M.S. Lap and Butt Welding Specimen Prof. Vaibhav V. Kulkarni, Vaishnavi Ghatge, Rutuja Kadam & Vaishnavi Karhale Hardness Testing: Hardness Testing: Prof. Vaibhav V. Kulkarni, Vaishnavi Ghatge, Rutuja Kadam & Vaishnavi Karhale 156 Vol.10 No.5 IJAAR measuring a material’s resistance to indentation. It involves applying a major load and a minor load to the material surface using different indenters and measuring the depth of penetration of the major load in relation to the minor load. The Rockwell scale employs different indenters and loads to provide various scales, which are identified by a single letter. The hardness value obtained from this test is represented by a dimensionless number with a letter indicating the corresponding Rockwell scale, such as HRA, HRB, HRC, and so on. The hardness of a material is not a fundamental physical property but rather a characteristic that describes its resistance to indentation. In non-destructive hardness testing, a controlled force is applied to a metal surface using a pointed or rounded object, and the permanent depth of the resulting indentation is measured. This depth measurement is then used to determine the material’s hardness. Indentation tests are a common method for measuring hardness, and there are various techniques available to perform such tests on a given material. The Rockwell hardness test is a widely used method for Prof. Vaibhav V. Kulkarni, Vaishnavi Ghatge, Rutuja Kadam & Vaishnavi Karhale Table .3. Hardness Testing Results. MS-MS lap joint (HRB) MS-MS LAP JOINT (100 kgf load) MS-MS LAP POINT 1 2 3 4 5 HRB 92 94 95 90 93 MS-MS butt joint (HRB) MS-MS BUTT JOINT (100 kgf load) MS-MS LAP POINT 1 2 3 4 5 HRB 84 75 79 82 84 Al-Al lap joint (BHN) Al-Al LAP JOINT (187.5 kgf load (P), 2.5 mm steel ball (D) Al-Al LAP POINT 1 2 3 Depression Dia. (d) (mm) 1.55 1.35 1.4 BHN 88.66 120.62 111.35 Al-Al Butt Joint Al-Al BUTT JOINT (187.5 kgf load (P), 2.5 mm steel ball (D) Al-Al LAP POINT 1 2 3 Depression Dia. (d) (mm) 1.58 1.6 1.55 BHN 84.87 82.45 88.66 MS-MS lap joint (HRB) MS-MS LAP JOINT (100 kgf load) MS-MS LAP POINT 1 2 3 4 5 HRB 84 75 79 82 84 MS-MS butt joint (HRB) MS-MS BUTT JOINT (100 kgf load) MS-MS LAP POINT 1 2 3 4 5 HRB 43 42 45 45.5 43 Table .3. Hardness Testing Results. 157 IJAAR Vol.10 No.5 ISSN – 2347-7075 Al-MS butt joint (BHN) AL-MS BUTT JOINT (187.5 kgf (P), 2.5 steel ball diameter (D)) Al-MS Butt point 1 2 3 Depression Dia. Results: Hardness test was performed on all the samples. A weight of 187.5 kg was applied for 15 seconds before being released to test the hardness (HRB) of several materials using an indentation/scratch cone, or “Indenter.” The ratio of the tester’s indenter’s depth during heavy loading to the preload’s penetration. The hardness measured is a number with one unit subtracted, abbreviated HRB, where B is the variable scale Conclusion: Al-MS was welded successfully with filler material Mild steel. When aluminium was used as filler metal joint was not welded as it cracked. This study aims to examine the variation in hardness that occurs when welding two dissimilar metals together. From the results shown above we can conclude that the hardness of weld increases when we weld two dissimilar metals. This increase in hardness can be very beneficial for industrial applications. We have performed welding of two mild steel plates and two aluminium plates. Welding was performed for butt as well as lap joint. In all four of these cases, we have observed that the hardness of the weld of aluminium and mild steel was greater than each sample IJAAR Al-Al LAP joint (BHN) test and microstructure to determine more about obtained specimen. Industrial uses of these types of welding can be huge if applied correctly and implemented without errors. Vol.10 No.5 (d) (mm) 1 1.05 1.02 BHN 228.60 206 219.47 Al-Al LAP joint (BHN) Al-Al LAP JOINT (187.5 kgf load (P), 2.5 mm steel ball (D) Al-MS Lap point 1 2 Depression Dia.(d) (mm) 1 0.95 BHN 228.60 254.60 Vol.10 No.5 ISSN – 2347-7075 IJAAR IJAAR Springer, Singapore. https://doi.org/10.1007/978-981- 287-420-8_27. Scope for Future Work: Prof. Vaibhav V. Kulkarni, Vaishnavi Ghatge, Rutuja Kadam & Vaishnavi Karhale 158 Research can be conducted using different type of metals or even alloys. By selecting a different type of welding process such as tungsten inert gas or plasma arc welding, advancements can be achieved. In this project we have taken lap and butt joint there are many more different types of joints we can explore and study. Hardness testing has been conducted on the given specimens. One can perform more such tests like tensile of weld increases when we weld two dissimilar metals. This increase in hardness can be very beneficial for industrial applications. We have performed welding of two mild steel plates and two aluminium plates. Welding was performed for butt as well as lap joint. In all four of these cases, we have observed that the hardness of the weld of aluminium and mild steel was greater than each sample Research can be conducted using different type of metals or even alloys. By selecting a different type of welding process such as tungsten inert gas or plasma arc welding, advancements can be achieved. In this project we have taken lap and butt joint there are many more different types of joints we can explore and study. Hardness testing has been conducted on the given specimens. One can perform more such tests like tensile Prof. Vaibhav V. Kulkarni, Vaishnavi Ghatge, Rutuja Kadam & Vaishnavi Karhale 158 Prof. Vaibhav V. Kulkarni, Vaishnavi Ghatge, Rutuja Kadam & Vaishnavi Karhale References: 1. Gupta, Y., Tanwar, A., & Gupta, R. (2016). Investigation of Microstructure and mechanical Properties of TIG and MIG Welding Using Aluminum Alloy. 5. Sharma, L. and Kaushal, C. (2015). To determine effects of gas metal arc welding (GMAW) parameters on mechanical properties of aluminum alloys. 2. Nguyen, Q.M., Huang, S.C. (2015). An investigation of the microstructure of an intermetallic layer in welding aluminum alloys to steel by MIG process. International Journal of Advanced Manufacturing Technology, 81(1- 4), 669-675. 6. Sharma, U., & Yadav, V. (2020). Study the effects of Welding Parameters on TIG welding of Aluminum plate. Journal of Physics: Conference Series, 1662(1), 012049. 7. Panchal, K. (2016). Experimental Investigation of TIG Welding on Stainless Steel and Mild Steel Plates. 3. Bhatt, H. (2018). Study of Effect of Process Parameters of Welding during TIG Welding of AA 7075 and its optimization. 8. Ibrahim, I.A., Mohmat, S.A., Amir, A., & Ghalib, A. (2012). The Effect of Gas Metal Arc Welding (GMAW) Processes on different Welding Parameters. International Journal of Mechanical, Aerospace, Industrial, Mechatronic and Manufacturing Engineering, 6(10), 1796-1802. 8. Ibrahim, I.A., Mohmat, S.A., Amir, A., & Ghalib, A. (2012). The Effect of Gas Metal Arc Welding (GMAW) Processes on different Welding Parameters. International Journal of Mechanical, Aerospace, Industrial, Mechatronic and Manufacturing Engineering, 6(10), 1796-1802. 4. Arun M., Ramachandran K. (2015) Effect of Welding Process on Mechanical and Metallurgical Properties of AA 6061 Aluminum Alloy Lap Joint. In: Murugan N. (eds) Materials Science and Manufacturing Engineering. Lecture Notes on Multidisciplinary Industrial Engineering, vol 2. 4. Arun M., Ramachandran K. (2015) Effect of Welding Process on Mechanical and Metallurgical Properties of AA 6061 Aluminum Alloy Lap Joint. In: Murugan N. (eds) Materials Science and Manufacturing Engineering. Lecture Notes on Multidisciplinary Industrial Engineering, vol 2. Prof. Vaibhav V. Kulkarni, Vaishnavi Ghatge, Rutuja Kadam & Vaishnavi Karhale 159
https://openalex.org/W1999741780	https://insu.hal.science/insu-02163277/document	English	null	Geological respiration of a mountain belt revealed by the trace element rhenium	Earth and planetary science letters	2,014	cc-by	11,477	To cite this version: Robert Hilton, Jérôme Gaillardet, Damien Calmels, Jean-Louis Birck. Geological respiration of a mountain belt revealed by the trace element rhenium. Earth and Planetary Science Letters, 2014, 403, pp.27-36. 10.1016/j.epsl.2014.06.021. insu-02163277 Geological respiration of a mountain belt revealed by the trace element rhenium Robert Hilton, Jérôme Gaillardet, Damien Calmels, Jean-Louis Birck Robert G. Hilton a,b,∗, Jérôme Gaillardet b, Damien Calmels c,d, Jean-Louis Birck b obert G. Hilton a,b,∗, Jérôme Gaillardet b, Damien Calmels c,d, Jean-Louis Birck b Robert G. Hilton a Department of Geography, Durham University, Science Laboratories, South Road, Durham, DH1 3LE, UK a Department of Geography, Durham University, Science Laboratories, South Road, Durham, DH1 3LE, UK b Equipe de Géochimie et Cosmochimie, Institut de Physique du Globe de Paris, Sorbonne Paris Cité, Univ Paris Diderot, UMR 7154 CNRS, F-75005 Paris, France c Equipe de Géochimie des Isotopes Stables, Institut de Physique du Globe de Paris, Sorbonne Paris Cité, Univ Paris Diderot, UMR 7154 CNRS, F-75005 Paris, France d Department of Earth Sciences, University of Cambridge, Cambridge, CB2 3EQ, UK a b s t r a c t Article history: Received 7 January 2014 Received in revised form 9 June 2014 Accepted 12 June 2014 Available online 9 July 2014 Editor: T.M. Harrison Keywords: carbon cycle organic carbon rhenium weathering erosion mountain rivers Article history: Received 7 January 2014 Received in revised form 9 June 2014 Accepted 12 June 2014 Available online 9 July 2014 Editor: T.M. Harrison Oxidation of rock-derived, petrogenic, organic carbon (OCpetro) during weathering of sedimentary rocks is a major source of carbon dioxide (CO2) to the atmosphere. This geological respiration is thought to be enhanced by physical erosion, suggesting that mountain belts could release large amounts of CO2 to counter the CO2 sequestration achieved by the erosion, riverine transfer and oceanic burial of organic carbon from the terrestrial biosphere. However, OCpetro oxidation rates in mountain belts have not been quantiﬁed. Here we use rhenium (Re) as a proxy to track OCpetro oxidation in mountain river catchments of Taiwan, where existing measurements of physical erosion rate allow the controls on OCpetro oxidation to be assessed. Re has been shown to be closely associated with OCpetro in rocks and following oxidation during chemical weathering forms a soluble oxyanion (ReO− 4 ) which contributes to the dissolved load of rivers. Soils on meta-sedimentary rocks in Taiwan show that Re loss is coupled to OCpetro loss during weathering, conﬁrming previous observations from soil proﬁles on sedimentary rocks elsewhere. In Taiwan rivers, dissolved Re ﬂux increases with the catchment-average sediment yield, suggesting that physical erosion rate is a major control on OCpetro oxidation. Based on our current understanding of Re mobility during weathering, the dissolved Re ﬂux can be used to quantify an upper bound on the OCpetro oxidation rate and the associated CO2 transfer. The estimated CO2 release from this mountain belt by OCpetro oxidation does not negate estimates of CO2 sequestration by burial of biospheric OC offshore. The ﬁndings are compared to OC transfers estimated for the Himalaya, where OCpetro oxidation in the mountain belt remains unconstrained. Together, these cases suggest that mountain building in the tropics can result in a net sink of OC which sequesters atmospheric CO2. Keywords: carbon cycle organic carbon rhenium weathering erosion mountain rivers Crown Copyright © 2014 Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/3.0/). Contents lists available at ScienceDirect Contents lists available at ScienceDirect * Corresponding author at: Department of Geography, Durham University, Science Laboratories, South Road, Durham, DH1 3LE, UK. Tel.: +44 0191 33 41970. E-mail address: r.g.hilton@durham.ac.uk (R.G. Hilton). HAL Id: insu-02163277 https://insu.hal.science/insu-02163277v1 Submitted on 24 Jun 2019 L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. Earth and Planetary Science Letters 403 (2014) 27–36 4.06.021 14 Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/3.0/). http://dx.doi.org/10.1016/j.epsl.2014.06.021 0012-821X/Crown Copyright © 2014 Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/3.0/). 1. Introduction atmospheric CO2 and O2 concentrations have evolved over geolog- ical time (Hayes et al., 1999; Hayes and Waldbauer, 2006). OCpetro oxidation is thought to occur when sedimentary rocks are exposed to aqueous and gas phase O2 (Chang and Berner, 1999), with the rate of CO2 release controlled by the supply of OCpetro to react (Petsch et al., 2000; Bolton et al., 2006). As such, mountain belts where high rates of physical erosion can supply abundant OCpetro to the surface (Galy et al., 2008a; Hilton et al., 2011) may be locations where this CO2 source is most potent. If true, OCpetro ox- idation may counter the CO2 consumption achieved by the erosion of OCbiosphere and its eﬃcient preservation and burial in sedimen- tary deposits of these settings (France-Lanord and Derry, 1997; Stallard, 1998; Galy et al., 2007; Kao et al., 2014). In order to assess the net impact of mountain building on CO2 ﬂuxes to and from the atmosphere (Caldeira et al., 1993; Derry and France-Lanord, 1996; Gaillardet and Galy, 2008), the rates of OCpetro oxidation during weathering in mountain belts must be quantiﬁed. Organic matter within sedimentary rocks constitutes a vast stock of carbon that was sequestered from the atmosphere in the geological past, containing ∼15 × 1015 gC which is ∼25 000 times the carbon content of the pre-industrial atmosphere (Sundquist and Visser, 2005). Oxidation of this rock-derived, or ‘petrogenic’, organic carbon (OCpetro) during weathering at Earth’s surface is a major source of CO2 to the atmosphere and sink of O2 (Berner and Canﬁeld, 1989; Derry and France-Lanord, 1996). Better under- standing the balance between OCpetro oxidation and the sedimen- tary burial of recently photosynthesized organic carbon from the terrestrial biosphere (OCbiosphere) is fundamental to assessing how R.G. Hilton et al. / Earth and Planetary Science Letters 403 (2014) 27–36 28 OCpetro oxidation has been shown to be signiﬁcant during the ﬂoodplain transport of clastic sediments eroded from the Himalaya and Andes mountain ranges in the Ganges and Amazon rivers, respectively (Galy et al., 2008a; Bouchez et al., 2010). However the method employed, which used the radiocarbon content of OC in river suspended load to determine OCpetro loss from the solid phase, did not have the resolution necessary to quantify OCpetro oxidation rate in the mountain catchments where it should be most rapid. 2.1. Setting The trace element rhenium (Re) is a strong candidate to act as a proxy for OCpetro oxidation during weathering as suggested by previous work (Dalai et al., 2002; Jaffe et al., 2002) due to: i) its association with OCpetro in rocks; and ii) its redox-dependent solu- bility, factors which are closely linked. In oxygenated waters (Eh > 0 V), Re is dominantly present as the soluble perrhenate oxyanion (ReO− 4 ) over a wide range of pH values (5.5 to 9.5) (Brookins, 1986) and in seawater Re behaves conservatively (Anbar et al., 1992; Colodner et al., 1993). In marine depositional environments, OC- bearing sediments experience reduction of ReO− 4 contained in seawater and Re forms a solid phase (Colodner et al., 1993; Crusius et al., 1996; Crusius and Thomson, 2000). Isolation of or- ganic matter from marine sedimentary rocks has suggested that Re is complexed to OCpetro in the depositional environment, favor- ing organic chelating sites to co-existing sulﬁde phases, both in marine sediments with high organic contents (Cohen et al., 1999; Selby and Creaser, 2003) and relatively organic-poor sediments (Pierson-Wickmann et al., 2002; Rooney et al., 2012). Taiwan formed by the collision of the Luzon Arc on the Philip- pine Sea Plate with the Eurasian continental margin. Rapid tec- tonic uplift and seismicity combine with a tropical cyclone climate to drive high physical erosion rates, coupling river incision and bedrock landsliding on steep slopes (Hovius et al., 2000). Sus- pended sediment yields have been quantiﬁed from 1970 for the major rivers and vary around the island, reaching amongst the highest in the world at >10 000 t km−2 yr−1 (Dadson et al., 2003). Rapid physical erosion and turnover of the landscape by landslides results in thin soils (<0.8 m) (Tsai et al., 2001) and high rates of erosion and ﬂuvial export of OCbiosphere (Hilton et al., 2012) which is exported to the ocean and preserved eﬃciently in marine sedi- ments (Kao et al., 2014). Exhumation in the Central Range has exposed passive margin sedimentary rocks which contain ∼0.2–0.4 weight% OCpetro (Hilton et al., 2010). Peak metamorphic temperatures decrease from east (∼500 ◦C) to west (<150 ◦C) across the range (Beyssac et al., 2007). Raman Spectroscopy reveals that OCpetro can be present as graphite in the Tananao Schist and Pilushan formations (Fig. S1) in the east- ern ﬂank of the Central Range mountains (Beyssac et al., 2007). 2.1. Setting Deep-seated bedrock landslides can mobilize unweathered OCpetro along with surface soils and contribute OCpetro to the suspended load of Taiwanese rivers (Hilton et al., 2010), as in other mountain river systems (Clark et al., 2013). Once in the ﬂuvial network, the suspended sediment and OCpetro is rapidly exported with minimal time for subsequent oxidation (Hilton et al., 2008a, 2011). As a re- sult, the solid load of mountain rivers in Taiwan is dominated by unweathered erosion products (Selvaraj and Chen, 2006) and Tai- wan delivers ∼1.7 MtC yr−1 of OCpetro to the ocean in solid form (Hilton et al., 2011). ( , ; y , ) When these sedimentary rocks are re-exposed, for example by erosion and exhumation during orogenesis (e.g. Hilton et al., 2011), Re-bearing OCpetro is subject to chemical weathering in oxic surface environments (Petsch et al., 2000; Jaffe et al., 2002; Pierson-Wickmann et al., 2002). Previous weathering studies from soils on OC-rich rocks have demonstrated that Re loss tracks the OCpetro loss (Jaffe et al., 2002). In contrast, oxidative weathering of sulﬁde minerals continued much deeper into the soil proﬁle where OCpetro and Re were still found at concentrations similar to the un- weathered bedrock (Jaffe et al., 2002). While the fate of gaseous CO2 released by OCpetro oxidation is diﬃcult to trace at the catch- ment scale (Keller and Bacon, 1998), Re will be oxidized to the sol- uble ReO− 4 (Brookins, 1986). Re is not thought to be cycled through the modern terrestrial biosphere (cf. to molybdenum which is used in nitrogenase), meaning that the mobilized ReO− 4 enters the hydrological network (Brookins, 1986; Colodner et al., 1993; Miller et al., 2011). Rivers integrate chemical reactions and hydro- logical sources across the landscape (Gaillardet et al., 1999), thus the dissolved Re ﬂux can be interpreted as the result of oxidative weathering reactions occurring upstream (Dalai et al., 2002). Build- ing on previous work, which has established the close aﬃnity of Re and OCpetro in sedimentary rocks and their coupled behavior dur- ing oxidative weathering, OCpetro oxidation rates may be quantiﬁed from the dissolved Re ﬂux, provided the Re to OCpetro ratio of the rocks which have undergone weathering is known. 1. Introduction This is because chemical denudation rates are usually ∼0.01–0.1 times lower than physical denudation rates in moun- tain belts (Jacobson and Blum, 2003; West et al., 2005), so sus- pended sediments in mountain rivers have a geochemical com- position similar to bedrocks (Hilton et al., 2010). Therefore it is diﬃcult to use the solid products of weathering carried by rivers to quantify OCpetro oxidation in mountain catchments (Hilton et al., 2011). An alternative approach is to track the products of chemical weathering carried in the dissolved load of rivers (Meybeck, 1987; Gaillardet et al., 1999) which has provided detailed constraints on chemical weathering processes and inorganic carbon ﬂuxes from mountain catchments (Jacobson and Blum, 2003; West et al., 2005; Calmels et al., 2011). oxidation, while estimating the associated CO2 release across a mountain belt for the ﬁrst time. We focus on major rivers in Taiwan, where river gauging records constrain the rates and pat- terns of physical erosion (Dadson et al., 2003) and previous work has quantiﬁed the erosion of OCbiosphere (Hilton et al., 2012) and tracked its preservation in marine sediments offshore (Kao et al., 2014). For the ﬁrst time, this allows the role of erosion on OCpetro oxidation to be assessed, and means that an OC budget is within reach for the mountain belt. We examine the mobility of Re from the solid phase during chemical weathering in soils from Taiwan, and measure the dissolved Re concentration in the major rivers. We discuss in detail the assumptions required to use Re as a quan- titative proxy of OCpetro oxidation, in light of the new data from Taiwanese soils and previous work on Re mobility during weather- ing. Our estimates of dissolved Re ﬂux provide new insight on the major controls on OCpetro oxidation and provide an upper bound estimate of the associated CO2 release, allowing us to better assess the impact of orogenesis on the carbon cycle. 2.3. Rhenium concentration Re concentrations in river waters ([Re]diss) are typically low (pmol L−1) and so an established anion-exchange column chem- istry technique was used to pre-concentrate Re and remove the sample matrix (Huffman et al., 1956; Birck et al., 1997). In sum- mary, 100 ml of sample was loaded onto 3 ml of AG1-X8 (200–400 mesh) resin at 0.1 N HNO3 and eluted to cleaned Teﬂon at 4 N HNO3. Re recovery was quantiﬁed using standard solutions (Re- ﬁlament dissolved in HNO3) and matrix effects assessed by Re addition to river samples. Puriﬁed eluted residues were re-taken in 3% HNO3 and analyzed by quadrupole inductively-coupled- plasma mass spectrometry (Q-ICP-MS). The full procedural blank was 1.8 pg, ∼2% of the typical sample mass of Re, similar to pub- lished methods (Dalai et al., 2002). Standards with similar concen- trations as samples not used to construct calibration curves had average precision of 3%. Sample reproducibility was 5% which is taken as the precision on the analyses. Recovery during column chemistry was 100% within this precision. To analyze smaller vol- ume water samples (<20 ml) we developed a standard-addition (SA) method for which pre-concentration is not required and the use of HNO3 is minimized, the major blank contributor (Birck et al., 1997). A set of samples analyzed by Q-ICP-MS following anion- exchange chemistry were also analyzed by SA, using a minimum of three additions, with strong agreement between the measure- ments across the dataset (y = 0.96 ± 0.02x, r = 0.99, P < 0.0001) suggesting [Re]diss by SA-Q-ICP-MS is precise to better than 10% in these samples. In the absence of a major hydrological control on [Re]diss, we examined other controls on [Re]diss. In contrast to previous work (Dalai et al., 2002; Miller et al., 2011), there was no signiﬁ- cant correlation between [SO4] and [Re]diss in Taiwan (Fig. 1A; P = 0.25). [Re]diss was correlated with [K] (r = 0.62, P = 0.008), supporting recent ﬁndings from Indian rivers with minimal an- thropogenic disturbance (Rahaman et al., 2012). In the ﬂood event samples where the instantaneous suspended sediment concen- tration (mg L−1) was measured (n = 4), there was no signiﬁcant correlation with [Re]diss (P = 0.07). The most signiﬁcant corre- lation with [Re]diss was with the catchment average suspended sediment concentration measured over 30 yr (Fig. 2; r = 0.84, P < 0.0001). 3.1. Dissolved rhenium in Taiwan rivers In river waters, [Re]diss range from 4.7 ± 0.2 pmol L−1 to 25.4 ± 1.2 pmol L−1 (Table S1) which are toward the higher end of pub- lished data from global rivers (Fig. 1A) (Colodner et al., 1993; Dalai et al., 2002; Miller et al., 2011; Rahaman et al., 2012). Dif- ferences in [Re]diss between catchments in Taiwan (RSD = 53%) are greater than the mean variability of [Re]diss in a single river catchment (RSD < 22%). Some of this variability is linked to the runoff at the time of sampling, with [Re]diss positively correlated with runoff (Fig. 1B; r = 0.65, P = 0.02). However, samples col- lected from the Liwu River during a ﬂood event have [Re]diss values which are relatively invariant (8% RSD) while runoff varied by up to >10 the long term mean (Fig. 1B). The ﬁndings are consistent with measurements from Arctic rivers, which did not show a sig- niﬁcant relationship between water discharge and [Re]diss (Miller et al., 2011). This contrasts strongly with some major elements which tend to show dilution at high runoff (e.g. Tipper et al., 2006), such as [Na] which is strongly negatively correlated with runoff in the sample set (Fig. 1C; r = −0.88, P = 0.004) and during the ﬂood event (Calmels et al., 2011). The dilution of [Na] is con- sistent with elements mobilized predominantly by acid hydrolysis reactions during silicate weathering (Calmels et al., 2011; Maher et al., 2011). 2.2. Samples River water samples were collected for this study and com- bined with existing samples of river water (Calmels et al., 2011), river bed materials (Hilton et al., 2010) and soils (Hilton et al., 2013). River waters for this study were collected from major rivers in Taiwan during the typhoon season in September 2007 following several days rainfall (Table S1). To establish the hy- drological controls on Re transfer, four additional samples were analyzed from a suite collected in 2004 (Calmels et al., 2011) during a typhoon ﬂood when daily runoff (water discharge nor- malized by the upstream area, mm day−1) reached >10 times Here we examine the transfer of Re in mountain rivers with high erosion rates with a view to assessing the controls on OCpetro R.G. Hilton et al. / Earth and Planetary Science Letters 403 (2014) 27–36 29 der was weighed to Teﬂon bombs and a 185Re enriched-spike was added to each sample. Following equilibration, digestion was achieved with 4 ml of concentrated distilled HF:HNO3 at 150 ◦C. Re was extracted using 3-methyl-1-butanol (iso-amylol) liquid–liquid extraction (Birck et al., 1997). The puriﬁed sample was re-taken in 3% HNO3, analyzed by Q-ICP-MS and corrected for the procedural blank. the long-term mean. Surface waters from turbulent rivers were collected using established methods for Re (Dalai et al., 2002; Miller et al., 2011) in thoroughly rinsed LDPE bottles and ﬁltered immediately through 0.2 μm nylon ﬁlters into acid-cleaned trace analysis grade Nalgene LDPE bottles rinsed with ﬁltered river wa- ter, then acidiﬁed to pH ∼2 (trace analysis grade HNO3). An addi- tional un-acidiﬁed aliquot was collected for major anions. Filtered samples were stored at 4 ◦C in the dark. Daily water discharge (m3/s) was obtained from the Water Resources Agency, Taiwan. 3. Results To assess whether the coupled loss of Re and OCpetro observed elsewhere in OC-rich bedrocks occurs during weathering of rocks in Taiwan, solid weathering products collected from upper soil horizons in the Wulu catchment were used (Fig. S1). These were collected by Hilton et al. (2013), as ∼500 cm3 of soil over a depth of ∼10 cm, corresponding to A and E Horizons of variable hu- miﬁed organic matter with coarse and ﬁne fractions bearing little structure of existing bedrock. Prior to analysis, soil and river bed materials samples were each homogenized to an integrated bulk sample. 2.4. Major ion and elemental concentrations To assess the geochemical composition of rocks at the scale of river catchments (e.g. Tipper et al., 2006; Galy et al., 2008b) river bed materials previously collected from three major river catch- ments (Hilton et al., 2010) were analyzed. These drain a range of geological formations which are representative of the wider Cen- tral Range, Taiwan (Fig. S1; Table S2). As described by Hilton et al. (2010), ∼500 cm3 of sand material were collected from alluvium deposited in bedrock channels at low ﬂow in March 2006. The majority of these sediments were deposited following the ﬂood associated with Typhoon LongWang (October 2005) (Turowski et al., 2008) and were likely to have been sourced by 100–1000s of bedrock landslides upstream (Hovius et al., 2000). Given the tran- sient nature of sediment storage in Taiwan rivers and the short ﬂuvial transit times (Hilton et al., 2008a), these river bed materials can provide constraint on catchment-averaged bedrock composi- tion. In river water, major ion concentrations were analyzed by Ion Chromatography (Table S1). In river bed materials and soils, total sulfur concentrations were determined by combustion and major elements (CaO and Al2O3) determined by ICP-OES at SARM, CRPG, Nancy (Table S2). The OC and nitrogen concentrations (%) were measured in previous work following inorganic carbon removal (Hilton et al., 2010, 2013). The radiocarbon activity of soil samples was also determined by Hilton et al. (2013) and is reported here as the fraction modern (Fmod) corrected to −25h δ13CVPDB based on measured stable isotope composition, where Fmod = 1 is the 14C content of 1950 atmosphere and Fmod = 0 is a sample which contains no measurable 14C (Stuiver and Polach, 1977). The OCpetro content of soil samples was quantiﬁed using Fmod measurements and a binary mixing model as outlined in Appendix A. 2.3. Rhenium concentration This is a measure of decadal physical erosion rate normalized by catchment runoff (Dadson et al., 2003). In addition to this trend, catchments draining higher grade metamorphic rocks In solid samples, [Re] was determined by isotope dilution fol- lowing established methods (Birck et al., 1997). 50 mg of pow- R.G. Hilton et al. / Earth and Planetary Science Letters 403 (2014) 27–36 30 30 R.G. Hilton et al. / Earth and Planetary Science Letters 403 (2014) 27–36 Fig. 1. A. Dissolved rhenium concentration, [Re]diss (pmol L−1), versus sulfate concentration, [SO4] (μmol L−1), for Taiwan rivers (black circles) and a global river compilation (Dalai et al., 2002; Miller et al., 2011; Rahaman et al., 2012). Across all measurements there is a positive correlation (r = 0.44; P < 0.0001) but considerable scatter, and values are not correlated for the Taiwan dataset (P = 0.25). B. [Re]diss versus instantaneous runoff (mm day−1) for Taiwan river samples (labeled by catchment). C. Dissolved sodium concentration, [Na] (μmol L−1), versus instantaneous runoff for the same samples shown in part B. Bars indicate the precision on the analysis where larger than the point size. R.G. Hilton et al. / Earth and P Fig. 1. A. Dissolved rhenium concentration, [Re]diss (pmol L−1), versus sulfate concentration, [SO4] (μmol L−1), for Taiwan rivers (black circles) and a global river compilation (Dalai et al., 2002; Miller et al., 2011; Rahaman et al., 2012). Across all measurements there is a positive correlation (r = 0.44; P < 0.0001) but considerable scatter, and values are not correlated for the Taiwan dataset (P = 0.25). B. [Re]diss versus instantaneous runoff (mm day−1) for Taiwan river samples (labeled by catchment). C. Dissolved sodium concentration, [Na] (μmol L−1), versus instantaneous runoff for the same samples shown in part B. Bars indicate the precision on the analysis where larger than the point size Fig. 2. Dissolved rhenium concentration ([Re]diss in pmol L−1) for Taiwan rivers sampled in 2004 and 2007 versus the catchment average suspended sediment con- centration for 1970–1999 (Dadson et al., 2003) (errors not provided) a measure of physical erosion per unit of water runoff. Vertical lines show analytical precision (if larger than the point size). in the northeast ﬂank of the Central Range (Beyssac et al., 2007; Hilton et al., 2010), such as the Liwu, Hualien, Hsiukuluan (Fig. 3.2. Rhenium in river bed materials and soils The river bed materials in Taiwan represent the solid prod- ucts of erosion and have [Re] values which range from 0.60 to 0.96 ppb (Table S2). These are lower than measured in one of the ﬁrst studies to propose Re as a tracer of OCpetro oxidation, with [Re] = 75–116 ppb in a Devonian black shale (Jaffe et al., 2002). The difference is mirrored by the low OCpetro content of river bed materials in Taiwan, [OCpetro] = 0.17–0.23%, compared to [OCpetro] = 6.0–7.8% in the Devonian shale (Jaffe et al., 2002). How- ever, [Re] in Taiwanese river bed sediments are similar to river bed sediments measured in the Himalaya of [Re] < 1.2 ppb (Rahaman et al., 2012) and Himalayan rocks (Pierson-Wickmann et al., 2002) where OCpetro contents of river bed loads are similar to Taiwan, with [OCpetro] < 0.2% (Galy et al., 2008b). The [Re] of river bed ma- terials from Taiwan are also similar to modern marine sediments in sub-oxic waters (Crusius and Thomson, 2000), where OC con- tents and [Re] are lower than sediments in anoxic basins (Ravizza et al., 1991). We note that [Re] may vary with the grain size of river bed materials in analogy to major elements (Bouchez et al., 2011), however, a larger sample set is necessary to establish these controls. Fig. 2. Dissolved rhenium concentration ([Re]diss in pmol L−1) for Taiwan rivers sampled in 2004 and 2007 versus the catchment average suspended sediment con- centration for 1970–1999 (Dadson et al., 2003) (errors not provided) a measure of physical erosion per unit of water runoff. Vertical lines show analytical precision (if larger than the point size). Fig. 2. Dissolved rhenium concentration ([Re]diss in pmol L−1) for Taiwan rivers sampled in 2004 and 2007 versus the catchment average suspended sediment con- centration for 1970–1999 (Dadson et al., 2003) (errors not provided) a measure of physical erosion per unit of water runoff. Vertical lines show analytical precision (if larger than the point size). rates likely to have occurred at the time of burial, which will in- ﬂuence the [Re] and [OCpetro] (McKay et al., 2007). In addition, the [Re]:[OCpetro] may vary with metamorphic grade and the ther- mal maturity of OCpetro, although this remains poorly constrained (Rooney et al., 2012; Cumming et al., 2014). 2.3. Rhenium concentration S1), appear to have slightly higher [Re]diss than catchments draining lower grade metamorphic rocks at similar erosion rates (Laonung, Chihpen). However, the scatter in the relationship (Fig. 2) may also be controlled by hydrological variability in [Re]diss (Fig. 1B) and longer-term records of [Re]diss are needed to examine second or- der trends in detail. 4.1. Insights on the source of dissolved rhenium The association of Re and OCpetro in sedimentary rocks and the solubility of Re upon oxidation has lead previous work to suggest that [Re]diss may provide a robust tracer of OCpetro oxi- dation reactions (e.g. Dalai et al., 2002). However, it is important to assess the source of [Re]diss and address the validity of as- sumptions which link dissolved Re ﬂux to OCpetro oxidation. First, the solid residue of chemical weathering provides insight on Re mobilization to the dissolved load. Bulk soil samples from Tai- wan have lower [OCpetro] and [Re] when compared to the river bed materials (Fig. 3), which are likely to have a composition close to unweathered bedrock following their recent mobilization by bedrock landslides (Hovius et al., 2000). This result is consis- tent with observations of coupled Re and OCpetro loss made on soil proﬁles elsewhere (Peucker-Ehrenbrink and Hannigan, 2000; Jaffe et al., 2002) and suggests coupled oxidative weathering of Re and OCpetro. Once oxidized to the perrhenate anion ReO− 4 , Re is soluble in oxygenated waters (Eh > 0 V) with pH ranges 5.5–9.5 (Brookins, 1986). As a result, Re loss from soils should be recorded in the [Re]diss of rivers and its concentration inﬂuenced by the rate of weathering of the Re-bearing substrate (Colodner et al., 1993). The decoupling between [Re]diss and major elements with in- creasing runoff, such as [Na] (Figs. 1B and C), is also consistent with this explanation and suggests Re is mobilized from sur- face soils. Having taken rainwater inputs into account, dissolved Na (and Mg, Ca) is predominantly sourced from the inorganic rock matrix by acid hydrolysis reactions (Gaillardet et al., 1999; Tipper et al., 2006; Calmels et al., 2011). Laboratory experiments on fresh minerals show that acid hydrolysis reactions occur <10 times slower than oxidation weathering reactions (Chang and Berner, 1999; White and Brantley, 2003). Therefore longer ﬂow paths which lengthen the timescales of ﬂuid–rock interaction are required to elevate [Na] (Maher et al., 2011) and so [Na] is strongly diluted with increasing runoff (Fig. 1C). As a result, groundwater inputs which are most signiﬁcant at low runoff can make up a large proportion (∼70%) of the Na ﬂux in Taiwan (Calmels et al., 2011). In contrast, Re is not diluted with runoff (Fig. 1B) and so groundwater inputs are likely to be relatively minor. Overall, the behavior of [Re]diss (Figs. 3.2. Rhenium in river bed materials and soils Therefore, we propose that the global correlation does not indicate a common source of Re and SO2− 4 , but instead reﬂects changes in evaporation and/or dilution which set a ﬁrst order control on all dissolved ion concentra- tions, especially in large rivers (Gaillardet et al., 1999). This would also explain why [Re]diss is better correlated with [Ca] (r = 0.57; P < 0.0001) in the same compilation of data shown in Fig. 1A. We are therefore wary of using correlations between dissolved ion concentrations to interpret a common source during chemical weathering (cf. Miller et al., 2011). The signiﬁcant correlation between [Re]diss in Taiwan rivers and the catchment-average suspended sediment concentration be- tween 1970 and 1999 (Fig. 2) suggests that the rate of physi- cal erosion over recent decades plays a major role in controlling [Re]diss of a catchment. This observation sheds new light on the weathering processes which mobilize Re from rocks to the dis- solved phase. Physical erosion in Taiwan is dominated by bedrock landslides which scour deep (>1 m) into the rock mass (Hovius et al., 2000). An increase in the decadal average erosion rate there- fore reﬂects an increase in the frequency at which unweathered bedrock is exposed at the surface by landslides. If weathering re- actions are mainly limited by the supply of new minerals, then increased physical erosion rate should result in an increase in the concentration of dissolved products in river waters. Previous work has hypothesized that OCpetro oxidation in soils under present at- mospheric O2 levels are limited by the supply of OCpetro to an oxidative weathering zone near the surface (Petsch et al., 2000; Bolton et al., 2006) and may be enhanced by microbial assimila- tion in surface soils (Petsch et al., 2001). Therefore, the observed increase in [Re]diss with erosion rate (Fig. 2) support this hypoth- esis and suggest that supply limits OCpetro oxidative weathering, even at the very high rates of physical erosion in Taiwan. Fig. 3. Solid phase samples, river bed materials (black squares) and homogenized surface (A and E) soil horizons (white circles) from Taiwan. Rhenium concentration, [Re], and petrogenic organic carbon concentration, [OCpetro], have been normalized to an immobile element (aluminum, Al2O3) to track weathering processes. Arrow indicates coupled loss of Re and OCpetro from soil horizons from Taiwan conﬁrming work on OCpetro-rich lithologies elsewhere (Jaffe et al., 2002). 3.2. Rhenium in river bed materials and soils Vertical lines show propagated uncertainty on [OCpetro] (Appendix A) if larger than the point size. [OCpetro], higher [Re] and a lower metamorphic grade than Taiwan (Jaffe et al., 2002). [ p ] g [ (Jaffe et al., 2002). 3.2. Rhenium in river bed materials and soils Nevertheless, the re- sults conﬁrm the importance of constraining a local [Re]:[OCpetro] in upstream lithologies when using Re to examine OCpetro weath- ering (Dalai et al., 2002). The mean ratio of [Re] to [OCpetro] in the three river bed ma- terial samples is 3.7 × 10−7 and the values range within a factor of two from 2.6 × 10−7 to 4.6 × 10−7. These ratios are lower than those measured on two sets of OC-rich Devonian black shale, with mean [Re]:[OCpetro] = 12.8 ±3.4 ×10−7 (n = 5, ±2 SE) (Jaffe et al., 2002) and mean [Re]:[OCpetro] = 12.2 ± 5.0 × 10−7 (n = 8, ±2 SE) (Selby and Creaser, 2003). However, they are similar to OC-rich Jurassic shale (mean [OCpetro] = 20 ± 8%) with mean [Re]:[OCpetro] = 3.5 ± 0.9 × 10−7 (n = 5, ±2 SE) (Cohen et al., 1999). The dif- ferences in [Re]:[OCpetro] between localities is expected given the different redox conditions, sedimentation rates and bioturbation The thin surface soils have low [OCpetro] < 0.05% (Table S2), consistent with oxidative loss of OCpetro during weathering. The bulk soils have [Re] = 0.07–0.19 ppb, lower than the river bed ma- terials. Soils are also depleted in Ca and S when compared to bed materials (Table S2), consistent with carbonate weathering and sul- ﬁde oxidation of the sedimentary rocks (Calmels et al., 2011). The coupled loss of Re and OCpetro relative to an immobile element is evident in Taiwanese soils (Fig. 3). The results are consistent with observations from soil proﬁles elsewhere, where rocks had higher R.G. Hilton et al. / Earth and Planetary Science Letters 403 (2014) 27–36 31 Fig. 3. Solid phase samples, river bed materials (black squares) and homogenized surface (A and E) soil horizons (white circles) from Taiwan. Rhenium concentration, [Re], and petrogenic organic carbon concentration, [OCpetro], have been normalized to an immobile element (aluminum, Al2O3) to track weathering processes. Arrow indicates coupled loss of Re and OCpetro from soil horizons from Taiwan conﬁrming work on OCpetro-rich lithologies elsewhere (Jaffe et al., 2002). Vertical lines show propagated uncertainty on [OCpetro] (Appendix A) if larger than the point size. P < 0.0001), it is extremely scattered (Fig. 1A). For [Re]diss val- ues similar to Taiwan, the global data shows [Re]diss varies by >50 pmol L−1 for a given [SO4]. 4.1. Insights on the source of dissolved rhenium 1B and 2) is more consistent with it being liberated from an oxygenated surface weathering zone by oxidation reactions which occur much quicker than acid hydrolysis. In river waters, previous work has suggested that a correla- tion between [Re]diss and [SO4] may imply a common source (i.e. sulﬁde minerals) during weathering (Miller et al., 2011). In Tai- wan, sulﬁde oxidation is an important weathering reaction and the dominant source of SO2− 4 to rivers (Yoshimura et al., 2001; Calmels et al., 2011) and so the weak correlation between [SO4] and [Re]diss in Taiwan (Fig. 1A) may suggest sulﬁde minerals are not the dominant host of Re (Selby and Creaser, 2003). While this is consistent with the available measurements from soils and river bed materials in Taiwan (Fig. 3), it contrasts with previous conclusions drawn from observations linking [Re]diss to [SO4] in Himalayan rivers (Dalai et al., 2002) and globally (Miller et al., 2011). The discrepancy can be reconciled by noting that while the global trend between [Re]diss and [SO4] is signiﬁcant (r = 0.44; q y y The observation that oxidative weathering of OCpetro is supply- limited may seem at odds with measurements of signiﬁcant export of unweathered OCpetro in the suspended load of Taiwanese rivers (Hilton et al., 2010, 2011). However, this reﬂects the dominant role of bedrock landslides in the erosion of mountain belts like Taiwan (Hovius et al., 2000). Landslides erode the weathering products from thin soils, which can be fully depleted in OCpetro (Fig. 3) and delivered them to river channels. However, the volume of bedrock landslides increases as a power law (exponent > 1.2) of the land- slide area (Larsen et al., 2010) and so large, deep-seated landslides can deliver signiﬁcant amounts of unweathered OCpetro to chan- nels (Hilton et al., 2008b), overwhelming soil inputs. Thus a large R.G. Hilton et al. / Earth and Planetary Science Letters 403 (2014) 27–36 32 Table 1 Rhenium-derived OCpetro oxidation rate estimates for river catchments in Taiwan. 4.2. Dissolved rhenium ﬂux and estimates of OCpetro oxidation The observed link between [Re]diss and sediment yield (Fig. 2), along with the coupled loss of Re and OCpetro from soils in Tai- wan (Fig. 3), suggest that [Re]diss provides a powerful proxy of OCpetro oxidation in Taiwan. Here we describe how the dissolved Re ﬂux can provide insight on the CO2 release by OCpetro oxida- tion, and consider the assumptions which need to be applied. The ﬁrst step is to quantify the dissolved Re ﬂux, accounting for sys- tematic changes in concentration with runoff (Calmels et al., 2011). [Re]diss was not diluted with increasing runoff (Fig. 1B) and so Re ﬂux can be quantiﬁed robustly using an average [Re]diss. Re ﬂuxes for individual catchments were calculated, based on between 1 and 6 samples per catchment (Table 1). While this is relatively small number of samples and catchment-scale Re ﬂux may be reﬁned by longer records, the number is similar to work quantifying Re ﬂuxes in large rivers (Miller et al., 2011). At the mountain belt scale, the area-weighted mean [Re]diss in the sampled catchments is 13.2 ± 3.7 pmol L−1 (n = 10, ±2 SE, Table 1) and for an annual water discharge of 49.8 km3 yr−1 (Dadson et al., 2003) the dis- solved Re ﬂux from Taiwan to the ocean is 700 ± 200 mol yr−1. The Re yield estimates from individual river catchments in Taiwan were correlated to the sediment yield measured independently from 1970 to 1999 (Fig. 4A). pp q To calculate OCpetro oxidation rates, the dissolved Re ﬂux (Fig. 4A) must be combined with measurements of the initial [Re] to [OCpetro] ratio of sedimentary rocks in the catchment. Individual bedrock samples may signiﬁcantly overestimate the heterogeneity present at the catchment scale and instead, river bed materi- als can provide a robust average sample of the major geological formations upstream (Galy et al., 2008a; Hilton et al., 2010). In Tai- wan, [Re]:[OCpetro] of the three river bed material samples ranged within a factor of two, from 2.6 × 10−7 to 4.6 × 10−7. While the sample set of river bed materials is small (n = 3), we note that the samples analyzed here have nitrogen to OC ratios between 0.1 and 0.2 (Table S2) which covers a range of compositions repre- sentative of OCpetro in the major geological formations (Hilton et al., 2010). Table 1 Table 1 Rhenium-derived OCpetro oxidation rate estimates for river catchments in Taiwan. proportion of the exhumed OCpetro does not spend signiﬁcant time in the weathering zone before being exported by rivers. Bedrock landslides can therefore explain both the high rates of OCpetro ox- idation by increasing mineral supply at landslide sites (Fig. 2) and also the rapid ﬂuvial export and re-burial of un-weathered OCpetro offshore Taiwan (Kao et al., 2014). Re during weathering (Fig. 3). However, soil proﬁles from else- where have shown that oxidation can result in slightly higher losses of Re from the solid phase (∼99%) compared to losses of OCpetro (∼80%) (Jaffe et al., 2002; Pierson-Wickmann et al., 2002). In addition, rocks in Taiwan can contain graphitic-OCpetro (Beyssac et al., 2007) which has been shown to be resilient to chemical and physical breakdown in river catchments (Galy et al., 2008a; Bouchez et al., 2010) but it is not known whether graphitic-OCpetro contains Re. Based on our current understanding of Re mobility, these assumptions mean that the dissolved Re ﬂux is likely to rep- resent an upper bound on the OCpetro oxidation rate. These details warrant future research to reﬁne the Re-proxy, however our new ﬁndings on the mobility of Re during weathering in Taiwan (Figs. 2 and 3) suggest that a Re-derived estimate of OCpetro oxidation rate offers a robust, upper bound quantiﬁcation. Re during weathering (Fig. 3). However, soil proﬁles from else- where have shown that oxidation can result in slightly higher losses of Re from the solid phase (∼99%) compared to losses of OCpetro (∼80%) (Jaffe et al., 2002; Pierson-Wickmann et al., 2002). In addition, rocks in Taiwan can contain graphitic-OCpetro (Beyssac et al., 2007) which has been shown to be resilient to chemical and physical breakdown in river catchments (Galy et al., 2008a; Bouchez et al., 2010) but it is not known whether graphitic-OCpetro contains Re. Based on our current understanding of Re mobility, these assumptions mean that the dissolved Re ﬂux is likely to rep- resent an upper bound on the OCpetro oxidation rate. These details warrant future research to reﬁne the Re-proxy, however our new ﬁndings on the mobility of Re during weathering in Taiwan (Figs. 2 and 3) suggest that a Re-derived estimate of OCpetro oxidation rate offers a robust, upper bound quantiﬁcation. 4.1. Insights on the source of dissolved rhenium Catchment Drainage area (km2) Suspended sediment yield (t km−2 yr−1)a Water discharge (g yr−1)a Decadal average SSC (g L−1)a nb [Re]diss (pmol L−1) RSDc (%) Re ﬂux (g yr−1) OCpetro oxidation rate (tC km−2 yr−1)d Wulu 639 17 371 1.41E+15 7.9 3 14 10 3629 12−22 Yenping 476 19 118 1.00E+15 9.1 1 13 − 2417 11−19 Hsiukuluan 249 12 851 6.23E+14 5.1 1 10 − 1192 10−18 Hualien 1506 20 850 3.16E+15 9.9 2 20 21 11 634 17−30 Liwu 435 33 103 1.04E+15 13.8 6 21 22 4010 20−35 Chenyoulan 367 8719 6.97E+14 4.6 3 9 2 1130 7−12 Upper Chenyoulan 205 2927 4.10E+14 1.5 1 5 – 358 4−7 Laonung 853 10 785 1.62E+15 5.7 1 5 – 1505 4−7 Taimarli 190 2105 4.37E+14 0.9 1 6 – 458 5−9 Chihpen 166 21 687 3.98E+14 9.0 1 8 – 571 7−13 Taiwan average 35 980 10 672 4.98E+16 – 10 13 53 122 308 7−13 a Averages from 1970 to 1999 (Dadson et al., 2003), with suspended sediment concentration (SSC) calculated from decadal mean sediment transfer and water discharge. b Number of [Re]diss analyses per catchment. c Relative Standard Deviation (RSD) on [Re]diss measurements. d Range of estimates derived from the range of measured [Re]:[OCpetro] ratios in river bed materials (Table S2). 4.2. Dissolved rhenium ﬂux and estimates of OCpetro oxidation In contrast, ∼0.8 MtC yr−1 of OCpetro is thought to be oxidized during ﬂoodplain transport once OCpetro has left the mountain belt (Galy et al., 2008a). OCpetro ox- idation rates have not been quantiﬁed in Himalayan catchments where CO2 emissions may be higher due to faster physical ero- sion (e.g. Fig. 4B). However, denudation rates in the Himalaya are lower than Taiwan (Dadson et al., 2003; Lupker et al., 2012), and the upward corrected OCpetro oxidation rate is unlikely to counter the CO2 sink by burial of OCbiosphere. The Re-derived OCpetro oxidation rates for individual catchments (Table 1) are positively related to the decadal suspended sediment yield (Fig. 4B). The catchment-scale OCpetro oxidation rates may be reﬁned using longer time-series of [Re]diss measurements and catchment-speciﬁc [Re]:[OCpetro] ratios of river bed materials. Nev- ertheless, the relationship is consistent with a Re-derived OCpetro oxidation rate made using the same methods in the Yamuna River draining the Himalaya (9600 km2), where OCpetro oxidation is esti- mated to release ∼4 tC km−2 yr−1 of CO2 (Dalai et al., 2002) with denudation ∼1700 t km−2 yr−1 (Lupker et al., 2012) (Fig. 4B). Our new data conﬁrm the hypothesis that physical erosion rate is a major control on OCpetro oxidation and moderates CO2 release dur- ing weathering (Petsch et al., 2000), even at the very high erosion rates experienced in Taiwan. p Together with our new data, these ﬁndings suggest that in mountain belts located in the subtropics, the CO2 sequestration achieved by OCbiosphere export and burial is greater than the CO2 emissions by geological respiration during oxidative weath- ering of OCpetro (Fig. 5). The frontal Himalaya and Taiwan expe- rience high physical erosion rates driven by ﬂuvial incision and bedrock landsliding, which can supply abundant OCpetro for ox- idation (Hilton et al., 2011). However, both orogenic belts fea- ture productive terrestrial biomass which is eroded rapidly and is buried with clastic sediment (Galy et al., 2007; Hilton et al., 2012; Kao et al., 2014). The balance of these OC ﬂuxes presently act as a sink of CO2 (Fig. 5). However, the net OC cycle in Taiwan appears to be sensitive to small increases in the CO2 release by OCpetro weath- ering, which will vary spatially (and temporally) in response to changes in the rate of physical erosion (Fig. 4B). 4.2. Dissolved rhenium ﬂux and estimates of OCpetro oxidation Dissolved rhenium (Re) yield (g km−2 yr−1) versus average suspended sediment yield 1970–1999 (Dadson et al., 2003) for individual catchments in Taiwan (symbols as in Fig. 2). Vertical bars denote the mean standard deviation in dissolved Re concentration in the sampled catchments (Table 1). B. Petrogenic organic carbon (OCpetro) oxidation rates (tC km−2 yr−1) calculated using the dissolved Re yield for Taiwan catchments (gray rectangles) versus suspended sediment yield. The range of estimates represent uncertainty based on the measured range of Re to OCpetro ratios in river bed materials. The black star shows OCpetro oxidation rate estimated previously using dissolved Re ﬂux from the Yamuna River catchment, Himalaya (Dalai et al., 2002). Fig. 4. A. Dissolved rhenium (Re) yield (g km−2 yr−1) versus average suspended sediment yield 1970–1999 (Dadson et al., 2003) for individual catchments in Taiwan (symbols as in Fig. 2). Vertical bars denote the mean standard deviation in dissolved Re concentration in the sampled catchments (Table 1). B. Petrogenic organic carbon (OCpetro) oxidation rates (tC km−2 yr−1) calculated using the dissolved Re yield for Taiwan catchments (gray rectangles) versus suspended sediment yield. The range of estimates represent uncertainty based on the measured range of Re to OCpetro ratios in river bed materials. The black star shows OCpetro oxidation rate estimated previously using dissolved Re ﬂux from the Yamuna River catchment, Himalaya (Dalai et al., 2002). ﬁnd that cycling of OC during the erosion and weathering of Tai- wan presently acts as a net sink of atmospheric CO2 (Fig. 5). (graphite) from the solid load (Galy et al., 2008a; Bouchez et al., 2010, 2014). In Taiwan, the combination of bedrock landslides (see Section 4.1), rapid ﬂuvial transit times and strong land–ocean cou- pling mean that despite the rapid OCpetro oxidation rates (Fig. 4B) the majority of exhumed OCpetro (>80%) is eroded and reburied offshore Taiwan (Kao et al., 2014). The ﬁndings can be compared to the Ganges–Brahmaputra river system draining the Himalaya, where source to sink sedi- ment and OC transfers have been partly quantiﬁed (France-Lanord and Derry, 1997; Galy et al., 2007, 2008a; Gaillardet and Galy, 2008). There terrestrial OCbiosphere is buried eﬃciently in rapidly accumulating, O2 depleted waters of the Bay of Bengal and se- questers 3.7 MtC yr−1 derived from a continental source area of ∼1.6 × 106 km2 (Galy et al., 2007). 4.2. Dissolved rhenium ﬂux and estimates of OCpetro oxidation The samples provide a plausible range of [Re]:[OCpetro] of the source and allow us to assess a range of CO2 emission estimates. With the knowledge that OCpetro oxidation rates de- rived from dissolved Re ﬂux are likely to be an upper bound, and that the initial [Re]:[OCpetro] of rocks is the main source of uncer- tainty, we estimate a maximum CO2 release by OCpetro oxidation of between 0.27 and 0.47 MtC yr−1 from this mountain belt. Nor- malized over the island source area (35 980 km2) the CO2 yield is 7.4–13.0 tC km−2 yr−1. This is a signiﬁcant carbon transfer and is higher than estimates of CO2 drawdown by silicate weathering in mountain belts of 3–4 tC km−2 yr−1 (Jacobson and Blum, 2003). The results suggest that in Taiwan the chemical denudation of OCpetro by oxidation represents <20% of the total OCpetro denuda- tion (weathering plus erosion) with ∼1.7 MtC yr−1 exported to the ocean in solid form (Hilton et al., 2011). In contrast, the extensive ﬂoodplains downstream of the Himalaya and Andes mountains (of the Ganges and Amazon rivers, respectively) provide time for addi- tional OCpetro oxidation, removing all but the most resilient OCpetro Having quantiﬁed the oxidized Re ﬂux from river catchments, the OCpetro oxidation rate can be estimated provided that the ini- tial [Re]:[OCpetro] of rocks is known. In addition, two assumptions are necessary: i) that Re is hosted primarily in OCpetro; and ii) that oxidation mobilizes [Re]diss and releases CO2 from OCpetro at the same rate. The ﬁrst assumption would appear to be valid based on the high Re contents of organic matter isolated from sedi- mentary rocks, with OCpetro shown to dominate the total mass of Re (Cohen et al., 1999; Pierson-Wickmann et al., 2002; Selby and Creaser, 2003; Rooney et al., 2012). However, some Re can be sourced from inorganic phases (Dalai et al., 2002), which may host ∼30% of Re in marine sediments (Selby and Creaser, 2003). A Re- derived OCpetro oxidation rate will be an overestimation if this is the case. Regarding the second assumption, weathering products from soil horizons in Taiwan show coupled loss of OCpetro and R.G. Hilton et al. / Earth and Planetary Science Letters 403 (2014) 27–36 33 R.G. Hilton et al. / Earth and Planetary Science Letters 403 (2014) 27–36 33 Fig. 4. A. 4.2. Dissolved rhenium ﬂux and estimates of OCpetro oxidation It is also sensitive to changes in the distribution of mountain forest and soil which might reduce the CO2 sequestration associated with erosion and burial of OCbiosphere (Hilton et al., 2012). These factors should be central in controlling the impact of mountain building on the long- term sequestration of CO2 in the lithosphere (Derry and France- Lanord, 1996) and the inﬂuence of orogenesis on the evolution of atmospheric CO2 and O2 concentrations (Caldeira et al., 1993; Berner and Canﬁeld, 1989; Gaillardet and Galy, 2008). The im- 5. Conclusions where f denotes the fraction of OC derived from petrogenic and biosphere sources ( fpetro and fbiosphere, respectively) where fpetro + fbiosphere = 1 for a binary mixture. The radiocarbon activity of petrogenic OC (Fmod)petro = 0 by deﬁnition (radiocarbon content indistinguishable from background). The C-weighted mean Fmod of upper soil horizons in Taiwan has been measured by Hilton et al. (2008a, 2008b) as Fmod = 0.98 ± 0.07 (n = 10, ±SD) and this is taken as the radiocarbon activity of the OCbiosphere end member composition (Fmod)biosphere. The rates and controls on CO2 release by OCpetro oxidation in a mountain belt were assessed using the trace element Re. In solid weathering products from Taiwan, Re loss was coupled to OCpetro loss (Fig. 3), conﬁrming previous work from soil proﬁles on OCpetro-rich rocks. In river waters from Taiwan, [Re]diss was posi- tively, signiﬁcantly correlated with the decadal average suspended sediment concentration (1970–1999), a measure of physical ero- sion rate normalized by runoff (Fig. 2). We estimate Re ﬂux for the sampled catchments and ﬁnd that the Re yield also increases with sediment yield (Fig. 4A). An increase in [Re]diss and Re ﬂux in basins which erode more rapidly is consistent with the dominant source of Re from a surﬁcial weathering zone, where oxidation re- actions are limited by mineral supply. p Eq. (A.1) can be solved for fpetro and the measured fpetro com- bined with measured total organic carbon concentration (Corg, %) to quantity the [OCpetro], in % (Kao et al., 2014). The uncertainty on [OCpetro] content of soils is derived from the propagation of the variability on the end member compositions (Fig. 3). The analysis provides an upper estimate of OCpetro content because aging of soil organic matter can also deplete 14C content, rather than OCpetro addition (Hilton et al., 2013). Previous work has established that OC in the river bed material samples is dominated by OCpetro us- ing element ratios and stable carbon isotopes (Hilton et al., 2010). y pp y The Re ﬂux measurements are used to estimate the CO2 release by OCpetro oxidation with knowledge of the initial [Re]:[OCpetro] of bedrocks, and assuming: i) that Re is hosted primarily in OCpetro; and ii) congruent release of [Re]diss and CO2 from OCpetro occurs during oxidative weathering. 4.3. The net OC budget of a mountain belt The Re-derived OCpetro oxidation rate can be used to assess the net OC budget of a mountain belt for the ﬁrst time. Physical erosion in Taiwan results in the mobilization of OCbiosphere from soils and vegetation and its transported to the oceans in river sus- pended load (Hilton et al., 2008a). The erosion of the OCbiosphere results in the export of ∼0.5 MtC yr−1 (1012 gC yr−1) of OCbiosphere from Taiwan, a value which is thought to be conservative (Hilton et al., 2012). Recent observations show that OCbiosphere is eﬃ- ciently preserved (70–100%) in sediments offshore Taiwan (Kao et al., 2014), in analogy with tectonic margins elsewhere (Galy et al., 2007; Blair and Aller, 2012), with an estimate of CO2 sequestra- tion of 0.5–0.6 MtC yr−1 by terrestrial OCbiosphere burial (Kao et al., 2014). Additional OC burial occurs as marine organic matter and has not been quantiﬁed. Therefore, given that the estimate of CO2 release provided here is likely to be an upper bound (Section 4.2), and that the estimates of CO2 drawdown are likely to be conserva- tive (Kao et al., 2014), the assessment of net OC sequestration will also be conservative. Taking the estimated bounds into account, we R.G. Hilton et al. / Earth and Planetary Science Letters 403 (2014) 27–36 34 Fig. 5. Net organic carbon balance in Taiwan expressed in terms of CO2 ﬂuxes to (positive) and from (negative) the atmosphere. Petrogenic organic carbon (OCpetro) oxidation rate derived from dissolved Re ﬂux (Fig. 4A, this study) is differenced from the transfer and burial of OC from the terrestrial biosphere (OCbiosphere) transfer and burial (Hilton et al., 2012; Kao et al., 2014) to assess the net OC cycle, with upper (dashed) and lower (dotted) bounds for net CO2 sequestration. to negate the estimated CO2 drawdown by erosion of OC from the terrestrial biosphere, its ﬂuvial transport and marine burial off- shore. Mountain building in Taiwan presently acts as an OC sink, sequestering atmospheric CO2 during weathering and erosion. Appendix A. Quantiﬁcation of OCpetro content in soil samples The OCpetro content (%) of soil samples was determined us- ing Fmod measurements (Hilton et al., 2013) and a binary mixing model which assumes contemporary inputs (OCbiosphere) mix with radiocarbon depleted OCpetro (e.g. Galy et al., 2008a; Hilton et al., 2008a; Clark et al., 2013): pact of changing physical erosion rates on the global carbon cycle has received renewed attention via the process of sulﬁde oxida- tive coupled to carbonate weathering (Calmels et al., 2007) over interglacial–glacial cycles (Georg et al., 2013) and during the Ceno- zoic (Torres et al., 2014). Our ﬁndings suggest a more direct link via the oxidative weathering of OCpetro and CO2 release (Fig. 4B) and better understanding the rates and patterns of OCpetro oxida- tion remains a research priority. (Fmod)petro × fpetro + (Fmod)biosphere × fbiosphere = (Fmod)sample (Fmod)petro × fpetro + (Fmod)biosphere × fbiosphere = (Fmod)sample (A.1) Acknowledgements Data associated with this research can be found in the sup- plementary materials. The research was funded by a Natural Environment Research Council (NERC), UK, New Investigator Grant to R.G. Hilton (NE/I001719/1) and by a CNRS EC2CO grant (‘OXYMORE’) to J. Gaillardet. R.G. Hilton acknowledges Research Leave from Durham University and funding from IPG, Paris, as a Visiting Professor. We thank F. Campas and J. Moureau for lab- oratory and analytical assistance and A. Galy, A.J. West, E. Tip- per, D. Selby and J. Bouchez for discussions, and L. Derry and three anonymous referees for their comments which improved this work. Fig. 5. Net organic carbon balance in Taiwan expressed in terms of CO2 ﬂuxes to (positive) and from (negative) the atmosphere. Petrogenic organic carbon (OCpetro) oxidation rate derived from dissolved Re ﬂux (Fig. 4A, this study) is differenced from the transfer and burial of OC from the terrestrial biosphere (OCbiosphere) transfer and burial (Hilton et al., 2012; Kao et al., 2014) to assess the net OC cycle, with upper (dashed) and lower (dotted) bounds for net CO2 sequestration. Appendix B. Supplementary material Supplementary material related to this article can be found on- line at http://dx.doi.org/10.1016/j.epsl.2014.06.021. 5. Conclusions We note that the ﬁrst assumption is validated by previous work, but may lead to an overestimation of CO2 ﬂux if some dissolved Re is derived from inorganic minerals. The second assumption appears to be valid based on our measure- ments of OCpetro and Re loss in weathering products in Taiwan (Fig. 3) and is supported by previous observations of coupled loss of Re and OCpetro in soil weathering proﬁles. However, based on our current understanding of Re mobility, it may lead to an over- estimation of OCpetro oxidation rate if Re is liberated more rapidly during weathering. The patterns of CO2 release across Taiwan are controlled by physical erosion rate (Fig. 4B), conﬁrming mountain belts as hotspots of CO2 release by geological respiration. The ab- solute rates of CO2 emission in river catchments estimated from dissolved Re ﬂux are likely to be an upper bound. Nevertheless, at the mountain belt scale, the OCpetro oxidation rate is not suﬃcient Anbar, A.D., Creasar, R.A., Papanastassiou, D.A., Wasserburg, G.J., 1992. Rhenium in seawater: conﬁrmation of generally conservative behaviour. Geochim. Cos- mochim. Acta 56, 4099–4103. Berner, R.A., Canﬁeld, D.E., 1989. A new model for atmospheric oxygen over Phanerozoic time. Am. J. Sci. 289, 333–361. Beyssac, O., et al., 2007. Late Cenozoic metamorphic evolution and exhumation of Taiwan. Tectonics 26, TC6001. http://dx.doi.org/10.1029/2006TC002064. Anbar, A.D., Creasar, R.A., Papanastassiou, D.A., Wasserburg, G.J., 1992. Rhenium in seawater: conﬁrmation of generally conservative behaviour. Geochim. Cos- mochim. Acta 56, 4099–4103. Berner, R.A., Canﬁeld, D.E., 1989. A new model for atmospheric oxygen over Phanerozoic time. Am. J. Sci. 289, 333–361. Beyssac, O., et al., 2007. Late Cenozoic metamorphic evolution and exhumation of Taiwan. Tectonics 26, TC6001. http://dx.doi.org/10.1029/2006TC002064. References Paciﬁc) – inﬂuence on the accumulation of redox-sensitive trace metals. Chem. Geol. 238, 180–196. Crusius, J., Calvert, S.E., Pedersen, T.F., Sage, D., 1996. Rhenium and molybdenum enrichments in sediments as indicators of oxic, suboxic and anoxic conditions of deposition. Earth Planet. Sci. Lett. 145, 65–79. Meybeck, M., 1987. Global chemical weathering of surﬁcial rocks estimated from river dissolved loads. Am. J. Sci. 287, 401–428. Cumming, V.M., Selby, D., Lillis, P.G., Lewan, M.D., 2014. Re–Os geochronology and Os isotope ﬁngerprinting of petroleum sourced from a Type I lacustrine kerogen: insights from the natural Green River petroleum system in the Uinta Basin and hydrous pyrolysis experiments. Geochim. Cosmochim. Acta 138, 32–56. Miller, C.A., Peucker-Ehrenbrink, B., Walker, B.D., Marcantonio, F., 2011. Re-assessing the surface cycling of molybdenum and rhenium. Geochim. Cosmochim. Acta 75, 7146–7179. Dadson, S.J., et al., 2003. Links between erosion, runoff variability and seismicity in the Taiwan orogen. Nature 426, 648–651. Petsch, S.T., Berner, R.A., Eglinton, T.I., 2000. A ﬁeld study of the chemical weathering of ancient sedimentary organic matter. Org. Geochem. 31, 475–487. 14 Dalai, T.K., Singh, S.K., Trivedi, J.R., Krishnaswami, S., 2002. Dissolved rhenium in the Yamuna River System and the Ganga in the Himalaya: role of black shale weathering on the budgets of Re, Os, and U in rivers and CO2 in the atmosphere. Geochim. Cosmochim. Acta 66, 29–43. Petsch, S.T., Eglinton, T.I., Edwards, K.J., 2001. 14C-dead living biomass: evidence for microbial assimilation of ancient organic matter during shale weathering. Sci- ence 292, 1127–1131. Peucker-Ehrenbrink, B., Hannigan, R.E., 2000. Effects of black shale weathering on the mobility of rhenium and platinum group elements. Geology 28, 475–478. k b d h f d Derry, L.A., France-Lanord, C., 1996. Neogene growth of the sedimentary organic car- bon reservoir. Paleoceanography 11, 267–275. Pierson-Wickmann, A.C., Reisberg, L., France-Lanord, C., 2002. Behavior of Re and Os during low-temperature alteration: results from Himalayan soils and altered black shales. Geochim. Cosmochim. Acta 66, 1539–1548. France-Lanord, C., Derry, L.A., 1997. Organic carbon burial forcing of the carbon cycle from the Himalayan erosion. Nature 390, 65–67. Rahaman, W., Singh, S.K., Shukla, A.D., 2012. Rhenium in Indian rivers: sources, ﬂuxes, and contribution to oceanic budget. Geochem. Geophys. Geosyst. 13, Q08019. Gaillardet, J., Galy, A., 2008. Himalaya – carbon sink or source? Science 320, 1727–1728. Gaillardet, J.B., Dupré, B., Louvat, P., Allègre, C.J., 1999. Global silicate weathering and CO2 consumption rates deduced from the chemistry of large rivers. Chem. References Cooling in the late Ceno- zoic. Nature 361, 123–124. Calmels, D., Gaillardet, J., Brenot, A., France-Lanord, C., 2007. Sustained sulﬁde ox- idation by physical erosion processes in the Mackenzie River basin: climatic perspectives. Geology 35, 1003–1006. Jacobson, A.D., Blum, J.D., 2003. Relationship between mechanical erosion and at- mospheric CO2 consumption in the New Zealand Southern Alps. Geology 31, 865–868. Calmels, D., et al., 2011. Contribution of deep groundwater to the weathering budget in a rapidly eroding mountain belt, Taiwan. Earth Planet. Sci. Lett. 303, 48–59. Jaffe, L.A., Peucker-Ehrenbrink, B., Petsch, S.T., 2002. Mobility of rhenium, platinum group elements and organic carbon during black shale weathering. Earth Planet. Sci. Lett. 198, 339–353. Chang, S., Berner, R.A., 1999. Coal weathering and the geochemical carbon cycle. Geochim. Cosmochim. Acta 63, 3301–3310. Kao, S.J., et al., 2014. Preservation of terrestrial organic carbon in marine sediments offshore Taiwan: mountain building and atmospheric carbon dioxide sequestra- tion. Earth Surf. Dyn. 2, 127–139. Clark, K.C., et al., 2013. New views on “old” carbon in the Amazon river: insight from the source of organic carbon eroded from the Peruvian Andes. Geochem. Geophys. Geosyst. 14, 1644–1659. http://dx.doi.org/10.1002/ggge.20122. Keller, C.K., Bacon, D.H., 1998. Soil respiration and georespiration distinguished by transport analyses of vadose CO2, 13CO2 and 14CO2. Glob. Biogeochem. Cy- cles 12, 361–372. Cohen, A.S., Coe, A.L., Bartlett, J.M., Hawkesworth, C.J., 1999. Precise Re–Os ages of organic-rich mudrocks and the Os isotope composition of Jurassic seawater. Earth Planet. Sci. Lett. 167, 159–173. Larsen, I.J., Montgomery, D.R., Korup, O., 2010. Landslide erosion controlled by hill- slope material. Nat. Geosci. 3, 247–251. http://dx.doi.org/10.1038/ngeo776. 10 Colodner, D., et al., 1993. The geochemical cycles of rhenium: a reconnaissance. Earth Planet. Sci. Lett. 117, 205–221. Lupker, M., et al., 2012. 10Be-derived Himalayan denudation rates and sediment budgets in the Ganga basin. Earth Planet. Sci. Lett. 333–334, 146–156. h h l f ﬂid id i d hi l i d Crusius, J., Thomson, J., 2000. Comparative behavior of authigenic Re, U, and Mo dur- ing reoxidation and subsequent long-term burial in marine sediments. Geochim. Cosmochim. Acta 64, 2233–2242. Maher, K., 2011. The role of ﬂuid residence time and topographic scales in deter- mining chemical ﬂuxes from landscapes. Earth Planet. Sci. Lett. 312, 48–58. d i A di d di i i McKay, J.L., Pedersen, T.F., Mucci, A., 2007. Sedimentary redox conditions in con- tinental margin sediments (N.E. References Berner, R.A., Canﬁeld, D.E., 1989. A new model for atmospheric oxygen over Phanerozoic time. Am. J. Sci. 289, 333–361. R.G. Hilton et al. / Earth and Planetary Science Letters 403 (2014) 27–36 35 Birck, J.L., Barman, M.R., Capmas, F., 1997. Re–Os isotopic measurements at the fem- tomole level in natural samples. J. Geostand. Geoanal. 20, 19–27. Hilton, R.G., Galy, A., Hovius, N., 2008b. Riverine particulate organic carbon from an active mountain belt: importance of landslides. Glob. Biogeochem. Cycles 22, GB1017. http://dx.doi.org/10.1029/2006GB002905. Blair, N.E., Aller, R.C., 2012. The fate of terrestrial organic carbon in the marine en- vironment. Annu. Rev. Mar. Sci. 4, 17.1–17.23. Hilton, R.G., Galy, A., Hovius, N., Horng, M.-J., Chen, H., 2010. The isotopic com- position of particulate organic carbon in mountain rivers of Taiwan. Geochim. Cosmochim. Acta 74, 3164–3181. Bolton, E.W., Berner, R.A., Petsch, S.T., 2006. The weathering of sedimentary organic matter as a control on atmospheric O2: II. Theoretical modeling. Am. J. Sci. 306, 575–615. Hilton, R.G., Galy, A., Hovius, N., Horng, M.-J., Chen, H., 2011. Eﬃcient transport of fossil organic carbon to the ocean by steep mountain rivers: an orogenic carbon sequestration mechanism. Geology 39, 71–74. Bouchez, J., et al., 2010. Oxidation of petrogenic organic carbon in the Amazon ﬂoodplain as a source of atmospheric CO2. Geology 38, 255–258. Hilton, R.G., et al., 2012. Climatic and geomorphic controls on the erosion of ter- restrial biomass from subtropical mountain forest. Glob. Biogeochem. Cycles 26, GB3014. Bouchez, J., Gaillardet, J., France-Lanord, C., Maurice, L., Dutra-Maia, P., 2011. Grain- size controls of river suspended sediment geochemistry: insights from Amazon River depth proﬁles. Geochem. Geophys. Geosyst. 12, GC00380. Hilton, R.G., Galy, A., West, A.J., Hovius, N., Roberts, G.G., 2013. Geomorphic control on the δ15N of mountain forests. Biogeosciences 10, 1693–1705. Bouchez, J., et al., 2014. Source, transport and ﬂuxes of Amazon River particulate or- ganic carbon: insights from river sediment depth-proﬁles. Geochim. Cosmochim. Acta 133, 280–298. Hovius, N., Stark, C.P., Chu, H.-T., Lin, J.-C., 2000. Supply and removal of sediment in a landslide-dominated mountain belt: Central Range, Taiwan. J. Geol. 108, 73–89. Brookins, D.G., 1986. Rhenium as an analog for ﬁssiogenic technetium: Eh–pH dia- gram (25 ◦C, 1 bar) constraints. Appl. Geochem. 1, 513–517. Huffman, E.H., Oswalt, R.L., Williams, L.A., 1956. Anion-exchange separation of molybdenum and technetium and of tungsten and rhenium. J. Inorg. Nucl. Chem. 3, 49–53. Caldeira, K., Arthur, M.A., Berner, R.A., Lasaga, A.C., 1993. White, A.F., Brantley, S.L., 2003. The effect of time on the weathering of silicate minerals: why do weathering rates differ in the laboratory and the ﬁeld? Chem. Geol. 202, 479–506. Yoshimura, K., et al., 2001. Geochemical and stable isotope studies on natural wa- ter in the Taroko Gorge karst area, Taiwan – chemical weathering of carbonate rocks by deep source CO2 and sulfuric acid. Chem. Geol. 177, 415–430. West, A.J., Galy, A., Bickle, M., 2005. Tectonic and climatic controls on silicate weath- ering. Earth Planet. Sci. Lett. 235, 211–228. West, A.J., Galy, A., Bickle, M., 2005. Tectonic and climatic controls on silicate weath- ering. Earth Planet. Sci. Lett. 235, 211–228. White, A.F., Brantley, S.L., 2003. The effect of time on the weathering of silicate minerals: why do weathering rates differ in the laboratory and the ﬁeld? Chem. Geol. 202, 479–506. Yoshimura, K., et al., 2001. Geochemical and stable isotope studies on natural wa- ter in the Taroko Gorge karst area, Taiwan – chemical weathering of carbonate rocks by deep source CO2 and sulfuric acid. Chem. Geol. 177, 415–430. Torres, M.A., West, A.J., Li, G., 2014. Sulphide oxidation and carbonate dissolution as a source of CO2 over geological timescales. Nature 507, 346–349. References Geol. 159, 3–30. Ravizza, G.E., Turekian, K.K., Hay, B.J., 1991. The geochemistry of rhenium and os- mium in recent sediments from the Black Sea. Geochim. Cosmochim. Acta 55, 3741–3752. Galy, V., et al., 2007. Eﬃcient organic carbon burial in the Bengal fan sustained by the Himalayan erosional system. Nature 450, 407–410. Rooney, A.D., Selby, D., Lewan, M.D., Lillis, P.G., Houzay, J.P., 2012. Evaluating Re–Os systematics in organic-rich sedimentary rocks in response to petroleum gen- eration using hydrous pyrolysis experiments. Geochim. Cosmochim. Acta 77, 275–291. Galy, V., Beyssac, O., France-Lanord, C., Eglinton, T.I., 2008a. Recycling of graphite during Himalayan erosion: a geological stabilization of carbon in the crust. Sci- ence 322, 943–945. Selby, D., Creaser, R.A., 2003. Re–Os geochronology of organic rich sediments: an evaluation of organic matter analysis methods. Chem. Geol. 200, 225–240. Galy, V., France-Lanord, C., Lartiges, B., 2008b. Loading and fate of particulate or- ganic carbon from the Himalaya to the Ganga–Brahmaputra delta. Geochim. Cosmochim. Acta 72, 1767–1787. Selvaraj, K., Chen, C.T.A., 2006. Moderate chemical weathering of subtropical Taiwan: constraints from solid-phase geochemistry of sediments and sedimentary rocks. J. Geol. 114, 101–116. Georg, R.B., West, A.J., Vance, D., Newmand, K., Halliday, A.N., 2013. Is the marine osmium isotope record a probe for CO2 release from sedimentary rocks? Earth Planet. Sci. Lett. 367, 28–38. Stallard, R.F., 1998. Terrestrial sedimentation and the carbon cycle: coupling weath- ering and erosion to carbon burial. Glob. Biogeochem. Cycles 12, 231–257. i l h i i i f 14 d di b ( ) Hayes, J.M., Waldbauer, J.R., 2006. The carbon cycle and associated redox processes through time. Philos. Trans. R. Soc. Lond. B 361, 931–950. Stuiver, M., Polach, H.A., 1977. Discussion: reporting of 14C data. Radiocarbon 19 (3), 355–363. Hayes, J.M., Strauss, H., Kaufman, A.J., 1999. The abundance of 13C in marine organic matter and isotopic fractionation in the global biogeochemical cycle of carbon during the past 800 Ma. Chem. Geol. 161, 103–125. Sundquist, E.T., Visser, K., 2005. In: Schlesinger, W. (Ed.), Biogeochemistry. Elsevier, Oxford, pp. 425–472. Tipper, E.T., et al., 2006. The short term climatic sensitivity of carbonate and silicate weathering ﬂuxes: insight from seasonal variations in river chemistry. Geochim. Cosmochim. Acta 70, 2737–2754. Hilton, R.G., et al., 2008a. Tropical-cyclone-driven erosion of the terrestrial biosphere from mountains. Nat. Geosci. 1, 759–762. R.G. Hilton et al. / Earth and Planetary Science Letters 403 (2014) 27–36 R.G. Hilton et al. / Earth and Planetary Science Letters 403 (2014) 27–36 36 Torres, M.A., West, A.J., Li, G., 2014. Sulphide oxidation and carbonate dissolution as a source of CO2 over geological timescales. Nature 507, 346–349. Tsai, C.C., Chen, Z.S., Duh, C.T., Horng, F.W., 2001. Prediction of soil depth using a soil–landscape regression model: a case study on forest soils in southern Tai- wan. Proc. Natl. Sci. Counc., Repub. China, Part B 26, 34–39. Turowski, J.M., Hovius, N., Hsieh, M.L., Lague, D., Chen, M.C., 2008. Distribu- tion of erosion across bedrock channels. Earth Surf. Process. Landf. 33, 353–363.
https://openalex.org/W3019301105	https://hal.univ-lille.fr/hal-04347974/document	English	null	Hexahydrofarnesyl as an original bio-sourced alkyl chain for the preparation of glycosides surfactants with enhanced physicochemical properties	RSC advances	2,020	cc-by	10,048	To cite this version: Guillaume Lemahieu, Julie Aguilhon, Henri Strub, Valerie Molinier, Jesus Ontiveros, et al.. Hex- ahydrofarnesyl as an original bio-sourced alkyl chain for the preparation of glycosides surfactants with enhanced physicochemical properties. RSC Advances, 2020, RSC Advances, 10, pp.16377-16389. 10.1039/d0ra02326d. hal-04347974 Distributed under a Creative Commons Attribution 4.0 International License Hexahydrofarnesyl as an original bio-sourced alkyl chain for the preparation of glycosides surfactants with enhanced physicochemical properties† Guillaume Lemahieu,a Julie Aguilhon,b Henri Strub,b Val´erie Molinier,c Jes´us F. Ontiverosa and Jean-Marie Aubry a Five new bio-based surfactants have been synthetized by coupling hexahydrofarnesol with mono and di- saccharides. Hexahydrofarnesol (3,7,11-trimethyl-dodecan-1-ol) is a by-product of the industrial production of farnesane, a sustainable aviation fuel obtained by a fermentation process from sugar feedstocks. Using hexahydrofarnesol as the lipophilic starting material allows obtaining 100% bio-based surfactants while valorizing an industrial by-product. Moreover, the C15-branched alkyl chain brings unique properties to the surfactants. This paper presents a physicochemical characterization of these new surfactants including their behaviors in water (water solubility, critical micellar concentration and surface tension) and in oil/water systems (interfacial tension against model oil and ternary phase behavior). Their hydrophilicities have been determined thanks to the PIT-slope method and compared to the ones of standard surfactants with linear alkyl chains, in order to distinguish the contributions of the sugar polar heads and of the branched hexahydrofarnesyl lipophilic chain. This novel class of surfactants combines the properties of sugar-based surfactants (low sensitivity to temperature and salinity, ability to form Winsor III microemulsion systems over a wide range of salinity), along with speciﬁcities linked to the branched alkyl chain (lower Kraﬀt temperature, low surface tension). Received 12th March 2020 Accepted 13th April 2020 aUniv. Lille, CNRS, Centrale Lille, ENSCL, Univ. Artois, UCCS – Unit´e de Catalyse et Chimie du Solide, UMR 8181, F-59000 Lille, France. E-mail: jean-marie.aubry@ univ-lille.fr bTotal Raﬃnage Chimie, Biofuels Division, Tour Coupole, 2 Place Jean Millier, 92400 Courbevoie, France. E-mail: julie.aguilhon@total.com cTotal Exploration Production, Pˆole d’Etudes et de Recherche de Lacq, B. P. 47, 64170 Lacq, France. E-mail: valerie.molinier@total.com † Electronic supplementary information (ESI) available. See DOI: 10.1039/d0ra02326d HAL Id: hal-04347974 https://hal.univ-lille.fr/hal-04347974v1 Submitted on 15 Dec 2023 L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. Distributed under a Creative Commons Attribution 4.0 International License aUniv. Lille, CNRS, Centrale Lille, ENSCL, Univ. Artois, UCCS – Unit´e de Catalyse et Chimie du Solide, UMR 8181, F-59000 Lille, France. E-mail: jean-marie.aubry@ univ-lille.fr Open Access Article. Published on 24 April 2020. Downloaded on 12/15/2023 8:36:02 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. Open Access Article. Published on 24 April 2020. Downloaded on 12/15/2023 8:36:02 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. s Attribution-NonCommercial 3.0 Unported Licence. View Article Online View Journal \| View Issue RSC Advances Cite this: RSC Adv., 2020, 10, 1637 Received 12th March 2020 Accepted 13th April 2020 DOI: 10.1039/d0ra02326d rsc.li/rsc-advances PAPER cle. Published on 24 April 2020. Downloaded on 12/15/2023 8:36:02 PM. article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. 1 Introduction and dispersing properties9,10 are also assets that motivate the replacement of polyethoxylated non-ionic surfactants classically used.11 Indeed, due to regulation evolution regarding poly- ethoxylated surfactants and their replacement in some markets, lots of academic and industrial researches are focused on the development of APG and other sugar-based surfactants as alternatives. 1 Alkyl-polyglucosides (APG) were the rst surfactants derived from sugars synthetized as early as in 1933.1–3 The rst appli- cation of these surfactants came 40 years later in the eld of detergency and was given up due to the relative high production cost compared to the existing polyethoxylated non-ionic surfactants. In the middle of the 80s, valorization strategies for agricultural resources and the awareness of the environ- mental footprint of chemicals led to a renewed interest in this class of surfactants.4 Since the beginning of the 90s, several companies have produced and used their own APG such as Kao Corp.,5 H¨uls6 and Henkel.7 Nowadays, APG are used not only in detergency but also in various applications elds such as pharmaceutical production or agricultural formulations due to their high compatibility with human skin and body as well as their low impact on the environment.8 Their higher foaming Sugar-based surfactants have also proved to be eﬀective dispersing agents in case of oil spills.12 They can also be used as non-ionic co-surfactants in association with anionic surfac- tants, to achieve WIII microemulsion systems by decreasing the interfacial tension between oil and water to ultralow values (around 103 mN m1).13 This is of particular interest for Enhanced Oil Recovery (EOR) applications. Interestingly, APG are able to form WIII microemulsion systems in high salinity and high temperature conditions14,15 and can be used to generate low interfacial tension systems that are largely inde- pendent of both salinity and temperature, contrarily to poly- ethoxylated surfactants classically used.16 The structure of the sugar-based surfactant itself is an important parameter to play with for adjusting the aﬃnity for both oil and water phases. 2 Material and methods g y y of linear chains with an even number of carbons. Petro-based alcohols generally have medium to highly branched alkyl chains, obtained through oligomerization of alkenes (“oxo”- type) or by Guerbet reaction. Both the length and the rami- cation of the chain play an important role on the physico- chemical properties of the nal surfactant.19 For instance, it is known that the ramication of the chain promotes the forma- tion of WIII microemulsion systems by bringing more disorder in the interfacial lm.20,21 Having in hands a well-dened bio- based branched alcohol is valuable for studying structure– properties relationship and envisaging innovative and 100% bio-based surfactants. In this context, ve well-dened starting sugars have been combined with hexahydrofarnesol (HHF, 3,7,11-trimethyl- dodecan-1-ol) (Table 1). HHF is a by-product of the industrial production of farnesane (2,6,10-trimethyldodecane), a sustain- able aviation fuel that can be used as a “drop-in” product blended with traditional petroleum-derived jet.22 Farnesane is itself obtained by hydrogenation of farnesene, a sesquiterpene obtained from sugar feedstocks by biotechnological processes. During this fermentation, farnesol is obtained as a side product, and converted into HHF in the hydrogenation process. As HHF is recovered in the heavy fraction during the distillation of farnesane, it can be considered as a by-product. Coupling HHF with sugars is a smart way to valorize this by-product into potentially eﬃcient and innovative 100% bio-sourced sugar surfactants. 2. H ne Fa ex xy w co di ac at be 10 ch la an in ha of th n N de pu Table 1 The ﬁve sugar-based surfactants from HHF used in this study Name Starting sugar MW (g mol1) HHFA Arabinose 360.53 HHFX Xylose 360.53 HHFG Glucose 390.55 HHFMan Mannose 390.55 HHFMalt Maltose 552.70 Open Access Article. Published on 24 April 2020. Downloaded on 12/15/2023 8:36:02 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. 1 Introduction In that way, QSPR models can be useful tools to design tailored surfactants.17 The diversity of available sugars and the multiplicity of the substitution posi- tions make it possible to adjust the polarity of the head.18 Regarding the lipophilic tail, bio-based hydrophobic RSC Adv., 2020, 10, 16377–16389 \| 16377 This journal is © The Royal Society of Chemistry 2020 Paper View Article Online Paper View Article Online View Article Online RSC Advances components generally come from acids or alcohols derived from vegetable or fat oils. They are usually obtained as mixtures of linear chains with an even number of carbons. Petro-based alcohols generally have medium to highly branched alkyl chains, obtained through oligomerization of alkenes (“oxo”- type) or by Guerbet reaction. Both the length and the rami- cation of the chain play an important role on the physico- chemical properties of the nal surfactant.19 For instance, it is known that the ramication of the chain promotes the forma- tion of WIII microemulsion systems by bringing more disorder in the interfacial lm.20,21 Having in hands a well-dened bio- based branched alcohol is valuable for studying structure– properties relationship and envisaging innovative and 100% bio-based surfactants. Open Access Article. Published on 24 April 2020. Downloaded on 12/15/2023 8:36:02 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. The hexahydrofarnesyl glycoside samples were prepared externally by coupling hexahydrofarnesol with arabinose, xylose, glucose, mannose and maltose. Alkyl monoglycosides HHFA, HHFX, HHFG and HHFMan were prepared by direct Fisher glycosylation starting from the corresponding commercial anhydrous sugars. The sugars were dispersed in a large excess of HHF (100 mL/10 g sugar) under acidic catalysis (H2SO4 – 2% mol) and under vacuum (10 mbar) at 105 C. At the end of the reaction, the samples obtained have been puried by silica gel chromatography with an elution of 100% dichloromethane followed by 100% ethyl acetate. The chemical pathway is exemplied in Scheme 1 for HHFA. In this context, ve well-dened starting sugars have been combined with hexahydrofarnesol (HHF, 3,7,11-trimethyl- dodecan-1-ol) (Table 1). HHF is a by-product of the industrial production of farnesane (2,6,10-trimethyldodecane), a sustain- able aviation fuel that can be used as a “drop-in” product blended with traditional petroleum-derived jet.22 Farnesane is itself obtained by hydrogenation of farnesene, a sesquiterpene obtained from sugar feedstocks by biotechnological processes. During this fermentation, farnesol is obtained as a side product, and converted into HHF in the hydrogenation process. As HHF is recovered in the heavy fraction during the distillation of farnesane, it can be considered as a by-product. Coupling HHF with sugars is a smart way to valorize this by-product into potentially eﬃcient and innovative 100% bio-sourced sugar surfactants. HHFMalt was synthetized from the commercial peracety- lated maltose. Glycosylation proceeded by activation of the anomeric position thanks to BF3EtO2, followed by deprotection in a methanolic solution, as displayed in Scheme 2. HHFMalt has been puried by silica gel chromatography with an elution of ethyl acetate and methanol. Each puried sample was dried under vacuum (0.5 mbar) in the presence of P2O5 to remove all solvent contaminants. The nal samples have been characterized by HPLC-MS, 1H and 13C NMR (Fig. 1–20 in ESI†) and exhibit a purity higher than 95%, as determined by 1H NMR. Table 2 recaps the isolated yields and purities of the samples. For HHFA (Fig. 2.1. Chemicals Hexahydrofarnesol was isolated during the distillation of far- nesane, which was prepared by hydrogenation of farnesene.23,24 Farnesene was obtained from Amyris. 16378 \| RSC Adv., 2020, 10, 16377–16389 Open Access Article. Published on 24 April 2020. Downloaded on 12/15/2023 8:36:02 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. 1 and 2 in ESI†), two Table 1 The ﬁve sugar-based surfactants from HHF used in this study Name Starting sugar MW (g mol1) Structure HHFA Arabinose 360.53 HHFX Xylose 360.53 HHFG Glucose 390.55 HHFMan Mannose 390.55 HHFMalt Maltose 552.70 Table 1 The ﬁve sugar-based surfactants from HHF used in this study This journal is © The Royal Society of Chemistry 2020 16378 \| RSC Adv., 2020, 10, 16377–16389 This journal is © The Royal Society of Chemistry 2020 Scheme 1 Synthetic pathway and puriﬁcation procedure for the preparation of hexahydrofarnesyl monoglycosides HHFA, HHFX, HHFG and ( ) Paper RSC Advances View Article Online Paper RSC Advances Open Access Article. Published on 24 April 2020. Downloaded on 12/15/2023 8:36:02 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. Scheme 1 Synthetic pathway and puriﬁcation procedure for the preparation of hexahydrofarnesyl monoglycosides HHFA, HHFX, HHFG and HHFMan (exempliﬁed with HHFA). Scheme 1 Synthetic pathway and puriﬁcation procedure for the preparation of hexahydrofarnesyl monoglycosides HHFA, HHFX, HHFG and HHFMan (exempliﬁed with HHFA) Scheme 2 Synthetic pathway and puriﬁcation for the preparation of hexahydrofarnesyl maltoside HHFMalt. ed under a Creative Commons Attribution-NonCommercial 3 Scheme 2 Synthetic pathway and puriﬁcation for the preparation of hexahydrofarnesyl maltoside HHFMalt. Scheme 2 Synthetic pathway and puriﬁcation for the preparation of hexahydrofarnesyl maltoside HHFMalt. Table 2 Retention times determined by HPLC-MS analysis (same chromatographic conditions, see ESI), isolated yields and purities of the diﬀerent sugar-based surfactants synthetized 2.2. Aqueous solubilization For estimating the aqueous solubility of the samples, 1% wt, 0.1% wt and 0.01% wt solutions were prepared in deionized or salted aqueous media. For homogenization, the samples were vigorously stirred with a magnetic stirrer in a water bath at 50 C for approximately 3 hours, and then leat 50 C during 15 hours in an oven. Surfactant Isolated yield (%) Retention time (min) Purity (%) determined by 1H NMR HHFA 22 4.4 (23%) and 4.9 (77%) >95 HHFX 46 4.7 >95 HHFG 33 3.1 >95 HHFMan 37 3.7 >95 HHFMalt 25 2.2 >95 For the 1% wt and 0.1% wt samples, the solubility was visually inspected at 50 C, and then aer 15 hours at room temperature and at 83 C. For the 0.01% wt samples, the solu- bility was visually inspected at room temperature, and aer 15 hours at 65.5 C and 83 C. For determining aqueous solubility limits of HHFA, HHFX, HHFG and HHFMan samples, the 0.01% wt solutions were ltered on 0.2 mm pore size syringe lter (Pall Corporation, Acrodisc PF syringe lters, diameter 0.8 cm/0.2 mm) two consecutive times at room temperature. The ltrate was dosed by HPLC/MS, thanks to a calibration line established before- hand on a 5–100 ppm scale. peaks with the same mass have been detected by HPLC-MS and correspond to the mixture of pyranose and furanose forms in the ratio 23/77. The samples were used as received without further puri- cation. At room temperature, they are either viscous liquids (HHFX), gelatinous solids (HHFG, HHFMan) or sticky powders (HHFA, HHFMalt). For comparison, an APG surfactant from Seppic (Simulsol SL26C) was used. It is a C10–16-polyglucoside, with a polymeri- sation degree of 1.3 and an active content of 50.7 % wt. The pure n-alkyl-b-D-glucosides (C8, C10 and C12) and n-alkyl-b-D-malto- sides (C10, C12, C14 and C16) used to characterize the HHF lipophilic tail with the PIT-slope method were provided by Sigma Aldrich with a high degree of purity ($99%). The C10E4 reference surfactant used in the PIT-slope method was synthe- tized and distilled in the lab to obtain an ultra-pure sample.25 This journal is © The Royal Society of Chemistry 2020 3.1. Amphiphilic properties in aqueous systems 3.1.1. Solubility. Alkyl monoglycosides HHFA, HHFX, HHFG and HHFMan are not soluble in water at 0.01% wt, from room temperature to 83 C. The 1% and 0.1% samples clearly show the presence of solid surfactant, which means that this is not a clouding phenomenon. Increasing temperature up to 83 C has no eﬀect on the solubility, which tends to indicate that the low solubility is due to the existence of a high Kra temperature. High Kratemperatures have been reported previously for alkyl-monosaccharides with alkyl chains longer than 12 carbons.30 To clearly identify the amphiphilicity of the surfactant under study, the variation of the PIT must be represented as a function of its molar fraction compared to the C10E4 reference surfactant as detailed below. The error on PIT-slope characterization is less than 10% in average.27 The diglycoside HHFMalt is soluble in water at 0.1% wt, from room temperature to 83 C. This means that no clouding phenomenon occurs in this temperature range. On the other hand, a 0.1% wt solution leduring 3 days at 4 C does not precipitate, which means that no Kratemperature exists down to 4 C due to an increased hydrogen bonding between water and disaccharide headgroups.31 However, this is also a direct eﬀect of the branched HHF chain which brings disorder into crystalline phases of surfactants and allows it to be soluble even at low temperatures. This property is a real advantage of the branched surfactants compared to their linear counterparts. In comparison, b-C10Malt has a Krapoint of 26 C and b- C14Malt of 32 C.32–34 Both surfactants are not soluble at room temperature whereas HHFMalt is. At 1% wt, a phase separation between a viscous phase (liquid crystal) and a clear aqueous phase (L1) is observed. Molar fraction x ¼ moles of added surfactant moles of C10E4 þ moles of added surfactant (1) (1) (1) 2.4. PIT-slope characterization The amphiphilicity of the diﬀerent sugar-based surfactants was determined according to the PIT-slope method already published.26,27 The Phase Inversion Temperature (PIT) can be precisely determined and expresses the amphiphilicity of the surfactant in its environment, including the contributions of oil hydropho- bicity and water salinity. However, the PIT cannot always be determined directly, that is why, in order to characterize all types of surfactants, it is more convenient to measure the PIT variation induced by adding increasing amounts of the surfactant under study to a well-dened SOW reference system. The reference system used is 3%C10E4/n-octane/102 M NaClaq because it shows a reproducible PIT close to room temperature (23.9 C).28 Open Access Article. Published on 24 April 2020. Downloaded on 12/15/2023 8:36:02 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. Interfacial tensions were measured at solubility limit for HHFA, HHFX, HHFG, and HHFMan, and at 0.01% wt for HHFMalt against n-octane (Acros Organics, >95%). In this study, n-octane was chosen as oil to perform a salinity scan by varying the NaCl concentration with various systems composed of 1.8% surfactant and 2.7% 1-BuOH as co-surfactant at 50 C and water–oil ratio (WOR) equal to 1. The surfactants were dissolved in 1-BuOH and diluted in deionized water. Once all components were introduced into the pipette, they were sealed under a nitrogen ow and placed at 50 C for 10 minutes. They were then mixed three times at rst and again aer 1 h, 2 h and one night of equilibration at 50 C. Observations of phase behavior were made aer complete stabilization. The supercial and interfacial tensions are given with an error of 0.5 mN m1. This journal is © The Royal Society of Chemistry 2020 Molar fraction x ¼ moles of added surfactant moles of C10E4 þ moles of added surfactant (1) 2.3. Supercial and interfacial tension Supercial and interfacial tensions were measured along time by drop shape analysis in the rising drop mode on a Tracker instrument from Teclis®. Measurements were performed at room temperature (20 2) C. The bubble/drop volume was maintained constant during the measurement and was set between 5 and 8 mL. Supercial tensions were measured at various concentrations for each sample. The surface tension at This journal is © The Royal Society of Chemistry 2020 RSC Adv., 2020, 10, 16377–16389 \| 16379 Paper View Article Online View Article Online View Article Online RSC Advances Paper equilibrium recorded for each solution allowed plotting g ¼ f(C) to workout CMC values. For HHFMalt, 0.01% wt, 0.005% wt and 0.0025% wt stock solutions were prepared. First, the surface tension of deionized water was measured, and then the concentration of the solution was gradually increased by dilu- tion with stock solutions. For HHFA, HHFX, HHFG and HHFMan, the same procedure was followed. The stock solu- tions were ltered on 0.2 mm lter. For these samples, fewer points were measured because of long equilibration times and of some instability during measurements. Therefore, only CMC ranges are given. The so-called “Optimum Formulation” is when the surfac- tant aﬃnities for oil and water are the same (R ¼ 1), corre- sponding to a Winsor III system in which the same amounts of oil and water are solubilized in the middle-phase micro- emulsion. In practice, the optimum formulation is located by performing salinity scans which varying the salinity and keeping all other parameters constant. To compare the hydrophilic/lipophilic balances of surfactants, one way is to determine their optimal salinities in a given O/W system. The higher the optimal salinity is, the more hydrophilic the surfactant is. Paper Paper Table 3 Solubility limits of the glycosides surfactants at room temperature determined by HPLC/MS. Comparison with solubility data for the maltoside derivative from visual inspection Table 3 Solubility limits of the glycosides surfactants at room temperature determined by HPLC/MS. Comparison with solubility data for the maltoside derivative from visual inspection 3.1.2. Surface tension. The surface tension is a key parameter to characterize the aqueous behavior of surfactants. The CMC determination in case of HHFMalt is presented in Fig. 1. The CMC values can be given with accuracy only for HHFMalt for which a suﬃcient number of points could be recorded. For the other samples, because of very low solubility limits, long equilibration times and instabilities during surface tension measurements, fewer points were recorded and only concentration ranges are given. The data are gathered in Table 4. Surfactant Polar head Solubility limit mmol L1 ppm HHFA Arabinose 0.0273 9.86 HHFX Xylose 0.0865 31.18 HHFG Glucose 0.0336 13.11 HHFMan Mannose 0.0197 7.68 HHFMalt Maltose >1.81a >1000a a Formation of liquid crystal phase at 1 wt%. Open Access Article. Published on 24 April 2020. Downloaded on 12/15/2023 8:36:02 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. All samples reduce signicantly the water surface tension. This means that all surfactants form micelles at concentrations below the solubility limit. It is diﬃcult to discuss diﬀerences in hydrophilicity for the various polar heads, because of the very low solubilities and uncertainties in CMC determination, except for HHFMalt which is the most hydrophilic surfactant in this series. Nevertheless, the results show that its CMC is close to some of the monoglycosides. This is consistent with the fact that within a homogeneous series, the CMC value is most impacted by the lipophilic chain length, which is the same for all surfactants under study.42 Some literature data on other sugar-based surfactants are collected in Table 5 for comparison. For these surfactants, the hydrophobic chains are linear and reliable CMC values are usually given for alkyl chains up to C12 because of insolubility issues aerwards.31 The data marked with a footnote should be taken with caution because of non- soluble samples. It is interesting to notice the important The solubility limits of the monoglycosides at room temperature are given in Table 3 and were determined by HPLC/ MS. 2.5. Ternary phase diagram behavior Alternatively, a qualitative tool to estimate the surfactant aﬃnity for oil or water in any Surfactant–Oil–Water (SOW) system was proposed by Winsor29 with the so-called R-ratio. R > 1 (respec- tively R < 1) indicates a surfactant aﬃnity for oil (respectively water), R ¼ 1 when the surfactant aﬃnity is the same for oil and water. The diﬀerent values of R correspond to diﬀerent equi- librium behaviors for the SOW systems, named Winsor I (O/W microemulsion in equilibrium with excess oil), Winsor II (W/ O microemulsion in equilibrium with excess water) and WIII (bicontinuous microemulsion with excess oil and water). The R- ratio was later extended to take into account all interactions occurring at the O/W interface20 that drive the aﬃnity towards water or oil. The volume of the liquid crystalline phase does not seem to vary when the temperature changes from room temperature to 83 C, which tends to indicate a very low temperature-sensitivity of the aqueous phase behavior. This observation is in accor- dance with previous studies on this class of surfactants.35–37 16380 \| RSC Adv., 2020, 10, 16377–16389 This journal is © The Royal Society of Chemistry 2020 RSC Advances View Article Online RSC Advances View Article Online Paper There are approximately 2 orders of magnitude diﬀerence in solubility between the maltoside derivative and the other glycosides. The solubility limits of all the monoglycosides are in the same range and depend on conformation and hydrogen bonding eﬀects. The mannose derivative seems to be the least soluble sample. The solubility has also been investigated in 25 wt% NaCl solution. No sample is soluble, down to 0.01% wt, from room temperature to 83 C. The results obtained are in agreement with the salting-out eﬀect of NaCl already highlighted in case of octyl-glucoside38–40 and dodecyl-maltoside.41 Fig. 1 (Left) Dynamic surface tension measurements as function of time for various concentrations of HHFMalt in aqueous media; (right) CMC determination for HHFMalt surfactant. Fig. 1 (Left) Dynamic surface tension measurements as function of time for various concentrations of HHFMalt in aqueous media; (right) CMC determination for HHFMalt surfactant. Fig. 1 (Left) Dynamic surface tension measurements as function of time for various concentrations of HHFMa determination for HHFMalt surfactant. Table 4 Critical micellar concentrations (CMC) and surface tensions at CMC (gCMC) determined for the diﬀerent surfactants studied by drop shape tensiometry Surfactant Polar head CMC gCMC, mN m1 mmol L1 ppm HHFA Arabinose (C5) 0.0037–0.0060 1.34–2.15 27.3 HHFX Xylose (C5) 0.0101–0.0304 3.65–10.98 28.6 HHFG Glucose (C6) 0.0092–0.0336 3.58–13.11 27.1 HHFMan Mannose (C6) 0.0041–0.0196 1.61–7.68 27.2 HHFMalt Maltose (C6–C6) 0.0298 16.49 28.3 This journal is © The Royal Society of Chemistry 2020 RSC Adv., 2020, 10, 16377–16389 \| 16381 Table 4 Critical micellar concentrations (CMC) and surface tensions at CMC (gCMC) determined for the diﬀe shape tensiometry View Article Online RSC Advances RSC Advances Table 5 Literature data on critical micellar concentrations (CMC) and surface tensions at CMC (gCMC) for various sugar-based surfactants Polar head Surfactant Chain length CMC (mmol L1) gCMC (mN m1) Reference Xylose (C5) 5-O-Xylose monoester (a, b isomers) C12 0.041 28.9 43 C14 a 0.015 36.0 43 C16 a 0.022 41.0 43 Glucose (C6) b-D-Glucoside C8 25.0 30.1 44 C10 2.2 27.7 44 C12 0.19 39.4 44 C12 0.13 33.1 43 a-D-Glucoside C12 a 0.042 41.2 43 Maltose (C6–C6) b-D-Maltoside C8 23.2 — 34 C9 6.50 — 34 C10 1.95 36.7 45 C10 2.19 — 34 C12 0.19 — 34 C12 0.15 35.3 45 C12 0.13 33.1 43 C14 0.0167 — 34 C16 0.00145 — 34 a-D-Maltoside C12 0.12 35.9 43 a Turbid solutions close and above CMC. ownloaded on 12/15/2023 8:36:02 PM. Paper Commons Attribution-NonCommercial 3.0 Unported Licence. ns at CMC (gCMC) for various sugar-based surfactants a Turbid solutions close and above CMC. anomeric eﬀect for dodecylglucoside a and b, which does not occur for dodecylmaltoside. Overall, the CMC value found for HHFMalt (branched C15) lies within the ones reported for the C14 and C16 derivatives. It is important to stress that the surface tension gCMC (mN m1) is lower for the farnesylglycosides compared to long chain linear n-alkyl-glucosides, which should indicate a better wetting ability. packing of branched alkyl chains into spherical or cylindrical micelles, which was shown by experimental measurements48,49 and model predictions.50 This eﬀect on surfactant packing was found to be eﬀective even for branching on the rst carbons of the alkyl chains near the sugar polar head, which is in accor- dance with our results.51–53 The dynamic tension proles on Fig. 1-leshow an expected evolution with the concentration of surfactant. For the most diluted solutions (0.0018 mM to 0.0145 mM), the induction time is quite high and it decreases when the concentration increases. For concentrations close to or larger than the CMC, the surface tension at 0.06s is already considerably lower than the initial value (g0) of 72.8 mN m1. The dynamic surface tension decay can be also exploited to approximate the self- diﬀusion coeﬃcient of surfactants in water. Indeed, the diﬀu- sion controlled adsorption G(t) of surface active species at air/ liquid interface is described by the Ward and Tordai equa- tion.54 For non-ionic surfactants, the decrease of the surface tension as a function of time g(t) has been evaluated from the classical Gibbs absorption equation48,55,56 by two diﬀerent asymptotic models for short and long times.57–59 In order to quantify the lipophilic contribution of the HHF carbon chain on the CMC, a linear correlation between the logarithm of the CMC with the linear alkyl chain length of n- alkyl-b-D-maltosides was made following the empirical equation of Klevens:46,47 log(CMC) ¼ A BN (2) (2) Using this equation and the CMC of HHFMalt (0.0298 mmol L1), the value of a linear equivalent chain length of 13.5 carbons can be estimated for hexahydrofarnesyl (Fig. 2). Branching tends to increase the CMC due to unfavorable Fig. 2 Correlation between the logarithm of the CMC and the alkyl chain length for n-alkyl-b-D-maltosides. Open Access Article. Published on 24 April 2020. Downloaded on 12/15/2023 8:36:02 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. gðtÞ ¼ geq þ g0 geq 1 þ t t n (5) expected because of less dense packing due to the branching. By modifying the diﬀusion coeﬃcient parameter, it is possible to t the model with experimental data of dynamic surface tension g(t). By changing the surfactant bulk concentration, the model should suit the two previous experimental data sets, with slight changes in the diﬀusion coeﬃcient as it appears to decrease slowly when surfactant concentration increases in case of n- octyl-b-D-glucoside.61 The trends obtained in case of HHFMalt in water at 0.0072 and 0.0145 mmol L1 conrm the previous statements and approximate the diﬀusion coeﬃcient at 2.9 1010 m2 s1 for a bulk concentration of 0.0072 mmol L1 and 1.9 1010 m2 s1 for 0.0145 mmol L1 (Fig. 21 in ESI†). (5) where t* and n are empirical constants which have the unity of time for t* whereas n is a dimensionless number. The param- eters t* and n can be also used to infer some properties of the surfactant under study such as the diﬀusion coeﬃcient.72–75 By evaluating the surface tension at the induction time ti dened by Rosen et al.65 using this empirical equation, the surfactant solution can be treated as dilute and a linear Henry isotherm can be applied in conjunction with the short time approxima- tion of the Ward and Tordai equation to deduce the diﬀusion coeﬃcient having the surfactant concentration c0 and surface tension at boundaries g0 and geq as follows: The values obtained are comparable to the ones found in literature for the same class of monomeric surfactant62 and logically higher than corresponding micelles for which a value of 7 1011 m2 s1 has been calculated from NMR experiments with n-dodecyl-b-D-maltoside in D2O solutions.63,64 gðtiÞ g0 ¼ geq gðtiÞ ti t* n ¼ RTGi with Gi ¼ GðtiÞ ¼ 2c0 ﬃﬃﬃﬃﬃﬃﬃ Dti p r (6) (6) Dynamic surface tension can be also tted by an empirical equation according to Rosen model.65–71 Rosen et al. divided the dynamic surface tension curves into four diﬀerent regions: induction, rapid fall, meso-equilibrium and equilibrium toward surface tension. The model ts the experimental data well for the most common dynamic surface tension curves. Paper gðtÞ ¼ g0 2RTC0 ﬃﬃﬃﬃﬃﬃ Dt p r for short times (3) gðtÞ ¼ g0 2RTC0 ﬃﬃﬃﬃﬃﬃ Dt p r for short times (3) gðtÞ ¼ geq þ Geq 2RT 2C0 ﬃﬃﬃﬃﬃﬃﬃﬃﬃ pDt p for long times (4) (3) gðtÞ ¼ geq þ Geq 2RT 2C0 ﬃﬃﬃﬃﬃﬃﬃﬃﬃ pDt p for long times (4) (4) where R is the perfect gas constant (8.314 J mol1 K1), T the temperature, C0 the bulk surfactant concentration, t the time, D the diﬀusion coeﬃcient of the surfactant and Geq the equili- brated surface excess. Geq has been evaluated to 2.6 106 mol m2 from Fig. 1-right and it is comparable but slightly lower than the values obtained for n-dodecyl-b-D-maltoside (3.32 106 mol m2) referenced in the literature,60 which was where R is the perfect gas constant (8.314 J mol1 K1), T the temperature, C0 the bulk surfactant concentration, t the time, D the diﬀusion coeﬃcient of the surfactant and Geq the equili- brated surface excess. Geq has been evaluated to 2.6 106 mol m2 from Fig. 1-right and it is comparable but slightly lower than the values obtained for n-dodecyl-b-D-maltoside (3.32 106 mol m2) referenced in the literature,60 which was Fig. 2 Correlation between the logarithm of the CMC and the alkyl chain length for n-alkyl-b-D-maltosides. 16382 \| RSC Adv., 2020, 10, 16377–16389 This journal is © The Royal Society of Chemistry 2020 This journal is © The Royal Society of Chemistry 2020 Paper Table 6 Values of diﬀusion coeﬃcients obtained from ﬁtting dynamic surface tension curves with the Rosen model for various HHFMalt concentrations C0 (mmol L1) ti (s) t* (s) n geq (mN m1) g(ti) (mN m1) Gi (106 mol m2) D (1010 m2 s1) 0.0145 25.6 43.8 2.45 32.1 63.2 1.08 1.7 0.0072 120.4 167.0 4.02 36.3 64.3 0.96 1.2 0.0036 180.5 335.3 2.13 41.4 65.5 0.83 2.3 0.0018 433.0 752.6 2.38 47.0 66.4 0.73 3.0 Open Access Article. Published on 24 April 2020. Downloaded on 12/15/2023 8:36:02 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. For HHFX and HHFMan, the drop comes oﬀnaturally before reaching equilibrium, indicating that interfacial tensions are lower than 1 mN m1 which is the limit of the instrument. For HHFMalt at 0.01% wt, no measurement was possible because of drop coming oﬀright from the beginning of the experiment. Table 7 recaps the values of interfacial tensions vs. n-octane. To compare all the glycosides derivatives, a value at 100 s is given, as an indication of the individual diﬀusion of surfactants at the interface. Fig. 4 Dynamic interfacial tensions vs. n-octane for the diﬀerent surfactants under study except HHFMalt (drops come oﬀbefore equilibrium for HHFX and HHFMan at solubility limit). Table 7 Interfacial tensions water/n-octane gOW at diﬀerent times Surfactant Polar head Concentration gow (mN m1) at 100 s at tmax HHFA Arabinose (C5) Solubility limit 21.9 7.6a (4000 s) HHFX Xylose (C5) Solubility limit 21.2 5.6a (1500 s) HHFG Glucose (C6) Solubility limit 7.3 6.1 (4000 s) HHFMan Mannose (C6) Solubility limit 4.2 2.9a (170 s) HHFMalt Maltose (C6–C6) 0.01% wt <1 <1 a Non stabilized values. Table 7 Interfacial tensions water/n-octane gOW at diﬀerent times 3.2.2. PIT-slope characterization. To measure their hydro- philicities, the PIT-slope values of the diﬀerent sugar-based surfactants have been determined. The gradual addition of a second surfactant “S2” increases or decreases the PIT of the reference system, allowing the evaluation of its true hydro- philic–lipophilic balance and its ranking against a series of well- dened nonionic and ionic surfactants. When the PIT increases, S2 is more hydrophilic than C10E4, and vice versa. The PIT varies linearly as a function of the added surfactant S2, expressed as a weight percentage or a molar fraction. The slope of the PIT vs. the concentration of the added surfactant is denoted dPIT/dC or dPIT/dx2 depending on whether the weight or the molar fraction is used for the X axis. The results are displayed in Fig. 5 and Table 8. Open Access Article. Published on 24 April 2020. Downloaded on 12/15/2023 8:36:02 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. Considering the surface tension of the solvent g0 and at equilibrium geq obtained from experimental data, the surface tension at each time g(t) can be deduced from the following equation: The results obtained highlight a good tting of the Rosen model with experimental dynamic surface tension curves for various HHFMalt concentrations (Fig. 3 and 22 in ESI†) as well as exhibiting diﬀusion coeﬃcients in the same range as the one determined from short and long times approximation models (Table 6). Fig. 3 Experimental dynamic surface tension curves versus Rosen model for 0.0072 and 0.0145 mmol L1 HHFMalt concentrations below the CMC. t* and n are obtained by minimizing the global error between experimental data with the Rosen model. The results for 0.0018 and 0.0036 mmol L1 HHFMalt concentrations are displayed in Fig. 22 in ESI.† Fig. 3 Experimental dynamic surface tension curves versus Rosen model for 0.0072 and 0.0145 mmol L1 HHFMalt concentrations below the CMC. t* and n are obtained by minimizing the global error between experimental data with the Rosen model. The results for 0.0018 and 0.0036 mmol L1 HHFMalt concentrations are displayed in Fig. 22 in ESI.† This journal is © The Royal Society of Chemistry 2020 RSC Adv., 2020, 10, 16377–16389 \| 16383 View Article Online RSC Advances Fig. 4 Dynamic interfacial tensions vs. n-octane for the diﬀerent surfactants under study except HHFMalt (drops come oﬀbefore equilibrium for HHFX and HHFMan at solubility limit). Fig. 4 Dynamic interfacial tensions vs. n-octane for the diﬀerent surfactants under study except HHFMalt (drops come oﬀbefore equilibrium for HHFX and HHFMan at solubility limit). Table 8 PIT-slope values of the diﬀerent sugar-based surfactants under study Table 8 PIT-slope values of the diﬀerent sugar-based surfactants under study Surfactant Polar head dPIT/dC (C/wt%) dPIT/dx (C) HHFA Arabinose (C5) 12.6 55.4 HHFX Xylose (C5) 15.3 68.8 HHFG Glucose (C6) 4.5 21.8 HHFMan Mannose (C6) 7.2 34.1 HHFMalt Maltose (C6–C6) 7.5 46.2 Open Access Article. Published on 24 April 2020. Downloaded on 12/15/2023 8:36:02 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. This journal is © The Royal Society of Chemistry 2020 3.2. Amphiphilic behavior in surfactant oil water (SOW) systems The same considerations hold for HHFG, which is more hydrophilic than HHFMan due to hydroxyl group 2 in equatorial position. Linear t from Fig. 7 can be used to estimate the “equivalent linear chain length” of the HHF tail. In both cases, the PIT-slope values of the HHF surfactants are lower than those measured for the linear corresponding surfactants and the data was extrapolated. Using n-alkyl-glucosides correlation, the approxi- mate value is 15.7 carbons, whereas it is 18.8 when the n-alkyl- maltosides results are used. These values should not be compared to the one obtained from CMC experiments, because in the PIT-slope experiments, the alkyl chain interacts with the oil phase. In the same vein, previous studies have highlighted the hydrophobic inuence of the ramication of the surfactant tail, which decreases the optimum salinity of SOW system.79 The fact that the PIT-slope correlation gives diﬀerent values of equivalent surfactant chain length in case of glucosides and maltosides can be linked to possible diﬀerences of partition coeﬃcients. In order to compare the HHF surfactants with linear surfactants and to establish the eﬀect of branching, the PIT- slope values of well-dened n-alkyl-b-D-glucosides and n-alkyl- b-D-maltosides were quantied. Fig. 6 shows the PIT-slope of n- alkyl-glucosides and n-alkyl-maltosides. As expected, the PIT-slope values obtained for n-alkyl- glucoside and maltoside surfactants are in accordance with Determining the PIT-slope values of these surfactants allows classifying them among other conventional surfactants already characterized by the method and especially the CiEj surfac- tants.26,27,80,81 Results displayed in Fig. 8 highlight that HHFMalt has a hydrophilicity between C12E6 and C12E7, whereas the HHFG is between the C12E3 and C12E4. This gure stresses again that the type of sugar used as polar head has a huge inuence on the hydrophilicity. Indeed, the HHFMan is found to be equivalent to the C12E2. Fig. 7 Evolution of the PIT-slope values with the number of carbons of the surfactant tail for n-alkyl-b-D-glucoside and n-alkyl-b-D-malto- side surfactants. In the same way, it is possible to determine the equivalent number of ethoxylated group to one glucoside or maltoside unit by comparison of the linear alkyl b-glucosides and maltosides to the CiEj surfactants. The trend of the PIT-slope values of C12Ej with the number of ethoxylated groups can be described by a second degree polynomial relation (Fig. 9). 3.2. Amphiphilic behavior in surfactant oil water (SOW) systems 3.2.1. Interfacial tension. The interfacial tension of the surfactant/octane/water system was studied. All surfactants reduce the n-octane/water interfacial tension as displayed by Fig. 4. The expected IFTn-octane/water value should be around 50 mN m1,76,77 which is what is measured at t ¼ 0 s, but the equilibrium value reaches 23.2 mN m1 because of the presence of impurities in the oil (95%). HHFMalt is the most hydrophilic surfactant because of its disaccharide head. This result conrms previous CMC and solubility experiments for which the HHFMalt surfactant has been dened as the most hydrophilic. Among the mono- saccharide derivatives, the ones derived from hexoses (HHFG Fig. 5 (Left) PIT variation of the reference system as function of the sugar-based surfactant concentration; (right) PIT variation of the reference system as function of the sugar-based surfactant molar fraction. Fig. 5 (Left) PIT variation of the reference system as function of the sugar-based surfactant concentration; (right) PIT variation of the reference system as function of the sugar-based surfactant molar fraction. This journal is © The Royal Society of Chemistry 2020 16384 \| RSC Adv., 2020, 10, 16377–16389 RSC Advances View Article Online RSC Advances View Article Online Paper Fig. 6 PIT variation of the reference system as function of the n-alkyl-b-D-glucoside (left) and the n-alkyl-b-D-maltoside (right) surfactants molar fraction. ution-NonCommercial 3.0 Unported Licence. Open Access Article. Published on 24 April 2020. Downloaded on 12/15/2023 8:36:02 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. Fig. 6 PIT variation of the reference system as function of the n-alkyl-b-D-glucoside (left) and the n-alkyl-b-D-maltoside (right) surfactants molar fraction. their linear carbon tail lengths. The increase of the length of alkyl chain from 8 to 12 diminishes the hydrophilicity of the surfactants and the PIT-slope decreases linearly, as shown in Fig. 7. and HHFMan) are more hydrophilic than the ones derived from pentoses (HHFA and HHFX). Glucose appears to be more hydrophilic than mannose and arabinose more hydrophilic than xylose, which shows again that the sugar conformation has a great inuence on its polarity. Indeed, the more isolated hydroxyl groups are, the higher the hydrophilicity of the surfactant is, due to favorable hydrogen bonding with water.78 It is the case for HHFA, which is more hydrophilic than HHFX due to hydroxyl groups 3 and 4 in equatorial and axial positions, compared to two equatorial positions for xylose. This journal is © The Royal Society of Chemistry 2020 linked to the HLD equation, and so to the PACN, by the relation below: dPIT dx ¼ ðPACN1 PACN2Þs1 ðs1 þ xðs2 s1ÞÞ2 /x/0 PACN1 PACN2 s1 with s ¼ ct k and PACN ¼ s k (7) (7) where PACN1 and s1 are the parameters of the normalized HLD for the reference surfactant C10E4. x, PACN2 at s2 are respectively the molar fraction and the parameters of the normalized HLD for the added surfactant. Fig. 8 The PIT-slope scale used for the classiﬁcation of CiEj and sugar-based surfactants. As the parameters of the reference surfactant PACN1 and s1 are known (PACN1 ¼ 8.2 and s1 ¼ 0.31 C1),85 the diﬀerent PACN of sugar-based surfactants can be calculated (Table 9). Values of PACN are inversely proportional to the PIT-slope values. The negative values of PACN for a surfactant indicate that a polar oil is needed to get a three phase behavior at 25 C without salt and alcohol. The high values of PACN indicate a lipophilic surfactant that needs very long alkanes to get a WIII behavior. Liquid alkanes at 25 C range from pentane to hex- adecane, and only HHFG reaches theoretically the optimum formulation with an alkane (pentadecane) at 25 C without salt. The diﬀerent PIT-slope values of sugar-based surfactants can be also useful to determine their Preferred Alkane Carbon Number (PACN), which is the linear alkane that gives a WIII microemulsion system with the surfactant at 25 C without salt or alcohol. The PACN is also a parameter of the normalized HLD equation, dened by Salager et al., which describes the phase behavior of SOW systems to reach the optimum formula- tion.82–84 As described in the literature,85 the PIT-slope can be 3.2.3. Salinity scans. In order to study the behavior of HHF surfactants in oil/water systems towards salinity, equilibrated scans have been performed at 50 C using n-octane as model oil and adding 1-butanol as cosurfactant (2.7% wt). The phase behaviors have been observed at equilibrium and are shown in the ESI (Fig. 23–28†). 3.2.3. Salinity scans. In order to study the behavior of HHF surfactants in oil/water systems towards salinity, equilibrated scans have been performed at 50 C using n-octane as model oil and adding 1-butanol as cosurfactant (2.7% wt). The phase behaviors have been observed at equilibrium and are shown in the ESI (Fig. 23–28†). Fig. 9 Evolution of the PIT-slope values of C12Ej with the number of ethoxylated groups of the surfactant. 3.2. Amphiphilic behavior in surfactant oil water (SOW) systems Using this corre- lation with nC12-b-glucoside and maltoside, the hydrophilicity contribution of one glucoside unit is equivalent to 5.9 ethoxy- lated groups, whereas a maltoside unit is equivalent to 8.3 ethoxylated groups. Fig. 7 Evolution of the PIT-slope values with the number of carbons of the surfactant tail for n-alkyl-b-D-glucoside and n-alkyl-b-D-malto- side surfactants. Fig. 7 Evolution of the PIT-slope values with the number of carbons of the surfactant tail for n-alkyl-b-D-glucoside and n-alkyl-b-D-malto- side surfactants. This journal is © The Royal Society of Chemistry 2020 RSC Adv., 2020, 10, 16377–16389 \| 16385 View Article Online RSC Advances Fig. 8 The PIT-slope scale used for the classiﬁcation of CiEj and sugar-based surfactants. e. Published on 24 April 2020. Downloaded on 12/15/2023 8:36:02 PM. ticle is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. Table 9 PIT-slope values and the corresponding PACN of sugar- based surfactants under study Table 9 PIT-slope values and the corresponding PACN of sugar- based surfactants under study Surfactant name dPIT/dx (C) PACN HHFA 55.4 25.4 HHFX 68.8 29.5 HHFG 21.8 15.0 HHFMan 34.1 18.8 HHFMalt 46.2 6.1 nC8-b-Glu 77.8 15.9 nC10-b-Glu 47.8 6.6 nC12-b-Glu 27.0 0.8 nC10-b-Malt 123.0 29.9 nC12-b-Malt 112.2 26.6 nC14-b-Malt 80.7 16.8 nC16-b-Malt 74.3 14.8 Open Access Article. Published on 24 April 2020. Downloaded on 12/15/2023 8:36:02 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. linked to the HLD equation, and so to the PACN, by the relation below: linked to the HLD equation, and so to the PACN, by the relation below: 4 Conclusions 3 D. Balzer and H. Luders, Nonionic surfactants: alkyl polyglucosides, Taylor & Francis, 2000. Five novel sugar surfactants prepared from bio-based hexahy- drofarnesol have been evaluated as regards to their water solubility, surface and interfacial tension, amphiphilicity and ternary phase behavior in n-octane/water systems. Only the disaccharide (maltose) derivative is signicantly soluble (>0.1 wt%) in deionized water whereas the water solubility of monosaccharide derivatives is lower than 50 ppm which might be a limitation for certain industrial applications. This study reveals that the hexahydrofarnesyl chain must be coupled with di- or poly-saccharides to achieve a high water solubility and to provide surfactants with well-balanced amphiphilicity. 4 C. Baron and T. E. Thompson, Biochim. Biophys. Acta, Biomembr., 1975, 382, 276–285. 5 A. Kiyoshi, F. Tadaaki, H. Kunizo, O. Hiroshi and T. Shinji, Process for producing alkyl glycoside, JP Pat., 04308599, 1992. 6 H. Luders, Process for the preparation of alkyloligoglycosides, US Pat., 4866165A, 1989. 7 M. Biermann, K. Hill, W. Wuest, R. Eskuchen, J. Wollmann, 7 M. Biermann, K. Hill, W. Wuest, R. Eskuchen, J. Wollmann, A. Bruns, G. Hellmann, K.-H. Ott, W. Winkle and K. Wollmann, Process for the production of surface active alkyl glycosides, US Pat., 5374716A, 1990. Despite their poor water-solubility, all sugar-based surfac- tants strongly reduce the surface tension of water even at very low concentrations (<20 ppm). In addition, the PIT-slope measurements allowed assessing their true amphiphilicities which are consistent with the known hydrophilicity of the sugar heads. As expected, the HHFMalt surfactant was found to be the most hydrophilic in this series. The PIT-slope of this surfactant is in agreement with its phase behavior in the HHFMalt/n- octane/water ternary system. The equivalent linear carbon chain of the HHF tail, as well as the equivalent ethoxylated group number of sugar head, have also been approximated by this method. 8 H.-D. Doerer, Angew. Chem., Int. Ed., 1997, 36, 2527. 9 O. J. Rojas, C. Stubenrauch, L. A. Lucia and Y. Habibi, Interfacial properties of sugar-based surfactants, AOCS Press, 2009, pp. 457–480. 10 M. K. Matsson, B. Kronberg and P. M. Claesson, Langmuir, 2004, 20, 4051–4058. 11 G. Bognolo, Techniques de l’ing´enieur, 2013, 2, J2265. 12 G. Dass, R. Tyagi and N. Kumar, J. Surf. Sci. Technol., 2015, 31, 156–163. 13 W. von Rybinski and K. Hill, Angew. Chem., Int. Ed., 1998, 37, 1328–1345. Conﬂicts of interest The authors declare that they have no conict of interest. 18 Y. Queneau, S. Chambert, C. Besset and R. Cheaib, Carbohydr. Res., 2008, 343, 1999–2009. References 1 H. T. Boehme AG, A process for the production of sulphuric acid esters of glucosides, Patent GB, 384230, 1932. 2 H. T. Boehme AG, Improvements in or relating to liquids or plastic preparations for the treatment of brous materials, Patent GB, 393769, 1933. 4 Conclusions In n-octane/water with 1-butanol as co-surfactant, HHFMalt forms Winsor III-type systems over a wide salinity range, whereas the monosaccharides form only Winsor II-type systems whatever the salinity. Thanks to this particular behavior, HHFMalt can be well suited for some enhanced oil recovery processes. 14 Z. Jeirani, B. Mohamed Jan, B. Si Ali, I. M. Noor, C. H. See and W. Saphanuchart, Ind. Crops Prod., 2013, 43, 6–14. 15 M. Santa, G. Alvarez-J¨urgenson, S. Busch, P. Birnbrich, C. Spindler and G. Brodt, in SPE Enhanced Oil Recovery Conference, Society of Petroleum Engineers, Kuala Lumpur, Malaysia, 2011. In the same way, to valorize monosaccharide derivatives, it would also be worthwhile to combine them with more water soluble surfactants (ionic ones or shorter alkyl chains for instance) to investigate possible synergistic formulations. 16 S. Iglauer, Y. Wu, P. Shuler, Y. Tang and W. A. Goddard, Colloids Surf., A, 2009, 339, 48–59. 17 T. Gaudin, H. Lu, G. Fayet, A. Berthauld-Drelich, P. Rotureau, G. Pourceau, A. Wadouachi, E. Van Hecke, A. Nesterenko and I. Pezron, Adv. Colloid Interface Sci., 2019, 270, 87–100. This journal is © The Royal Society of Chemistry 2020 linked to the HLD equation, and so to the PACN, by the relation below: All monosaccharide surfactants form Winsor II systems at 50 C, whatever the salinity, which means that the aﬃnity of these surfactants is higher for the oil phase, even if the aqueous phase does not contain electrolyte. For instance, the most lipophilic compound, HHFX, is soluble neither in water nor in oil, and oil droplets are observed on the pipette wall (Fig. 23 in ESI†). This is consistent with the negative PIT-slope values and the very low water solubility of these compounds. HHFMalt, a hydrophilic surfactant following the PIT-slope criteria, changes from a Winsor I system at 0% wt NaCl to a Winsor III system from 5 to 15% wt NaCl. Equal amounts of oil and water are solubilized in the middle-phase microemulsion at Fig. 9 Evolution of the PIT-slope values of C12Ej with the number of ethoxylated groups of the surfactant. 16386 \| RSC Adv., 2020, 10, 16377–16389 This journal is © The Royal Society of Chemistry 2020 RSC Advances View Article Online RSC Advances View Article Online Paper Paper a salinity around 12% wt NaCl, which can be identied as the “optimal salinity”. The APG SL26 (commercial C12–16Glu1.3) exhibits the same behavior using the same co-surfactant (Fig. 28 in ESI†). These results prove that sugar-based surfactant are really interesting for formulating WIII microemulsion systems along a wide range of salinity, which can be a real benet for enhanced oil recovery applications for example. This ternary phase behavior conrms also the PIT-slope, CMC and solubility measurements exposed previously by identifying HHFMalt as the most hydrophilic surfactant. performed the solubility, phase behavior, surface and interfacial tension experiments, and to Ms. Leoni Koller who performed some of the PIT-slope experiments during her Master internship. Open Access Article. Published on 24 April 2020. Downloaded on 12/15/2023 8:36:02 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unp Lipids, 2000, 104, 75–91. 67 X. Y. Hua and M. J. Rosen, J. Colloid Interface Sci., 1991, 141, 180–190. 37 H. M. von Minden, K. Brandenburg, U. Seydel, M. H. J. Koch, V. Garamus, R. Willumeit and V. Vill, Chem. Phys. Lipids, 2000, 106, 157–179. 68 M. J. Rosen, Z. H. Zhu and X. Y. Hua, J. Am. Oil Chem. Soc., 1992, 69, 30–33. 38 P. Mukerjee, J. Phys. Chem., 1965, 69, 4038–4040. 69 T. Gao and M. J. Rosen, J. Am. Oil Chem. Soc., 1994, 71, 771– 776. 39 P. Mukerjee, J. Phys. Chem., 1970, 74, 3824–3826. 70 T. Gao and M. J. Rosen, J. Colloid Interface Sci., 1995, 172, 242–248. 40 P. Mukerjee and C. C. Chan, Langmuir, 2002, 18, 5375–5381. 41 L. Zhang, P. Somasundaran and C. Maltesh, Langmuir, 1996, 12, 2371–2373. 71 M. J. Rosen and L. D. Song, J. Colloid Interface Sci., 1996, 179, 261–268. 42 Sugar-Based Surfactants: Fundamentals and Applications, ed. C. Carnero Ruiz, CRC Press, 2008. 72 L. K. Filippov, J. Colloid Interface Sci., 1994, 163, 49–60. 43 G. Garofalakis, B. S. Murray and D. B. Sarney, J. Colloid Interface Sci., 2000, 229, 391–398. 73 L. K. Filippov, J. Colloid Interface Sci., 1994, 164, 471–482. 74 L. K. Filippov and N. L. Filippova, J. Colloid Interface Sci., 1996, 178, 571–580. 44 K. Shinoda, T. Yamaguchi and R. Hori, Bull. Chem. Soc. Jpn., 1961, 34, 237–241. 75 L. K. Filippov, J. Colloid Interface Sci., 1996, 182, 330–347. 45 M. J. Rosen and S. B. Sulthana, J. Colloid Interface Sci., 2001, 239, 528–534. 76 B.-Y. Cai, J.-T. Yang and T.-M. Guo, J. Chem. Eng. Data, 1996, 41, 493–496. 46 H. B. Klevens, J. Phys. Colloid Chem., 1948, 52, 130–148. 77 T. Al-Sahhaf, A. Elkamel, A. Suttar Ahmed and A. R. Khan, Chem. Eng. Commun., 2005, 192, 667–684. 47 H. B. Klevens, J. Am. Oil Chem. Soc., 1953, 30, 74–80. 78 Z. Mosapour Kotena, R. Behjatmanesh-Ardakani and R. Hashim, Liq. Cryst., 2014, 41, 784–792. 48 M. J. Rosen and J. T. Kunjappu, Surfactants and interfacial phenomena, Wiley, Hoboken, N.J., 4th edn, 2012. 79 C. E. Hammond and E. J. Acosta, J. Surfactants Deterg., 2012, 15, 157–165. 49 J. A. Molina-Bol´ıvar and C. C. Ruiz, in Sugar-Based Surfactants, CRC Press, 2008, pp. 80–123. 50 T. Gaudin, P. Rotureau, I. Pezron and G. Fayet, Ind. Eng. Chem. Res., 2016, 55, 11716–11726. 80 G. Lemahieu, J. F. Ontiveros, V. Open Access Article. Published on 24 April 2020. Downloaded on 12/15/2023 8:36:02 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unp 25 S. Queste, J. L. Salager, R. Strey and J. M. Aubry, J. Colloid Interface Sci., 2007, 312, 98–107. 57 J. Eastoe, J. S. Dalton, P. G. A. Rogueda, E. R. Crooks, A. R. Pitt and E. A. Simister, J. Colloid Interface Sci., 1997, 188, 423–430. 26 J. F. Ontiveros, C. Pierlot, M. Catt´e, V. Molinier, J.-L. Salager and J.-M. Aubry, Colloids Surf., A, 2014, 458, 32–39. 27 J. F. Ontiveros, C. Pierlot, M. Catt´e, V. Molinier, J.-L. Salager and J.-M. Aubry, J. Colloid Interface Sci., 2015, 448, 222–230. 58 V. B. Fainerman, A. V. Makievski and R. Miller, Colloids Surf., A, 1994, 87, 61–75. 28 A. Pizzino, V. Molinier, M. Catt´e, J. F. Ontiveros, J.-L. Salager and J.-M. Aubry, Ind. Eng. Chem. Res., 2013, 52, 4527–4538. 59 C.-H. Chang and E. I. Franses, Colloids Surf., A, 1995, 100, 1– 45. 60 C. J. Drummond, G. G. Warr, F. Grieser, B. W. Ninham and 29 P. A. Winsor, Trans. Faraday Soc., 1948, 44, 376. 60 C. J. Drummond, G. G. Warr, F. Grieser, B. W. Ninham and D. F. Evans, J. Phys. Chem., 1985, 89, 2103–2109. J , , , D. F. Evans, J. Phys. Chem., 1985, 89, 2103–2109. 30 R. G. Laughlin, The Aqueous Phase Behaviour of Surfactants, Academic Press, San Diego, 1994. , J y , , , 61 F. Nilsson, O. S¨oderman and I. Johansson, Langmuir, 1996, 12, 902–908. 31 C. Stubenrauch, Curr. Opin. Colloid Interface Sci., 2001, 6, 160–170. 62 A. J. M. Valente, M. Nilsson and O. S¨oderman, J. Colloid Interface Sci., 2005, 281, 218–224. 32 C. A. Ericsson, O. S¨oderman and S. Ulvenlund, Colloid Polym. Sci., 2005, 283, 1313–1320. 63 Q. Yang, Q. Zhou and P. Somasundaran, Colloids Surf., A, 2007, 305, 22–28. 33 R. A. Salkar, H. Minamikawa and M. Hato, Chem. Phys. Lipids, 2004, 127, 65–75. 64 Q.-Q. Yang, Q. Zhou and P. Somasundaran, J. Mol. Liq., 2009, 146, 105–111. 34 W. J. De Grip and P. H. M. Bovee-Geurts, Chem. Phys. Lipids, 1979, 23, 321–335. 65 Xi Yuan and M. J. Rosen, J. Colloid Interface Sci., 1988, 124, 652–659. 35 B. J. Boyd, C. J. Drummond, I. Krodkiewska and F. Grieser, Langmuir, 2000, 16, 7359–7367. 66 M. J. Rosen and X. Y. Hua, J. Colloid Interface Sci., 1990, 139, 397–407. 36 V. Vill, H. M. von Minden, M. H. J. Koch, U. Seydel and K. Brandenburg, Chem. Phys. Acknowledgements 19 Q. Zhang, Y. Li, Y. Song, J. Li and Z. Wang, Colloids Surf., A, 2018, 538, 361–370. The authors would like to thank TOTAL for nancing this work and for allowing its publication. Chevreul Institute (FR 2638), Minist`ere de l’Enseignement Sup´erieur et de la Recherche, R´egion Nord-Pas de Calais and FEDER are also acknowledged for sup- porting this work. We are deeply grateful to M. Christian Dur who 20 M. Bourrel and C. Chambu, Soc. Pet. Eng. J., 1983, 23, 327– 338. 21 R. S. Schechter and M. Bourrel, Microemulsions and related systems: formulation, solvency, and physical properties, ed. M. Dekker, 1988. This journal is © The Royal Society of Chemistry 2020 RSC Adv., 2020, 10, 16377–16389 \| 16387 View Article Online View Article Online RSC Advances Paper Paper 52 F. Nilsson, O. S¨oderman and I. Johansson, J. Colloid Interface Sci., 1998, 203, 131–139. 22 International Renewable Energy Agency, Renewable Energy Statistics, 2017, 52. 53 S. Matsumura, K. Imai, S. Yoshikawa, K. Kawada and T. Uchibori, J. Jpn. Oil Chem. Soc., 1991, 40, 709–714. 23 N. S. Renninger and D. J. Mcphee, Fuel compositions comprising farnesane and farnesane derivatives and method of making and using same, US pat., US7399323B2, 2008. 54 A. Ward and L. Tordai, J. Chem. Phys., 1946, 14, 453–461. 55 P. C. Hiemenz and P. C. Hiemenz, Principles of colloid and surface chemistry, ed. M. Dekker, New York, 1986, vol. 188. 24 P. P. Peralta-Yahya, F. Zhang, S. B. del Cardayre and J. D. Keasling, Nature, 2012, 488, 320–328. 56 J. T. Davies and E. K. Rideal, Interfacial Phenomena, Academic Press, New York, 1st edn, 1961, pp. 343–450. This journal is © The Royal Society of Chemistry 2020 81 G. Lemahieu, J. F. Ontiveros, V. Molinier and J.-M. Aubry, J. Colloid Interface Sci., 2019, 557, 746–756. 82 J. L. Salager, J. C. Morgan, R. S. Schechter, W. H. Wade and E. Vasquez, Soc. Pet. Eng. J., 1979, 19, 107–115. 83 J.-L. Salager, A. M. Forgiarini and J. Bull´on, J. Surfactants Deterg 2013 16 449 472 84 J.-L. Salager, A. M. Forgiarini, L. M´arquez, L. Manchego and J. Bull´on, J. Surfactants Deterg., 2013, 16, 631–663. 85 J. F. Ontiveros, C. Pierlot, M. Catt´e, J.-L. Salager and J.-M. Aubry, Colloids Surf., A, 2018, 536, 30–37. 81 G. Lemahieu, J. F. Ontiveros, V. Molinier and J.-M. Aubry, J. Colloid Interface Sci., 2019, 557, 746–756. 82 J. L. Salager, J. C. Morgan, R. S. Schechter, W. H. Wade and 84 J.-L. Salager, A. M. Forgiarini, L. M´arquez, L. Manchego and J. Bull´on, J. Surfactants Deterg., 2013, 16, 631–663. 85 J. F. Ontiveros, C. Pierlot, M. Catt´e, J.-L. Salager and RSC Adv., 2020, 10, 16377–16389 \| 16389 82 J. L. Salager, J. C. Morgan, R. S. Schechter, W. H. Wade and E. Vasquez, Soc. Pet. Eng. J., 1979, 19, 107–115. 85 J. F. Ontiveros, C. Pierlot, M. Catt´e, J.-L. Salager and J.-M. Aubry, Colloids Surf., A, 2018, 536, 30–37. Open Access Article. Published on 24 April 2020. Downloaded on 12/15/2023 8:36:02 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unp Molinier and J. Aubry, IOR 2019 – 20th European Symposium on Improved Oil Recovery, Pau, France, 2019. 51 H. Minamikawa and M. Hato, Chem. Phys. Lipids, 2005, 134, 151–160. 16388 \| RSC Adv., 2020, 10, 16377–16389 This journal is © The Royal Society of Chemistry 2020 RSC Advances View Article Online RSC Advances View Article Online Paper RSC Advances 82 J. L. Salager, J. C. Morgan, R. S. Schechter, W. H. Wade and E. Vasquez, Soc. Pet. Eng. J., 1979, 19, 107–115. 83 J.-L. Salager, A. M. Forgiarini and J. Bull´on, J. Surfactants Deterg., 2013, 16, 449–472. RSC Adv., 2020, 10, 16377–16389 \| 16389
https://openalex.org/W4307518966	https://www.researchsquare.com/article/rs-1881732/latest.pdf	English	null	Thermodynamic Parameters of Phenylglycine Interaction with UO22+, La3+ and Zr4+	Research Square (Research Square)	2,022	cc-by	4,481	Thermodynamic Parameters of Phenylglycine Interaction with UO22+, La3+ and Zr4+ Farid I. El-Dossoki (  feldossoki64@sci.psu.edu.eg ) Port Said University Abd Allah Abd El-Wahab Mohamed Al-Azhar University Research Article License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Additional Declarations: No competing interests reported. Additional Declarations: No competing interests reported. Version of Record: A version of this preprint was published at BMC Chemistry on June 15th, 2023. See the published version at https://doi.org/10.1186/s13065-023-00966-7. Page 1/13 Page 1/13 Abstract The interactions between phenylglycine and some transition metal ions (UO2 2+,La3+ and Zr4+) were studied at three ionic strengths [(µ = 0.05, 0.10 and 0.15) M KNO3] and three temperatures [(25, 35 and 45) ºC] in aqueous media by using Bjerrum’s potentiometric technique. The stoichiometric and thermodynamic stability constants as well as the degree of complexation for all studied systems ( ) were determined and discussed. The standard thermodynamic parameters (Δ pKº and Δlog Kº) and the thermodynamic parameter differences (ΔGº, ΔΔGº, ΔHº, ΔΔHº, ΔSº, and ΔΔSº) were calculated for all possible reactions. Advanced discussions of these functions as well as the factors that may control the complexation processes from the thermodynamic point of view have been reported. −n The interactions between phenylglycine and some transition metal ions (UO2 2+,La3+ and Zr4+) were studied at three ionic strengths [(µ = 0.05, 0.10 and 0.15) M KNO3] and three temperatures [(25, 35 and 45) ºC] in aqueous media by using Bjerrum’s potentiometric technique. The stoichiometric and thermodynamic stability constants as well as the degree of complexation for all studied systems ( ) were determined and discussed. The standard thermodynamic parameters (Δ pKº and Δlog Kº) and the thermodynamic parameter differences (ΔGº, ΔΔGº, ΔHº, ΔΔHº, ΔSº, and ΔΔSº) were calculated for all possible reactions. Advanced discussions of these functions as well as the factors that may control the complexation processes from the thermodynamic point of view have been reported. −n Introduction Today, at this time, and because of the importance of amino acids that serve as building blocks in protein molecules and hormones, and as transitional metal ions are of great importance as they are found in all plant and animal organisms, for these reasons many authors have been interested in studying the interactions between them [1–8]. Therefore, this research represents an essential stage in the base plan of our lab [9–12], this article considers the application of a novel thermodynamic treatment to the complexing processes between phenylglycine and some transition metal ions (UO2 2+,La3+ and Zr4+). The conditions for studying these reactions were determined at three ionic strengths [(µ = 0.05, 0.10 and 0.15) M KNO3] and at three temperatures [(25, 35 and 45) OC] in aqueous solutions by performing the Bjerrum pH titration technique [13 & 14]. ]. The Bjerrum pH titration procedure requires the determination of the ionization constants (pK011 and pK012) for pure phenylglycine at the same ionic strength and temperature conditions. We have previously assigned the values of pK011 and pK012 [9]. The degree of complexation of the studied systems ( ) was determined and using this data to calculate the stoichiometric stability constants (K), and thus the thermodynamic stability constants (Kº) in addition to calculating the standard thermodynamic parameters (ΔGº,, ΔHº, and ΔSº) as well as the standard thermodynamic differences, ( ΔΔGº, ΔΔHº, and ΔΔSº). Through the discussion of these data, the factors that control the complexity processes studied from the point of view of thermodynamics have been identified. −n (B) Determination Of The Ionization Constants Of Phenylglycine: The procedure of the ionization constants determination of phenylglycine and the values of pK011 and pK012 were as reported in our previous work [9]. (a) Instrumentation: The pH measurements were carried out by means of an Orion research ionalyzer pH meter fed with a glass pH electrode.. The pH meter was calibrated by using three buffers at pH 7, 4, and 10. The pH values with accuracy of ± 0.01 pH units. The correction factor for the pH meter was calculated using the following equation [9, 10, 15]: pH = 7.00 + (pHmeter − 7.00)Cmeter (1) Accordingly, there will be two correction factors (Cmeter), the first factor in the acidic range and the other factor in the alkaline range [9, 10, 15]. (C) Determination Of The Stability Constants The procedure of the stability constants determination of phenylglycine complexes with UO2 2+, La3+ and Zr4+ was as reported in our previous work [9]. 1. Materials All the used compound in this study (phenylglycine, NaOH, La (NO3)3, Zr(NO3)4, and UO2(NO3)2) were pure chemicals from Riedel-deHäen Co. The conductivity water was distilled twice and deionized by passing through mixed resin (anion – cation exchangers). The solutions were prepared by dissolving the accurate weight of each salt in the conductivity water just before the pH measurements. Sodium phenylglycinate was prepared by the addition of equivalent amounts of pure phenylglycine and sodium Page 2/13 Page 2/13 hydroxide solution. The exact molarities of transition metal nitrates solutions were standardized by titration with EDTA [15&16]. hydroxide solution. The exact molarities of transition metal nitrates solutions were standardized by titration with EDTA [15&16]. (Ii) Calculation Of The Stability Constants Of Phenylglycine-m Complexes: (Ii) Calculation Of The Stability Constants Of Phenylglycine-m Complexes: The calculation process of the stability constants of phenylglycine complexes with UO2 2+, La3+ and Zr4+ and the suggested mechanism of interaction were as reported in previous works [9, 10, 12, 13, 15]. The determined ionization constants K011 and K012 for pure phenylglycine were used to calculate the concentrations of H2L+, HL and L− which are present in the reaction medium. Therefore, the most probable overall reaction was carried out according to the suggested mechanisms [9]. The K110 and K120 are the overall stability constants of the complexes formed from the suggested reaction (I) and K111 and K122 are the overall stability constants of the complexes formed from the suggested reaction (II) [16–19]. The observed values of are in the range 0.0< ≤1.05) indicate the formation of a 1:1 stoichiometric complexes. −n −n The standard free energy change, ΔGº, the standard enthalpy change, ΔH°, and the standard entropy change, ΔS°, were calculated using a temperature dependence method [9–12. The standard free energy change, ΔGº, the standard enthalpy change, ΔH°, and the standard entropy change, ΔS°, were calculated using a temperature dependence method [9–12. (i) Ionization constants of phenylglycine The ionization constants of phenylglycine (K011 and K012) were reported in our previous work [9] and their values are presented in Table 1. Table (1): The ionization constants values of pure phenylglycine Page 3/13 (Ii) Calculation Of The Stability Constants Of Phenylglycine-m Complexes: (Ii) Calculation Of The Stability Constants Of Phenylglycine-m Complexes: Results And Discussion Due to the richness of the collected ( ) data in this article, a chosen simple examples are summarized in Tables (2–4). −n Table (2): data for the UO2 2+ complex with phenylglycine at 25 ºC and µ = 0.15. −n Page 4/13 Page 4/13 pH [H2L+] [HL] [L−] 4.38 4.91 5.10 5.26 5.45 0.07 0.40 0.61 0.83 1.05 2.57 x 10− 4 1.28x 10− 4 9.13 x 10− 5 6.70 x 10− 5 4.54 x 10− 5 4.17 x 10− 4 7.04x 10− 4 7.78 x 10− 4 8.25 x 10− 4 8.66 x 10− 4 4.47 x 10− 9 2.56 x 10− 8 4.38 x 10− 8 6.71 x 10− 8 1.09 x 10− 7 − n Table (3): data for La3+ complex with phenylglycine at 25 ºC and µ = 0.15. pH [H2L+] [HL] [L−] 8.65 9.63 9.66 0.04 0.41 0.51 3.25 x 10− 8 3.44 x 10− 9 3.21 x 10− 9 9.82 x 10− 4 9.92 x 10− 4 9.91 x 10− 4 1.96 x 10− 4 1.89 x 10− 3 2.03 x 10− 3 −n − n Table (3): data for La3+ complex with phenylglycine at 25 ºC and µ = 0.15. −n Table (3): data for La3+ complex with phenylglycine at 25 ºC and µ = 0.15 −n Table (4): data for Zr4+ complex with phenylglycine at 25 ºC and µ = 0.15. pH [H2L+] [HL] [L−] 3.60 3.92 4.21 5.19 5.77 0.13 0.49 0.63 0.84 1.03 2.89 x 10− 4 3.20 x 10− 4 2.89 x 10− 4 7.67 x 10− 5 2.29 x 10− 5 7.78 x 10− 5 1.80 x 10− 4 3.17 x 10− 4 8.04 x 10− 4 9.14 x 10− 4 1.39 x 10− 10 6.70 x 10− 10 2.30 x 10− 9 5.57 x 10− 8 2.41 x 10− 7 −n − n able (4): data for Zr4+ complex with phenylglycine at 25 ºC and µ = 0.15. −n Inspecting these data in Tables (2–4), we observed that at low pH range (4.38 ≤ pH ≤ 4.91 for UO2 2+ and 3.60 ≤ pH ≤ 3.92 for Zr4+complexes), H2L+ has only weak interactions as it is shown from the very low values (less than 0.5). Results And Discussion At high pH values (4.91 < pH ≤ 5.45 for UO2 2+, 9.63 < pH ≤ 9.66 for La3+ and 3.92 < pH ≤ 5.77 for Zr4+), the concentration of HL is high and seem to be constant, while the concentration of −n Inspecting these data in Tables (2–4), we observed that at low pH range (4.38 ≤ pH ≤ 4.91 for UO2 2+ and 3.60 ≤ pH ≤ 3.92 for Zr4+complexes), H2L+ has only weak interactions as it is shown from the very low values (less than 0.5). At high pH values (4.91 < pH ≤ 5.45 for UO2 2+, 9.63 < pH ≤ 9.66 for La3+ and 3.92 < −n Page 5/13 Page 5/13 L− increased and H2L+ concentration decreased. In this pH region, we note that 0.5< ≤ 1.05 (i.e. 1:1 complexes were formed). −n L− increased and H2L+ concentration decreased. In this pH region, we note that 0.5< ≤ 1.05 (i.e. 1:1 complexes were formed). −n Since complex formation is observed as pH increases and insoluble precipitates are formed, then the titrations were stopped. At high pH values, no chance for complexes can be formed due to the precipitation of the hydroxide. The stoichiometric and thermodynamic stability constants are tabulated in Tables (5 and 6). Page 6/13 titrations were stopped. At high pH values, no chance for complexes can be formed due to t precipitation of the hydroxide. The stoichiometric and thermodynamic stability constants a Tables (5 and 6). Table (5): Stability constants of MZ+ - phenylglycine complexes according to the reaction sc scheme II. Metal ion µ Log K110(Scheme I) Log K111 (Scheme II) 25ºC 35ºC 45ºC 25ºC 35ºC 45ºC UO2 2+ 0.05 0.10 0.15 7.13 7.23 7.22 7.52 7.48 7.31 7.75 7.71 7.70 -2.03 -2.06 -2.01 -1.76 -1.74 -1.84 -1.52 -1.50 -1.45 La3+ 0.05 0.10 0.15 3.44 2.92 2.47 4.28 3.94 3.48 5.00 4.70 4.37 -5.94 -6.37 -6.76 -5.00 -5.27 -5.66 -4.27 -4.51 -4.78 Zr4+ 0.05 0.10 0.15 10.14 9.67 9.74 9.95 9.75 9.68 9.77 9.73 9.62 0.76 0.38 -0.09 0.67 0.43 -0.06 0.51 0.52 0.48 Table (6): Thermodynamic stability constants of MZ+- phenylglycine complexes. Results And Discussion Metal ion Log Kº110(Scheme I) Log Kº111(Scheme II) 25ºC 35ºC 45ºC 25ºC 35ºC 45ºC UO2 2+ La3+ Zr4+ 7.10 3.91 10.25 7.65 4.70 10.06 7.78 5.52 9.86 -2.05 -5.54 1.08 -1.70 -4.65 1.00 -1.56 -4.01 0.53 Table (5): Stability constants of MZ+ - phenylglycine complexes according to the reaction scheme I and scheme II. scheme II. Metal ion µ Log K110(Scheme I) Log K111 (Scheme II) 25ºC 35ºC 45ºC 25ºC 35ºC 45ºC UO2 2+ 0.05 0.10 0.15 7.13 7.23 7.22 7.52 7.48 7.31 7.75 7.71 7.70 -2.03 -2.06 -2.01 -1.76 -1.74 -1.84 -1.52 -1.50 -1.45 La3+ 0.05 0.10 0.15 3.44 2.92 2.47 4.28 3.94 3.48 5.00 4.70 4.37 -5.94 -6.37 -6.76 -5.00 -5.27 -5.66 -4.27 -4.51 -4.78 Zr4+ 0.05 0.10 0.15 10.14 9.67 9.74 9.95 9.75 9.68 9.77 9.73 9.62 0.76 0.38 -0.09 0.67 0.43 -0.06 0.51 0.52 0.48 T bl (6) Th d i bili f MZ+- phenylglycine complexes. mole) UO2 2+ 25 35 45 9.15 9.35 9.34 52.21 55.15 56.87 17.44 233.72 235.68 233.68 La3+ 25 35 45 9.45 9.35 9.53 53.92 55.14 56.81 -10.96 144.16 143.44 144.18 Zr4+ 25 35 45 9.16 9.06 9.33 52.32 53.43 56.81 14.14 233.02 219.39 223.11 º values for reaction of the metal ion with L−(Scheme I) indicate that these The high negative ΔGº values for reaction of the metal ion with L−(Scheme I) indicate that these The high negative ΔGº values for reaction of the metal ion with L−(Scheme I) indicate that these complexation processes are spontaneous, while the complexation processes for reaction of the metal ion with HL (Scheme II) are nonspontaneous due to their positive ΔGº values [with exception of complexation of Zr4+ which characterized by its negative ΔHº values (i.e. have exothermic nature), therefore, it can be described asenthalpy favored processes] [9]. The differences of the thermodynamic parameters ΔlogK1º, ΔΔGº, ΔΔHº and ΔΔSº indicate that the most predominant reaction is the reaction of the metal ion with L−. Although L− reacting species have a lower concentration than HL reacting species in most cases [see Table (2–4)], we can interpret the high spontaneity of reaction of the metal ion with L− due to the electrostatic attraction between the two opposite charged reactants according to the following Eq. (6) which supported with the high ΔlogK1º values. Mz+ + L−K110 ←−→ MLz−1 6 Inspecting ΔSº values, we found that ΔSº values were positive (with exception of some cases in reaction of the metal ion with HL such as the complexation of Zr4+ which have negative ΔSº values). The high positive ΔΔSº values indicate that the standard entropy changes for reaction of the metal ion with L− were higher than those for reaction of the metal ion with HL in all cases under study. This reflects the more spontaneity of reaction of the metal ion with L−. Therefore, the high negative ΔGº values for reaction of the metal ion with L− is attributed to the higher contribution of ΔSº term which indicate that these complexation processes are entropy favored processes. This conclusion was supported by the positive ΔΔSº. Inspecting ΔSº values, we found that ΔSº values were positive (with exception of some cases in reaction of the metal ion with HL such as the complexation of Zr4+ which have negative ΔSº values). MZ+- phenylglycine complexes. MZ+- phenylglycine complexes. Metal ion Log Kº110(Scheme I) Log Kº111(Scheme II) 25ºC 35ºC 45ºC 25ºC 35ºC 45ºC UO2 2+ La3+ Zr4+ 7.10 3.91 10.25 7.65 4.70 10.06 7.78 5.52 9.86 -2.05 -5.54 1.08 -1.70 -4.65 1.00 -1.56 -4.01 0.53 From Tables 5 and 6, we can observe that logK110 >>> logK111. The suggested equations were to get an idea about the difference between the two suggested reactions of the metal ion with L− and HL: ΔlogKº= log Kº110 (reaction I) - log Kº111 (reaction II) (2) ΔΔGº= ΔGº(reaction I) - ΔGº(reaction II) (3) ΔΔHº= ΔHº(reaction I) - ΔHº(reaction II) (4) ΔΔSº= ΔSº(reaction I) - ΔSº(reaction II) (5) ΔlogKº= log Kº110 (reaction I) - log Kº111 (reaction II) (2) ΔΔGº= ΔGº(reaction I) - ΔGº(reaction II) (3) ΔΔHº= ΔHº(reaction I) - ΔHº(reaction II) (4) ΔΔSº= ΔSº(reaction I) - ΔSº(reaction II) (5) ΔlogKº= log Kº110 (reaction I) - log Kº111 (reaction II) (2) ΔlogKº= log Kº110 (reaction I) - log Kº111 (reaction II) (2) ΔΔGº= ΔGº(reaction I) - ΔGº(reaction II) (3) ΔΔHº= ΔHº(reaction I) - ΔHº(reaction II) (4) ΔΔSº= ΔSº(reaction I) - ΔSº(reaction II) (5) Tables (7 and 8) illustrate the standard thermodynamic parameters ΔGº, ΔHº and ΔSº (as well as the differences of the thermodynamic parameters ΔlogK1º, ΔΔGº, ΔΔHº and ΔΔSº) for all possible complexation processes. Table (7): Standard thermodynamic parameters of MZ+-phenylglycine complexes Metal ion T (ºC) Reaction with L−(Scheme I) Reaction with HL(Scheme II) -ΔGº (KJ/mole) ΔHº (KJ/mole) ΔSº (J/deg. mole) ΔGº (KJ/mole) ΔHº (KJ/mole) ΔSº (J/deg. mole) UO2 2+ 25 35 45 40.51 45.12 47.37 62.08 344.26 348.05 344.18 11.70 10.03 9.50 44.64 110.54 112.37 110.50 La3+ 25 35 45 22.31 27.72 32.39 128.04 504.53 505.71 504.50 31.61 27.42 24.42 139.00 360.37 362.27 360.32 Zr4+ 25 35 45 58.48 59.33 60.04 -35.36 77.58 77.83 77.61 -6.16 -5.90 -3.23 -49.50 -145.44 -141.56 -145.50 Table (8): The difference of thermodynamic parameters for MZ+-phenylglycine complexes tandard thermodynamic parameters of MZ+-phenylglycine complexes Table (7): Standard thermodynamic parameters of MZ+-phenylglycine co Table (8): The difference of thermodynamic parameters for MZ+-phenylglycine complexes Page 7/13 Page 7/13 MZ+ T (ºC) Δ logKº -ΔΔGº (KJ/mole) ΔΔHº (KJ/mole) ΔΔSº (J/ deg. MZ+- phenylglycine complexes. The high positive ΔΔSº values indicate that the standard entropy changes for reaction of the metal ion with L− were higher than those for reaction of the metal ion with HL in all cases under study. This reflects the more spontaneity of reaction of the metal ion with L−. Therefore, the high negative ΔGº values for reaction of the metal ion with L− is attributed to the higher contribution of ΔSº term which indicate that these complexation processes are entropy favored processes. This conclusion was supported by the positive ΔΔSº. Page 8/13 In order that, we shall direct our attention to reaction of the metal ion with L− to determine another factor that control the complexation processes under study. Table (9): Thermodynamic stability constants of M2+- phenylglycine complexes as presented in our previous paper [9] Table (9): Thermodynamic stability constants of M2+- phenylglycine complexes as presented in our previous paper [9] Table (9): Thermodynamic stability constants of M2+- phenylglycine complexes as presented in our previous paper [9] Metal ion Log Kº110(Scheme I) Log Kº111(Scheme II) 25ºC 35ºC 45ºC 25ºC 25ºC 25ºC Ni2+ Cu2+ Zn2+ Cd2+ 5.08 9.22 5.44 4.44 5.06 8.96 5.66 5.03 5.00 8.62 5.10 3.94 -4.31 -0.24 -4.02 -5.03 -4.40 -1.38 -3.90 -4.31 -5.00 -1.12 -4.23 -5.39 The nature of metal ion plays a great role in the complexation processes. Inspecting ΔGº values, we notice that the spontaneity and stability of the complex formations were increased in the following order Zr4+>UO2 2+>La3+. The nature of metal ion plays a great role in the complexation processes. Inspecting ΔGº values, we notice that the spontaneity and stability of the complex formations were increased in the following order Zr4+>UO2 2+>La3+. Comparing the thermodynamic stability constants, Log Kº110(Scheme I) of MZ+- phenylglycine in the present work (forZr4+>UO2 2+>La3+) (Tables 5 and 6) with that in our previous paper [9] (for Cu2+> Zn2+> Ni2+> Cd2+) Table 9), we can conclude that the stability of the formed complexes were increased in the following order: Comparing the thermodynamic stability constants, Log Kº110(Scheme I) of MZ+- phenylglycine in the present work (forZr4+>UO2 2+>La3+) (Tables 5 and 6) with that in our previous paper [9] (for Cu2+> Zn2+> Ni2+> Cd2+) Table 9), we can conclude that the stability of the formed complexes were increased in the following order: Zr4+>Cu2+>UO2 2+>Zn2+>Ni2+> Cd2+>La3+ This is a good agreement with Irving-Wiliams order [20]. Conclusions The interactions between phenylglycine and some transition metal ions (UO2 2+,La3+ and Zr4+) were studied at three ionic strengths [(µ = 0.05, 0.10 and 0.15) M KNO3] and three temperatures [(25, 35 and 45) ºC] in aqueous media by using the Bjerrum’s potentiometric technique. The factors that control the interaction between phenylglycine and the metal ions under study have been determined through the calculations of the stability constants and thermodynamic functions. It was found that these factors related to the nature of the reacting spices of the amino acid and also related to the nature of metal ions such as the valance and radius of the ion in addition to some other factors such as Jahn-Teller effect. It was noted that, the most probable reactions were carried out between the metal ion and L−. It was concluded that the degree of complex formation as well as the formation of different reacting spices depend on the pH values. The degree of formation range was 0.5< <1.05(i.e. 1:1 stoichiometric complexes were formed). It was also concluded that the stability of the formed complexes between phenylglycine and MZ+ were increased in good agreement with Irving-Wiliams order as in the following order: −n −n Zr4+>Cu2+>UO2 2+>Zn2+>Ni2+> Cd2+>La3+ Zr4+>Cu2+>UO2 2+>Zn2+>Ni2+> Cd2+>La3+ The high negative ΔGº values for reaction of the metal ion with L− (Scheme I) indicate that these complexation processes are spontaneous, while the complexation processes for reaction of the metal ion with HL (Scheme II)are nonspontaneous due to their positive ΔGº values (with exception of the enthalpy favored complexation processes of Zr4+which characterized by their negative ΔHº values) at different temperatures under study. MZ+- phenylglycine complexes. According to Davies, a simple relation exists between the stability of the formed complexes and the radius, r, of the unhydrated ion and its valancy, z, as reported in Table (10). Therefore, the less stable complexes of La3+ than the other studied transition metal complexes are attributed to the high value of its radius in comparison with the other studied metal ions [21 &22]. It can be concluded that Zr4+complexes are more stable than that for other metal ions complexes due to the larger ratio of (valance/radius) as well as Jahn-Teller effect [23–25]. Table (10): Ionic radii of studied metal ions Table (10): Ionic radii of studied metal ions Page 9/13 Metal ion Radius (Ǻ) z/r Ni2+ Cu2+ Zn2+ Cd2+ Zr4+ UO2 2+ La3+ 0.72 0.69 0.74 0.97 0.86 0.66 1.06 2.78 2.90 2.70 2.06 4.65 3.03 2.83 C l i Metal ion Radius (Ǻ) z/r Ni2+ Cu2+ Zn2+ Cd2+ Zr4+ UO2 2+ La3+ 0.72 0.69 0.74 0.97 0.86 0.66 1.06 2.78 2.90 2.70 2.06 4.65 3.03 2.83 Declarations Page 10/13 -Ethics approval and consent to participate: Not applicable. -Ethics approval and consent to participate: Not applicable. -Ethics approval and consent to participate: Not applicable. -Consent for publication: Not applicable. -Consent for publication: Not applicable. -Availability of data and materials: All data generated or analyzed during this study are included in this published article. -Availability of data and materials: All data generated or analyzed during this study are included in this published article. -Competing interests: Authors declare no competing financial interest. -Funding: No funding is present. El-Dossoki, wrote and discusses the main manuscript text. El-Dossoki, wrote and discusses the main manuscript text. Mohamed, make the experimental part and wrote part of the text. Mohamed, make the experimental part and wrote part of the text. All authors reviewed the manuscript -Acknowledgements: Authors acknowledgement the faculty of science and engineering collage for their contributions in finishing the experimental part of this work. -Acknowledgements: Authors acknowledgement the faculty of science and engineering collage for their contributions in finishing the experimental part of this work. References 1. A. S. Orabi, A. M. Abbas, M. A. AbdElraoof, G. M. Khairy, (2022) Modern view for binary and ternary complexes of metal ions with amoxicillin and some amino acids.Advance in Environmental and Life Science, 1, 22-39. 1. A. S. Orabi, A. M. Abbas, M. A. AbdElraoof, G. M. Khairy, (2022) Modern view for binary and ternary complexes of metal ions with amoxicillin and some amino acids.Advance in Environmental and Life Science, 1, 22-39. 2. A.M. Radalla, (2022) Complexation Equilibria and Determination of Stability Constants of Some Divalent Metal Ion Complexes of L-Cysteine and Diphenylamine in Aqueous Media.American Journal of Chemical Engineering, 10(2), 23-31 3. X. Liu, M. Wu, C. Li, P. Yu, S.Feng, Y. Li, Q. Zhang, (2022) Interaction Structure and Affinity of Zwitterionic Amino Acids with Important Metal Cations (Cd2+, Cu2+, Fe3+, Hg2+, Mn2+, Ni2+ and Zn2+) in Aqueous Solution.Molecules, 27(8), 2407. 4. N.Chetry, T.G. Devi, T Karlo, (2022) Synthesis and characterization of metal complex amino acid using spectroscopic methods and theoretical calculation.Journal of Molecular Structure, 1250(2),131670 5. M Gharakhloo, D Jagleniec, JRomanski, M. Karbarz, (2022) A Novel Self-Healing Hydrogel Based on Derivatives of Natural α-Amino Acids with Potential Applications as a Strain Sensor.Journal of Materials Chemistry B, 10, 4463-4472. 6. . Yao, Z. Yang, L, Huang, C, Su, (2022) Complexation of Amino Acids with Cadmium and Their Application for Cadmium-Contaminated Soil Remediation.Applied Sciences,12(3), 1114. Page 11/13 Page 11/13 7. A.M. Radalla, Abdelatty Mohamed, (2021) Potentiometric Investigation on Complex Formation and Stabilities of Some Divalent Metal Ions with l-Cysteine and Glycine as Ligands in Aqueous Solutions. Journal of Solution Chemistry, 50(7), 954-967. 8. .Rayah, B.Absar, M.R.Ghezzar, F. Abed, (2021) Complexation Studyof Heavy metal ion Cu (II) with Amino Acids before Elimination by Ultrafiltration.Rasayan Journal Chem ,14(3), 1573-1580. . 9. A. A. Mohamed, F. I. El-Dossoki and H.A. Gumaa, (2010) Thermodynamic Study on the Interaction between Phenylglycine and Some Transition Metal Ions.J. Chem. Eng. Data, 55(2), 673– 678. 10. Mohamed, A. A.; Bakr, M. F.; Abd El-Fattah, H. M.( 2003) Thermodynamicstudies on the interaction between some amino acids with some rareearth metal ions in aqueous solutions. Thermochim. Acta, 405, 235–253. 11. Abd El-bary, H. M.; Shata, H. A.; EzElarab, M. A. F.; Mohamed,A. A.; Emara, M. M. (1996) Thermodynamics of amino acid ionization inaqueous solutions using pH-titration. J. Indian Chem. Soc, 73,25. 12. Bakr, M. F.; Mohamed, A. A.; Abd El-Fattah, K. A. 25. rigoryan,R.A., Grigoryan, L.A., Babayan, G.G., (2001) Synthesis of ZnFe2O4–Zn2ZrO4 Solid Solutions, Inorganic materials, Springer. References (2000) Thermodynamic Studies on the Complexation Processes between some Amino Acidsand some Lanthanide Metal Ions in Aqueous Solutions. Al- AzharBull. Sci., 11 (2), 81. 13. (a) Bjerrum, J. Metal Amine Formation in Aqueous Solutions; P.Hasseand Sons: Copenhagen, 1957. (b) Albert, A.; Serjeant, E. P. The Determination of ionization Constants; Chapman and Hall: London,1971; p 7. 14. Mohamed, A. A. An Advanced (2003) Thermodynamic Study on theIonization Processes of some Amino Acids. Al-Azhar Bull. Sci.,14 (1), 237. 15. (a) Christian, D. G. Analytical Chemistry, 3rd ed.; Wiley: New York,1980. (b) Meities, L.; Thomas, L. C. Advanced Analytical Chemistry; McGraw-Hill: New York, 1958. 16. Vogel, A. I. Text Book of QuantitatiVe Inorganic Analysis, 3rd ed.;Longman: London, 1978. 16. Vogel, A. I. Text Book of QuantitatiVe Inorganic Analysis, 3rd ed.;Longm I. Text Book of QuantitatiVe Inorganic Analysis, 3rd ed.;Longman: London, 1 17. Purcell, K. F.; Kotz, J. C. Inorganic Chemistry; W. B. Saunders:Philadelphia, PA, 1977; pp 899-905. 18. Irving, H.; Rossotti, H. S. (1953) Methods for Computing SuccessiveStabilityConstants from Experimental Formation Curves. J. Chem. Soc., 3397–3405. 19. Hartley, F. R.; Burgess, C.; Alcock, R. M. Solution Equilibria; Wiley:New York, 1980. Hartley, F. R.; Burgess, C.; Alcock, R. M. Solution Equilibria; Wiley:New York, 1 20. Barrett, J. Understanding Inorganic Chemistry; Ellis Horwood: London, 1993. 20. Barrett, J. Understanding Inorganic Chemistry; Ellis Horwood: London, 1993. 21. Cotton, F. A.; Wilkinson, G. Advanced Inorganic Chemistry, AcomprehensiVe text, 4th ed.; John Wiley & Sons: New York, 1980. 22. Emara, M. M.; AbdElbary, H. M.; Shehata, H. A.; Ez El-Arab,M. A. F.; Mohamed, A. A. (1995) L-Serine Interaction with some Divalent and Trivalent Metals. Egypt J. Appl. Sci., 10 (9), 332. 23. ZhouJ.,Li R, Introducing a New d0 Sc3+ Asymmetric Ion for Functional Materials: Large Birefringence Enhancement by ScO6 in Ba3Sc2(BO3)4 ,Chem. Phys. Chem., 2022, Wiley Online Library. 24. Reddy, K. V.,Symmetry and Spectroscopy of Molecules; New Age International Publishers, Ltd.: New Delhi, 2002. Page 12/13 Page 12/13 Page 13/13
https://openalex.org/W2788386109	https://europepmc.org/articles/pmc5829336?pdf=render	English	null	The Role of Minimally Invasive Techniques in Scoliosis Correction Surgery	Minimally invasive surgery	2,018	cc-by	4,948	Hindawi Minimally Invasive Surgery Volume 2018, Article ID 4185840, 6 pages https://doi.org/10.1155/2018/4185840 Hindawi Minimally Invasive Surgery Volume 2018, Article ID 4185840, 6 pages https://doi.org/10.1155/2018/4185840 Hindawi Minimally Invasive Surgery Volume 2018, Article ID 4185840, 6 pages https://doi.org/10.1155/2018/4185840 Correspondence should be addressed to Michael B. Cloney; michael.cloney@yahoo.com Received 28 August 2017; Accepted 11 December 2017; Published 24 January 2018 Copyright © 2018 Michael B. Cloney et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Objective. Recently, minimally invasive surgery (MIS) has been included among the treatment modalities for scoliosis. However, literature comparing MIS to open surgery for scoliosis correction is limited. The objective of this study was to compare outcomes for scoliosis correction patients undergoing MIS versus open approach. Methods. We retrospectively collected data on demographics, procedure characteristics, and outcomes for 207 consecutive scoliosis correction surgeries at our institution between 2009 and 2015. Results. MIS patients had lower number of levels fused (𝑝< 0.0001), shorter surgeries (𝑝= 0.0023), and shorter overall lengths of stay (𝑝< 0.0001), were less likely to be admitted to the ICU (𝑝< 0.0001), and had shorter ICU stays (𝑝= 0.0015). On multivariable regression, number of levels fused predicted selection for MIS procedure (𝑝= 0.004), and multiple other variables showed trends toward significance. Age predicted ICU admission and VTE. BMI predicted any VTE, and DVT specifically. Comorbid disease burden predicted readmission, need for transfusion, and ICU admission. Number of levels fused predicted prolonged surgery, need for transfusion, and ICU admission. Conclusions. Patients undergoing MIS correction had shorter surgeries, shorter lengths of stay, and shorter and fewer ICU stays, but there was a significant selection effect. Accounting for other variables, MIS did not independently predict any of the outcomes. Michael B. Cloney ,1 Jack A. Goergen,2 Angela M. Bohnen,1 Zachary A. Smith Tyler Koski,1 and Nader Dahdaleh1 1Department of Neurological Surgery, Feinberg School of Medicine, Chicago, IL, USA 2Feinberg School of Medicine, Chicago, IL, USA Correspondence should be addressed to Michael B. Cloney; michael.cloney@yahoo.com 3. Results 3.1. Demographic Characteristics. There was no difference in age (64.5 ± 1.4 versus 58.1 ± 1.2, Δ−6.44 [−15.11, 2.22], 𝑝= 0.1442), gender (35.7% male versus 25.9% male, 1.588889 [0.5081890, 4.967774], 𝑝= 0.5304), race (𝑝= 0.3243), insurance type (𝑝= 0.6694), smoking status (𝑝= 0.4284), BMI (26.1 ± 0.6 versus 26.8 ± 0.4, Δ0.72 ± 1.47 [−2.17, 3.61], 𝑝= 0.6242), or comorbid disease burden (𝑝= 0.4499). On multivariable regression, age (OR 1.076323 [0.9977851, 1.161042], 𝑝= 0.057) and having private insurance (OR 3.735077 [0.9058839, 15.40021], 𝑝= 0.068) showed trends toward selection for MIS surgery (see Table 2). 2.2. Clinical and Demographic Data. Data on patients’ age, sex, race, BMI, smoking status (ever smoker versus never smoker), number of comorbid diagnoses, and insurance type (private versus other, including Medicare, Medicaid, and disability insurance) at the time of presentation were retrospectively collected for analysis. Data were also collected on the number of levels fused and length of surgery, as well as whether the patients’ scoliosis correction involved a staged procedure, interbody fusion, laminectomy, or osteotomy. Data on surgical techniques and approaches used were also collected. The approaches used included lateral, posterior, and anterior. Of the 14 minimally invasive operations, most utilized an interbody fusion with percutaneous screws; see Table 1 for details. 3.2. Procedure Data. MIS patients were equally likely to have staged surgery (OR 0.6507177 [0.1749019, 2.420978], 𝑝= 0.5186), decompression (OR 0.3597285 [0.04548798, 2.844809], 𝑝 = 0.3127), osteotomy (OR 0.2604895 [0.03313165, 2.048035], 𝑝 = 0.1703), and allograft (OR 0.6491885 [0.2098068, 2.008732], 𝑝= 0.4505). There was a trend toward significance in surgical approach (𝑝= 0.0857) and surgery involving the thoracic spine (OR 0.1539582 [0.008965574, 2.643795], 𝑝= 0.0805). MIS patients were less likely to have autograft (OR 0.1382386 [0.01769585, 1.079909], 𝑝= 0.0289) and had a lower number of levels fused (4.0 versus 9.0, Δ5.0 [2.0, 7.0], 𝑝< 0.0001, Figure 1). There was significantly more variance in the number of levels fused among patients undergoing open surgery (𝑝< 0.0001). On multivariable regression, the number of levels fused predicted selection for MIS procedure (OR 0.6079009 [0.4340611, 0.8513629], 𝑝= 0.004), and there was a trend toward significance for selection for MIS among patients undergoing a posterior approach (OR 3.43426 [0.8365153, 14.09913], 𝑝= 0.087) and not requiring surgical decompres- sion (OR 0.1319887 [0.0147237, 1.183196], 𝑝= 0.070) (see Table 3). 2.3. Outcome Measures. 2. Methods Figure 1: Number of levels fused for patients undergoing MIS approach versus open approach. 2.1. Patient Population. Patients were identified using the Northwestern University Electronic Data Warehouse (EDW). The EDW is an institution-specific registry clinical data repository jointly funded by Northwestern Memorial Hos- pital (NMH), Northwestern Medical Faculty Foundation (NMFF), and Northwestern University Feinberg School of Medicine. We identified all patients who underwent surgery for scoliosis in the Departments of Neurological Surgery or Orthopedic Surgery at Northwestern University between January 1, 2009, and May 31, 2015, as determined by the preoperative indication for surgery provided by the surgeon. for candidate variables on single-variable logistic regression. A value of 𝑝< 0.05 was considered statistically significant. for candidate variables on single-variable logistic regression. A value of 𝑝< 0.05 was considered statistically significant. 1. Introduction of MIS for adult degenerative scoliosis have already been established [7]. Also, results have previously been reported that showed similar clinical improvement for patients who underwent open surgery versus MIS [8]. Furthermore, the patients who underwent MIS had lower morbidity and complication rates and significantly shorter hospital stays [8]. Adult scoliosis is a spinal deformity typically caused by asymmetrical disc degeneration, osteoporosis, and vertebral body compression fractures [1]. When nonsurgical treatment fails, there are multiple surgical techniques that can be used [2]. The goals of surgery are to improve functionality, relieve pain, improve cosmesis, and prevent curve progression [3]. Whether performed posteriorly or anteriorly, open tech- niques are associated with large blood loss, muscle injury and denervation, significant postoperative pain, and other complications [4, 5]. While these initial results are promising, these studies were on small subsets of patients with many confounding variables. The literature comparing open surgery versus MIS for scoliosis correction is limited; therefore, the need exists for further investigation to determine the efficacy of MIS. Here, we compared MIS and open scoliosis surgery with respect to selection for surgical technique and outcomes, including readmission rates, reoperation rates, bleeding, and clotting complications. Minimally invasive surgery (MIS) potentially avoids or lessens these complications due to its ability to reduce intra- operative blood loss, soft tissue damage, infection, postoper- ative pain, and recovery time [6]. The safety and feasibility Minimally Invasive Surgery 2 MIS Open 0 2 4 6 8 10 12 14 Number of levels fused Figure 1: Number of levels fused for patients undergoing MIS approach versus open approach. Table 1: Approaches and techniques used in minimally invasive operations. Number of operations using technique Anterior approach 2 Posterior approach 11 Lateral approach 1 Percutaneous screws 9 Interbody fusion 10 3. Results Odds ratio for MIS patients/open approach patients Confidence interval 𝑝value Posterior approach 3.43426 [0.8365153, 14.09913] 0.087 Number of levels fused 0.6079009 [0.4340611, 0.8513629] 0.004 Staged procedure 0.6507177 [0.1749019, 2.420978] 0.5186 Osteotomy 0.2604895 [0.03313165, 2.048035] 0.1703 Decompression 0.3597285 [0.04548798, 2.844809] 0.3127 Allograft 0.6491885 [0.2098068, 2.008732] 0.4505 Autograft 0.1382386 [0.01769585, 1.079909] 0.0289 Operation involving Thoracic spine 0.1539582 [0.008965574, 2.643795] 0.0805 MIS Open 0 50 100 Percent in surgery 0 800 200 400 600 1000 (Minutes) Figure 2: Comparison in surgery length between MIS and open approach patients. MIS Open 0 50 100 Percent in surgery 0 800 200 400 600 1000 (Minutes) Figure 2: Comparison in surgery length between MIS and open approach patients. 3.3. Outcomes for MIS versus Open Scoliosis Correction. MIS surgery was significantly shorter (287.0 minutes versus 433.0 minutes, HR 2.319 [1.604, 8.342], 𝑝= 0.0023, Figure 2) and was less likely to last ≥6 hours (OR 0.2280405 [0.0751441, 0.6920369], 𝑝= 0.0051). MIS patients had shorter overall lengths of stay (4.5 days versus 8.0 days, HR 3.032 [3.725, 22.61], 𝑝< 0.0001, Figure 3), were less likely to be admitted to the ICU (OR 0.08779576 [0.02348702, 0.3281854], 𝑝< 0.0001), and had shorter ICU stays (19.0 hours versus 48.5 hours, HR 5.174 [5.200, 866.7], 𝑝= 0.0015). 3.3. Outcomes for MIS versus Open Scoliosis Correction. MIS surgery was significantly shorter (287.0 minutes versus 433.0 minutes, HR 2.319 [1.604, 8.342], 𝑝= 0.0023, Figure 2) and was less likely to last ≥6 hours (OR 0.2280405 [0.0751441, 0.6920369], 𝑝= 0.0051). MIS patients had shorter overall lengths of stay (4.5 days versus 8.0 days, HR 3.032 [3.725, 22.61], 𝑝< 0.0001, Figure 3), were less likely to be admitted to the ICU (OR 0.08779576 [0.02348702, 0.3281854], 𝑝< 0.0001), and had shorter ICU stays (19.0 hours versus 48.5 hours, HR 5.174 [5.200, 866.7], 𝑝= 0.0015). 𝑝 On single-variable analysis, MIS patients were equally likely to experience the following within 30 days of surgery: readmission (OR 0.3271202 [0.01872778, 5.713847], 𝑝 = 0.2318), reoperation (OR 1.181818 [0.06218783, 22.45929], 𝑝= 1.0000), DVT (OR 0.3477832 [0.01986991, 6.087253], 𝑝= 0.2464), PE (OR 3.634615 [0.3782204, 34.92786], 𝑝= 0.2328), any VTE (OR 0.8509616 [0.1044311, 6.934095], 𝑝= 0.8800), and postoperative death (OR 2.668966 [0.1222150, 58.28564], 𝑝= 0.7034). MIS patients were less likely to require transfusion (OR 0.1231231 [0.02681325, 0.5653661], 𝑝= 0.0017). Figure 2: Comparison in surgery length between MIS and open approach patients. 3. Results Odds ratio for MIS patients/open approach patients Confidence interval 𝑝value Posterior approach 3.43426 [0.8365153, 14.09913] 0.087 Number of levels fused 0.6079009 [0.4340611, 0.8513629] 0.004 Staged procedure 0.6507177 [0.1749019, 2.420978] 0.5186 Osteotomy 0.2604895 [0.03313165, 2.048035] 0.1703 Decompression 0.3597285 [0.04548798, 2.844809] 0.3127 Allograft 0.6491885 [0.2098068, 2.008732] 0.4505 Autograft 0.1382386 [0.01769585, 1.079909] 0.0289 Operation involving Thoracic spine 0.1539582 [0.008965574, 2.643795] 0.0805 3.3. Outcomes for MIS versus Open Scoliosis Correction. MIS surgery was significantly shorter (287.0 minutes versus 433.0 minutes, HR 2.319 [1.604, 8.342], 𝑝= 0.0023, Figure 2) and was less likely to last ≥6 hours (OR 0.2280405 [0.0751441, 0.6920369], 𝑝= 0.0051). MIS patients had shorter overall lengths of stay (4.5 days versus 8.0 days, HR 3.032 [3.725, 22.61], 𝑝< 0.0001, Figure 3), were less likely to be admitted to the ICU (OR 0.08779576 [0.02348702, 0.3281854], 𝑝< 0.0001), and had shorter ICU stays (19.0 hours versus 48.5 hours, HR 5.174 [5.200, 866.7], 𝑝= 0.0015). On single-variable analysis, MIS patients were equally likely to experience the following within 30 days of surgery: readmission (OR 0.3271202 [0.01872778, 5.713847], 𝑝 = 0.2318), reoperation (OR 1.181818 [0.06218783, 22.45929], 𝑝= 1.0000), DVT (OR 0.3477832 [0.01986991, 6.087253], 𝑝= 0.2464), PE (OR 3.634615 [0.3782204, 34.92786], 𝑝= 0.2328), any VTE (OR 0.8509616 [0.1044311, 6.934095], 𝑝= 0.8800), and postoperative death (OR 2.668966 [0.1222150, 58.28564], 𝑝= 0.7034). MIS patients were less likely to require transfusion (OR 0.1231231 [0.02681325, 0.5653661], 𝑝= 0.0017). 3.4. Predictors of Outcomes after Scoliosis Correction. On multivariable regression, BMI predicted DVT within 30 days postoperatively (OR 1.130749 [1.021063, 1.252219], 𝑝= 0.018), and age (OR 1.048752 [0.9975131, 1.102623], 𝑝= 0.063) showed a trend toward significance. Number of levels MIS Open 0 50 100 Percent in surgery 0 800 200 400 600 1000 (Minutes) Figure 2: Comparison in surgery length between MIS and open approach patients. fused showed a trend toward significance in predicting PE 30 days postop (OR 1.251054 [0.9814253, 1.594758], 𝑝 = 0.071). Age (OR 1.053943 [1.000293, 1.11047], 𝑝= 0.049) and BMI (OR 1.143371 [1.037006, 1.260645], 𝑝= 0.007) predicted VTE within 30 days postop. Comorbid disease burden predicted readmission within 30 days postop (OR Table 3: Surgical procedure data for MIS approach versus open approach. Table 3: Surgical procedure data for MIS approach versus open approach. 3. Results Information about complications within 30 days after the surgery included the cumulative 30- day incidence and timing of VTEs (defined as either DVT or PE), all-cause readmissions, reoperations, ICU admission, length of ICU stay, length of hospital stay, and incidence of death. 2.4. Statistical Methods. Microsoft Excel 2011 (Microsoft, Redmond, WA, USA) was used to manage data. Statistical analysis was performed using Stata 12.0 (StataCorp, College Station, TX, USA) and Prism 6.0b (GraphPad Software, Inc., La Jolla, CA, USA). Parametric data was given as mean ± standard deviation and compared using a t-test. Nonparametric data was compared using Mann–Whitney 𝑈 test or Chi-square test, as appropriate. Regression analysis was performed using stepwise logistic regression, with an inclusion threshold for the multivariable model of 𝑝< 0.10 2.4. Statistical Methods. Microsoft Excel 2011 (Microsoft, Redmond, WA, USA) was used to manage data. Statistical analysis was performed using Stata 12.0 (StataCorp, College Station, TX, USA) and Prism 6.0b (GraphPad Software, Inc., La Jolla, CA, USA). Parametric data was given as mean ± standard deviation and compared using a t-test. Nonparametric data was compared using Mann–Whitney 𝑈 test or Chi-square test, as appropriate. Regression analysis was performed using stepwise logistic regression, with an inclusion threshold for the multivariable model of 𝑝< 0.10 3 Minimally Invasive Surgery Table 2: Patient demographic data for MIS approach versus open approach. Table 2: Patient demographic data for MIS approach versus open approach. Table 2: Patient demographic data for MIS approach versus open approach. MIS patients Open patients 𝑝value Age 64.5 years (mean) 58.1 years (mean) 0.1442 Gender 35.7% male 25.9% male 0.5304 Race 100% Caucasian 86.01% Caucasian, 4.15% African American, 9.84% other 0.3243 Insurance type 10 private, 4 Medicare 108 private, 77 Medicare, 7 Medicaid, 1 other 0.6694 Smoking status 11 never smoked, 3 quit more than 12 months ago 119 never smoked, 53 quit more than 12 months ago, 21 current smokers or smoked within last 12 months 0.4284 BMI 26.1 (mean) 26.8 (mean) 0.6242 Comorbidities 2.43 comorbidities per patient (cardiac, renal, pulmonary, endocrine, or hypertension) 2.8 comorbidities per patient (cardiac, renal, pulmonary, endocrine, or hypertension) 0.4499 or hypertension) hypertension) Table 3: Surgical procedure data for MIS approach versus open approach. 3. Results Number of levels fused predicted prolonged surgery (surgery > 6 h) (OR 1.452142 [1.233844, 1.709062], 𝑝 < 0.001), and a history of smoking showed a trend toward significance (OR 0.0040854 [7.33𝑒−06, 2.275654], 𝑝 = 0.088). Number of levels fused (OR 1.297174 [1.182993, 1.422377], 𝑝 < 0.001) and comorbid disease burden (OR 1.297174 [1.182993, 1.422377], 𝑝< 0.001) predicted the need for transfusion, and osteotomy showed a trend toward significance (OR 2.359625 [0.9492208, 5.865686], 𝑝 = 0.065). Age (OR 0.9512098 [0.9089738, 0.9954083], 𝑝= 0.031), gender (OR 3.299076 [1.06138, 10.25448], 𝑝= 0.039), comorbid disease burden (OR 1.686387 [1.123766, 2.530687], 𝑝= 0.012), number of levels fused (OR 2.089387 [1.615015, 2.703095], 𝑝< 0.001), and undergoing a staged procedure (OR 5.321398 [1.470397, 19.25826], 𝑝= 0.011) all predicted ICU admission. 2.543268 [1.376737, 4.698219], 𝑝= 0.003), and involvement of the thoracic spine showed a trend toward significance (OR 0.1084136 [0.0109545, 1.072934], 𝑝= 0.057). Number of levels fused predicted prolonged surgery (surgery > 6 h) (OR 1.452142 [1.233844, 1.709062], 𝑝 < 0.001), and a history of smoking showed a trend toward significance (OR 0.0040854 [7.33𝑒−06, 2.275654], 𝑝 = 0.088). Number of levels fused (OR 1.297174 [1.182993, 1.422377], 𝑝 < 0.001) and comorbid disease burden (OR 1.297174 [1.182993, 1.422377], 𝑝< 0.001) predicted the need for transfusion, and osteotomy showed a trend toward significance (OR 2.359625 [0.9492208, 5.865686], 𝑝 = 0.065). Age (OR 0.9512098 [0.9089738, 0.9954083], 𝑝= 0.031), gender (OR 3.299076 [1.06138, 10.25448], 𝑝= 0.039), comorbid disease burden (OR 1.686387 [1.123766, 2.530687], 𝑝= 0.012), number of levels fused (OR 2.089387 [1.615015, 2.703095], 𝑝< 0.001), and undergoing a staged procedure (OR 5.321398 [1.470397, 19.25826], 𝑝= 0.011) all predicted ICU admission. MIS surgery was significantly shorter (287.0 minutes versus 433.0 minutes, 𝑝= 0.0023, Figure 2) and patients undergoing MIS were less likely to have surgery last >6 hours (𝑝= 0.0051) based on single-variable analysis. Anand et al. noted that their surgical outcomes data for MIS scoliosis correction was similar to open correction outcomes data when compared to the literature [8]. However, a meta- analysis on MIS versus open approach in degenerative lumbar disease revealed significant variability in operating times [15]. For example, one study found an average operating time of 161 minutes for the MIS approach compared to 375 minutes for the open approach [16]. In contrast, a study in the same meta-analysis found an average operating time of 159.2 minutes for the MIS approach versus 113.06 minutes for the open approach [17]. 3. Results We suspect that con- founding variables have an important impact on operating time, which would explain the significant variability between studies. 3. Results fused showed a trend toward significance in predicting PE 30 days postop (OR 1.251054 [0.9814253, 1.594758], 𝑝 = 0.071). Age (OR 1.053943 [1.000293, 1.11047], 𝑝= 0.049) and BMI (OR 1.143371 [1.037006, 1.260645], 𝑝= 0.007) predicted VTE within 30 days postop. Comorbid disease burden predicted readmission within 30 days postop (OR 3.4. Predictors of Outcomes after Scoliosis Correction. On multivariable regression, BMI predicted DVT within 30 days postoperatively (OR 1.130749 [1.021063, 1.252219], 𝑝= 0.018), and age (OR 1.048752 [0.9975131, 1.102623], 𝑝= 0.063) showed a trend toward significance. Number of levels 4 MIS Open 0 50 100 Percent in hospital 0 20 10 40 30 (Days) Figure 3: Comparison in LOS between MIS and open approach patients. Minimally Invasive Surgery 4 that both older patients and patients with less severe defor- mities were more likely to be selected for an MIS technique [13]. Our analysis agrees with this finding. i Furthermore, having private insurance (𝑝 = 0.068), undergoing a posterior approach (𝑝= 0.087), and not requir- ing surgical decompression (𝑝= 0.070) each showed a trend toward selection for MIS. Similarly to our finding that private insurance was an important determinant of surgery choice, a study by Park and Ha revealed that the cost of the MIS technique determined which patients underwent MIS versus open technique [6]. Another group separated the patients by allocating the private hospital patients to receive MIS and the public hospital patients in their study to receive the open technique [14]. This trend of selection biases between the two groups was consistent across most of the studies that reported patient selection information, limiting the ability to conclude a true difference in outcomes between the MIS and open techniques.i Percent in hospital Figure 3: Comparison in LOS between MIS and open approach patients. In our study, there was significantly more variance in the number of levels fused among patients undergoing open surgery (𝑝< 0.0001). Dangelmajer et al. came to the same conclusion in their systematic review and attributed it to the fact that patients undergoing open procedure had a larger preoperative scoliosis [13]. This result shows that an open technique can be used for a broader range of spinal levels than MIS.i 2.543268 [1.376737, 4.698219], 𝑝= 0.003), and involvement of the thoracic spine showed a trend toward significance (OR 0.1084136 [0.0109545, 1.072934], 𝑝= 0.057). References [1] M. Aebi, “The adult scoliosis,” European Spine Journal, vol. 14, no. 10, pp. 925–948, 2005. Our finding that comorbid disease burden independently predicted ICU admission (𝑝 = 0.012) and readmission within 30 days (𝑝= 0.003) is consistent with the existing spine surgery literature [20, 21]. Cardiac, GI, and respiratory issues that were present before the operation are frequent causes of ICU admission and readmission and appear to be an important factor when comparing MIS versus open technique for scoliosis correction as well. [2] S. Yadla, M. G. Maltenfort, J. K. Ratliff, and J. S. Harrop, “Adult scoliosis surgery outcomes: a systematic review,” Neurosurg Focus, vol. 28, no. E3, 2010. [3] N. Anand, E. M. Baron, and B. Khandehroo, “Is circumfer- ential minimally invasive surgery effective in the treatment of moderate adult idiopathic scoliosis?” Clinical Orthopaedics and Related Research, vol. 472, pp. 1762–1768, 2014. [4] J. J. Regan, H. Yuan, and P. C. McAfee, “Laparoscopic fusion of the lumbar spine: Minimally invasive spine surgery: A prospective multicenter study evaluating open and laparoscopic lumbar fusion,” The Spine Journal, vol. 24, no. 4, pp. 402–411, 1999. On multivariable regression analysis, age (𝑝= 0.049) and BMI (𝑝= 0.007) predicted VTE within 30 days postop. These variables are typically found to be strongly associated with such outcomes in spinal surgery, as evidenced in numerous previous studies [22–24]. Importantly, MIS surgery was not found to be an independent predictor of any outcome analyzed during multivariable regression. So, although we found that the typical outcome predictors (age, BMI, and comorbid disease burden) were significant in this study, we did not find any significant difference in patient outcome based on MIS versus open technique alone.h [5] M. Y. Wang and P. V. Mummaneni, “Minimally invasive surgery for thoracolumbar spinal deformity: initial clinical experience with clinical and radiographic outcomes,” Neurosurg Focus, vol. 28, p. E9, 2010. [6] Y. Park and J. W. Ha, “Comparison of one-level posterior lumbar interbody fusion performed with a minimally invasive approach or a traditional open approach,” The Spine Journal, vol. 32, no. 5, pp. 537–543, 2007. Our study has a number of important limitations. The study was conducted retrospectively and is subject to the biases inherent to this study design. 4. Discussion Minimally invasive surgical techniques could potentially reduce the morbidity associated with traditional open sur- gical techniques in scoliosis correction [9–11]. Currently, the literature on MIS for scoliosis correction is limited. Many of the studies performed to date observed relatively few patients, and multiple systematic reviews have concluded that more research is needed [9, 12]. Our study examined 207 scoliosis correction surgeries and identified selection factors for MIS, how MIS outcomes compare to open surgery outcomes, and predictors of outcomes for each technique. MIS patients also had shorter overall lengths of stay based on single-variable analysis (4.5 days versus 8.0 days, 𝑝< 0.0001, Figure 3). While it seems promising that MIS patients typically had a shorter length of stay, the results from other studies are variable, potentially indicating a selection effect [6, 14, 18]. Our length of stay results were consistent with a meta-analysis by Phan et al. that found a median length of stay of 4.7 days for the MIS approach and 8 days for the open approach [15]. Although our results are consistent with the meta-analysis, the presence of confounding variables makes it uncertain if MIS versus open surgery is the etiology of the variability in length of stay. Importantly, the median number of levels fused predicted selection for MIS technique with MIS patients having fewer levels fused (4 versus 9, 𝑝< 0.0001). Many previous studies did not search for a selection effect for MIS versus open surgery selection. One meta-analysis on scoliosis correction, by Dangelmajer et al., did examine selection bias and found Compared to patients undergoing open surgery, MIS patients were less likely to be admitted to the ICU (𝑝< 0.0001) but did have shorter ICU stays when it was required 5 Minimally Invasive Surgery MIS versus open surgery, as well as a variety of perioperative outcomes. MIS versus open surgery, as well as a variety of perioperative outcomes. (19.0 hours versus 48.5 hours, 𝑝= 0.0015). These results are consistent with previous groups who found that open surgery patients typically had more complications following surgical treatment [13, 19]. However, similarly to our study, these groups also noted the presence of confounding variables such as age and preoperation severity of deformity that could have attributed to these results. In fact, in our analysis, open surgery did not predict ICU admission on multivariable regression. (19.0 hours versus 48.5 hours, 𝑝= 0.0015). 4. Discussion These results are consistent with previous groups who found that open surgery patients typically had more complications following surgical treatment [13, 19]. However, similarly to our study, these groups also noted the presence of confounding variables such as age and preoperation severity of deformity that could have attributed to these results. In fact, in our analysis, open surgery did not predict ICU admission on multivariable regression. 5. Conclusion Patients undergoing MIS scoliosis correction had shorter surgeries, shorter lengths of stay, and shorter and fewer ICU stays, but there was a significant selection effect. Accounting for other clinical variables, undergoing MIS surgery did not independently predict any of the outcomes analyzed. While MIS or open correction was not independently associated with ICU admission, the number of levels fused did independently predict ICU admission (𝑝< 0.001). As the number of levels fused was also independently associated with the selection of technique, it may be a confounding factor that accounts for the significant difference in likelihood of ICU admission on single-variable analysis. Similarly, age was likely a confounding factor, as age showed a trend toward significance in predicting technique, and was a significant predictor of ICU admission (𝑝= 0.031).i References Mobbs, P. Sivabalan, and J. Li, “Minimally invasive surgery compared to open spinal fusion for the treatment of degenera- tive lumbar spine pathologies,” Journal of Clinical Neuroscience, vol. 19, pp. 829–835, 2012. [15] K. Phan, P. J. Rao, A. C. Kam, and R. J. Mobbs, “Minimally invasive versus open transforaminal lumbar interbody fusion for treatment of degenerative lumbar disease: systematic review and meta-analysis,” European Spine Journal, vol. 24, pp. 1017– 1030, 2015. [16] W. A. R. Sulaiman and M. Singh, “Minimally invasive versus open transforaminal lumbar interbody fusion for degenerative spondylolisthesis grades 1-2: patient-reported clinical outcomes and cost-utility analysis,” The Ochsner Journal, vol. 14, no. 1, pp. 32–37, 2014. [17] N.-F. Tian and F.-M. Mao, “Minimally invasive versus open transforaminal lumbar interbody fusion: a meta-analysis based on the current evidence,” European Spine Journal, vol. 23, no. 4, pp. 929-930, 2014. [18] F. Shunwu, Z. Xing, Z. Fengdong, and F. Xiangqian, “Minimally invasive transforaminal lumbar interbody fusion for the treat- ment of degenerative lumbar diseases,” The Spine Journal, vol. 35, no. 17, pp. 1615–1620, 2010. [19] R. E. Isaacs, J. Hyde, J. A. Goodrich, W. B. Rodgers, and F. M. Phillips, “A prospective, nonrandomized, multicenter evaluation of extreme lateral interbody fusion for the treatment of adult degenerative scoliosis: perioperative outcomes and complications,” The Spine Journal, vol. 35, pp. S322–330, 2010. [20] H. F. Kay, S. Chotai, J. B. Wick, D. P. Stonko, M. J. McGirt, and C. J. Devin, “Preoperative and surgical factors associated with postoperative intensive care unit admission following operative treatment for degenerative lumbar spine disease,” European Spine Journal, vol. 25, no. 3, pp. 843–849, 2016. [21] R. A. McCormack, T. Hunter, N. Ramos, R. Michels, L. Hutzler, and J. A. Bosco, “An Analysis of Causes of Readmission After Spine Surgery,” The Spine Journal, vol. 37, no. 14, pp. 1260–1266, 2012. [22] M. P. Glotzbecker, C. M. Bono, K. B. Wood, and M. B. Har- ris, “Thromboembolic disease in spinal surgery: a systematic review,” Spine (Phila Pa 1976), vol. 34, pp. 291–303, 2009. [23] S. Z. Goldhaber and H. Bounameaux, “Pulmonary embolism and deep vein thrombosis,” Lancet, vol. 379, pp. 1835–1846, 2012.h [24] K. Papadimitriou, A. G. Amin, R. M. Kretzer et al., “Throm- boembolic events and spinal surgery,” Journal of Clinical Neuro- science, vol. 19, no. 12, pp. 1617–1621, 2012. [25] S. H. Berven, V. Deviren, J. A. Smith, S. H. Hu, and D. S. References A prospective study would enable us to further understand if the trends we discov- ered (private insurance, number of levels fused showing selec- tion for MIS) were actively affecting the surgeon’s decision whether to use an MIS or open approach. The operations we collected data on varied in the minimally invasive technique and approach used, which made it a less homogenous population to draw conclusions from. As a single-institution study, it only reflects the clinical decision-making of our spine surgeons with respect to patient selection and management. Our series is limited by its size, and a larger series would allow for a more thorough comparison between MIS and open surgery. Our study does not provide radiographic compar- isons of corrections, as is common in the scoliosis literature. However, multiple prior studies have compared radiographic outcome for MIS and open scoliosis correction, the results of which have been meta-analyzed [2, 5, 7, 13, 25–27]. Despite its limitations, our study contributes to the existing literature on scoliosis correction by examining selection factors for [7] E. Dakwar, R. F. Cardona, D. A. Smith, and J. S. Uribe, “Early outcomes and safety of the minimally invasive, lateral retroperi- toneal transpsoas approach for adult degenerative scoliosis,” Neurosurg Focus, vol. 28, no. 3, p. E8, 2010. [8] N. Anand, R. Rosemann, B. Khalsa, and E. M. Baron, “Mid- term to long-term clinical and functional outcomes of mini- mally invasive correction and fusion for adults with scoliosis,” Neurosurg Focus, vol. 28, E6, 2010. [9] K. Bach, A. Ahmadian, A. Deukmedjian, and J. S. Uribe, “Minimally invasive surgical techniques in adult degenerative spinal deformity: a systematic review,” Clinical Orthopaedics and Related Research, vol. 472, pp. 1749–1761, 2014. [10] E. M. Baron and T. J. Albert, “Medical complications of surgical treatment of adult spinal deformity and how to avoid them,” Spine (Phila Pa 1976), vol. 31, pp. S106–118, 2006. [11] S. S. Hu, “Blood loss in adult spinal surgery,” European Spine Journal, vol. 13, no. S01, pp. S3–S5, 2004. [12] N. Anand, E. M. Baron, and S. Kahwaty, “Evidence basis/ outcomes in minimally invasive spinal scoliosis surgery,” Neu- rosurgery Clinics of North America, vol. 25, pp. 361–375, 2014. Minimally Invasive Surgery 6 [13] S. Dangelmajer, P. L. Zadnik, S. T. Rodriguez, Z. L. Gokaslan, and D. M. Sciubba, “Minimally invasive spine surgery for adult degenerative lumbar scoliosis,” Neurosurgical Focus, vol. 36, no. 5, p. E7, 2014. [14] R. J. References Bradford, “Management of fixed sagittal plane deformity: outcome of combined anterior and posterior surgery,” Spine (Phila Pa 1976), vol. 28, no. 15, pp. 1710–1715, 2003. [26] M. O. Kelleher, M. Timlin, O. Persaud, and Y. R. Rampersaud, “Success and failure of minimally invasive decompression for focal lumbar spinal stenosis in patients with and without deformity,” The Spine Journal, vol. 35, no. 19, pp. E981–E987, 2010. [27] W. Liu, X. S. Chen, L. S. Jia, and D. W. Song, “The clinical features and surgical treatment of degenerative lumbar scoliosis: a review of 112 patients,” Orthopaedic Surgery, vol. 1, pp. 176–183, 2009.
https://openalex.org/W4280640736	https://www.frontiersin.org/articles/10.3389/fpsyg.2022.745608/pdf	English	null	Risk Matrix for Violent Radicalization: A Machine Learning Approach	Frontiers in psychology	2,022	cc-by	14,644	ORIGINAL RESEARCH published: 12 May 2022 doi: 10.3389/fpsyg.2022.745608 ORIGINAL RESEARCH published: 12 May 2022 doi: 10.3389/fpsyg.2022.745608 Risk Matrix for Violent Radicalization: A Machine Learning Approach Krisztián Ivaskevics and József Haller* Department of Criminal Psychology, Faculty of Law Enforcement, University of Public Service, Budapest, Hungary Krisztián Ivaskevics and József Haller* Department of Criminal Psychology, Faculty of Law Enforcement, University of Public Service, Budapest, Hungary Hypothesis-driven approaches identiﬁed important characteristics that differentiate violent from non-violent radicals. However, they produced a mosaic of explanations as they investigated a restricted number of preselected variables. Here we analyzed without a priory assumption all the variables of the “Proﬁles of Individual Radicalization in the United States” database by a machine learning approach. Out of the 79 variables considered, 19 proved critical, and predicted the emergence of violence with an accuracy of 86.3%. Typically, violent extremists came from criminal but not radical backgrounds and were radicalized in late stages of their life. They were followers in terrorist groups, sought training, and were radicalized by social media. They belonged to low social strata and had problematic social relations. By contrast, non-violent but still criminal extremists were characterized by a family tradition of radicalism without having criminal backgrounds, belonged to higher social strata, were leaders in terrorist organizations, and backed terrorism by supporting activities. Violence was also promoted by anti-gay, Sunni Islam and Far Right, and hindered by Far Left, Anti- abortion, Animal Rights and Environment ideologies. Critical characteristics were used to elaborate a risk-matrix, which may be used to predict violence risk at individual level. Edited by: Heng Choon (Oliver) Chan, City University of Hong Kong, Hong Kong SAR, China Reviewed by: Jason Roach, University of Huddersﬁeld, United Kingdom Corine de Ruiter, Maastricht University, Netherlands *Correspondence: József Haller Haller.Jozsef@uni-nke.hu Edited by: Heng Choon (Oliver) Chan, City University of Hong Kong, Hong Kong SAR, China Edited by: Heng Choon (Oliver) Chan, City University of Hong Kong, Hong Kong SAR, China Reviewed by: Jason Roach, University of Huddersﬁeld, United Kingdom Corine de Ruiter, Maastricht University, Netherlands Keywords: machine learning, terrorism, violent extremism, risk assessment, XGBoost INTRODUCTION Research into terrorism can be divided into several major areas. One approach views radicalization as a process (Borum, 2011) and addresses the pathways that lead to extremism in general and violent extremism by investigating the life history of individuals in the light of underlying individual characteristics, life events and social conditions (Borum, 2003; Moghaddam, 2005; McCauley and Moskalenko, 2008; Aghabi et al., 2017). The aim of this approach is to understand the process of radicalization, which may reveal opportunities for interrupting or reversing its progression. The delineation of terrorist typologies—the second popular approach—is mainly based on structured professional judgments (Pressman and Flockton, 2012; Sarma, 2017; Meloy, 2018), which was indicated as highly preferable by scholars of the ﬁeld (Monahan, 2012). The third major approach aims at the identiﬁcation of risk factors which can be used in counterterrorism activities (Pressman and Flockton, 2012; Cook, 2014; Lloyd and Dean, 2015; Meloy, 2018). Ideally, this third approach— adopted in this study—would serve as an empirical basis for the other approaches and may provide means to improve both our views on radicalization pathways and expert opinion-based terrorist typologies. Specialty section: This article was submitted to Forensic and Legal Psychology, a section of the journal Frontiers in Psychology Received: 22 July 2021 Accepted: 14 April 2022 Published: 12 May 2022 Specialty section: This article was submitted to Forensic and Legal Psychology, a section of the journal Frontiers in Psychology Specialty section: This article was submitted to Forensic and Legal Psychology, a section of the journal Frontiers in Psychology Received: 22 July 2021 Accepted: 14 April 2022 Published: 12 May 2022 Citation: The main statistical tools employed were bivariate correlations, often supplemented with multivariate approaches. Those who pursued this route were regularly focusing on diﬀerences between violent and non-violent extremists, and preferentially relied on the “Proﬁles of Individual Radicalization in the United States (PIRUS)” database developed by the National Consortium for the Study of Terrorism and Responses to Terrorism (START), University of Maryland (Jensen et al., 2015). The basic assumption of the mathematical approach is that the features of terrorists and their organizations are organized into patterns that can be described mathematically and used for prediction. Instead of relying on hypotheses, researchers employ data mining techniques to extract hidden predictive knowledge without a priori assumptions (Gassebner and Luechinger, 2011; Tolan and Soliman, 2015; Kerodal et al., 2016; Basuchoudhary and Bang, 2018). Mathematical approaches include the development of equations that may describe, for instance the spreading and ampliﬁcation of online support for terrorism (Johnson et al., 2016) or predict the group responsible for a terrorist attack (Sachan and Roy, 2012). In recent years, however, machine learning algorithms seem to take the lead in this ﬁeld. They automatically evaluate the relationships between independent variables on a training dataset and apply the recognized rules to a test dataset to make predictions regarding the dependent variable. For instance, the algorithm may allocate individuals of the test dataset into violent and non-violent populations. The percent success rate of allocation indicates the degree of applicability of the procedure for predictions. y y The main advantage of this approach is its solid theoretical framework, which helps avoiding the statistical pitfalls of “ﬁshing expeditions.” At the same time, however, such studies deal with a relatively low number of factors, i.e., those relevant for the theoretical framework addressed. For instance, three studies addressed violent extremism based on 23, 29, and 21 variables (Jasko et al., 2017; LaFree et al., 2018; Becker, 2019). Albeit these numbers were by far not negligible, they constituted about one ﬁfth of the number of factors contained in the PIRUS database from where data were extracted. More importantly, the factors showed little overlap across studies due to the diﬀerential theoretical approach. For instance, Becker (2019)— who focused on social control and social learning theories— included only 30% of the factors studied by LaFree et al. (2018) who employed a criminological approach. On its turn, Jasko et al. Citation: Ivaskevics K and Haller J (2022) Risk Matrix for Violent Radicalization: A Machine Learning Approach. Front. Psychol. 13:745608. doi: 10.3389/fpsyg.2022.745608 y g It was suggested that risk assessment should have four components “(1) identifying empirically- valid risk factors, (2) determining a method for measuring (scoring) risk factors, (3) establishing May 2022 \| Volume 13 \| Article 745608 1 Frontiers in Psychology \| www.frontiersin.org Machine Learning and Violent Terrorism Ivaskevics and Haller (LaFree et al., 2018), revealed the role of social control and social learning (Becker, 2019), helped understanding interactions between social phenomena and individual factors (Youngblood, 2020), allowed the description of ﬁve terrorist types (Ligon et al., 2019), revealed diﬀerences between terrorist ideologies (Jensen et al., 2015; Freilich et al., 2018; Varaine, 2019), outlined the role of prison radicalization (LaFree et al., 2019), revealed factors aﬀecting plot success (Austin, 2019), and uncovered diﬀerences between criminal and terrorist groups (Pyrooz et al., 2017). Despite important achievements, however, hypothesis-driven approaches provide a mosaic of explanations rather than a unitary system, for which their predictions seem incomplete. More precise predictions may be achieved by mathematical approaches. a procedure for combining scores on the risk factors, and (4) producing an estimate of violence risk” (Skeem and Monahan, 2011). However, the identiﬁcation of empirically valid risk factors was not accomplished fully so far. Thus, the need of identifying such factors is still current. Attempts to reach the goal can roughly be grouped into two major categories: hypothesis-driven and mathematics-driven approaches. a procedure for combining scores on the risk factors, and (4) producing an estimate of violence risk” (Skeem and Monahan, 2011). However, the identiﬁcation of empirically valid risk factors was not accomplished fully so far. Thus, the need of identifying such factors is still current. Attempts to reach the goal can roughly be grouped into two major categories: hypothesis-driven and mathematics-driven approaches. Hypothesis-driven approaches borrowed concepts from various scientiﬁc disciplines to select critical factors that may underlie violent radicalization (LaFree et al., 2018). Such studies used concepts of communication science (Youngblood, 2020), criminology (LaFree et al., 2018, 2019), economics (Varaine, 2019), political science (Abrahms, 2012), social psychology (Jasko et al., 2017; Carson et al., 2019), sociology (Becker, 2019; Ligon et al., 2019), etc. Typically, studies start by a thorough analysis of the theoretical explanations of violence, which serve as criteria for the selection of the variables chosen for analysis. Frontiers in Psychology \| www.frontiersin.org Dependent and Independent Variables Dependent and Independent Variables The dependent variable was the ﬁeld “Violent” of the database. This covered two characteristics: lack of violent activities (non- violent) and violent acts. The latter were deﬁned in the Codebook as “actively participate in ideologically motivated operations/actions that resulted in causalities/injuries or clearly intended to result in causalities/injuries” (Codebook page 12). Thus, violent individuals perpetrated terrorist attacks that resulted in, or were clearly planned to result in injuries and/or fatalities. Noteworthy, non-violent extremists also engaged in unlawful activities by supporting extremist organizations and/or terrorist plots by, e.g., by providing ﬁnancial or legal help or by recruiting new extremists. Thus, non/violent individuals were extremists who refrained from committing violent attacks but contributed to their completion. The variable “Violent” diﬀerentiated a non-violent (N = 908) and a violent group (N = 1240). Citation: The likely reason was that the accuracy of prediction depended to a great extent on the variables chosen for the model and the handling of missing data (Gruenewald and Pridemore, 2011; Tolan and Soliman, 2015). were reviewed by full-time project researchers. In the third stage, researchers coded the relevant background, contextual, and ideological information. Random sampling techniques were used to maximize the representativeness of the dataset at all points in time that were covered by the project (START, 2018). y p j The database contains information on individuals who (1) were fully or partially radicalized in the United States, (2) espoused ideological motives, (3) engaged in ideologically motivated acts, and (4) were either arrested, indicted, or killed because of their ideological activities in the United States (Jensen et al., 2016). Individuals fulﬁlling the ﬁrst three criteria were also included if they were members of a designated terrorist organization or were associated with an organization whose leaders or founders were indicted of an ideologically motivated violent oﬀense. Consequently, neither foreign ﬁghters nor United States extremists active abroad were included in the database. The version used in this study contained information on 2,148 extremists characterized by 112 variables (called “ﬁelds” in the codebook of the database). These were arranged by the authors of the database into groups of variables which they clustered into six “supergroups” (Table 1). The variables were chosen to represent the radicalization processes as proposed by ﬁve core theories, e.g., cost/beneﬁt theory, psychological models, recruitment theory, social identity theory, and social movement theory (Jensen et al., 2016). Each variable could take several values, which represented terrorist characteristics. For instance, the variable “Broad ethnicity” covered the following seven characteristics in the database: Hispanic/Latino, Black/African American, White, Middle Eastern/North African, Native American, Asian, and Other. As all variables covered several characteristics, the individual contribution of these were evaluated after the identiﬁcation of critical variables. Taken together, the mathematical approach lacks a theoretical background but seems to provide practically useful predictions. To our best knowledge, however, no study employed machine learning to diﬀerentiate violent from non-violent extremists so far. Here we employed a machine learning approach to diﬀerentiate non-violent but still criminal extremists from those who were engaged in violent attacks. More conventional Multiple Regression models were also studied for comparison. We considered all extremists and all variables available in the PIRUS database without a priori assumption. Citation: (2017)—who addressed the psychological concept of “quest for signiﬁcance”—studied only 38% of the factors that were also studied by LaFree et al. (2018). Naturally, diﬀerent theoretical approaches ask for diﬀerent sets of factors; yet hypothesis-based approaches still provide a fragmented picture of the factors that engender violence. It is also worth to note that the context of the investigation inﬂuences the signiﬁcance of interactions, especially in the case of multivariate analyses, where the role of a particular factor may be made conspicuous or may be masked by the factors investigated concurrently. For instance, marital status inﬂuenced violence in two but failed to inﬂuence it in the third study referred to above. Likewise, criminal history had a signiﬁcant impact on violence in only one of the three studies. The preferentially employed database in this area was the Global Terrorism Database, also developed and maintained by START. The aims of studies was to predict: terrorism risk by analyzing internet communications (Cheong and Lee, 2011; Iskandar, 2017; Pelzer, 2018; Kalaiarasi et al., 2019), the location, timing and/or the type of future attacks (Ding et al., 2017; Verma et al., 2018; Hao et al., 2019; Singh et al., 2019; Uddin et al., 2020), the terrorist group responsible for an attack (Tolan and Soliman, 2015; Talreja et al., 2017; Alfatih et al., 2019), the weapons which may be used in forthcoming attacks (Verma et al., 2018; Narula et al., 2020) and global social determinants of terrorism (Gassebner and Luechinger, 2011). Two studies used the PIRUS database to identify Islamist radicals based on data unrelated to ideological backgrounds (Al-Zewairi and Naymat, 2017) and to predict the use of chemical/biological weapons (Guarrieri and Meisel, 2019). Prediction accuracy was around ∼80%. Algorithms employed include decision trees (e.g., C4.5 algorithm), Deep Learning, Extreme Bound Analysis, Gradient Boosting, Iterative Dichotomizer, K-Nearest Neighbor, Maximum Likelihood Estimation, Naïve Bayes, Neural Network models, Random Forest, Rare Event Logit Model and Support Overall, theory-based approaches that employed the PIRUS database made important contributions to understanding terrorism. They helped understand motivations (Jasko et al., 2017), outlined criminological concomitants May 2022 \| Volume 13 \| Article 745608 Frontiers in Psychology \| www.frontiersin.org 2 Machine Learning and Violent Terrorism Ivaskevics and Haller Vector Machine. Although the identiﬁcation of the “best” machine learning algorithm was evaluated in some of these papers, data remained inconsistent: the most accurate model in one study proved to be less accurate in another. 1http://www.start.umd.edu/pirus Citation: Although a few variables were excluded, this was done exclusively for technical reasons as shown below. This approach deprived our analysis of the solid theoretical framework of hypothesis-driven approaches but oﬀered the possibility of comparing the predictive value of variables that were usually not studied together as shown above. Similarly, all extremists were included in the analysis irrespective of their ideological background, which allowed the direct comparison of ideologies regarding their potential to generate violence. The primary aim of the study was to identify those variables and individual characteristics that predicted the emergence of violence in extremists, and to evaluate their predictive power. The secondary aim was to evaluate the relative importance of individual characteristics, by producing a rank order, which may help diﬀerentiate more from less important but still predictive characteristics. Finally, we aimed at generating a matrix of proviolence and antiviolence characteristics, which may be used in the future for risk assessment purposes. Ideally, one could ﬁnd the place of an extremist within the matrix based on his/her characteristics and predict violence risk based on this place. The Database Investigated Group Nature (“super-group” 2) Group details Group_membership (0) Not member (1) Informal group (2) Formal extremist group (3) Above-ground group Terrorist_Group_Name Not outcome measure Recruitment details Actively_Recruited (0) No (1) Yes Recruiter Not outcome measure Actively_Connect (0) No (1) Prior to (2) After radical behaviors Group activities and dynamics Group_Competition (0) No (1) Yes Role_Group (0) Loose Associate (1) Follower (2) Leader Length_Group Month in group Clique (0) No (1) Yes Clique_Radicalize Not outcome measures Clique_Connect Radicalization (“super-group” 3) Internet and media Internet_Radicalization (0) No (1) Some (2) Primary importance Media_Radicalization Social_Media Social_Media_Frequency Not outcome measures Social_Media_Platform Social_Media_Activities Radicalization ideology Radicalization_Islamist (0) No (1) Yes Radicalization_Far_Right Radicalization_Far_Left Radicalization_Single_Issue Ideological_Sub _Category (1) Militia/gun rights (2) White supremacist (3) Xenophobic (4) Anti-government (5) Christian Identity (6) Animal rights/Environmentalist (7) New Left (8) Black Nationalist (9) Anti-capitalist Anarchist (11) Islamist (12) Puerto Rican nationalist (13) Irish Republican Army (14) Cult/idiosyncratic (15) Anti-abortion (16) Jewish Defense League (17) Anti-gay (18) Other (19) Male supremacist Radicalization location and timing Loc_Habitation Not outcome measure Itinerant (0) No (1) Yes External_Rad Rad_Duration Month radicalized Extent of radicalization Radical_Behaviors (0) No (1) Associates with radicals (2) Changing lifestyle (3) Converting others (4) Distancing fr past relationships (5) Legal activism (6) Material/ﬁnancial support (7) Logistical support (8) Se training [(9) Active participation in plots (10) Active participation in violent plots] Radical_Beliefs (0) No evidence (1) Exposure to (2) Pursues information (3) Full knowledge of tenets (4) Share beliefs (5) Deep commitment Radicalizing events US_Govt_Leader (0) No (1) Yes Foreign_Govt_Leader Event_Inﬂuence (0) None (1) September 11 attacks (2) Vietnam War (3) Cold War (4) First Gulf War (5) Afghanistan/Iraq War (6) Ruby Ridge (7) Arab Spring (8) Other Radicalization process Beliefs_Trajectory (1) Gradual (2) Key moments Behaviors_Trajectory Radicalization _Sequence (1) Beliefs preceded radical behaviors (2) Beliefs followed radical behaviors (3) Concomitant TABLE 1 \| The structure of, and terrorist characteristics in, the PIRUS 2018 Database. The Database Investigated Data were drawn from the public release version of the PIRUS 2018 database (START, 2018). This is freely available at https://www.start.umd.edu/proﬁles-individual-radicalization- united-states-pirus-keshif, and was created by 15 members of the PIRUS group listed on the PIRUS homepage.1 The database was based on open-sources and extant START research products. The preliminary list contained 4,000 individuals from which the ﬁnal list of 2,148 was created by the evaluation of inclusion requirements (see below) by full-time project researchers, or trained research assistants whose evaluations The remaining 111 variables were evaluated as predictors. Out of these, 32 were excluded for either of the following reasons: (1) The variable was directly indicative of violence. For instance, all variables included in the “Plot and Consequences” section described the plot, i.e., directly indicated whether the individual was violent; (2) data were missing for ≥85% of the sample. May 2022 \| Volume 13 \| Article 745608 Frontiers in Psychology \| www.frontiersin.org 3 Machine Learning and Violent Terrorism Ivaskevics and Haller TABLE 1 \| The structure of, and terrorist characteristics in, the PIRUS 2018 Database. Variable groups Variables (“ﬁelds”) Terrorist characteristics (values) Plot and Consequences (“super-group” 1) This super-group of variables were directly indicative of the dependent variable “Violence”; consequently, it was not considered in analysis and was omitted from this table. The Database Investigated of children Age_Child Not outcome measure Gender (1) Female (2) Male Religious background Religious_ Background (1) Sunni (2) Shi’a (3) Suﬁ(4) Other (5) Unspeciﬁed Islam (6) Evangelical Protestant (7) Mainline Protestant (8) Catholic (9) Orthodox (10) Other (11) Unspeciﬁed Christianity (12) Jewish (13) Buddhist (14) Hindu (15) New religion (16) Agnostic (17) Atheist (18) Other Convert (0) No (1) Prior to (2) During (3) After radicalization Convert_Date Not outcome measure Reawakening (0) No (1) Prior to (2) During (3) After radicalization Reawakening_date Not outcome measure Citizenship history Citizenship Not outcome measure Residency_Status (1) Born Citizen (2) Naturalized Citizen (3) Permanent Resident (4) Temporary Resident (5) Undocumented resident Nativity Not outcome measure Time_US_Months Month in USE Immigrant_Generation (0) 3 + (1) First (2) Second Immigrant_Source Not outcome measure Ties to society Language_English (0) No (1) Yes Diaspora_Ties (0) None (1) Weak (2) Strong Socioeconomic Status (“super-group” 5) Education Education (1) No high school (2) Some High school (3) High school (4) Some College (5) College degree (6) Some vocational school (7) Vocational school degree (8) Some Master’s school (9) Master’s degree 1(0) Some Doctoral/Professional training (11) Doctoral/Professional degree Student (0) No (1) Yes Education_Change Not outcome variable Finances and employment Employment_Status (1) Employed (2) Self-employed (3) Unemployed, seeking work (4) Unemployed, not seeking work (5) Student (6) Retired Change_Performance (0) No (1) Yes Work_History (1) Long-term Unemployed (2) Underemployed (3) Serially Employed (4) Regularly Employed Military Military (0) No (1) Inactive, unknown deployment (2) Inactive, never deployed (3) Inactive, previously deployed (4) Active, unknown deployment (5) Active, never deployed (6) Active, deployed Foreign_Military (0) No (1) Yes Socioeconomic stratum Social_Stratum_Childhood (1) Low (2) Middle (3) High Social_Stratum_Adulthood Aspirations (0) No (1) Yes, no attempt to achieve (2) failed to achieve (3) achieved prior to radicalization Personal (“super-group” 6) Abuse and psychological concerns Abuse_Child (0) No (1) By non-family (2) Family (3) Both Abuse_Adult Abuse_Type Not outcome variable Psychological (0) No (1) Speculation (2) Diagnosed Alcohol_Drug (0) No (1) Yes Family and Absent_Parent (0) No (1) Mother (2) Father (3) Both TABLE 1 \| (Continued) Variable groups Variables (“ﬁelds”) Terrorist characteristics (values) Demographics (“super-group” 4) General details Broad_Ethnicity (1) Hispanic/Latino (2) Black/African-American (3) White (4) Middle Eastern/North African (5) Native American (6) Asian (7) Other Age Years of age Marital_Status (1) Single (2) Married (3) Divorced or Separated (4) Widowed Children No. The Database Investigated (1) Militia/gun rights (2) White supremacist (3) Xenophobic (4) Anti-government (5) Christian Identity (6) Animal rights/Environmentalist (7) New Left (8) Black Nationalist (9) Anti-capitalist (10) Anarchist (11) Islamist (12) Puerto Rican nationalist (13) Irish Republican Army (14) Cult/idiosyncratic (15) Anti-abortion (16) Jewish Defense League (17) Anti-gay (18) Other (19) Male supremacist Radicalization location and timing Loc_Habitation Itinerant External_Rad Rad_Duration Extent of radicalization Radical_Behaviors Radical_Beliefs Radicalizing events US_Govt_Leader Foreign_Govt_Leader Event_Inﬂuence Radicalization process Beliefs_Trajectory Behaviors_Trajectory Radicalization _Sequence Radicalizing sites Radicalization_Place Prison_Radicalize (0) No (1) Associates with radicals (2) Changing lifestyle (3) Converting others (4) Distancing from past relationships (5) Legal activism (6) Material/ﬁnancial support (7) Logistical support (8) Seeks training [(9) Active participation in plots (10) Active participation in violent plots] (0) No evidence (1) Exposure to (2) Pursues information (3) Full knowledge of tenets (4) Shares beliefs (5) Deep commitment (1) Beliefs preceded radical behaviors (2) Beliefs followed radical behaviors (3) Concomitant (0) No radicalization (1) Place of worship (2) Educational institution (3) Social club (0) Full radicalization before prison (1) Increased in prison (2) Maximum after prison (3) Maximum in prison (Continued) May 2022 \| Volume 13 \| Article 745608 Frontiers in Psychology \| www.frontiersin.org Machine Learning and Violent Terrorism Ivaskevics and Haller TABLE 1 \| (Continued) Variable groups Variables (“ﬁelds”) Demographics (“super-group” 4) General details Broad_Ethnicity Age Marital_Status Children Age_Child Gender Religious background Religious_ Background Convert Convert_Date Reawakening Reawakening_date Citizenship history Citizenship Residency_Status Nativity Time_US_Months Immigrant_Generation Immigrant_Source Ties to society Language_English Diaspora_Ties Socioeconomic Status (“super-group” 5) Education Education Student Education_Change Finances and employment Employment_Status Change_Performance Work_History Military Military Foreign_Military Socioeconomic stratum Social_Stratum_Childhood Social_Stratum_Adulthood Aspirations Personal (“super-group” 6) Abuse and psychological concerns Abuse_Child Abuse_Adult Abuse_Type Psychological Alcohol_Drug Family and relationships Absent_Parent Overseas_Family Close_Family TABLE 1 \| (Continued) Variable groups Variables (“ﬁelds”) Terrorist characteristics (values) Demographics (“super-group” 4) General details Broad_Ethnicity (1) Hispanic/Latino (2) Black/African-American (3) White (4) Middle Eastern/North African (5) Native American (6) Asian (7) Other Age Years of age Marital_Status (1) Single (2) Married (3) Divorced or Separated (4) Widowed Children No. The Database Investigated Problematic social relations were depicted as platonic troubles in the database. Variables excluded from analysis were shown in small font. For other explanations see text. (1) Homicide (2) Rape (3) Robbery (4) Aggravated Assault (5) Burglary (6) Larceny-Theft (7) Motor Vehicle Theft (8) Arson (9) Simple Assault (10) Fraud (11) Forgery (12) Embezzlement (13) Driving under inﬂuence (14) Prostitution (15) Vandalism (16) Drug related (17) Parole violation (18) Firearm violation (19) Domestic violence (20) Other (0) No (1) timing vis-à-vis radicalization unknown (2) Long before (3) Shortly before (0) No (1) Yes, no personal connection (2) Personal connection (3) Direct experience (0) No (1) Timing vis-à-vis radicalization unknown (2) Long before (3) Shortly before Variables (“Field Names”) were shown as in the database; characteristics (values) were abbreviated to ﬁt table. Problematic social relations were depicted as platonic troubles in the database. Variables excluded from analysis were shown in small font. For other explanations see text. Variables (“Field Names”) were shown as in the database; characteristics (values) were abbreviated to ﬁt table. Problematic social relations were depicted as platonic troubles in the database. Variables excluded from analysis were shown in small font. For other explanations see text. Variables (“Field Names”) were shown as in the database; characteristics (values) were abbreviated to ﬁt table. Problematic social relations were depicted as platonic troubles in the database. Variables excluded from analysis were shown in small font. For other explanations see text. as in the database; characteristics (values) were abbreviated to ﬁt table. Problematic social relations were depicted as platonic luded from analysis were shown in small font. For other explanations see text. Variables (“Field Names”) were shown as in the database; characteristics (values) were abbreviated to ﬁt table. Problematic soc troubles in the database. Variables excluded from analysis were shown in small font. For other explanations see text. Missing data with the remaining 79 variables were handled by two diﬀerent methodologies as shown below. When all the variables were considered together, Multiple Regression could not be performed because the number of variables was too large for the sample size. Therefore, a three- step approach was adopted (Figure 1A). In Step 1, Multiple Regression was run separately on each group of variables contained in the database (see Table 1 for variable groups). The Database Investigated of children Age_Child Not outcome measure Gender (1) Female (2) Male Religious background Religious_ Background (1) Sunni (2) Shi’a (3) Suﬁ(4) Other (5) Unspeciﬁed Islam (6) Evangelical Protestant (7) Mainline Protestant (8) Catholic (9) Orthodox (10) Other (11) Unspeciﬁed Christianity (12) Jewish (13) Buddhist (14) Hindu (15) New religion (16) Agnostic (17) Atheist (18) Other Convert (0) No (1) Prior to (2) During (3) After radicalization Convert_Date Not outcome measure Reawakening (0) No (1) Prior to (2) During (3) After radicalization Reawakening_date Not outcome measure Citizenship history Citizenship Not outcome measure Residency_Status (1) Born Citizen (2) Naturalized Citizen (3) Permanent Resident (4) Temporary Resident (5) Undocumented resident Nativity Not outcome measure Time_US_Months Month in USE Immigrant_Generation (0) 3 + (1) First (2) Second Immigrant_Source Not outcome measure Ties to society Language_English (0) No (1) Yes Diaspora_Ties (0) None (1) Weak (2) Strong Socioeconomic Status (“super-group” 5) (1) Hispanic/Latino (2) Black/African-American (3) White (4) Middle Eastern/North African (5) Native American (6) Asian (7) Other (1) No high school (2) Some High school (3) High school (4) Some College (5) College degree (6) Some vocational school (7) Vocational school degree (8) Some Master’s school (9) Master’s degree 1(0) Some Doctoral/Professional training (11) Doctoral/Professional degree (0) No (1) Yes, no attempt to achieve (2) failed to achieve (3) achieved prior to radicalization (0) No (1) By non-family (2) Family (3) Both (0) No (1) Yes, no attempt to achieve (2) failed to achieve (3) achieved prior to radicalization Not outcome variable (0) No (1) Speculation (2) Diagnosed (0) No (1) Yes (0) No (1) Mother (2) Father (3) Both (0) No, (1) Yes (1) Distant (2) Close (Continued) May 2022 \| Volume 13 \| Article 745608 May 2022 \| Volume 13 \| Article 745608 Frontiers in Psychology \| www.frontiersin.org Machine Learning and Violent Terrorism Ivaskevics and Haller TABLE 1 \| (Continued) Variable groups Variables (“ﬁelds”) Terrorist characteristics (values) Family_Religiosity (0) Secular (1) Somewhat (2) Very Religious Family_Ideology (0) None (1) Islamist (2) Far right (3) Far left (4) Other (5) Single-Issue Family_Ideological_Level Not outcome variable Prison_Family_Friend (0) No (1) Yes Crime_Family_Friend (0) No (1) Victim (2) Perpetrator (3) both Radical_Friend (0) No (1) Yes, only legal activities (2) Non-violent illegal activities (3) Extremist violence Radical_Family Radical_Signif_Othe Relationship_Troubles (0) No (1) Yes Problematic Social Relations Unstructured_Time Friendship_Source (1) School (2) Work (3) Family (4) Religious group (5) Social Organization (6) Other Kicked_Out (marginalized) (0) No (1) Yes Criminal activity Previous_Criminal _Activity (0) No (1) Non-violent, minor (2) Non-violent, serious (3) Violent crime Previous_Criminal _Activity_Type (1) Homicide (2) Rape (3) Robbery (4) Aggravated Assault (5) Burglary (6) Larceny-Theft (7) Motor Vehicle Theft (8) Arson (9) Simple Assault (10) Fraud (11) Forgery (12) Embezzlement (13) Driving under inﬂuence (14) Prostitution (15) Vandalism (16) Drug related (17) Parole violation (18) Firearm violation (19) Domestic violence (20) Other Previous_Criminal_ Activity_Age Not outcome variable Gang (0) None (1) Street (2) Organized (3) Both Gang_Age_Joined Not outcome variable Mindset prior to radicalization Trauma (0) No (1) timing vis-à-vis radicalization unknown (2) Long before (3) Shortly before Other_Ideologies (prior radicalization) (0) No (1) Yes Angry_US Group_Grievance (0) No (1) Yes, no personal connection (2) Personal connection (3) Direct experience Standing (diminution) (0) No (1) Timing vis-à-vis radicalization unknown (2) Long before (3) Shortly before Variables (“Field Names”) were shown as in the database; characteristics (values) were abbreviated to ﬁt table. The Database Investigated Step 2 involved the repeating of the Multiple Regression analysis with those variables, which had a signiﬁcant individual contribution to predictions in Step 1. In Step 3, Multiple Regression analysis was run with those variables, which did not have a signiﬁcant contribution in Step 1. This analysis checked for the accuracy of variable selection in Step 1. Finally, the same analyses were performed after replacing missing data by mean substitution, a way oﬀered by the Statistica software to address missing data. Missing data with the remaining 79 variables were handled by two diﬀerent methodologies as shown below. Frontiers in Psychology \| www.frontiersin.org Statistical Strategy Multiple Regression and Bivariate Analyses FIGURE 1 \| Strategy employed in Multiple Regression and machine learning analyses. (A) The stepwise approach employed in Multiple Regression. Note that the sample size did not allow the evaluation of all 79 variables in one single analysis, which explains the stepwise approach. [. . .], variables not shown. (B) The partition of the database into training and testing data subsets with the XGBoost algorithm. The small squares represent the extremists included in the PIRUS database; the larger rectangles demarcate data subsets. White, training subset; gray, testing subset; [. . .], data subsets not shown. below 1 indicated that the characteristic was more and less frequent, respectively, in violent as compared to non-violent individuals. Signiﬁcance level was established at p < 0.05. Bonferroni adjustment was used to correct for errors deriving from multiple comparisons. To account for the initial choice of the 10 partitions, the whole process was repeated 10 times by rotating the training and test datasets (Figure 1B). This procedure uses all observations for both training and testing and uses each for testing just once. The individual contribution of predictor variables to the models was evaluated by their permutation importance (Strobl et al., 2008). This consists of calculating predictions after permutating the values of predictor variables one by one. The importance of a variable is inversely proportional to the decrease of prediction accuracy after the shuﬄing of its values (Molnar, 2019). Statistical Strategy Multiple Regression and Bivariate Analyses p g y Multiple Regression analysis was performed by means of the Statistica software (TIBCO Software Inc, Palo Alto CA, United States). This extends regression to predict the value of the dependent variable based on the values of several predictor variables. Key measures include the coeﬃcient of multiple correlation (Multiple R), F and p values indicative of the signiﬁcance of prediction and the adjusted R-square. This indicates that percentage of variance in the dependent variable, which was explained by the independent variables. The latter is called prediction in statistical terms. The individual contribution of predictor variables to overall predictions is indicated by individual regression coeﬃcients (β) and their t-statistics. The overall prediction by independent variables is usually attributed to those that have signiﬁcant individual contributions. To clarify the impact of individual characteristics covered by the variables, a bivariate post hoc comparison was performed (non-violent vs. violent). This was made by crosstabulation. To evaluate the extent of group diﬀerences, risk ratios were also calculated; this was the ratio of the prevalence of a characteristic in violent and non-violent individuals. Risk ratios above and May 2022 \| Volume 13 \| Article 745608 Frontiers in Psychology \| www.frontiersin.org Frontiers in Psychology \| www.frontiersin.org 6 Machine Learning and Violent Terrorism Ivaskevics and Haller FIGURE 1 \| Strategy employed in Multiple Regression and machine learning analyses. (A) The stepwise approach employed in Multiple Regression. Note that the sample size did not allow the evaluation of all 79 variables in one single analysis, which explains the stepwise approach. [. . .], variables not shown. (B) The partition of the database into training and testing data subsets with the XGBoost algorithm. The small squares represent the extremists included in the PIRUS database; the larger rectangles demarcate data subsets. White, training subset; gray, testing subset; [. . .], data subsets not shown. FIGURE 1 \| Strategy employed in Multiple Regression and machine learning analyses. (A) The stepwise approach employed in Multiple Regression. Note that the sample size did not allow the evaluation of all 79 variables in one single analysis, which explains the stepwise approach. [. . .], variables not shown. (B) The partition of the database into training and testing data subsets with the XGBoost algorithm. The small squares represent the extremists included in the PIRUS database; the larger rectangles demarcate data subsets. White, training subset; gray, testing subset; [. . .], data subsets not shown. Frontiers in Psychology \| www.frontiersin.org Machine Learning and Bivariate Analyses The open-source ensemble machine learning algorithm XGBoost (eXtreme Gradient Boosting, 47) was used in the Python 3.6 programming interface. This was chosen for its outstanding performance in a wide range of classiﬁcation and regression problems in various disciplines, and successful applications in scientiﬁc studies (Torlay et al., 2017; Aaboud et al., 2018). Training and testing datasets, i.e., those that enabled the algorithm to learn associations and to test predictions underwent 10 times repeated 10-fold cross-validation (James et al., 2013). This involves the division of the extremists of the PIRUS database into 10 equally sized subsets. In each iteration, nine subsets were used for training, whereas the 10th was used for testing. Albeit XGBoost was able to deal with all variables at the same time, analysis was performed in three steps, because during the optimization process (“boosting”) the algorithm keeps only one of those variables that correlate strongly, e.g., the one that has the highest permutation importance (predictive power). Consequently, the models contained three diﬀerent sets of variables: (i) predictor variables that strongly predicted the dependent variable (permutation importance >1%); (ii) variables May 2022 \| Volume 13 \| Article 745608 Frontiers in Psychology \| www.frontiersin.org 7 Machine Learning and Violent Terrorism Ivaskevics and Haller that were not selected because of their high correlation with the former, but which potentially may have become predictor variables if the former were missing from the database (predictor- like variables), and (iii) non-predictor variables. In Step 1, we allowed the algorithm to work with all the variables. In Step 2, we evaluated predictive power after excluding the predictor variables identiﬁed in Step 1 but keeping predictor-like and non-predictor variables. In Step 3, we included non- predictor variables. This approach answered two questions: (1) Can the most important variables be replaced by predictor-like variables, if they were missing from the database? and (2) Do variables with no permutation importance possess any predictive value? indicating that violence was interactively predicted by variables belonging to the diﬀerent groups of variables (Table 3). In Step 3, Multiple Regression was rerun on those 47 variables that had no signiﬁcant contributions in Step 1. This combination of variables did not explain the variance in the dependent variable (Table 3). Thus, the selection of critical variables in Step 1 was adequate. Machine Learning and Bivariate Analyses The mean substitution of missing data decreased the average prediction power to 1.9% in Step 1, and to 19.1% in Step 2 (Table 3). Thus, the interactions between variables were non- linear; consequently, mean substitution appeared inappropriate to control for missing data in this database. Missing predictors were replaced by using Multiple Imputation with Chained Equations, with the mice package (van Buuren and Groothuis-Oudshoorn, 2010) of the R statistical software (R Development Core Team, 2013). Following the guidelines, diﬀerent imputation methods were speciﬁed for diﬀerent variable types: predictive mean matching (pmm) was used for continuous variables, logistic regression (logreg) for binary variables, polytomous logistic regression (polyreg) for unordered (nominal) categorical variables and proportional odds model (polr) for ordered (ordinal) categorical variables. A total of 20 imputed datasets were created over 20 iterations, which were then analyzed separately, and the results were pooled to obtain an overall estimation. The reported values for area under the curves (AUCs), precisions and recalls therefore represent the mean of each statistic across the 10 imputed datasets. Critical Characteristics—Multiple Regression Pairwise comparisons were performed within those variables that signiﬁcantly predicted violence in Multiple Regression analysis Step 2. Only a subgroup of characteristics showed signiﬁcant group diﬀerences (Figure 2). It also occurs that neither characteristic alone could predict violence. Some were rather frequent, which made them potentially useful, but these had low risk ratios. See, for instance, “none” (Media Radicalization) and “white” (Broad ethnicity) in Figure 2. With other characteristics, risk ratios were high yet their share in the population was too low to be used as individual predictor characteristics. See, for instance, anti-gay (Ideological Subcategory) and Other Christian (Religious background) in Figure 2. Therefore, we investigated in the followings how pro-violence and anti-violence characteristics were combined in individual extremists to see whether such combinations may be used for predictions. Performance measures were the Receiver Operating Characteristic Area Under the Curve (ROC-AUC) which represents the relationship between the probability of ﬁnding a true positive (an actual violent individual that was predicted violent by the model) and a false positive (a non-violent individual that was predicted violent by the model). The AUC limits for chance, acceptable, moderate, good, and excellent discrimination were 0.5–0.6, 0.61–0.7, 0.71–0.8, 0.81–0.9, and 0.91–1.0, respectively (Sjöstedt and Grann, 2002). We also included other metrics, e.g., precision, a measure of classiﬁcation exactness and recall, a measure of classiﬁcation completeness. In this case, 100% indicates perfect precision/recall, whereas 50% represents chance level. Finally, the importance of individual characteristics was also analyzed. This was done as described above. y p The largest number of pro-violence and anti-violence characteristics concomitantly present in one and the same individual was 7 and 5, because alternative characteristics were mutually exclusive within variables (e.g., the individual could show only one of the seven characteristics covered by Broad Ethnicity). Although the maximum number of critical characteristics was already low compared to their relatively large number (25, see Figure 3), most individuals presented with even less critical characteristics namely with one or two. In addition, anti-violence and pro-violence characteristics were often present concomitantly in the same individual. As this resulted in a very high number of individual combinations, a characteristic- focused analysis could not be performed. It occurred, however, that the number of critical characteristics was relevant for predictions as shown below. RESULTS Concerns (Alcohol/Drug abuse, Psychological) 0.132 Finances and Employment (Change Performance, Employment Status, Work History) 0.115 Ties to society (Diaspora Ties) 0.093 Radicalization process (none) 0.094 Military (none) 0.031 Mindset prior to radicalization (none) 0.032 Average variance explained TABLE 2 \| Explanations of variance in violence according to the multiple regression analysis. STEP 1 Analysis by variable groups (see Table 1) Predictor variable groups (Variables with signiﬁcant contributions) Multiple R F statistics Variance explained Family and relationships (Family Ideology, Problematic Social Relations) 0.404 F(13,161) = 2.42; p < 0.01 9.6% Religious background (Convert, Religious Background) 0.322 F(3,319) = 12.31; p < 0.0001 9.5% Extent of radicalization (Radical Behaviors, Radical Beliefs) 0.305 F(2,607) = 31.19; p < 0.0001 9.0% Radicalization ideology (Ideological Sub Category, Radicalization Far Left, Radicalization Far Right, Radicalization Islamist) 0.293 F(4,2143) = 50.51; p < 0.0001 8.4% Group nature (Clique, Length Group, Role Group) 0.250 F(7,625) = 5.96; p < 0.0001 5.2% Socioeconomic stratum (Aspirations) 0.236 F(3,184) = 3.64; p < 0.05 4.1% General details (demographics) (Age, Broad Ethnicity, Gender) 0.201 F(5,1070) = 9.03; p < 0.0001 3.6% Radicalization location and timing (External Radicalization, Radicalization duration) 0.171 F(3,660) = 6.667; p < 0.005 2.5% Criminal Activity (Criminal Activity) 0.166 F(3,1222) = 11.48; p < 0.0001 2.5% Citizenship history (Immigrant Generation, Time US Months) 0.153 F(3,1630) = 13.04; p < 0.0001 2.1% Internet and Media (Media Radicalization) 0.161 F(3,514) = 4.592; p < 0.005 2.0% Radicalizing events (Event Inﬂuence, Foreign Government Leader) 0.154 F(3,856) = 6.93; p < 0.001 2.0% Education (Education) 0.150 F(2,708) = 8.18; p < 0001 1.9% Radicalizing sites (Prison Radicalize) 0.137 F(2,566) = 5.42; p < 0.005 1.5% Abuse and Psychol. Concerns (Alcohol/Drug abuse, Psychological) 0.132 F(4,2143) = 9.50; p < 0.0001 1.5% Finances and Employment (Change Performance, Employment Status, Work History) 0.115 F(3,2145) = 9.64; p < 0.0001 1.1% Ties to society (Diaspora Ties) 0.093 F(2,1659) = 7.34; p < 0.001 0.7% Radicalization process (none) 0.094 F(3,730) = 2.13; p < 0.1 None Military (none) 0.031 F(2,1427) = 0.726; p < 0.5 None Mindset prior to radicalization (none) 0.032 F(5,304) = 1.063; p < 0.4 None Average variance explained 3.95% Average variance explained of the 9 that were identiﬁed by Multiple Regression (Table 5). Moreover, most variables indicated as critical by the latter had high permutation importance according to machine learning. RESULTS Nevertheless, XGBoost analysis revealed the importance of 12 variables that were not identiﬁed by Multiple Regression. one third of the cells, the prediction of violence was poor, i.e., was between 34 and 65%. The weighted mean of all these predictions was 61.2%, which was in line with the 66.7% explanation of variance by Multiple-Regression analysis of variables. RESULTS When no anti-violence characteristic was present in the individual (N = 384), the number of pro-violence characteristics predicted violence well (Figure 3A). Likewise, when anti- violence characteristics were unaccompanied by pro-violence ones (N = 227), the chance of violence was very low (Figure 3B). In most individuals, however, both categories of characteristics were present (N = 1137), and the share of violent individuals depended on the relationship between anti- and proviolence characteristics (Figure 3C). In about one third of the sample the analysis correctly identiﬁed violent individuals; in numerically the same number of cells non-violent individuals were identiﬁed correctly (in this case the chance of violence was low). In about Critical Variables—Multiple Regression In Step 1, we studied separately the predictive power of independent variables within variable groups. Out of the 22 such groups (see Table 1), 17 predicted violence signiﬁcantly (Table 2). Out of the 79 variables belonging to these 17 groups, 32 variables had signiﬁcant individual contributions to prediction (Table 2). Albeit predictions were signiﬁcant, they explained a low share of variance in violence (∼4% on average). In Step 2, a Multiple Regression analysis was run on those 32 variables that had signiﬁcant contributions in Step 1. For these variables, prediction power increased by one order of magnitude, May 2022 \| Volume 13 \| Article 745608 Frontiers in Psychology \| www.frontiersin.org 8 Machine Learning and Violent Terrorism Ivaskevics and Haller TABLE 2 \| Explanations of variance in violence according to the multiple regression analysis. STEP 1 Analysis by variable groups (see Table 1) Predictor variable groups (Variables with signiﬁcant contributions) Multiple R Family and relationships (Family Ideology, Problematic Social Relations) 0.404 Religious background (Convert, Religious Background) 0.322 Extent of radicalization (Radical Behaviors, Radical Beliefs) 0.305 Radicalization ideology (Ideological Sub Category, Radicalization Far Left, Radicalization Far Right, Radicalization Islamist) 0.293 Group nature (Clique, Length Group, Role Group) 0.250 Socioeconomic stratum (Aspirations) 0.236 General details (demographics) (Age, Broad Ethnicity, Gender) 0.201 Radicalization location and timing (External Radicalization, Radicalization duration) 0.171 Criminal Activity (Criminal Activity) 0.166 Citizenship history (Immigrant Generation, Time US Months) 0.153 Internet and Media (Media Radicalization) 0.161 Radicalizing events (Event Inﬂuence, Foreign Government Leader) 0.154 Education (Education) 0.150 Radicalizing sites (Prison Radicalize) 0.137 Abuse and Psychol. Critical Variables—Machine Learning Prediction slightly improved when the analysis was restricted to the 19 variables with permutation importance larger than 1%, suggesting that the removal of highly redundant variables increased the predictive accuracy of the model. In both cases, the recall of the models was considerably higher compared to precision, suggesting that the model predicted fewer false The baseline XGBoost model that considered all the variables (Step 1) provided a ﬁnal model that diﬀerentiated violent and non-violent extremists with good discrimination accuracy (Table 4). The 19 most important variables, i.e., those that had permutation importance larger than 1% contained 7 out May 2022 \| Volume 13 \| Article 745608 Frontiers in Psychology \| www.frontiersin.org 9 Machine Learning and Violent Terrorism Ivaskevics and Haller TABLE 3 \| Multiple Regression analysis of variables having or not having signiﬁcant contributions to prediction in Step 1. TABLE 3 \| Multiple Regression analysis of variables having or not having signiﬁcant contributions to prediction in Step 1. Original data Step 2 Analyzed dataset: Signiﬁcant contributors in Step 1 Step 3 Non-signiﬁcant contributors in Step 1 Multiple R = 0.841 F(10,102) = 24.61; p < 0.0001 Variance explained: 67.8% Signiﬁcant contributor variables in Step 2 Broad Ethnicity, Family Ideology, Ideological Sub Category, Media Radicalization, Previous Criminal Activity, Radical Behaviors, Radicalization Far Left, Radicalization Far Right, Religious Background Multiple R = 0.342 F(32,165) = 0.687; p < 0.9 Variance explained: 0.0% Signiﬁcant contributors in Step 3 None After mean substitution Step 2 Analyzed dataset: Signiﬁcant contributors in Step 1 Step 3 Non-signiﬁcant contributors in Step 1 Multiple R = 0.454 F(22,2107) = 13.36 p < 0.0001 Variance explained: 19.1% Signiﬁcant contributor variables in Step 2 Broad Ethnicity, Family Ideology, Ideological Sub Category, Media Radicalization, Previous Criminal Activity, Radical Behaviors, Radicalization Far Left, Radicalization Far Right, Religious Background Multiple R = 0.212 F (34,2114) = 0.81; p < 0.8 Variance explained: 0.0% Signiﬁcant contributor variables None Note that Multiple Regression was run by activating the pairwise deletion and forward stepwise modes of function of the Statistica module, to minimize the impact of missing data. the same individual possessed only a few of these. The largest number of coexisting pro- and anti-violence characteristics was 14 and 12, respectively, but such high numbers were present in one individual for each the violent and non- violent category. Typically, individuals expressed 4–5 pro- and/or anti-violence characteristics concomitantly. Main Findings We identiﬁed an array of critical variables that allowed the prediction of violence with an accuracy of over 85%. Machine learning performed better in this respect than the more conventional regression model. Within each critical variable, certain characteristics predicted violence and others predicted non-violence among radicalized individuals. Within the variable “Ideological Subcategory” for instance, the “Anti- gay” characteristic was the second strongest predictor of violence, whereas the “Animal Rights and Environmentalist” characteristic was the strongest predictor of non-violence. Most extremists displayed both pro- and antiviolence characteristics. These were displayed by extremists in many distinct combinations, which precluded their detailed analysis at this time. Yet the ratio of proviolent/antiviolent characteristics emerged as a strong predictor of violence and allowed the elaboration of a risk matrix that may be used to predict individual violence risk provided that enough critical characteristics are known for the individual. The characteristics of violent and non-violent Critical Variables—Machine Learning As the number of diﬀerent combinations of characteristics was even larger than for the characteristics identiﬁed by Multiple Regression, we again used the number of characteristics to evaluate their predictive value at individual level. If one of the characteristic types (e.g., pro- and antiviolence) was missing from the individual, the other characteristic type predicted violence with rather high precision (Figures 5A,B). However, such individuals were rare. Sample size for individuals with no anti-violence characteristics was 135, whereas for individuals without pro-violence characteristics was 26. Consequently, the prediction of violence depended on the interaction between pro- and anti-violence characteristics in 92% of the sample. The representation of interactions between pro-violence characteristics, anti-violence characteristics, and the proportion of violent individuals in the cells of the interaction matrix (Figure 5C) was diﬀerent from that seen with the characteristics identiﬁed by multiple Regression in two respects: (1) the pro- and anti-violence interaction matrix was considerably larger due to the larger number of critical characteristics; (2) predictions were more polarized. In most cases, the interaction reliably predicted the categorization of individuals into non-violent and violent groups (Figure 5C). Cells with low prediction power, i.e., where predicted group assignment was correct in 34–65% of cases only, were rare. Such indecisive predictions were seen in 14.7% of the cells. This was in sharp contrast with the characteristic matrix that was derived from Multiple Regression, where around one third of the cells provided indecisive predictions. negatives (non-violent extremists who were in fact violent) than false positives (non-violent extremists predicted as violent). When only predictor-like and non-predictor variables were evaluated, all three, prediction power, precision and recall considerably decreased to acceptable/moderate performance. Finally, the elimination of both predictor and predictor-like variables reduced prediction accuracy considerably. DISCUSSION Note that data imputation did not aﬀect performance notably (Table 4), which on one hand shows that the imputation method employed was able to correctly represent the non- linearity of interactions between variables but on the other hand it also shows that missing data did not aﬀect prediction accuracies signiﬁcantly. Frontiers in Psychology \| www.frontiersin.org Critical Characteristics—Machine Learning The 19 most important variables covered in total 113 characteristics, out of which 53 showed signiﬁcant group diﬀerences (χ2 was between 5.85 and 180.62; corresponding p values were between 0.015 and 0.00001). About half of the characteristics increased whereas the other half decreased the risk of violence (Figure 4). Risk ratios were high with relatively rare characteristics, whereas risk ratios were low with frequent characteristics. Consequently, none of the critical characteristics on its own was able to predict violence. Although the number of critical characteristics was relatively large (27 pro- and 25 anti-violence characteristics) one and May 2022 \| Volume 13 \| Article 745608 Frontiers in Psychology \| www.frontiersin.org 10 Machine Learning and Violent Terrorism Ivaskevics and Haller FIGURE 2 \| The share of various characteristics in non-violent and violent individuals of the PIRUS database. Characteristics were selected based on Multiple Regression analysis. The violent and non-violent groups were different in all cases after Bonferroni adjustment for repeated comparisons (p < 0.05 at least). Characteristic names (as shown in the database) were followed by the abbreviation of the variable where the characteristic belonged. BE, Broad ethnicity; FI, Family FIGURE 2 \| The share of various characteristics in non-violent and violent individuals of the PIRUS database. Characteristics were selected based on Multiple Regression analysis. The violent and non-violent groups were different in all cases after Bonferroni adjustment for repeated comparisons (p < 0.05 at least). Characteristic names (as shown in the database) were followed by the abbreviation of the variable where the characteristic belonged. BE, Broad ethnicity; FI, Family Ideology; ISC, Ideological Sub-Category; MR, Media Radicalization; PCA, Previous Criminal Activity; RadB, Radical Behaviors; RFL, Radicalization Far Left; RFR, Radicalization Far Right; RelB, Religious Background. Risk ratios were also shown as graphs on a scale of 1–10 (pro-violence characteristics) and 0–1 (anti-violence characteristics). The smallest and largest risk ratios were numerically shown as reference values. FIGURE 2 \| The share of various characteristics in non-violent and violent individuals of the PIRUS database. Characteristics were selected based on Multiple Regression analysis. The violent and non-violent groups were different in all cases after Bonferroni adjustment for repeated comparisons (p < 0.05 at least). Characteristic names (as shown in the database) were followed by the abbreviation of the variable where the characteristic belonged. Critical Characteristics—Machine Learning Predictor variables Variable group Permutation importance (%) Radical Behaviors Radicalization 14.64 Radicalization Far Left Radicalization 5.82 Ideological Sub Category Radicalization 3.58 Broad Ethnicity Demographics 1.25 Radical Friend Personal 0.78 Religious Background Demographics 0.54 Prison Radicalize Radicalization 0.45 Radicalization Islamist Radicalization 0.45 Group Membership Group 0.35 Radical Family Personal 0.35 Role Group Group 0.33 Social Media Radicalization 0.25 Convert Personal 0.18 Previous Criminal Activity Demographics 0.15 Problematic Social Relations Personal 0.14 Radical Beliefs Radicalization 0.13 Clique Personal 0.11 Media Radicalization Radicalization 0.10 Social Stratum Adulthood Demographics 0.10 Higher permutation importance values indicate stronger contribution to the prediction of the model. Underlined and italic, predictor variable also identiﬁed by Multiple Regression. TABLE 4 \| Performance metrics of the XGBoost models with different predictor sets. Original data Predictor dataset ROC-AUC % Precision % Recall % M (SD) M (SD) M (SD) All variables (79) 86.3 (2.3) 78.2 (3.5) 88.4 (3.0) Predictor variables (19) 87.2 (2.2) 78.5 (3.6) 88.2 (3.0) Predictor-like variables (?) 70.9 (3.5) 67.6 (4.3) 78.4 (4.0) Not included in the model (?) 69.1 (3.7) 66.4 (4.4) 78.8 (3.9) After multiple imputation Predictor dataset ROC-AUC % Precision % Recall % M (SD) M (SD) M (SD) All variables (79) 86.9 (3.2) 78.7 (7.5) 88.1 (4.3) Predictor variables (19) 87.2 (3.0) 79.0 (7.3) 87.8 (4.7) Predictor-like variables (?) 67.6 (5.9) 65.8 (10.1) 78.0 (8.1) Not included in the model (?) 66.6 (5.9) 64.8 (11.4) 75.9 (9.6) M, mean; SD, standard deviation; ROC-AUC, Receiver Operating Characteristics – Area Under the Curve. Note ROC-AUC represents the discrimination accuracy of the models with 50% indicating chance level prediction and 100% indicating perfect prediction. Precision refers to the proportion of true violent extremists over the number of extremists who were predicted as violent. Recall refers to the proportion of true violent extremists over all actual violent extremists. TABLE 5 \| Average permutation importance of the 19 most important predictors listed in descending order. TABLE 4 \| Performance metrics of the XGBoost models with different predictor sets. Higher permutation importance values indicate stronger contribution to the prediction of the model. Underlined and italic, predictor variable also identiﬁed by Multiple Regression. Higher permutation importance values indicate stronger contribution to the prediction of the model. Underlined and italic, predictor variable also identiﬁed by Multiple Regression. extremists suggested that the two groups played diﬀerent roles in terrorist attacks, suggesting the existence of work division among criminal extremists. Critical Characteristics—Machine Learning BE, Broad ethnicity; FI, Family Ideology; ISC, Ideological Sub-Category; MR, Media Radicalization; PCA, Previous Criminal Activity; RadB, Radical Behaviors; RFL, Radicalization Far Left; RFR, Radicalization Far Right; RelB, Religious Background. Risk ratios were also shown as graphs on a scale of 1–10 (pro-violence characteristics) and 0–1 (anti-violence characteristics). The smallest and largest risk ratios were numerically shown as reference values. FIGURE 3 \| The prediction of violence by the number of pro- and anti-violence characteristics. (A,B) The dependence of the share of violent individuals on the number of pro-violence (“pro”) and anti-violence (“anti”) characteristics, respectively, when the opposite characteristic was not present in the individual. (C) The interactive dependence of violence on the relation between the number of pro-violence and anti-violence characteristics present in the individual. Gray, the area of chance prediction (34–66%, i.e., close to 50%). FIGURE 3 \| The prediction of violence by the number of pro- and anti-violence characteristics. (A,B) The dependence of the share of violent individuals on the number of pro-violence (“pro”) and anti-violence (“anti”) characteristics, respectively, when the opposite characteristic was not present in the individual. (C) The interactive dependence of violence on the relation between the number of pro-violence and anti-violence characteristics present in the individual. Gray, the area of chance prediction (34–66%, i.e., close to 50%). FIGURE 3 \| The prediction of violence by the number of pro- and anti-violence characteristics. (A,B) The dependence of the share of violent individuals on the number of pro-violence (“pro”) and anti-violence (“anti”) characteristics, respectively, when the opposite characteristic was not present in the individual. (C) The interactive dependence of violence on the relation between the number of pro-violence and anti-violence characteristics present in the individual. Gray, the area of chance prediction (34–66%, i.e., close to 50%). May 2022 \| Volume 13 \| Article 745608 11 Frontiers in Psychology \| www.frontiersin.org Machine Learning and Violent Terrorism Ivaskevics and Haller TABLE 5 \| Average permutation importance of the 19 most important predictors listed in descending order. Critical Characteristics—Machine Learning These studies used the same PIRUS database but preselected variables according to their relevance for the hypothesis under scrutiny. As the hypotheses tested were diﬀerent, the sets of selected variables showed little overlaps. One can hypothesize that their discrepant ﬁndings were primarily due to the constellation of variables under study, because multivariate analyses investigate interactions between variables. For instance, achieving aspirations correlated negatively with violence when considered together with employment (Becker, 2019) but positively when considered together with having a radical family and radical peers (Jasko et al., 2017). One can hypothesize that achieving aspirations in these two contexts results in diﬀerent predictions regarding violence. As such, the predictive value of a variable can be changed by the constellation of variables investigated. This indirectly suggests that the more variables are considered the more equilibrated the result of the analysis becomes. Frontiers in Psychology \| www.frontiersin.org Approaches and Variable Constellations Approaches and Variable Constellations We found ﬁve earlier studies that employed the PIRUS database, which together identiﬁed 8 out of the 19 variables that predicted violence in our study (Jasko et al., 2017; LaFree et al., 2018, 2019; Becker, 2019; Jensen et al., 2020). Other factors recognized by us were indicated by studies on smaller databases (Berrebi, 2007; Altunbas and Thornton, 2011; Bartlett and Miller, 2012; Benmelech and Berrebi, 2007; Krueger, 2008; Meloy and Gill, 2016; Knight et al., 2017). These outlined, for instance, the roles of ethnicity, radical behaviors and beliefs, roles assumed by extremists within groups and economic status. The role of media was also revealed earlier (Cohen et al., 2014; Etudo, 2017; Youngblood, 2020). Taken together, this suggests that our study identiﬁed only 5 new variables and conﬁrmed the importance of 14 (Table 6). Although there were important overlaps in our and earlier studies, we failed to conﬁrm the predictive value of 15 variables that were found predictive in the above mentioned ﬁve studies that employed the PIRUS database. Importantly, the ﬁndings from these earlier studies were inconsistent themselves. For instance, younger age predicted violence in the studies by LaFree et al. (2018) and Jasko et al. (2017) but did not do so in the study by Becker (2019). On their turn, achieving aspirations and employment were negatively associated with violence in some studies (LaFree et al., 2018; Becker, 2019), but positively in others (Jasko et al., 2017). Similar examples are abundant. Primary and Secondary Typologies Primary and Secondary Typologies The development of extremist typologies is essential for counterterrorism, and several systems were developed to address this issue (Borum, 2003; Moghaddam, 2005; McCauley and Moskalenko, 2008; Aghabi et al., 2017). These were to a large extent based on the individual characteristics of terrorists. For instance, eight types were identiﬁed in a sample of 56 terrorists by studying 10 factors/variables (Jensen et al., 2020). None of the types presented with all the 10 factors; they were characterized by their unique combination. Not surprisingly, in May 2022 \| Volume 13 \| Article 745608 Frontiers in Psychology \| www.frontiersin.org 12 Machine Learning and Violent Terrorism Ivaskevics and Haller FIGURE 4 \| Critical individual characteristics identiﬁed by machine learning and their share in the population. The names of variables were abbreviated (bold); those of characteristics were shortened to ﬁt ﬁgure. The share of the characteristic was shown for the non-violent group. Abbreviations for variables. BE, Broad ethnicity; C, Clique; Con, Convert; FI, Family Ideology; GM, Group Membership; ISC, Ideological Sub Category; MR, Media Radicalization; PCA, Previous Criminal Activity; PR, Prison Radicalize; PSR, Problematic Social Relations; RadB, Radical Behaviors; RB, Radical Beliefs; RBeh, Radical Behaviors; RelB, Religious Background; RFam, Radical Family; RFL, Radicalization Far Left; RFr, Radical friend; RFR, Radicalization Far Right; RG, Role Group; RI, Radicalization Islamist; SM, Social Media; SSA, Social Stratum Adulthood. FIGURE 4 \| Critical individual characteristics identiﬁed by machine learning and their share in the population. The names of variables were abbreviated (bold); those of characteristics were shortened to ﬁt ﬁgure. The share of the characteristic was shown for the non-violent group. Abbreviations for variables. BE, Broad ethnicity; C, Clique; Con, Convert; FI, Family Ideology; GM, Group Membership; ISC, Ideological Sub Category; MR, Media Radicalization; PCA, Previous Criminal Activity; PR, Prison Radicalize; PSR, Problematic Social Relations; RadB, Radical Behaviors; RB, Radical Beliefs; RBeh, Radical Behaviors; RelB, Religious Background; RFam, Radical Family; RFL, Radicalization Far Left; RFr, Radical friend; RFR, Radicalization Far Right; RG, Role Group; RI, Radicalization Islamist; SM, Social Media; SSA, Social Stratum Adulthood. our much larger sample and with a considerably larger number of factors considered, the number of characteristic-combinations was much larger, which prevented the development of a detailed typology. Nevertheless, our analysis did delineate two major types of extremists. to low social strata and had problematic social relations. These extremists may be termed executive/violent. Frontiers in Psychology \| www.frontiersin.org The Risk Matrix Note that sample sizes were low; symbols in gray indicate N = 1 for the given “pro”/”anti” characteristic combination. (C) The interactive dependence of violence on the relation between the number of pro-violence and anti-violence characteristics present in the individual. Gray in panels (A–C), the area of chance prediction (34–65%, i.e., close to 50%). FIGURE 5 \| Violence prediction by the number of pro- and anti-violence characteristics identiﬁed by machine learning. (A,B) When the opposite characteristic was not present, pro-violence (“pro”) and anti-violence (“anti”) characteristics predicted violence with high accuracy. Note that sample sizes were low; symbols in gray indicate N = 1 for the given “pro”/”anti” characteristic combination. (C) The interactive dependence of violence on the relation between the number of pro-violence and anti-violence characteristics present in the individual. Gray in panels (A–C), the area of chance prediction (34–65%, i.e., close to 50%). to violent extremists fully, or mainly radicalized in prison, violent extremists who had no prison sentences or were fully radicalized before prison were less likely to be United States citizens (born or naturalized), had higher levels of education but poorer employment histories, belonged to higher social strata, and committed lesser crimes before radicalization. The same signiﬁcant predictor of violence, particularly radicalization in prisons, also diﬀerentiated subgroups among non-violent extremists, but these subgroups were diﬀerentiated by a diﬀerent set of variables. As compared to non-violent extremists radicalized in prison, those who were fully radicalized before prison were more likely to be students, had better employment histories, belonged to lower social strata, committed more severe Primary and Secondary Typologies By contrast, the non-violent but still extremists were characterized by a family tradition of radicalism, were less likely to have criminal backgrounds, belonged to higher social strata, played leadership roles in terrorist organizations, and instead of committing attacks, backed terrorism by supporting activities. This type may be termed organizing/non-violent. Typically, violent extremists came from criminal but not radical backgrounds and were converted and radicalized in later stages of their life, e.g., in early or mid-adulthood. They played minor roles in terrorist groups, sought training, and were radicalized largely by the social media. Also, they belonged A rough analysis of data suggests that both these major types are likely to be composed of subtypes. As compared May 2022 \| Volume 13 \| Article 745608 Frontiers in Psychology \| www.frontiersin.org 13 Machine Learning and Violent Terrorism Ivaskevics and Haller FIGURE 5 \| Violence prediction by the number of pro- and anti-violence characteristics identiﬁed by machine learning. (A,B) When the opposite characteristic was not present, pro-violence (“pro”) and anti-violence (“anti”) characteristics predicted violence with high accuracy. Note that sample sizes were low; symbols in gray indicate N = 1 for the given “pro”/”anti” characteristic combination. (C) The interactive dependence of violence on the relation between the number of pro-violence and anti-violence characteristics present in the individual. Gray in panels (A–C), the area of chance prediction (34–65%, i.e., close to 50%). crimes before radicalization, were less likely to have radical family friends, and were less likely to belong to a far-left ideology. It occurs that in some instances the subtypes of the two major types had opposite characteristics. For instance, prison radicalization in violent and non-violent extremists seemed to be associated with higher and lower, respectively, social status as compared to their peers belonging to the same type. The analysis of such subtypes would stretch the boundaries of this study; yet the rough analysis presented above suggests that: (1) The critical variables identiﬁed in this study may be used to reﬁne the current typologies and (2) the subtypes of the major types (the “executive/violent” and “organizing/non- violent” types) may be diﬀerentiated based on a diﬀerent set of variables. As shown above for instance, violent extremists may be categorized into subgroups based on citizenship and prison radicalization, whereas the subtypes of non-violent extremists may be diﬀerentiated based education and prison radicalization. Frontiers in Psychology \| www.frontiersin.org The Risk Matrix Unsurprisingly, neither variable alone was able to predict violence accurately. In addition to the well-known complexity of the trait, this was also due to the nature of individual characteristics. These either had a large impact but were rare or the other way round were frequent, but their impact was low. For instance, the ideological subcategory “animal rights/environmentalist” predicted non-violence rather reliably yet only 7% of extremists belonged to this subcategory. By contrast, whites were frequent among radicals, but their risk ratio for violence was low (Figure 5). The power of interactions was shown by the fact that prediction accuracy increased by one order of magnitude when variables with signiﬁcant contributions were considered together. In this regard, our study showed that: (1) one and the same individual may show a mixture of pro- and anti-violence factors and (2) the number and ratio of these predicts either violence or non-violence with high accuracy when one prevails over the other. Putatively, the simple counting of pro- and anti-violence characteristics present in an individual and calculating their ratio or ﬁnding the place of the individual in the matrix shown in Figure 5C provides a good estimate of violence risk in the individual. Missing data may make such endeavors diﬃcult, but one can assume that missingness is due to a certain extent to chance, and as such may not aﬀect considerably the ratio of proviolence and antiviolence characteristics. As such, the risk matrix may be used by calculating ratios rather than counting characteristics. Many missing data makes analysis impossible, but this is a technical rather than a theoretical limitation. FIGURE 5 \| Violence prediction by the number of pro- and anti-violence characteristics identiﬁed by machine learning. (A,B) When the opposite characteristic was not present, pro-violence (“pro”) and anti-violence (“anti”) characteristics predicted violence with high accuracy. Note that sample sizes were low; symbols in gray indicate N = 1 for the given “pro”/”anti” characteristic combination. (C) The interactive dependence of violence on the relation between the number of pro-violence and anti-violence characteristics present in the individual. Gray in panels (A–C), the area of chance prediction (34–65%, i.e., close to 50%). FIGURE 5 \| Violence prediction by the number of pro- and anti-violence characteristics identiﬁed by machine learning. (A,B) When the opposite characteristic was not present, pro-violence (“pro”) and anti-violence (“anti”) characteristics predicted violence with high accuracy. Earlier Hypotheses and the Findings With Machine Learning As shown above, violent extremism was addressed from the point of view of communication science (Youngblood, 2020), criminology (LaFree et al., 2018, 2019), economics (Varaine, 2019), political science (Abrahms, 2012), social psychology (Jasko et al., 2017; Carson et al., 2019), and sociology (Becker, 2019; Ligon et al., 2019), to mention the most popular May 2022 \| Volume 13 \| Article 745608 Frontiers in Psychology \| www.frontiersin.org Frontiers in Psychology \| www.frontiersin.org 14 Machine Learning and Violent Terrorism Ivaskevics and Haller TABLE 6 \| Predictor variables in our and earlier studies. Signiﬁcant predictorthis study Becker, 2019 Jasko et al., 2017 Jensen et al., 2020 LaFree et al., 2018 LaFree et al., 2019 Other RadicalBehaviors X RadicalizationFar Left X IdeologicalSub Category BroadEthnicity – X RadicalFriend X X ReligiousBackground PrisonRadicalize X RadicalizationIslamist X GroupMembership – RadicalFamily X X RoleGroup X SocialMedia X Convert PreviousCriminal Activity – X X – ProblematicSocial Relations – RadicalBeliefs X X Clique X X – MediaRadicalization X Social StratumAdulthood – X This rough analysis shows that our predictor variables were identiﬁed earlier but in different studies, which precluded their direct comparison. X, signiﬁcant predictor in earlier studies; –, non-signiﬁcant predictor in earlier studies; Other, studies that used databases other then PIRUS; for references see text; white characters on black background, predictors not identiﬁed by earlier studies. TABLE 6 \| Predictor variables in our and earlier studies. X This rough analysis shows that our predictor variables were identiﬁed earlier but in different studies, which precluded their direct comparison. X, signiﬁcant predictor in earlier studies; –, non-signiﬁcant predictor in earlier studies; Other, studies that used databases other then PIRUS; for references see text; white characters on black background, predictors not identiﬁed by earlier studies. approaches. Our ﬁndings conﬁrm all these but reveal a group of major factors that were not investigated earlier, suggest a “rank order” of the relative importance of theoretical assumptions, and show that various factors have diﬀerent connotations in violent and non-violent extremists. the social media, were converted, sought training, and were followers if members in groups, non-violent extremists were the deliverers of information by support, sharing beliefs, and leadership roles. Deprivation contributed to violent extremism as problematic social relationships and belonging to low social strata emerged as important predictors of violence. Extremism per se, however, did not seem to be associated with deprivation as non-violent extremists usually belonged to high social strata. Earlier Hypotheses and the Findings With Machine Learning Finally, social learning from family and friends played surprisingly little role as violent extremists usually came from non-extremist backgrounds. On the other hand, non- violent extremists usually came from backgrounds favoring legal or non-violent illegal extremism. As such, non-violent rather than violent extremism seemed to be associated with social learning. Over one third of critical characteristics relate to ideologies, religion, and ethnicity (Figure 4), which received little attention so far. Revealing such factors was the main reason for performing the analysis on the entire PIRUS sample, i.e., without breaking it down to ideological motivations. Noteworthy, the three often interconnected categories of factors may either decrease or increase the likelihood extremist violence, depending on their features. For instance, anti-gay, Sunni Islamist, and white supremacist extremism increase the likelihood of violence, whereas animal rights/environmentalist, and anti-abortion extremism decrease it (Figure 4). To our knowledge, such diﬀerences were not evidenced earlier by hypothesis-driven approaches, yet they appear as major factors of violence according to our machine learning approach. These ﬁndings and conclusion naturally need further experimental support by similar studies made on diﬀerent terrorist databases. Such studies may clarify the generalizability of conclusions, which at present are tightly bound to the particularities of one single database. Furthermore, the machine learning approach should also be employed to investigate separately extremists belonging to diﬀerent ideologies. Such studies may reveal ideology-speciﬁc risk factors, and risk matrices. Regarding the “rank order” of factors, criminological factors appear highly important as prison radicalization emerged as the strongest predictor of violent extremism. Previous violent crime also emerged as an important predictor, whereas never being in prison increased the likelihood of non-violence. Factors related to communication seemed to be the next most important category of factors as about one ﬁfth of critical characteristics were related to communication. Interestingly, however, violent, and non-violent extremists seemed to be on the opposite side of the communication channel. While violent extremists were recipients who were strongly inﬂuenced by Frontiers in Psychology \| www.frontiersin.org DATA AVAILABILITY STATEMENT Publicly available datasets were analyzed in this study. This data can be found here: https://www.start.umd.edu/proﬁles- individual-radicalization-united-states-pirus-keshif. ACKNOWLEDGMENTS We would like to thank the START team for making their database available to the public and scientists interested in their study. We would like to thank the START team for making their database available to the public and scientists interested in their study. The XGBoost algorithm delineated a set of 19 variables that covered in total 53 individual characteristics based on which the risk of extremist violence could be predicted with more than 85% FUNDING This work was supported by the New National Excellence Program of the Ministry of Human Capacities under Grant ÚNKP-18-3-I-NKE-80. This work was supported by the New National Excellence Program of the Ministry of Human Capacities under Grant ÚNKP-18-3-I-NKE-80. Limitations The largest limitations of our analysis were the missing data, which, however, cannot be amended presently. All studies that use large datasets are faced with this problem (Safer-Lichtenstein et al., 2017; Knight and Keatley, 2020). Several methods were proposed to address this issue. Jasko et al. (2017) for instance, May 2022 \| Volume 13 \| Article 745608 15 Machine Learning and Violent Terrorism Ivaskevics and Haller analyzed both the original PIRUS dataset and the same after replacing missing data by multiple imputation. LaFree et al. (2018) compared four diﬀerent methods to substitute missing data—including mean substitution and multiple imputations— and concluded that they were equivalent. In sharp contrast, Safer-Lichtenstein et al. (2017) found that the results of analyses greatly depend on the amount of missing data. In our study, mean substitution worsened whereas multiple imputation did not aﬀect the outcome of analyses. accuracy. This is in the upper region of the accuracies reported for a variety of algorithms that evaluated violence-unrelated features of radicalization (see paragraph 9 in Introduction). We submitted the critical factors and characteristics to a deeper analysis and based on this we suggest that machine learning algorithms may be successfully used to evaluate the relative weight of, and relationship between variables and characteristics, and may be used to improve both terrorist typologies and risk assessment guidelines. Another limitation was the inclusion of all ideological backgrounds despite that Far Right, Far Left, Islamist, and Single-Issue extremists diﬀer from each other in multiple ways (Jensen et al., 2015; Al-Zewairi and Naymat, 2017; Freilich et al., 2018). We studied all extremists together because this allowed the comparison of ideologies within the same analysis. This showed that among ideologies anti- gay and Sunni Islam were the most likely indicators for violence. These were followed by Islam ideology in general, and Far Right ideology, whereas certain ideologies, for instance, Far Left, Anti-abortion, Single-issue, Animal Rights and Environment ideologies were protective against violence. Naturally, this does not decrease the potential value of the separate analysis of diﬀerent ideologies. This, however, remains for forthcoming studies. We suggest that the risk matrix developed here may already be tested as a risk assessment tool for violent extremism. More importantly, however, this study shows the power of machine learning in studying the complex relationships between various factors that endanger violence in extremists, which confers both theoretical and practical perspectives to this approach. AUTHOR CONTRIBUTIONS The individual combinations of various risk and protective factors was insuﬃciently addressed in this study. The reason was that we aimed at comparing ideological backgrounds within the same analysis, which inherently made the sample heterogenous. The individual combinations of various risk and protective factors was insuﬃciently addressed in this study. The reason was that we aimed at comparing ideological backgrounds within the same analysis, which inherently made the sample heterogenous. Separate analyses along ideological backgrounds would likely result in more homogenous samples, which will enable us to analyze individual combinations of risk factors in future studies. KI conceptualized the study and performed the statistical analysis of machine learning. JH performed the statistical analysis of multiple regression. Both authors contributed to the article and approved the submitted version. Separate analyses along ideological backgrounds would likely result in more homogenous samples, which will enable us to analyze individual combinations of risk factors in future studies. Finally, the database was probably not free of errors, which cannot be checked presently. However, the database was compiled by a restricted number of researchers based on a unitary system of carefully established principles. Our ﬁndings need of course crosschecking with a similar study to be done on other databases. REFERENCES Understanding the dynamics of terrorism events with multiple-discipline datasets and machine learning approach. PLoS One 12:e0179057. doi: 10.1371/journal.pone.0179057 Ligon, G., Windisch, S., Braun, C. L., Logan, M. K., Derrick, D. C., and Armstrong, G. (2019). SalaﬁJihadist Inspired Proﬁles and Radicalization Clusters (SPARC). Final Report to the United States Department of Homeland Security. Los Angeles, CA: University of Southern California. Etudo, U. O. (2017). Automatically Detecting the Resonance of Terrorist Movement Frames on the Web. Ph.D. dissertation. Richmond, VA: Virginia Commonwealth University. Lloyd, M., and Dean, C. (2015). The development of structured guidelines for assessing risk in extremist oﬀenders. J. Threat Assess. Manag. 2:40. doi: 10.1037/ tam0000035 Freilich, J. D., Chermak, S. M., Gruenewald, J., Parkin, W. S., and Klein, B. R. (2018). Patterns of fatal extreme-right crime in the United States. Perspect. Terror. 12, 38–51. McCauley, C., and Moskalenko, S. (2008). Mechanisms of political radicalization: pathways toward terrorism. Terror. Polit. Violence 20, 415–433. doi: 10.1080/ 09546550802073367 Gassebner, M., and Luechinger, S. (2011). Lock, stock, and barrel: a comprehensive assessment of the determinants of terror. Public Choice 149:235. doi: 10.1007/ s11127-011-9873-0 Meloy, J. R. (2018). The operational development and empirical testing of the terrorist radicalization assessment protocol (TRAP–18). J. Pers. Assess. 100, 483–492. doi: 10.1080/00223891.2018.1481077 Gruenewald, J., and Pridemore, W. A. (2011). A comparison of ideologically- motivated homicides from the new extremist crime database and homicides from the supplementary homicide reports using multiple imputation by chained equations to handle missing values. J. Quant. Criminol. 28, 141–162. doi: 10.1007/s10940-011-9155-5 Meloy, J. R., and Gill, P. (2016). The lone-actor terrorist and the TRAP-18. J. Threat Assess. Manag. 3:37. doi: 10.1007/s11757-020-00596-y Moghaddam, F. M. (2005). The staircase to terrorism: a psychological exploration. Am. Psychol. 60:161. doi: 10.1037/0003-066X.60.2.161 Guarrieri, T. R., and Meisel, C. J. (2019). Extremists and unconventional weapons: examining the pursuit of chemical and biological agents. Behav. Sci. Terror. Polit. Aggress. 13, 23–42. doi: 10.1080/19434472.2019.1698633 Molnar, C. (2019). Interpretable Machine Learning: A Guide for Making Black Box Models Explainable. Available online at: https://christophm.github.io/ interpretable-ml-book/ (accessed June 20, 2021). Hao, M., Jiang, D., Ding, F., Fu, J., and Chen, S. (2019). Simulating spatio-temporal patterns of terrorism incidents on the Indochina peninsula with GIS and the random forest method. ISPRS Int. J. Geo Inform. 8:133. doi: 10.3390/ijgi8030133 Monahan, J. (2012). The individual risk assessment of terrorism. Psychol. Public Policy Law 18, 167–205. Iskandar, B. (2017). Terrorism detection based on sentiment analysis using machine learning. J. REFERENCES International Conference on Artiﬁcial Intelligence and Computer Science (New York, NY: Association for Computing Machinery), 320–324. doi: 10.1145/ 3349341.3349424 Aaboud, M., Aad, G., Abbott, B., Abdinov, O., Abeloos, B., Abidi, S., et al. (2018). Evidence for the associated production of the Higgs boson and a top quark pair with the ATLAS detector. Phys. Rev. D 97:072003. doi: 10.1103/PhysRevD.97. 072003 Altunbas, Y., and Thornton, J. (2011). Are Homegrown Islamic terrorists diﬀerent? Some UK evidence. Southern Econ. J. 78, 262–272. Abrahms, M. (2012). The political eﬀectiveness of terrorism revisited. Comp. Polit. Stud. 45, 366–393. doi: 10.1177/0010414011433104 Al-Zewairi, M., and Naymat, G. (2017). Spotting the Islamist radical within: religious extremists proﬁling in the United States. Procedia Comput. Sci. 113, 162–169. doi: 10.1016/j.procs.2017. 08.336 Aghabi, L., Bondokju, N., Osborne, A., and Wilkinson, K. (2017). Social Identity and Radicalisation: A Review of Key Concepts. Amman: West Asia-North Africa Institute. Austin, J. (2019). Revisiting the Far Right Violent Extremist Threat: Violent Extremist Plot Success From 1948 Through 2017. Honors Theses. Durham NH: University of New Hampshire. Alfatih, M., Li, C., and Saadalla, N. E. (2019). “Prediction of groups responsible for terrorism attack using tree based models,” in Proceedings of the 2019 May 2022 \| Volume 13 \| Article 745608 Frontiers in Psychology \| www.frontiersin.org 16 Machine Learning and Violent Terrorism Ivaskevics and Haller online at: https://www.start.umd.edu/pubs/PIRUS%20Research%20Brief_Jan% 202015.pdf (accessed June 20, 2021). Bartlett, J., and Miller, C. (2012). The edge of violence: towards telling the diﬀerence between violent and non-violent radicalization. Terror. Polit. Violence 24, 1–21. doi: 10.1080/09546553.2011.594923 Jensen, M., Lafree, G., James, P., Atwell-Seate, A., Pisoiu, D., Stevenson, J., et al. (2016). Empirical Assessment of Domestic Radicalization (EADR). Final Report of the PIRUS Project, National Consortium for the Study of Terrorism and Responses to Terrorism (START), College Park, MD. Available online at: https: //www.ncjrs.Gov/Pdﬃles1/Nij/Grants/250481.Pdf (accessed January 10, 2021). Basuchoudhary, A., and Bang, J. T. (2018). Predicting terrorism with machine learning: lessons from “predicting terrorism: a machine learning approach. Peace Econ. Peace Sci. Publ. Policy 24, 1–8. doi: 10.1515/peps-2018- 0040 Becker, M. H. (2019). When extremists become violent: examining the association between social control, social learning, and engagement in violent extremism. Stud. Conﬂ. Terror. 44, 1–21. doi: 10.1080/1057610X.2019.1626093 Johnson, N. F., Zheng, M., Vorobyeva, Y., Gabriel, A., Qi, H., Velásquez, N., et al. (2016). New online ecology of adversarial aggregates: ISIS and beyond. Science 352, 1459–1463. doi: 10.5120/6303-8516 Benmelech, E., and Berrebi, C. (2007). REFERENCES Human capital and the productivity of suicide bombers. J Econ. Perspect. 21, 223–238. doi: 10.1257/jep.21.3.223 Kalaiarasi, S., Mehta, A., Bordia, D., and Sanskar. (2019). Using global terrorism database (GTD) and machine learning algorithms to predict terrorism and threat. Int. J. Eng. Adv. Technol. 9, 5995–6000. doi: 10.35940/ijeat.A1768.109119 Berrebi, C. (2007). Evidence about the link between education, poverty and terrorism among palestinians. Peace Econ. Peace Sci. Public Policy 13:2. doi: 10.2139/ssrn.487467 Kerodal, A. G., Freilich, J. D., and Chermak, S. M. (2016). Commitment to extremist ideology: using factor analysis to move beyond binary measures of extremism. Stud. Conﬂ. Terror. 39, 687–711. doi: 10.1080/1057610X.2016. 1141012 Borum, R. (2003). Understanding the terrorist mind-set. FBI Law Enforc. Bull. 72, 7–10. Knight, S., and Keatley, D. A. (2020). How can the literature inform counter- terrorism practice? Recent advances and remaining challenges. Behav. Sci. Terror. Polit. Aggress. 12, 217–230. doi: 10.1080/19434472.2019.1666894 Borum, R. (2011). Radicalization into violent extremism I: a review of social science theories. J. Strateg. Security 4:2. doi: 10.5038/1944-0472.4.4.1 Carson, J. V., James, P. A., and O’Neal, T. A. (2019). The radicalization of the Kanes: family as a primary group inﬂuence? Dyn. Asymmetric Conﬂ. 12, 67–89. doi: 10.1080/17467586.2019.1568513 Knight, S., Woodward, K., and Lancaster, G. L. (2017). Violent versus nonviolent actors: an empirical study of diﬀerent types of extremism. J. Threat Assess. Manag. 4:230. doi: 10.1037/tam0000086 Cheong, M., and Lee, V. C. (2011). A microblogging-based approach to terrorism informatics: exploration and chronicling civilian sentiment and response to terrorism events via Twitter. Inform. Syst. Front. 13, 45–59. doi: 10.1007/ s10796-010-9273-x Krueger, A. B. (2008). What makes a homegrown terrorist? Human capital and participation in domestic Islamic terrorist groups in the U.S.A. Econ. Lett. 101, 293–296. doi: 10.1016/j.econlet.2008.09.008 Cohen, K., Johansson, F., Kaati, L., and Mork, J. C. (2014). Detecting linguistic markers for radical violence in social media. Terror. Polit. Violence 26, 246–256. doi: 10.1080/09546553.2014.849948 LaFree, G., Jensen, M. A., James, P. A., and Safer-Lichtenstein, A. (2018). Correlates of violent political extremism in the United States. Criminology 56, 233–268. doi: 10.1111/1745-9125.12169 Cook, A. N. (2014). Risk Assessment and Management of Group-Based Violence. Ph.D. thesis. Burnaby, BC: Simon Fraser University. LaFree, G., Jiang, B., and Porter, L. C. (2019). Prison and violent political extremism in the United States. J. Quant. Criminol. 36, 473–498. doi: 10.1007/s10940-019- 09412-1 Ding, F., Ge, Q., Jiang, D., Fu, J., and Hao, M. (2017). REFERENCES Revisiting the economics and terrorism nexus: collective deprivation, ideology and domestic radicalization in the US (1948–2016). J. Quant. Criminol. 36, 667–699. doi: 10.1007/s10940-019-09422-z Singh, K., Chaudhary, A. S., and Kaur, P. (2019). “A machine learning approach for enhancing defence against global terrorism,” in Proceedings of the 2019 Twelfth International Conference on Contemporary Computing (IC3) (Piscataway, NJ: IEEE), 1–5. doi: 10.1109/IC3.2019.8844947 Verma, C., Malhotra, S., and Verma, V. (2018). Predictive modeling of terrorist attacks using machine learning. Int. J. Pure Appl. Math. 119, 49–61. Youngblood, M. (2020). Extremist ideology as a complex contagion: the spread of far-right radicalization in the United States between 2005-2017. arXiv [Preprint]. arXiv:2004.04265 Sjöstedt, G., and Grann, M. (2002). Risk assessment: what is being predicted by actuarial prediction instruments? Int. J. Forensic Ment. Health 1, 179–183. doi: 10.1080/14999013.2002.10471172 Skeem, J. L., and Monahan, J. (2011). Current directions in violence risk assessment. Curr. Dir. Psychol. Sci. 20, 38–42. doi: 10.1177/0963721410397271 Skeem, J. L., and Monahan, J. (2011). Current directions in violence risk assessment. Curr. Dir. Psychol. Sci. 20, 38–42. doi: 10.1177/0963721410397271 START (2018). Proﬁles of Individual Radicalization in the United States [Data ﬁle]. Available online at: http://www.start.umd.edu/pirus (accessed June 20, 2021). Conﬂict of Interest: The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. START (2018). Proﬁles of Individual Radicalization in the United States [Data ﬁle]. Available online at: http://www.start.umd.edu/pirus (accessed June 20, 2021). Strobl, C., Boulesteix, A.-L., Kneib, T., Augustin, T., and Zeileis, A. (2008). Conditional variable importance for random forests. BMC Bioinformatics 9:307. doi: 10.1186/1471-2105-9-307 Publisher’s Note: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their aﬃliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. Talreja, D., Nagaraj, J., Varsha, N. J., and Mahesh, K. (2017). “Terrorism analytics: learning to predict the perpetrator,” in Proceedings of the 2017 International Conference on Advances in Computing, Communications and Informatics (ICACCI) (Piscataway, NJ: IEEE), 1723–1726. doi: 10.1109/ICACCI.2017. 8126092 Copyright © 2022 Ivaskevics and Haller. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). REFERENCES Eng. Appl. Sci. 12, 691–698. doi: 10.36478/jeasci.2017. 691.698 Narula, M., Kumar, S., Kaul, S., and Saurav, K. S. (2020). “Predictive analysis on global terrorism. In international conference on innovative computing and communications,” in Proceedings of the ICICC 2020, (Singapore: Springer). James, G., Witten, D., Hastie, T., and Tibshirani, R. (2013). An Introduction to Statistical Learning with Applications in R. New York, NY: Springer. doi: 10. 1007/978-1-4614-7138-7 Pelzer, R. (2018). Policing of terrorism using data from social media. Eur. J. Security Res. 3, 163–179. doi: 10.1007/s41125-018-0029-9 Pressman, D. E., and Flockton, J. (2012). Calibrating risk for violent political extremists and terrorists: the VERA 2 structured assessment. Br. J. Forensic Pract. 14, 237–251. doi: 10.1108/14636641211283057 Jasko, K., LaFree, G., and Kruglanski, A. (2017). Quest for signiﬁcance and violent extremism: the case of domestic radicalization. Polit. Psychol. 38, 815–831. doi: 10.1111/pops.12376 Pyrooz, D. C., LaFree, G., Decker, S. H., and James, P. A. (2017). Cut from the same cloth? A comparative study of domestic extremists and gang members in the United States. Justice Q. 35, 1–32. doi: 10.1080/07418825.2017.1 311357 Jensen, M. A., Atwell, S. A., and James, P. A. (2020). Radicalization to violence: a pathway approach to studying extremism. Terror. Polit. Violence 32, 1067–1090. doi: 10.1080/09546553.2018.1442330 Jensen, M., James, P., and Tinsley, H. (2015). Proﬁles of Individual Radicalization in the United States: Preliminary Findings. START Research Brief. Available R Development Core Team (2013). R: A Language and Environment for Statistical Computing. Vienna: R Foundation for Statistical Computing. May 2022 \| Volume 13 \| Article 745608 17 Frontiers in Psychology \| www.frontiersin.org Machine Learning and Violent Terrorism Ivaskevics and Haller Sachan, A., and Roy, D. (2012). TGPM: terrorist group prediction model for counter terrorism. In. J. Comput. Appl. 44, 49–52. Uddin, M. I., Zada, N., Aziz, F., Saeed, Y., Zeb, A., and Ali Shah, S. A. (2020). Prediction of future terrorist activities using deep neural networks. Complexity 2020:1373087. doi: 10.1155/2020/1373087 Safer-Lichtenstein, A., LaFree, G., and Loughran, T. (2017). Studying terrorism empirically: what we know about what we don’t know. J. Contemp. Crim. Justice 33, 273–291. doi: 10.1177/1043986217697873 van Buuren, S., and Groothuis-Oudshoorn, K. (2010). Multivariate imputation by chained equations in R. J. Stat. Softw. 45, 1–68. doi: 10.18637/jss.v045.i03 Sarma, K. M. (2017). Risk assessment and the prevention of radicalization from nonviolence into terrorism. Am. Psychol. 72:278. doi: 10.1037/amp0000121 Varaine, S. (2019). Frontiers in Psychology \| www.frontiersin.org REFERENCES The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Tolan, G. M., and Soliman, O. S. (2015). An experimental study of classiﬁcation algorithms for terrorism prediction. Int. J. Knowl. Eng. 1, 107–112. doi: 10.7763/ IJKE.2015.V1.18 Torlay, L., Perrone-Bertolotti, M., Thomas, E., and Baciu, M. (2017). Machine learning–XGBoost analysis of language networks to classify patients with epilepsy. Brain Inf. 4, 159–169. doi: 10.1007/s40708-017-0065-7 May 2022 \| Volume 13 \| Article 745608 Frontiers in Psychology \| www.frontiersin.org 18
https://openalex.org/W1855988937	https://univ-rennes.hal.science/hal-01196604/file/Dystrophin%20and%20the%20two%20related%20genetic%20diseases_accepted.pdf	English	null	Dystrophin and the two related genetic diseases, Duchenne and Becker muscular dystrophies	Biomolecules & biomedicine	2,015	cc-by	6,538	To cite this version: Elisabeth Le Rumeur. Dystrophin and the two related genetic diseases, Duchenne and Becker muscular dystrophies. Bosnian Journal of Basic Medical Science, 2015, 15 (3), pp.14–20. 10.17305/bjbms.2015.636. hal-01196604 Dystrophin and the two related genetic diseases, Duchenne and Becker muscular dystrophies Elisabeth Le Rumeur To cite this version: Elisabeth Le Rumeur. Dystrophin and the two related genetic diseases, Duchenne and Becker muscular dystrophies. Bosnian Journal of Basic Medical Science, 2015, 15 (3), pp.14–20. 10.17305/bjbms.2015.636. hal-01196604 Distributed under a Creative Commons Attribution 4.0 International License INTRODUCTION A high number of mutations of the DMD gene has been reported with ∼65% being deletions of one or several exons, ∼10% of duplications of exons and ∼15% of single point muta tions [4, 5]. Depending on the fact that mutations maintain or not of the reading frame, dystrophin will or not be present according to the Monaco rule [6]. In case of out-of-frame muta tions, dystrophin is mostly deficient and this leads to the severe DMD disease. In case of in-frame mutations, dystrophin will be expressed as a mutated protein either with missense sub stitution or deletions or duplications of an internal part of the protein (Figure 1). These in-frame mutations mostly lead to the less severe BMD disease in accord with the Monaco rule but in certain cases, exceptions to the Monaco rule appear where a DMD phenotype with an in-frame mutation is observed. This is particularly prominent when mutations involve the N-terminal actin-binding-domain or the Cys-rich domain affecting the binding of dystrophin to F-actin or β-dystroglycan, respectively. Mutations of the dystrophin DMD gene are the cause of two devastating and to date incurable diseases, Duchenne (DMD) and Becker (BMD) muscular dystrophies [1]. DMD gene is the longest human gene with 2.4 Megabases of DNA representing ∼1% of the chromosome X DNA [2, 3]. It is localized on the locus p21 of chromosome X and codes for the protein dystrophin. This large gene comprises 79 exons separated by very large introns which explain the giant size of the gene; the mRNA of the largest isoform of dystrophin is 14 kb which is only 0.6% of the total weight of the gene. The intron 44 is 170 kb on its own. Several promoters are active in a tissue-specificity manner and lead to the expression of full length or shorter dystrophins. The full length dystrophin is expressed in all striated skeletal, smooth and cardiac muscles. Shorter isoforms are expressed in brain cells and in retina. HAL Id: hal-01196604 https://univ-rennes.hal.science/hal-01196604v1 Submitted on 22 Mar 2016 L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. Distributed under a Creative Commons Attribution 4.0 International License BOSNIAN JOURNAL OF BASIC MEDICAL SCIENCES REVIEW WWW.BJBMS.ORG Abstract Mutations of the dystrophin DMD gene, essentially deletions of one or several exons, are the cause of two devastating and to date incurable dis eases, Duchenne (DMD) and Becker (BMD) muscular dystrophies. Depending upon the preservation or not of the reading frame, dystrophin is completely absent in DMD, or present in either a mutated or a truncated form in BMD. DMD is a severe disease which leads to a premature death of the patients. Therapy approaches are evolving with the aim to transform the severe DMD in the BMD form of the disease by restoring the expression of a mutated or truncated dystrophin. These therapies are based on the assumption that BMD is a mild disease. However, this is not completely true as BMD patients are more or less severely affected and no molecular basis of this heterogeneity of the BMD form of the disease is yet understood. The aim of this review is to report for the correlation between dystrophin structures in BMD deletions in view of this heterogeneity and to emphasize that examining BMD patients in details is highly relevant to anticipate for DMD therapy effects. KEYWORDS: Dystrophin; Becker muscular dystrophy; Duchenne muscular dystrophy; exon skipping; micro-dystrophin DOI: http://dx.doi.org/10.17305/bjbms.2015.636 Bosn J Basic Med Sci. 2015;15(3):14-20. © 2015 ABMSFBIH Dystrophin and the two related genetic diseases, Duchenne and Becker muscular dystrophies Elisabeth Le Rumeur1, 2* 1Université de Rennes 1, Faculté de Médecine, 35043, Rennes cedex, France, 2UMR CNRS 6290, Institut de génétique et développement de Rennes, Campus Biologie-Santé, 35043 Rennes Cedex, France *Corresponding author: Elisabeth Le Rumeur, IGDR, Faculté de Médecine, CS34317, 35043 Rennes cedex, France, E-mail: Elisabeth.lerumeur@uni-rennes1.fr © 2015 Le Rumeur E.; licensee ABMSFBIH. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. Submitted: 16 July 2015 / Accepted: 20 July 2015 Submitted: 16 July 2015 / Accepted: 20 July 2015 DYSTROPHIN Nuclei are in the normal muscle at the cell periphery and with the ongoing regeneration in DMD muscles, nuclei become centrally localized. All these features vary depending on the age of the DMD patients with fibrosis and fatty infiltration increasing with age and being highly variable in BMD. In DMD, electron microscopy reveals lesions of the plasma membrane [13] which indicates that the primary role of dystrophin is to maintain plasma membrane integrity. Dystrophin is anchored to plasma membrane by phospho lipids and β-dystroglycan which in turn interacts with extracel lular matrix proteins (Figure 2). As such, dystrophin constitutes a major scaffolding protein of normal muscle which links cyto skeletal actin, microtubules and intermediate filaments to the extracellular matrix. This dystrophin scaffolding network is pres ent at specific structures of the skeletal and heart muscle named costameres situated at the periphery of the fibers along with the transverse M- and Z-lines [44, 45]. Costameres are specifically involved in the lateral transmission of forces from the cytosol to the extracellular matrix by which they prevent plasma mem brane ruptures during muscle contractions [41, 46-49]. DYSTROPHIN Dystrophin consists of four major structural domains (Figure 2) [18, 19]. The N-terminal domain coded by exons 1 to 8 is mostly an actin-binding domain with two calpo nin-homology domains CH1 and CH2 [20-22]. The central part of dystrophin coded by exons 8 to 61 is made of 24 spec trin-like repeats interspaced by four hinges H1 to H4 and at the origin of the rod-shaped filament nature of dystrophin [23]. This long domain interacts with a high number of proteins among which they are the filamentous actin [24], interme diate filaments [25, 26] and microtubules [27, 28] and finally the muscular isoform of nitric oxide synthase (nNOS)[29-31] and PAR1-b [32]. In addition, this domain interacts with mem brane phospholipids[33-35]. The third domain coded by exons 62 to 69 is the Cys-rich domain composed of a WW domain, two EF hands and a ZZ domain[36, 37]. This domain binds to the most important partner of dystrophin i.e. the membrane protein β-dystroglycan [38, 39] and to several other proteins such as plectin[40], ankyrin [41] and the intermediate fila ment protein synemin[25]. The C-terminal domain of dystro phin coded by exons 69 to 79 binds to the two cytoplasmic proteins syntrophin and dystrobrevin[42, 43]. FIGURE 1. Examples of mutations of the DMD gene and their consequences on the production of dystrophin and the corre sponding phenotypes. The boxes represent exons. Right faces of boxes indicate that the exon codes for an entire protein sequence based on 3-bases codons (examples: exons 47, 48, 49). Curved faces indicate that the exon does not code for an entire protein sequence but that either the first or the last bases need the pre ceding or following exon to code a full 3-bases codon (examples: exons 50, 51, 52). BMD for Becker muscular dystrophy; DMD for Duchenne muscular dystrophy. FIGURE 1. Examples of mutations of the DMD gene and their consequences on the production of dystrophin and the corre sponding phenotypes. The boxes represent exons. Right faces of boxes indicate that the exon codes for an entire protein sequence based on 3-bases codons (examples: exons 47, 48, 49). Curved faces indicate that the exon does not code for an entire protein sequence but that either the first or the last bases need the pre ceding or following exon to code a full 3-bases codon (examples: exons 50, 51, 52). DYSTROPHIN BMD for Becker muscular dystrophy; DMD for Duchenne muscular dystrophy. never run and become wheelchair confined before their 12 years. Respiratory and cardiac impairments appear progressively and in most cases, patients do not survive after their 30 years (for a review focusing on clinics, see [7]). These clinical signs are due to a progressive muscle weakness involving all striated skeletal and cardiac muscles and are accompanied by highly elevated cre atine kinase blood levels [8-10]. Similar clinical signs are observed for BMD patients but with very variable time course and sever ity. Some BMD patients are highly asymptomatic while some become wheelchair confined around 16 years of age. They could survive until very old ages or some of them die from an early heart failure [11, 12]. On the histological point of view, the DMD muscles show cycles of fiber necrosis and regeneration. However, regener ation is overtaken by fiber loss mechanisms and fibrosis and adi pose tissue replacement are increased. Nuclei are in the normal muscle at the cell periphery and with the ongoing regeneration in DMD muscles, nuclei become centrally localized. All these features vary depending on the age of the DMD patients with fibrosis and fatty infiltration increasing with age and being highly variable in BMD. In DMD, electron microscopy reveals lesions of the plasma membrane [13] which indicates that the primary role of dystrophin is to maintain plasma membrane integrity. never run and become wheelchair confined before their 12 years. Respiratory and cardiac impairments appear progressively and in most cases, patients do not survive after their 30 years (for a review focusing on clinics, see [7]). These clinical signs are due to a progressive muscle weakness involving all striated skeletal and cardiac muscles and are accompanied by highly elevated cre atine kinase blood levels [8-10]. Similar clinical signs are observed for BMD patients but with very variable time course and sever ity. Some BMD patients are highly asymptomatic while some become wheelchair confined around 16 years of age. They could survive until very old ages or some of them die from an early heart failure [11, 12]. On the histological point of view, the DMD muscles show cycles of fiber necrosis and regeneration. However, regener ation is overtaken by fiber loss mechanisms and fibrosis and adi pose tissue replacement are increased. CLINICS, HISTOPATHOLOGY AND DYSTROPHIN IN DMD AND BMD The first clinical signs of DMD are difficulties for young boys to walk and to climb stairs as early as 2-3 years of age. These boys 14 Elisabeth Le Rumeur: Dystrophin in Duchenne and Becker muscular dystrophies BMD muscles, the expression of mutated dystrophin is observed but with highly variable extents from less than 10 % to as high as 75% of the full length expression of normal muscles [14, 15]. The direct correlation between dystrophin amount in BMD and clin ical severity is not proven. However, it is recognized that mea suring accurately the dystrophin level is highly challenging [16] partly explaining the lack of clear correlation between pheno type severity and dystrophin level. Other parameters could also be involved as there exist certain unexplained results which appear in contrast with the accepted rules [17]. FIGURE 1. Examples of mutations of the DMD gene and their consequences on the production of dystrophin and the corre sponding phenotypes. The boxes represent exons. Right faces of boxes indicate that the exon codes for an entire protein sequence based on 3-bases codons (examples: exons 47, 48, 49). Curved faces indicate that the exon does not code for an entire protein sequence but that either the first or the last bases need the pre ceding or following exon to code a full 3-bases codon (examples: exons 50, 51, 52). BMD for Becker muscular dystrophy; DMD for Duchenne muscular dystrophy. DYSTROPHIN AND DMD AND BMD The primary defects in both DMD and BMD are disrup tions of this dystrophin scaffolding network by the absence of Immunoblotting of dystrophin in DMD muscles reveals the total absence of dystrophin except in some revertant fibers. In 15 Elisabeth Le Rumeur: Dystrophin in Duchenne and Becker muscular dystrophies FIGURE 2. Dystrophin domains and partners. CH1 and CH2: calponin homology domains 1 and 2; H: unstructured regions called hinge; R: spectrin like repeat; WW: Tryptophan-rich domain; Z: ZZ domain; EFH: EF hand domains; PAR-1b: polarity regulating kinase 1b; nNOS: n-nitrous oxide synthase; Ifs: intermediate filaments. FIGURE 2. Dystrophin domains and partners. CH1 and CH2: calponin homology domains 1 and 2; H: unstructured regions called hinge; R: spectrin like repeat; WW: Tryptophan-rich domain; Z: ZZ domain; EFH: EF hand domains; PAR-1b: polarity regulating kinase 1b; nNOS: n-nitrous oxide synthase; Ifs: intermediate filaments. FIGURE 2. Dystrophin domains and partners. CH1 and CH2: calponin homology domains 1 and 2; H: unstructured regions called hinge; R: spectrin like repeat; WW: Tryptophan-rich domain; Z: ZZ domain; EFH: EF hand domains; PAR-1b: polarity regulating kinase 1b; nNOS: n-nitrous oxide synthase; Ifs: intermediate filaments. FIGURE 3. The Monaco rule and how it is used to generate ther apy patterns. dystrophin in DMD or due to mutations that ablate part of the protein as in BMD. The transmission of forces from the cyto sol to the extracellular matrix at the costamere is impaired, leading to frequent ruptures of plasma membrane during con tractions [13, 50]. This leads to leakage of cellular components such as creatine kinase from the interior of the muscle cells which explains the high plasma membrane creatine kinase levels in blood of children suffering from DMD. In addition, these ruptures of plasma membrane increase inside fluxes of calcium which subsequently activate calcium-dependent pro teases [51]. It is remarkable that a primary defect i.e. the dystrophin deficit or mutation induces a progressive muscle impairment demonstrating that muscle is able to function without dys trophin but is not able to resist forces in the long time. DYSTROPHIN AND DMD AND BMD That signifies that the therapy strategies could be very large from inducing dystrophin expression to compensatory therapies such as increasing blood flow with NO mimetic, increasing regeneration potential, surrogate protein expression… As well, the observation of the progression of the disease in BMD patients indicates that dystrophin could sustain partly its func tion even with mutations that ablate part of its central domain. In the reverse, missense mutations in the N-terminal ABD or in the Cys-rich domain could be accompanied by the DMD severe phenotype [4, 5]. Therefore, it is clear that certain parts of the molecule are more indispensable than others and this led to the idea that a therapy strategy could be to transform the severe DMD phenotype into a mild BMD phenotype, expressing mutation in the less indispensable part of dystro phin (Figure 3). FIGURE 3. The Monaco rule and how it is used to generate ther apy patterns. emerged [52]. Truncated gene coding sequences (micro-dys trophins) inspired from the truncated dystrophin coding sequences observed in mild BMD patients [53](Figure 3) were designed [54-57]. One of these has been well used in animal models such as the mdx mouse and the grmd dog. This highly simplified micro-dystrophin only consists of the N- and Cys-rich domains with two hinges and four repeats (Figure 4)[58, 59]. However, this micro-dystrophin does not recapitulate all the functions of dystrophin [54] and further improvements are needed essentially by addition of other binding domains of dystrophin such as for example, the nNOS and microtubule binding sites (Figure 2) [56]. DYSTROPHIN AND DMD THERAPY Indeed, the injection of naked whole cDNA has been proven to be inefficient to produce significant dystrophin lev els in diseased muscles and the cDNA has to be vectorised to specifically reach muscles. For this purposes, different serotypes of adeno-associated viruses (AAV) have been used. However, the whole DMD cDNA is too large to be inserted in the AAVs. Therefore, the concept of micro-dystrophin has The exon skipping therapy aims at recover an in-frame mutation equivalent to a BMD mutation. In case of the frame-shifting deletion of exon, modifying the splicing of mRNA allows the exclusion of one or several additional exons and the restoration of a reading frame (Figure 5). To 16 Elisabeth Le Rumeur: Dystrophin in Duchenne and Becker muscular dystrophies FIGURE 4. Two examples of micro-dystrophins. FIGURE 5. Exon skipping for restoration of the reading frame. Same legend as in Figure 1. FIGURE 4 Two examples of micro dystrophins FIGURE 4. Two examples of micro-dystrophins. FIGURE 4. Two examples of micro-dystrophins. FIGURE 5. Exon skipping for restoration of the reading frame. Same legend as in Figure 1. FIGURE 5. Exon skipping for restoration of the reading frame. Same legend as in Figure 1. produce no dystrophin (personal communication). On con trast, there are no patients reported for the deletion of exons 44-45, signifying that the deletion may be asymptomatic. Therefore, the therapy of these two DMD deletions may not result in the same cure effect even though we anticipate that patients will all be transformed in BMD-like patients. this end, small antisense oligonucleotides (AONs) sequences are designed to bind to exon splice junctions, masking them from the spliceosome. This modifies the splicing and skips the exon (s) focused on from the mRNA. Two types of AONs have been designed, 2’O-methyl phosphorothioate (2’O-Me) or phosphorodiamidate morpholino (PMO) oligomers. In that field, we recently studied the clinical status of BMD patients with deletions starting by exon 45 deletion e.g. dele tions of exons 45-47, 45-48, 45-49 and 45-51 and the in vitro biochemical status of the proteins [76]. We showed that the structure of the new dystrophins as produced in these BMD patients could partly explain the differences in the clinical severity of the patients. The patients with deletions 45-47 and 45-49 were confined to wheelchair or were diagnosed with a cardiomyopathy about 10 years earlier than patients with deletions 45-48 and 45-51. DYSTROPHIN AND DMD THERAPY The new junction at the deletion site of these two last deletions reconstitutes a spectrin-like structure while it is not the case of the two first deletions. Therefore, the choice for skipping one exon or the other for restoring the reading frame to a DMD deletion means that the genotype - phenotype correlation data in BMD patients is well established. In addition, this proves that examining the status of large cohorts of BMD patients together with the status of dystrophin is of high interest for the therapy design. For the clinicians, this also signifies that it is essential to diagnose all the BMD patients by genetics and to follow the time course The proof of concept has been largely demon strated [60-66] [67] and there are now human trials in prog ress using oligonucleotides injection using AAV [68-72](see the excellent reviews about therapy [56, 73-74]). For example, the deletions of exon 46 or 44 both shift the reading frame and result in no dystrophin production. The two deletions can be enlarged by skipping exon 45 resulting in the two in-frame deletions 45-46 and 44-45, respectively (Figure 5). Theoretically, the two enlarged deletions should lead to the production of a BMD-like truncated and partially functional dystrophin. However, these schemes are more complex than at first glance since a lot of BMD patients have a more severe disease than others. This high variability of phenotypes has long been recognized but rather recently examining BMD patients in details has emerged to be relevant to anticipate for DMD ther apy effects. In the two examples here in Figure 5, two BMD-like deletions may be produced using the same drug in differ ent patients. However, the deletion 45-46 has been recently reported to be accompanied by a DMD phenotype [75] and to 17 Elisabeth Le Rumeur: Dystrophin in Duchenne and Becker muscular dystrophies of the disease by regularly evaluating respiratory and cardiac muscle status and walking ability of the BMD patients. [7] Wein N, Alfano L, Flanigan KM. Genetics and Emerging Treatments for Duchenne and Becker Muscular Dystrophy. Pediatr Clin N Am. 2015;62(3):723-42. http://dx.doi.org/10.1016/j.pcl.2015.03.008. [7] Wein N, Alfano L, Flanigan KM. Genetics and Emerging Treatments for Duchenne and Becker Muscular Dystrophy. Pediatr Clin N Am. 2015;62(3):723-42. http://dx.doi.org/10.1016/j.pcl.2015.03.008. [8] Kohler M, Clarenbach CF, Boni L, Brack T, Russi EW, Bloch KE. Quality of life, physical disability, and respiratory impairment in Duchenne muscular dystrophy. ACKNOWLEDGEMENTS [15] Anthony K, Cirak S, Torelli S, Tasca G, Feng L, Arechavala-Gomeza V, et al. Dystrophin quantification and clinical correlations in Becker muscular dystrophy: implications for clinical trials. Brain. 2011 Dec;134:3544-56. http://dx.doi.org/10.1093/brain/ awr291. The author thanks all the colleagues and collaborators who participated in France and worldwide to her work about dys trophin structure and function and the AFM-Téléthon for its constant support. [16] Aartsma-Rus A. Dystrophin analysis in clinical trials. Journal of Neuromuscular diseases. 2014;1:41-53. [17] Dubowitz V, Cohn RD. Dystrophin and Duchenne dystro phy. Neuromuscul Disord. 2015 May;25:361-2. http://dx.doi. org/10.1016/j.nmd.2015.03.013. CONCLUSION Finally, it appears that therapies in progress will not cure the DMD disease but slow down its progression. The purpose now is to focus therapy on the BMD-like dystrophin sequences with the less severe disease either for gene or exon skipping therapy. We have now to increase our knowledge about BMD disease genetics and time course and dystrophin structure consequences after BMD deletions. The effects of genetics modifiers have also to be studied in details such as non-synon ymous polymorphism (SNPs) [4], noncoding RNAs [77, 78] or other genetic modifiers [79-83] as they all may modulate the expression of dystrophin or the severity of the disease. The diagnosis of BMD is highly relevant to anticipate and to understand data resulting from human clinical trials. It is also likely that certain severe BMD patients should benefit from therapy strategies or from compensatory strategies elaborated for DMD patients. [9] Davies KE, Nowak KJ. Molecular mechanisms of muscular dystro phies: old and new players. Nat Rev Mol Cell Biol. 2006 Oct;7:762- 73. http://dx.doi.org/10.1038/nrm2024. g [10] Rahimov F, Kunkel LM. The cell biology of disease: cellular and molecular mechanisms underlying muscular dystrophy. J Cell Biol. 2013 May 13;201:499-510. http://dx.doi.org/10.1083/jcb.201212142h g [10] Rahimov F, Kunkel LM. The cell biology of disease: cellular and molecular mechanisms underlying muscular dystrophy. J Cell Biol. 2013 May 13;201:499-510. http://dx.doi.org/10.1083/jcb.201212142 [11] Bushby KM, Gardner-Medwin D. The clinical, genetic and dys trophin characteristics of Becker muscular dystrophy. I. Natural history. J Neurol. 1993 Feb;240:98-104. http://dx.doi.org/10.1007/ BF00858725. [12] Bushby KM, Gardner-Medwin D, Nicholson LV, Johnson MA, Haggerty ID, Cleghorn NJ, et al. The clinical, genetic and dystro phin characteristics of Becker muscular dystrophy. II. Correlation of phenotype with genetic and protein abnormalities. J Neurol. 1993 Feb;240:105-12. http://dx.doi.org/10.1007/BF00858726. [13] Petrof BJ, Shrager JB, Stedmann HH, Kelly AM, Sweeney HL. Dystrophin protects the sarcolemma from stresses developed during muscle contraction. Proceedings of the National Academy of Sciences, USA. 1993;90:3710-4. http://dx.doi.org/10.1073/ pnas.90.8.3710. [14] Neri M, Torelli S, Brown S, Ugo I, Sabatelli P, Merlini L, et al. Dystrophin levels as low as 30% are sufficient to avoid muscular dystrophy in the human. Neuromuscul Disord. 2007 Dec;17:913-8. http://dx.doi.org/10.1016/j.nmd.2007.07.005. DYSTROPHIN AND DMD THERAPY Am J Respir Crit Care Med. 2005 Oct 15;172:1032-6. http://dx.doi.org/10.1164/ rccm.200503-322OC. REFERENCES [18] Ervasti J, Campbell K. Dystrophin and the membrane skele ton. Curr Op Cell Biol. 1993;5:82-7. http://dx.doi.org/10.1016/ S0955-0674(05)80012-2. [1] Koenig M, Beggs A, Moyer M, Scherpf S, Heindrich K, Bettecken T, et al. The molecular basis for Duchenne versus Becker muscular dystrophy: correlation of severity with type of deletion. Am J Hum Genet. 1989;45:498-506. [19] Le Rumeur E, Winder SJ, Hubert JF. Dystrophin: More than just the sum of its parts. Biochim Biophys Acta. 2010 Sep;1804:1713-22. [2] Koenig M, Hoffman EP, Bertelson CJ, Monaco AP, Feener C, Kunkel LM. Complete cloning of the Duchenne muscular dystrophy (DMD) cDNA and preliminary genomic organization of the DMD gene in normal and affected individuals. Cell. 1987 Jul 31;50:509-17. http://dx.doi.org/10.1016/0092-8674(87)90504-6. [20] Levine BA, Moir AJG, Patchell VB, Perry SV. Binding sites involved in the interaction of actin with the N-terminal region of dystrophin. FEBS Letters. 1992;298:44-8. http://dx.doi. org/10.1016/0014-5793(92)80019-D. [21] Keep NH, Norwood FLM, Moores CA, Winder SJ, Kendrick-Jones J. The 2.0 A Structure of the Second Calponin Homology Domain from the Actin-binding Region of the Dystrophin Homologue Utrophin. J Mol Biol. 1999;285:1257-64. http://dx.doi.org/10.1006/ jmbi.1998.2406. [21] Keep NH, Norwood FLM, Moores CA, Winder SJ, Kendrick-Jones J. The 2.0 A Structure of the Second Calponin Homology Domain from the Actin-binding Region of the Dystrophin Homologue Utrophin. J Mol Biol. 1999;285:1257-64. http://dx.doi.org/10.1006/ jmbi.1998.2406. [3] Kenwrick S, Patterson M, Speer A, Fischbeck K, Davies K. Molecular analysis of the Duchenne muscular dystrophy region using pulsed field gel electrophoresis. Cell. 1987 Jan 30;48:351-7. http://dx.doi.org/10.1016/0092-8674(87)90438-7. [4] Flanigan KM, Dunn DM, von Niederhausern A, Soltanzadeh P, Gappmaier E, Howard MT, et al. Mutational spectrum of DMD mutations in dystrophinopathy patients: application of modern diag nostic techniques to a large cohort. Hum Mutat. 2009 Dec;30:1657- 66. http://dx.doi.org/10.1002/humu.21114. j [22] Norwood F, Sutherland-Smith A, Keep N, Kendrick-Jones J. The structure of the N-terminal actin-binding domain of human dys trophin and how mutations in this domain may cause Duchenne or Becker muscular dystrophy. Structure. 2000;8:481-91. http://dx.doi. org/10.1016/S0969-2126(00)00132-5. [22] Norwood F, Sutherland-Smith A, Keep N, Kendrick-Jones J. The structure of the N-terminal actin-binding domain of human dys trophin and how mutations in this domain may cause Duchenne or Becker muscular dystrophy. Structure. 2000;8:481-91. http://dx.doi. org/10.1016/S0969-2126(00)00132-5. [5] Tuffery-Giraud S, Beroud C, Leturcq F, Yaou RB, Hamroun D, Michel- Calemard L, et al. Genotype-phenotype analysis in 2,405 patients with a dystrophinopathy using the UMD-DMD database: a model of nationwide knowledgebase. Hum Mutat. 2009 Jun;30:934-45. http://dx.doi.org/10.1002/humu.20976. Elisabeth Le Rumeur: Dystrophin in Duchenne and Becker muscular dystrophies Molecular clues to the dystrophin–nNOS interac tion: a theoretical approach. Biochemistry. 2013;52:7777-84. http:// dx.doi.org/10.1021/bi400794p. http://dx.doi.org/10.1097/00003677-200304000-00004. [49] Kumar A, Khandelwal N, Malya R, Reid MB, Boriek AM. Loss of dystrophin causes aberrant mechanotransduction in skeletal mus cle fibers. Faseb J. 2004 Jan;18:102-13. http://dx.doi.org/10.1096/ fj.03-0453com. [32] Yamashita K, Suzuki A, Satoh Y, Ide M, Amano Y, Masuda-Hirata M, et al. The 8th and 9th tandem spectrin-like repeats of utrophin cooperatively form a functional unit to inter act with polarity-regulating kinase PAR-1b. Biochem Biophys Res Commun. 2010 Nov 26;391:812–7. http://dx.doi.org/10.1016/j. bbrc.2009.11.144. j [50] Bloch RJ, Reed P, O’Neill A, Strong J, Williams M, Porter N, et al. Costameres mediate force transduction in healthy skeletal muscle and are altered in muscular dystrophies. J Muscle Res Cell Motil. 2004;25:590-2. [33] Le Rumeur E, Fichou Y, Pottier S, Gaboriau F, Rondeau-Mouro C, Vincent M, et al. Interaction of dystrophin rod domain with mem brane phospholipids: evidence of a close proximity between trypto phan residues and lipids. J Biol Chem. 2003;278:5993-6001. http:// dx.doi.org/10.1074/jbc.M207321200. [50] Bloch RJ, Reed P, O’Neill A, Strong J, Williams M, Porter N, et al. Costameres mediate force transduction in healthy skeletal muscle and are altered in muscular dystrophies. J Muscle Res Cell Motil. 2004;25:590-2. [51] Mariol MC, Segalat L. Muscular degeneration in the absence of dystrophin is a calcium-dependent process. Curr Biol. 2001 Oct 30;11:1691-4. http://dx.doi.org/10.1016/ S0960-9822(01)00528-0. [34] Legardinier S, Hubert J-F, Le Bihan O, Tascon C, Rocher C, Raguénès-Nicol C, et al. Sub-domains of the dystrophin rod domain display contrasting lipid-binding and stability properties. Biochim Biophys Acta. 2008;1784:672-82. http://dx.doi.org/10.1016/j. bbapap.2007.12.014. [52] Harper SQ, Hauser MA, DelloRusso C, Duan D, Crawford RW, Phelps SF, et al. Modular flexibility of dystrophin: implications for gene therapy of Duchenne muscular dystrophy. Nature Med. 2002;8:253-61. http://dx.doi.org/10.1038/nm0302-253. [35] Legardinier S, Raguénès-Nicol C, Tascon C, Rocher C, Hardy S, Hubert JF, et al. Mapping of the lipid-binding and stability prop erties of the central rod domain of human dystrophin. J Mol Biol. 2009;389:546-58. http://dx.doi.org/10.1016/j.jmb.2009.04.025. [53] England S, Nicholson L, Johnson M, Forrest S, Love D, Zubrzycka-Gaarn E, et al. Very mild muscular dystrophy associated with the deletion of 46% of the dystrophin. Nature. 1990;343:180-2. http://dx.doi.org/10.1038/343180a0. [36] Campbell K, Kahl S. Association of dystrophin and an integral membrane glycoprotein. Nature. 1989;338:259-62. http://dx.doi. org/10.1038/338259a0. [54] Seto JT, Ramos JN, Muir L, Chamberlain JS, Odom GL. Gene replacement therapies for duchenne muscular dystrophy using adeno-associated viral vectors. Curr Gene Ther. 2012 Jun;12:139-51. Elisabeth Le Rumeur: Dystrophin in Duchenne and Becker muscular dystrophies utrophin. Biochem Biophys Res Commun. 2006 Aug 4;346:768-77. http://dx.doi.org/10.1016/j.bbrc.2006.05.192. [42] Sadoulet-Puccio H, Rajala M, Kunkel L. Dystrobrevin and dystro phin: An interaction through coiled-coil motifs. Proc Natl Acad Sci USA. 1997;94:12413-8. http://dx.doi.org/10.1073/pnas.94.23.12413. g j [26] Stone MR, O’Neill A, Catino D, Bloch RJ. Specific interaction of the actin-binding domain of dystrophin with intermediate filaments containing keratin 19. Mol Biol Cell. 2005 Sep;16:4280-93. http:// dx.doi.org/10.1091/mbc.E05-02-0112. [43] Newey SE, Benson MA, Ponting CP, Davies KE, Blake DJ. Alternative splicing of dystrobrevin regulates the stoichiome try of syntrophin binding to the dystrophin protein complex. Curr Biol. 2000 Oct 19;10:1295-8. http://dx.doi.org/10.1016/ S0960-9822(00)00760-0. [27] Prins KW, Humston JL, Mehta A, Tate V, Ralston E, Ervasti JM. Dystrophin is a microtubule-associated protein. J Cell Biol. 2009 Aug 10;186:363-9. http://dx.doi.org/10.1083/ jcb.200905048. [44] Zubrzycka-Gaarn EE, Bulman DE, Karpati G, Burghes AH, Belfall B, Klamut HJ, et al. The Duchenne muscular dystrophy gene prod uct is localized in sarcolemma of human skeletal muscle. Nature. 1988 Jun 2;333:466-9. http://dx.doi.org/10.1038/333466a0. j [28] Belanto JJ, Mader TL, Eckhoff MD, Strandjord DM, Banks GB, Gardner MK, et al. Microtubule binding distinguishes dystrophin from utrophin. Proc Natl Acad Sci U S A. 2014 Apr 15;111:5723-8. http://dx.doi.org/10.1073/pnas.1323842111.h [45] Jaka O, Casas-Fraile L, Lopez de Munain A, Saenz A. Costamere proteins and their involvement in myopathic processes. Expert Rev Mol Med. 2015;17:e12. http://dx.doi.org/10.1017/erm.2015.9. [29] Lai Y, Thomas GD, Yue Y, Yang HT, Li D, Long C, et al. Dystrophins carrying spectrin-like repeats 16 and 17 anchor nNOS to the sar colemma and enhance exercise performance in a mouse model of muscular dystrophy. J Clin Invest. 2009 Mar;119:624-35. http://dx. doi.org/10.1172/JCI36612. p g [46] Ervasti JM. Costameres: the Achilles’ heel of Herculean muscle. J Biol Chem. 2003 Apr 18;278:13591-4. http://dx.doi.org/10.1074/jbc. R200021200. [47] Bloch RJ, Capetanaki Y, O’Neill A, Reed P, Williams MW, Resneck WG, et al. Costameres: repeating structures at the sarco lemma of skeletal muscle. Clin Orthop Relat Res. 2002 Oct: S203- 10. http://dx.doi.org/10.1097/00003086-200210001-00024. g [30] Lai Y, Zhao J, Yue Y, Duan D. alpha2 and alpha3 helices of dystro phin R16 and R17 frame a microdomain in the alpha1 helix of dys trophin R17 for neuronal NOS binding. Proc Natl Acad Sci U S A. 2013 Nov 26;110:525-30. http://dx.doi.org/10.1073/pnas.1211431109 [48] Bloch RJ, Gonzalez-Serratos H. Lateral force transmission across costameres in skeletal muscle. Exerc Sport Sci Rev. 2003 Apr;31:73-8. http://dx.doi.org/10.1097/00003677-200304000-00004. [31] Giudice E, Molza A-E, Laurin Y, Nicolas A, Le Rumeur E, Delalande O. REFERENCES [23] Koenig M, Kunkel LM. Detailed analysis of the repeat domain of dystrophin reveals four potential hinge segments that may confer flexibility. J Biol Chem. 1990;265:4560-6. [23] Koenig M, Kunkel LM. Detailed analysis of the repeat domain of dystrophin reveals four potential hinge segments that may confer flexibility. J Biol Chem. 1990;265:4560-6. l [24] Amann KJ, Renley BA, Ervasti JM. A Cluster of Basic Repeats in the Dystrophin Rod Domain Binds F-actin through an Electrostatic Interaction. J Biol Chem. 1998;273:28419-23. http://dx.doi. org/10.1074/jbc.273.43.28419. l [24] Amann KJ, Renley BA, Ervasti JM. A Cluster of Basic Repeats in the Dystrophin Rod Domain Binds F-actin through an Electrostatic Interaction. J Biol Chem. 1998;273:28419-23. http://dx.doi. org/10.1074/jbc.273.43.28419. [6] Monaco A, Bertelson C, Liechti-Gallati S, Moser H, Kunkel L. An explanation for the phenotypic differences between patients bear ing partial deletions of the DMD locus. Genomics. 1988;2:90-5. http://dx.doi.org/10.1016/0888-7543(88)90113-9. [25] Bhosle RC, Michele DE, Campbell KP, Li Z, Robson RM. Interactions of intermediate filament protein synemin with dystrophin and [25] Bhosle RC, Michele DE, Campbell KP, Li Z, Robson RM. Interactions of intermediate filament protein synemin with dystrophin and 18 Elisabeth Le Rumeur: Dystrophin in Duchenne and Becker muscular dystrophies Elisabeth Le Rumeur: Dystrophin in Duchenne and Becker muscular dystrophies transgenes improves expression and physiological outcome in dystrophic mdx mice following AAV2/8 gene transfer. Mol Ther. 2008 Nov;16:1825-32. http://dx.doi.org/10.1038/mt.2008.186. [71] Cirak S, Arechavala-Gomeza V, Guglieri M, Feng L, Torelli S, Anthony K, et al. Exon skipping and dystrophin restoration in patients with Duchenne muscular dystrophy after systemic phos phorodiamidate morpholino oligomer treatment: an open-label, phase 2, dose-escalation study. Lancet. 2011 Aug 13;378:595-605. http://dx.doi.org/10.1016/S0140-6736(11)60756-3. [60] Goyenvalle A, Vulin A, Fougerousse F, Leturcq F, Kaplan J, Garcia L, et al. Rescue of dystrophic muscle through U7 snR NA-mediated exon skipping. Science. 2004;306. http://dx.doi. org/10.1126/science.1104297. [72] Cirak S, Feng L, Anthony K, Arechavala-Gomeza V, Torelli S, Sewry C, et al. Restoration of the Dystrophin-associated Glycoprotein Complex After Exon Skipping Therapy in Duchenne Muscular Dystrophy. Mol Ther. 2011 Nov 15. [61] Aartsma-Rus A, Janson AA, Heemskerk JA, De Winter CL, Van Ommen GJ, Van Deutekom JC. Therapeutic modulation of DMD splicing by blocking exonic splicing enhancer sites with antisense oligonucleotides. Ann N Y Acad Sci. 2006 Oct;1082:74-6. http://dx.doi.org/10.1196/annals.1348.058. h [73] Guiraud S, Aartsma-Rus A, Vieira NM, Davies KE, van Ommen GJ, Kunkel LM. The Pathogenesis and Therapy of Muscular Dystrophies. Annu Rev Genomics Hum Genet. 2015 Jun 4. http://dx.doi. org/10.1146/annurev-genom-090314-025003. [62] Aartsma-Rus A, Janson AA, Kaman WE, Bremmer-Bout M, van Ommen GJ, den Dunnen JT, et al. Antisense-induced multiexon skipping for Duchenne muscular dystrophy makes more sense. Am J Hum Genet. 2004 Jan;74:83-92. http://dx.doi.org/10.1086/381039. [74] Wilton SD, Veedu RN, Fletcher S. The emperor’s new dystrophin: finding sense in the noise. Trends Mol Med. 2015 May 23. http:// dx.doi.org/10.1016/j.molmed.2015.04.006. p g [63] Takeshima Y, Yagi M, Wada H, Ishibashi K, Nishiyama A, Kakumoto M, et al. Intravenous infusion of an antisense oligonu cleotide results in exon skipping in muscle dystrophin mRNA of Duchenne muscular dystrophy. Pediatr Res. 2006 May;59:690-4. http://dx.doi.org/10.1203/01.pdr.0000215047.51278.7c. [75] Findlay AR, Wein N, Kaminoh Y, Taylor LE, Dunn DM, Mendell JR, et al. Clinical phenotypes as predictors of the outcome of skipping around DMD exon 45. Ann Neurol. 2015 Jan 21;77:668-74. http:// dx.doi.org/10.1002/ana.24365. [64] McClorey G, Moulton HM, Iversen PL, Fletcher S, Wilton SD. Antisense oligonucleotide-induced exon skipping restores dystro phin expression in vitro in a canine model of DMD. Gene Ther. 2006 Oct;13:1373-81. http://dx.doi.org/10.1038/sj.gt.3302800. [76] Nicolas A, Raguenes-Nicol C, Ben Yaou R, Ameziane-Le Hir S, Cheron A, Vie V, et al. Becker muscular dystrophy severity is linked to the structure of dystrophin. Hum Mol Genet. 2015 Oct 27;24:1267-79. http://dx.doi.org/10.1093/hmg/ddu537. Elisabeth Le Rumeur: Dystrophin in Duchenne and Becker muscular dystrophies [65] Vulin A, Barthelemy I, Goyenvalle A, Thibaud JL, Beley C, Griffith G, et al. Muscle function recovery in golden retriever muscular dystro phy after AAV1-U7 exon skipping. Mol Ther. 2012 Nov;20:2120-33. http://dx.doi.org/10.1038/mt.2012.181. [77] Cacchiarelli D, Martone J, Girardi E, Cesana M, Incitti T, Morlando M, et al. MicroRNAs involved in molecular circuit ries relevant for the Duchenne muscular dystrophy pathogene sis are controlled by the dystrophin/nNOS pathway. Cell Metab. 2010 Oct 6;12:341-51. http://dx.doi.org/10.1016/j.cmet.2010.07.008. [66] Aartsma-Rus A, Muntoni F. 194th ENMC international work shop. 3rd ENMC workshop on exon skipping: towards clinical application of antisense-mediated exon skipping for Duchenne muscular dystrophy 8-10 December 2012, Naarden, The Netherlands. Neuromuscul Disord. 2013 Nov;23:934-44. http://dx. doi.org/10.1016/j.nmd.2013.06.369. g j [78] Cacchiarelli D, Incitti T, Martone J, Cesana M, Cazzella V, Santini T, et al. miR-31 modulates dystrophin expression: new implications for Duchenne muscular dystrophy therapy. EMBO Rep. 2011 Feb;12:136-41. http://dx.doi.org/10.1038/embor.2010.208. [79] Kerst B, Mennerich D, Schuelke M, Stoltenburg-Didinger G, von Moers A, Gossrau R, et al. Heterozygous myogenic factor 6 mutation associated with myopathy and severe course of Becker muscular dystrophy. Neuromuscul Disord. 2000;10:572-7. http:// dx.doi.org/10.1016/S0960-8966(00)00150-4. [67] Le Guiner C, Montus M, Servais L, Cherel Y, Francois V, Thibaud JL, et al. Forelimb treatment in a large cohort of dystrophic dogs supports delivery of a recombinant AAV for exon skipping in Duchenne patients. Mol Ther. 2014 Nov;22:1923-35. http://dx.doi. org/10.1038/mt.2014.151. [80] Davis DB, Delmonte AJ, Ly CT, McNally EM. Myoferlin, a candi date gene and potential modifier of muscular dystrophy. Hum Mol Genet. 2000 Jan 22;9:217-26. http://dx.doi.org/10.1093/hmg/9.2.217 [68] van Deutekom JC, Janson AA, Ginjaar IB, Frankhuizen WS, Aartsma-Rus A, Bremmer-Bout M, et al. Local dystrophin res toration with antisense oligonucleotide PRO051. N Engl J Med. 2007 Dec 27;357:2677-86. http://dx.doi.org/10.1056/NEJMoa073108. [81] Pegoraro E, Hoffman EP, Piva L, Gavassini BF, Cagnin S, Ermani M, et al. SPP1 genotype is a determinant of disease severity in Duchenne muscular dystrophy. Neurology. 2011 Jan 18;76:219-26. http://dx.doi. org/10.1212/WNL.0b013e318207afeb. [69] Kinali M, Arechavala-Gomeza V, Feng L, Cirak S, Hunt D, Adkin C, et al. Local restoration of dystrophin expression with the morpholino oligomer AVI-4658 in Duchenne muscular dystro phy: a single-blind, placebo-controlled, dose-escalation, proof-of- concept study. Lancet Neurol. 2009 Oct;8:918-28. http://dx.doi. org/10.1016/S1474-4422(09)70211-X. g [82] Bello L, Piva L, Barp A, Taglia A, Picillo E, Vasco G, et al. Importance of SPP1 genotype as a covariate in clinical trials in Duchenne mus cular dystrophy. Neurology. 2012 Jul 10;79:159-62. http://dx.doi. org/10.1212/WNL.0b013e31825f04ea. Elisabeth Le Rumeur: Dystrophin in Duchenne and Becker muscular dystrophies http://dx.doi.org/10.2174/156652312800840603. [37] Ervasti J, Campbell K. A Role for the Dystrophin-Glycoprotein Complex as aTransmembrane Linker between Laminin and Actin. J Cell Biol. 1993;122:809-23. http://dx.doi.org/10.1083/jcb.122.4.809. p g j [38] Huang X, Poy F, Zhang R, Joachimiak A, Sudol M, Eck MJ. Structure of a WW domain containing fragment of dystrophin in complex with beta-dystroglycan. Nat Struct Biol. 2000 Aug;7:634-8. http:// dx.doi.org/10.1038/77923. g [55] Fairclough RJ, Wood MJ, Davies KE. Therapy for Duchenne muscu lar dystrophy: renewed optimism from genetic approaches. Nat Rev Genet. 2013 Jun;14:373-8. http://dx.doi.org/10.1038/nrg3460. [56] Jarmin S, Kymalainen H, Popplewell L, Dickson G. New develop ments in the use of gene therapy to treat Duchenne muscular dys trophy. Expert Opin Biol Ther. 2014 Feb;14:209-30. http://dx.doi.org /10.1517/14712598.2014.866087. [39] Hnia K, Zouiten D, Cantel S, Chazalette D, Hugon G, Fehrentz JA, et al. ZZ domain of dystrophin and utrophin: topol ogy and mapping of a beta-dystroglycan interaction site. Biochem J. 2007 Feb 1;401:667-77. http://dx.doi.org/10.1042/BJ20061051. [57] McGreevy JW, Hakim CH, McIntosh MA, Duan D. Animal mod els of Duchenne muscular dystrophy: from basic mechanisms to gene therapy. Dis Model Mech. 2015 Mar;8:195-213. http://dx.doi. org/10.1242/dmm.018424. [40] Rezniczek GA, Konieczny P, Nikolic B, Reipert S, Schneller D, Abrahamsberg C, et al. Plectin 1f scaffolding at the sarcolemma of dystrophic (mdx) muscle fibers through multiple interactions with beta-dystroglycan. J Cell Biol. 2007 Mar 26;176:965-77. http://dx. doi.org/10.1083/jcb.200604179. [58] Gregorevic P, Blankinship MJ, Allen JM, Chamberlain JS. Systemic microdystrophin gene delivery improves skeletal mus cle structure and function in old dystrophic mdx mice. Mol Ther. 2008 Apr;16:657-64. http://dx.doi.org/10.1038/mt.2008.28. [41] Ayalon G, Davis JQ, Scotland PB, Bennett V. An ankyrin-based mechanism for functional organization of dystrophin and dystro glycan. Cell. 2008 Dec 26;135:1189-200. http://dx.doi.org/10.1016/j. cell.2008.10.018. [59] Foster H, Sharp PS, Athanasopoulos T, Trollet C, Graham IR, Foster K, et al. Codon and mRNA sequence optimization of microdystrophin 19 Elisabeth Le Rumeur: Dystrophin in Duchenne and Becker muscular dystrophies Elisabeth Le Rumeur: Dystrophin in Duchenne and Becker muscular dystrophies [70] Goemans NM, Tulinius M, van den Akker JT, Burm BE, Ekhart PF, Heuvelmans N, et al. Systemic administration of PRO051 in Duchenne’s muscular dystrophy. N Engl J Med. 2011 Apr 21;364:1513-22. http://dx.doi.org/10.1056/NEJMoa1011367. [83] Flanigan KM, Ceco E, Lamar KM, Kaminoh Y, Dunn DM, Mendell JR, et al. LTBP4 genotype predicts age of ambulatory loss in Duchenne muscular dystrophy. Ann Neurol. 2013 Apr;73:481-8 http://dx.doi.org/10.1002/ana.23819. 20
https://openalex.org/W2125448293	https://europepmc.org/articles/pmc3466267?pdf=render	English	null	Extra-Hippocampal Subcortical Limbic Involvement Predicts Episodic Recall Performance in Multiple Sclerosis	PloS one	2,012	cc-by	7,246	Received March 16, 2012; Accepted August 15, 2012; Published October 8, 2012 Copyright: 2012 Dineen et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: The study was part funded by Nottingham University Hospitals Charity. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. No additional external funding received for this study. Competing Interests: The authors have declared that no competing interests exist. Competing Interests: The authors have declared that no competing interests exist. * E-mail: rob.dineen@nottingham.ac.uk * E-mail: rob.dineen@nottingham.ac.uk [9,10]. However, familiarity-based recognition is not thought to be dependent on the integrity of this system [9]. Extra-Hippocampal Subcortical Limbic Involvement Predicts Episodic Recall Performance in Multiple Sclerosis Robert A. Dineen1, Christopher M. Bradshaw2, Cris S. Constantinescu3, Dorothee P. Auer1 1 Division of Radiological and Imaging Sciences, University of Nottingham, Nottingham, United Kingdom, 2 Department of Psychopharmacology, University of Nottingham, Nottingham, United Kingdom, 3 Division of Clinical Neurology, University of Nottingham, Nottingham, United Kingdom October 2012 \| Volume 7 \| Issue 10 \| e44942 Abstract Background: Episodic memory impairment is a common but poorly-understood phenomenon in multiple sclerosis (MS). We aim to establish the relative contributions of reduced integrity of components of the extended hippocampal-diencephalic system to memory performance in MS patients using quantitative neuroimaging. Methodology/Principal Findings: 34 patients with relapsing-remitting MS and 24 healthy age-matched controls underwent 3 T MRI including diffusion tensor imaging and 3-D T1-weighted volume acquisition. Manual fornix regions-of-interest were used to derive fornix fractional anisotropy (FA). Normalized hippocampal, mammillary body and thalamic volumes were derived by manual segmentation. MS subjects underwent visual recall, verbal recall, verbal recognition and verbal fluency assessment. Significant differences between MS patients and controls were found for fornix FA (0.38 vs. 0.46, means adjusted for age and fornix volume, P,.0005) and mammillary body volumes (age-adjusted means 0.114 ml vs. 0.126 ml, P,.023). Multivariate regression analysis identified fornix FA and mammillary bodies as predictor of visual recall (R2 = .31, P = .003, P = .006), and thalamic volume as predictive of verbal recall (R2 = .37, P,.0005). No limbic measures predicted verbal recognition or verbal fluency. Conclusions/Significance: These findings indicate that structural and ultrastructural alterations in subcortical limbic components beyond the hippocampus predict performance of episodic recall in MS patients with mild memory dysfunction. ation: Dineen RA, Bradshaw CM, Constantinescu CS, Auer DP (2012) Extra-Hippocampal Subcortical Limbic Involvement Predicts Episo Multiple Sclerosis. PLoS ONE 7(10): e44942. doi:10.1371/journal.pone.0044942 Editor: Christian Beaulieu, University of Alberta, Canada Received March 16, 2012; Accepted August 15, 2012; Published October 8, 2012 Received March 16, 2012; Accepted August 15, 2012; Published October 8, 2012 Introduction Cognitive impairment occurs in 30%–70% of multiple sclerosis (MS) patients and commonly manifests as disturbances in memory, complex attention, efficiency of information processing, executive functioning, and processing speed [1]. Cognitive impairment in MS impacts on quality of life, vocational status and social functioning [2,3,4]. Impairment of episodic memory is one of the most common cognitive deficits in MS, with particular involve- ment of recent memory and to a lesser extent immediate and remote memory [5]. Both visual and verbal memory functions are affected being impaired in 56% and 34% of MS patients respectively [6]. However, conventional radiological measures of the extent of disease involvement, such as T2 lesion load, remain poor predictors of cognitive impairment in MS [7]. A number of studies have used imaging to study the impact of damage of isolated components of the extended hippocampal - diencephalic system on cognitive performance in MS, but there is a lack of studies employing a comparative quantitative approach to identify the relative importance of damage to components of the integrated system. Selective atrophy of the hippocampus has been described in relapsing-remitting multiple sclerosis (RRMS) [11], and this has been found to correlate with performance in tests of verbal learning/recall in MS patients [11,12]. Atrophy of the thalamus occurs in MS [13,14], and shows a significant relationship to cognitive performance when measured directly or indirectly [15,16,17,18]. Mammillary body atrophy has also been identified in MS [19], but the cognitive relevance of this finding has not, to our knowledge, been tested. Evidence from animal and human studies has shown that episodic memory is dependent on the extended hippocampal - diencephalic system [8]. This system consists of a series of direct and indirect parallel temporo-diencephalic pathways with damage to individual components leading to impairments in learning and recall, clinically manifesting as anterograde amnesia syndromes The fornix is the major efferent pathway from the hippocampus projecting to the mammillary bodies via the post-commissural fornix, as well as numerous other sites including the nucleus accumbens, septal nuclei and prefrontal cortex (via the pre- October 2012 \| Volume 7 \| Issue 10 \| e44942 PLOS ONE \| www.plosone.org 1 October 2012 \| Volume 7 \| Issue 10 \| e44942 Limbic Involvement and Episodic Memory in MS commissural fornix), the hypothalamus and anterior thalamic nuclei. Participants Thirty-four patients with definite relapsing remitting MS (RRMS) (median age 42.6 years, range 31.1 to 56.1, M:F = 1:2.1) were prospectively recruited from a local database of RRMS patients after giving written informed consent. Participants with MS had median disease duration 10 years (range 3 to 27 years) and median Expanded Disability Status Scale (EDSS) of 2.5 (range 1.5 to 6.5) at the time of the study. Participants with MS were not taking immunomodulatory therapy and had not had a relapse or steroid therapy for at least two months prior to inclusion. Twenty-four healthy controls (median age 38.7, age range 28.3 to 55.3, M:F = 1:1.7) were included. Participants were excluded if they had any other significant neurological, medical, psychiatric or cognitive disorder. Analyses based on this participant cohort have been reported in a previous publication [21]. MRI acquisition and image processing The measurements were made by a single trained academic neuroradiologist (RAD) blinded to disease, cognitive and demographic status. Anatomical boundaries of the hippocampus were defined according to Pantel Figure 1. Volumetric segmentation of brain images. (a–c) Thalamic segmentation. Axial MPRAGE image from a control subject (a) before and (b) after the addition of FA weighting from the co- registered FA image. Note the increase in clarity of the lateral border of the thalamus that facilitates (c) manual segmentation of the thalamus. (d–g) axial, coronal and sagittal images from a participant with MS demonstrating ROIs drawn for the mammillary bodies (blue) and right (yellow) and left (green) hippocampi. (h) Surface rendered representa- tion of the segmented hippocampi and mammillary bodies from a participant with MS. The manually defined fornix ROI used to derive fornix FA is also shown (grey). doi:10.1371/journal.pone.0044942.g001 MRI acquisition and image processing MRI acquisition and image processing All subjects underwent MRI scanning at 3 Tesla (Achieva, Philips, Eindhoven, NL), including axial DTI (Single-shot diffusion weighted EPI, b = 1000 s/mm2, 15 directions, TE = 56 ms, TR = 9700 ms, 2 mm62 mm62.5 mm voxel size interpolated to 1 mm61 mm62.5 mm voxels, 45 interleaved slices with no gap, four averages), sagittal MPRAGE (TR= 7.5 ms, TE = 2.2 ms, Flip= 8, Matrix 2566256, voxel size= 0.8 mm60.8 mm60.8 mm, TFE factor = 236, FOV = 205), and axial FLAIR (TR = 11 000 ms TE = 125 ms TI = 2800 ms, matrix 256_256, slice thickness 2.5 mm/0 mm gap, FOV= 256). DTI data underwent eddy current correction and removal of non-brain voxels [31], and was further post-processed using the fMRIB diffusion toolbox (FSL version 3.3 [32]), to give individual subject maps of FA, axial diffusivity (AD) and radial diffusivity (RD). Volumetric measurements (Figure 1). Semi-automated brain extraction of the MPRAGE images was performed using the Brain Finder tool within JIM 3.0 software (Xinapse Medical Systems, UK) to provide total intracranial volume (TICV) and brain parenchymal volume, from which brain parenchymal fraction (BPF) was derived (BPV/TICV). Volumetric measurements (Figure 1). Semi-automated brain extraction of the MPRAGE images was performed using the Brain Finder tool within JIM 3.0 software (Xinapse Medical Systems, UK) to provide total intracranial volume (TICV) and brain parenchymal volume, from which brain parenchymal fraction (BPF) was derived (BPV/TICV). The purpose of this study is thus to test the relative contributions of loss of structural integrity in components of the extended hippocampal–diencephalic system to cognitive performance in MS. In particular we aim to identify the relative contribution of reduced hippocampal, thalamic and mammillary body volumes and reduced fornix tract integrity to performance in tests of visual memory/recall, verbal learning and recall, verbal recognition and a non-memory dependent test of verbal fluency. We go beyond our previous whole brain DTI analysis of this cohort [21] and the work of others who have studied integrity of components of this system in MS, by creating regression models using the limbic integrity measures (and including potential confounders) to identify the limbic structures for which reduced integrity most predicts episodic memory performance. Volumes of the hippocampi, mammillary bodies and thalami were obtained manually from the MPRAGE images using 3D- Slicer software (MIT, Cambridge, MA, USA). Introduction Three recent studies using tract-based spatial statistics [20] have identified reduced fractional anisotropy (FA, a measure of white matter integrity derived from diffusion tensor imaging, DTI) in the fornices of patients with multiple sclerosis compared to healthy controls [21,22,23]. Roosendaal et al report increases in both axial and radial diffusivity within the fornix of MS patients compared to the controls, with the elevation of radial diffusivity being more significant than axial diffusivity [22]. Kern et al have conducted a combined fMRI and DTI study in 18 MS patients and 16 controls and again demonstrated the relationship between structural damage to the fornix an poorer performance in a verbal memory task [24]. Furthermore, they found differences in hippocampal activation during a verbal memory task that were related to fornix FA, interpreted as being a limiting effect of damage to the fornix on disease-related compensatory activity. Neuropsychological testing To aid visualization of the lateral edge of the thalamus (which merges with the internal capsule on MPRAGE images), MPRAGE images were weighted by FA as follows: The subjects FA map was first coregistered to the MPRAGE using FSL FLIRT, following which the accuracy of the coregistration was checked by careful visual inspection. The MPRAGE images were then multiplied by (1+FA) to give the FA- weighted MPRAGE image. The high FA in the internal capsule provides a clear position of the lateral thalamic boundary, which aided the manual thalamic outlining (Figure 1 a–c). Hippocampal, mammillary body and thalamic volumes were normalized to head size according to the analysis of covariance approach proposed by Jack Jr et al [35]. Figure 2. Identification of focal MS lesions of the fornix. (a) Axial FLAIR image from one of the participants with MS showing focal T2- hyperintesity in crus of the left fornix (LFx) with accompanying T1- hypointensity on the corresponding MPRAGE image (b). Crus of the right fornix labelled for comparison (RFx). doi:10.1371/journal.pone.0044942.g002 Hippocampal, thalamic and mammillary body volumes were re- measured in a subset of 10 MS patients (thus 20 of each structure) after an interval of at least one week to allow assessment of intra- observer reliability (Type A intraclass correlation coefficients using an absolute agreement definition). Inter-observer reliability was performed in a similar way, with the repeat measurements made by a second trained researcher. parametric tests. Comparison of age between the MS patients and controls was made using Independent-Samples T-Test. Between group comparisons of normalized hippocampal volumes, normal- ized thalamic volumes and normalized mammillary body volumes were performed using analysis of covariance (ANCOVA) treating group as a fixed factor and age as a covariate. Group comparisons of forniceal ROI FA, AD and RD values were performed using ANCOVA with age and fornix volume included as covariates, the latter included to account for account for partial volume edge effects on the FA measures [36]. Group comparison was also planned to test whether fornix FA differed in those with and without focal T2 hyperintense lesion in the fornix. Derivation of the fornix FA. The Fornix ROI was drawn manually onto the MPRAGE images as follows. Using the sagittal plane, a marker was placed at the point where the fimbria-fornix detached from the hippocampal tail to become the fornix proper. Neuropsychological testing All MS patients underwent standardized neuropsychological testing by a single researcher on the day of the MRI scan. The tests performed were the Benton Visual Retention Test (BVRT), a test of visuospatial memory [25], California Verbal Learning Test version II (CVLT-II), a test of verbal memory [26] and Controlled Oral Word Association Test (COWAT), a test of verbal fluency [27]. CVLT-II subscores used were the short delay free recall (CVLT-II recall) and the total recognition discriminability (CVLT-II recognition). The National Adult Reading Test (NART) was performed in all MS subjects and used to estimate pre-morbid IQ (WAIS FS-IQ) [28]. Unadjusted scores or Z-transforms (based on published datasets [26,29,30]) were used as described previously [21]. Figure 1. Volumetric segmentation of brain images. (a–c) Thalamic segmentation. Axial MPRAGE image from a control subject (a) before and (b) after the addition of FA weighting from the co- registered FA image. Note the increase in clarity of the lateral border of the thalamus that facilitates (c) manual segmentation of the thalamus. (d–g) axial, coronal and sagittal images from a participant with MS demonstrating ROIs drawn for the mammillary bodies (blue) and right (yellow) and left (green) hippocampi. (h) Surface rendered representa- tion of the segmented hippocampi and mammillary bodies from a participant with MS. The manually defined fornix ROI used to derive fornix FA is also shown (grey). doi:10.1371/journal.pone.0044942.g001 PLOS ONE \| www.plosone.org October 2012 \| Volume 7 \| Issue 10 \| e44942 October 2012 \| Volume 7 \| Issue 10 \| e44942 2 Limbic Involvement and Episodic Memory in MS Figure 2. Identification of focal MS lesions of the fornix. (a) Axial FLAIR image from one of the participants with MS showing focal T2- hyperintesity in crus of the left fornix (LFx) with accompanying T1- hypointensity on the corresponding MPRAGE image (b). Crus of the right fornix labelled for comparison (RFx). doi:10.1371/journal.pone.0044942.g002 and co-workers [33]. The mammillary bodies were outlined using axial, sagittal and coronal planes according to previously described boundaries [34]. The thalami were outlined in the axial plane, commencing on the slice at the level of the anterior commissure (below which reliable separation of the thalamus from subthalamic structures is difficult) and proceeding dorsally to the most dorsal slice on which the thalamus could clearly be visualized. Neuropsychological testing This marker was then identified in the axial plane as the commencement point for the fornix ROI. The ROI was then drawn to include the crus of the fornix using successively more dorsal slices, before switching to the coronal plane to continue the ROI rostrally through the bodies of the fornix. The axial plane was then used to define the ROI down the fornix columns, terminating at a point immediately dorsal to the mammillary bodies. Care was taken to restrict the ROI to voxels within the fornix, with exclusion of voxels lying on the margin of the fornix which would be prone to CSF contamination due to partial volume effects. Both left and right sides of the fornix were defined in this way, resulting in a single fornix ROI that included left, right and midline fornix components. Finally, the placement of the ROI was checked the axial, sagittal and coronal planes, and edited as appropriate. The ROI was then used to extract mean fornix FA, AD and RD from the coregistered DTI image, and to calculate the fornix ROI volume. To test for independent contributions of structural and ultrastructural alterations on neuropsychological performance we carried out standard multivariate stepwise regression treating the neuropsychological scores as the dependent variable. The main imaging variables of interest (fornix FA and normalized hippo- campal, thalamic and mammillary body volumes), along with age, brain parenchymal volume and estimated premorbid IQ (consid- ered as potential confounding variables), and normalized fornix volume (included to control for partial volume effects of CSF signal on fornix DTI measures) were included. The model used entry and exit criteria of P,0.05 and P,0.10 respectively. The regressions were repeated with fornix AD and RD substituted in place of fornix FA. In addition, the FLAIR images were carefully inspected for the presence of focal T2 hyperintense lesions. When focal T2 hyperintensity was detected in the fornix, the relevant MPRAGE images were inspected for corresponding T1, which was found to help confirm the presence of a genuine focal lesion (Figure 2). Ethics The study was approved by the UK National Health Service Research Ethics Committee (Nottingham 1 Committee), and all participants in this study gave written informed consent. Comparison of MS patients and controls The difference in age between MS patients and controls was not statistically significant (42.6 years vs 38.7, P = .07). Mean fornix FA was strongly reduced in MS patients vs. controls (means adjusted for age and normalised fornix volume: 20.37 vs 0.46, F (1,55) = 34.6, P,.0005, Figure 3a). Age was a significant covariate in the ANCOVA model (F (1,55) = 5.7, P = .021), but normalised fornix volume was not (P = .062). A group difference was also identified for mean fornix RD (means adjusted for age and normalised fornix volume: 0.00129 vs 00100, F (1,55) = 27.8, P,.0005), but not for mean fornix AD. Statistical analysis Statistical analysis was performed using IBM SPSS Statistics 19.0 (IBM Corporation. Somers, NY). A significance threshold of a = 0.05 was used, unless stated otherwise. Normality of data distribution was assessed by the Shapiro-Wilk test prior to use of Intra- and inter-observer reliability measures of volumes derived by manual segmentation were good (table 1). The group difference in normalized hippocampal volume fell just outside the predefined October 2012 \| Volume 7 \| Issue 10 \| e44942 PLOS ONE \| www.plosone.org October 2012 \| Volume 7 \| Issue 10 \| e44942 3 Limbic Involvement and Episodic Memory in MS significance threshold (age-adjusted means: 5.38 ml vs 5.71 ml, F (1,56) = 3.8, P = .057, Figure 3b). Age was not a significant covariate in the ANCOVA model (P = .59). A significant group difference was found for normalized mammillary body volumes (age-adjusted means: 0.114 ml vs 0.126 ml, F (1,56) = 5.5, P = .023, Figure 3c). Age was not a significant covariate in the Figure 3. Boxplots showing mean (a) mean fornix FA, (b) normalized total hippocampal volumes, (c) normalized total mammillary body volumes and (d) normalized total thalamic volumes for the MS and control groups. doi:10.1371/journal.pone.0044942.g003 PLOS ONE \| www plosone org 4 October 2012 \| Volume 7 \| Issue 10 \| e44942 significance threshold (age-adjusted means: 5.38 ml vs 5.71 ml, F (1 56) = 3 8 P = 057 Figure 3b) Age was not a significant difference was found for normalized mammillary body volumes (age adjusted means: 0 114 ml vs 0 126 ml F (1 56) = 5 5 Figure 3. Boxplots showing mean (a) mean fornix FA, (b) normalized total hippocampal volumes, (c) normalized total mammillary body volumes and (d) normalized total thalamic volumes for the MS and control groups. doi:10.1371/journal.pone.0044942.g003 Figure 3. Boxplots showing mean (a) mean fornix FA, (b) normalized total hippocampal volumes, (c) normalized total mammillary body volumes and (d) normalized total thalamic volumes for the MS and control groups. doi:10.1371/journal.pone.0044942.g003 significance threshold (age-adjusted means: 5.38 ml vs 5.71 ml, F (1,56) = 3.8, P = .057, Figure 3b). Age was not a significant covariate in the ANCOVA model (P = .59). A significant group difference was found for normalized mammillary body volumes (age-adjusted means: 0.114 ml vs 0.126 ml, F (1,56) = 5.5, P = .023, Figure 3c). Age was not a significant covariate in the Figure 3. Statistical analysis Boxplots showing mean (a) mean fornix FA, (b) normalized total hippocampal volumes, (c) normalized total mammillary body volumes and (d) normalized total thalamic volumes for the MS and control groups. doi:10.1371/journal.pone.0044942.g003 difference was found for normalized mammillary body volumes (age-adjusted means: 0.114 ml vs 0.126 ml, F (1,56) = 5.5, P = .023, Figure 3c). Age was not a significant covariate in the difference was found for normalized mammillary body volumes (age-adjusted means: 0.114 ml vs 0.126 ml, F (1,56) = 5.5, P = .023, Figure 3c). Age was not a significant covariate in the difference was found for normalized mammillary body volumes (age-adjusted means: 0.114 ml vs 0.126 ml, F (1,56) = 5.5, P = .023, Figure 3c). Age was not a significant covariate in the significance threshold (age-adjusted means: 5.38 ml vs 5.71 ml, F (1,56) = 3.8, P = .057, Figure 3b). Age was not a significant covariate in the ANCOVA model (P = .59). A significant group PLOS ONE \| www.plosone.org PLOS ONE \| www.plosone.org October 2012 \| Volume 7 \| Issue 10 \| e44942 October 2012 \| Volume 7 \| Issue 10 \| e44942 4 Limbic Involvement and Episodic Memory in MS Table 1. Intra- and inter-observer reliability assessment for the volumetric measurement, made in a subset of 10 MS patients (thus 20 of each structure). Intra-observer reliability Inter-observer reliability Structure Intraclass correlation coefficient 95% confidence intervals Significance Intraclass correlation coefficient* 95% confidence intervals Significance Hippocampal volume 0.93 0.43–0.98 P,0.0005 0.86 0.67–0.94 P,0.0005 Thalamic volume 0.96 0.71–0.98 P,0.0005 0.83 0.17–0.94 P,0.0005 Mammillary body volume 0.92 0.78–0.97 P,0.0005 0.83 0.54–0.94 P,0.0005 Type A intraclass correlation coefficients using an absolute agreement definition. doi:10.1371/journal.pone.0044942.t001 Table 1. Intra- and inter-observer reliability assessment for the volumetric measurement, made in a subset of 10 MS patients (thus 20 of each structure). Type A intraclass correlation coefficients using an absolute agreement definition. doi:10.1371/journal.pone.0044942.t001 Regression analysis of local and global imaging-derived measure for neuropsychological performance (Table 3) ANCOVA model (P = .12). Normalized thalamic volumes were not different between MS patients and controls (age adjusted means: 9.53 ml vs 10.36 ml, P = .10) (Figure 3d). For BVRT, fornix FA and normalised mammillary body volumes were retained in the model (R2 = .31, P = .003, Figure 5). For CVLT-II recall, normalized thalamic volume was retained in the final regression (R2 = .37, P,.0005). Neuropsychological test performance (table 2) The MS subjects showed no impairment of estimated premor- bid intelligence (WAIS FS_IQ score mean = 102.5, range = 81 to 118, s.d. = 10, as derived from NART). Mean neuropsychological scores fell within one standard deviation of the normative mean in keeping with a minimal overall neuropsychological burden, although the cohort included individuals with a range of performance from above to more than two standard deviations below the normative mean for each test. Figure 4. Scatterplot showing mean fornix FA for the MS participants with and without visible T2 hyperintense lesions in the fornix. doi:10.1371/journal.pone.0044942.g004 Statistical analysis For CVLT-II recognition and COWAT scores, the final regression models consisted of only a single non-limbic variable (BPF and estimated premorbid IQ respectively) and accounted for only a small proportion of the variance. Substitution of mean fornix AD or RD in place of fornix FA in the regression did not improve the strength Discussion This study identifies reduced structural integrity in the limbic system beyond the hippocampus as being an important substrate underlying episodic memory performance in MS. Fornix FA and thalamic volume contributed significantly to regression models of episodic memory performance corrected for age and estimated premorbid IQ, whereas global neuroimaging measure of brain atrophy was not predictive. This relationship was not found for verbal recognition or a verbal fluency task. Our data therefore support an anatomical dissociation between visual/verbal recall Recent evidence demonstrating that fornix FA is reduced in both Alzheimer Disease (AD) and MCI, and that fornix FA can discriminate between MCI patients and controls [42], is pertinent to the current results; it is possible that the cognitively-relevant fornix FA reduction in the MS patients reflects a shared anatomical substrate for cognitive dysfunction with MCI patients. Furthermore, attention has recently focused on the role of disruption of large scale brain networks in neurodegenerative disorders [43]. In AD, episodic memory and visuospatial functions are affected early in the disease course, and this is known to be associated with disruption of the default mode network (DMN), which includes medial prefrontal cortex, posterior hippocampus, and posterior cingulate cortex/retrosplenial cortex [44,45]. The DMN has been shown to be altered in MCI patients [46] and those at risk of developing AD [47]. It is possible that the cognitively-relevant fornix FA reduction shown here contributes to altered hippocampal connectivity to other nodes of the DMN. The recent finding by Roosendaal et al that reduced functional connectivity of the hippocampus is detectable in MS patients prior to the onset of measurable hippocampal atrophy [48] supports this notion. Reduced ultrastructural integrity of the fornix provides a potential substrate for the altered hippocampal connectivity but longitudinal studies will be of importance to establish the complex temporal relationships between fornix degeneration, hippocampal atrophy, hippocampal functional connectivity and deterioration in episodic memory performance in MS. Figure 5. Scatterplot showing the relationship between mean fornix FA and BVRT scores in the MS group. doi:10.1371/journal.pone.0044942.g005 Unlike previous studies, we found no relationship between hippocampal volume and memory performance. Relationships between fornix FA and normalized hippocampal, mammillary body and thalamic volumes Considering all subjects, fornix FA correlated with thalamic volumes (R = .45, P,0.0005), but not with hippocampal volumes. When subjects with MS were considered separately, correlation was found between fornix FA and normalized thalamic volumes (R = .55, P,0.0005), but no other correlation was found. It was not possible to formally test whether fornix FA differed in those with and without focal T2 hyperintense lesion in the fornix, as only four MS subjects had a confirmed T2 hyperintense lesion, but the scatterplot of FA for these two groups shows no clear difference in fornix FA between these groups (Figure 4). Table 2. Neuropsychological test Z-scores (mean and range) for the MS cohort. One sample T- test* Test n Mean Z score Z-score range t p BVRT 34 20.35 23.1 to 2.1 21.48 n/s CVLT-II recall 34 20.16 22.5 to 1.5 20.81 n/s CVLT-II recognition 34 0.02 23.0 to 1.5 0.08 n/s COWAT 34 20.56 23.8 to 1.7 22.78 0.009 *against normative mean value of 0. doi:10.1371/journal.pone.0044942.t002 Figure 4. Scatterplot showing mean fornix FA for the MS participants with and without visible T2 hyperintense lesions in the fornix. doi:10.1371/journal.pone.0044942.g004 doi:10.1371/journal.pone.0044942.g004 October 2012 \| Volume 7 \| Issue 10 \| e44942 5 Limbic Involvement and Episodic Memory in MS Table 3. Results of the regression analysis for performance in the four neuropsychological tests. Table 3. Results of the regression analysis for performance in the four neuropsychological tests. Final regression model Unstandardized coefficients Neuropsychological test R2 Variables retained B (95% CI) Standard error Standardized coefficient (b) Significance (P) BVRT 0.31 Fornix FA 12.24 (3.44–21.04) 4.32 0.42 .008 nMBV 0.30 (0.00–0.06) 0.14 0.32 0.038 (constant) 21.09 (25.50–3.32) 2.16 - .618 CVLT-II recall 0.37 nTV 0.95 (0.50–1.39) 0.22 0.61 ,.0005 (constant) 1.70 (22.62–6.01) 2.12 - .428 CVLT-II recognition 0.18 BPF 7.88 (1.42–14.34) 3.16 0.42 .019 (constant) 23.63 (29.16–1.91) 2.71 - .191 COWAT 0.14 Estimated premorbid IQ 0.41 (0.04–0.79) 0.18 0.37 .032 (constant) 27.91 (246.31–30.49) 18.85 - .678 Abbreviations: nMBV = normalized mammillary body volume; nTV = normalized thalamic volume; BPF = brain parenchymal fraction; CI = confidence interval. doi:10.1371/journal.pone.0044942.t003 = normalized mammillary body volume; nTV = normalized thalamic volume; BPF = brain parenchymal fraction; CI = confidence interval. one.0044942.t003 or significance of the regression models for any of the neuropsy- chological tests. and verbal recognition, with the latter showing no statistical relationship with fornix and diencephalic integrity. Relationships between fornix FA and normalized hippocampal, mammillary body and thalamic volumes Animal models have shown impaired recall but preserved recognition memory of fornix disruption in animals with fornix lesions [37,38,39]. Data from human studies is more limited but a similar pattern of recall impairment with spared recognition has been observed in post- surgical patients with damage to the fornix and mammillary body atrophy [40,41]. October 2012 \| Volume 7 \| Issue 10 \| e44942 Discussion As the volume reduction of the hippocampus in the MS patients was modest and did not reach significance, our findings do not necessarily contradict previous studies showing a relationship between hippocampal atrophy and memory function in MS; instead, it is possible that alterations of the fornix and diencephalic structures predominates for mild degrees of cognitive dysfunction in MS. For Figure 5. Scatterplot showing the relationship between mean fornix FA and BVRT scores in the MS group. doi:10.1371/journal.pone.0044942.g005 PLOS ONE \| www.plosone.org October 2012 \| Volume 7 \| Issue 10 \| e44942 October 2012 \| Volume 7 \| Issue 10 \| e44942 6 Limbic Involvement and Episodic Memory in MS these processes. Indeed, as Rudebeck and colleagues have shown fornix microstructure relates to episodic recall in healthy volunteers [49]. example, in the study by Sicotte et al which found a relationship between hippocampal volume and verbal memory in MS patients, both the RRMS (n = 23) and SPMS (n = 11) groups showed significant impairment of verbal memory compared to the control group (n = 18). In line with previous studies, we found a relationship between cognitive performance and thalamic volumes [14,18], although contrary to previously published studies the reduction in normalised thalamic volumes in the current MS cohort did not reach statistical significance. A further limitation that we acknowledge is that there is potential for fornix FA to be altered by partial volume effects from adjacent CSF. We defined the fornix ROI on MPRAGE images and then extracted the DTI metrics from linearly coregistered images. We acknowledge that potential registration errors may arise that could affect fornix diffusion measurements, but have attempted to minimise this by careful inspection of the coregistered images for errors, particularly in the region of the fornix. While many studies reporting DTI metrics do not correct for partial volume effects, we have attempted to control for this by including normalised fornix volume in the statistical analyses that include FA measures. A similar but less anatomically-specific approach has previously been used by Rashid et al, who used whole brain parenchymal volumes to perform a CSF correction on DTI data from MS patients [36]. Recently this approach has been criticised [50] and a more robust partial volume correction using free water elimination has been advocated [51]. However, this approach has not yet been widely implemented in clinical studies. Limitations The 34 MS participants had range of cognitive performance levels, from those with normal performance through to those with performance at and beyond 2.5 standard deviations below the normative population mean, but only a small number had evidence of significant cognitive dysfunction (for example, only four had BVRT Z-score #22). It is therefore possible that the observed relationship between episodic memory performance and the imaging metrics of limbic integrity may be contributed to by inherent premorbid variations in these measures between individ- uals, as opposed to being the result of disease-specific changes in The Authors would like to acknowledge the input of Anis Fozi and Emily Tombs for assistance with image analysis, of Dr Janek Vilisaar and Dr Hari Krishnan for EDSS scoring of MS patients and of Dr Paul S. Morgan for assistance with the MRI protocol and scanning. Discussion It is notable that the current analysis did not find a relationship between fornix FA and CVLT-II recall performance while our previous TBSS analysis of this dataset identified clusters in the left fornix where reduced FA correlated with CVLT-II recall performance. This difference presumably relates to the use of a whole fornix ROI including both left and right sides of the fornix in the current study compared to the voxelwise approach employed by TBSS, The final regression models for visual and verbal recall only accounted for 31% and 21% of variance in performance respectively, thus a large proportion of the variance remains unaccounted for which there are a number of possible explana- tions. Firstly the measures of integrity chosen for this analysis (fornix FA and limbic subcortical gray matter volumes) may not have sufficient sensitivity to measure subtle alterations in tissue integrity of these structures. Secondly, pre-morbid memory performance is likely to account in part for the variability, and we have attempted to control for pre-morbid cognitive ability using an estimate of premorbid IQ. Thirdly, as shown in our previous analysis of this cohort brain white matter tract integrity outside the limbic system related to both verbal and visual memory performance, and hence extra-limbic brain integrity may also contribute to the unexplained variance in episodic memory performance. In summary, the study establishes the contribution of reduced integrity of the fornix and diencephalic limbic structures to episodic memory performance in MS, a finding which is in line with the current literature from animal studies and human studies. This is, to our knowledge, the first study to demonstrate that reduced integrity in these structures plays a predominant role over hippocampal volume loss in episodic memory performance in MS. Our results provide further evidence for the notion that cognitive performance in MS is influenced by a combination of white matter disruption resulting in disconnection phenomena, and grey matter atrophy. 7. Barkhof F (2002) The clinico-radiological paradox in multiple sclerosis revisited. Current Opinion in Neurology 15: 239–245. Author Contributions Conceived and designed the experiments: RAD CMB CSC DPA. Performed the experiments: RAD DPA. Analyzed the data: RAD DPA. Contributed reagents/materials/analysis tools: RAD CMB CSC DPA. Wrote the paper: RAD CMB CSC DPA. Conceived and designed the experiments: RAD CMB CSC DPA. Performed the experiments: RAD DPA. Analyzed the data: RAD DPA. Contributed reagents/materials/analysis tools: RAD CMB CSC DPA. Wrote the paper: RAD CMB CSC DPA. 14. Houtchens MK, Benedict RHB, Killiany R, Sharma J, Jaisani Z, et al. (2007) Thalamic atrophy and cognition in multiple sclerosis. Neurology 69: 1213–1223. 13. Cifelli A, Arridge M, Jezzard P, Esiri MM, Palace J, et al. (2002) Thalamic neurodegeneration in multiple sclerosis. Annals of Neurology 52: 650–653. 12. Benedict RH, Ramasamy D, Munschauer F, Weinstock-Guttman B, Zivadinov R (2009) Memory impairment in multiple sclerosis: correlation with deep grey matter and mesial temporal atrophy. J Neurol Neurosurg Psychiatry 80: 201– 206. 11. Sicotte NL, Kern KC, Giesser BS, Arshanapalli A, Schultz A, et al. (2008) Regional hippocampal atrophy in multiple sclerosis. Brain 131: 1134–1141. 13. Cifelli A, Arridge M, Jezzard P, Esiri MM, Palace J, et al. (2002) Thalamic neurodegeneration in multiple sclerosis. Annals of Neurology 52: 650–653. 14. Houtchens MK, Benedict RHB, Killiany R, Sharma J, Jaisani Z, et al. (2007) Thalamic atrophy and cognition in multiple sclerosis. Neurology 69: 1213–1223. Limbic Involvement and Episodic Memory in MS Vann SD, Denby C, Love S, Montaldi D, Renowden S, et al. (2008) Memory loss resulting from fornix and septal damage: impaired supra-span recall but preserved recognition over a 24-hour delay. Neuropsychology 22: 658–668. 22. Roosendaal SD, Geurts JJG, Vrenken H, Hulst HE, Cover KS, et al. (2009) Regional DTI differences in multiple sclerosis patients. Neuroimage 44: 1397– 1403. 42. Bozoki AC, Korolev IO, Davis NC, Hoisington LA, Berger KL (2011) Disruption of limbic white matter pathways in mild cognitive impairment and Alzheimer’s disease: A DTI/FDG-PET Study. Human Brain Mapping: n/a–n/ a. 23. Yu HJ, Christodoulou C, Bhise V, Greenblatt D, Patel Y, et al. (2012) Multiple white matter tract abnormalities underlie cognitive impairment in RRMS. Neuro Image 59: 3713–3722. 43. Seeley WW, Crawford RK, Zhou J, Miller BL, Greicius MD (2009) Neurodegenerative Diseases Target Large-Scale Human Brain Networks. Neuron 62: 42–52. 24. Kern KC, Ekstrom AD, Suthana NA, Giesser BS, Montag M, et al. (2012) Fornix damage limits verbal memory functional compensation in multiple sclerosis. Neuroimage 59: 2932–2940. 44. Greicius MD, Srivastava G, Reiss AL, Menon V (2004) Default-mode network activity distinguishes Alzheimer’s disease from healthy aging: evidence from functional MRI. Proceedings of the National Academy of Sciences 101: 4637– 4642. g 25. Benton AL (1974) The Revised Visual Retention Test (4th ed). New York: Psychological Corporation. 26. Delis DC, Kramer JH, Kaplan E, Ober BA (2000) California Verbal Learning Test Manual. Second Edition. San Antonio: The Psychological Corporation. 45. Buckner RL, Snyder AZ, Shannon BJ, LaRossa G, Sachs R, et al. (2005) Molecular, Structural, and Functional Characterization of Alzheimer’s Disease: Evidence for a Relationship between Default Activity, Amyloid, and Memory. The Journal of Neuroscience 25: 7709–7717. y g p 27. Benton AL, Hamsher KD, Sivan AB (1983) Multilingual Aphasia Examination (3rd ed.). Iowa: AJA Associates. 28. Nelson HE, Willison J (1991) National Adult Reading Test (NART): Test Manual (Part II). 46. Petrella JR, Sheldon FC, Prince SE, Calhoun VD, Doraiswamy PM (2011) Default mode network connectivity in stable vs progressive mild cognitive impairment. Neurology 76: 511–517. 29. Youngjohn JR, Larrabee GJ, Crook III TH (1993) New adult age- and education-correction norms for the Benton Visual Retention Test. The Clinical Neuropsychologist 7: 155–160. 47. Sorg C, Riedl V, Mu¨hlau M, Calhoun VD, Eichele T, et al. (2007) Selective changes of resting-state networks in individuals at risk for Alzheimer’s disease. Proceedings of the National Academy of Sciences 104: 18760–18765. 30. Limbic Involvement and Episodic Memory in MS 15. Turner B, Ramli N, Blumhardt LD, Jaspan T (2001) Ventricular enlargement in multiple sclerosis: a comparison of three-dimensional and linear MRI estimates. Neuroradiology 43: 608–614. 35. Jack Jr CR, Twomey CK, Zinsmeister AR, Sharbrough FW, Petersen RC, et al. (1989) Anterior Temporal Lobes and Hippocampal Formations: Normative Volumetric Measurements from MR Images in Young Adults. Radiology 172: 549–554. gy 16. Rao SM, Glatt S, Hammeke TA, McQuillen MP, Khatri BO, et al. (1985) Chronic progressive multiple sclerosis. Relationship between cerebral ventricular size and neuropsychological impairment. Archives of Neurology 42: 678–682. 36. Rashid W, Hadjiprocopis A, Griffin CM, Chard DT, Davies GR, et al. (2004) Diffusion tensor imaging of early relapsing-remitting multiple sclerosis with histogram analysis using automated segmentation and brain volume correction. Multiple Sclerosis 10: 9–15. p y g p gy 17. Tsolaki M, Drevelegas A, Karachristianou S, Kapinas K, Divanoglou D, et al. (1994) Correlation of dementia, neuropsychological and MRI findings in multiple sclerosis. Dementia 5: 48–52. 37. Easton A, Zinkivskay A, Eacott MJ (2009) Recollection is impaired, but familiarity remains intact in rats with lesions of the fornix. Hippocampus 19: 837–843. p 18. Benedict RH, Weinstock-Guttman B, Fishman I, Sharma J, Tjoa CW, et al. (2004) Prediction of neuropsychological impairment in multiple sclerosis: comparison of conventional magnetic resonance imaging measures of atrophy and lesion burden. Archives of Neurology 61: 226–230. 38. Charles DP, Gaffan D, Buckley MJ (2004) Impaired Recency Judgments and Intact Novelty Judgments after Fornix Transection in Monkeys. J Neurosci 24: 2037–2044. and lesion burden. Archives of Neurology 61: 226–230. 19. Pagani E, Rocca MA, Gallo A, Rovaris M, Martinelli V, et al. (2005) Regional brain atrophy evolves differently in patients with multiple sclerosis according to clinical phenotype. AJNR: American Journal of Neuroradiology 26: 341–346. 39. Buckley MJ, Wilson CR, Gaffan D (2008) Fornix transection impairs visuospatial memory acquisition more than retrieval. Behav Neurosci 122: 44–53. 20. Smith SM, Jenkinson M, Johansen-Berg H, Rueckert D, Nichols TE, et al. (2006) Tract-based spatial statistics: voxelwise analysis of multi-subject diffusion data. Neuroimage 31: 1487–1505. 40. Tsivilis D, Vann SD, Denby C, Roberts N, Mayes AR, et al. (2008) A disproportionate role for the fornix and mammillary bodies in recall versus recognition memory. Nat Neurosci 11: 834–842. g 21. Dineen RA, Vilisaar J, Hlinka J, Bradshaw CM, Morgan PS, et al. (2009) Disconnection as a mechanism for cognitive dysfunction in multiple sclerosis. Brain 132: 239–249. 41. References 8. Aggleton JP, Brown MW (1999) Episodic memory, amnesia, and the hippocampal-anterior thalamic axis. Behavioral & Brain Sciences 22: 425– 444; discussion 444–489. 1. Chiaravalloti ND, DeLuca J (2008) Cognitive impairment in multiple sclerosis. The Lancet Neurology 7: 1139–1151. 2. Mitchell AJ, Benito-Leo´n J, Gonza´lez J-MM, Rivera-Navarro J (2005) Quality of life and its assessment in multiple sclerosis: integrating physical and psychological components of wellbeing. The Lancet Neurology 4: 556–566. 9. Aggleton JP, O’Mara SM, Vann SD, Wright NF, Tsanov M, et al. (2010) Hippocampal-anterior thalamic pathways for memory: uncovering a network of direct and indirect actions. European Journal of Neuroscience 31: 2292–2307. 3. Benedict RHB, Wahlig E, Bakshi R, Fishman I, Munschauer F, et al. (2005) Predicting quality of life in multiple sclerosis: accounting for physical disability, fatigue, cognition, mood disorder, personality, and behavior change. Journal of the Neurological Sciences 231: 29–34. 10. Aggleton JP, Pearce JM (2001) Neural systems underlying episodic memory: insights from animal research. Philos Trans R Soc Lond B Biol Sci 356: 1467– 1482. 11. Sicotte NL, Kern KC, Giesser BS, Arshanapalli A, Schultz A, et al. (2008) Regional hippocampal atrophy in multiple sclerosis. Brain 131: 1134–1141. 4. Rao SM, Leo GJ, Ellington L, Nauertz T, Bernardin L, et al. (1991) Cognitive dysfunction in multiple sclerosis. II. Impact on employment and social functioning. Neurology 41: 692–696. 12. Benedict RH, Ramasamy D, Munschauer F, Weinstock-Guttman B, Zivadinov R (2009) Memory impairment in multiple sclerosis: correlation with deep grey matter and mesial temporal atrophy. J Neurol Neurosurg Psychiatry 80: 201– 206. 5. Rao SM, Leo GJ, Bernardin L, Unverzagt F (1991) Cognitive dysfunction in multiple sclerosis. I. Frequency, patterns, and prediction. Neurology 41: 685– 691. 6. Benedict RH, Cookfair D, Gavett R, Gunther M, Munschauer F, et al. (2006) Validity of the minimal assessment of cognitive function in multiple sclerosis (MACFIMS). Journal of the International Neuropsychological Society 12: 549– 558. 6. Benedict RH, Cookfair D, Gavett R, Gunther M, Munschauer F, et al. (2006) Validity of the minimal assessment of cognitive function in multiple sclerosis (MACFIMS). Journal of the International Neuropsychological Society 12: 549– 558. October 2012 \| Volume 7 \| Issue 10 \| e44942 7 PLOS ONE \| www.plosone.org Limbic Involvement and Episodic Memory in MS Crawford JR, Moore JW, Cameron IM (1992) Verbal fluency: a NART-based equation for the estimation of premorbid performance. British Journal of Clinical Psychology 31: 327–329. 48. Roosendaal SD, Hulst HE, Vrenken H, Feenstra HE, Castelijns JA, et al. (2010) Structural and functional hippocampal changes in multiple sclerosis patients with intact memory function. Radiology 255: 595–604. 31. Smith SM (2002) Fast robust automated brain extraction. Human Brain Mapping 17: 143–155. 49. Rudebeck SR, Scholz J, Millington R, Rohenkohl G, Johansen-Berg H, et al. (2009) Fornix Microstructure Correlates with Recollection But Not Familiarity Memory. J Neurosci 29: 14987–14992. 32. Smith SM, Jenkinson M, Woolrich MW, Beckmann CF, Behrens TE, et al. (2004) Advances in functional and structural MR image analysis and implementation as FSL. Neuroimage 23 Suppl 1: S208–219. y J 50. Metzler-Baddeley C, O’Sullivan MJ, Bells S, Pasternak O, Jones DK (2012) H d h f CSF i i i diff i MRI 50. Metzler-Baddeley C, O’Sullivan MJ, Bells S, Pasternak O, Jones DK (2012) How and how not to correct for CSF-contamination in diffusion MRI. Neuroimage 59: 1394–1403. p g pp 33. Pantel J, O’Leary DS, Cretsinger K, Bockholt HJ, Keefe H, et al. (2000) A new method for the in vivo volumetric measurement of the human hippocampus with high neuroanatomical accuracy. Hippocampus 10: 752–758. 51. Pasternak O, Sochen N, Gur Y, Intrator N, Assaf Y (2009) Free water elimination and mapping from diffusion MRI. Magnetic Resonance in Medicine 62: 717–730. g y pp p 34. Goncalves CV, Duarte G, Costa JS, Marcolin AC, Bianchi MS, et al. (2009) Diagnosis and treatment of cervical cancer during pregnancy. Sao Paulo medical journal = Revista paulista de medicina 127: 359–365. October 2012 \| Volume 7 \| Issue 10 \| e44942 PLOS ONE \| www.plosone.org 8
https://openalex.org/W4226186856	https://figshare.com/articles/thesis/Many-Objective_Genetic_Programming_for_Job-Shop_Scheduling/17151722/1/files/31713872.pdf	English	null	Many-Objective Genetic Programming for Job-Shop Scheduling	null	2,021	cc-by	70,303	Many-Objective Genetic Programming for Job-Shop Scheduling A thesis submitted to the Victoria University of Wellington in fulﬁlment of the requirements for the degree of Doctor of Philosophy in Computer Science. Victoria University of Wellington 2020 Abstract The Job Shop Scheduling (JSS) problem is considered to be a challenging one due to practical requirements such as multiple objectives and the com- plexity of production ﬂows. JSS has received great attention because of its broad applicability in real-world situations. One of the prominent solu- tion approaches to handling JSS problems is to design effective dispatch- ing rules. Dispatching rules are investigated broadly in both academic and industrial environments because they are easy to implement (by comput- ers and shop ﬂoor operators) with a low computational cost. However, the manual development of dispatching rules is time-consuming and requires expert knowledge of the scheduling environment. The hyper-heuristic approach that uses genetic programming (GP) to solve JSS problems is known as GP-based hyper-heuristic (GP-HH). GP-HH is a very useful ap- proach for discovering dispatching rules automatically. Although it is technically simple to consider only a single objective op- timization for JSS, it is now widely evidenced in the literature that JSS by nature presents several potentially conﬂicting objectives, including the maximal ﬂowtime, mean ﬂowtime, and mean tardiness. A few studies in the literature attempt to solve many-objective JSS with more than three objectives, but existing studies have some major limitations. First, many- objective JSS problems have been solved by multi-objective evolutionary algorithms (MOEAs). However, recent studies have suggested that the performance of conventional MOEAs is prone to the scalability challenge and degrades dramatically with many-objective optimization problems (MaOPs). Many-objective JSS using MOEAs inherit the same challenge as MaOPs. Thus, using MOEAs for many-objective JSS problems often fails to select quality dispatching rules. Second, although the reference points method is one of the most prominent and efﬁcient methods for diversity maintenance in many-objective problems, it uses a uniform distribution of reference points which is only appropriate for a regular Pareto-front. However, JSS problems often have irregular Pareto-front and uniformly distributed reference points do not match well with the irregular Pareto- front. It results in many useless points during evolution. These useless points can signiﬁcantly affect the performance of the reference points- based algorithms. They cannot help to enhance the solution diversity of evolved Pareto-front in many-objective JSS problems. Third, Pareto Local Search (PLS) is a prominent and effective local search method for handling multi-objective JSS optimization problems but the literature does not dis- cover any existing studies which use PLS in GP-HH. Abstract To address these limitations, this thesis’s overall goal is to develop GP- HH approaches to evolving effective rules to handle many conﬂicting ob- jectives simultaneously in JSS problems. To achieve the ﬁrst goal, this thesis proposes the ﬁrst many-objective GP-HH method for JSS problems to ﬁnd the Pareto-fronts of nondomi- nated dispatching rules. Decision-makers can utilize this GP-HH method for selecting appropriate rules based on their preference over multiple conﬂicting objectives. This study combines GP with the ﬁtness evalua- tion scheme of a many-objective reference points-based approach. The experimental results show that the proposed algorithm signiﬁcantly out- performs MOEAs such as NSGA-II and SPEA2. To achieve the second goal, this thesis proposes two adaptive reference point approaches (model-free and model-driven). In both approaches, the reference points are generated according to the distribution of the evolved dispatching rules. The model-free reference point adaptation approach is inspired by Particle Swarm Optimization (PSO). The model-driven ap- proach constructs the density model and estimates the density of solutions from each deﬁned sub-location in a whole objective space. Furthermore, the model-driven approach provides smoothness to the model by apply- ing a Gaussian Process model and calculating the area under the mean function. The mean function area helps to ﬁnd the required number of the reference points in each mean function. The experimental results demon- strate that both adaptive approaches are signiﬁcantly better than several state-of-the-art MOEAs. To achieve the third goal, the thesis proposes the ﬁrst algorithm that combines GP as a global search with PLS as a local search in many- objective JSS. The proposed algorithm introduces an effective ﬁtness- based selection strategy for selecting initial individuals for neighborhood exploration. It deﬁnes the GP’s proper neighborhood structure and a new selection mechanism for selecting the effective dispatching rules during the local search. The experimental results on the JSS benchmark problem show that the newly proposed algorithm can signiﬁcantly outperform its baseline algorithm (GP-NSGA-III). iv Acknowledgments I would like to express my sincere gratitude to the Victoria University of Wellington for providing the Victoria Doctoral Scholarship for me to con- tinue my research studies in New Zealand. I would like to express my sincere gratitude to my supervisors, Dr. Gang Chen, Dr. Yi Mei, Dr. Harith Al-Sahaf, and Professor Mengjie Zhang for their exemplary guidance, support, and encouragement throughout the progress of my Ph.D. study. They have provided invaluable feedback to improve my research. I also take this opportunity to express my gratitude to Dr. John Park, Dr. Deepak Singh, Dr. Michael Ikechi Emmanuel, Philippa Becroft, Dr. TJ Boutorwick, and Karen Commons for their cordial support, valuable infor- mation, and guidance, which helped me in my research activities through various stages. I would also like to thank my research colleagues, the Evolutionary Computation Research Group (ECRG), and the Evolutionary Computa- tion for Scheduling and Combinatorial Optimisation group (ECCO) for their constructive feedback and comments in my study and discussions. My sincere gratitude also goes to my beloved husband, Masood, for his constant support, encouragement, and understanding throughout this research work. I would like to thank my mother for her prayer, uncondi- tional love and care. My sincere thanks as well to my uncles, Dr. Jaffery and Muhammad Kalim, for their constant support, prayer, and encour- agement throughout my study. v vi 7. A. Masood, G. Chen,Y. Mei, H. Al-Sahaf, M. Zhang, ”Adaptive Ref- erence Point Generation based on Gaussian Process Model for Many- Objective Optimization”. Submitted to IEEE Transactions on Evolu- tionary Computation. (2020), 14pp. (under review). in Many-Objective Job Shop Scheduling”, (Evolutionary Computation (CEC) 2020 IEEE Congress on, pp. 1-8, 2020). Contents 1 Introduction 1 1.1 Problem statement . . . . . . . . . . . . . . . . . . . . . . . . 1 1.2 Motivations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.3 Research goals . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 1.4 Major contributions . . . . . . . . . . . . . . . . . . . . . . . . 13 1.5 Organisation of thesis . . . . . . . . . . . . . . . . . . . . . . . 16 2 Literature Review 19 2.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 2.1.1 Basic concepts . . . . . . . . . . . . . . . . . . . . . . . 20 2.1.2 JSS Problems . . . . . . . . . . . . . . . . . . . . . . . 23 2.1.3 Evolutionary computation . . . . . . . . . . . . . . . . 27 2.1.4 Multi-objective EAs (MOEAs) . . . . . . . . . . . . . . 30 2.1.5 Many-objective optimization (MaOPs) . . . . . . . . . 32 2.1.6 MaOPs for irregular Pareto-front . . . . . . . . . . . . 38 2.1.7 Model-based EMO . . . . . . . . . . . . . . . . . . . . 40 2.1.8 Gaussian process modelling . . . . . . . . . . . . . . 41 2.1.9 Genetic programming (GP) . . . . . . . . . . . . . . . 43 2.1.10 Basic GP algorithm . . . . . . . . . . . . . . . . . . . . List of Publications 1. A. Masood, Y. Mei, G. Chen, M. Zhang, ”Many-objective genetic pro- gramming for job-shop scheduling”, Evolutionary Computation (CEC) 2016 IEEE Congress, pp. 209-216, 2016. 2. Masood, A., Mei, Y., Chen, G., Zhang, M.: A PSO-Based Reference Point Adaption Method for Genetic Programming Hyper-Heuristic in Many-Objective Job Shop Scheduling. ACALCI. Lecture Notes in Computer Science, vol. 10142, pp. 326–338 (2017). 3. Masood, A., Chen, G., Mei, Y., Zhang, M.: Reference point adaption method for genetic programming hyper-heuristic in many-objective job shop scheduling. In: European Conference on Evolutionary Compu- tation in Combinatorial Optimization. pp. 116–131. Springer (2018). 4. Masood A., Chen G., Mei Y., Zhang M.: Adaptive Reference Point Generation for Many-Objective Optimization Using NSGA-III. In: Advances in Artiﬁcial Intelligence. AI 2018. Lecture Notes in Computer Science, vol 11320. pp. 358-370. Springer(2018). 5. Masood A., Chen G., Mei Y., Al-Sahaf H., Zhang M.: Genetic Pro- gramming with Pareto Local Search for Many-Objective Job Shop Scheduling. In: Advances in Artiﬁcial Intelligence. AI 2019. Lecture Notes in Computer Science, vol 11919. pp. 536-548. Springer(2019). 6. A. Masood, G. Chen,Y. Mei, H. Al-Sahaf, M. Zhang, ”Fitness-based Selection Method for Genetic Programming with Pareto Local Search vii viii in Many-Objective Job Shop Scheduling”, (Evolutionary Computation (CEC) 2020 IEEE Congress on, pp. 1-8, 2020). 7. A. Masood, G. Chen,Y. Mei, H. Al-Sahaf, M. Zhang, ”Adaptive Ref- erence Point Generation based on Gaussian Process Model for Many- Objective Optimization”. Submitted to IEEE Transactions on Evolu- tionary Computation. (2020), 14pp. (under review). Contents 47 2.1.11 Local search . . . . . . . . . . . . . . . . . . . . . . . . 48 2.1.12 Pareto local search (PLS) . . . . . . . . . . . . . . . . . 49 2.2 Related work . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 2.2.1 JSS techniques . . . . . . . . . . . . . . . . . . . . . . . 52 ix x 2.2.2 GP-HH for JSS . . . . . . . . . . . . . . . . . . . . . . 57 2.2.3 Multi-objective and many-objective JSS . . . . . . . . 59 2.2.4 Local search for JSS . . . . . . . . . . . . . . . . . . . . 63 2.2.5 PLS for JSS . . . . . . . . . . . . . . . . . . . . . . . . 63 2.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 3 Experimental Methodology 69 3.1 Benchmark problems for JSS . . . . . . . . . . . . . . . . . . . 69 3.2 GP terminals and function set . . . . . . . . . . . . . . . . . . 71 3.3 Performance measures . . . . . . . . . . . . . . . . . . . . . . 71 4 Many-Objective GP-HH for JSS 75 4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 4.2 Problem description . . . . . . . . . . . . . . . . . . . . . . . 78 4.3 Many-objective-GP-HH for JSS . . . . . . . . . . CONTENTS xi 5.4.1 Reference point Adaptation by density-based model 119 5.5 Gaussian process-based probabilistic model . . . . . . . . . . 124 5.5.1 Modelling by Gaussian process . . . . . . . . . . . . . 126 5.5.2 Calculate the area under the mean function . . . . . . 128 5.5.3 Reference points generation . . . . . . . . . . . . . . . 130 5.5.4 Computational Complexity of One Generation of GP-MARP-NSGA-III . . . . . . . . . . . . . . . . . . . 130 5.5.5 Design of experiment . . . . . . . . . . . . . . . . . . 131 5.5.6 Results and discussion . . . . . . . . . . . . . . . . . . 132 5.6 Selection of adaptive reference points approach . . . . . . . . 136 5.7 Chapter summary . . . . . . . . . . . . . . . . . . . . . . . . . 139 6 GP with Pareto Local Search for Many-Objective JSS 141 6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 6.2 GP-PLS structure . . . . . . . . . . . . . . . . . . . . . . . . . 144 6.2.1 General framework of GP-PLS . . . . . . . . . . . . . 144 6.2.2 GP-PLS-I overview . . . . . . . . . . . . . . . . . . . 144 6.2.3 GP-PLS-II overview . . . . . . . . . . . . . . . . . . . 148 6.3 Design of experiment . . . . . . . . . . . . . . . . . . . . . . . 155 6.3.1 Sensitivity analysis . . . . . . . . . . . . . . . . . . . . 156 6.4 Results and discussions . . . . . . . . . . . Contents . . . . . . . 79 4.3.1 Representation of rules . . . . . . . . . . . . . . . . . . 80 4.3.2 General framework of GP-NSGA-III . . . . . . . . . . 80 4.4 Experimental studies . . . . . . . . . . . . . . . . . . . . . . . 83 4.4.1 Parameter settings . . . . . . . . . . . . . . . . . . . . 84 4.5 Results and discussions . . . . . . . . . . . . . . . . . . . . . 85 4.5.1 Relationship among scheduling objectives . . . . . . 85 4.5.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 4.5.3 Discussions . . . . . . . . . . . . . . . . . . . . . . . . 89 4.6 Chapter summary . . . . . . . . . . . . . . . . . . . . . . . . . 100 5 Reference Points Adaptation for Many-Objective JSS 105 5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 5.2 General framework for adaptive reference points generation 109 5.3 Model-free adaptive reference points generation . . . . . . 110 5.3.1 Design of experiment . . . . . . . . . . . . . . . . . . 112 5.3.2 Results and discussions . . . . . . . . . . . . . . . . . 113 5.4 Model-based adaptive reference points generation . . . . . 118 CONTENTS . . . . . . . . . . 159 6.4.1 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 6.4.2 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . 161 6.5 Chapter summary . . . . . . . . . . . . . . . . . . . . . . . . . 173 7 Conclusions 175 7.1 Achieved objectives . . . . . . . . . . . . . . . . . . . . . . . . 175 7.2 Main conclusions . . . . . . . . . . . . . . . . . . . . . . . . . 178 7.2.1 Many-Objective GP for JSS . . . . . . . . . . . . . . . 178 7.2.2 Non-uniform Pareto-front . . . . . . . . . . . . . . . . 179 7.2.3 Pareto local search (PLS) . . . . . . . . . . . . . . . . . 181 7.3 Future work . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182 7 Conclusions 175 7.1 Achieved objectives . . . . . . . . . . . . . . . . . . . . . . . . 175 7.2 Main conclusions . . . . . . . . . . . . . . . . . . . . . . . . . 178 7.2.1 Many-Objective GP for JSS . . . . . . . . . . . . . . . 178 7.2.2 Non-uniform Pareto-front . . . . . . . . . . . . . . . . 179 7.2.3 Pareto local search (PLS) . . . . . . . . . . . . . . . . . 181 7.3 Future work . . . . . . . . . . . . . . . . . . . CONTENTS . . . . . . . . . 182 CONTENTS xii CONTENTS 7.3.1 Incorporate user preferences to many-objective JSS . 183 7.3.2 Incorporate locality of search operators to Genetic Programming . . . . . . . . . . . . . . . . . . . . . . . 183 7.3.3 Incorporate effective crossover operator to many- objective JSS . . . . . . . . . . . . . . . . . . . . . . . . 184 7.3.4 Incorporate adaptive terminal selection to many- objective JSS . . . . . . . . . . . . . . . . . . . . . . . . 185 7.3.5 Dispatching rules for many-objective dynamic JSS . . 185 7.3.6 Dispatching rules for many-objective ﬂexible JSS . . . 186 7.4 General Considerations . . . . . . . . . . . . . . . . . . . . . . 187 7.4.1 Main components of thr Proposed algorithm . . . . . 187 7.4.2 Cloud task scheduling problem . . . . . . . . . . . . . 189 A Further studies 193 A.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193 A.2 Benchmark functions on MaOPs . . . . . . . . . . . . . . . . 194 A.2.1 IDTLZ1 and IDTLZ2 problems . . . . . . . . . . . . . 194 A.2.2 DTLZ-4, DTLZ5, and IDTLZ-7 Problems . . . . . . . 195 A.2.3 MAF1 and MAF2 Problems . . . . . . . . . . . . . . . 196 A.2.4 WFG1,WFG2 and WFG9 Problems . . . . . . . . . . . 196 A.3 Experiment design . . . . . . . . . . . . . . . . . . . . . . . . 197 A.3.1 Parameter setting of benchmark problems . . . . . . 197 A.3.2 Algorithms parameter settings . . . . . . . . . . . . . 198 A.4 Results and discussion . . . . . . . CONTENTS . . . . . . . . . . . . . . . 199 A.4.1 Performance of obtained solutions . . . . . . . . . . . 199 A.4.2 Further analysis . . . . . . . . . . . . . . . . . . . . . . 199 List of Tables 2.1 Notations in JSS . . . . . . . . . . . . . . . . . . . . . . . . . . 24 2.2 Example of a static JSS problem instances (N = 3,M =2) . . . 27 3.1 Static JSS data sets. . . . . . . . . . . . . . . . . . . . . . . . . 69 3.2 Terminal set of GP for JSS. . . . . . . . . . . . . . . . . . . . . 72 4.1 The mean and standard deviation over the average HV and IGD values on training instances of the compared algo- rithms in the four-objective experiment. . . . . . . . . . . . . 88 4.2 The mean and standard deviation over the HV values on the test instances of the compared algorithms. . . . . . . . . 90 4.3 The mean and standard deviation over the IGD values on the test instances of the compared algorithms. . . . . . . . . 91 5.1 The mean and standard deviation over the HV and IGD val- ues on the test instances of the compared algorithms in the 4-obj experiment. . . . . . . . . . . . . . . . . . . . . . . . . . 114 5.2 The mean and standard deviation of HV and IGD values of the 30 independent runs on training instances of the com- pared algorithms on four-objective JSS problems. . . . . . . . 132 5.3 The mean and standard deviation of HV and IGD achieved by all competing algorithms on test instances on four- objective JSS problems. . . . . . . . . . . . . . . . . . . . . . 132 xiii xiii LIST OF TABLES xiv 6.1 The mean and standard deviation over the average HV and IGD values on training instances of the compared algo- rithms in the four-objective experiment. . . . . . . . . . . . . 157 6.2 The mean and standard deviation over the average HV and IGD values on training instances of the compared algo- rithms in the four-objective experiment. . . List of Tables . . . . . . . . . . 159 6.3 The mean and standard deviation over the HV values on the test instances of the compared algorithms. . . . . . . . . 162 6.4 The mean and standard deviation over the IGD values on the test instances of the compared algorithms. . . . . . . . . 163 A.1 The characteristics of benchmark problems . . . . . . . . . . 194 A.2 Number of Reference Points and Population Size for DTLZ and MAF. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197 A.3 Number of Reference Points and Population Size for WFG. . 198 A.4 The mean and standard deviation (¯x ± σ) over the aver- age HV values on M-objectives on IDTLZ1, IDTLZ2, DTLZ4,DTLZ5, DTLZ7, MAF1, MAF2, WFG1, WFG2 prob- lems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200 A.5 The mean and standard deviation (¯x ± σ) over the aver- age IGD values on M-objectives on IDTLZ1, IDTLZ2, DTLZ4, DTLZ5, and DTLZ7, MAF1, MAF2, WFG1, WFG2 problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201 List of Figures 2.1 The schedule generated by non-delay SPT α = 0. . . . . . . . 28 2.2 The schedule generated by active SPT α = 1. . . . . . . . . . 28 2.3 An example of a GP :PR + (RO × DD)). . . . . . . . . . . . . 44 2.4 An example of a crossover operation in GP. . . . . . . . . . . 47 2.5 An example of a mutation operation in GP. . . . . . . . . . . 47 2.6 Overview of a tree-based GP-HH applied to JSS. . . . . . . . 57 4.1 Illustration of a dispatching rule in JSS. . . . . . . . . . . . . 79 4.2 The GP tree representation of the 2PT+WINQ+NPT rule. . . 80 4.3 Pareto-front for pairwise objective combination. . . . . . . . 86 4.4 Pareto-front for pairwise objectives combination between (mWT −maxWT). . . . . . . . . . . . . . . . . . . . . . . . . 87 4.5 Computational time to evolve dispatching rules . . . . . . . 93 4.6 Length of rules of each generation in GP-NSGA-II, GP- SPEA2, and GP-NSGA-III. . . . . . . . . . . . . . . . . . . . . 93 4.7 Length of the best rules of each run in GP-NSGA-II, GP- SPEA2, and GP-NSGA-III. . . . . . . . . . . . . . . . . . . . . 94 4.8 Frequency of terminals in GP-NSGA-II, GP-SPEA2, and GP- NSGA-III. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 4.9 Frequency of terminals (after simpliﬁcation) in GP-NSGA- III, GP-NSGA-II, GP-SPEA2. . . . . . . . . . . . . . . . . . . . 97 xv LIST OF FIGURES xvi 4.10 The average relevance of each terminal over the 30 indepen- dent runs of GP-NSGA-III, GP-NSGA-II and GP-SPEA2 on the training set. . . . . . . . . . . . . . . . . . List of Figures . . . . . . . . . 98 4.11 The average HV value of the non-dominated solutions on the training set during the 30 independent GP runs. . . . . . 99 4.12 The average IGD value of the non-dominated solutions on the training set during the 30 independent GP runs. . . . . . 99 4.13 Parallel coordinate plot for the distribution of the reference points and the ﬁtness values of the population at genera- tions 50 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 4.14 Box plots on instance 6 of the compared algorithms. . . . . . 101 4.15 Box plots on instance 21 of the compared algorithms. . . . . 101 4.16 Box plots on instance 32 of the compared algorithms. . . . . 102 5.1 Associated and contributing solutions on a convex curva- ture of Pareto-front . . . . . . . . . . . . . . . . . . . . . . . . 107 5.2 The curve of average number of useless reference points in GP-NSGA-III on the training instances during the 30 inde- pendent GP runs . . . . . . . . . . . . . . . . . . . . . . . . . 108 5.3 The curves of the average number of useless reference points in GP-NSGA-III, GP-A-NSGA-III(PSO). . . . . . . . . 116 5.4 The curves of the HV and IGD values of the non-dominated solutions on the training set during the 30 independent GP runs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 5.5 Parallel coordinate plots of GP-A-NSGA-III(PSO) at gener- ations 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 5.6 Parallel coordinate plots of GP-A-NSGA-III(PSO) at gener- ations 50. . . . . . . . . . . . . . . . . . . . . . . . List of Figures . . . . . . . 117 5.7 Solutions are closest to the reference points. . . . . . . . . . . 120 5.8 Density of solution at each sub-location of the simplex. . . . 122 5.9 Generate reference points until M −1 times . . . . . . . . . . 125 5.10 Train a Gaussian process on density-based model. . . . . . . 127 LIST OF FIGURES xvii 5.11 Area under the mean function of the Gaussian process model130 5.12 HV values of the non-dominated solutions on the training set during the 30 independent GP runs. . . . . . . . . . . . . 133 5.13 IGD values of the non-dominated solutions on the training set during the 30 independent GP runs. . . . . . . . . . . . . 134 5.14 The curves of the average number of useless refer- ence points in GP-NSGA-III, GP-NSGA-III-DRA, and GP- MARP-NSGA-III. . . . . . . . . . . . . . . . . . . . . . . . . . 135 5.15 Parallel coordinate plot for the distribution of the reference points and the distribution of the ﬁtness values at genera- tions 50 of GP-NSGA-III. . . . . . . . . . . . . . . . . . . . . . 136 5.16 Parallel coordinate plot for the distribution of the reference points and the distribution of the ﬁtness values at genera- tions 50 of GP-A-NSGA-III. . . . . . . . . . . . . . . . . . . . 137 5.17 Parallel coordinate plot for the distribution of the reference points and the distribution of the ﬁtness values at genera- tions 50 of GP-NSGA-III-DRA. . . . . . . . . . . . . . . . . . 137 5.18 Parallel coordinate plot for the distribution of the reference points and the distribution of the ﬁtness values at genera- tions 50 of GP-MARP-NSGA-III. . . . . . . . . . . . . . . . . 138 6.1 General Framework of GP-PLS-I. . . . . . . . . . . . . . . . . List of Figures 146 6.2 Examples of possible neighbor rules(shaded sub trees rep- resented newly generated sub tree). . . . . . . . . . . . . . . 147 6.3 Framework of GP-PLS-II. . . . . . . . . . . . . . . . . . . . . 150 6.4 Example showing how to associate an individual r with a reference points. In this example, w1 and w2 are two unit reference points, theta1 and theta2 are the angles between r and w1 and w2, respectively. Since theta2 < theta1, the indi- vidual denoted by r is associated with reference points w2. . 152 6.5 Distance measure in the context of minimization with re- spect to a reference direction. . . . . . . . . . . . . . . . . . . 153 LIST OF FIGURES xviii xviii 6.6 Computational time of whole population . . . . . . . . . . . 158 6.7 Computational time of sub-population. . . . . . . . . . . . . 158 6.8 Frequency of terminals in GP-NSGA-III, GP-PLS-I-s, GP- PLS-I-r, GP-PLS-II, and GP-PLS-II-A. . . . . . . . . . . . . . . 161 6.9 Length of rules from each generation in GP-NSGA-III, GP- PLS-I-s, GP-PLS-I-r, GP-PLS-II. . . . . . . . . . . . . . . . . . 165 6.10 The curves of the average number of HV value of the non- dominated solutions on the training set during the 30 inde- pendent GP runs. . . . . . . . . . . . . . . . . . . . . . . . . . 166 6.11 The curves of the average number of IGD value of the non- dominated solutions on the training set during the 30 inde- pendent GP runs. . . . . . . . . . . . . . . . . . . . . . . . . . 166 6.12 Parallel coordinate plot of GP-NSGA-III. . . . . . . . . . . . . 167 6.13 Parallel coordinate plot of GP-PLS-I-s. . . . . . . . . . . . . . 167 6.14 Parallel coordinate plot of GP-PLS-I-r. . . . . . . . . List of Figures . . . . . 168 6.15 Parallel coordinate plot of GP-PLS-II-U. . . . . . . . . . . . . 168 6.16 Parallel coordinate plot of GP-PLS-II-A. . . . . . . . . . . . . 169 6.17 Distribution of solutions of GP-PLS-II-U on instance 26. . . . 170 6.18 Distribution of solutions of GP-PLS-II-A on instance 26. . . . 171 6.19 Box-plots of the HV values. . . . . . . . . . . . . . . . . . . . 172 6.20 Box-plots of the HV values. . . . . . . . . . . . . . . . . . . . 172 6.21 Box-plots of the IGD values. . . . . . . . . . . . . . . . . . . . 173 6.22 Box-plots of the IGD values. . . . . . . . . . . . . . . . . . . . 173 7.1 Main Components of Algorithm . . . . . . . . . . . . . . . . 187 A.1 The Pareto-front with three objectives on IDTLZ problems. . 195 A.2 The Pareto-front with three objectives on DTLZ problems. . 196 A.3 The Pareto-front with three objectives on MAF problems. . . 197 A.4 The Pareto-front with three objectives on WFG problems. . . 198 A.5 Approximate Pareto-front for 3-objective IDTLZ1 problem . 202 A.6 Approximate Pareto-front for 3-objective DTLZ4 problem . . 203 LIST OF FIGURES xix A.7 Approximate Pareto-front for 3-objective DTLZ5 problem . . 204 A.8 Approximate Pareto-front for 3-objective DTLZ7 problem . . 205 A.9 Parallel coordinate plot for the ﬁtness values of the popula- tion for 5-objective DTLZ7 problem . . . . . . . . . . . . . . . 205 A.10 Parallel coordinate plot for the ﬁtness values of the popula- tion for 5-objective MAF1 problem . . . . . . . . . . . . . . . 206 A.11 Parallel coordinate plot for the ﬁtness values of the popula- tion for 5-objective MAF2 problem . . . . . . . . . . . . . . . 207 A.12 Parallel coordinate plot for the ﬁtness values of the popula- tion for 5-objective WFG2 problem . . . List of Figures . . . . . . . . . . . . 208 A.13 Parallel coordinate plot for the ﬁtness values of the popula- tion for eight-objective WFG9 problem . . . . . . . . . . . . . 208 LIST OF FIGURES xx List of Abbreviations ACO Ant colony optimization A-NSGA-III Adaptive-NSGA-III COVERT Cost over time DD Due date DRA Density model-based reference point adapta- tion EAs Evolutionary algorithms EC Evolutionary computation EDAs Estimation of distribution algorithms EMO Evolutionary multi-objective optimization FDD Flow due date GAs Genetic algorithms GaP Gaussian process GP Genetic programming GP-A-NSGAIII GP-adaptive-Non-dominated sorting GA-III GP-HH Genetic programming based hyper-heuristic GP-NSGA-III GP-non-dominated sorting GA-III xxi xxi xxii HV IGD JSS MaOPs maxF maxWT MARP MBEAs mF MOEAs MOPs MRT mWT NMRT NOIQ NOINQ NOPT NP NSGA-II NSGA-III PLS PSO PT LIST OF FIGURES Hypervolume Inverted generational distance Job shop scheduling Many-objective optimization problems Maximal ﬂowtime Maximal weighted tardiness Model-based adaptive reference points Model-based evolutionary algorithms Mean ﬂowtime Multi-objective evolutionary algorithms Multi-objective optimization problems Ready time of the machine Mean weighted tardiness Ready time of the next machine Number of operations in the queue Number of operations in the next queue Processing time of the next operation Non-deterministic polynomial-time Non-dominated sorting genetic algorithm-II Non-dominated sorting GA-III Pareto local search Particle swarm optimization Processing time of the operation xxiii RVEA Reference vector-guided evolutionary algo- rithm SPEA2 Strength Pareto evolutionary algorithm SPT Shortest processing time SI Swarm intelligence RVEA* Adaptive version of RVEA TWT Total weight tardiness W Weight WIQ Work in the queue WINQ Work in the next queue WKR Work remaining RVEA Reference vector-guided evolutionary algo- rithm SPEA2 Strength Pareto evolutionary algorithm SPT Shortest processing time SI Swarm intelligence RVEA* Adaptive version of RVEA TWT Total weight tardiness W Weight WIQ Work in the queue WINQ Work in the next queue WKR Work remaining LIST OF FIGURES xxiv Chapter 1 This chapter provides an introduction to this thesis and its motivations, goals, contributions, and organization. The problem statement is provided ﬁrst, followed by a discussion of the main limitations of existing litera- ture. The research goals and major contributions of this thesis are then discussed. This chapter also provides a brief discussion of how this thesis structured. 1.1 Problem statement Job Shop Scheduling (JSS) [165] is a non-deterministic polynomial-time (NP) hard combinatorial optimization problem [53] in which various man- ufacturing jobs are assigned to machines at particular times while trying to minimize several objectives such as makespan, mean ﬂowtime, mean tardiness. JSS problems have received signiﬁcant attention from both aca- demics and industry experts. From an industry perspective, JSS is con- sidered to be a good model for many manufacturing scenarios because scheduling has direct impacts on the production efﬁciency and costs of a manufacturing system [94]. As reported by Johns and Rabelo [94], thou- sands of manufacturers contribute billions of dollars to the United State’s economy. Furthermore, JSS is considered one of the signiﬁcant production 1 CHAPTER 1. INTRODUCTION 2 scheduling problems in practice. It has a wide range of applications in many industries such as cloud computing [188] and management and op- erations research [115]. JSS has received substantial research attention due to its high computational challenges and strong practical value [143, 165]. A JSS problem usually has a set of machines on the shop ﬂoor that can be used to process a set of jobs [165]. Each job has a predetermined se- quence of operations, which needs to be carried out in order to complete the job. Each job has a predetermined route through the machines before it leaves the shop. The machine can only process one operation at a time. JSS aims to process all arriving jobs by the machines in an optimal way so that the predeﬁned objectives (e.g., makespan, and tardiness) are op- timized in order to maximize the total revenue subject to the constraints (e.g., order of the operations of the jobs and the available machines for the jobs). The quality of a schedule in a job shop depends on the objective(s) of the problem. JSS problems can be classiﬁed into two categories: static and dynamic [165]. The ﬁrst subset belongs to static JSS problems where a problem has a ﬁxed amount of jobs with known processing requirements such as processing time [165]. The second subset is dynamic JSS problems. In dynamic JSS problems, jobs arrive on the shop ﬂoor at various instances of time with no prior process information, e.g., release date, due date, and processing time [165]. Approaches for solving static JSS problems can be broken down into two main categories. 1.1 Problem statement The ﬁrst category is the exact mathematical optimi- sation approaches [2, 19, 25, 55, 165]. The exact mathematical optimization approaches generate optimal solutions for JSS problem instances. The ex- act algorithms aim at searching through the full solution space and only work well on small-scale JSS problems [13]. For larger JSS problems, these algorithms are infeasible due to the exponential size of the solution space [19]. It is reported in [13] that the JSS problem with up to 100 jobs and 20 machines is computationally intractable for any exact mathematical opti- 1.1. PROBLEM STATEMENT 3 3 mization approach. For such large-scale problems, ﬁnding near-optimal solutions (not optimal) within a reasonable amount of computation time is more feasible than ﬁnding optimal solutions with high computation cost. In this context, the second category of heuristic approaches is con- sidered more desirable. These approaches are fast, although they cannot guarantee the optimality of the solution for JSS problems. These heuris- tic approaches are “rules-of-thumb” with the hope of generating ”good” schedules [71]. Heuristic algorithms are mostly divided into two groups: the ﬁrst group consists of meta-heuristics which directly targets the solu- tion space (schedules). The second group explores the space of scheduling heuristics, typically in the form of dispatching rules represented as prior- ity functions. Meta-heuristic algorithms are search-based methods (e.g., genetic algo- rithms, particle swarm optimization). They can be used to search the so- lution space [9, 152, 157]. These methods are higher-level heuristics which provide a general framework to guide low-level heuristics that make local decisions. Meta-heuristic algorithms have the advantage that they can ﬁnd high-quality solutions and successfully tackle large scheduling problems [152, 157]. A downside of meta-heuristic algorithms is that they require a substantial amount of time to explore high-quality solutions but are faster than exact algorithms, particularly for large problems [152]. Dispatching rules have been applied extensively to JSS problems due to their computational efﬁciency [176, 185]. Dispatching rules can be seen as a priority function which is used to assign priority to each job wait- ing to be processed by a machine. Then, the next job to process will be selected based on the priority value. Such computation is carried out at each decision point (e.g., when a machine becomes idle) and can be done efﬁciently [185]. There are two broad types of dispatching rules: (1) non- delay and (2) active rules. 1.1 Problem statement As the name indicates, non-delay rules do not allow any delay on the idle machines as long as the waiting queue is not empty [165]. On the other hand, active rules [165] allow some reasonable CHAPTER 1. INTRODUCTION 4 delay (which is no more than the minimal processing time of the wait- ing jobs) to handle the potential new job arrivals with an urgent due date. Nguyen et al. [141] explored three different genetic programming (GP) representations of dispatching rules and proved that evolved dispatching rules could outperform benchmark human-made dispatching rules for JSS problems. Hunt et al. [75] evolved effective dispatching rules for the two- machine JSS problem to minimize the makespan. Dispatching rules are attractive to both researchers and practitioners because of their simplicity and high scalability in comparison to most of the direct optimization meth- ods [141, 159]. However, in designing a dispatching rule for JSS, there are two main challenges. First, dispatching rules are time-consuming to de- sign manually, especially for optimizing multiple potentially conﬂicting objectives, a frequent demand in a manufacturing environment. Mean- while, any dispatching rules to be designed by domain experts will have to go through a lengthy and costly process [70]. Second, it is not always clear which of the existing dispatching rules is suitable for the given JSS problems [90]. In order to deal with these disadvantages, hyper-heuristics have been adopted in this thesis to design the dispatching rules automatically [70, 103]. GP has been a promising approach for designing dispatching rule heuristics automatically because GP has an ability to evolve priority functions with its ﬂexible representation [143, 103]. The hyper-heuristic approach that uses GP to solve JSS problems is known as GP based hyper- heuristic (GP-HH ) [17, 143]. A GP-HH evolves dispatching rules for both static and dynamic JSS problems automatically [17, 142, 143, 147, 145]. A GP-HH bypasses the need for human experts and extensive trial-and-error to construct dis- patching rules. Dispatching rules evolved by GP can be directly and in- tuitively represented as tree-based priority functions. Moreover, there are a wide variety of evolutionary algorithms that deal with multi-objective (two or three objectives) and many-objective optimization (four or more 1.1. PROBLEM STATEMENT 5 5 objectives), which shows to have good performances on many optimiza- tion problems. Therefore, by coupling GP with such multi-objective and many-objective algorithms can design a group of non-dominated dis- patching rules automatically. 1.1 Problem statement These dispatching rules are expected to achieve a wide range of trade-offs over many conﬂicting objectives. Since the early 2000s, the research on multi-objective JSS has started gaining popularity. Therefore, many independent studies have been car- ried out for multi-objective JSS with two ore three conﬂicting objectives, including the makespan, mean ﬂowtime, maximum tardiness, maximum lateness, total workload, and proportion of tardy jobs [147, 184, 201]. In the literature, multi-objective optimization problems in a job shop were gener- ally treated in two ways. The ﬁrst approach aggregates multiple objectives together into a single objective through a weighted sum (for example, the linear weighted summation function). It then applies the single-objective optimization method to ﬁnd the (single) optimal solution [79]. The sec- ond category of approach aims at ﬁnding a set of optimal solutions based on the Pareto-dominance concept [42] instead of ﬁnding a single optimal solution using aggregation functions. The Pareto-dominance concept deﬁnes that a solution x dominates an- other solution y if no objective of x is worse than the corresponding one of y and at least one objective of x is better than y. Based on this concept, Pareto-optimal rules in JSS refer to the rules that cannot be dominated by others based on their performance on concerned scheduling objec- tives [116]. Multi-objective scheduling problems with the goal to ﬁnd the Pareto-optimal front was ﬁrst considered by Ishibuchi et al. [79] in 1998. After that, many evolutionary computation (EC) algorithms have also been proposed to evolve the Pareto-front for multi-objective JSS [44, 211]. Recently, many-objective optimization has become an active research topic [43, 116]. As emphasized by Deb in [43], a large proportion of real- world problems can be described naturally as many-objective optimiza- tion problems (MaOPs). Many-objective optimization refers to a class CHAPTER 1. INTRODUCTION 6 of optimization problems that have more than three objectives. The last decade has witnessed the emergence of many-objective optimization as a booming topic in a wide range of complex modern real-world scenarios [84]. There is also a growing interest in industries to tackle problems with many objectives [43, 86]. According to our knowledge, the research on many-objective JSS algorithms is limited in the literature. Only a few algo- rithms in the literature tackles JSS problems consisting of more than three scheduling objectives [51, 147]. 1.1 Problem statement These algorithms utilize the conventional multi-objective evolutionary algorithms (MOEAs) (NSGA-II and SPEA2) but MOEAs experienced substantial difﬁculties when they were adopted to tackle MaOPs [43, 84]. These conventional MOEAs lost their selection pressure in solving MaOPs which is essential for the population to con- verge toward the Pareto-front [84]. Because of the lack of research on many-objective JSS, it is both theoretically and practically important to de- velop innovative GP-HH algorithms for many-objective JSS. Hence, the overall goal of this thesis is to develop GP-HH approaches to evolve effective dispatching rules for many conﬂicting objectives in JSS problems. 1.2 Motivations Scheduling theory has been established over the years, but most of the ex- isting literature on the automatic design of dispatching rules mainly con- centrates on single-objective JSS [67, 159] and multi-objective JSS [79, 200]. However, in practice, a good schedule for a JSS problem is expected to sat- isfy many objectives such as minimizing: tardiness, makespan, ﬂowtime, and percentage of tardy jobs. Most studies in JSS considered decomposition-based approaches to op- timize these objectives together [67, 185]. In decomposition-based ap- proaches, objectives are aggregated into a single objective (ﬁtness func- tion) and the problem is treated as a single-objective optimization prob- lem. Even in the case where normalized objective functions are used, 1.2. MOTIVATIONS 7 the assigned weights still need to be predeﬁned by the decision-makers and decision-makers should have good information about the trade-offs among different objective functions. To alleviate the strong dependency on domain knowledge, evolving the whole Pareto-front of non-dominated dispatching rules is naturally a better target for JSS problems with two or more than two objectives. It will also assist decision-makers in deciding of selection for quality dispatching rules which based on the trade-offs rep- resented by the obtained Pareto-front. Very few studies [129, 147] attempted to solve the many-objective JSS using MOEAs such as non-dominated sorting genetic algorithm II (NSGA- II) [44] and the strength Pareto evolutionary algorithm 2 (SPEA2) [211]. However, recent studies have suggested that conventional MOEAs are prone to the scalability challenge, i.e., the performance of these MOEAs degrades dramatically with the MaOPs [84, 43]. This is because MOEAs cannot provide sufﬁcient selection pressure towards the Pareto-frontfor MaOPs. The main reason is that the number of non-dominated solutions increases exponentially as the number of objectives increases [84]. Many- objective JSS using MOEAs faces the same scalability challenges. Thus, using MOEAs for many-objective JSS often fail to select quality dispatch- ing rules. Although the effectiveness of GP-HH has been extensively stud- ied in the ﬁeld of JSS [143, 147] (single JSS and multi-objective JSS), it was seldom explored in the context of many-objective JSS. In MOEAs, when the primary selection criterion (convergence) can- not readily ﬁnd good solutions, the diversity-based secondary criterion plays a vital role in the selection of the ﬁttest solutions. For good solution diversity, MOEAs requires uniformly distributed candidate solutions to approximate the Pareto-front jointly [31]. 1.2 Motivations RESEARCH GOALS 9 is one of the key research issues for many-objective JSS problems. Pareto Local Search (PLS) is a simple and effective local search method for tackling multi-objective combinatorial optimization problems [26]. It is an extension of iterative local search algorithms for single-objective prob- lems [149] to the multi-objective domain. PLS is a crucial component of memetic algorithms. Researchers have studied the application of PLS to multi-objective evolutionary algorithms with some success [26]. In fact, by hybridizing global search with local search, the performance of many MOEAs can be noticeably improved [26, 79]. Despite the preliminary suc- cess, the practical use of PLS on many-objective JSS is relatively limited. Moreover, our comprehensive search of the literature does not discover any existing studies involving the use of PLS in GP-HH for many-objective JSS. In this thesis, one of our goals is to enhance the quality of evolved dis- patching rules for many-objective JSS through hybridizing GP with the PLS technique. There are three challenges herein. First, the neighborhood structure in GP is not deﬁned. Second, the PLS algorithm may suffer from low efﬁciency in the case of more than two objectives [89] because the number of optimal solutions grows exponentially with a number of objec- tives. Thus to search the neighborhood of a massive number of solutions becomes very time-consuming. For this reason, the size of the initial solu- tions for neighborhood exploration should be determined carefully. Third, the hybridized algorithm for many-objective JSS is how to divide the avail- able computation time between the local search and the global search. A thorough investigation is required to determine how to achieve a desirable balance between local search and global search. 1.2 Motivations However, in the case of a high dimension in terms of a number of objectives, it is quite challenging to maintain solution diversity because of the sparse distribution of the candi- date solutions. The diversity maintenance can be controlled by providing multiple predeﬁned reference points [43]. The reference points method CHAPTER 1. INTRODUCTION 8 is one of the prominent and efﬁcient ways for diversity maintenance in many-objective problems. With widely distributed reference points, solu- tions associated with these points are also encouraged to be distributed widely in the objective space. Several recently developed algorithms such as NSGA-III [43], RVEA [31], and SPEAR [92] employed a predeﬁned set of uniformly distributed reference points located on a hyperplane. These algorithms have suc- cessfully solved various practical MaOPs. One of the reasons for using a uniform distribution of reference points is because candidate solutions corresponding to each reference point can be emphasized to ﬁnd a set of widely distributed sets of Pareto-optimal solutions. It is also reported by Ishibuchi et al. [80] that good results can be obtained by MOEAs algo- rithms only if the distribution of reference points is consistent with the Pareto-front shape of the problem to be solved. In this case, uniformly dis- tributed reference points are inappropriate when the true Pareto-front is non-uniform, discontinuous, and irregular [86]. JSS problems often have irregular Pareto-front [142]. Therefore, many uniformly distributed reference points may be located in a region where no Pareto-optimal solutions exist and they are never associated with any dispatching rules on the evolved Pareto-front and become useless refer- ence points. Clearly, if only a few reference points are truly associated with the dispatching rules evolved by GP-HH at the current generation, it will not be easy to distinguish and select these rules to improve diversity for future generations. In other words, many reference points will be as- sociated with a large number of candidate dispatching rules. Some rules can be far from the corresponding reference point. Such a dispatching rule should enjoy higher selection opportunities but may not be selected during evolution simply because it is associated with a popular reference point. This issue has been clearly highlighted in [85, 86]. In particular, to evolve high-quality dispatching rules, it is essential to match the distribu- tion of the reference points with the distribution of the Pareto-front. This 1.3. 2. Investigate how to develop GP-HH approaches for the non-uniform Pareto- front of many-objective JSS problems which can evolve high-quality Pareto- optimal dispatching rules. 1.3 Research goals The overall goal of this thesis is to develop GP-HH methods to evolve reusable and effective dispatching rules for many-objective JSS problems. CHAPTER 1. INTRODUCTION 10 This research is also focused on investigating the conﬂicting nature among many related objectives that often should be considered together for ef- fective JSS. This research aims to develop GP-HH methods that alleviate issues related to many-objective optimization in JSS problems and evolve new effective dispatching rules that are capable of enhancing the produc- tivity of job shops. It is expected that the evolved rules can be reusable in unseen situations and outperform state-of-the-art multi-objective and many-objective algorithms in the literature. This research will be broken down into the following key objectives: 1. Investigate how GP can be used to handle many-objective JSS problems. 1. Investigate how GP can be used to handle many-objective JSS problems. This research objective explores whether the standard multi- objective optimization algorithms, such as NSGA-II and SPEA2, are competent at tackling many-objective JSS problems. To carry out this research effectively, we adopt a popular reference points- based approach (NSGA-III) [43] for many-objective optimization JSS . In particular, this research objective combines GP with the many- objective ﬁtness evaluation scheme of NSGA-III and compares the performance of the resulting hybridized algorithm with NSGA-II and SPEA2 in solving many-objective JSS. It is expected that the newly proposed method will evolve high-quality Pareto-optimal dis- patching rules that can effectively tackle not only the training prob- lem instances but also unseen problem instances. The new algorithm is also investigating factors that inﬂuence the trade-offs among dif- ferent objectives. In this objective, we will further study whether many typically considered scheduling objectives such as mean ﬂow- time, mean total tardiness, maximum tardiness, and maximum ﬂow- time are mutually conﬂicting or not. 2. Investigate how to develop GP-HH approaches for the non-uniform Pareto- front of many-objective JSS problems which can evolve high-quality Pareto- optimal dispatching rules. 1.3. RESEARCH GOALS 11 This study is mainly focused on addressing the diversity issue in many-objective JSS. In the literature, diversity maintenance is of- ten controlled by providing multiple predeﬁned reference points [31, 43]. These points are widely distributed. Therefore, solutions associated with these points are also widely distributed in the objec- tive space. Most of the algorithms, such as NSGA-III and RVEA [31, 43] employ a predeﬁned set of uniformly distributed reference points. However, as discussed in Section 1.1, the adoption of uniformly distributed reference points affect the solution diversity and deteriorate algo- rithm’s performance. To address this vital issue of useless reference points, the main goal of this research objective is to develop new and effective mechanisms for reference point generation. These mech- anisms improve the association between reference points and the Pareto-front during the whole evolutionary search process. Further, these new algorithms will be developed to increase the chance of discovering well-distributed solutions on the Pareto-optimal fronts. In this research objective, we explore both model-free and model- driven techniques to approximate the Pareto-front based on evolved dispatching rules accurately. In order to evaluate the performance of the proposed model-free and model-driven methods, we evaluate JSS problems and ten mathematical optimization benchmark prob- lems. 1. Investigate how GP can be used to handle many-objective JSS problems. These problems have irregular, disconnected, degenerate, and inverted shapes of the Pareto-front where the number of objectives is scaled from three to eight. The proposed algorithms will be com- pared with four state-of-the-art algorithms as well as adaptive refer- ence point algorithms. 3. Investigate how to hybridize a local search with a global search and improve the quality of the evolved rules in many-objective JSS. 3. Investigate how to hybridize a local search with a global search and improve the quality of the evolved rules in many-objective JSS. Many existing research studies analyzed the potential combination CHAPTER 1. INTRODUCTION 12 of the local search with evolutionary multi-objective optimization (EMO) algorithms [26, 79, 108]. PLS has three phases: (1) the se- lection of initial solutions phase from the non-dominated solution’s archive; (2) the exploration phase, in which the neighborhood of the selected solutions is explored, and candidate neighbors are ex- tracted; and (3) the Pareto-archive phase, in which the archive is up- dated with the candidate neighboring solutions. The main goal of this study is to build an effective many-objective PLS algorithm for JSS problems. Driven by this goal, this study thoroughly investigates the inclu- sion of PLS within GP-HH algorithms. Firstly, this research objec- tive builds an effective strategy for selecting initial individuals. In particular, a ﬁtness-based selection strategy will be proposed to bal- ance convergence and diversity measures properly. By using this ﬁtness-based selection strategy, we can select the representatives of each group of solutions. Secondly, the neighborhood structure of a tree (dispatching rule) in GP is deﬁned. Speciﬁcally, this can be done by using restricted subtree mutation. This mutation can pre- vent a new neighboring rule discovered during the local search pro- cess from being signiﬁcantly different from the original rule. Lastly, for comparing neighbor rule with its immediate parent, we consider the following two different strategies: (1) the scalarization strategy [79] and (2) the replacement strategy [26]. In the scalarization strat- egy, the objective vector of each rule is aggregated into a scalar using the weighted sum. The replacement strategy is based on the domi- nance relation. It is evidenced in the literature [79] that the total number of gener- ations for global search and the maximum number of local search steps is highly inﬂuential on the performance of hybridized algo- rithms. 1. Investigate how GP can be used to handle many-objective JSS problems. To further enhance the effectiveness of our new PLS algo- rithms, we investigate how to keep the balance between the number 1.4. MAJOR CONTRIBUTIONS 13 of generations (for global search) and local search steps in our pro- posed algorithms. It is expected that our proposed algorithms help to enhance the exploitation ability and increase the chance of discov- ering promising dispatching rules. 1.4 Major contributions This thesis makes the following major contributions. 1. This study proposes the ﬁrst many-objective GP-HH method for JSS problems to ﬁnd the Pareto-front of non-dominated dispatching rules by many conﬂicting objectives. The experimental results show that the combination of GP and NSGA-III produces a competitive al- gorithm as compared with NSGA-II and SPEA2 for evolving a set of trade-offs rules in many-objective JSS. The detailed analysis of the evolved Pareto-front reveals in-depth knowledge about the useful terminals for a large proportion of evolved dispatching rules. The rules evolved by the new algorithm also exhibit good generalization abilities as veriﬁed by performance on testing benchmark instances. 1. This study proposes the ﬁrst many-objective GP-HH method for JSS problems to ﬁnd the Pareto-front of non-dominated dispatching rules by many conﬂicting objectives. The experimental results show that the combination of GP and NSGA-III produces a competitive al- gorithm as compared with NSGA-II and SPEA2 for evolving a set of trade-offs rules in many-objective JSS. The detailed analysis of the evolved Pareto-front reveals in-depth knowledge about the useful terminals for a large proportion of evolved dispatching rules. The rules evolved by the new algorithm also exhibit good generalization abilities as veriﬁed by performance on testing benchmark instances. Part of this contribution has been published in: A. Masood, Y. Mei, G. Chen, and M. Zhang, ”Many-objective ge- netic programming for job-shop scheduling”, Evolutionary Computa- tion (CEC) 2016 IEEE Congress on, pp. 209-216, 2016. 2. This thesis presents two new approaches to dealing with the chal- lenges caused by the non-uniform distribution of Pareto-front in ob- jective space and proposes new adaptation mechanisms which im- prove the performance of evolutionary search and promote popula- tion diversity for better exploration. This contribution has the fol- lowing sub-contributions: CHAPTER 1. INTRODUCTION 14 (a) A new adaptive reference point strategy is proposed by us- ing the adaptive model-free approach and the reference points are generated according to the distribution of the solutions. A new reference point adaptation mechanism has been success- fully developed based on Particle Swarm Optimization (PSO). Essential changes to particle dynamics in PSO have also been in- troduced in our algorithm to prevent the majority of reference points from converging to small areas in the objective space. This reference point adaptation mechanism helps to reduce use- less reference points and signiﬁcantly improve the performance of the proposed algorithms. 1.4 Major contributions In the experimental evaluations based on the Taillard benchmark set, we successfully showed that the proposed reference point adaptation mechanism could signiﬁcantly improve the performance of GP-HH and NSGA-III in terms of both HV and IGD. (a) A new adaptive reference point strategy is proposed by us- ing the adaptive model-free approach and the reference points are generated according to the distribution of the solutions. A new reference point adaptation mechanism has been success- fully developed based on Particle Swarm Optimization (PSO). Essential changes to particle dynamics in PSO have also been in- troduced in our algorithm to prevent the majority of reference points from converging to small areas in the objective space. This reference point adaptation mechanism helps to reduce use- less reference points and signiﬁcantly improve the performance of the proposed algorithms. In the experimental evaluations based on the Taillard benchmark set, we successfully showed that the proposed reference point adaptation mechanism could signiﬁcantly improve the performance of GP-HH and NSGA-III in terms of both HV and IGD. Part of this contribution has been published in: Masood, A., Mei, Y., Chen, G., and Zhang, M.: A PSO- Based Reference Point Adaption Method for Genetic Program- ming Hyper-Heuristic in Many-Objective Job Shop Schedul- ing. ACALCI. Lecture Notes in Computer Science, vol. 10142, pp. 326–338 (2017). Masood, A., Mei, Y., Chen, G., and Zhang, M.: A PSO- Based Reference Point Adaption Method for Genetic Program- ming Hyper-Heuristic in Many-Objective Job Shop Schedul- ing. ACALCI. Lecture Notes in Computer Science, vol. 10142, pp. 326–338 (2017). (b) (b) This thesis develops a new adaptive strategy for generating ref- erence points based on a model-driven technique. This model- driven technique estimates the density of solutions from each pre-deﬁned sub-location in the entire objective space. Fur- thermore, the proposed algorithm provides smoothness to the model by applying a Gaussian Process on the density-based model which gives the ability to reduce the noise of the model and then calculate the area under the mean function. This area under the curve ﬁnds the required number of reference points in each sub-location. The experimental results demonstrate that 1.4. MAJOR CONTRIBUTIONS 15 a new adaptive mechanism can outperform several state-of-the- art MOEAs such as RVEA and NSGA-III as well as cutting-edge MOEAs that support adaptive reference points. Part of this contribution has been published in: Part of this contribution has been published in: Part of this contribution has been published in: Masood, A., Chen, G., Mei, Y., and Zhang, M.: Reference point adaption method for genetic programming hyper-heuristic in many-objective job shop scheduling. In: European Conference on Evolutionary Computation in Combinatorial Optimization. pp. 116–131. Springer (2018). Masood A., Chen G., Mei Y., and Zhang M.: Adaptive Refer- ence Point Generation for Many-Objective Optimization Using NSGA-III. In: Advances in Artiﬁcial Intelligence. AI 2018. Lecture Notes in Computer Science, vol 11320. pp. 358-370. Springer(2018). A. Masood, G. Chen,Y. Mei, H. Al-Sahaf, and M. Zhang, ”A Model-Based Approach for Many-Objective Optimization Al- gorithm with Adaptive Reference Point Generation”, To be sub- mitted for review to Genetic and Evolutionary Computation Confer- ence (GECCO 2020). A. Masood, G. Chen,Y. Mei, H. Al-Sahaf, and M. Zhang, Adap- tive Reference Point Generation based on Gaussian Process Model for ManyObjective Optimization”. Submitted to IEEE Transactions on Evolutionary Computation. (2020), 14pp, (un- der review. 3. This research objective investigates the usefulness of the PLS in GP- HH and develops two new algorithms that combine GP-HH as a global search with PLS as a local search. This is the ﬁrst algorithm of its kind for many-objective JSS that combines GP with PLS-based local search. In this objective, we propose the new selection mecha- nism for the initial solutions for neighborhood exploration in PLS. This mechanism deﬁnes representatives from each group of solu- 3. This research objective investigates the usefulness of the PLS in GP- HH and develops two new algorithms that combine GP-HH as a global search with PLS as a local search. This is the ﬁrst algorithm of its kind for many-objective JSS that combines GP with PLS-based local search. In this objective, we propose the new selection mecha- nism for the initial solutions for neighborhood exploration in PLS. This mechanism deﬁnes representatives from each group of solu- 3. This research objective investigates the usefulness of the PLS in GP- HH and develops two new algorithms that combine GP-HH as a global search with PLS as a local search. This is the ﬁrst algorithm of its kind for many-objective JSS that combines GP with PLS-based local search. In this objective, we propose the new selection mecha- nism for the initial solutions for neighborhood exploration in PLS. This mechanism deﬁnes representatives from each group of solu- CHAPTER 1. INTRODUCTION 16 tions and selects a solution based on their ﬁtness value. Part of this contribution has been published in: Masood A., Chen G., Mei Y.,H. Al-Sahaf, and Zhang M.: Genetic Programming with Pareto Local Search for Many-Objective Job Shop Scheduling. In: Advances in Artiﬁcial Intelligence. AI 2019. Lecture Notes in Computer Science. A. Masood, G. Chen,Y. Mei, H. Al-Sahaf, and M. Zhang, ”Fitness- based Selection Method for Genetic Programming with Pareto Local Search”, Evolutionary Computation (CEC) 2020 IEEE Congress on, pp. 1-8, 2020. Part of this contribution has been published in: Further, the proposed algorithm deﬁnes the proper neighborhood structure for tree-based dispatching rules. A new selection mechanism for se- lecting the suitable dispatching rules during the neighborhood ex- ploration is also deﬁned. A detailed sensitivity analysis is also per- formed. The analysis ﬁnds the total number of generations and the maximum number of local search steps in GP-PLS. 1.5 Organisation of thesis The remainder of this thesis is organized as follows. Chapter 2 presents a literature review of related works. Chapter 3 presents the experimental methodology of this thesis. Chapters 4–6, present the main contributions of the thesis. Chapter 7 concludes the thesis. • Chapter 2 presents descriptions of the JSS problem, many-objective optimization, local search, Pareto local search, GP, and dispatching rules used for this thesis. The basic concepts of meta-heuristics, heuristics, and hyper-heuristics for the automatic generation of heuristics are also presented. This chapter reviews current re- search methodologies using GP for the automatic generation of new 1.5. ORGANISATION OF THESIS 17 dispatching rules in single-objective, multi-objective, and many- objective JSS problems. • Chapter 3 presents the core experimental methodology of this thesis. It starts with the benchmark problems used throughout the thesis, the terminals and function set used to evolved dispatching rules. At the end of the performance measures for experimental results analy- sis are discussed in detail. • Chapter 4 develops a new many-objective GP-HH method to deal with the four conﬂicting objectives of JSS simultaneously. The pro- posed GP-HH method is used to evolve the non-dominated dis- patching rules that are represented by the obtained Pareto-front. An extensive comparison between the evolved rules from GP-NSGA- III and the two famous MOEAs (GP-NSGA-II and GP-SPEA2) have been carried out. Popular multi-objective performance measures are also used to help assess the performance of the compared algo- rithms. • Chapter 5 identiﬁes a key research issue involved in using the uni- formly distributed reference points, i.e., the failure to promote so- lution diversity during evolution which affected the performance of GP-HH. To solve these issues, two new reference point adaptation mechanisms are proposed which improve the association between reference points and the evolved Pareto-front, enhancing solution diversity, and hence the performance of the algorithm. This chap- ter consists of two sections: (i) the ﬁrst section develops an adaptive model-free approach and (ii) the second section proposes a model- driven approach for learning irregular distributions of the Pareto- front and systematically sampling reference points. The proposed algorithm introduces the model-based technique based on the Gaus- sian process model which estimates the density of solutions from each deﬁned sub-location in a whole objective space. 1.5 Organisation of thesis This density- • Chapter 5 identiﬁes a key research issue involved in using the uni- formly distributed reference points, i.e., the failure to promote so- lution diversity during evolution which affected the performance of GP-HH. To solve these issues, two new reference point adaptation mechanisms are proposed which improve the association between reference points and the evolved Pareto-front, enhancing solution diversity, and hence the performance of the algorithm. This chap- ter consists of two sections: (i) the ﬁrst section develops an adaptive model-free approach and (ii) the second section proposes a model- driven approach for learning irregular distributions of the Pareto- front and systematically sampling reference points. The proposed algorithm introduces the model-based technique based on the Gaus- sian process model which estimates the density of solutions from each deﬁned sub-location in a whole objective space. This density- CHAPTER 1. INTRODUCTION 18 based model learns the distribution of the candidate solutions and accurately approximates the Pareto-front based on the evolved so- lutions. To verify the performance of the proposed algorithm, it is applied to JSS problems and ten benchmark tests with three to eight- objective optimization problems. The proposed algorithm is com- pared with four state-of-the-art many-objective algorithms (RVEA) [31], IMMOEA [30], NSGA-III [43], Two-Arch [193]), as well as with adaptive reference many-objective algorithms [85]. • Chapter 6 develops hybrid GP-HH algorithms based on PLS to dis- cover high-quality dispatching rules for JSS problems. An innova- tive hybridization of global search (GP) with PLS techniques is pro- posed. In the proposed algorithm, a new selection mechanism for the initial individuals is also proposed. Further, a neighborhood struc- ture is described for GP. Moreover, a sensitivity analysis is performed for ﬁnding the total number of generations required to search the solution space extensively and a sufﬁcient number of local search steps utilizing the local search effectively. The key idea of this pro- posed algorithm is to applied local search effectively and explore the neighborhood of non-dominated dispatching rules. Extensive ex- periments are performed to understand the effectiveness of the pro- posed algorithm as compared to the other algorithms (without local search). The experimental studies are carried out using the Taillard static job-shop benchmark set. • Chapter 6 develops hybrid GP-HH algorithms based on PLS to dis- cover high-quality dispatching rules for JSS problems. An innova- tive hybridization of global search (GP) with PLS techniques is pro- posed. Chapter 2 This chapter presents the literature for many-objective job shop schedul- ing (JSS). The literature review starts with a background that covers fun- damental concepts, ideas, and algorithms. The literature review then dis- cusses the relevant research works in Section 2.2, ranging from genetic programming based hyper-heuristic (GP-HH ) approaches to scheduling problems, many-objective JSS to Pareto local search (PLS). Finally, this chapter concludes with a summary of the literature review in Section 2.3. 1.5 Organisation of thesis In the proposed algorithm, a new selection mechanism for the initial individuals is also proposed. Further, a neighborhood struc- ture is described for GP. Moreover, a sensitivity analysis is performed for ﬁnding the total number of generations required to search the solution space extensively and a sufﬁcient number of local search steps utilizing the local search effectively. The key idea of this pro- posed algorithm is to applied local search effectively and explore the neighborhood of non-dominated dispatching rules. Extensive ex- periments are performed to understand the effectiveness of the pro- posed algorithm as compared to the other algorithms (without local search). The experimental studies are carried out using the Taillard static job-shop benchmark set. • Chapter 7 summarizes the research goals that have been achieved, followed by the key ﬁndings of the thesis. Furthermore, the chapter highlights the potential research areas of future works. . • Chapter 7 summarizes the research goals that have been achieved, followed by the key ﬁndings of the thesis. Furthermore, the chapter highlights the potential research areas of future works. . 2.1 Background This section covers the basic concepts related to the research works dis- cussed in this thesis. This includes a deﬁnition of sequencing and schedul- ing problems, a deﬁnition of machine learning, the conceptual differ- ences between heuristics, meta-heuristics, hyper-heuristics, and dispatch- ing rules. It is then followed by the introduction of evolutionary compu- tation, many-objective optimization, and PLS. 19 CHAPTER 2. LITERATURE REVIEW 20 Sequencing/Scheduling Sequencing and scheduling are a decision-making process that aims to uti- lize limited resources as effectively as possible. Sequencing and schedul- ing studies on production planning so that the manufacturing systems generate output with minimal waste in time and money [111]. Sequencing selects a job to process next on a speciﬁc machine. Schedul- ing determines the resource (e.g., machine, work center) to handle each job and the time when the machine begins processing the selected job. Although sequencing and scheduling are considered two of the crucial de- cisions for generating a schedule, sometimes, it is not straightforward to utilize them separately due to the complexity of scheduling problems. In our thesis, we use dispatching rules. These rules choose the sequence of jobs at each decision point and begin the process to complete a plan or schedule. In general, sequencing and scheduling problems can be used to model a wide range of real-world scenarios. JSS is an essential example of scheduling which is brieﬂy discussed in Section 2.1.2. The other types of scheduling problems include ﬂowshop scheduling problems [53] and ﬂexible scheduling problems [165]. Machine learning Machine learning is a ﬁeld of computer science that focuses on the area of artiﬁcial intelligence [6]. Tom Mitchell [134] gave a “well-deﬁned” math- ematical and relational deﬁnition for machine learning which is “A com- puter program is said to learn from experience E with respect to some task T and some performance measure P, if its performance on T, as mea- sured by P, improves with experience E. In other words, machine learning allows computer programs to improve through experience automatically. Machine learning techniques are categorized into four major categories: 2.1. BACKGROUND 21 1. Supervised learning: The supervised machine learning algorithms handle labeled data with explicit values. The machine learning algo- rithms in supervised learning environments try to create a function to map inputs to desired outputs which are known in advance. Ex- amples of problems that are handled by supervised machine learn- ing algorithms are classiﬁcation and regression [172]. 1. Supervised learning: The supervised machine learning algorithms handle labeled data with explicit values. The machine learning algo- rithms in supervised learning environments try to create a function to map inputs to desired outputs which are known in advance. Ex- amples of problems that are handled by supervised machine learn- ing algorithms are classiﬁcation and regression [172]. 2. Unsupervised learning: The unsupervised machine learning algo- rithms deal with the unlabelled data and capture the patterns from a set of data. An example of problems which are handled by unsuper- vised machine learning algorithms is clustering and neural networks [172]. 3. Semi-supervised learning: The semi-supervised learning is a ma- chine learning algorithm that carries out both supervised and unsu- pervised learning. 4. Reinforcement learning: The reinforcement machine learning algo- rithm interacts with a dynamic environment and receives feedback in terms of rewards and penalties [172]. The goal of the learner, in this case, is to maximize the rewards and minimize the penalties [172]. Reinforcement learning is connected to applications for which the machine learning algorithm must make decisions as the learner is not trained to take action and must discover the actions that max- imize the rewards (e.g., playing a chess game). Heuristics/Meta-heuristics/Hyper-heuristics There are a number of deﬁnitions for heuristics, meta-heuristics, and hyper-heuristics. For this thesis, we deﬁne heuristics as ‘rules-of-thumb. These can be applied to problems directly [172]. These heuristics are useful in cases where it is impractical to ﬁnd exact solutions to complicated and large-scale problems [133]. For a particular problem, heuristics often incor- CHAPTER 2. LITERATURE REVIEW 22 porate existing knowledge about the problem to ﬁnd good solutions for the problem instances [17]. It means that heuristics are problem-speciﬁc [194]. In other words, a heuristic designed for one scheduling problem (a heuristic for minimizing makespan) may not be effective for another (a heuristic for minimizing ﬂowtime) scheduling problem. Meta-heuristics are deﬁned as problem independent methods that can be applied to solve a wide range of different problems [152]. Meta- heuristics solve problems by directly searching the problem’s solution space by incorporating lower level and problem-speciﬁc heuristics, which facilitate the search to ﬁnd good solutions [152]. There are many meta- heuristic techniques developed in the literature and they can be classiﬁed into two main categories: (1) local search-based [1] and (2) population- based [152]. The details of these meta-heuristics categories can be seen in Section 2.2.1. Hyper-heuristics are problem independent techniques that do not di- rectly solve problems, unlike heuristics and meta-heuristics. Hyper- heuristics combine low-level heuristic components and construct a suit- able heuristic for any speciﬁc problem [17, 20]. The details of these hyper- heuristics categories can be seen in subsection 2.2.1. Dispatching rules Most popular heuristics in the literature for solving scheduling problems are referred to as dispatching rules [73, 145]. Dispatching rules are also called priority rules that decide the processing order of the jobs in a queue. The priorities depend on several factors, such as the due date, arrival time, and processing time. The job with the highest priority is processed next to the available machine. Simplicity and ﬂexibility are the two advantages of dispatching rules [165]. Dispatching rules are perhaps the most straightforward method to deal with both static and dynamic JSS problems [20]. In the literature [90, 91], dispatching rules are divided into simple and composite. A simple rule is 2.1. BACKGROUND 23 a human-made priority function made using various scheduling param- eters (processing times, waiting times) such as SPT (shortest-processing- time). On the other hand, a composite dispatching rule is a combination of two or more simple rules. In this thesis, we use genetic programming (GP) to evolve dispatching rules (priority functions), comprising distinc- tive composite dispatching rules. These rules are applied to JSS based on the machine and job attributes. 2.1.2 JSS Problems JSS [165] is a signiﬁcant scheduling problem with a wide range of applica- tions in many industries, such as manufacturing [165] and cloud comput- ing [188]. JSS has been intensively investigated in the literature [62] and has been proven to be NP-hard [13]. This subsection covers the deﬁnitions and mathematical notations for JSS problems and JSS objective functions. Afterward, active and non-delay scheduling is deﬁned. Table 2.1: Notations in JSS processing information of all jobs is available such as arrival time and due date. On the other hand, in a dynamic environment, jobs continuously arrive and no prior information related to them is available in advance. When an operation has the ﬂexibility of being processed on more than one machine, it is called ﬂexible JSS problems. One of the simplest forms of JSS is a ﬂowshop, where all jobs follow the same processing order of op- erations. Table 2.1 shows the basic deﬁnitions and notations in JSS. In this thesis, we will use the GP to evolve dispatching rules for static JSS because many-objective research is comparatively new and the existing research only focus on optimizing single or multi objectives. Furthermore, other forms of JSS environments (dynamic and ﬂexible) are more complex than static JSS environments. This thesis evolved effective dispatching rules for static JSS and understand how optimization approaches can be adapted to improve many-objective JSS algorithms’ performance. JSS: Deﬁnition The general JSS problems are deﬁned based on a set of N jobs and a set of M machines. Each job ji, 1 ≤i ≤N has a sequence of m operations to be performed, i.e. {o1 i , o2 i . . . , om i }. Stringent restrictions apply to the processing order’s overall operations of a job. Speciﬁcally, for any job ji, the operation ok+1 i cannot start until its previous operation ok i has been completed. In other words, operations follow the precedence constraint. Besides, the operations are non-preemptive, i.e., once an operation starts to be processed on a machine, the process cannot be interrupted. Every operation ok i , 1 ≤k ≤m, is further associated with a ﬁxed processing time pk i > 0 and has to be processed on a speciﬁc machine mk i , 1 ≤i ≤M. Each job j has its pre-determined route and processing time on a spe- ciﬁc machine. Classiﬁcation of JSS depends on machine conﬁgurations and the nature of jobs [11]. In static JSS, the number of jobs is ﬁxed and 24 CHAPTER 2. LITERATURE REVIEW CHAPTER 2. LITERATURE REVIEW Table 2.1: Notations in JSS Notation Deﬁnition m1, m2, m3 . . . , mM Set of M machines j1, j2, j3 . . . , jN Set of N jobs ok i kth operation of job i Di due date for job i Ri release time of job i wi the weight given to job i rk i ready time of kth operation of job i Ci completion time of job i Ti Tardiness of job i Table 2.1: Notations in JSS Notation Deﬁnition m1, m2, m3 . . . , mM Set of M machines j1, j2, j3 . . . , jN Set of N jobs ok i kth operation of job i Di due date for job i Ri release time of job i wi the weight given to job i rk i ready time of kth operation of job i Ci completion time of job i Ti Tardiness of job i Objectives for JSS A schedule is evaluated by several criteria that can be classiﬁed as process- focused performance criteria and customer-focused due date criteria [59]. Process-focused performance criteria pertain to information about the start and end time of jobs and focus on shop performance, such as the utilization of machines. Flowtime and makespan are the most commonly 2.1. BACKGROUND 25 used performance criteria. Flowtime is the amount of time a job spends in a shop. Let Ci be the completion time and Ri be the release time of job i, then meanflowtime and then maxﬂowtime [165] are given as: meanﬂowtime = 1 N N X i=1 (Ci −Ri), (2.1) meanﬂowtime = 1 N N X i=1 (Ci −Ri), (2.1) maxﬂowtime = max ( 1 N N X i=1 (C1 −R1), 1 N N X i=1 (C2 −R21), . . . 1 N N X i=1 (Ci −Ri) ) (2.2) where N is a number of jobs (2.1) ime = max ( 1 N N X i=1 (C1 −R1), 1 N N X i=1 (C2 −R21), . . . 1 N N X i=1 (Ci −Ri) ) (2.2) i b f j b (2.2) where N is a number of jobs. where N is a number of jobs. where N is a number of jobs. Makespan is the total length of the schedule. In order to optimize the schedule, a minimum makespan (Cmax) is desired which is given as: Makespan is the total length of the schedule. In order to optimize the schedule, a minimum makespan (Cmax) is desired which is given as: Cmax = max {C1, C2, . . . , CN} (2.3) (2.3) Besides performance criteria, some customer-focused criteria are also frequently considered. For example, tardiness is one of the common customer-focused criteria. It measures the amount of time by which the completion time exceeds the due date. Tardiness of job i is deﬁned as Ti = max {Ci −Di, 0}. The TotalTardiness, TotalWeightTardiness, and MeanWeightTardiness are shown below TotalTardiness = N X i=1 {max(Ci −Di, 0)} , (2.4) (2.4) = N X i=1 wi {max(Ci −Di, 0)} . (2.5) (2.5) MeanWeightTardiness = 1 N N X i=1 wi {max(Ci −Di, 0)}. (2.6) (2.6) The maximum tardiness (Tmax) is another difﬁcult objective to minimize due to its sensitivity to the shop condition [147] which is deﬁne as: The maximum tardiness (Tmax) is another difﬁcult objective to minimize due to its sensitivity to the shop condition [147] which is deﬁne as: Tmax = max {T1, T2, . . . , TN} (2.7) (2.7) CHAPTER 2. LITERATURE REVIEW 26 Different studies have considered different objectives for JSS problems. Banu et al. [23] used a range of artiﬁcial intelligence approaches and explored their effectiveness in the JSS environment with different popu- lar objectives such as mean ﬂowtime, maximum ﬂowtime, percentage of tardy jobs, mean tardiness and maximum tardiness. schedules. The non-delay factor does not go beyond 1. schedules. The non-delay factor does not go beyond 1. The following is an example of a static JSS problem instance with the minimization of the makespan objective. This example shows the non- delay and an active schedule in JSS with α = 0 and α = 1, respectively. Table 2.2 shows job properties (processing order and processing time). Figures 2.1 and 2.2 show the non-delay and active scheduling for a JSS problem, respectively. In this problem, there are three jobs (j1, j2, j3) and two machines (m1 and m2). SPT (where operations with shorter processing times have a higher priority) is used to generate different types of schedules (active or non-delay). For the non-delay schedule, we can see that job j1 has the shortest processing time than j3. Therefore, m1 process j1 before j3. Then m1 begins processing j3 as soon as j1 has been completed. On the other hand, in the active schedule, machine m1 processes j2 before j3, as j2’s second operation has a shorter processing time than j3’s ﬁrst operation, and j2 arrives at machine m1 before the expected completion time of j3. For this problem instance, we can see that the non-delay SPT generates a better solution than active SPT. Active schedules and non-delay schedules When tackling JSS problems, it is essential to distinguish between active and non-delay scheduling [165]. In a non-delay scheduling, no waiting occurs before the machine starts processing the next job [165, 179]. In other words, a non-delay scheduling algorithm does not allow any delay on the idle machines as long as the waiting queue is not empty. For example, if machine m is available, the scheduling algorithm starts the new job ji from the waiting queue without any delay. In the case where no jobs are waiting at the machine m, the non-delay scheduling algorithm will wait until the next jobs arrive and select one of the jobs at the decision point which has the earliest arrival time. An active schedule allows some reasonable delay (which is no more than the minimal processing time of the waiting jobs) to handle any newly arriving jobs with an urgent due date [165]. Suppose that when a machine m ﬁnishes processing a job in an active schedule, a job that has the earliest expected completion time will be processed next by machine m. The non-delay factor α ∈[0,1] determines how many jobs should be considered to be processed next. If a scheduling algorithm has α = 0, then the scheduling algorithm is considered a non-delay schedule where the machine will immediately begin processing some operation as soon as there are jobs at the machine m. if α = 1, then the scheduling algorithm generates an active schedule for jobs that have arrived at the machine and gives preference to these jobs according to their earliest completion time. Finally, if α is between 0 and 1, then the scheduling algorithm is considered a hybrid schedule which has characteristics of both active and non-delay 27 2.1. BACKGROUND Table 2.2: Example of a static JSS problem instances (N = 3,M =2) job machine sequence processing time j1 m1, m2 1,2 j2 m2, m1 2,1 j3 m1, m2 3,1 Evolutionary algorithms (EAs) EAs are inspired by the principles of natural evolution such as selection, reproduction, crossover, and mutation. In EA, a population of individ- uals s selected by their ﬁtness value (i.e., individuals’ survival depends on their ﬁtness). After the selection, offspring are generated by crossover and mutation. The genetic algorithms (GAs) [69] and GP [105] are the two well-known members of EA that are brieﬂy introduced here. • Genetic algorithms (GAs): GAs work on ﬁnite populations. Each population consists of N chromosomes (solutions) which are typi- cally represented as a ﬁxed-length array [69]. These arrays can rep- resent as bits, integer numbers, or real numbers that carry the infor- mation. However, these typical arrays can be decoded to solutions before solving any problem. Then solutions are evaluated according to a speciﬁed ﬁtness function and parents are selected to produce off- spring through the genetic operators (crossover and mutation). The resulting offspring inherit properties directly from their parents. The offspring are evaluated and a ﬁtter individual is placed in the new population which replaces the weaker solution. The GAs mechanism consists of three phases: evaluation of each solution’s ﬁtness, selec- tion of the parent, and applications of mutation and recombination (crossover) operators to the parent for producing offspring. The pro- cess is repeated until to have the desired quality of solution or reach the maximum generation. • Genetic algorithms (GAs): GAs work on ﬁnite populations. Each population consists of N chromosomes (solutions) which are typi- cally represented as a ﬁxed-length array [69]. These arrays can rep- resent as bits, integer numbers, or real numbers that carry the infor- mation. However, these typical arrays can be decoded to solutions before solving any problem. Then solutions are evaluated according to a speciﬁed ﬁtness function and parents are selected to produce off- spring through the genetic operators (crossover and mutation). The resulting offspring inherit properties directly from their parents. The offspring are evaluated and a ﬁtter individual is placed in the new population which replaces the weaker solution. The GAs mechanism consists of three phases: evaluation of each solution’s ﬁtness, selec- tion of the parent, and applications of mutation and recombination (crossover) operators to the parent for producing offspring. The pro- cess is repeated until to have the desired quality of solution or reach the maximum generation. • Genetic programming (GP): GP is a “search-based automatic pro- gramming” technique [105]. 2.1.3 Evolutionary computation Evolutionary computation (EC) is a sub-ﬁeld of artiﬁcial intelligence which focuses on algorithms inspired by the principle of the Darwinian theory of biological evolution [95]. This includes nature-inspired algo- rithms or population-based systems to deal with various problems. In general, EC techniques are divided into two main categories: (1) evolu- 28 CHAPTER 2. LITERATURE REVIEW Figure 2.1: The schedule generated by non-delay SPT α = 0. Figure 2.2: The schedule generated by active SPT α = 1. tionary algorithms (EAs) [95] and (2) swarm intelligence (SI) [101]. There is also a third category called miscellaneous algorithms, including the learn- ing classiﬁer system [134]. Figure 2.1: The schedule generated by non-delay SPT α = 0. Figure 2.1: The schedule generated by non-delay SPT α = 0. Figure 2.2: The schedule generated by active SPT α = 1. Figure 2.2: The schedule generated by active SPT α = 1. tionary algorithms (EAs) [95] and (2) swarm intelligence (SI) [101]. There is also a third category called miscellaneous algorithms, including the learn- ing classiﬁer system [134]. tionary algorithms (EAs) [95] and (2) swarm intelligence (SI) [101]. There is also a third category called miscellaneous algorithms, including the learn- ing classiﬁer system [134]. 2.1. BACKGROUND 29 Swarm Intelligence (SI) SI algorithms are inspired by the collective intelligence of a group of sim- ple agents [12]. In SI, the swarm is an abstract representation of a group of biological organisms that interact with each other to reach a speciﬁc goal, e.g., ﬁnding food. There are ﬁve basic principles of swarm intelli- gence: proximity, quality, diverse response, stability, and adaptability. A well-known SI algorithm is: • Particle swarm optimization (PSO): PSO [101] is a swarm intelli- gence technique where the particles represent a group of organisms such as a ﬂock of birds. The movement of an individual in PSO is inﬂuenced by the search direction of the local best (position of the one-particle) and global best (found by all particles). The feedback between the individuals allows PSO to identify high-quality solu- tions to a given problem quickly. Evolutionary algorithms (EAs) GP is an extension of GAs where a pop- ulation of computer programs is generated with variable length. The application of reproduction operators (elitism, crossover, and mu- tation) produces new programs applied in GP. These programs are then tested against ﬁtness function and those who have higher ﬁt- ness values are more likely to survive to future generations. In GP, • Genetic programming (GP): GP is a “search-based automatic pro- gramming” technique [105]. GP is an extension of GAs where a pop- ulation of computer programs is generated with variable length. The application of reproduction operators (elitism, crossover, and mu- tation) produces new programs applied in GP. These programs are then tested against ﬁtness function and those who have higher ﬁt- ness values are more likely to survive to future generations. In GP, CHAPTER 2. LITERATURE REVIEW 30 the programs are traditionally represented as tree structures. Other data structures (linear,graph) have also been employed to construct computer programs. GP is explained in detail in Subsection 2.1.9. 2.1.4 Multi-objective EAs (MOEAs) Over the last three decades, MOEAs have been successfully developed to handle various constrained and unconstrained multi-objective optimiza- tion problems (MOPs). In general, MOEAs have two ultimate goals with respect to performance which are convergence (minimizing the distance to the Pareto-front ) and diversity (widely and uniformly distributed over the Pareto-front front) [42]. MOEAs have been applied to those problems with two or three objectives. Without loss of generality, in general, MOPs, can be formulated as: min f−→ (x) = {f1(−→x ) . . . fm(−→x )} : s.t. −→x ∈F f ∈Y, (2.8) (2.8) 2.1. BACKGROUND 31 where F is the complete set of feasible solutions and x ∈F is one of the feasible solution, Y ⊂Rm is the objective space and f(x) = (f1(−→x ) . . . fm(−→x )) ∈Y is the objective vector. Here m stands for the number of objectives and m either has values two or three. Because objectives are mutually conﬂicting, no single solution can op- timize all objectives simultaneously. Instead, we aim to obtain a set of op- timal solutions that represent different trade-offs between these objectives. In multi-objective optimization, the comparison of two solutions (x1 and x2) is based on the concept of domination [42]. The solution x1 dominates x2 if and only if ∀i, 1 ≤i ≤m, fi(x1) ≤fi(x2), (2.9) ∃i, fi(x1) < fi(x2). (2.10) (2.9) (2.10) If any other solution does not dominate a solution x1, then it is called a Pareto-optimal solution (x∗). The set of all Pareto-optimal solutions jointly forms the Pareto-front in the objective space and the Pareto Set (PS) in the decision space. Two very popular MOEAs are: If any other solution does not dominate a solution x1, then it is called a Pareto-optimal solution (x∗). The set of all Pareto-optimal solutions jointly forms the Pareto-front in the objective space and the Pareto Set (PS) in the decision space. Two very popular MOEAs are: • Non-dominated sorting GA-II (NSGA-II): The ﬁrst popular MOEA is NSGA-II, proposed by Deb et al. [44]. NSGA-II is a parameterless elitist strategy which ﬁrst combines the parent and offspring pop- ulations and then performs non-dominated sorting on a combined population followed by the crowding distance. • Non-dominated sorting GA-II (NSGA-II): The ﬁrst popular MOEA is NSGA-II, proposed by Deb et al. [44]. 2.1.4 Multi-objective EAs (MOEAs) NSGA-II is a parameterless elitist strategy which ﬁrst combines the parent and offspring pop- ulations and then performs non-dominated sorting on a combined population followed by the crowding distance. Non-dominated sorting uses a dominance rank. The solutions that are not dominated by any other population solutions are assigned the dominance rank of one. Then, the next set of solutions which are only dominated by solutions with a dominance rank of one, is assigned a dominance rank of two, and so forth. Crowding distance calculates the average distance of two neighbor- ing solutions. The best solutions (by ﬁtness and spread) are selected through tournament selection for the genetic operators to create the new population. CHAPTER 2. LITERATURE REVIEW 32 • Strength Pareto evolutionary algorithm 2 (SPEA2): The second popular MOEA is SPEA2 proposed by Zitzler et al. [211]. Like NSGA-II, is also an elitist method. SPEA2 has a separate ﬁxed-size archive where the population is updated at each generation with non-dominated individuals. The archive can eliminate individuals if the non-dominated fronts exceed the ﬁxed size. Individuals’ ﬁt- ness is a combination of a dominance strength (which is the number of solutions in the archive and current population dominated by the individual) and density information. There are many other MOEAs in the literature, such as MOEA/D [205]. This thesis evolves trade- off (non-dominated) dispatching rules for many-objective JSS prob- lems, therefore, we focus on Pareto dominance based MOEAs such as NSGA-II and SPEA2. • Strength Pareto evolutionary algorithm 2 (SPEA2): The second popular MOEA is SPEA2 proposed by Zitzler et al. [211]. Like NSGA-II, is also an elitist method. SPEA2 has a separate ﬁxed-size archive where the population is updated at each generation with non-dominated individuals. The archive can eliminate individuals if the non-dominated fronts exceed the ﬁxed size. Individuals’ ﬁt- ness is a combination of a dominance strength (which is the number of solutions in the archive and current population dominated by the individual) and density information. There are many other MOEAs in the literature, such as MOEA/D [205]. This thesis evolves trade- off (non-dominated) dispatching rules for many-objective JSS prob- lems, therefore, we focus on Pareto dominance based MOEAs such as NSGA-II and SPEA2. • MOEA based on decomposition (MOEA/D): MOEA/D [205] de- composes MOP to single-objective problems by employing three possible decomposition functions [205]. 2.1.4 Multi-objective EAs (MOEAs) These functions are the weighted sum function, the Chebyshev function, and the penalty-based boundary intersection function, which decompose the multiple-objective problems into a set of scalarizing subproblems. It maintains population diversity by a predeﬁned set of weight vectors. Several variants of MOEA/D have been proposed for enhancing the selection strategy of each sub-problem [205]. • MOEA based on decomposition (MOEA/D): MOEA/D [205] de- composes MOP to single-objective problems by employing three possible decomposition functions [205]. These functions are the weighted sum function, the Chebyshev function, and the penalty-based boundary intersection function, which decompose the multiple-objective problems into a set of scalarizing subproblems. It maintains population diversity by a predeﬁned set of weight vectors. Several variants of MOEA/D have been proposed for enhancing the selection strategy of each sub-problem [205]. 2.1. BACKGROUND 2.1. BACKGROUND 33 Prominent examples include water resource engineering [138], nurse ros- tering [180], car control system design [140], and air trafﬁc control [65]. In recent years, MaOPs have drawn great attention in the EMO commu- nity, intending to tackle various new challenges caused by handling many concurrent objectives. Prominent examples include water resource engineering [138], nurse ros- tering [180], car control system design [140], and air trafﬁc control [65]. In recent years, MaOPs have drawn great attention in the EMO commu- nity, intending to tackle various new challenges caused by handling many concurrent objectives. MaOPs pose a number of challenges to Pareto-based MOEAs [84]. First, the number of non-dominated solutions increases at an exponential scale with an increasing number of objectives. Therefore, most Pareto- based MOEAs, such as NSGA-II and SPEA2, cannot provide sufﬁcient selection pressure towards the Pareto-front. Second, commonly used diversity-preservation operators such as crowding distance [44] and clus- tering [211]. These operators become computationally expensive when they used in MaOPs [84]. 2.1.5 Many-objective optimization (MaOPs) Over the years, evolutionary multi-objective (EMO) algorithms have two performance-centric considerations: convergence (minimizing the distance to the Pareto-front) and diversity (widely and uniformly distributed over the Pareto-front) on a problem with two or three conﬂicting optimization objectives [42]. If MOPs have four or more than four objectives (i.e., m ≥ 4), then it is called MaOPs [116]. Many real-world problems have been naturally deﬁned as MaOPs [43]. Categories of many-objective optimization algorithms Many-objective optimization approaches are roughly classiﬁed into the three classes: (1) dominance-based approaches, (2) decomposition-based approaches, and (3) indicator-based approaches. • Dominance-based approaches solve the MaOPs by enhancing the convergence pressure towards the Pareto-front. Since most of the MOEAs lose their selection pressure toward the Pareto- frontbecause of Pareto-dominance’s inability to distinguish solu- tions. To solve this issue, most MOEAs use the convergence en- hancement which modiﬁes the dominance relationship to increase the Pareto-frontselection pressure. Several dominance modiﬁca- tion approaches have been proposed, such as grid-dom [199] and α-dom.[77]. Many other algorithms are also used to combine the additional convergence-related metrics with the Pareto-dominance- based criterion. An example is an indicator-based preference com- bined with the dominance criterion for speeding up the convergence CHAPTER 2. LITERATURE REVIEW 34 of NSGA-II [27]. One of the recently proposed algorithms, the knee point-driven EA (KnEA) [208], enhances the convergence pressure by assigning the higher selection to the non-dominated fronts’ knee points in the current population. Dominance-based approaches en- large the dominating area of the non-dominated solution but make diversity maintenance more difﬁcult [116]. • Decomposition-based approaches convert a complex MOP to a number of sub-problems and simultaneously solve them. They are another promising method for MaOPs. This class of MaOPs is further divided into two other types of decomposition-based ap- proaches. • Decomposition-based approaches convert a complex MOP to a number of sub-problems and simultaneously solve them. They are another promising method for MaOPs. This class of MaOPs is further divided into two other types of decomposition-based ap- proaches. The ﬁrst type of decomposition-based approach decomposes multi- ple objectives into a single objective using scalarizing functions and weighting vectors. These approaches use the aggregated criteria to distinguish solutions. Multiple single-objective Pareto sampling [72] and MOEAs based on decomposition (MOEA/D) [205] are the two famous aggregation-based algorithms. These approaches have two major challenges: selecting weighting vectors and diversity main- tenance subject to the limitation of adopted scalarizing functions. Several variants of MOEA/D provide a better balance between con- vergence and diversity by enhancing the selection strategy for each sub-problem [205]. In the second type of decomposition-based approach, an MOP is de- composed into other sub-MOPs. For instance, the research studies in [31, 43] partition the whole Pareto-front into a Pareto-front sub- set, and each sub-Pareto-front can be considered a sub-problem. Categories of many-objective optimization algorithms An- other MOEA that mainly falls under this category is NSGA-III [43] that employs a set of uniformly distributed reference points to pre- serve the diversity of the candidate solutions. The results on several test problems and practical problems have shown NSGA-III’s useful- 2.1. BACKGROUND 35 ness in solving three- to 15-objective constrained and unconstrained optimization problems [43], [86]. The reference vector-guided EA (RVEA) [31] is one of the promi- nent decomposition-based approaches. RVEA uses reference vec- tors to decompose the MOPs into a number of single-objective sub- problems via objective function aggregations and use user prefer- ences to target preferred subspaces of the whole Pareto-front. • Indicator-based approaches Indicator-based approaches use indica- tor values to guide the search process. An example of indicator- based approaches is the S-metric selection-based evolutionary multi- objective algorithm [132] and a dynamic neighborhood MOEA based on HV indicator [117, 18]. These approaches do not have dominance- based MOEAs issue for solving MaOPs. Unfortunately, the calcu- lation of the performance indicator (HV) becomes computationally expensive especially in the case of many-objective [18]. A few other approaches do not fall into any of the above three main categories. These approaches include the diversity-based approaches, preference-based approaches, and dimensionality reduction-based ap- proaches. The diversity-based approaches are used in the selection of a solution when the primary selection criterion fails. Diversity-based approaches try to minimize the adverse impact of diversity in MaOPs. Many diversity measures have been proposed, such as the shift density estimator [119] and the grid-based neighborhood niching [199]. The Preference-based approaches select the biased region of search space with the help of user preference. This approach can be further categorized into three classes, a priori, interactive, and posterior methods, according to the timing when users give their preference (before, during, or after the optimization). Dimensionality reduction-based approaches try to reduce dimensionality CHAPTER 2. LITERATURE REVIEW 36 by reducing the number of objectives. They might lose some important information as a result of collectively handling several relevant objectives. It is expected from population-based optimization algorithms to con- vergence its population near the Pareto-optimal front and distributed uni- formly around the entire front. In this thesis, we evolve dispatching rules which will help decision-makers by providing them with potential trade-offs among different objectives. In this scenario, indicator-based ap- proaches can not be a good choice because they have high computational cost. Categories of many-objective optimization algorithms Further, few indicators-based approaches are not strictly monotonic with Pareto dominance and might lead to performance degradation. Next, dominance-based approaches modify the Pareto-dominance idea by en- larging the non-dominated solutions’ dominating area and improving the convergence criteria. However, the drive toward a more aggressive se- lection pressure seems to make diversity maintenance more difﬁcult in a dominance-based MOEAs. In contrast, decomposition-based approaches distribute reference points evenly in the whole objective space and man- age the candidate solutions’ diversity, eventually contributing to enhanced convergence of the algorithm and ﬁnding well-converge and widely dis- tributed sets of Pareto-optimal solutions. Non-dominated sorting GA-III (NSGA-III) The basic framework of NSGA-III is similar to the NSGA-II algorithm. The difference between NSGA-II and NSGA-III is the maintenance of diversity among population members. NSGA-II uses the crowding distance mea- sure for selecting a well-distributed set of Pareto-optimal solutions. In contrast, NSGA-III is aided by the supplied well-spread reference points (Zr) for selecting the Pareto-optimal solutions. In NSGA-III, the parent population (Pg) of size N is randomly gener- ated in the speciﬁed problem domain which is then followed by the selec- tion of binary tournament, application of crossover and mutation opera- tors to Pg. Then, an offspring population Qg of size N is created. This is 2.1. BACKGROUND 37 then followed by the combination of Pg and Qg which is called Rg (Pg ∪ Qg) of size 2N. The combined population (Rg) is sorted according to different non- domination ranks (F1, F2, . . . Fk). Each rank (F1, F2, . . . Fk) is selected one at a time and stored in a new population Sg of size N. After the non- dominated sorting, members of all the acceptable levels and the last level (lth level) are stored in Sg. The lth level is usually accepted partially. The niche-preservation operator selects the members of the well-distributed solution of Fl. The supplied reference points (Zr) are used to select the remaining members in NSGA-III. The reference points are generated systematically [39] and place on the normalized hyper-plan. Objective values are also normalized so that they have an identical range of reference points. There- after, the perpendicular distance between a member in Sg and each of the reference point is calculated. The member is then associated with the ref- erence point having the smallest perpendicular distance. Next, the niche counts ρ for each reference point is calculated. The reference point having the minimum associated solution is identiﬁed and the associated mem- bers of the last rank Fl enter in the ﬁnal population. The selected reference point’s niche count is increased by one and the procedure is repeated to ﬁll up the population (Pg+1) of a new generation. Issues in handling many-objective optimization The literature has discussed that EMO algorithms may face the difﬁcul- ties of scalability and diversity measures [43]. In this thesis, we focus on scalability and diversity issues which are brieﬂy discussed below: • Scalability (in terms of a number of objectives) is considered as one of the major challenges where MOEAs cannot provide sufﬁcient pressure toward the Pareto-front. Conventional Pareto-based EMO algorithms encounter difﬁculties when the number of objectives in- CHAPTER 2. LITERATURE REVIEW 38 creases [43]. This is because all the candidate solutions become non- dominated and the Pareto-based selection criterion cannot clearly distinguish among these candidate solutions. Therefore, NSGA-II and it is contemporary fail to handle more than three objectives prob- lems. • Diversity Maintenance plays an important role in the selection of the ﬁttest solutions when the primary selection criterion (conver- gence) cannot readily ﬁnd good solutions in EMO. For the sake of enhancing diversity, EMO promotes the evolution of uniformly dis- tributed candidate solutions [43]. However, in the case of a high dimension, it is quite challenging to maintain solution diversity be- cause solutions distribute very sparsely in a high-dimensional space. Moreover, diversity-preservation operators of EMO such as crowd- ing distance [44] and clustering [211] cannot strengthen the selection pressure toward the Pareto-front and also suffers from high compu- tational cost. Diversity maintenance can be controlled by providing multiple predeﬁned reference points. Most of the algorithms [31, 43] employ a predeﬁned set of uniformly distributed reference points to encourage the generation of wide distributed solutions on the whole objective space. 2.1.6 MaOPs for irregular Pareto-front It has been veriﬁed that the many-objective algorithms’ performance is highly dependent on the curvature of the Pareto-front [80, 118]. Hence many algorithms that use uniformly distributed reference points can per- form very well on the MaOPs when Pareto-front has regular shapes [86] such as DTLZ-1 [43]. RVEA [31] and NSGA-III [43] show remarkable results, especially when Pareto-front is uniform. However, their perfor- mance deteriorates considerably when these algorithms are applied to ir- regular, non-uniform, and discontinuous Pareto-front [85, 86]. Recently, 2.1. BACKGROUND 39 Ishibuchi et al. [80] found that the reference-point-based algorithms can perform effectively if the distribution of reference points is consistent with the distribution of Pareto-optimal solutions. Motivated by this discov- ery, several algorithms such as A-NSGA-III [85] and RVEA* (extension of RVEA) [31] have been proposed which adjust reference points adaptively. These reference-point-adaptation algorithms mainly learn the distribu- tion of the candidate solutions and adjust the reference points accordingly. The reference points based EA for many-objective optimization (REPA) [123] is an extension of NSGA-III which generates a set of reference points according to the current population. In the environmental selection pro- cess, the ﬁttest individuals are selected by calculating the Euclidean dis- tance between the reference points and individuals. A-NSGA-III [85] is a notable improvement in NSGA-III. This extension of NSGA-III is designed to relocate the reference points by two major op- erations dynamically: (1) inclusion and (2) exclusion. In the inclusion pro- cedure, multiple reference points are added around the crowded reference points with a high niche count. These reference points are generated by a centroid method where the original points are considered as a centroid. However, in A-NSGA-III, it is not easy to introduce new points around the vertices of simplex because this violates the positive quadrant’s condition [85]. Due to this reason, A-NSGA-III cannot avoid useless points com- pletely and fails to guide the evolution of a well-distributed set of Pareto- optimal solutions. RVEA* [31] is a well-known state-of-the-art algorithm for reference point adaptation. RVEA* uses the replacement approach instead of the deletion and addition operations. RVEA* ﬁrst partitions the current popu- lation into different subspaces. Next, reference points around these empty subspaces are replaced by the new non-empty subspace locations based on objective values’ ideal and nadir points. CHAPTER 2. LITERATURE REVIEW 40 Generation of reference points Without loss of generality, this thesis’s reference points are positive (they are all in the ﬁrst quadrant). In this approach, a set of uniformly dis- tributed reference points (z) is generated by the canonical simplex-lattice design method [35] as    zi k ∈ 0 H , 1 H , . . . , H H , Pm i=1 zi k = 1 ⃗zk = (z1 k, z2 k, . . . , zm k ), k = 1, 2, . . . , K (2.11) (2.11) where i = 1, 2, . . . , K are the K uniformly distributed points, m is the num- ber of objectives, and H is a positive integer used in the simplex-lattice design. where i = 1, 2, . . . , K are the K uniformly distributed points, m is the num- ber of objectives, and H is a positive integer used in the simplex-lattice design. 2.1. BACKGROUND 41 In the earlier research of EDA, conditional probabilistic modeling has been used which enables piece-wise interaction between candidate solu- tions. The research studies in [41, 161] show an elementary work of pair- wise intersection in the linear model. On the other hand, complex interac- tions among the solutions can also be handled through the Bayesian net- work as a multivariate model [160]. The Bayesian network is also used to build a multi-objective model and to balance convergence and diversity. For example, Bayesian multi-objective optimization algorithm (BMOA) [110] introduces the new selection operator ϵ which is based on a ϵ-archive that keeps track of a minimal set of solutions that dominates all other so- lutions generated so far. One of the research studies in [151] proposed a Voronoi-based EDA which helps to put the candidate solutions on differ- ent fronts. The multi-objective EDAs can effectively preserve the diversity of the population is shown in [15, 186, 35]. Multi-objective EDA usually builds the model in a decision space but the study in [30, 36] suggests directly controlling the distribution of the solutions through an inverse model. Notably, the Gaussian process-based inverse model aims to construct the Gaussian process model that maps the optimal solutions from the objec- tive space to the decision space [30]. This model is subsequently utilized to sample candidate solutions with the help of a reproduction operator. Similar to the related works summarized above, in this thesis, we will adopt a modeling technique to tackle one of the challenges of MaOPs. 2.1.7 Model-based EMO Traditional EAs such as NSGA-II and SPEA2 generate new individuals for the next generation using genetic operators (e.g., crossover and muta- tion) [42]. These EAs generate solutions without utilizing any problem- speciﬁc knowledge. Unlike these algorithms, model-based evolutionary algorithms (MBEAs) are designed to learn desired features or structure problems explicitly. Recently, many MBEAs have been proposed and one of the recent sur- veys in [29] classiﬁed MBEAs into three categories: (i) Estimation of Dis- tribution Algorithms (EDAs) [109], (ii) Inverse Modelling [30], and (iii) Surrogate Modelling [93]. EDA is the most popular category of MBEA and it mainly focuses on estimating the solution distribution. EDA estimates the distribution of the Pareto-optimal solutions by training and sampling in the decision space [109]. In EDA, a machine learning model such as a regression model, is iteratively reﬁned to estimate the distribution of solutions as the evolution proceeds. 2.1. BACKGROUND 2.1.8 Gaussian process modelling A Gaussian process (GaP) is a stochastic process which has a collection of random variables and can be seen as a generalization of multivariate Gaussian distribution in the function space [170]. In particular, if any ﬁnite sub-collection of random variables r(x) : x ∈X are drawn from a GaP, these variables are fully speciﬁed by a CHAPTER 2. LITERATURE REVIEW 42 mean function µ(x) and a covariance function K(xi, xj). The following equa- tion can denote the GaP:: r(x) ∼GaP(µ(X), K(xi, xj)), (2.12) (2.12) where the covariance function K(xi, xi) measures the correlation between any of the two arbitrary xi and xj which is expressed as: K =   K (x1, x1) · · · K (x1, xj) ... ... ... K (xi, x1) · · · K (xi, xj)  . (2.13) (2.13) One of the most commonly used covariance functions is the squared exponential in which the correlation of xi and xj is measured as a function of their distance: K (xi, xj) = σ2 f exp − n X k=1 θk xk i −xk j 2 ! , (2.14) (2.14) where σ2 f and θk are hyperparameters of the covariance functions. Assuming the regression model in equation (2.15) which shows the re- lationship between x (input vector) and y (output vector), we have: y (xi) = r (xi) + ϵ (xi) , (2.15) (2.15) where ϵ (x) ∼N (0, σ2) and i = 1, 2 · · · , m. Training of GaP focuses on the mean function µ(x) and the covariance function K(xi, xj). The prior distribution of the Gaussian process regres- sion model is shown below as a latent function. Training of GaP focuses on the mean function µ(x) and the covariance function K(xi, xj). The prior distribution of the Gaussian process regres- sion model is shown below as a latent function. r (·) ∼GaP (0, K(·, ·)) , (2.16) (2.16) In order to calculate this posterior distribution of the model, a distribution over the hyperparameters p(θ\|y, X) needs to be deﬁned. In general, log- likelihood could be used for estimating hyperparameter values which is explained as: BACKGROUND 43 BACKGROUND 43 2.1. BACKGROUND 43 logp(θ\|y, X) (θ) = −1 2 log\|K\| + y⊤K−1y + N log σ2 , (2.17) (2.17) Based on the training set X, a covariance matrix K of size N × N can be calculated. 2.1.8 Gaussian process modelling With respect to any given input vector x∗, the corresponding predicted output y∗can be obtained by a Gaussian probability distribution with the mean and covariance given as: µ(y∗) =K⊤ ∗ K + (σn)2I −1 y∗, (σ(y∗))2 =K∗∗−K⊤ ∗ K + (σn)2I −1 K∗. (2.18) (2.18) where K∗=[K (x1, x∗) , ...., K (xN, x∗)] is the N × 1 vector of covariances between the test and the training cases, and K∗∗is the covariance between the test input itself. CHAPTER 2. LITERATURE REVIEW 44 Representation In GP, solutions are normally represented as Lisp S-expression. Tree-based GP is represented by a tree structure. Trees are constructed by internal nodes and leaf nodes. In general, leaf nodes are called terminals and they consist of variables and constant numbers. On the other hand, internal nodes are functions deﬁned by user in terms of different operators such as arithmetic operators {+, −, ∗, /} and logical operators {OR, AND, XOR} [178]. An example of a tree-based GP is given in Figure 2.3. In the GP tree, the selection of a terminal set and function set can be problem-speciﬁc. For example, ready time {RJ}, number of remaining operations {RO}, work remaining of job {RT}, operation processing time {PR}, weight {w}, due date {DD}, and machine ready time {RM} are commonly considered job shop attributes [145]. + PR * RO DD Figure 2.3: An example of a GP :PR + (RO × DD)). RO Figure 2.3: An example of a GP :PR + (RO × DD)). 2.1.9 Genetic programming (GP) GP [105] is an EC-based meta-heuristic search method which can evolve computer programs to perform speciﬁc tasks automatically. GP individ- uals are represented as variable-length data structures. Different data structures have been used for GP representation, such as tree-based [105], graph-based [120] and linear-based [203] representations. The tree-based representation is ﬂexible in nature because the size of the evolved tress can vary from minimum depth to maximum depth [40]. This means GP has the potential to cover a larger search space of potential functions GAs (because of ﬁxed-length representation). Due to its ﬂexibility, GP can be applied to various machine learning and optimization problems, including numerous combinatorial optimization problems such as JSS. Tree-based GP has been widely used in JSS literature to evolve dispatching rules [145, 147]. This thesis will focus on using tree- based GP for evolving dispatching rules to construct schedules. CHAPTER 2. LITERATURE REVIEW Evaluation A driving force of GP is the ﬁtness function. It helps guide the search to ﬁnd good individuals from the population-based on the ﬁtness values of those individuals. The uses of the ﬁtness function depend on the problem domain. For example, consider a JSS problem to minimize tardiness. In this problem, the ﬁtness value of each individual is calculated as follows. First, the individual from a tree-based GP population is applied to train- ing instances as a priority dispatching rule and generates solutions. After- ward, the ﬁtness function calculates the mean tardiness of the solutions. If an individual A1 has a ﬁtness value lower (in a minimization problem) than an individual A2, A1 is considered better than individual A2. Initialization The initial individuals are usually generated by two earliest initialization methods: full and grow methods [105]. In the full method, nodes are se- lected from the function set and appended to the tree until the tree’s depth reaches a predeﬁned maximum depth. After that, the members of the ter- minal set are used as leaf nodes. In the full initialization procedure, all the terminals are located at a maximum depth of the trees. In the grow approach, the leaf nodes do not need to be at maximum depth. Therefore, the nodes are randomly picked from both the terminal 2.1. BACKGROUND 45 set and the function set. The grow method generates trees with more va- riety in shape and size as compare to the full method. Another initialization method is the ramped half-and-half method [104]. In order to generate a wide variety of trees of various depths, lengths, and shapes, the full and grow methods are often combined. In the ramped half-and-half method, half of the trees are created by the full method and the other half are constructed by the grow method. Selection In a natural evolution, next-generation is produced by selecting the ﬁttest individuals from the current population. GP follows a similar approach and after the evaluation phase, ﬁtness values are used to select individu- als to generate offsprings for the next generation. It should be noted that better individuals are more likely to survive in the evolutionary process. Two popular selection methods are the roulette wheel selection and the tour- nament selection. In the roulette wheel selection [105], an individual is randomly selected based on the probability determined by its ﬁtness value. Thus, an indi- CHAPTER 2. LITERATURE REVIEW 46 vidual having a higher ﬁtness value will have a higher chance of being selected. In the tournament selection method [105], a number of individuals are sampled randomly from the entire population into the tournament. The best individual from the tournament is then selected and is placed into the mating pool for mating by genetic operators. The mutation uses one tournament because of the requirement of a single parent individual. The Crossover uses two tournaments because of the requirement of two-parent individuals. Genetic operators There are three main genetic operators used in GP which are crossover, mutation, and reproduction. They together produce new programs from the selected individuals. The crossover is applied to two selected individuals known as the par- ents. A subtree is then randomly picked from each parent. Next, follow- ing the subtree crossover method, the selected subtrees are exchanged and two offsprings will be created as a result. A node is randomly picked from each parent who is identiﬁed as a crossover point. This is known as a sub- tree crossover, where two offsprings are produced by swapping the sub- tree from the two parents. Figure 2.4 represents an example of a subtree crossover. The mutation operator selects only one parent to create an offspring. The most common mutation approach is the subtree mutation, in which a random node is chosen as a mutation point from a selected parent and the subtree rooted from the node is removed. After that, a new subtree is grown from that mutation point. Figure 2.5 shows a subtree mutation example. In the case of reproduction, all the selected individuals are copied to the new population. This operator ensures the good individual will survive during the evolutionary process. 2.1. BACKGROUND 47 Parent A + * PR RT - / RO w PR Parent B + PR - + DD RT RT Child A + * PR RT - + DD RT PR Child B + PR - / RO w RT Figure 2.4: An example of a crossover operation in GP. Parent A Figure 2.4: An example of a crossover operation in GP. Parent + / PR RT - + DD RT RT Replacement * RO DD Child + / PR RT - + DD * RO DD RT Figure 2.5: An example of a mutation operation in GP. PR Parent Parent Figure 2.5: An example of a mutation operation in GP. Input : problem data Output: the best evolved program ∆∗ 1 randomly initialize the population P ←{∆1, . . . , ∆S}; 2 set ∆∗←null (best rule); 3 set fitness(∆∗) ←+∞; 4 generation ←0; 5 while generation < maxGeneration do 6 foreach ∆i ∈P do 7 evaluate and assign ﬁtness to ∆i; 8 end 9 foreach ∆i ∈P do 10 if fitness(∆i) < fitness(∆∗) then 11 ∆∗←∆i ; 12 fitness(∆∗) ←fitness(∆i); 13 end 14 end 15 set P new ←∅; 16 while \|P new\| < S do 17 ∆new ←apply genetic operators crossover, mutation and reproduction to selected programs from P; 18 P new ←P new ∪∆new ; 19 end 20 P ←P new ; 21 generation ←generation + 1; 22 end 23 return the best individual ∆∗; ∆new ←apply genetic operators crossover, mutation and reproduction to selected programs from P; 23 return the best individual ∆∗; 23 return the best individual ∆∗; Algorithm 1: Basic GP algorithm. Algorithm 1: Basic GP algorithm. Algorithm 1: Basic GP algorithm. Input : problem data 2.1.10 Basic GP algorithm Algorithm 1 shows a basic GP algorithm, which aims to ﬁnd the best per- forming individual program from the evolved programs. The algorithm starts with the initialization of a randomly generated population of a size speciﬁed, S. Then, evaluation, selection, and genetic operations (discussed above) are carried out. Every individual ∆i in the population of programs P is evaluated using a pre-determined ﬁtness function. Suppose the eval- uated program is better (has smaller ﬁtness value since we focus on min- imization problems in this research) than the best program ∆∗. In that case, it will be assigned to the best program found so far ∆∗and the best ﬁtness value fitness(∆∗) is also updated. Once all the best individuals in the population are evaluated, new individuals for the next generation are generated by applying mutation, crossover, and elitism to individuals selected using the selection mechanism. The algorithm repeats the evolu- tionary process until the maximum number of generations is attained and returns the best found rule ∆∗. CHAPTER 2. LITERATURE REVIEW 48 2.1.11 Local search Local search [38] is a meta-heuristic approach to improves the search effec- tiveness and quality of individuals. The local search method is typically 2.1. BACKGROUND 49 presented as an iterative process that starts with an initial solution gener- ated randomly or by some constructive heuristic. The method iteratively improves the current solution by moving to better neighboring solutions deﬁned by the neighborhood function. If the current solution s0 found a better neighborhood solution s1, it replaces the current solution s0 and the search is continued from s1. If no better solution is found, the algo- rithm terminates at a local minimum. The exploration strategies in local search determine the number of the neighborhood to be explored [149]. The most often used strategies for exploration in the literature are the ﬁrst- improvement and the best-improvement [79]. In the ﬁrst-improvement strat- egy, the ﬁrst best neighboring solution is selected. The best-improvement strategy explores the neighborhood entirely and returns the best neighbor- ing solution. The combinatorial optimization problem’s solution is represented by discrete structures like graphs, sequences, and genetic programs The neighborhood function introduces suitable perturbation operators tailored for the corresponding representations to generate ’neighboring’ solutions via swapping, moving, or replacement. In subsection 2.2.4, we will discuss the works related to local search methods applied to JSS problems. 2.1.12 Pareto local search (PLS) PLS can be considered as a direct extension of local search from single- objective problems to the multi-objective case. Generally, PLS maintains the Pareto archive of non-dominated solutions. Algorithm 2 illustrates the PLS framework. The input to PLS is an ini- tial set of solutions S0 that are mutually non-dominated. The solutions in S0 are initially marked as unexplored (line 2). PLS updates an archive of non dominated solutions S which is initially equal to S0. PLS starts with an initial Pareto archive and then solution s is chosen randomly among all unexplored ones in the archive. Next, the neighborhood of solution of s, 50 CHAPTER 2. LITERATURE REVIEW Algorithm 2: Pseudo-code for Pareto Local Seach Algorithm 2: Pseudo-code for Pareto Local Seach Algorithm 2: Pseudo-code for Pareto Local Seach g Input : An initial set of solutions S0 Output: S 1 explored(s)←FALSE ∀s ∈S0; 2 S←S0; 3 while S0! = ∅do 4 s ←select randomly a solution from S0; 5 foreach s′ ∈N(s) do 6 if s′ non-dominated with respect to the solutions in S then 7 explored(s′)←FALSE; 8 S ←Update(S, s′); 9 end 10 end 11 explored(s) ←TRUE; 12 S0 ←{s ∈S\|explored(s) = False} ; 13 end 14 return the best individual ∆∗; 1 explored(s)←FALSE ∀s ∈S0; 2 S←S0; 3 while S0! = ∅do 4 s ←select randomly a solution from S0; 5 foreach s′ ∈N(s) do 6 if s′ non-dominated with respect to the solutions in S then 7 explored(s′)←FALSE; 8 S ←Update(S, s′); 9 end 10 end 11 explored(s) ←TRUE; 12 S0 ←{s ∈S\|explored(s) = False} ; 13 end 14 return the best individual ∆∗; N(s) is explored. In PLS, the acceptance of a new neighborhood solution is based on the dominance relation. The procedure Update (line 7) adds an s′ ∈N(s) candidate solution to the archive if s′ is not dominated by any solution in the archive and removes all the archive solutions that become dominated by s′. Whenever a new neighbor solution s1 not dominated by the current solution is found, the Pareto archive needs to be updated. The s1 is accepted if it is not dominated by all solutions in the archive and all dominated solutions by s1 are removed from the archive. Furthermore, PLS has a natural stopping condition when the neighborhoods of all so- lutions in the archive have been explored (line 11). Liefooghe et al. [121] generalize PLS algorithms into several categories based on the different ways of deﬁning its basic components: (1) initial solutions strategy for neighborhood scanning, (2) exploring the neighbors of a non-dominated solution, fully vs. partially, (3) archiving method, e.g., bounded vs. un- bounded, and (4) stopping conditions. 2.1. BACKGROUND 51 Selection The selection strategy selects the solutions of the current archive whose neighborhood will be explored. Either all of the solutions are selected from the archive or only a subset of solutions may be selected. The subset of solutions is selected either uniformly at random or according to the ﬁtness of solutions. Archiving Two types of archiving strategies are considered. In the ﬁrst strategy, the archive can have a set of all non-dominated solutions. In the second strat- egy, the archive can have a subset of non-dominated solutions. The second strategy limits the archive’s size using some criteria such as diversity cri- teria (crowding distance). Exploration Explorations strategies control the size of the neighborhood for explo- ration. PLS can either explore the neighborhood entirely (fully) or only partially until their termination criterion is met. Termination criteria A common termination criterion is the one when all the solutions are fully explored. Otherwise, other popular termination criteria consider the time spent by the algorithm, the number of iterations, and a number of iter- ations without improvement [50]. Several PLS-based methods have also been developed for solving multi-objective JSS [14]. CHAPTER 2. LITERATURE REVIEW 52 2.2 Related work This section covers the related works that have been carried out in vari- ous ﬁelds of research that are relevant to the research goals and objectives. First, we cover the exact optimisation, heuristic, meta-heuristic, and the GP-HH approaches that have been proposed in scheduling research. Af- terward, we cover the multi-objective and many-objective JSS. Finally, we cover local search and PLS that have been applied to JSS problems. 2.2.1 JSS techniques This section covers the exact optimisation, heuristic and meta-heuristic ap- proaches to JSS techniques. Exact optimisation techniques Exact optimization techniques such as Jackson’s algorithm [83] can solve two-machine JSS problems to minimize makespan. Garey et al. [53] showed that static makespan minimization JSS problems with the num- ber of machines (M > 2) are NP-hard. This means that no algorithm can spend polynomial time in the worst-case scenario (M > 2) to minimize the makespan in a job shop. Researchers have suggested exhaustive search techniques to handle more difﬁcult JSS problems. Branch-and-bound [107] is one of the search techniques which have been used extensively in the lit- erature [5, 7, 19, 24, 25]. Carlier and Pinson [25] have proposed a notable branch-and-bound technique where they were able to ﬁnd an optimal so- lution for a JSS problem instances proposed by Muth and Thompson [139] with N = 10 jobs, M = 10 machines, and 10 operations per job. Branch- and-bound techniques are covered in detail in a survey paper by Potts and Strusevich [166]. Dynamic programming is an exact optimization technique that has been applied to static JSS problems [60, 112, 165]. Dynamic programming 2.2. RELATED WORK 53 approaches divide a JSS problem instance into sub-problems and attempt to solve each sub-problem separately. Lawler and Moore [112] proposed a technique of applying dynamic programming to a single machine schedul- ing problem with total weighted tardiness minimization. They suggested methods of extending it to parallel and two machine ﬂowshop problems. A more recent dynamic programming approach to JSS with makespan minimization has been proposed by Gromicho et al. [60] in 2012, where they adopt an approach proposed by Held and Karp [64] for the traveling salesman problem (TSP). Their analysis shows that the dynamic program- ming approach can generate optimal solutions for moderate benchmark instances where problem instances have up to N = 10 jobs and M = 5 machines. Although exact mathematical optimization techniques guarantee an optimal solution for a JSS problem instance, they generally take too long to generate optimal solutions for problem instances that have more than 10 to 15 jobs [60]. In addition, exact mathematical optimization techniques are not suitable for more complex problems because they generally take too long to generate optimal solutions such as many-objective JSS [5]. Heuristic techniques Heuristic approaches (e.g., shifting bottleneck heuristic) have been ap- plied to large static JSS problems, where problem instances have up to N = 8700 jobs and M = 9 machines [154]. They have also been ap- plied to dynamic JSS problems where it is not appropriate to use exact optimization techniques as a schedule may not be optimal as new un- foreseen events occur during processing [181]. Examples of heuristics are small and simple dispatching rules such as Shortest Processing Time (SPT), First In First Out (FIFO), and Earliest Due Date (EDD) [165]. The EDD rule prioritizes jobs based on the earliest due and time. Vepsalainen et al. [191] covers multiple composite dispatching rule heuristics for JSS problems with Total Weight Tardiness (TWT) minimization objective, in- CHAPTER 2. LITERATURE REVIEW 54 cluding the COVERT (cost over time) and ATC (apparent tardiness cost) priority-based dispatching rules. Gupta et al. [61] use dispatching rules in semiconductor manufacturing. They show that dispatching heuristics provides schedules quickly that are easy to understand, easy to apply, and require relatively short computation time. The primary disadvantage of manually dispatching rules is that these cannot hope for optimal solutions for all performance measures in the dynamic job shop [147]. Jayamohan et al. [90] compares various dispatching rules which have been proposed in the literature, such as rules proposed by Holthaus et al. [70] and also introduced several new dispatching rules. They per- formed a computational analysis of the rules on dynamic JSS problems for minimizing various aspects of ﬂowtime and tardiness. They further ex- tended the comparative study to incorporate weighted COVERT [90] and weighted ATC rules [191] and propose new dispatching rules. Nguyen et al. [145] proposed three different representations of dispatching rules and suggested that the representation that integrates the system and machine attributes can improve evolved rules’ quality. Holthaus et al. [70] proposed the 2PT + WINQ + NPT priority rule and variations of the rule for the dynamic JSS problems with ﬂowtime and tardiness minimization objectives. PT denotes the processing time for the job’s operation. WINQ denotes the work-in-next-queue, which is the sum of processing time of other jobs waiting at the next machine that the job will be processed on. NPT denotes the processing time of the job’s operation on the next machine, it needs to be processed or 0 otherwise. Heuristic techniques They show that 2PT + WINQ + NPT consistently outperformed the other benchmark dispatching rules, such as the RR rule (Raghu and Rajendran’s rule) [168] for minimizing mean tardiness for dynamic JSS problem instances. [63] considers multiple shop scenarios in a dynamic shop environment and apply Gaussian process regression to switch dispatching rules between EDD, MOD, 2PTPlusWinqPlusNPT [169]. A more complex example of a heuristic approach is the shifting bot- 2.2. RELATED WORK 55 tleneck (SB) heuristic which was proposed by Adams et al. [3] in 1988 for static JSS problem with makespan minimization objective. Adams et al. showed that the heuristic approach can ﬁnd high-quality optimal solu- tions for problem instances with up to N = 15 jobs and M = 15 machines. Meta-heuristic techniques The research on meta-heuristics for JSS problems has also been extensively studied in the literature where both local search-based techniques and population-based techniques have been effectively applied to different JSS problems. • Local search Local search based methods such as simulated anneal- ing [190] and Tabu search [162] have shown very promising results in JSS. These methods begin with a complete schedule and try to improve the schedule by developing efﬁcient and effective neigh- borhood structures, mainly based on the concept of critical paths and critical blocks. They also have diversifying strategies to escape from local optima. Many of the research work in [3, 7, 139] applied tabu search to various benchmark JSS problems and showed that Tabu search approaches can found optimal solutions for the static JSS problem instances. Like Tabu search, simulated annealing [190] approaches have been applied effectively to JSS problems in the lit- erature. Kreipl [106] proposed a large step random walk (LSRW) method. This is considered one of the best local search methods for dealing with JSS problems with tardiness objectives. The detailed information of local search in JSS can be found in subsection 2.2.4. • Local search Local search based methods such as simulated anneal- ing [190] and Tabu search [162] have shown very promising results in JSS. These methods begin with a complete schedule and try to improve the schedule by developing efﬁcient and effective neigh- borhood structures, mainly based on the concept of critical paths and critical blocks. They also have diversifying strategies to escape from local optima. Many of the research work in [3, 7, 139] applied tabu search to various benchmark JSS problems and showed that Tabu search approaches can found optimal solutions for the static JSS problem instances. Like Tabu search, simulated annealing [190] approaches have been applied effectively to JSS problems in the lit- erature. Kreipl [106] proposed a large step random walk (LSRW) method. This is considered one of the best local search methods for dealing with JSS problems with tardiness objectives. The detailed information of local search in JSS can be found in subsection 2.2.4. • Population-based techniques Cheng et al. [28] surveyed the GAs to JSS problems, showing that GAs consists of a population of individ- uals represented by a ﬁxed-length chromosome. One of the notable examples of a GA to JSS problems with tardiness objective is pro- posed by Zhou et al. [206]. Meta-heuristic techniques The GA algorithm is hybridized with • Population-based techniques Cheng et al. [28] surveyed the GAs to JSS problems, showing that GAs consists of a population of individ- uals represented by a ﬁxed-length chromosome. One of the notable examples of a GA to JSS problems with tardiness objective is pro- posed by Zhou et al. [206]. The GA algorithm is hybridized with CHAPTER 2. LITERATURE REVIEW 56 the existing heuristics for JSS. The experimental result showed that the hybrid GA outperforms pure GA approaches and signiﬁcantly reduces the computation time required to generate a solution. A hy- brid GA method and neighborhood local search for JSS to minimize makespan was proposed in [173]. In this method, the random keys were used for chromosome representation which represents the pri- orities of operations. Other meta-heuristic approaches which use EC techniques include ACO [34] and PSO [177, 196]. Sha and Hsu [177] used PSO and Tabu search (HPSO) for JSS problems. HPSO modiﬁed the particle posi- tion based on priority-based and a new preference list-based repre- sentation. The experimental results showed that HPSO has better solutions than other meta-heuristics methods. Xia and Wu [197] developed a hybridization of PSO and simulated annealing. In this work, a hybridization approach is applied to the multi-objective ﬂexible job-shop problems with makespan, total workload, and critical machine workload. The results showed that the hybridized approach effectively solved problems of up to 20 ma- chines and 20 jobs. Hyper-heuristics As described by Cowling et al.[37] as “heuristics to design heuristics”, hyper-heuristics have gained the attention of researchers whose goals are to design generic but effective heuristic methods to problems [20]. Burke et al. [20] describe one of the main motivations for developing hyper-heuristic approaches to handle the ”challenge of automating the de- sign and tuning of heuristic methods to solve hard computational search problems.” Hyper-heuristics are designed to work on low-level heuristic components with the help of suitable operators that combine low-level heuristics and a method of evaluating the performance of heuristics. The 2.2. RELATED WORK 57 Figure 2.6: Overview of a tree-based GP-HH applied to JSS. Figure 2.6: Overview of a tree-based GP-HH applied to JSS. hyper-heuristic methods hence search in the heuristic space [21]. This enables us to easily distinguish hyper-heuristics from meta-heuristics as meta-heuristics search in the solution space. One popular hyper-heuristic approach is GP-HH [20]. Figure 2.6 shows a high-level overview of a tree-based GP being applied to a JSS problem. Figure 2.6 shows a tree-based GP-HH takes in the base heuristics {PT (processing time), W (weight), and DD (due date)} and the operators {+, −, ×} to initialize the individuals in the population. The GP system then passes the heuristic into the JSS problem domain and gets a goodness measure back in the individual’s ﬁtness. The ﬁnal output is a heuristic which represents a priority-based dispatching rule PT + DD −PT. 2.2.2 GP-HH for JSS The main aim of hyper-heuristic is to automate designing and selecting the heuristics to solve hard and complex problems such as JSS problems [20]. GP is able to evolve complex programs or rules. It becomes a suitable paradigm for learning heuristics known as GP-HH [20]. CHAPTER 2. LITERATURE REVIEW 58 Representation of GP as a Hyper-Heuristic In GP, the most prominent representation and traditionally represented as tree structures. Other program structures have been investigated in the literature, such as linear sequences of instructions or grammar. The individual of tree-based GP is represented as arithmetic function trees [105, 145, 159] which are called priority dispatching rules in the scheduling environment. Hunt et al. [74] used arithmetic representation to evolve dispatching rules to minimize tardiness objectives in the static JSS problem. The evolved rules from GP were more effective than manual design dispatching rules. Nguyen et al. [145] investigated three GP representations for JSS. The ﬁrst representation (R1) evolved decision trees, the second representa- tion (R2) used arithmetic representation, and the third representation (R3) showed a mixed representation, which combined (R1 and R2). Experimen- tal results showed that the evolved rules outperformed the existing rules on the static JSS problems [145]. GP-based hyper-heuristics for scheduling The ﬁrst comprehensive review on hyper-heuristics is shown in [17]. This study focused on the key design choices and critical issues involved in the process of developing scheduling heuristics. Hunt et al. [75] proposed a GP-HH approach to evolve dispatching rules for both the static and dy- namic two-machine job shop environments. This was considered the ﬁrst study that used GP to evolve dispatching rules for reducing makespan. Park et al. [158] improved the robustness of evolving ensembles of dis- patching rules from a single population of GP individuals. They investi- gate four ensemble approaches based on majority voting, linear combina- tion, weighted linear combination and weighted majority voting. Through experimentation, they found that linear combination schemes are better than other ensemble techniques. Hunt et al. [73] revealed that local infor- 2.2. RELATED WORK 59 mation is one of the drawbacks in dispatching rules. Therefore, they used lookahead terminals for GP to evolve “less myopic”rules for dynamic job shop scheduling problems. Nguyen et al. [145] proposed a tree-based GP representation for evolv- ing rules for static JSS problems. They proposed different representations of dispatching rules and suggested that the representation that integrates the system and machine attributes can improve the quality of the evolved rules. Nguyen et al.[141] also developed methods for evolving iterative dispatching rules using GP. Tay and Ho [185] used GP-HH for multi-objective FJSS problems. They tried to minimize makespan, mean tardiness, and mean ﬂowtime. They used an aggregation method to solve the multi-objective FJSS problems. Nguyen et al.[147] developed a GP-HH method for multi-objective JSS problems that optimize ﬁve conﬂicting objectives simultaneously. A literature survey shows that dispatching rules are prominently used for solving JSS problems with a single objective or multiple objectives. There are no many works in the literature to deal with many-objective JSS problems. However, the previous section mentioned that Pareto-based EMO algorithms may not effectively handle many objectives due to a high proportion of non-dominated solutions. Therefore, in this research, we will use GP-HH with the selection scheme of NSGA-III for evolving the dispatching rules more effectively. 2.2.3 Multi-objective and many-objective JSS In the literature, many research works have been used EC algorithms for JSS but these algorithms mainly focus on optimizing a single objective [67, 159]. For example, GAs have been utilized for minimizing makespan while scheduling identical parallel machines [99]. Park et al. have also considered using cooperative evolutionary technologies to minimize the TWT in dynamic JSS problems [159]. 60 CHAPTER 2. LITERATURE REVIEW It is becoming clear that JSS essentially has multiple (or many) different objectives. In general, there are two alternative approaches for handling multiple objectives in a job shop, i.e., the aggregation method and the Pareto- dominance method [96]. In a typical aggregation method, multiple optimiza- tion objectives have to be aggregated together to form a scalar-valued ﬁt- ness function through a weighted sum [116]. Obviously, this method’s usefulness is restricted to the situation when the preferences over differ- ent objectives can be quantiﬁed before applying any EC techniques. On the other hand, without using any aggregation functions, the Pareto- dom- inance concept can be exploited to deﬁne the optimization criteria for a guided search of Pareto-optimal schedules [116]. Based on this idea, many EC algorithms have been proposed with the aim of evolving the Pareto- front [43, 44, 211, 27, 76]. In the literature, the Pareto-dominance method has attracted substan- tial research attention. Prominent examples include the NSGA-II and SPEA2. Speciﬁc techniques have also been successfully developed for multi-objective JSS. For instance, Murata et al. [137] proposed a multi- objective GA for ﬂowshop scheduling problems. Their research speciﬁ- cally considered problem instances with concave Pareto-front and at most three optimization objectives (i.e., the makespan, total tardiness, and to- tal ﬂowtime). A trade-off between the makes and the availability of ma- chines identiﬁes in bi-objective JSS problems [200]. Nguyen et al. [147, 144] used several multi-objective GP approaches in order to evolve schedul- ing rules consisting of a dispatching rule and due-date assignment rule. Those approaches evolved two expression trees, one of which would be used for due-date assignment to jobs, while the other would be used as a standard dispatching rule. The results showed that the evolved rules outperformed various combinations of existing scheduling rules from the literature. Evolving dispatching rules for multi-objective criteria were also analyzed by Nguyen et al. [146]. In [73], a scheduling problem consisting of two and three scheduling criteria was optimized by the use of two pro- 2.2. 2.2.3 Multi-objective and many-objective JSS RELATED WORK 61 posed heuristics and a genetic algorithm. Moreover, A short overview of some other multi-objective problems in the unrelated machine’s environment can be found in [163]. Moslehi and Mahnam [128] present an approach based on a hybridization of the parti- cle swarm and local search algorithm to solve the multi-objective ﬂexible JSS problem. Zhang et al. [207] have formulated the textile dyeing pro- cess scheduling problem as a bi-objective optimization model, in which one objective is related to tardiness cost while the other objective reﬂects the level of pollutant emission. And they proposed a multi-objective par- ticle swarm optimization algorithm enhanced by problem-speciﬁc local search techniques (MO-PSO-L) to seek high-quality non-dominated solu- tions. Luo et al. [126] proposed a distributed ﬂexible JSS Problem with transfers (DFJSPT), in which operations of a job can be processed in differ- ent factories. In order to expand the search space and accelerate the con- vergence speed of the solution, an efﬁcient memetic aAlgorithm (EMA) is proposed to solve the DFJSPT with the objectives of minimizing the makespan, maximum workload, and total energy consumption of facto- ries. Zhao et al. [209] proposed an improved multi-objective evolution- ary algorithm, which is based on decomposition (IMOEA/D) for multi- objective JSS problem which minimized three objectives – the maximum completion time (makespan), the total ﬂow time and the tardiness time are considered simultaneously. Several prior rules are presented in the proposed algorithm to construct the initial population with a high-quality level. Karunakaran et al. [98] proposed the sampling heuristic for GP-HH and presented the algorithm based on the island model approach. In addition to the research works mentioned above, some researchers have started considering JSS problems with more than three objectives. For example, Nguyen et al. [147] have considered designing dispatch- ing rules for general JSS problems with up to ﬁve different objectives. In [137], four different objectives have been considered for evolving op- timal schedules. Fowler et al. [51] considered the problem for scheduling CHAPTER 2. LITERATURE REVIEW 62 a printed wiring board manufacturer’s drilling operation subject to ﬁve optimization objectives. They have used several approaches to solve the given problem. Kolahan and Kayvanfar [103] used the simulated anneal- ing approach to solve a scheduling problem consisting of the makespan, earliness, and tardiness objectives. Wang et al. [192] proposed an algo- rithm for many-objective ﬂexible job shop scheduling problems. 2.2.3 Multi-objective and many-objective JSS In this algorithm, the mathematical model was established considering six ob- jectives: makespan, workload balance, mean of earliness and tardiness, cost, quality, and energy consumption simultaneously. XU et al. [198] proposed an algorithm for many-objective ﬂowshop scheduling problems based on fuzzy sets’ relative entropy. This algorithm built a mathematical model of the many-objective ﬂow-shop scheduling problems, including four objectives, makespan, tardiness, total inventory cost and total tardi- ness cost. A many-objective permutation ﬂow shop scheduling problems model with four objectives, namely, the makespan, total tardiness, inven- tory holding cost, and energy consumption cost, was established in [210]. Gong et al. [56] considered many-objective ﬂexible job shop scheduling problem (MaOFJSP) with ﬁve objectives under dynamic electricity pricing and applied non-dominated genetic algorithm-III to solve it. Our survey found that there has been relatively little attention to solv- ing many-objective JSS problems. There are a few studies found in the lit- erature of many-objective JSS but these studies used conventional MOEAs. MaOPs pose a number of challenges to Pareto-based MOPs [43, 31]. First, the proportion of non-dominated solutions in a population rises rapidly with the number of objectives. This issue also becomes a part of JSS prob- lems whenever optimization of more than three objectives is attempted. Second, a commonly used diversity-preservation operator such as crowd- ing distance [44], clustering [211] cannot strengthen the Pareto front’s se- lection pressure. These operators also become computationally expensive [84]. 2.2. RELATED WORK 63 2.2.4 Local search for JSS Local search is a popular method for dealing with NP-hard problems — the ﬁrst application of a local search that was studied in [38]. There are several reasons for interest in local search algorithms. One important as- pect is that local search algorithms are more easily understandable and implementable than exact algorithms. In practice, local search has been shown to be very effective at solving complicated problems [1]. The local search algorithm is widely used for solving JSS [166]. It was ﬁrst introduced for JSS in 1980. A year later, Talliard et al. [183] intro- duced the tabu search for JSS problems. One of the studies [189] discussed the local search method to emphasize deterministic and randomized JSS problems. They also discussed existing neighborhood techniques for local search methods. Hunt et al. [75] applied a local search in the ﬁtness eval- uation process to improve the performance of dispatching rules. Further- more, JSS problems with alternative sequence and sequence-dependent setup are addressed in [33]. They developed mixed-integer program mod- els with local search to reduce computational time. The research study in [58] solved stochastic JSS problems using several neighborhood functions from the literature. A few pieces of literature show that local search is also used with GP to improve the quality of evolved dispatching rules [148, 173]. Nguyen et al. [148] improved the quality of evolved dispatching rules for dynamic JSS through Iterated Local Search (ILS) [124]. The key idea of ILS is to iteratively apply local search heuristics to solutions obtained by perturba- tions of previously visited locally optimal solutions [10]. The study in [173] combined ILS, and GP. The goal of GP in these algorithms was to evolve perturbation operators based on a range of low-level operators and rules. 2.3. SUMMARY 65 The research in [89] applied PLS to the many-objective combinatorial optimization problems and showed high effectiveness of the proposed many-objective Pareto local search algorithm. In this algorithm, they used three new mechanisms: (1) the efﬁcient update of large Pareto archives with ND-Tree data structure, (2) a new mechanism for the selection of the promising solutions for the neighborhood exploration, and (3) partial ex- ploration of the neighborhoods. Furthermore, Seada et al. [175] combined NSGA-III with local search and Karush-Kuhn-Tucker Proximity Measure (KKTPM). The local search in this algorithm used two distinct operators. One operator enhances overall population diversity and the other pro- motes convergence.Meanwhile, KKTPM speeds up the convergence and guides the second local search operator. 2.2.5 PLS for JSS PLS is one of the heuristic algorithms for handling combinatorial opti- mization problems and maintaining the Pareto archive [155]. The goal of CHAPTER 2. LITERATURE REVIEW 64 PLS is to approximate a high-quality Pareto-optimal set. It can be used as a standalone algorithm [156] or hybridized with the EMO algorithm [127]. Hybridization of local search with EMO is often called a memetic algorithm [81]. This hybridization improves the searchability of the pure EMO algorithms [88]. The hybrid algorithms have been found in the literature for solving combinatorial optimization problems effectively [14, 49, 82, 88, 81, 137]. The evaluation mechanisms in the local search of memetic EMO algorithms are performed by either weighted scalar ﬁt- ness function [81] or Pareto ranking [102]. The effectiveness of PLS as a component of the hybrid algorithms may be explained by the fact that it has different characteristics than most other multi-objective algorithms. As noticed by Lara et al. [108], each multi-objective metaheuristic should search both towards and along the Pareto-front. Several methods of PLS have been proposed in the multi-objective JSS problems. Murata et al. [137] proposed a multi-objective GA to minimize the makespan, total tardiness, and ﬂowtime. Ishibuchi et al. [81] proposed a genetic local search algorithm. This algorithm uses a weighted sum of multiple objectives as a ﬁtness function to guide parents’ selection for gen- erating offspring solutions through crossover and mutation operations. A local search is then employed to each solution generated by genetic opera- tors to improve its ﬁtness value further. This algorithm was tested on two instances of ﬂowshop with two pairs of objectives (makespan and total tardiness) and (makespan and maximum tardiness). Moslehi et al. [136] proposed a hybridization of a particle swarm algorithm and local search methods for multi-objective JSS problems. In this algorithm, local search is employed to enhance convergence speed. The research study in [4] de- ﬁned a variable neighborhood search (VNS) for a multi-objective dynamic JSS problem. In this approach, VNS is hybridized with an artiﬁcial neural network (ANN), where ANN is used to optimize the algorithmic param- eters for VNS. They showed that VNS can outperform common human- made dispatching rules such as SPT and FIFO. 2.3. SUMMARY 2.3 Summary This chapter has discussed the key concepts of GP, JSS, hyper-heuristics, multi-objective optimization, many-objective optimization, model-based EMO, Gaussian process, local search, and PLS. GP is one of the most pop- ular hyper-heuristic because of its ﬂexibility, which automatically evolves dispatching rules in single and multi-objective JSS. However, the research in the direction of many-objective JSS is relatively new and many aspects need to be explored to enhance the quality of GP-HH methods and cope with practical requirements. There are some identiﬁed limitations of the existing body of work in the literature, which motivate this thesis’s work. These limitations are: • There are a number of different approaches that have been proposed in the literature for various JSS problems. Most of the works in the literature on JSS do not focus on many-objective JSS. Our survey found that there has been relatively little attention to solving many- objective JSS problems. Heuristic approaches are prominently used • There are a number of different approaches that have been proposed in the literature for various JSS problems. Most of the works in the literature on JSS do not focus on many-objective JSS. Our survey found that there has been relatively little attention to solving many- objective JSS problems. Heuristic approaches are prominently used CHAPTER 2. LITERATURE REVIEW 66 to solve JSS problems because of its advantages over other meth- ods. Literature shows that heuristic approaches are better for solving multi-objective JSS problems. Moreover, GP-HH approaches have been successfully used to design effective dispatching rules automat- ically. However, GP-HH approaches are relatively new for many- objective JSS and require further investigations. There are a few stud- ies found in the literature of many-objective JSS but these studies used conventional MOEAs. These MOEAs fail to work properly for MaOPs. Due to the issue of using MOEAs in solving many-objective JSS, new GP-HH approaches are required that can potentially work with many-objective JSS. • Literature survey shows that decomposition-based approaches are good for diversity maintenance [43]. The decomposition-based ap- proaches use uniformly distributed reference points and perform better on problems which have uniformly distributed Pareto-front such as the DTLZ1 problem. However, they face difﬁculty when it is applied to non-uniform and irregular Pareto-front problems such as the DTLZ7 and JSS problems. Deb and Jain [86] also notice this lim- itation. 2.3 Summary Deb and Jain have witnessed several many-objective prob- lems where reference points can never be associated with any good solutions, while others are associated with more than one candidate solutions. The distribution of the Pareto-front of JSS problems in the objective space is also non-uniform. Algorithms for many-objective JSS are required to consider this limitation and generate reference points according to the distribution of the Pareto-front. • Literature survey shows that decomposition-based approaches are good for diversity maintenance [43]. The decomposition-based ap- proaches use uniformly distributed reference points and perform better on problems which have uniformly distributed Pareto-front such as the DTLZ1 problem. However, they face difﬁculty when it is applied to non-uniform and irregular Pareto-front problems such as the DTLZ7 and JSS problems. Deb and Jain [86] also notice this lim- itation. Deb and Jain have witnessed several many-objective prob- lems where reference points can never be associated with any good solutions, while others are associated with more than one candidate solutions. The distribution of the Pareto-front of JSS problems in the objective space is also non-uniform. Algorithms for many-objective JSS are required to consider this limitation and generate reference points according to the distribution of the Pareto-front. • Literature survey shows that local search is a widely used technique for the classical combinatorial optimization problem. There are sev- eral existing studies using PLS (extension of the local search) for multi-objective JSS problems. These algorithms hybridized the lo- cal search with the global search and used to improve the quality of 2.3. SUMMARY 67 the dispatching rules. This hybridization is called the memetic al- gorithm. PLS enhances multi-objective GA performance; however, no works in the literature hybridize GP with PLS techniques. In this thesis, we enhance the quality of evolved dispatching rules through PLS in many-objective JSS. The following contribution chapters in this thesis will propose new ap- proaches that mainly aim to address these limitations. CHAPTER 2. LITERATURE REVIEW CHAPTER 2. LITERATURE REVIEW CHAPTER 2. LITERATURE REVIEW 68 Chapter 3 Experimental Methodology This chapter presents the experimental methodology of this thesis. First, it discusses the benchmark problems for JSS that will be used throughout the thesis. Next, it describes the GP functions and terminals for JSS. Then it discusses the performance measures used for evaluating many-objective optimization algorithms. 3.1 Benchmark problems for JSS There are four widely used static job shop benchmark sets in the literature. These are Lawrence static job shop benchmark set (LA 01-40 Lawrence) [113], Applegate and Cook benchmark set (ORB 01-10) [8], Taillard static job shop benchmark set (TA 01-80) [182], and Demirkol et al. benchmark set (DMU 01-80) [46]. Details of the benchmark sets can be seen in Table 3.1. Table 3.1: Static JSS data sets. Data set Notation # of instances Size ( #J #M) LA LA01-LA40 40 10 5 to 15 15 ORB ORB01-ORB10 10 10 10 TA TA01-TA80 80 15 15 to 100 20 DMU DMU01-DMU80 80 20 15 to 50 20 Table 3.1: Static JSS data sets. 69 70 CHAPTER 3. EXPERIMENTAL METHODOLOGY The goal of this thesis is to evolve effective dispatching rules (prior- ity functions) for many-objective JSS. We used static JSS environments be- cause many-objective research is comparatively newer than single objec- tive JSS and multi-objective JSS. Further, dynamic JSS environments are more complex than static JSS environments. The TA is the most used benchmark problem for solving static JSS prob- lems in the literature [131, 159]. The benchmark targets permutation ﬂow- shop, ﬂowshop, open shop, and JSS problems. It contains job shop in- stances of sizes up to 100 jobs on 20 machines, resulting in 2000 job opera- tions. This benchmark problem is still using to use to solve multi-objective optimization problems [97]. Therefore, the TA has been selected as the dataset for experiments in this thesis. There are 80 indexed problem instances in the TA set (i.e., each problem instance has a different ID from 1 to 80). These can be further divided into eight groups (denoted as TA-1, TA-1, ..., TA-8). The problem instances in the same group have the same number of jobs and machines but the processing time matrices are generated by using different random seeds. In the experiments, the total 80 instances are divided into the training set and the test set, each consisting of 40 instances. The training set consists of all the instances with odd IDs and the test set contains all the instances with even IDs. The GP uses the training instances in order to evolve dispatching rules for some given objectives. Once obtained, the effectiveness of these evolved dispatching rules will be evaluated by using the test instances. 3.2 GP terminals and function set The terminal set is described in Table 3.2. The non-terminal set is set to {+, −, ∗, /, min, max, ifthe}, where ”+”, ”−” and “∗” are basic arithmetic operators. The function ”/” is the protected division operator which re- turns one if the denominator is zero. The functions ”min” and ”max” take two arguments and return either the smallest and the largest value, re- spectively. The function ifthe takes three arguments a, b and c, and returns b if a > 0 and c otherwise. These non-terminals [141] used by existing GP- HH methods in the literature and also used in manually designed rules [20, 70]. These rules determine the priority of jobs by computing the ex- pression. For example, the simple shortest processing time (SPT) rule [176] and the apparent tardiness cost (ATC) rule [70]. 3.1 Benchmark problems for JSS The release times of all jobs in all problem instances are safely set to zero. Since there is no due date information included in the original dataset, the due date assignment rule [11] will assign the due dates of each insistence of jobs. That is, dd(ji) = λ × m X k=1 pk i . where dd(ji) stands for the due date of the job ji, and λ is used to indicate where dd(ji) stands for the due date of the job ji, and λ is used to indicate GP TERMINALS AND FUNCTION SET 71 3.2. the tightness of due dates. We chose λ value is 1.3 for all instances in the training set and test set in our experiments because it is most commonly used value in the JSS literature. The job weights are set according to the 2:6:2 rule [165]. That is, the weight is set to 4, 2, and 1 with the probabilities of 20%, 60%, and 20%, respectively. These weighted probabilities in this research are inspired by Pinedo and Singer [165]. Their study showed that the ﬁrst 20% of jobs are assigned a weight of 4, the weight of 2 is assigned to the next 60% of jobs, and weight of 1 is assigned the last 20% of jobs. Four objectives are considered in the experiment. These are the mean ﬂowtime (mF) (see equation (2.1) of Chapter 2) , maximal ﬂowtime (maxF) (see equation (2.2) of Chapter 2), mean weighted tardiness (mWT) (see equation (2.6) of Chapter 2) and maximal weighted tardiness (maxWT) (see equation (2.7) of Chapter 2). Table 3.2: Terminal set of GP for JSS. Table 3.2: Terminal set of GP for JSS. Table 3.2: Terminal set of GP for JSS. Attribute Symbol Job Properties Processing time of the operation PT Processing time of the next operation NOPT Ready time of the next machine NMRT Work Remaining WKR Work in the next queue WINQ Number of operations in the next queue NOINQ Flow due date FDD Due Date DD Weight W Machine Properties Number of operations in the queue NOIQ Work in the queue WIQ Ready time of the machine MRT this thesis, we choose the HyperVolume [212] and Inverted Generational Dis- tance [206]. They are the two main performance measures in the literature. Theoretically, a set of trade-off dispatching rules with better performance should have a larger HV value and a smaller IGD value. All the objectives have been normalized into the range between 0 and 1 before calculating the above performance measures. 3.3 Performance measures There are a variety of performance measures proposed for evaluating multi-objective optimization algorithms from different perspectives. In 72 CHAPTER 3. EXPERIMENTAL METHODOLOGY CHAPTER 3. EXPERIMENTAL METHODOLOGY HyperVolume (HV) Given a nadir points z∗and a set of non-dominated solutions {y1, . . . , yn} in the objective space. HV indicates the area covered by {y1, . . . , yn} with respect to z∗. The selection of the nadir points is an important issue. It is better for z∗to be slightly larger than the maximum value of each objective to em- phasize the balance between convergence and diversity. In the thesis, all algorithms have normalized objective values. So that the ideal point and the nadir point are (0, 0) and (1, 1) respectively in the objective space of a two-objective problem. After this normalization, the nadir points (z∗, z∗) with z∗=1 + 1 (n−1) (where n is the size of solutions) may be a good choice 3.3. PERFORMANCE MEASURES 73 when the Pareto front is nonlinear or irregular [78]. when the Pareto front is nonlinear or irregular [78]. For HV, the nadir point is set to (1, 1, 1, 1) in this thesis since all the objectives are to be minimized. Inverted generational distance (IGD) The IGD value is based on the true Pareto-front {y∗ 1, . . . , y∗ u}. Speciﬁcally, it is deﬁned as the average distance of each Pareto-optimal solution y∗ i to its closest point in {y1, . . . , ys}. That is, IGD = 1 u u X i=1 min y∈{y1,...,ys}{dist(y∗ i , y)}. (3.1) (3.1) Since the true Pareto-front is unknown in JSS, an approximate Pareto- front will be adopted for IGD by selecting the non-dominated solutions among the ﬁnal solutions acquired from all the runs of all the tested algo- rithms. CHAPTER 3. EXPERIMENTAL METHODOLOGY CHAPTER 3. EXPERIMENTAL METHODOLOGY 74 4.1 Introduction A study of the literature shows that most of the past research on JSS fo- cuses on optimizing job shop scheduling (JSS)-relevant objectives [67]. For example, Park et al. have considered cooperative evolutionary technolo- gies to minimize the total weighted tardiness (TWT) in dynamic JSS prob- lems [159]. It is technically simple to consider only a single optimization objective for JSS. However, it is now widely evidenced in the literature that JSS by nature presents conﬂicting objectives, including the makespan, mean ﬂowtime, maximum tardiness, maximum lateness, total workload, and proportion of tardy jobs [137, 122]. Over the last few years, JSS has been frequently considered as a multi- objective optimization problem with several potentially conﬂicting objec- tives [201]. The research study in [185] presented the ﬁrst work on genetic programming (GP) that focuses on multi-objective JSS problems. In this study, an aggregation method was used which aggregated the multiple optimization objectives together and formed a scalar-valued ﬁtness func- tion through a weighted sum [79]. However, this method is restricted to the situation when the preferences over different objectives can be quanti- ﬁed before applying any evolutionary computation (EC) techniques. The 75 CHAPTER 4. MANY-OBJECTIVE GP-HH FOR JSS 76 Pareto-dominance approach can be exploited to deﬁne the optimization criteria for a guided search of Pareto-optimal schedules. In the literature, the Pareto-dominance approach has attracted substantial research atten- tion. Prominent examples include the non-dominated sorting genetic al- gorithm II (NSGA-II) [44] and the strength Pareto evolutionary algorithm 2 (SPEA2) [211]. Despite the rapid development of EC techniques for multi-objective JSS, existing studies of JSS problems that have more than three objectives (so-called many-objective ) are still very limited. There are only a few stud- ies on many-objective JSS [53, 147]. In [147], Nguyen et al. considered ﬁve conﬂicting objectives to design dispatching rules for JSS problems. These studies used multi-objective evolutionary algorithms (MOEAs) such as NSGA-II [44] and SPEA2 [211] for solving the many-objective JSS Prob- lems. However, MOEAs noticeably deteriorate their search ability when more than three objectives are involved [31, 43]. One major reason for this is that the proportion of non-dominated solutions in a population rises rapidly with the number of objectives. This issue also becomes a part of JSS problems, whenever optimization of more than three objectives is at- tempted. 4.1 Introduction Therefore, we aim to explicitly investigate JSS problems with many conﬂicting objectives and consider the challenges of many-objective optimization. As we discussed in Chapter 1, rather than directly evolving Pareto- optimal schedules, we evolve dispatching rules. We hope that the evolved dispatching rules could achieve a good balance over many conﬂicting ob- jectives. This is essentially a machine learning approach where rules can be evolved based on a training set of problem instances and then further evaluated on a different set of testing instances. In the literature, the most effective technique is to evolve dispatching rules using (GP-HH ) [143, 148] which is also be adopted in this study. This chapter aims to use GP-HH for evolving dispatching rules for many-objective JSS problems, which can provide potential trade-offs among different objectives. The sub-objectives of this chap- 4.1. INTRODUCTION 77 ter are: ter are: 1. To develop a many-objective GP-HH method to evolve dispatching rules for JSS, 1. To develop a many-objective GP-HH method to evolve dispatching rules for JSS, 2. To investigate the relationship among JSS objectives, 3. To perform detailed comparisons with other multi-objective evolu- tionary algorithms (i.e. NSGA-II [44] and SPEA2 [211]) for verifying the efﬁcacy of the proposed algorithm in solving many-objective JSS. The ﬁrst sub-objective provides the algorithm for evolving a set of trade-off dispatching rules in many-objective JSS. A few interesting many- objective algorithms have already been developed in recent years [116, 43]. Among all the different technologies, we have particular interests in the decomposition-based approach (see in Section 2.1.5 of Chapter 2). NSGA- III is a state-of-the-art algorithm that can be seen as a decomposition-based approach and has shown leading performance on many benchmark prob- lems [43]. In this thesis, we used decomposition-based approaches specif- ically for obtaining Pareto-optimal solutions for many-objective JSS prob- lems. The second sub-objective will give an answer to the question of how different scheduling objectives inﬂuence each other when they are optimized together. Finally, the last sub-objective will examine multi- objective optimization algorithms such as NSGA-II and SPEA2 are compe- tent at tackling many-objective JSS problems. The ﬁrst research objective’s main goal is to use GP-HH for JSS problems to ﬁnd the Pareto fronts of non-dominated dispatching rules to deal with many conﬂicting objectives. NSGA-II and SPEA2 were selected since they represent one of the most popular Pareto dominance based multi-objective algorithm. 4.1 Introduction Also, the ba- sic framework of the NSGA-III remains similar to the original NSGA-II algorithm. Further, NSGA-II and SPEA2 used in [147] and we will explore the performance of the proposed many-objective algorithm as compared to the NSGA-II and SPEA2 algorithms in solving many-objective JSS. In CHAPTER 4. MANY-OBJECTIVE GP-HH FOR JSS 78 Section 4.2, a problem description is presented. Section 4.3 describes the proposed algorithm for many-objective JSS. Section 4.4 gives a description of the experimental studies. Section 4.5 describes the achieved results and contains the discussion. Finally, Section 4.6 concludes this chapter. 4.2 Problem description Based on the JSS problem described in Subsection 2.1.2 of Chapter 2, our goal is to evolve useful dispatching rules that can build the com- plete schedules incrementally. The quality of the schedules created by using these rules will then be evaluated with respect to D objectives f = {(f1, f2, . . . , fd)}. Without losing generality, we assume that all the dimensions of f are minimized. Here, we consider the case with D ≥4, i.e., there are four or more objectives ( many-objective JSS). Given two schedules ∆1 and ∆2, it is said that ∆1 dominates ∆2 if and only if ∀i, 1 ≤i ≤D, fi(∆1) ≤fi(∆2) (4.1) ∀i, 1 ≤i ≤D, fi(∆1) ≤fi(∆2) (4.1) ∀i, 1 ≤i ≤D, fi(∆1) ≤fi(∆2) (4.1) and ∃i, fi(∆1) < fi(∆2). (4.2) (4.2) Consequently, if P1 is the dispatching rule that produces schedule ∆1 and P2 is the dispatching rule that produces schedule ∆2, then we can also say that, for the given problem instance I, P1 dominates P2. Based on the above explanations, it should now be clear that our goal is to evolve a collection of Pareto-optimal dispatching rules, jointly known as the Pareto- front. Each dispatching rule in the Pareto-front is non-dominated by any other rules evolved by using our proposed algorithm. In the job shop, when a new job arrives at an idle machine that job will be processed immediately. The information about that job will be ex- tracted from the terminals in Table 3.2 of Chapter 3. Then, the dispatching rule will be evaluated and after the evaluation, priority will be assigned 79 4.3. MANY-OBJECTIVE-GP-HH FOR JSS Figure 4.1: Illustration of a dispatching rule in JSS. Figure 4.1: Illustration of a dispatching rule in JSS. to the considered job. This procedure will be applied until priorities are assigned to all waiting jobs on the machine and the job in the queue with the highest priority will be processed next. This can be seen in Figure 4.1. 4.3.1 Representation of rules As with previous studies of GP for JSS problems [143, 147], dispatching rules are also represented by GP trees [105] in this thesis. Basically, dis- patching determines the priorities of jobs waiting in the queue. Figure 4.2 shows the GP tree representation of the 2PT+WINQ+NPT rule [70]. Figure 4.2: The GP tree representation of the 2PT+WINQ+NPT rule. Figure 4.2: The GP tree representation of the 2PT+WINQ+NPT rule. In Figure 4.2, the terminals in that tree are {2, PT, WINQ, NPT}, and the non-terminals are {+, ∗}. Proper terminal and non-terminal sets are cru- cial for constructing promising and concise search space for GP. GP-based approaches in this thesis use a mixture of terminal sets. These terminals are used by existing GP-HH approaches in the literature and evolved qual- ity rules [141, 143, 147]. These terminals range is from common attributes (e.g., operation processing time PT) to more complex terminals (e.g., pro- cessing time of the next operation NPT). WINQ and NPT reﬂect the status of the current machines. Whereas WINQ denotes the sum of imminent process times of all jobs waiting at the queue of the next operation of job i. The commonly considered job shop attributes are included in this thesis, which has been summarized in Table 3.2 of Chapter 3. 4.3 Many-objective-GP-HH for JSS This section shows how the proposed Many-Objective-GP-HH method is used to solve many-objective JSS problems. It is common to use hyper- heuristics for evolving dispatching rules for JSS and this has achieved great success [68, 87]. Section 4.3.1 will show how GP represents dispatch- ing rules. The newly proposed algorithm which is named GP-NSGA-III is given in Algorithm 3. GP-NSGA-III combines the initialization, evalua- tion, and evolutionary operators of GP and the selection scheme of NSGA- III. Then, the proposed Many-Objective-GP-HH algorithm is presented. CHAPTER 4. MANY-OBJECTIVE GP-HH FOR JSS 80 4.3.2 General framework of GP-NSGA-III In this thesis, we evolve dispatching rules to minimize four popular JSS objectives which include the mean ﬂowtime (see equation (2.1) of Chapter 4.3. MANY-OBJECTIVE-GP-HH FOR JSS 81 Algorithm 3: The framework of GP-NSGA-III. Input : A training set Itrain Output: A set of non-dominated rules P ∗ 1 Initialize and evaluate the population P 0 of rules by the ramped-half-and-half method; 2 Calculate set of reference points Z; 3 Set Pg ←P0 and generation ←0; 4 while generation < maxGeneration do 5 Generate the offspring population Qg using the crossover, mutation and reproduction of GP; 6 foreach Q ∈Qg do 7 Evaluate rule Q 8 end 9 Rg ←P g ∪Qg; 10 (F1, F2, · · · ) ←Non-dominated-sort(Rg); 11 Form the new population P g+1 from (F1, F2, · · · ) by the NSGA-III selection; 12 generation ←generation + 1; 13 end 14 return The non-dominated individuals P ∗⊆Pgmax; Algorithm 3: The framework of GP-NSGA-III. Input : A training set Itrain Output: A set of non-dominated rules P ∗ 1 Initialize and evaluate the population P 0 of rules by the ramped-half-and-half method; 2 Calculate set of reference points Z; 3 Set Pg ←P0 and generation ←0; 4 while generation < maxGeneration do 5 Generate the offspring population Qg using the crossover, mutation and reproduction of GP; 6 foreach Q ∈Qg do 7 Evaluate rule Q 8 end 9 Rg ←P g ∪Qg; 10 (F1, F2, · · · ) ←Non-dominated-sort(Rg); 11 Form the new population P g+1 from (F1, F2, · · · ) by the NSGA-III selection; 12 generation ←generation + 1; 13 end 14 return The non-dominated individuals P ∗⊆Pgmax; 1 Initialize and evaluate the population P 0 of rules by the ramped-half-and-half method; 5 Generate the offspring population Qg using the crossover, mutation and reproduction of GP; 14 return The non-dominated individuals P ∗⊆Pgmax; 2), maximal ﬂowtime (see equation (2.2) of Chapter 2), mean weighted tardiness (see equation (2.6) of Chapter 2), maximal weighted tardiness (see equation (2.7) of Chapter 2). The GP-NSGA-III algorithm explores the Pareto-front of non- dominated dispatching rules regarding the four objectives mentioned above. Algorithm 3 shows the workﬂow of our proposed many-objective algorithm. In this algorithm, ﬁrst, a number of training instances are given as input. The initial GP population P0 is created using the ramped-half- and-half method [105]. 4.3.2 General framework of GP-NSGA-III The ramped half-and-half is the best generative method that works best over a broad range of problems and generates a CHAPTER 4. MANY-OBJECTIVE GP-HH FOR JSS 82 wide variety of tree sizes and shapes [105]. All individuals in the GP popu- lation P0 will be evaluated by applying the rules to training instances. This ﬁtness evaluation of GP population P0 is required in line 1 of Algorithm 3. Algorithm 4: The evaluation of a dispatching rule. Input : A training set Itrain and a rule P Output: The ﬁtness f(P) of the rule P 1 foreach I in Itrain do 2 Construct a schedule ∆(P, I) by applying the rule P to the JSS instance I; 3 Calculate the objective values f(∆(P, I)); 4 end 5 f(P) ← 1 \|Itrain\| P I∈Itrain f(∆(P, I)); 6 return f(P); Algorithm 4: The evaluation of a dispatching rule. Input : A training set Itrain and a rule P Output: The ﬁtness f(P) of the rule P 1 foreach I in Itrain do 2 Construct a schedule ∆(P, I) by applying the rule P to the JSS instance I; 3 Calculate the objective values f(∆(P, I)); 4 end 5 f(P) ← 1 \|Itrain\| P I∈Itrain f(∆(P, I)); 6 return f(P); Algorithm 4: The evaluation of a dispatching rule. 2 Construct a schedule ∆(P, I) by applying the rule P to the JSS instance I; 3 Calculate the objective values f(∆(P, I)); 3 Calculate the objective values f(∆(P, I)); 4 end The detailed process for ﬁtness evaluation is presented in Algorithm 4 where f(.) is the objective vector. The quality of each individual (dis- patching rule) in the population will be measured by the average value of the objectives across all training instances. This is shown in line 5 of Algorithm 4. After all, individuals have been evaluated then the reference points Z are generated in Algorithm 3. In this study, we use Das and Dennis’s systematic approach [39]. In this approach, reference points to place on a normalized hyperplane (for more detail, refer to Subsection 2.1.6 in Chap- ter 2). Next, in the algorithm, offspring (Qg) of generation g is generated by applying subtree crossover and subtree mutation to the current popu- lation (Pg). The ﬁtness evaluation of Qg is required in line 7 of Algorithm 3. The detailed process for ﬁtness evaluation is presented in Algorithm 4. 4.3.2 General framework of GP-NSGA-III After the ﬁtness evaluation of Qg, GP-NSGA-III combines the offspring population and the parent Pg together. Later, all the individuals in the combined population Rg are sorted according to different non-domination EXPERIMENTAL STUDIES 83 4.4. ranks (F1, F2 ..., Fl−1, Fl). Then, each non-domination rank is selected one at a time to construct a new population of next-generation Pg+1, starting from F1 to Fl or until the size of the population reaches to N (population size). All solutions from F(l+1) to onwards are rejected from the combined population Rg. If Fl cannot be fully inserted into the population, then each individual of Fl is assigned to a reference point Z ∈Z and selected one by one using the niching method (details of the niching method can be seen in Subsection 2.1.5 of Chapter 2) and in [43]. 4.4 Experimental studies In this section, experimental studies will be conducted to compare the pro- posed GP-NSGA-III to GP-NSGA-II and GP-SPEA2 which are two well- known MOEAs. All compared algorithms are combined with GP as well, i.e., they adopted the solution representation and evolutionary operators of GP and selection mechanisms of NSGA-II and SPEA2 (The detailed in- formation of NSGA-II and SPEA2 can be seen in Section 2.1.4 of Chapter 2). In the following, we will describe the parameter settings of the algo- rithms. For each run of each compared algorithm, the experiment consists of two steps as follows. 1. In the ﬁrst step, the algorithm (GP-NSGA-III, GP-NSGA-II, or GP- SPEA2) is applied to the training set and obtain a set of trade-offs dispatching rules; 2. In the second step, for every problem instance in the test set, we apply every newly evolved rule to produce a schedule and calculate all the considered objective values with respect to the schedule. The above steps are repeated for the compared GP-NSGA-III, GP- NSGA-II and GP-SPEA2 algorithms. In terms of objectives, four objec- tives, i.e. the mean ﬂowtime (mF) (see equation (2.1) of Chapter 2) , max- imal ﬂowtime (maxF) (see equation (2.2) of Chapter 2), mean weighted CHAPTER 4. MANY-OBJECTIVE GP-HH FOR JSS 84 tardiness (mWT) (see equation (2.6) of Chapter 2) and maximal weighted tardiness (maxWT) (see equation (2.7) of Chapter 2) are considered to min- imize in the experiment. We selected the Taillard (TA) static JSS bench- mark instances [182] as the testbed. ( The detailed information of TA can be found in Section 3.1 of Chapter 3). 4.5 Results and discussions In the experiment, for each of the three algorithms, 30 independent runs are carried out. Then, the mean and the standard deviation of HV and IGD values are reported. First, we will analyze how the different schedul- ing objectives are correlated with each other. Then, we will present the training performance results followed by testing performance results. 4.4.1 Parameter settings The population size for GP-NSGA-II and GP-SPEA2 is set to 1024. This is a reasonably large population size that ensures enough diversity in the population [135, 145]. The elitist archive of size in GP-SPEA2 is set to 512 by following the ECJ [125] guidelines (we used the ECJ platform for implementing our algorithms). GP-NSGA-II builds the archive separately which maintains the current best individuals. On the other hand, GP- SPEA2 archive consists of the speciﬁed proportion of its population. NSGA-III is suggested that the population size is equal to the number of reference points [43]. In addition, the reference points of NSGA-III de- pends on the number of objectives. In this study, we chose the population size for GP-NSGA-III that are not signiﬁcantly different from GP-NSGA-II and GP-SPEA2. The population size chooses for GP-NSGA-III is 1025. The total number of generations is set to 50, commonly used in the lit- erature [145, 164]. The crossover rate is set to 85% and the mutation rate is set to 10%. The reproduction rate is set to 5%. These rates (reproduc- tion, crossover, and mutation) have been previously used in [145]. The maximum depth of dispatching rules is set to eight. This depth is used to restrict the dispatching rule from becoming too large [145]. However, we can choose this maximum depth to decrease the computational times of the GP system and make the evolved rules easier to analyze. Tourna- ment selection with a tournament size of seven is used to select individuals for genetic operations [105]. In the experiments, the two commonly used measures in multi-objective optimization, i.e., IGD [206] and HV [212] are RESULTS AND DISCUSSIONS 85 4.5. used to compare the algorithms (The detailed information can be seen in Section 3.3 of Chapter 3). 4.5.1 Relationship among scheduling objectives This section will analyze how the different scheduling objectives are cor- related with each other. In order to visualize the interdependencies be- tween different objectives, we used the aggregated Pareto-front, including the non-dominated evolved dispatching rules extracted from Pareto-front generated by all compared algorithms in 30 independent runs on the train- ing instances. Figure 4.3 shows a scatter plot matrix that contains all the pairwise of four objectives (any two plots in a scatter plot matrix are simi- lar with respect to the diagonal and these two axes are interchanged). This plot helps us to understand the trade-offs between objectives and how op- timizing one objective inﬂuences optimizing another objective. Figure 4.3 shows that ﬂowtime objectives (maxF-mF) are conﬂicting in nature meaning that by optimizing one objective, the other objective be- comes worse. Therefore, maxF compromises its objective values because of the mF and vice versa. This is because trying to minimize the maxF could negatively affect several other jobs in the shop and thus increase the total mF of the entire system. Also, if we are trying to minimize the maxWT it may increase the mWT of other jobs which are late in the entire system. CHAPTER 4. MANY-OBJECTIVE GP-HH FOR JSS 86 Scatter Plot Matrix mF 440 460 480 500 440 460 480 500 380 400 420 440 380 400 420 440 maxF 620 640 660 680 620 640 660 680 560 580 600 620 560 580 600 620 mWT 600 700 600 700 400 500 400 500 maxWT 1500 2000 1500 2000 500 1000 500 1000 Scatter Plot Matrix Figure 4.3: Pareto-front for pairwise objective combination. 4.5.2 Results Table 4.1 shows the mean and standard deviations of the training perfor- mance in terms of HV and IGD of the rules obtained by GP-NSGA-III, GP-NSGA-II, and GP-SPEA2. For each instance, Wilcoxon’s rank-sum test [195] is conducted to compare the best performing algorithm against the other two algorithms. The signiﬁcance level is set to 0.05. If both p-values are smaller than 0.05, then the best algorithm is considered signiﬁcantly better than the other two algorithms. The signiﬁcantly better results are marked in bold. 4.5. RESULTS AND DISCUSSIONS 4.5. RESULTS AND DISCUSSIONS 87 180 185 190 195 200 260 280 300 320 340 meanW_T max_T Figure 4.4: Pareto-front for pairwise objectives combination between (mWT −maxWT). Figure 4.4: Pareto-front for pairwise objectives combination between (mWT −maxWT). It is interesting to note from Figure 4.3 that mF and mWT have a strong relationship when the values are high. Thus, by optimizing one objective, the other objective will also be indirectly optimized. However, trade-offs between mF and mWT can be witnessed when these objectives reach their lowest extreme. We can see from the Pareto-front that reducing maxWT will deteriorate other objectives. On the other hand, maxF optimized well with mWT and mF. It can also be observed from Figure 4.4 that the due date related objec- tives (maxWT-mWT) have overlapping solutions. This could be possible because the jobs which contribute late to the objective value are mostly redundant. CHAPTER 4. MANY-OBJECTIVE GP-HH FOR JSS 88 Table 4.1: The mean and standard deviation over the average HV and IGD values on training instances of the compared algorithms in the four- objective experiment. HV (¯x ± σ) GP-NSGA-III GP-NSGA-II GP-SPEA2 0.694(0.0152) 0.688(0.0220) 0.564(0.0233) IGD (¯x ± σ) GP-NSGA-III GP-NSGA-II GP-SPEA2 0.00139(0.00021) 0.00163(0.00034) 0.00235(0.00019) HV (¯x ± σ) GP-NSGA-III GP-NSGA-II GP-SPEA2 0.694(0.0152) 0.688(0.0220) 0.564(0.0233) IGD (¯x ± σ) GP-NSGA-III GP-NSGA-II GP-SPEA2 0.00139(0.00021) 0.00163(0.00034) 0.00235(0.00019) Performance of Obtained Dispatching Rules Table 4.1 reveals that GP-NSGA-III outperforms GP-NSGA-II and GP- SPEA2 in terms of both HV and IGD on training instances. However, GP-NSGA-II is very competitive with GP-NSGA-III in terms of HV. For more detail, Tables 4.2 and 4.3 show the test performance on the 40 test instances. Tables 4.2 and 4.3 show the mean and standard deviation over the HV and IGD values of the 30 independent runs of the compared algorithms with respect to the 40 test instances in the 4-objective experiment. In the tables, “ID” is the instance ID, and #J and #M indicate the number of jobs RESULTS AND DISCUSSIONS 89 4.5. and machines, respectively. One can see that as the instance ID increases, the problem size increases as well. Tables 4.2 and 4.3 together suggest that GP-NSGA-III performed sig- niﬁcantly better in terms of both HV and IGD than GP-NSGA-II and GP- SPEA2 in most of the tested instances with four objectives have signiﬁ- cantly deteriorated each other. In terms of HV, GP-NSGA-III performed the best with statistical signiﬁcance in 31 out of the total 40 instances. GP- NSGA-II outperformed the other algorithms in the remaining 8 instances. GP-SPEA2 failed to achieve the best performance in any instance. In terms of IGD, Table 4.3 reveals a similar pattern. GP-NSGA-III per- formed signiﬁcantly better in 37 out of the 40 instances, and GP-NSGA-II achieved the best performance in the remaining 3 instances. Again, there is no instance in which GP-SPEA2 performed the best. When taking a closer look at the tables, it can be found that GP-NSGA- II performed better on smaller instances (no more than 20 jobs and 20 ma- chines). On the other hand, GP-NSGA-III appeared to be more effective on larger instances. This demonstrates the advantage of GP-NSGA-III, es- pecially on challenging problems. In addition, GP-NSGA-III appeared to be more effective at improving IGD than HV. This might be GP-NSGA-III seem to contribute most of the non-dominated solutions in the approxi- mate Pareto-front. 4.5.3 Discussions In this section, we further analyze the evolved dispatching rules and ob- tained distribution of the optimal solutions. We also show the average HV and IGD of the non-dominated solutions evolved by compared algorithms across multiple generations. Computational Complexity of the Algorithm The proposed algorithm (GP-NSGA-III) is hybridized, combines GP with the evaluation scheme of NSGA-III. So, the main computational complex- CHAPTER 4. MANY-OBJECTIVE GP-HH FOR JSS 90 Table 4.2: The mean and standard deviation over the HV values on the test instances of the compared algorithms. ID #J #M GP-NSGA-III GP-NSGA-II GP-SPEA2 1 15 15 .0292(.0068) .0035(.0007) .0170(.0028) 2 15 15 .0541(.0049) .3430(.0493) .0345(.0037) 3 15 15 .0668(.0062) .0197(.0021) .0179(.0017) 4 15 15 .0329(.0049) .1136(.0250) .0126 (.0012) 5 15 15 .0344(.0034) .0122(.0012) .0116(.0017) 6 20 15 .1027(.0087) .1796(.0094) .0410(.0032) 7 20 15 .0668(.0085) .2354(.0122) .0425(.0017) 8 20 15 .1193(.0118) .2506(.0138) .0457(.0029) 9 20 15 .1287(.0109) .0593(.0035) .0602(.0047) 10 20 15 .1106(.0118) .0876(.0038) .0533(.0032) 11 20 20 .0945(.0056) .2706(.0139) .0307(.0031) 12 20 20 .0873(.0064) .5793(.0209) .0361(.0021) 13 20 20 .0348(.0057) .0332(.0015) .0346(.0024) 14 20 20 .0826(.0047) .0241(.0017) .0330(.0025) 15 20 20 .0030(.0032) .0161(.0012) .0158(.0017) 16 30 15 .2235(.0188) .1729(.0048) .1263(.0062) 17 30 15 .2092(.0132) .1514(.0046) .1139(.0047) 18 30 15 .1932(.0132) .1402(.0041) .1194(.0057) 19 30 15 .1976(.0161) .1525(.0057) .1058(.0040) 20 30 15 .1863(.0157) .1533(.0044) .1163(.0046) 21 30 20 .1863(.0184) .1188(.0038) .0839(.0047) 22 30 20 .0777(.0067) .0230(.0019) .0249(.0031) 23 30 20 .1739(.0127) .1119(.0034) .0880(.0058) 24 30 20 .1687(.0147) .1401(.0028) .0900(.0034) 25 30 20 .1692(.0127) .1252(.0035) .0920(.0044) 26 50 15 .2609(.0139) .1999(.0062) .1713(.0070) 27 50 15 .3092(.0229) .2035(.0047) .1750(.0061) 28 50 15 .2453(.0159) .1820(.0070) .1577(.0058) 29 50 15 .2811(.0216) .1789(.0062) .1578(.0067) 30 50 15 .2708(.0177) .1845(.0053) .1599(.0098) 31 50 20 .2143(.0135) .1738(.0062) .1366(.0054) 32 50 20 .3182(.0193) .2001(.0042) .1582(.0045) 33 50 20 .2568(.0162) .1829(.0066) .1383(.0051) 34 50 20 .2955(.0189) .1876(.0046) .1522(.0055) 35 50 20 .2737(.0130) .1851(.0043) .1500(.0064) 36 100 20 .3333(.0163) .2169(.0062) .2131(.0081) 37 100 20 .2687(.0164) .0896(.0037) .0796(.0048) 38 100 20 .2659(.0145) .2141(.0057) .2132(.0061) 39 100 20 .2345(.0134) .1141(.0668) .1111(.0055) 40 100 20 .3564(.0190) .2217(.0046) .2211(.0059) RESULTS AND DISCUSSIONS 91 4.5. Table 4.3: The mean and standard deviation over the IGD values on the test instances of the compared algorithms. Computational Complexity of the Algorithm ID #J #M GP-NSGA-III GP-NSGA-II GP-SPEA2 1 15 15 .0071(4.0E-05) .0077(4.9E-05) .0087(1.4E-05) 2 15 20 .0071(6.9E-05) .0037(2.1E-05) .0105(9.3E-05) 3 15 15 .0009(1.9E-05) .0018(6.1E-06) .0017(1.8E-05) 4 15 15 .0057(3.9E-05) .0036(1.2E-05) .0119(4.7E-05) 5 15 15 .0013(1.6E-05) .0023(1.4E-05) .0037(4.9E-05) 6 20 15 .0024(2.2E-05) .0028(3.3E-05) .0051(6.2E-05) 7 20 15 .0110(7.5E-05) .0088(5.9E-05) .0167(1.4E-05) 8 20 15 .0027(4.0E-05) .0038(3.3E-05) .0060(9.2E-05) 9 20 15 .0023(2.3E-05) .0033(3.9E-05) .0054(7.5E-05) 10 20 15 .0029(3.0E-05) .0044(5.6E-05) .0072(9.0E-05) 11 20 20 .0017(1.4E-05) .0025(1.5E-05) .0046(6.3E-05) 12 20 20 .0023(1.8E-05) .0031(2.4E-05) .0048(5.9E-05) 13 20 20 .0016(9.7E-06) .0027(1.0E-05) .0042(8.1E-05) 14 20 20 .0021(2.3E-05) .0027(2.2E-05) .0045(5.7E-05) 15 20 20 .0032(3.6E-05) .0050(4.0E-05) .0083(9.3E-05) 16 30 15 .0009(1.7E-05) .0013(1.9E-05) .0020(2.5E-05) 17 30 15 .0024(3.3E-05) .0032(4.9E-05) .0054(5.5E-05) 18 30 15 .0017(2.0E-05) .0023(3.0E-05) .0036(3.4E-05) 19 30 15 .0025(3.4E-05) .0033(3.8E-05) .0052(6.3E-05) 20 30 15 .0015(2.3E-05) .0018(3.1E-05) .0030(2.8E-05) 21 30 20 .0026(4.1E-05) .0033(4.7E-05) .0050(1.0E-05) 22 30 20 .0040(2.5E-05) .0048(2.7E-05) .0080(7.5E-05) 23 30 20 .0027(2.0E-05) .0031(3.1E-05) .0048(4.4E-05) 24 30 20 .0033(5.1E-05) .0037(5.0E-05) .0059(6.3E-05) 25 30 20 .0030(4.0E-05) .0037(3.6E-05) .0059(6.1E-05) 26 50 15 .0013(1.3E-05) .0018(2.9E-05) .0031(3.5E-05) 27 50 15 .0007(1.2E-05) .0013(2.2E-05) .0022(2.7E-05) 28 50 15 .0041(4.2E-05) .0047(8.3E-05) .0081(9.8E-05) 29 50 15 .0017(3.0E-05) .0023(4.2E-05) .0041(4.9E-05) 30 50 15 .0020(3.6E-05) .0024(4.0E-05) .0041(4.6E-05) 31 50 20 .0010(1.0E-05) .0015(4.2E-05) .0024(4.3E-05) 32 50 20 .0008(9.6E-06) .0013(2.1E-05) .0021(3.2E-05) 33 50 20 .0033(5.0E-05) .0035(4.5E-05) .0058(6.6E-05) 34 50 20 .0012(2.4E-05) .0016(2.8E-05) .0026(2.7E-05) 35 50 20 .0013(2.2E-05) .0020(3.3E-05) .0034(3.4E-05) 36 100 20 .0011(1.6E-05) .0016(1.9E-05) .0030(3.8E-05) 37 100 20 .0005(8.4E-06) .0012(1.9E-05) .0010(1.4E-05) 38 100 20 .0015(1.7E-05) .0018(3.3E-05) .0035(8.8E-05) 39 100 20 .0012(1.3E-05) .0027(3.7E-05) .0047(5.8E-05) 40 100 20 .0009(1.4E-05) .0015(2.3E-05) .0028(4.6E-05) CHAPTER 4. MANY-OBJECTIVE GP-HH FOR JSS 92 ity resulted from an evaluation scheme of NSGA-III and computational time for evolving dispatching rules. The evaluation scheme of NSGA-III holds a computational complexity of O(MN 2) [43], where M is the objec- tive number and N is the population size. In addition, the computational complexity of two compared algorithms (NSGA-II and SPEA2) are also having O(MN 2) [44, 211]. To know the computational cost of each algorithm to evolve rules, we performed several experiments. Figure 4.5 shows each algorithm’s com- putational cost (GP-NSGA-III, GP-NSGA-II, GP-SPEA2) to evolve rules. Figure 4.5 reveals that the computational cost of GP-NSGA-III is lower than GP-NSGA-II and GP-SPEA2. In addition, the computational time of GP-NSGA-III is quite consistent in all generations as compared to GP- NSGA-II and GP-SPEA2. Computational Complexity of the Algorithm One possible explanation is that GP-NSGA-II and GP-SPEA2 have trouble ﬁnding effective compact rules in most gen- erations and good performances can only be obtained by large and so- phisticated rules in the GP-NSGA-II and GP-SPEA2 populations. As these rules are evolved, they will take GP-NSGA-II and GP-SPEA2 more time for evaluations. The lengths of rules obtained by GP-NSGA-III, GP-NSGA-II, and GP- SPEA2 are shown in Figure 4.6. The length of a rule is the total num- ber of nodes, including terminal and function nodes. In most genera- tions, GP-NSGA-III produces shorter rules as compared to GP-NSGA-II, and GP-SPEA2. This is also consistent with our results that the compu- tational time of GP-NSGA-III is lower than GP-NSGA-II and GP-SPEA2. This is because GP-NSGA-III can ﬁnd more compact rules which require less time for evaluations. Further, the rules generated by GP-SPEA2 is rel- atively short as compared to GP-NSGA-II. Given the poor performance of the rules generated by GP-SPEA2 in terms of HV and IGD (see Section 4.5.2), the results here show that GP-SPEA2 has failed to develop effective rules. Figure 4.7 shows the length of the best rules of each run. Overall the 4.5. RESULTS AND DISCUSSIONS 93 Figure 4.5: Computational time to evolve dispatching rules Figure 4.6: Length of rules of each generation in GP-NSGA-II, GP-SPEA2, and GP-NSGA-III. Figure 4.5: Computational time to evolve dispatching rules Figure 4.5: Computational time to evolve dispatching rules Figure 4.5: Computational time to evolve dispatching rules Figure 4.5: Computational time to evolve dispatching rules Figure 4.6: Length of rules of each generation in GP-NSGA-II, GP-SPEA2, and GP-NSGA-III. Figure 4.6: Length of rules of each generation in GP-NSGA-II, GP-SPEA2, and GP-NSGA-III. the length of the best rules of GP-NSGA-II and GP-SPEA2 are quite long as compared to GP-NSGA-III which may directly affect the computational time of the algorithm. CHAPTER 4. MANY-OBJECTIVE GP-HH FOR JSS 94 0 10000 20000 30000 40000 50000 60000 70000 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Length of rules Runs GP-NSGA-III GP-NSGA-II GP-SPEA2 Figure 4.7: Length of the best rules of each run in GP-NSGA-II, GP-SPEA2, and GP-NSGA-III. Figure 4.7: Length of the best rules of each run in GP-NSGA-II, GP-SPEA2, and GP-NSGA-III. Analysis of dispatching rules This section analyzes each terminal’s frequency that was counted in the set of 30 independent runs of evolved rules from GP-NSGA-II, GP-SPEA2, and GP-NSGA-III. From the frequency of terminals, we can further an- alyze the relevant and irrelevant terminals. The bar chart in Figure 4.8 shows the frequency of terminals in each algorithm. We can be seen from Figure 4.8 that W and PT are the most frequently chosen terminals in all algorithms. However, we need to investigate the relevance of terminals to optimize tardiness and ﬂowtime objectives. According to the existing studies in the literature [70, 114], MRT, PT, WKR, WINQ, NOINQ, FDD, and NOPT are useful terminals (relevant) for optimizing ﬂowtime objectives. Speciﬁcally, PT, WINQ, and WKR are the most important three terminals for optimizing the ﬂowtime objective. On the other hand, WINQ, NOINQ, NOPT, W, PT, MRT, and DD are the RESULTS AND DISCUSSIONS 95 4.5. useful terminals for optimizing tardiness objectives. PT, DD, and W are the most useful terminals for optimizing tardiness objectives. It can be seen from Figure 4.8 that W, PT are the most frequently chosen terminals in all algorithms. The frequency of occurrences of PT and W terminals is higher in the GP-NSGA-III than GP-NSGA-II. It can also be seen that WINQ, MRT, NOPT, and DD are also more frequent terminals in GP-NSGA-III. NOINQ and WKR are more frequent terminals in GP- NSGA-II. GP-NSGA-III and GP-NSGA-II have an identical frequency of occurrences of FDD. It is not very clear whether the remaining terminals (WIQ, NMRT) are useful or not. It can be seen from Figure 4.8 that GP-NSGA-III has a greater frequency of useful terminals which positively affects the performance of the GP- NSGA-III. As mentioned in the last section 4.5.2, GP-NSGA-III evolved signiﬁcantly better rules as compared to the other algorithms in terms of HV and IGD. Therefore, GP-NSGA-III selects well-converged and well- diversiﬁed rules. In this sense, these selected rules are well-optimized and have more occurrences of these useful terminals. Feature relevance The contribution of a feature (terminal) to a priority function on a set of JSS instances show the relevancy of each terminal. In other words, the terminal (ti) contribution to the priority function is the difference between the ﬁtness values before and after removing ti from the priority function. A positive value of contribution indicates that after removing ti, the new dispatching rule will generate worse schedules for the tested JSS instances. On the other hand, a negative value implies that ti makes a negative contri- bution to a priority function, and removing ti can lead to an improvement in the performance. If the value is zero, ti makes no contribution. Intu- itively, if ti is a more relevant or useful terminal, it contributes to the rules that perform well in tested JSS instances. To perform this experiment, we are using an ofﬂine terminal-selection CHAPTER 4. MANY-OBJECTIVE GP-HH FOR JSS 96 Figure 4.8: Frequency of terminals in GP-NSGA-II, GP-SPEA2, and GP- NSGA-III. Figure 4.8: Frequency of terminals in GP-NSGA-II, GP-SPEA2, and GP- NSGA-III. strategy. We used the following steps for the selection of terminals. strategy. We used the following steps for the selection of terminals. 1. simplify the obtained rules(see Section 4.5.2) according to the princi- ples which were proposed in [131], 2. calculate the relevance of each terminal which is based on the 30 best rules. The bar chart in Figure 4.9 shows the frequency of each terminal (after simpliﬁcation of rules) that was counted in the set of 30 independent runs of evolved rules from GP-NSGA-III, GP-NSGA-II, GP-SPEA2. Figure 4.8 reveals that W and PT are the most frequently chosen terminals in all al- gorithms which is consistent with Figure 4.8 where the W and PT are also the most frequent terminals. Figure 4.9 also shows the PT and W termi- nals’ frequencies are higher in the GP-NSGA-III than GP-NSGA-II and GP-SPEA2. It can be seen from Figure 4.8 that the WKR is a more fre- quent terminal in GP-NSGA-II but Figure 4.9 shows after simpliﬁcation 4.5. RESULTS AND DISCUSSIONS 4.5. RESULTS AND DISCUSSIONS 97 0 10000 20000 30000 40000 50000 60000 70000 80000 90000 100000 W PT FDD DD WKR MRT WINQ NOIQ NOINQ NMRT NOPT WIQ Frequency of Terminals Terminals GP-NSGA-III GP-NSGA-II GP-SPEA2 Figure 4.9: Frequency of terminals (after simpliﬁcation) in GP-NSGA-III, GP-NSGA-II, GP-SPEA2. Figure 4.9: Frequency of terminals (after simpliﬁcation) in GP-NSGA-III, GP-NSGA-II, GP-SPEA2. of rules, WKR is the most frequent terminals in GP-NSGA-III. Moreover, GP-NSGA-III and GP-NSGA-II have an identical frequency of occurrences of FDD in Figure 4.8 but after simpliﬁcation of rules, FDD is higher in the GP-NSGA-II. Figure 4.10 shows the average relevance of terminals from all three al- gorithms to optimize tardiness and ﬂowtime objectives. Figure 4.10 re- veals that PT has the largest relevance (nearly 0.48), followed by PT, which is about 0.4. There are 4 features whose relevance is almost close to zero (FDD, NOIQ, NMRT, and WIQ), indicating that they made no contribution to the best rule most of the time. Hence W, PT, DD, WKR, MRT, WINQ, NOINQ, and NOPT are selected as relevant terminals to tardiness and ﬂowtime objectives. Most of these terminals are higher in the GP-NSGA- III than GP-NSGA-II and GP-SPEA2 which can positively affect the per- formance of the GP-NSGA-III. CHAPTER 4. MANY-OBJECTIVE GP-HH FOR JSS 98 0 0.1 0.2 0.3 0.4 0.5 0.6 W PT FDD DD WKR MRT WINQ NOIQ NOINQ NMRT NOPT WIQ Relevance Terminals Figure 4.10: The average relevance of each terminal over the 30 indepen- dent runs of GP-NSGA-III, GP-NSGA-II and GP-SPEA2 on the training set. 0.6 0 0.1 0.2 0.3 0.4 0.5 0.6 W PT FDD DD WKR MRT WINQ NOIQ NOINQ NMRT NOPT WIQ Relevance Terminals Figure 4.10: The average relevance of each terminal over the 30 indepen- dent runs of GP-NSGA-III, GP-NSGA-II and GP-SPEA2 on the training set. 0 0.1 0.2 0.3 0.4 0.5 W PT FDD DD WKR MRT WINQ NOIQ NOINQ NMRT NOPT WIQ Relevance Terminals Terminals Figure 4.10: The average relevance of each terminal over the 30 indepen- dent runs of GP-NSGA-III, GP-NSGA-II and GP-SPEA2 on the training set. Further analysis To further investigate how the reference point scheme affects the GP search process, the average HV and average IGD of the non-dominated solutions obtained on training instances are plotted for each generation during the 30 independent runs of the three compared algorithms, as given in Figures 4.11 and 4.12. Figures 4.11 and 4.12 show that GP-NSGA-III obtained better conver- gence curves in terms of both HV and IGD as compared to the GP-NSGA- II and GP-SPEA2. The convergence curve shows that in terms of HV, GP- NSGA-II performs better in the early generations but then GP-NSGA-III has better HV value than the other two compared algorithms. GP-SPEA2 obtained the worst performance in both HV and IGD, which is consistent with the results reported in Table 4.1. more diversiﬁed rules than GP-NSGA-II. more diversiﬁed rules than GP-NSGA-II. Figures 4.14 (a) to 4.16 (b) represents the box plots of the HV and IGD values on different test instances. For the HV metric, GP-NSGA-III achieved signiﬁcantly better results on complex instances (ID-21 and ID- 32). On the other hand, GP-NSGA-II performs signiﬁcantly better on the smallest instance (ID-6). As for IGD, GP-NSGA-III has again proven to be signiﬁcantly better than the other two algorithms on both small and complex instances. 4.5. RESULTS AND DISCUSSIONS 4.5. RESULTS AND DISCUSSIONS 99 Figure 4.11: The average HV value of the non-dominated solutions on the training set during the 30 independent GP runs. Figure 4.11: The average HV value of the non-dominated solutions on the training set during the 30 independent GP runs. Figure 4.12: The average IGD value of the non-dominated solutions on the training set during the 30 independent GP runs. Figure 4.12: The average IGD value of the non-dominated solutions on the training set during the 30 independent GP runs. Figures 4.13 (a) and 4.13 (b) depict the populations’ distribution of the ﬁtness values of GP-NSGA-II and GP-NSGA-III in generations 50 with 4 objectives. It can be seen that GP-NSGA-III managed to achieve better values in objectives 3 and 4 and covered a wider range of trade-offs cor- responding to objective 1. This also shows that GP-NSGA-III can evolve CHAPTER 4. MANY-OBJECTIVE GP-HH FOR JSS 100 (a) GP-NSGA-II (b) GP-NSGA-III. Figure 4.13: Parallel coordinate plot for the distribution of the reference points and the ﬁtness values of the population at generations 50 . (b) GP-NSGA-III. (b) GP-NSGA-III. Figure 4.13: Parallel coordinate plot for the distribution of the reference points and the ﬁtness values of the population at generations 50 . 4.6 Chapter summary Dispatching rules for JSS problems are mainly designed in the literature for single-objective and multi-objective optimization problems. However, several industries (e.g., manufacturing and cloud) require suitable dis- patching rules that can balance several objectives instead of simply focus- 4.6. CHAPTER SUMMARY 101 4.6. CHAPTER SUMMARY 101 (a) HV (b) IGD. Figure 4.14: Box plots on instance 6 of the compared algorithms. (a) HV (b) IGD. Figure 4.15: Box plots on instance 21 of the compared algorithms. (a) HV (b) IGD. Figure 4.14: Box plots on instance 6 of the compared algorithms. (a) HV (b) IGD. Figure 4 14: Box plots on instance 6 of the compared algorithms Figure 4.14: Box plots on instance 6 of the compared algorithms. (a) HV (b) IGD. Figure 4.15: Box plots on instance 21 of the compared algorithms. (a) HV (b) IGD. Figure 4.15: Box plots on instance 21 of the compared algorithms. Figure 4.15: Box plots on instance 21 of the compared algorithms. ing only on one objective. Thus, this chapter studies the importance of considering many (four) potentially conﬂicting objectives for effective JSS. Aiming at solving JSS problems with many objectives, the new algorithm CHAPTER 4. MANY-OBJECTIVE GP-HH FOR JSS CHAPTER 4. MANY-OBJECTIVE GP-HH FOR JSS 102 (a) HV (b) IGD. Figure 4.16: Box plots on instance 32 of the compared algorithms. Figure 4.16: Box plots on instance 32 of the compared algorithms. was developed that seamlessly combines GP-HH for evolving dispatching rules with the selection technique introduced in NSGA-III. Extensive experiments have been performed to investigate the pro- posed algorithm’s effectiveness compared with the other two popular multi-objective algorithms, i.e., NSGA-II and SPEA2. The experimental studies utilized the Taillard static job-shop benchmark set. Our experi- ment results gave evidence that the proposed algorithm performed con- sistently better than GP-NSGA-II and GP-SPEA2 on the majority of the problem instances. Several interesting conclusions were also drawn. The results have shown on both training and test instances, GP-NSGA-III can evolve more effective rules than GP-NSGA-II and GP-SPEA2 in terms of HV and IGD. Further, GP-NSGA-III has well-optimized rules; therefore, it has more ef- fective terminals than GP-NSGA-II and GP-SPEA2. Further, the results show that the evolved rules enjoy high generalization capability and can effectively tackle both training instances and unseen problem instances. In the next chapter, we will focus on addressing the second issue: di- 103 .6. CHAPTER SUMMARY 4.6. versity maintenance [199]. In the literature, diversity maintenance is of- ten controlled by providing multiple predeﬁned reference points. These points are widely distributed. However, uniformly distributed reference points are inappropriate when the true Pareto-front is non-uniform and irregular [85, 86], such as JSS problems. These problems may have some reference points that are never associated with any of the Pareto-optimal solutions and will become useless reference points during evaluation. The existence of these useless reference points in algorithms signiﬁcantly af- fects its performance. To address this issue, we have proposed a new ref- erence point adaptation mechanisms in our next chapter. CHAPTER 4. MANY-OBJECTIVE GP-HH FOR JSS 104 5.1 Introduction This chapter addresses the second issue of many-objective optimization problems (MaOPs), which is diversity maintenance [199]. In the literature, diversity maintenance is often controlled by providing multiple prede- ﬁned reference points such as NSGA-III [43] and SPEAR [92]. These points are widely distributed on a normalized hyperplane (simplex) [31, 43] and they are expected to guide the population to be distributed evenly over the true Pareto-front [43]. Reference points based algorithms such as NSGA-III [43], RVEA [31], and SPEAR [92] use a predeﬁned set of uniformly distributed reference points and they have successfully solved various practical many-objective optimization problems [43]. This set of widely distributed reference points promotes the diversity of evolved solutions in the Pareto-front with three to ﬁfteen objectives. Even though these algorithms have successfully solved various practi- cal MaOPs, they still have challenges when applying to real-world prob- lems such as the car cab design problem and the design of vehicle prob- 105 CHAPTER 5. REFERENCE POINTS ADAPTATION METHODS 106 lems. These real-world problems usually have non-uniform and irregular Pareto-front. Therefore, the adoption of uniformly distributed reference points affects algorithms’ performance adversely [85, 86]. This is because many of these reference points are never associated with any of the opti- mal solutions and become useless. Also, useless reference points will no- tably affect the performance of reference points based algorithms which is hard to achieve whenever the true Pareto-front is irregular (e.g., non- convex or non-uniform). This issue has been highlighted in the literature [85, 86]. Particularly, in problems with an irregular, non-uniform and discon- nected Pareto-front, a few reference points are associated with more than one optimal solution in their closest proximity. If the reference point is as- sociated with many solutions, they may not be selected during evolution. Therefore, candidate solutions that should enjoy higher selection oppor- tunities may not be selected because they are associated with a popular reference point [86]. Figure 5.1 shows that reference point F is associated with many solutions than the other useful points (A, C, and D). Consider the population size (N=10) in this example. Therefore, reference point ( A or C or D) has the minimum associated solutions selected ﬁrst and mem- bers associated with these points become a part of the next generation. The reference point F has a lesser selection opportunity than the other ref- erence points. 5.1 Introduction This is because a number of solutions associated with A, C, and D are equal to the population size. Thus, solutions associated with F may not be part of the next generation and highly affect the solution diversity. In many combinatorial optimization problems, such as job shop scheduling (JSS) problems, the true Pareto-front is usually irregular and discontinuous [142]. JSS problems have also inherited the issue of use- less reference points, i.e., the solution diversity’s demotion. Despite the promising results of GP-NSGA-III ( the detailed information can be seen in Subsection 4.3.2 of Chapter 4), we have found (shown in Figure 5.2) that 5.1. INTRODUCTION 107 Figure 5.1: Associated and contributing solutions on a convex curvature of Pareto-front Figure 5.1: Associated and contributing solutions on a convex curvature of Pareto-front many reference points are useless. They are never associated with any dispatching rules on the evolved Pareto-front in JSS. It can also be seen in Figure 5.2 that the number of useless references points in GP-NSGA-III kept increasing from 910 to about 980. Suppose only a few reference points are truly associated with the Pareto-optimal dispatching rules at the cur- rent generation. In that case, it is not easy to distinguish and select these rules to improve diversity for future generations. Due to the above rea- son, we should ﬁnd how to get better matches between reference points and the evolved Pareto-front that can help to enhance solution diversity and, therefore, the performance of GP-NSGA-III. To enhance the solutions’ diversity and reduce the useless points, the distribution of reference points must have a better match with the candi- date solutions’ distribution. For this purpose, A-NSGA-III [85], RVEA* [31] proposed reference points adaptation strategies for generating refer- ence points according to the approximated Pareto-front which can be seen in Section 2.1.6 of Chapter 2. RVEA* and A-NSGA-III mechanisms require the number of refer- 108 CHAPTER 5. REFERENCE POINTS ADAPTATION METHODS Figure 5.2: The curve of average number of useless reference points in GP- NSGA-III on the training instances during the 30 independent GP runs Figure 5.2: The curve of average number of useless reference points in GP- NSGA-III on the training instances during the 30 independent GP runs ence points to be changed dynamically. In high-dimensional objective space, when the number of reference points becomes large, adding or replacing reference points may affect algorithms’ efﬁciency. 5.2. GENERAL FRAMEWORK FOR ADAPTIVE REFERENCE POINTS GENERATION109 ERAL FRAMEWORK FOR ADAPTIVE REFERENCE POINTS GENERATION109 1. To develop a new adaptive reference point strategy using the model- free approach (heuristics approach). In the model-free approach, we proposed a new adaptation mechanism inspired by particle swarm optimization (PSO). This reference point adaptation mechanism is expected to signiﬁcantly improve our proposed algorithm’s perfor- mance by reducing useless reference points. The proposed approach will be evaluated on the Taillard benchmark JSS problem and is com- pared with the baseline algorithm (GP-NSGA-III). 2. To develop an adaptation mechanism using a model-based tech- nique that estimates the density of solutions from each deﬁned sub- location in a whole objective space. Subsequently, it applies a Gaus- sian Process on the model, which gives smoothness to the model. The performance of the proposed adaptive model-based technique is also veriﬁed on JSS instances and compared with GP-A-NSGA-III [85] and GP-NSGA-III. In this section, we will further show the gen- eral applicability and usefulness of the model-based reference point adaptation technique by applying them to the benchmarks prob- lems. This chapter is divided into two major parts. First, we provide the model- free reference points generation algorithm, followed by the experimental design and the results compared with GP-NSGA-III approaches. After- ward, the model-based algorithms in the context of adaptive reference point generation are discussed. Next, we present the experimental design and the results. Finally, a chapter summary is provided. 5.1 Introduction Moreover, both RVEA* and A-NSGA-III use some heuristics for generating reference points. These heuristics do not have any concrete knowledge of the objec- tive space and generated the reference points based on the last generation solutions. This chapter aims to develop two new effective mechanisms (model-free and model-driven) for reference point generation and improve the association be- tween reference points and the Pareto-front during evaluation to address the issues mentioned above. Furthermore, better matches between refer- ence points and the evolved Pareto-front can help to enhance the solution diversity and quality of dispatching rules. The objective of this chapter is further broken down into two major sub-objectives: 5.2 General framework for adaptive reference points generation The proposed algorithms for generating adaptive reference points can be considered as a major enhancement of GP-NSGA-III (GP-NSGA-III is de- CHAPTER 5. REFERENCE POINTS ADAPTATION METHODS 110 veloped in Section 4.3.2 of Chapter 4). Both model-based and model- free mechanisms utilize the adaptive reference point’s general framework, which is similar to that of the GP-NSGA-III. In GP-NSGA-III, the popula- tion is created after the niching operation Pt+1. Further, the niche counts ρi (the number of population members associated with ith reference point, zi ∈Z) for each reference point is updated then reference points are gen- erated adaptively. More details of adaptive reference point mechanisms will be discussed in the next two sections. 5.3 Model-free adaptive reference points gener- ation In GP-NSGA-III, the size of reference points is almost equal to the popu- lation’s size (N). Additionally, in an ideal condition, every reference point is expected to be associated with one population member. Thus, if ρ ≥2, some other reference point has ρ ( niche count) equal to zero and is con- sidered to be a useless point. In this model-free approach, each reference point has its ﬁtness value which is deﬁned as the number of individuals in the population associ- ated with reference points. In other words, the ﬁtness value of the useless reference point has a niche count (ρ) which is equal to zero. These use- less reference points are relocated to those reference points which have maximum ﬁtness value. This relocation strategy provides a better match between reference points and the Pareto-front evolved by GP-HH. This is expected to help to enhance the solution diversity and, therefore, the performance of GP-NSGA-III. PSO has been proven in the literature to be highly effective for approx- imating arbitrary distributions such as the ﬁtness landscape [153, 204]. Our model-free approach GP-Adaptive-NSGA-III(PSO). (GP-A-NSGA- III(PSO)) is inspired by the principle of PSO, which has a swarm of parti- 5.3. MODEL-FREE ADAPTIVE REFERENCE POINTS GENERATION 111 cles (reference points). The particle’s size is similar to the size of the refer- ence point (nref). Particularly, each reference point is seen as a particle in the swarm and each reference point consists of its own velocity vector V along with the position vector X. Like the PSO, GP-A-NSGA-III(PSO) also has the concept of the global best particle. The global best particle is the one with the most individuals associated with it. The swarm consists of only one single best position in the PSO, but our proposed algorithm can have more than one global best with no elitism strategy being adopted for the global best. This can be seen in Algorithm 5 (line 4). This way, the reference points can have sufﬁcient diversity. Every useless reference point can be optionally attracted to its closest global best location. This GP-A-NSGA-III(PSO) reference point up- date scheme (update(Z)) is described in Algorithm 5 where Z is a set of reference points. Each useless particle updates its position according to the new velocity shows in line 11 in Algorithm 5. 5.3 Model-free adaptive reference points gener- ation 3 for i = 1 →nref do 4 Calculate the global best Zk ∗= arg max{fit(Zi)} ; // global best 5 end 6 for i = 1 →nref do 7 if ρi ̸= 0 then 8 for j = 1 →M do 9 V g i,j = w ∗V g−1 i,j + c2 ∗rand() ∗(Zk ∗−Xg−1 i,j ); // velocity vector of generation g. 10 ; 11 Xg i,j = Xg−1 i,j + V g i,j; // position vector of generation g. 12 ; 13 end 14 end 15 end 16 return Zg = (Xg, V g); Algorithm 5: Update of the reference points. 1 Calculate ﬁtness fit(Zi)= ρi (number of individuals associated for each ; // inertia weight of generation(g), gmax is the maximum generation. 3 for i = 1 →nref do 4 Calculate the global best Zk ∗= arg max{fit(Zi)} ; // global best 5 end 16 return Zg = (Xg, V g); 5.3 Model-free adaptive reference points gener- ation Note that when updating the velocity (line 9) in Algorithm 5, the term for the local best position (PSO uses of memory of each particle’s best location) is ignored. It can be seen that the ﬁtness of particles (reference points) depends on the distribution of the whole swarm and the positions of other particles. Thus, it may not be meaningful to move towards the local best, which can become worse upon particles’ movements prevent the majority of reference points from converging to small areas in the objective space. The new velocity is deﬁned as V g i,j = w ∗V g−1 i,j + c2 ∗rand() ∗(Z∗−Xg−1 i,j ). (5.1) (5.1) where Z∗denotes the global best position and the parameters w, c2 are inertia weight and positive constants respectively. In the baseline, PSO algorithm w is chosen as [0.5 0.9]. Also, V g−1 i,j is a velocity vector of i- th element of the preceding generation, which includes the M objective values. 112 CHAPTER 5. REFERENCE POINTS ADAPTATION METHODS Algorithm 5: Update of the reference points. Input : Reference points Z = (X, V ), ρ Output: Updated reference points Zg 1 Calculate ﬁtness fit(Zi)= ρi (number of individuals associated for each Zi ∈Z); 2 Calculate w = wmax −g · (wmax −wmin)/gmax; ; // inertia weight of generation(g), gmax is the maximum generation. 3 for i = 1 →nref do 4 Calculate the global best Zk ∗= arg max{fit(Zi)} ; // global best 5 end 6 for i = 1 →nref do 7 if ρi ̸= 0 then 8 for j = 1 →M do 9 V g i,j = w ∗V g−1 i,j + c2 ∗rand() ∗(Zk ∗−Xg−1 i,j ); // velocity vector of generation g. 10 ; 11 Xg i,j = Xg−1 i,j + V g i,j; // position vector of generation g. 12 ; 13 end 14 end 15 end 16 return Zg = (Xg, V g); Algorithm 5: Update of the reference points. Input : Reference points Z = (X, V ), ρ Output: Updated reference points Zg 1 Calculate ﬁtness fit(Zi)= ρi (number of individuals associated for each Zi ∈Z); 2 Calculate w = wmax −g · (wmax −wmin)/gmax; ; // inertia weight of generation(g), gmax is the maximum generation. 5.3.1 Design of experiment The experimental design is carried out to evaluate the GP-A-NSGA-III (PSO) and GP-NSGA-III approaches. In the experimental studies, we compared the performance of GP-A-NSGA-III(PSO) with the baseline GP- NSGA-III. The JSS benchmark, the Taillard (TA) static JSS benchmark in- stances (see in Section 3.3 of Chapter 3) is selected as a testbed. In the ex- periments, we considered four potentially conﬂicting objectives:the mean ﬂowtime (mF) (see equation (2.1) of Chapter 2) , maximal ﬂowtime (maxF) (see equation (2.2) of Chapter 2), mean weighted tardiness (mWT) (see 5.3. MODEL-FREE ADAPTIVE REFERENCE POINTS GENERATION 113 equation (2.6) of Chapter 2) and maximal weighted tardiness (maxWT) (see equation (2.7) of Chapter 2). The terminal set and function set are described in Table 3.2 of Chapter 3. Both GP-A-NSGA-III(PSO) and GP-NSGA-III adopt the GP representa- tion (tree-based) and evolutionary operators (e.g., initialization, crossover, and mutation). For both compared algorithms, the population size is set to 1025. The crossover, mutation and reproduction rates are set similar to Subsection 4.4.1 of Chapter 4. The maximal number of generations (gmax) is set to 51. For the PSO parameters, we set c2 = 2, wmin = 0.4 and wmax = 0.9, which are standard settings used in many existing studies [100]. In the experiments, we again use IGD [206] and HV [212] to com- pare the algorithms (The detailed information can be seen in Section 3.3 Chapter 3). 5.3.2 Results and discussions During the GP search process, a rule is evaluated on the 40 training in- stances. The ﬁtness function for each objective is deﬁned as the average normalized objective value of the schedule obtained by applying that rule to each of the 40 training instances. For each algorithm, 30 GP runs ob- tained 40 ﬁnal dispatching rules. Then, the rules were tested on the 40 test instances. Overall results Table 5.1 shows the mean and standard deviations of the test performance (HV and IGD) of the rules obtained by GP-NSGA-III and GP-A-NSGA- III(PSO). In addition, for each test instance, the Wilcoxon rank-sum test with the signiﬁcance level of 0.05 was conducted separately on both the HV and IGD of the rules obtained by the two compared algorithms. That is, if the p-value is smaller than 0.05, then the best algorithm is considered signiﬁ- cantly better than the other algorithm. The signiﬁcantly better results are marked in bold. CHAPTER 5. REFERENCE POINTS ADAPTATION METHODS 114 Table 5.1: The mean and standard deviation over the HV and IGD values on the test instances of the compared algorithms in the 4-obj experiment. Overall results Problem Instances HV IGD ID #J #M GP-NSGA-III GP-A-NSGA-III(PSO) GP-NSGA-III GP-A-NSGA-III(PSO) 1 15 15 .0096(.0145) .2414(.0368) .0265(.0008) .0088(.0113) 2 15 15 .1600(.0182) .1666(.0110) .0254(.0005) .0218(.0004) 3 15 15 .0806(.0125) .0906(.0127) .0205(.0008) .0292(.0006) 4 15 15 .1263(.0183) .0692(.0091) .01765(.0007) .0218(.0006) 5 15 15 .1661(.0202) .1752(.0192) .0190(.0005) .0237(.0007) 6 20 15 .1271(.0176) .1421(.0139) .0154(.0004) .0211(.0004) 7 20 15 .2488(.0683) .2538(.0695) .0079(.0013) .0077(.0014) 8 20 15 .1115(.0201) .2015(.0253) .0186(.0006) .01819(.0026) 9 20 15 .1700(.0148) .1839(.0219) .0170(.0005) .0149(.0005) 10 20 15 .1086(.0151) .3580(.0366) .0156(.0002) .0160(.0003) 11 20 20 .0114(.0038) .1087(.0118) .0299(.0006) .0240(.0008) 12 20 20 .0852(.0133) .1206(.0134) .0173(.0004) .0222(.0006) 13 20 20 .1540(.0151) .1569(.0326) .0188(.0003) .0137(.0002) 14 20 20 .0504(.01103) .0704(.0118) .0328(.0008) .0272(.0007) 15 20 20 .3985(.0225) .2454(.0199) .0109(.0004) .0150(.0002) 16 30 15 .2465(.030) .2375(.0234) .0144(.0006) .0140(.0002) 17 30 15 .1813(.0096) .2278(.0229) .0138(.0003) .0097(.0004) 18 30 15 .3198(.0233) .1957(.0132) .0131(.0004) .0152(.0003) 19 30 15 .2789(.0126) .3004(.0197) .0150(.0004) .0114(.0004) 20 30 15 .2575(.0312) .2124(.0312) .0144(.0005) .0195(.0003) 21 30 20 .1347(.0657) .2325(.0792) .0135(.0022) .0098(.0014) 22 30 20 .2365(.0472) .3027(.0470) .0065(.0010) .0052(.0006) 23 30 20 .2944(.0398) .2984(.0410) .0046(.00004) .0045(.0005) 24 30 20 .3812(.0503) .6161(.0174) .0070(.0009) .0018(.0004) 25 30 20 .5199(.0477) .5290(.0396) .0059(.0012) .0046(.0005) 26 50 15 .4563(.0417) .4872(.0270) .0051(.0009) .0039(.0004) 27 50 15 .5710(.0361) .5685(.0304) .0040(.0009) .0032(.0003) 28 50 15 .4598(.0398) .4966(.0250) .0049(.0010) .0036(.0003) 29 50 15 .4862(.0372) .5125(.0251) .0045(.0007) .0037(.0003) 30 50 15 .4510(.0406) .4732(.0240) .0033(.0005) .0026(.0002) 31 50 20 .5085(.0424) .5147(.0295) .0053(.0008) .0042(.0004) 32 50 20 .4378(.0476) .4266(.0375) .0046(.0006) .0041(.0004) 33 50 20 .3422(.0266) .4383(.0838) .0125(.0006) .0069(.0051) 34 50 20 .3828(.0384) .4089(.0262) .0036(.00005) .0030(.00029) 35 50 20 .5558(.0349) .5763(.0165) .0025(.0005) .0020(.0001) 36 100 20 .3648(.0179) .2972(.0093) .010(.0003) .0130(.0006) 37 100 20 .3442(.0142) .2844(.0101) .0086(.0003) .0107(.0003) 38 100 20 .3006(.0196) .3025(.0136) .0067(.0009) .0066(.0002) 39 100 20 .6495(.0185) .6515(.0191) .0010(.0001) .0010(.0001) 40 100 20 .3658(.0158) .3828(.0100) .0085(.0003) .0088(.0002) Table 5.1 reveals that GP-A-NSGA-III(PSO) performed signiﬁcantly better than GP-NSGA-III in most of the test instances. In the case of HV, GP-A-NSGA-III(PSO) performed signiﬁcantly better in 24 out of 40 test instances. On the other hand, GP-NSGA-III performed signiﬁcantly bet- 5.3. MODEL-FREE ADAPTIVE REFERENCE POINTS GENERATION 115 ter only in 6 instances. For the remaining 10 instances, the two compared algorithms performed statistically the same. In regard to IGD, Table 5.1 ex- hibits the same pattern. GP-A-NSGA-III(PSO) achieved signiﬁcantly bet- ter performance in 21 out of the 40 test instances. In contrast, GP-NSGA-III only performed signiﬁcantly better in 13 instances. Overall results Table 5.1 also shows that GP-A-NSGA-III(PSO) not only performed better on smaller instances compared to the baseline algorithms GP- NSGA-III. Also, it is more effective in more challenging and larger in- stances. For some test instances (e.g., instances 1, 10, 24), GP-A-NSGA- III(PSO) achieves a signiﬁcant improvement. This demonstrates the use- fulness of the proposed adaptive reference point scheme, which can ﬁnd a better association with population members and obtain well-distributed reference points. Further analysis To further investigate how the adaptive reference point scheme affects the GP search process, we plot the average number of useless reference points (those associated with no individual in the population). Moreover, we plot the average HV and IGD of the non-dominated solutions obtained for each generation during the 30 independent runs of the two compared algorithms, as given in Figures 5.3 and 5.4, respectively. Figure 5.4 shows the convergence curves of the HV and IGD values of the non-dominated solutions on the training set. Figure 5.4 shows that the adaptive reference point scheme can signiﬁcantly reduce the number of useless points during the GP search process. Without adaptive points, the number of useless references points in GP-NSGA-III increased from 910 to about 980. On the contrary, in GP-A-NSGA-III(PSO), the number of useless reference points ﬁrst increased and then decreased to 810. This ﬁgure reveals that the adaptive reference point scheme at the later stage of the search led to less useless reference points. Thus, it was a better reﬁnement of the densely populated regions of the population. Figure 116 CHAPTER 5. REFERENCE POINTS ADAPTATION METHODS 0 10 20 30 40 50 60 Generation 800 820 840 860 880 900 920 940 960 980 # of Useless Points Uniform Reference Points Adaptive Reference Points Figure 5.3: The curves of the average number of useless reference points in GP-NSGA-III, GP-A-NSGA-III(PSO). 30 Generation Figure 5.3: The curves of the average number of useless reference points in GP-NSGA-III, GP-A-NSGA-III(PSO). Figure 5.4: The curves of the HV and IGD values of the non-dominated solutions on the training set during the 30 independent GP runs. Figure 5.4: The curves of the HV and IGD values of the non-dominated solutions on the training set during the 30 independent GP runs. 5.4 reveals that the GP-A-NSGA-III obtained better convergence curves in terms of both HV and IGD. This ﬁgure reveals that the reduction of useless reference points can lead to better non-dominated sets. Figures 5.5 (a) to 5.6 (b) show the distribution of the reference points and the population’s ﬁtness values in generations 1 and 50 of GP-ANSGA- 5.3. MODEL-FREE ADAPTIVE REFERENCE POINTS GENERATION 117 MF MaxWT MaxF MWT 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (a) Distribution of the reference points. MF MaxWT MaxF MWT 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (b) Distribution of the ﬁtness values. Further analysis MF MaxWT MaxF MWT 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (a) Distribution of the reference points. (b) Distribution of the ﬁtness values. Figure 5.5: Parallel coordinate plots of GP-A-NSGA-III(PSO) at genera- tions 1. Figure 5.5: Parallel coordinate plots of GP-A-NSGA-III(PSO) at genera- tions 1. MF MaxWT MaxF MWT 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values MF MaxWT MaxF MWT 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (a) Distribution of the reference points. MF MaxWT MaxF MWT 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (b) Distribution of the ﬁtness values. MF MaxWT MaxF MWT 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (a) Distribution of the reference points. (b) Distribution of the ﬁtness values. Figure 5.6: Parallel coordinate plots of GP-A-NSGA-III(PSO) at genera- tions 50. III(PSO). It can be seen that in generation 1, the reference points are close to the initial uniform distribution, and the ﬁtness distribution of the pop- ulation is relatively uniform as well. On the other hand, at generation 50, the distributions of the reference points and the population’s ﬁtness val- CHAPTER 5. REFERENCE POINTS ADAPTATION METHODS 118 ues become very similar. This is consistent with our expectation which is to use a similar distribution of reference points as that of the population to ﬁne-tune the promising area around the Pareto-front. From this study, we identify a key research issue of having non- uniform Pareto-front, i.e., the simple adoption of uniformly distributed reference points failed to promote solution diversity during evolution and affected the performance of algorithms. In conclusion, the proposed reference point adaptation mechanism has the potential to decrease useless reference points. It can signiﬁcantly im- prove the performance of GP-HH and NSGA-III in terms of both HV and IGD. However, this research direction requires further investigation to re- duce more useless reference points and improve the quality of evolved rules. 5.4.1 Reference point Adaptation by density-based model Our proposed algorithm (GP-NSGA-III with density model-based refer- ence point adaptation (GP-NSGA-III-DRA)) can be considered as a major enhancement of GP-NSGA-III. GP-NSGA-III-DRA starts after the popu- lation update mechanism of GP-NSGA-III which can be seen in Subsec- tion 4.3.2 of Chapter 4. GP-NSGA-III-DRA utilized the Pg+1 population from GP-NSGA-II and predeﬁned simplex locations W. 5.4 Model-based adaptive reference points gen- eration In the literature, NSGA-III has been extended to add and delete reference points in an adaptive manner [85, 86]. A-NSGA-III is an extended version of NSGA-III and has implicitly guided the distribution of solutions. A- NSGA-III ﬁrst generates the uniformly distributed reference points then adds and removes these points in a high dimension space. However, in a high dimension space, the removal condition (ρ is less than one) is difﬁcult to achieve. The algorithm keeps on including additional reference points that clearly affect the algorithm’s performance. Most of the adaptive refer- ence point approaches cannot explicitly construct the distribution model. The model’s construction helps the concrete knowledge of the Pareto-front that can provide a close match between the reference points with the dis- tribution of Pareto-optimal solutions. Our model-based approach in this thesis adopts a modeling technique .4. MODEL-BASED ADAPTIVE REFERENCE POINTS GENERATION 119 [30] that learns the distribution of the Pareto-optimal solution and gener- ates the reference points according to the solution distribution. This sec- tion introduces the model with and without the Gaussian process model [30, 170] (see in Subsection 2.1.8 of Chapter 2). The model without the Gaussian model is called a density-based model that estimates the density of solutions from each pre-deﬁned sub-region. After the density-based model, we build a Gaussian process-based probabilistic model which can reduce the density noise and provides a reliable approximation of the true shape of the Pareto-front. In this manner, this section will show the model effectiveness of the density-based model and the Gaussian process model. Formation of density-based model The density-based model estimates the density of solutions at each sub- location (a location on a normalized hyperplane) w ∈W. Building this density-based probabilistic model consists of two steps. First, the density- based model evenly decomposes the simplex into several sub-locations w1, w2, w3.., wk ∈W. This decomposition uses Das and Dennis’s [39] sys- tematic approach (The detailed information can be seen in Subsection 2.1.6 of Chapter 2). Then the association between each solution s with w is ob- tained based on their perpendicular distance (⊥). As a result, a solution is associated with a sub-location( ˆw) where the perpendicular distance be- tween the s with w has a minimum value ˆw. The solutions associated with CHAPTER 5. REFERENCE POINTS ADAPTATION METHODS 120 ˆw is recorded in archive E( ˆw). ˆw is recorded in archive E( ˆw). The number of associated solutions in any sub-location ˆw is obtained by dividing the number of the associated solutions in E( ˆw) by the total number of non-dominated solutions (∥S ∥). The density-based proba- bilistic model is therefore deﬁned as P(D\| ˆw ∈W) = ∥(E( ˆw)) ∥ ∥S ∥ , (5.2) (5.2) The formation of the density model is shown in Algorithm 6. Figures 5.7 and 5.8 show an example to illustrate the above procedure of formation of density model. First, solutions are associated with their closest refer- ence points, which is presented in Figure 5.7. For instance, three solutions are closest to w3. Then the algorithm calculates the density of solutions at each sub-simplex location and builds a density-based model using equa- tion (5.2). This can be seen in Figure 5.8. Figure 5.7: Solutions are closest to the reference points. Figure 5.7: Solutions are closest to the reference points. 5.4. MODEL-BASED ADAPTIVE REFERENCE POINTS GENERATION 121 Algorithm 6: Construct DensityModel(Sg, W ) Input : Sg, W Output: Z∗ g 1 foreach w ∈W do 2 E(w) = ∅; 3 D(w) = ∅; 4 end 5 foreach s ∈Sg do 6 foreach w ∈W do 7 compute d⊥(s, w); // perpendicular distance of each solution from w 8 end 9 Assign ˆw = argmins∈Sd⊥(s, w) ; // associate the solution with the sub-location w 10 Save s in E( ˆw); 11 end 12 for i= 1 to ∥E( ˆw) ∥do 13 Assign P(D\| ˆw) =∥E( ˆw)i ∥÷∥S ∥; // probability of the associated solution 14 Assign D( ˆw)=∥P( ˆw) ∥length of reference points; // return solution’s density 15 end 16 Z∗ g= Generate (E( ˆw), D( ˆw), Sg, W); 17 return Z∗ g; Algorithm 6: Construct DensityModel(Sg, W ) Algorithm 6: Construct DensityModel(Sg, W ) Algorithm 6: Construct DensityModel(Sg, W ) Reference points generation Our proposed algorithm is broken into two sub-routines: (1) references points in proximity to the vertex and (2) the internal reference points. The generation of reference points is activated in line 16 of Algorithm 6. 1. References points close to the vertex: This method of the proposed algorithm handles the issue of A-NSGA-III [85] which relates to the generation of the reference points close to the vertices of the simplex. Our proposed algorithm generates reference points in close proxim- ity to the simplex vertices. As a result, our algorithm enhances the Our proposed algorithm generates reference points in close proxim- ity to the simplex vertices. As a result, our algorithm enhances the 122 CHAPTER 5. REFERENCE POINTS ADAPTATION METHODS Figure 5.8: Density of solution at each sub-location of the simplex. Figure 5.8: Density of solution at each sub-location of the simplex. Figure 5.8: Density of solution at each sub-location of the simplex. ability to generate reference points that match the distribution model closely. The detailed procedure is explained below: (a) Obtain the center location (centroid) from the existing solutions in the sub-location ˆw ∈W where ˆw is one of the locations close to the vertices of the hyperplane. (b) Calculate the perpendicular distance from the centroid to asso- ciated solutions of ˆw. (c) Select a solution s based on a minimum perpendicular distance. (d) Calculate a mid-point value between the selected solution and the central location for generating a corresponding reference point. This mid-point of each dimension is considered as one of the reference points around the vertices. (e) Selected solution s will not be the part of a centroid. Thus, the 5.4. Reference points generation REFERENCE POINTS ADAPTATION METHODS reference points generated by the following two equations: pointsi = Zj −Zp, (5.3) (5.3) Zi new = pointsi + (Interval)/M. (5.4) (5.4) where the interval is the difference between two consecutive refer- ence points on the hyperplane. The above two equations kept the j-th internal locations at the center of the newly generated reference point. These newly generated refer- ence points can be inserted in the reference points archive called Znew if they satisfy the two main conditions: (i) a reference point must be inside the entire simplex boundary; (ii) every reference point must occupy a unique location on the simplex. Once new reference points are added into archive Znew, then the association between existing members of Znew and solutions in E( ˆw) must be checked. If the i-th reference point from Znew still has ρi ≥2, again new reference points are generated around the i-th reference points, but this time, the in- terval’s parameter value is halved. Therefore, new reference points are getting closer to the associated solutions with the original refer- ence points and increase the chance of getting an ideal association (ρ =1). This process is also shown in Figure 5.9. Figure 5.9 demon- strates that the ith reference point is kept as a centroid location for newly generated reference points and the reference points are lay- ered. These layers are also shown in Figure 5.9 with two different colors (blue and black lines). Thus, we named this method a layered centroid approach. Reference points generation MODEL-BASED ADAPTIVE REFERENCE POINTS GENERATION 123 Algorithm 7: E( ˆw), D( ˆw), Sg, W Input : E( ˆw), D( ˆw), Sg, W Output: Z∗ g 1 foreach ˆw ∈W do 2 set nref= ∥D( ˆw) ∥; // number of reference points required at location ˆw 3 Assign Zr= ˆw ; // set ˆw as a first reference point 4 if Zr!=Vertex Points then 5 Assign Z∗ g= IntermediatePoints(E( ˆw), D( ˆw), nref, Sg, W, Zr) ; // call intermediate points method 6 end 7 if Vertex points then 8 Assign Z∗ g= V ertexPoints(E( ˆw), D( ˆw), nref, Sg, W, Zr); // call vertex points method 9 end 10 end 11 return Z∗ g; Algorithm 7: E( ˆw), D( ˆw), Sg, W Input : E( ˆw), D( ˆw), Sg, W Output: Z∗ g 1 foreach ˆw ∈W do 2 set nref= ∥D( ˆw) ∥; // number of reference points required at location ˆw 3 Assign Zr= ˆw ; // set ˆw as a first reference point 4 if Zr!=Vertex Points then 5 Assign Z∗ g= IntermediatePoints(E( ˆw), D( ˆw), nref, Sg, W, Zr) ; // call intermediate points method 6 end 7 if Vertex points then 8 Assign Z∗ g= V ertexPoints(E( ˆw), D( ˆw), nref, Sg, W, Zr); // call vertex points method 9 end 10 end 11 return Z∗ g; next centroid is calculated from existing solutions that are still in the race of acquiring reference points. (f) Repeat steps (a) to (e) until the required number of reference points are generated based on the size of the number of associ- ated solutions with ˆw. 2. Generation of reference points internal to the simplex: An exam- ple below shows the reference points generation internal to the sim- plex. Consider the j-th internal simplex location which is associated with more than one solution. If the situation has M = 3 objectives, then M points are generated at the vertices of j-th internal locations on the simplex. The simplex side length (interval) is equal to the dis- tance between two neighboring internal locations on the originally speciﬁed hyperplane and store in the archive Zp. This example is shown in Figure 5.9. In this example, Z1, Z2, Z3 ∈Znew reference points are generated around the j-th internal locations (Zj ∈Zr). The 124 CHAPTER 5. CHAPTER 5. REFERENCE POINTS ADAPTATION METHODS 126 5.5 Gaussian process-based probabilistic model In Algorithm 6, we deﬁned the density-based model which is built to the model on the solution’s density at each location of the simplex. This sec- 5.5. GAUSSIAN PROCESS-BASED PROBABILISTIC MODEL 125 Figure 5.9: Generate reference points until M −1 times Figure 5.9: Generate reference points until M −1 times tion predicts a mean value in sub-location for generating reference points and subsequently reducing the density noise. Gaussian processes (GaP) (The detailed information can be seen in Subsection 2.1.8 of Chapter 2) as a powerful method to model unknown functions. In particular, in this research objective, we used GaP-based modeling. The GaP is deﬁned as a mean and a covariance function. f(x) ∼GaP(µ(x), K(xi, xj)), (5.5) (5.5) K deﬁnes the shape of GaP and the choice of an appropriate kernel. The kernels’ selection is based on assumptions such as patterns to be expected in the data. However, in this study, our assumption provides smoothness to the model which can reduce the noise of the model. Next, we calculate the area under the mean function. This area under the curve helps us to ﬁnd the required number of reference points in each sub-location. In our proposed algorithm, model-based adaptive reference points (GP-MARP-NSGA-III), we use the squared exponential kernel [170]. In this kernel, if xi and xj behave similarly then the function values at these points, f(xi) and f(xj), can be expected to be similar. The squared expo- nential kernel is also known as the Gaussian kernel. K (xi, xj) = σ2 f exp −1 2l \|xi −xj\|2 . (5.6) (5.6) where σ2 f and l are hyperparameters of the covariance functions. CHAPTER 5. REFERENCE POINTS ADAPTATION METHODS 5.5.1 Modelling by Gaussian process In GP-MARP-NSGA-III, ﬁrst, we construct the density model. The details of the construction of the density model of the simplex are described in Subsection 5.4.1. However, we construct the model by using mid-point location(wc) of the two neighborhoods ˆwi ∈W and ˆwj ∈W on simplex. The reason for selecting mid-point locations because, in later this algo- rithm, we approximate areas under the mean function by using Riemann sum with the midpoints [48]. For the model construction, we need to ﬁnd two sub-locations ˆwi and ˆwi those are close in perpendicular distance with each other. These neigh- boring locations of ˆwi are further saved in an archive Negh(wi). Next, the mid-point between ˆwi and ˆwj is determined and stored in Wc. This center location at wc ∈Wc serves as an input of our density-based model. Then, the model based on the association between the center locations and their respective solutions is stored in E( ˆwc). The detailed information can be seen in Algorithm 8. Speciﬁcally, the total number of the associated solutions with ˆwc ( ˆwc ˆwc means minimum perpendicular distance with population members) is recorded in archive D( ˆwc). In accordance with D( ˆwc), the solution density of each sub-location ˆwc is calculated by the associated solutions in E( ˆwc) divided by the total number of non-dominated solutions (∥S ∥) of each generation. The density-based probabilistic model(P(D\| ˆwc ∈Wc)) is de- ﬁned in Equation (5.2). The GaP is a sample of a stochastic process. In this way, P(D\| ˆwc ∈ Wc) can be seen as a latent function [170] where the joint distribution of a inﬁnite number of these variables P(D(wc1), . . . , D(wck)) is itself Gaussian: P(D\| ˆwc ∈Wc) = N(D\|µ, K), (5.7) (5.7) After the construction of the density model, we train the GaP means to estimate the conditional probability of P(D\| ˆw ∈W) which is fully speci- ﬁed by the mean function µ(wi) and K that calculates the covariance be- 5. GAUSSIAN PROCESS-BASED PROBABILISTIC MODEL 5.5. 127 tween any two sub-locations(wci, wcj). Without the loss of generality, we set zero to the prior mean function common in practice. P(D\| ˆwc ∈Wc) = N(D\|0, K), (5.8) (5.8) With equation (5.8), we make predictions D∗of given inputs Wc∗from the posterior distribution P(D∗\| ˆwc ∈Wc, D, ˆ wc∗∈Wc∗) .The posterior dis- tribution can be obtained as a Gaussian distribution with mean and vari- ance. 5.5.1 Modelling by Gaussian process REFERENCE POINTS ADAPTATION METHODS Algorithm 8: Gaussion Modelling(Sg, W ) Input : Sg, W Output: E( ˆ wc) ( at ˆ wc) &µ( ˆ wc) (mean prediction at ˆ wc ) 1 foreach wi ∈W do 2 foreach wj ∈W do 3 compute d⊥(wi, wj) ; // find neighbors of wi 4 end 5 Assign ˆwneg = argminwj∈W d⊥(wi, wj) Save ˆwneg in Neg(wi); ; // save neighbors of wi location in Neg(wci) 6 end 7 foreach wneg ∈Neg(wi) do 8 Calculate the center location wc between wneg ∈Neg(wi) and wi ; // find center loction between wi and wneg location in Neg(wi) 9 Save wci in E(wci) ; // save center location in E(wci) 10 end 11 foreach s ∈Sg do 12 foreach wci in E(wci) do 13 compute d⊥(s, wci) 14 end 15 Assign ˆ wc = argmins∈Sd⊥(s, wci) ; // associate the solution with the center location 16 Save s in A( ˆ wci) 17 end 18 foreach s ∈A( ˆ wci) do 19 Calculate the number of associated solutions with ˆ wci and store in D( ˆ wci) 20 end 21 Assign P(D\| ˆ wci) =∥A( ˆ wci) ∥÷∥S ∥; // build a density model 22 Train a Gaussian process on density-based model.P(D\| ˆ wci) ∼N(0, K) 23 Test input ˆ wc∗and obtain the Gaussian distribution with mean µ( ˆ wc∗) return E( ˆ wc) &µ( ˆ wc) Algorithm 8: Gaussion Modelling(Sg, W ) Input : Sg, W Output: E( ˆ wc) ( at ˆ wc) &µ( ˆ wc) (mean prediction at ˆ wc ) 1 foreach wi ∈W do 2 foreach wj ∈W do 3 compute d⊥(wi, wj) ; // find neighbors of wi 4 end 5 Assign ˆwneg = argminwj∈W d⊥(wi, wj) Save ˆwneg in Neg(wi); ; // save neighbors of wi location in Neg(wci) 6 end 7 foreach wneg ∈Neg(wi) do 8 Calculate the center location wc between wneg ∈Neg(wi) and wi ; // find center loction between wi and wneg location in Neg(wi) 9 Save wci in E(wci) ; // save center location in E(wci) 10 end 11 foreach s ∈Sg do 12 foreach wci in E(wci) do 13 compute d⊥(s, wci) 14 end 15 Assign ˆ wc = argmins∈Sd⊥(s, wci) ; // associate the solution with the center location 16 Save s in A( ˆ wci) 17 end 18 foreach s ∈A( ˆ wci) do 19 Calculate the number of associated solutions with ˆ wci and store in D( ˆ wci) 20 end 21 Assign P(D\| ˆ wci) =∥A( ˆ wci) ∥÷∥S ∥; // build a density model 22 Train a Gaussian process on density-based model.P(D\| ˆ wci) ∼N(0, K) 23 Test input ˆ wc∗and obtain the Gaussian distribution with mean µ( ˆ wc∗) return E( ˆ wc) &µ( ˆ wc) Algorithm 8: Gaussion Modelling(Sg, W ) Input : Sg, W Output: E( ˆ wc) ( at ˆ wc) &µ( ˆ wc) (mean prediction at ˆ wc ) 1 foreach wi ∈W do 2 foreach wj ∈W do 3 compute d⊥(wi, wj) ; // find neighbors of wi 4 end 5 Assign ˆwneg = argminwj∈W d⊥(wi, wj) Save ˆwneg in Neg(wi); ; // save neighbors of wi location in Neg(wci) 6 end 7 foreach wneg ∈Neg(wi) do 8 Calculate the center location wc between wneg ∈Neg(wi) and wi ; // find center loction between wi and wneg location in Neg(wi) 9 Save wci in E(wci) ; // save center location in E(wci) 10 end 11 foreach s ∈Sg do 12 foreach wci in E(wci) do 13 compute d⊥(s, wci) 14 end 15 Assign ˆ wc = argmins∈Sd⊥(s, wci) ; // associate the solution with the center location 16 Save s in A( ˆ wci) 17 end 18 foreach s ∈A( ˆ wci) do 19 Calculate the number of associated solutions with ˆ wci and store in D( ˆ wci) 20 end 21 Assign P(D\| ˆ wci) =∥A( ˆ wci) ∥÷∥S ∥; // build a density model 22 Train a Gaussian process on density-based model.P(D\| ˆ wci) ∼N(0, K) 23 Test input ˆ wc∗and obtain the Gaussian distribution with mean µ( ˆ wc∗) return E( ˆ wc) &µ( ˆ wc) 5.5.1 Modelling by Gaussian process P(D∗\| ˆwc ∈Wc, D, ˆ wc∗∈Wc∗)) = N(D∗\|µ∗, K∗). (5.9) (5.9) The detailed information of the GaP can be seen from the Subsection 2.1.8 of Chapter 2. The detailed process of GaP modeling is shown in Algo- rithm 8 and in Figure 5.10. In this Figure, we show Gaussian process model and get a smooth function by reducing model-noise. This function helps us to calculate the area under a mean function of the GaP. Figure 5.10: Train a Gaussian process on density-based model. Figure 5.10: Train a Gaussian process on density-based model. 128 CHAPTER 5. 5.5.2 Calculate the area under the mean function In this method, we approximately calculate the area under the mean- function through the mid-point rule. This area under the mean-function helps us to ﬁnd the required number of the reference points in each sub- 5.5. GAUSSIAN PROCESS-BASED PROBABILISTIC MODEL 129 Algorithm 9: Calculate Area(µ( ˆ w), , E( ˆ wc)) Input : µ( ˆ wc), E( ˆ wc) Output: Z( ˆ wc) 1 foreach ˆ wc ∈E( ˆ wc) do 2 Assign area = b−a n ∗(µ(wc)) 3 Save area in Ar(wc) 4 end 5 Total Area: R b a µ( ˆ wc)dx = b−a n (µ(wc1) + µ(wc2) + µ(wc3) + · · · + µ(wn)) = Ar(wc1) + Ar(wc2) + · · · + Ar(wcn)) 6 foreach ˆ wc ∈Ar( ˆ wc) do 7 Calculate the number of reference points (nref ) in each ˆ wc : Ar(wc) T otalArea ; 8 Save nref in Z( ˆwc) ; // save number of reference points in Z( ˆwc) 9 end 10 return Z( ˆ wc); Algorithm 9: Calculate Area(µ( ˆ w), , E( ˆ wc)) Input : µ( ˆ wc), E( ˆ wc) Output: Z( ˆ wc) 1 foreach ˆ wc ∈E( ˆ wc) do 2 Assign area = b−a n ∗(µ(wc)) 3 Save area in Ar(wc) 4 end 5 Total Area: R b a µ( ˆ wc)dx = b−a n (µ(wc1) + µ(wc2) + µ(wc3) + · · · + µ(wn)) = Ar(wc1) + Ar(wc2) + · · · + Ar(wcn)) 6 foreach ˆ wc ∈Ar( ˆ wc) do 7 Calculate the number of reference points (nref ) in each ˆ wc : Ar(wc) T otalArea ; 8 Save nref in Z( ˆwc) ; // save number of reference points in Z( ˆwc) 9 end 10 return Z( ˆ wc); Algorithm 9: Calculate Area(µ( ˆ w), , E( ˆ wc)) location. We decided to use the mid-point rule for calculating the area under the mean function. This is deﬁned in equation (5.10) location. We decided to use the mid-point rule for calculating the area under the mean function. This is deﬁned in equation (5.10) TotalArea ≈ Z b a µ( ˆwc)dx = b −a n (µ(wc1) + µ(wc2) + · · · + µ(wn)), (5.10) (5.10) Next we have found the number of reference points in each speciﬁed area and they can be determined as below referencepoints = Ar(wc) TotalArea . 5.5.3 Reference points generation The generation of reference points is described in this section. For each ˆwc in the reference points archive, if the number of reference points, i.e., Z( ˆwc) has a niche count equal to one, there is no need to generate new ref- erence points. Otherwise, reference points are generated using the equa- tions (5.3) and (5.4). The rest of the procedure for the newly generated reference points is similar to the ”Generation of reference points internal to the simplex” of Subsection 5.4.1. CHAPTER 5. REFERENCE POINTS ADAPTATION METHODS CHAPTER 5. REFERENCE POINTS ADAPTATION METHODS 130 0 0.2 0.4 0.6 0.8 1 −1 0 1 2 3 4 5 Simplex Location Solution Density w2 w3 w4 w5 w6 wc3 wc2 wc1 b a Simplex locations w1 U(wc2) U(wc3) U(wc1) Figure 5.11: Area under the mean function of the Gaussian process model Figure 5.11: Area under the mean function of the Gaussian process model 5.5.2 Calculate the area under the mean function (5.11) (5.11) where Ar(wc) is a speciﬁc area under the curve. The detailed process for area calculation is further presented in Algorithm 9. This can also be seen in Figure 5.11. Next, the algorithm estimates the number of refer- ence points at each sub-location wc using equation (5.11) and generates the required reference points in each sub-area. mean µ (line 23) is O(N). All operations in Algorithm 9 (lines 2, 5, and 8 ) in calculating area and calculating the number of reference points would require O(N) computa- tions. Thereafter, the reference points generation (the procedure shows in Algorithm 7) requires O(MN) comparisons. Considering all the above considerations and computations, the over- all worst-case complexity of one generation of GP-MARP-NSGA-III is O(MN 3). 5.5.4 Computational Complexity of One Generation of GP-MARP-NSGA-III In line 3 of Algorithm 8, we are computing a neighboring location of W, where its size is equal to the size of population N. So, computing a neigh- boring location takes O(MN 2) where M is the objective number. After that, the center location of two neighboring locations of W (line 8 of Algo- rithm 8) requires O(MN). In line 13 of Algorithm 8, associating a popula- tion of N individuals to N reference directions takes O(MN 2). However, to build a density model (line 21) requires only O(N). In addition, to train a GaP on the density-based model (line 22) O(MN 3), which totally depends on the kernel (KNxN). Finally, the computational complexity of obtaining . GAUSSIAN PROCESS-BASED PROBABILISTIC MODEL 131 5.5.5 Design of experiment To verify the effectiveness of the proposed model-driven reference points, we compare the performance of our algorithm GP-MARP-NSGA-III and GP-NSGA-III-DRA with the GP-NSGA-III, GP-A-NSGA-III in the exper- imental studies. We selected the Taillard (TA) static JSS benchmark in- stances [182] as the testbed. The details of TA can be found in Section 3.1 of Chapter 3. In the experiments, we considered four potentially conﬂicting objectives:the mean ﬂowtime (mF) (see equation (2.1) of Chapter 2) , max- imal ﬂowtime (maxF) (see equation (2.2) of Chapter 2), mean weighted tardiness (mWT) (see equation (2.6) of Chapter 2) and maximal weighted tardiness (maxWT) (see equation (2.7) of Chapter 2). For all the compared algorithms, 30 independent runs were conducted. All the compared algorithms adopt the tree-based representation of dispatching rules. The terminal set and function set are described in Sec- tion 3.2 of Chapter 3. GP-MARP-NSGA-III, GP-NSGA-III-DRA and GP- A-NSGA-III adopt the same parameter setting of the baseline algorithm GP-NSGA-III. In the experiments, the two commonly used measures, i.e. IGD [206] and HV [212] are used to compare the algorithms (see in Section 3.3 of Chapter 3). CHAPTER 5. REFERENCE POINTS ADAPTATION METHODS 132 Table 5.2: The mean and standard deviation of HV and IGD values of the 30 independent runs on training instances of the compared algorithms on four-objective JSS problems. HV Statistic GP-NSGA-III GP-A-NSGA-III GP-NSGAIII-DRA GP-MARP-NSGA-III (¯x ± σ) 0.67234(0.02245) 0.67561(0.02043) 0.67655(0.01955) 0.68498(0.01961) IGD Statistic GP-NSGA-III GP-A-NSGA-III GP-NSGA-III-DRA GP-MARP-NSGA-III (¯x ± σ) 0.00139(0.00012) 0.00131(0.00012) 0.00130(0.00012)) 0.00127(0.00008) HV Statistic GP-NSGA-III GP-A-NSGA-III GP-NSGAIII-DRA GP-MARP-NSGA-III (¯x ± σ) 0.67234(0.02245) 0.67561(0.02043) 0.67655(0.01955) 0.68498(0.01961) IGD Statistic GP-NSGA-III GP-A-NSGA-III GP-NSGA-III-DRA GP-MARP-NSGA-III (¯x ± σ) 0.00139(0.00012) 0.00131(0.00012) 0.00130(0.00012)) 0.00127(0.00008) formed signiﬁcantly better than GP-A-NSGA-III and GP-NSGA-III. formed signiﬁcantly better than GP-A-NSGA-III and GP-NSGA-III. Table 5.3 summarizes the testing performance of all algorithms in terms of IGD and HV. The obtained test results exhibit the same patterns as the training performance results. In the case of HV, GP-MARP-NSGA- III performed signiﬁcantly better than the other compared algorithms. In the case of IGD, GP-MARP-NSGA-III is competitive with GP-A-NSGA-III. However, both our proposed model-driven algorithms performance is still signiﬁcantly better in IGD than the GP-A-NSGA-III and GP-NSGA-III. 0 5 10 15 20 25 30 35 40 45 50 Generation 0.3 0.4 0.5 0.6 0.7 0.8 HV Values HV GP-NSGA-III GP-A-NSGA-III GP-NSGA-III-DRA GP-MARP-NSGA-III Figure 5.12: HV values of the non-dominated solutions on the training set during the 30 independent GP runs. Figure 5.12: HV values of the non-dominated solutions on the training set during the 30 independent GP runs. 5.5.6 Results and discussion For each algorithm, 30 GP runs have been performed to obtain 30 ﬁnal sets of dispatching rules. Afterward, the rules are tested on the 40 test instances. Tables 5.2 and 5.3 show the mean and standard deviation of the average HV and average IGD values obtained by GP-NSGA-III, GP- A-NSGA-III, GP-NSGA-III-DRA, and GP-MARP-NSGA-III. The Wilcoxon rank-sum test [195] with the signiﬁcance level of 0.05 has been applied separately to compare both the HV and IGD achieved by all algorithms. The signiﬁcantly better results are bolded. Table 5.3: The mean and standard deviation of HV and IGD achieved by all competing algorithms on test instances on four-objective JSS problems. HV Statistic GP-NSGA-III GP-A-NSGA-III GP-NSGAIII-DRA GP-MARP-NSGA-III (¯x ± σ) 0.47422(0.024231) 0.48896(0.02336) 0.48625(0.02445) 0.49844(0.02147) IGD Statistic GP-NSGA-III GP-A-NSGA-III GP-NSGA-III-DRA GP-MARP-NSGA-III (¯x ± σ) 0.00177(0.00020) 0.00162(0.00025) 0.00164(0.00027) 0.00161(0.00016) Table 5.3: The mean and standard deviation of HV and IGD achieved by all competing algorithms on test instances on four-objective JSS problems. Table 5.2 reveals that GP-MARP-NSGA-III achieved signiﬁcantly better performance in HV than the other algorithms. In terms of IGD, GP-MARP- NSGA-III clearly outperformed GP-A-NSGA-III, GP-NSGA-III-DRA, and GP-NSGA-III. Moreover, our proposed algorithm GP-NSGA-III-DRA per- GAUSSIAN PROCESS-BASED PROBABILISTIC MODEL 133 5.5. Further analysis For further analysis of the algorithms’ performance, we plotted the aver- age HV and IGD of the non-dominated solutions obtained so far for each generation during the 30 independent runs of the four compared algo- rithms. Figures 5.12 and 5.13 show convergence curves on the training sets. Figures 5.12 and 5.13 reveal that during the ﬁrst few generations of evolu- tion, all algorithms exhibited similar HV and IGD values. However, GP- MARP-NSGA-III starts to outperform other competing algorithms. More- over, when the solutions are very close to the Pareto-front, GP-MARP- CHAPTER 5. REFERENCE POINTS ADAPTATION METHODS CHAPTER 5. REFERENCE POINTS ADAPTATION METHODS CHAPTER 5. REFERENCE POINTS ADAPTATION METHODS 134 0 5 10 15 20 25 30 35 40 45 50 Generation 0 0.5 1 1.5 IGD Values 10-3 IGD GP-NSGA-III GP-A-NSGA-III GP-NSGA-III-DRA GP-MARP-NSGA-III Figure 5.13: IGD values of the non-dominated solutions on the training set during the 30 independent GP runs. Figure 5.13: IGD values of the non-dominated solutions on the training set during the 30 independent GP runs. NSGA-III achieved signiﬁcantly better HV and IGD. This result clearly demonstrates that GP-MARP-NSGA-III can converge to a high-quality Pareto-front. NSGA-III achieved signiﬁcantly better HV and IGD. This result clearly demonstrates that GP-MARP-NSGA-III can converge to a high-quality Pareto-front. A useless reference points plot can help us to understand how well the reference points are generated from model-based approaches that match the distribution of solutions. From Figure 5.14 shows that the number of useless points during the GP search process is constantly decreasing in both model-based reference points adaptation methods. Figure 5.14 shows that GP-MARP-NSGA-III has less useless reference points than GP-NSGA- III-DRA. This reveals that reducing the density noise led to having fewer useless reference points and enhanced the better matches between the ref- erence points and candidate solutions. Further, the evolved Pareto-front can help enhance solution diversity and, therefore, the algorithms’ perfor- mance. The parallel coordinate plots for all competing algorithms are shown in Figures 5.15 (a) to 5.18 (b). It is well known that the number of useless reference points will increase if the distribution of reference points is not fully matched with the distribution of the candidate solutions. This can be seen in Figures 5.15 (a) and 5.15 (b) for GP-NSGA-III. In these ﬁgures, 135 5.5. GAUSSIAN PROCESS-BASED PROBABILISTIC MODEL 0 10 20 30 40 50 60 Generation 650 700 750 800 850 900 950 1000 # of Useless Points GP-NSGA-III GP-NSGA-III-DRA GP-MARP-NSGA-III Figure 5.14: The curves of the average number of useless reference points in GP-NSGA-III, GP-NSGA-III-DRA, and GP-MARP-NSGA-III. 0 10 20 30 40 50 60 Generation 650 700 750 800 850 900 950 1000 # of Useless Points # of Useless Points Figure 5.14: The curves of the average number of useless reference points in GP-NSGA-III, GP-NSGA-III-DRA, and GP-MARP-NSGA-III. we can see that the distribution of the reference points is uniform and that some reference points are far away from the solution locations, especially on MWT and MaxWT objectives. CHAPTER 5. REFERENCE POINTS ADAPTATION METHODS These useless reference points will cause a negative impact on algorithm performance, as conﬁrmed by experiment results reported in Table 5.2. Figure 5.16(a) shows that GP-A-NSGA-III still has a few useless points. Figures 5.16 (a) and 5.16 (b) show that solutions are also not well associ- ated with reference points on MAxF and MaxWT objectives. In contrast, Figures 5.18 (a), and 5.18 (b) show that the reference points’ distributions and the population closely match each other in the GP-MARP-NSGA-III CHAPTER 5. REFERENCE POINTS ADAPTATION METHODS 136 which is better than GP-NSGAIII-DRA. The distributions of the reference points and the population of GP-NSGA-III-DRA shown in Figures 5.17 (a) and 5.17 (b). MF MaxF MWT MaxWT 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (a) Distribution of the reference points. MF MaxF MWT MaxWT 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (b) Distribution of the ﬁtness values. Figure 5.15: Parallel coordinate plot for the distribution of the reference points and the distribution of the ﬁtness values at generations 50 of GP- NSGA-III. MF MaxF MWT MaxWT 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values MF MaxF MWT MaxWT 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (b) Distribution of the ﬁtness values. (a) Distribution of the reference points. Figure 5.15: Parallel coordinate plot for the distribution of the reference points and the distribution of the ﬁtness values at generations 50 of GP- NSGA-III. Figure 5.15: Parallel coordinate plot for the distribution of the reference points and the distribution of the ﬁtness values at generations 50 of GP- NSGA-III. 2. If the decision-maker prefers an effective approach other than the 5.6 Selection of adaptive reference points ap- proach In this chapter, we proposed three adaptive reference points approaches. Each reference point approach is equally important but the selection of approach depends on the decision-makers speciﬁc requirements as follow: 1. If the decision-maker prefers a very simple and efﬁcient approach with reasonable effectiveness, the PSO-based adaptive reference points approach is the right choice for generating reference points adaptively. 5.6. SELECTION OF ADAPTIVE REFERENCE POINTS APPROACH 137 MF MaxF MWT MaxWT 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Objectives Objectives Values (a) Distribution of the reference points. MF MaxF MWT MaxWT 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (b) Distribution of the ﬁtness values. Figure 5 16: Parallel coordinate plot for the distribution of the reference MF MaxF MWT MaxWT 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (b) Distribution of the ﬁtness values. MF MaxF MWT MaxWT 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Objectives Objectives Values (b) Distribution of the ﬁtness values. (a) Distribution of the reference points. Figure 5.16: Parallel coordinate plot for the distribution of the reference points and the distribution of the ﬁtness values at generations 50 of GP-A- NSGA-III. Figure 5.16: Parallel coordinate plot for the distribution of the reference points and the distribution of the ﬁtness values at generations 50 of GP-A- NSGA-III. MF MaxF MWT MaxWT 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Objectives Objectives Values (a) Distribution of the reference points. MF MaxF MWT MaxWT 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (b) Distribution of the ﬁtness values. Figure 5.17: Parallel coordinate plot for the distribution of the reference points and the distribution of the ﬁtness values at generations 50 of GP- NSGA-III-DRA. MF MaxF MWT MaxWT 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values MF MaxF MWT MaxWT 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Objectives Objectives Values (a) Distribution of the reference points. (b) Distribution of the ﬁtness values. Figure 5.17: Parallel coordinate plot for the distribution of the reference points and the distribution of the ﬁtness values at generations 50 of GP- NSGA-III-DRA. Figure 5.17: Parallel coordinate plot for the distribution of the reference points and the distribution of the ﬁtness values at generations 50 of GP- NSGA-III-DRA. CHAPTER 5. 5.6 Selection of adaptive reference points ap- proach REFERENCE POINTS ADAPTATION METHODS 138 MF MaxF MWT MaxWT 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Objectives Objectives Values (a) Distribution of the reference points. MF MaxF MWT MaxWT 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (b) Distribution of the ﬁtness values. Figure 5.18: Parallel coordinate plot for the distribution of the reference points and the distribution of the ﬁtness values at generations 50 of GP- MARP-NSGA-III. MF MaxF MWT MaxWT 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values MF MaxF MWT MaxWT 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Objectives Objectives Values (b) Distribution of the ﬁtness values. (a) Distribution of the reference points. Figure 5.18: Parallel coordinate plot for the distribution of the reference points and the distribution of the ﬁtness values at generations 50 of GP- MARP-NSGA-III. PSO-based approach, the density-based model adaptive reference points approach is a good choice for generating reference points adaptively. Further, the density model is more efﬁcient than our GaP model because it does not need to maintain the Gaussian matrix. Also, the density model is more ﬂexible than the GaP model because it introduces simple sampling techniques. The density-based model is a better trade-off approach between the GaP model approach and the PSO-based approach. 3. If the decision-maker prefers an effective approach which approxi- mates the Pareto-front more accurately than the PSO-based approach and density-based model approaches, the GaP model is the right choice for the reference points generation. However, the GaP model approach compromises simplicity, efﬁciency and ﬂexibility as com- pared to the other two approaches. 3. If the decision-maker prefers an effective approach which approxi- mates the Pareto-front more accurately than the PSO-based approach and density-based model approaches, the GaP model is the right choice for the reference points generation. However, the GaP model approach compromises simplicity, efﬁciency and ﬂexibility as com- pared to the other two approaches. 5.7. CHAPTER SUMMARY 139 5.7 Chapter summary This chapter aims to identify key research issues of those algorithms that use uniformly distributed reference points, including NSGA-III. The sim- ple adoption of uniformly distributed reference points failed to promote solution diversity during evolution and affected the performance of the many-objective optimization problems, which irregular, disconnected, de- generate, and inverted shapes of Pareto-front. To achieve this goal, we proposed model-free and model-driven adap- tive generation mechanisms for reference points. In our model-free mech- anism, we introduced a new reference point adaptation mechanism in- spired by PSO. Essential changes to particle dynamics in PSO have also been introduced in our mechanism to prevent the majority of reference points from converging to small areas in the objective space. This algo- rithm outperformed the baseline algorithm GP-NSGA-III. Model-based techniques explicitly construct the distribution model and give the learning ability to the algorithms. These algorithms have been designed to approximate the Pareto-front accurately and generate reference points that closely match the distribution of Pareto-optimal so- lutions. Additionally, improvements have also been made to effectively reduce the modeling noise in our density-based model using GaP model- ing. The GP-MARP-NSGA-III algorithm deﬁnes the area under the mean- locations and generates the required number of reference points in each sub-area. The proposed algorithm was applied to the static JSS problem. We compared our proposed algorithms with NSGA-III and previously pro- posed reference point adaptive approaches. Experimental results on the benchmark JSS problems show that our proposed algorithm reduces the useless reference points and provides a better distribution of Pareto- optimal solutions on the entire Pareto-front. Further, a better distribution of reference points also helps to improve the diversity of the solutions that 140 CHAPTER 5. REFERENCE POINTS ADAPTATION METHODS can be observed visually and in terms of HV and IGD. can be observed visually and in terms of HV and IGD. The experimental results have demonstrated that the GP-MARP- NSGA-III performs signiﬁcantly better problems with irregular Pareto- front because it reduces the number of useless reference points and pro- vides well-diversiﬁed Pareto-optimal solutions on the entire Pareto-front. Further, a better distribution of reference points also helps to improve the diversity of the solutions that can be observed visually as well as in terms of HV and IGD. This ﬁnding leads us to believe that our algorithm perfor- mance is promising on non-uniformly and irregularly distributed Pareto- front. 5.7 Chapter summary Also, in the future, these adaptive reference point methods might be easily deployed with any other reference points-based EMO algorithms to improve their performance on problems with irregular Pareto-front. In the next chapter, we will enhance the exploitation ability of the GP- NSGA-III. This is realized by combining GP-HH with Pareto local search. 6.1 Introduction Genetic programming (GP) is considered the most popular method for discovering and constructing dispatching rules for scheduling problems [141, 149]. Previous studies have shown that GP has been successfully used to evolve very effective dispatching rules for job shop scheduling (JSS) problems automatically [159, 143]. Researchers have studied the application of Pareto local search (PLS) to multi-objective evolutionary algorithms (MOEAs) with some success [50, 26]. By hybridizing global (such as genetic algorithms (GA)) search with local search, the performance of many evolutionary algorithms (EAs) can be improved because the local search can help to enhance the exploitation ability of EAs [26, 79]. This hybridization motivates us to integrate GP as a primary global search method with PLS and improve the quality of the evolved rules for many-objective JSS. PLS can be considered a direct extension of local search from single- objective problems to the multi-objective domain [26, 50]. PLS has three main algorithmic components [137]: (1) selection of solutions for neigh- 141 142CHAPTER 6. GENETIC PROGRAMMING WITH PARETO LOCAL SEARCH borhood exploration, (2) exploring the neighbors of the selected solutions from (1), and (3) describe the conditions under which a new solution stores in the PLS archive. The goal of PLS is to obtain a good approximation of the Pareto-optimal set. In particular, the study in [79] showed that suit- able candidates for local search should be carefully selected based on a weighted sum of multiple objectives as a ﬁtness function. One promising approach for improving the searchability of evolutionary multi-objective optimization (EMO) algorithms to ﬁnd near Pareto-optimal solutions is the hybridization with local search [26, 79]. The hybridization with local search algorithms is often referred to as memetic algorithms [79]. Exam- ples of such a hybrid algorithm can be found in multi-objective traveling salesman problems and ﬂowshop problems [14, 137]. PLS is very effective for tackling NP-hard multi-objective JSS problems [14]. However, no research works have been dedicated to studying PLS in genetic programming based hyper-heuristic (GP-HH ) for many-objective JSS. Based on our survey, there is only one existing work [149] studied us- ing of local search techniques in GP-HH for single-objective JSS. Due to this limitation, we investigate the effectiveness of PLS in GP-HH in this chapter. This study is expected to inspire many future studies on PLS in GP-HH for many-objective optimization. Based on the investigation in this chapter, new GP-PLS algorithms will be developed. 6.1 Introduction While developing our algorithm, we aimed to address three challenges that are vital for the seamless integration of PLS with GP-HH. First, the selection of initial solu- tions for neighborhood exploration. Second, a tree’s neighborhood struc- ture (dispatching rule) in GP is not deﬁned in the literature. Third, the acceptance criteria during the local search have to be carefully designed to guide the search properly for trade-offs solutions in many-objective JSS problems. A new ﬁtness-based selection mechanism is proposed to address the ﬁrst challenge, which uses a decomposition-based approach. To address the second challenge, a restricted mutation, is introduced to modify dis- 143 6.1. INTRODUCTION patching rules evolved by GP. The restricted mutation tries to avoid large changes, making the neighboring rules too different from their parent rule. Multiple consecutive local search steps will be performed to encourage the discovery of better rules surrounding an existing one. Meanwhile, the third challenge is tackled by adopting a dominance-based [26] and ﬁtness- guided [79] acceptance strategy. These three challenges will be explained more in Section 6.2. In this chapter, we ﬁrst perform the empirical study to determine whether the inclusion of local search improves the algorithms’ performance. Then we propose the two new memetic algorithms that in- tegrate GP with PLS. The chapter aims at developing GP-PLS and enhance the quality of evolved dispatching rules for many-objective JSS. The sub- objectives of this chapter are: 1. To investigate whether the inclusion of PLS within a GP-HH algo- rithm can increase the chance of discovering highly effective dis- patching rules for many-objective JSS, 2. To develop a new ﬁtness-based selection mechanism and neighbor- hood structure in GP-PLS, and 3. To compare the proposed GP-PLS with the state-of-the-art GP- NSGA-III algorithm on a group of benchmark JSS problems. In this chapter, we start with an explanation of our proposed algo- rithm, GP-PLS-I which, uses a restricted neighborhood structure and the partial acceptance mechanism for many-objective optimization problems (MaOPs). This is followed with the explanation of our other proposed al- gorithm, GP-PLS-II, which is the extension of GP-PLS-I. Following this, we will expand on the overview of proposed algorithms, GP-PLS-I and GP-PLS-II. After the overviews of proposed algorithms, the experimental design, the results, the analysis, and the discussion for the efﬁcacy investi- gation are covered. Finally, a chapter summary that wraps up this chapter is provided. 6.2.1 General framework of GP-PLS GP-PLS starts with the initialization by using the ramped-half-and-half method. The quality of dispatching rules evaluate in terms of each objec- tive (lines 1 and 19 of Algorithm 10). These rules then apply to a set of JSS training instances Itrain to generate schedules for them. Then, for each objective, the quality of a rule p is deﬁned as the average objective value of the schedules generated across all training instances. Next, PLS features the use of an archive to keep track of candidate rules for local search. The archive initially has either a randomly-selected sub- set of rules (Pk) in the population (used in GP-PLS-I) or a selection of the subset of rules (Pk) based on their ﬁtness value (used in GP-PLS-II). In both PLS algorithms, ﬁrst, PLS selects K individuals from the population to form the archive. Then, the proposed algorithms iteratively search the neighboring solutions of every rule (p) in the archive with the mutation operator’s help. A maximum of stepmax neighbors can be generated and the best neighbors ( pnew) are compared with p. If the neighbor rule is bet- ter than p, then it is added into Pbest. This (Pbest)g archive represents the best-performing dispatching rules evolved so far. 6.2 GP-PLS structure This section describes the general framework of the proposed algorithms which combine GP with PLS. First, we propose GP-PLS-I, which combine PLS with GP. Then, we propose an extension of GP-PLS-I, GP-PLS-II. . 6.1 Introduction In this chapter, we start with an explanation of our proposed algo- rithm, GP-PLS-I which, uses a restricted neighborhood structure and the partial acceptance mechanism for many-objective optimization problems (MaOPs). This is followed with the explanation of our other proposed al- gorithm, GP-PLS-II, which is the extension of GP-PLS-I. Following this, we will expand on the overview of proposed algorithms, GP-PLS-I and GP-PLS-II. After the overviews of proposed algorithms, the experimental design, the results, the analysis, and the discussion for the efﬁcacy investi- gation are covered. Finally, a chapter summary that wraps up this chapter is provided. 144CHAPTER 6. GENETIC PROGRAMMING WITH PARETO LOCAL SEARCH 6.2.2 GP-PLS-I overview Algorithm 10 outlines the framework of GP-PLS-I. The algorithm of GP- PLS-I has three signiﬁcant components, which can be seen in Figure 6.1. These components are 1) initialization and evaluation of dispatching rules, 2) PLS, and 3) NSGA-III selection. The initialization and evaluation of GP 6.2. GP-PLS STRUCTURE 145 Input : training set Itrain Input : training set Itrain Output: A set of non-dominated solutions(rules) P ∗ 1 Initialize of rules and Evaluate the population P0; Output: A set of non-dominated solutions(rules) P ∗ 1 Initialize of rules and Evaluate the population P0; p p 2 g ←0; 2 g ←0; 3 while g < gmax do 4 Pbest ←∅; 5 Randomly select K individuals from Pg to form archive; 6 foreach p ∈archive do 7 pnew ←p; 8 for step = 1 →stepmax do 9 p′ ←mutate(p); // neighbors 10 evaluate(p′); 11 if p′ is better than pnew then pnew ←p′ ; 12 end 13 if pnew is better than p then 14 Pbest \ p ; 15 (Pbest)g ←(Pbest)g ∪pnew 16 end 17 end 18 Apply genetic operators to (Pbest)g ∪(Pg \ (Pbest)g) to generate offspring Qg; 19 foreach Q ∈Qg do Evaluate rule Q; 20 g ←g + 1; 21 end 22 return The non-dominated individuals P ∗⊆Pgmax; 3 while g < gmax do are similar to the general framework of GP-NSGA-III ( the detailed infor- mation can be seen in Subsection 4.3.2 of Chapter 4). PLS components are the same as the algorithmic components that have been described in 146CHAPTER 6. GENETIC PROGRAMMING WITH PARETO LOCAL SEARCH Figure 6.1: General Framework of GP-PLS-I. Figure 6.1: General Framework of GP-PLS-I. Subsection 2.1.12 of Chapter 2. In GP-PLS-I, ﬁrst, K individuals are selected randomly from the popu- lation to form the archive. Then, each p in the archive (Pk) is selected for the neighborhood exploration by using the restricted mutation operator. When the restricted mutation is applied to rule p, the algorithm ran- domly selects a node in p whose corresponding sub-tree has a depth of 2. This restriction tries to avoid signiﬁcant change that makes the neighbor rules too different from the parent rules. A randomly generated depth-2 sub-tree then replaces the selected node and its sub-tree. With the help of this restricted mutation, GP-PLS-I can try to avoid signiﬁcant changes that make the neighboring rules signiﬁcantly different from the original rule [149]. This restricted mutation procedure is also described in Figure 6.2 and Algorithm 11. Exploring neighboring rules requires the exploration strategy, which determines the neighborhood’s size for exploration and the selection strategy for best dispatching rules. One can either explore the neighborhood entirely (best-improvement) [50]. Alternatively, only partially until the termination criterion is met [50]. 147 6.2. GP-PLS STRUCTURE Figure 6.2: Examples of possible neighbor rules(shaded sub trees repre- sented newly generated sub tree). Figure 6.2: Examples of possible neighbor rules(shaded sub trees repre- sented newly generated sub tree). 3. If pnew is dominated by p′, do nothing. By replacing rule p with rule pnew that dominates it in the archive, we can impose selection pressure on the archive and push it towards the Pareto- front. In order to maintain a well-diversiﬁed collection of rules in the archive, we combine the GP-PLS-I with the niching mechanism used by NSGA-III (the detailed information can be seen in Subsection 2.1.5 of Chapter 2). Algorithm 11: Restricted Mutation Algorithm 11: Restricted Mutation Input : p Output: pnew 1 Select a node at maximum depth of two in a p; 2 Randomly generate new subtree ; 3 pnew ←replace a selected node with new subtree; 4 return (pnew); 1 Select a node at maximum depth of two in a p; GP-PLS-I randomly samples a neighbor from the neighborhood repeti- tively until the maximum number of steps (stepmax) is reached. Whenever the new program is better than the current program, it will replace the current program. During the neighborhood exploration of rule p′, pnew sampled from the neighborhood will be compared with p′ (e.g., line 11 of Algorithm 10). Ac- cordingly, the following two strategies are considered: (1) the scalarization strategy [79] and (2) the replacement strategy [26]. In the scalarization strategy, the objective vector of each rule is aggregated into a scalar using 148CHAPTER 6. GENETIC PROGRAMMING WITH PARETO LOCAL SEARCH weighted sum, i.e. weighted sum, i.e. weighted sum, i.e. fit(x) = w1 · f1(x) + w2 · f2(x) + · · · + wm · fm(x), (6.1) (6.1) fit(x) = w1 · f1(x) + w2 · f2(x) + · · · + wm · fm(x), (6.1) where w = (w1, . . . , wm) is a random weight vector such that wi ≥0 (∀i = 1, . . . , m) and w1 + · · · + wm = 1. where w = (w1, . . . , wm) is a random weight vector such that wi ≥0 (∀i = 1, . . . , m) and w1 + · · · + wm = 1. The replacement strategy is based on the dominance relation. When we compare two rules pnew and p′; there are three possible outcomes based on the replacement strategy: 1. If pnew dominates p′, p′ is replaced by pnew. 2. If pnew and p′ are incomparable to each other, randomly choose one of them. 3. If pnew is dominated by p′, do nothing. 6.2.3 GP-PLS-II overview GP-PLS-II algorithm is an extension of GP-PLS-I. GP-PLS-II uses the ref- erence points and partitions the whole objective space into several sub- regions. The idea of decomposing the objective space has also been em- ployed extensively in recent literature [31, 43]. This decomposition of the objective space determines the appropriate search direction of each solu- tion. Thus solutions that belong to the same search direction are resided in the same sub-region of the objective space and called representatives of this subgroup. With the help of Equation (6.2), the number of repre- sentative solutions can be identiﬁed for each reference point. So, Kr best 6.2. GP-PLS STRUCTURE 149 solutions are selected according to their ﬁtness values from each subspace. The number of solutions to be selected for neighborhood exploration from each sub-region is: solutions are selected according to their ﬁtness values from each subspace. The number of solutions to be selected for neighborhood exploration from each sub-region is: Number of Representatives(Kr) = number of solution from each subspace Total population (6 2 ×K. ×K (6.2) (6.2) Representatives who have the highest ﬁtness value are given prior- ity to enter the archive of the initial solution. Then, each individual in the archive is selected for the neighborhood exploration. The selection of the representatives from each sub-region guides the search toward the Pareto-front while guaranteeing good population diversity in the objective space. GP-PLS-II has two variations, the GP-PLS-II-Uniform and GP-PLS- II-Adaptive. GP-PLS-II-Uniform uses a set of uniform reference points that are deﬁned by Das and Dennis’s systematic approach [39]. For GP-PLS-II-adaptive, we adopt the adaptive reference points ap- proach. For generating the adaptive reference points, we select the GP- MARP-NSGA-III algorithm from Section 5.5 of Chapter 5. GP-MARP- NSGA-III outperformed other adaptive reference points approaches (A- NSGA-III, NSGA-III-DRA).Therefore, we select GP-MARP-NSGA-III al- gorithm in GP-PLS-II. The details of the GP-PLS-II components and its framework are described below. In GP-PLS-I, K solutions are selected randomly from the whole popu- lation. However, in GP-PLS-II, individuals are selected as representatives from each sub-region according to their local ﬁtness value. In GP-PLS-II, we used the following steps for the selection of K solutions. 1. Combine the parent (Pg) and offspring (Qg) population and obtain the combined population (Rg). 2. Use a decomposition-based approach to split the objective space into a number of independent sub-regions according to a set of reference points. 6.2.3 GP-PLS-II overview Input : A set of non-dominated solutions (rules) Pg Output: archive of selected solutions for neighbourhood 1 Apply genetic operators to Pg to generate offspring Qg; 2 foreach Q ∈Qg do Evaluate rule Q; 3 Combine Pg and Qg (Rg=Pg ∪Qg); 4 Generate reference points W for j =1 to ∥R ∥do 5 Calculate the ideal point Zmin j =minr∈R fj(r); 6 Calculate the worst point Zmax j =maxr∈R fj(r) 7 end 8 for i = 1 to PopSize do 9 ˆ f(i)= fi−Zmin j Zmax j −Zmin j ; 10 end 11 foreach r ∈Rg do 12 foreach w ∈W do 13 compute the acute angle ⟨ˆ f(r), w⟩; 14 end 15 Assign ˆw = w : argminw∈W ⟨ˆ f(r), w⟩; 16 /∗Population Partition ∗/; 17 Assign θr = ⟨ˆ f(r), w⟩; 18 save r in E( ˆw) 19 end 20 /∗ﬁtness of individual from each sub-region ∗/; 21 foreach w ∈W do 22 foreach r ∈E( ˆw) do 23 Compute the convergence criteria C(r) ; 24 Compute the diversity criteria D(r) ; 25 Compute the ﬁtness of each individual FV (r) by using equation (6.5) 26 end 27 end 28 foreach w ∈W do 29 Select solution according to the FV (r) add selected solution from each subspace into the archive 30 end 31 return archive; 6.2.3 GP-PLS-II overview Solutions associated with similar reference points having an identical search direction. 150CHAPTER 6. GENETIC PROGRAMMING WITH PARETO LOCAL SEARCH Figure 6.3: Framework of GP-PLS-II. Figure 6.3: Framework of GP-PLS-II. 3. Assign ﬁtness values to each solution and N solutions are selected based on their ﬁtness values. 4. Choose K solutions from the N selected solutions as a representa- tive according to their ﬁtness value. Algorithm 12 outlines the selec- tion process of solutions in GP-PLS-II and the complete workﬂow is shown in Figure 6.3. For partitioning the objective spaces into many subspaces, GP-PLS-II gen- erates reference points (either uniformly or adaptively). The generation of adaptive reference points is shown in Algorithm 8 of Chapter 5. For partitioning the objective spaces into many subspaces, GP-PLS-II gen- erates reference points (either uniformly or adaptively). The generation of adaptive reference points is shown in Algorithm 8 of Chapter 5. Through decomposition, the search direction of each solution can be determined. Two solutions, s1 and s2, have identical search directions if they are associated with the same reference point. Reference points are positive and inside the ﬁrst quadrant, therefore, population Rg are normalized (see line 9 of Algorithm 12) before partitioned into 2N sub- populations Rg1, Rg2,. . . RgN by associating each individual with its closest reference point. The association of r is described in Figure 6.4. 6.2. GP-PLS STRUCTURE 151 Framework of GP-PLS-II Algorithm 12: Solution selection in GP-PLS-II. Framework of GP-PLS-II Input : A set of non-dominated solutions (rules) Pg Output: archive of selected solutions for neighbourhood 1 Apply genetic operators to Pg to generate offspring Qg; 2 foreach Q ∈Qg do Evaluate rule Q; 3 Combine Pg and Qg (Rg=Pg ∪Qg); 4 Generate reference points W for j =1 to ∥R ∥do 5 Calculate the ideal point Zmin j =minr∈R fj(r); 6 Calculate the worst point Zmax j =maxr∈R fj(r) 7 end 8 for i = 1 to PopSize do 9 ˆ f(i)= fi−Zmin j Zmax j −Zmin j ; 10 end 11 foreach r ∈Rg do 12 foreach w ∈W do 13 compute the acute angle ⟨ˆ f(r), w⟩; 14 end 15 Assign ˆw = w : argminw∈W ⟨ˆ f(r), w⟩; 16 /∗Population Partition ∗/; 17 Assign θr = ⟨ˆ f(r), w⟩; 18 save r in E( ˆw) 19 end 20 /∗ﬁtness of individual from each sub-region ∗/; 21 foreach w ∈W do 22 foreach r ∈E( ˆw) do 23 Compute the convergence criteria C(r) ; 24 Compute the diversity criteria D(r) ; 25 Compute the ﬁtness of each individual FV (r) by using equation (6.5) 26 end 27 end 28 foreach w ∈W do 29 Select solution according to the FV (r) add selected solution from each subspace into the archive 30 end 31 return archive; Framework of GP-PLS-II Framework of GP-PLS-II Algorithm 12: Solution selection in GP-PLS-II. Input : A set of non-dominated solutions (rules) Pg Output: archive of selected solutions for neighbourhood 1 Apply genetic operators to Pg to generate offspring Qg; 2 foreach Q ∈Qg do Evaluate rule Q; 3 Combine Pg and Qg (Rg=Pg ∪Qg); 4 Generate reference points W for j =1 to ∥R ∥do 5 Calculate the ideal point Zmin j =minr∈R fj(r); 6 Calculate the worst point Zmax j =maxr∈R fj(r) 7 end 8 for i = 1 to PopSize do 9 ˆ f(i)= fi−Zmin j Zmax j −Zmin j ; 10 end 11 foreach r ∈Rg do 12 foreach w ∈W do 13 compute the acute angle ⟨ˆ f(r), w⟩; 14 end 15 Assign ˆw = w : argminw∈W ⟨ˆ f(r), w⟩; 16 /∗Population Partition ∗/; 17 Assign θr = ⟨ˆ f(r), w⟩; 18 save r in E( ˆw) 19 end 20 /∗ﬁtness of individual from each sub-region ∗/; 21 foreach w ∈W do 22 foreach r ∈E( ˆw) do 23 Compute the convergence criteria C(r) ; 24 Compute the diversity criteria D(r) ; 25 Compute the ﬁtness of each individual FV (r) by using equation (6.5) 26 end 27 end 28 foreach w ∈W do 29 Select solution according to the FV (r) add selected solution from each subspace into the archive 30 end 31 return archive; Framework of GP-PLS-II Algorithm 12: Solution selection in GP-PLS-II. Framework of GP-PLS-II Input : A set of non-dominated solutions (rules) Pg Output: archive of selected solutions for neighbourhood 1 Apply genetic operators to Pg to generate offspring Qg; 2 foreach Q ∈Qg do Evaluate rule Q; 3 Combine Pg and Qg (Rg=Pg ∪Qg); 4 Generate reference points W for j =1 to ∥R ∥do 5 Calculate the ideal point Zmin j =minr∈R fj(r); 6 Calculate the worst point Zmax j =maxr∈R fj(r) 7 end 8 for i = 1 to PopSize do 9 ˆ f(i)= fi−Zmin j Zmax j −Zmin j ; 10 end 11 foreach r ∈Rg do 12 foreach w ∈W do 13 compute the acute angle ⟨ˆ f(r), w⟩; 14 end 15 Assign ˆw = w : argminw∈W ⟨ˆ f(r), w⟩; 16 /∗Population Partition ∗/; 17 Assign θr = ⟨ˆ f(r), w⟩; 18 save r in E( ˆw) 19 end 20 /∗ﬁtness of individual from each sub-region ∗/; 21 foreach w ∈W do 22 foreach r ∈E( ˆw) do 23 Compute the convergence criteria C(r) ; 24 Compute the diversity criteria D(r) ; 25 Compute the ﬁtness of each individual FV (r) by using equation (6.5) 26 end 27 end 28 foreach w ∈W do 29 Select solution according to the FV (r) add selected solution from each subspace into the archive 30 end 31 return archive; Algorithm 12: Solution selection in GP-PLS-II. Algorithm 12: Solution selection in GP-PLS-II. Algorithm 12: Solution selection in GP-PLS-II. Algorithm 12: Solution selection in GP-PLS-II. Input : A set of non-dominated solutions (rules) Pg Input : A set of non-dominated solutions (rules) Pg Input : A set of non-dominated solutions (rules) Pg p g Output: archive of selected solutions for neighbourhood Output: archive of selected solutions for neighbourhood 1 Apply genetic operators to Pg to generate offspring Qg; 1 Apply genetic operators to Pg to generate offspring Qg; 152CHAPTER 6. GENETIC PROGRAMMING WITH PARETO LOCAL SEARCH Figure 6.4: Example showing how to associate an individual r with a ref- erence points. In this example, w1 and w2 are two unit reference points, theta1 and theta2 are the angles between r and w1 and w2, respectively. Since theta2 < theta1, the individual denoted by r is associated with refer- ence points w2. Figure 6.4: Example showing how to associate an individual r with a ref- erence points. In this example, w1 and w2 are two unit reference points, theta1 and theta2 are the angles between r and w1 and w2, respectively. Since theta2 < theta1, the individual denoted by r is associated with refer- ence points w2. The acute angle measures the association between individual and ref- erence points. During the recent few years, the vector angle has attracted a high level of interest in evolutionary many-objective optimization [92, 30]. In our algorithm, the vector angle reﬂects the similarity of search direc- tions between two individuals and latter, the angle information between two individuals in the objective space is used to maintain the diversity. The acute angle can be calculated as: cosθi,j = ri,j · wi,j \|ri,j\| . (6.3) (6.3) where ri,j is an individual from the combined population of the size 2N and wi,j is a reference point. If an individual ri,j and wi,j have a minimal acute angle among all the reference points, ri,j becomes the member of the subpopulation Rg,k. 6.2. GP-PLS STRUCTURE 153 Once the population Rg is partitioned into 2N subpopulations, N solu- tions are selected through their ﬁtness value. The selection criteria based on the ﬁtness value (FV) are designed based on two sub-criteria: (1) the convergence criteria (d1 in Figure 6.5) and (2) the diversity criteria (d2 in Figure 6.5). d1 is represented by the distance from solution (ri,j) to the ideal point (Z∗) i.e, ∥ri,j −Z∗∥. Similarly, d2 is represented by the in- verse of the acute angle between (ri,j) and wi,j, i.e., θi,j. Input : A set of non-dominated solutions (rules) Pg In order to balance between the convergence criterion and the diversity criterion total FV of each individual can be formulated as a scalarization function: FV = d1 + d2 θm . (6.4) (6.4) since our motivation is to ﬁnd the solution on each reference point that is closest to the ideal point Z∗. Therefore, θm is used in equation (6.4) to Figure 6.5: Distance measure in the context of minimization with respect to a reference direction. Figure 6.5: Distance measure in the context of minimization with respect to a reference direction. normalize θi,j. This angle normalization process is adopted from RVEA [31]. This process is meaningful when some of these reference points are sparsely distributed or densely distributed. As a result, angles between 154CHAPTER 6. GENETIC PROGRAMMING WITH PARETO LOCAL SEARCH the candidate solutions and the reference points are either extremely small or extremely large. It is the best idea to apply high selection pressure on convergence dur- ing the exploration phase and push the population toward the Pareto- frontof the search process. However, during the exploitation phase, the constant pressure applies to diversity. Therefore, we introduce the penalty parameter g gmax in equation (6.5) which can better regulate the proportion of convergence and diversity information. So the new FV is expressed as: FV = d1 + g gmax ∗d2 θm . (6.5) (6.5) In the early stage, FV determines the convergence value (d1) because g << gmax, therefore d2 ≈0. However, when g approaches gmax, the penalty parameter gradually increases to emphasize the importance of the diver- sity criterion θi,j. After getting the ﬁtness values of each individual, then N solutions are selected. From N, K solutions are selected for the neigh- borhood exploration. Here parameter K is set identically as GP-PLS-I. • Neighborhood exploration: Neighborhood explorations strategies govern the size of the explored neighborhood and the selection of neighboring solutions. GP-PLS-II supports a partial exploration of the neighborhood until the maximum number of steps (stepmax) is reached. This number of steps is selected after the sensitivity analy- sis, which will discussed in subsection 6.3.1. The neighborhood solu- tion is obtained from any given rule p using the restricted mutation operator during neighborhood exploration. The detailed informa- tion of the restricted sub-tree mutation can be found in Section 6.2.2. Computational Complexity of the GP-PLS-II To analyze the computational complexity of the GP-PLS-II, we consider the main steps in one generation in the main loop of Algorithm 12. Apart from genetic operations such as crossover and mutation, the main com- putational cost has resulted from the objective normalization, population partition (line 16 of Algorithm 12), calculation of ﬁtness value (line 25 of Algorithm 12), and elitism selection (line 29 of Algorithm 12). The time complexity for the objective normalization (line 9 of Algorithm 12) is O(MN), where M is the objective number and N is the population size. The time complexity for the population partition of the 2N subpopulation is O(MN 2). In addition, the calculation of ﬁtness value by using equation (6.4) holds a computational complexity of O(MN 2). Further, for elitism selection, computational resources are mainly consumed by both conver- gence and diversity selection. The computational complexity of elitism selection is O(MN 2) and O(N 2) in the worst-case scenario. To summarize, apart from the genetic variations, the worst-case, that is, all the 2N individuals get trapped into one subspace and other subspaces do not contain any member, the overall computational complexity of GP- PLS-II within one generation is O(MN 2), which is the same as the average complexity. Input : A set of non-dominated solutions (rules) Pg • Comparison: The replacement strategy as discussed in Subsection 6.2.2 using the dominance relation to compare any two rules (p and pnew). The replacement strategy is also used in GP-PLS-II. It is obvi- ous that replacing the current rule p with any neighborhood rule that 6.3. DESIGN OF EXPERIMENT 155 dominates pnew would help in the coverage of the Pareto-optimal so- lutions. Selection pressure is applied to the solutions by performing PLS schema on the population Pk and new population Pbest is cre- ated. The Pbest combined with PN/best and created a new population Pg+1. 156CHAPTER 6. GENETIC PROGRAMMING WITH PARETO LOCAL SEARCH 156CHAPTER 6. GENETIC PROGRAMMING WITH PARETO LOCAL SEARCH lected as a testbed. In the experiments, we considered four potentially conﬂicting objectives: (1) the mean ﬂowtime (Obj1) (see equation (2.1) of Chapter 2), (2) maximal ﬂowtime (Obj2) (see equation (2.2) of Chapter 2), (3) mean weighted tardiness (Obj3) (see equation (2.6) of Chapter 2) , and (4) maximal weighted tardiness (Obj4) (see equation (2.7) of Chapter 2) which have been deﬁned in Section 3.2 of Chapter 2. The terminal set and function set for the tree-based GP have been de- scribed in Section 3.2 of Chapter 3. The crossover, mutation, and repro- duction rates are set identical to Subsection 2.1.2 of Chapter 4. In the experiments, the two commonly used measures in multi- objective optimisation, i.e. IGD [206] and HV [212] are used to compare the algorithms (see section of in Section 3.3 of Chapter 3). 6.3 Design of experiment The JSS benchmark which is a Taillard (TA) static JSS benchmark instances (the detailed information can be seen in Section 3.1 of Chapter 3) is se- 6.3.1 Sensitivity analysis In a hybridized algorithm, it is important to understand how to divide the available computation time between the local search and the global search. In order to prevent the local search from spending almost all available computation time, we decide to use the partial strategy, which restricts the number of iteration in the local search. If we use a very small value of stepmax (e.g., stepmax=1 ), the local search procedure may be terminated sooner than desired. On the contrary, if we use a large value of stepmax (e.g., stepmax=10 ), the local search procedure tends to evaluate more solu- tions than necessary. We need to carefully adjust the computation time spent by the local search procedure in our hybrid algorithm because of the above. There- fore, in this experiment, we examined different combinations of the pa- rameters where population size is equal to 1000. These combinations are: (K, stepsmax, generations) = (1000,3,25), (500,2,50), and (250,4,50). The Sen- sitivity analysis applies to GP-PLS-I and selecting parameters are later used in the GP-PLS-I algorithm. We also used selected parameters in GP- 157 6.3. DESIGN OF EXPERIMENT 6.3. Table 6.1: The mean and standard deviation over the average HV and IGD values on training instances of the compared algorithms in the four- objective experiment. HV (¯x ± σ) Comb1-(1000,3,25) Comb2-(500,2,50) Comb3-(250,4,50) GP-PLS-I-s GP-PLS-I-r GP-PLS-I-s GP-PLS-I-r GP-PLS-I-s GP-PLS-I-r 0.634(0.013) 0.684(0.015) 0.630(0.020) 0.676(0.015) 0.690(0.018) 0.705(0.016) IGD (¯x ± σ) GP-PLS-I-s GP-PLS-I-r GP-PLS-I-s GP-PLS-I-r GP-PLS-I-s GP-PLS-I-r 0.00131(0.00017) 0.00127(0.00012) 0.00134(0.00017) 0.00130(0.00024) 0.00127(0.00013) 0.00122(0.00012) PLS-II because GP-PLS-II is an extension of GP-PLS-I. The three-parameter settings have the same total number of ﬁtness evaluations (100000). For a fair comparison, the number of ﬁtness evaluations is kept identical to GP-NSGA-III. In the sensitivity analysis, 30 independent runs were per- formed to produce 30 ﬁnal sets of dispatching rules for each combination of the parameters. For the case of (1000,3,25), GP-PLS-I uses the whole population during the local search with three stepmax for exploring the neighborhood solu- tions. For the case of (500,2,50), GP-PLS-I can explore the solution space very well through 50 generations of evolution but has a small number of local searches. In contrast with the ﬁrst two parameter combinations, (250,4,50) has a proper balance between global search (50 generations) and local search (4 steps during the local search) capabilities. 6.3.1 Sensitivity analysis From the results summarized in Table 6.1, we found that the total num- ber of generations and the maximum number of local search steps highly inﬂuenced the performance of GP-PLS-I and their computational time. They together provide varied trade-offs between global and local searches in GP-PLS-I. The result showed that GP-PLS-I could not search the solution space extensively with a small number of generations in (1000,3,25). On the other hand, if GP-PLS-I cannot perform a sufﬁcient number of local search steps in (500,2,50), the power of local search cannot be effectively utilized. 158CHAPTER 6. GENETIC PROGRAMMING WITH PARETO LOCAL SEARCH 158CHAPTER 6. GENETIC PROGRAMMING WITH PARETO LOCAL SEARCH Figure 6.6: Computational time of whole population Figure 6.7: Computational time of sub-population. Table 6 1 shows that a combination (250 4 50) signiﬁcantly performed bet Figure 6.6: Computational time of whole population Figure 6.6: Computational time of whole population Figure 6.6: Computational time of whole population Figure 6.7: Computational time of sub-population. Figure 6.7: Computational time of sub-population. Table 6.1 shows that a combination (250,4,50) signiﬁcantly performed bet- . RESULTS AND DISCUSSIONS 159 6.4. ter in terms of HV and IGD as compared to the other two combinations for PLS. Figures 6.6 and 6.7 reveal that the local search on the whole popula- tion is more computationally expensive than the subset of solutions. For GP-PLS-I to achieve excellent performance, we select (250,4,50) in the sub- sequent experiments. GP-PLS-II is an extension of GP-PLS-I. Therefore, GP-PLS-II will also select similar parameters of GP-PLS-I. 6.4.1 Results Table 6.2: The mean and standard deviation over the average HV and IGD values on training instances of the compared algorithms in the four- objective experiment. HV (¯x ± σ) GP-NSGA-III GP-PLS-I-s GP-PLS-I-r GP-PLS-II-U GP-PLS-II-A 0.68850(0.0221) 0.69048(0.0160) 0.70513(0.0130) 0.70967(0.0115) 0.7133(0.0102) IGD (¯x ± σ) GP-NSGA-III GP-PLS-I-s GP-PLS-I-r GP-PLS-II-U GP-PLS-II-A 0.00125(0.00015) 0.00127(0.00013) 0.00122(0.00012) 0.00126(0.00016) 0.00123(0.000059) Tables 6.2 shows the mean and standard deviation of the training per- formance in terms of HV and IGD of the rules obtained by GP-NSGA-III, GP-PLS-I-s, GP-PLS-I-r, GP-PLS-II-Uniform (GP-PLS-II-U), and GP-PLS- II-Adaptive (GP-PLS-II-A). Here GP-PLS-I-s refers to the variation of GP- PLS-I where the scalarization approach is used for selection in Algorithm 10. On the other hand, GP-PLS-I-r represents the variation where the re- placement strategy is used for selection in Algorithm 10. For each algorithm in the experiment, 30 GP runs are conducted to obtain 30 sets of dispatching rules. Then, the rules are tested on the 40 test instances. The Wilcoxon rank-sum test [195], with the signiﬁcance level of 160CHAPTER 6. GENETIC PROGRAMMING WITH PARETO LOCAL SEARCH 0.05 is applied to the HV and IGD of the Pareto-front evolved by the ﬁve compared algorithms. Table 6.2 reveals that GP-PLS-II-A performs signiﬁcantly better than other competing algorithms in terms of both HV and IGD. Table 6.2 shows that GP-PLS-I-r is also highly competitive with GP-PLS-II-U. Therefore, we can conﬁrm that the replacement strategy is more effective than the scalarization strategy. Table 6.2 further shows that the selection of an ini- tial solution based on FV in PLS improves the algorithm’s performance compared to those that adopt the random selection. Therefore, both ver- sions of GP-PLS-II performed signiﬁcantly better than GP-PLS-I in terms of HV and IGD. Tables 6.3 and 6.4 show the mean and standard deviations of the test performance on each of the 40 test instances. In the case of HV, GP-PLS- II-A performed the best in 17 instances out of the 40 test instances. GP- PLS-II-U outperformed the other algorithms in 8 instances, GP-PLS-I-r performed the best in 5 instances, and GP-PLS-I-s outperformed other al- gorithms in 2 instances. GP-NSGA-III performed the best only on one instance. Regarding IGD, GP-PLS-II-A performed signiﬁcantly better than other algorithms in 14 instances out of the 40 test instances. GP-PLS-II-U out- performed the other algorithms in 9 instances. GP-PLS-I-r performed the best in 4 test instances. In contrast, GP-NSGA-III performed the best in 4 instances. 6.4.1 Results GP-PLS-I-s outperformed the other algorithms in only 2 in- stances. Tables 6.3 and 6.4 show that GP-PLS-I-r performed signiﬁcantly better in more instances than GP-PLS-I-s and GP-NSGA-III. The results indicate that GP-PLS-I-r is more effective in utilizing the dominance relation dur- ing the local search. Upon taking a closer look at Tables 6.3 and 6.4, it can be observed that GP-PLS-II algorithms not only performed well on small- scale problem instances but also on larger and more challenging instances in terms of HV and IGD. Tables 6.3 and 6.4 further show that GP-PLS-II al- 6.4. RESULTS AND DISCUSSIONS 161 Figure 6.8: Frequency of terminals in GP-NSGA-III, GP-PLS-I-s, GP-PLS- I-r, GP-PLS-II, and GP-PLS-II-A. Figure 6.8: Frequency of terminals in GP-NSGA-III, GP-PLS-I-s, GP-PLS- I-r, GP-PLS-II, and GP-PLS-II-A. gorithms performed signiﬁcantly better than GP-PLS-I-r and GP-PLS-I-s. The results indicate the effectiveness of selecting solutions based on the FV for neighborhood exploration. The results also demonstrate that overall, GP-PLS algorithms performed signiﬁcantly better than GP-NSGA-III. gorithms performed signiﬁcantly better than GP-PLS-I-r and GP-PLS-I-s. The results indicate the effectiveness of selecting solutions based on the FV for neighborhood exploration. The results also demonstrate that overall, GP-PLS algorithms performed signiﬁcantly better than GP-NSGA-III. Analysis of dispatching rules Previous results have shown that GP-PLS-II is a very effective approach to discovering dispatching rules for JSS. This section will explore its behav- iors to understand how it can effectively search for dispatching rules. The bar chart in Figure 6.8 shows the percentage of terminals in evolved rules from each algorithm. Useful terminals for optimizing ﬂowtime and tardi- ness objectives have been discussed in Subsection 4.5.3 of Chapter 4. It can be seen in Figure 6.8 that more than 10 percent of evolved rules from each 2CHAPTER 6. GENETIC PROGRAMMING WITH PARETO LOCAL SEARCH Table 6.3: The mean and standard deviation over the HV values on the test instances of the compared algorithms. nstances of the compared algorithms. Analysis of dispatching rules ID #J #M GP-NSGAIII GP-PLS-I-s GP-PLS-I-r GP-PLS-II-U GP-PLS-II-A 1 15 15 .1868(.0368) .1698(.0125) .2723(.0145) .2631(.0173) .1060(.0632) 2 15 15 .3296(.0182) .3070(.0145) .4091(.0110) .4141(.0161) .3271(.0782) 3 15 15 .2175(.0225) .2683(.0125) .2508(.0127) .3191(.0371) .3955(.0671) 4 15 15 .3159(.0183) .1259(.0133) .4561(.0091 .1771(.0087) .396(.0107) 5 15 15 .2370(.0202) .1799(.0302) .3479(.0192) .3360(.0462 .2719(.0577) 6 20 15 .4147(.0176) .4024(.0186) .4280(.0139) .3166(.0162) .4719(.0537) 7 20 15 .2396(.0683) .2551(.0695) .3146(.0683) .3636(.0613) .3779(.0582) 8 20 15 .2935(.0201) .3201(.0253) .2158(.0501) .2508(.0590) .2969(.0612) 9 20 15 .1787(.0198) .3232(.0198) .1606(.0219) .4021(.0832) .2019(.0502) 10 20 15 .3180(.0151) .3847(.0141) .2743(.0366) .3533(.0167) .4187(.0609) 11 20 20 .2114(.0038) .2214(.0138) .3087(.0118) .0086(.0064) .2218(.0605) 12 20 20 .3652(.0133) .2452(.0123) .3206(.0134) .3206(.0134) .3974(.0159) 13 20 20 .4540(.0151) .4541(.0326) .4550(.0221) .4971(.0163) .2469(.0175) 14 20 20 .1658(.01103) .2945(.01403) .3318(.0118) .3658(.0558) .2146(.0155) 15 20 20 .1742(.0225) .1840(.0199) .1352(.0125) .1558(.0158) .3704(.0119) 16 30 15 .2964(.0300) .3508(.0234) .3198(.0110) .3733(.0428) .2008(.0157) 17 30 15 .3741(.0096) .3385(.0229) .4076(.0093) .3915(.0529) .4729(.0145) 18 30 15 .3825(.0233) .3710(.3032) .3274(.0103) .4571(.0279) .3988(.0139) 19 30 15 .4353(.0126) .3808(.0197) .3988(.0146) .3495(.0264) .4119(.0302) 20 30 15 .3800(.0312) .3801(.0312) .3762(.0212) .3331(.0296) .3198(.0339) 21 30 20 .1983(.0657) .3340(.0792) .2190(.0492) .3635(.0719) .3768(.0449) 22 30 20 .2385(.0472) .2954(.0470) .3177(.0372) .3095(.0417) .3599(.0602) 23 30 20 .2020(.0398) .2672(.0410) .3675(.0294) .3510(.0267) .3468(.0449) 24 30 20 .4420(.0503) .4652(.0443) .4529(.0174) .3510(.0267) .3468(.0449) 25 30 20 .3372(.0477) .3854(.0396) .2864(.0427) .4555(.0359) .4947(.0418) 26 50 15 .4563(.0417) .4672(.0170) .4872(.0270) .5905(.0275) .6439(.0338) 27 50 15 .5610(.0361) .5685(.0304) .5555(.0304) .5510(.0529) .5849(.0238) 28 50 15 .4598(.0398) .4966(.0250) .4798(.0333) .5250(.0619) .6029(.0182) 29 50 15 .5749(.0372) .5702(.0251) .4323(.0413) .4418(.0274) .6260(.0241) 30 50 15 .4510(.0406) .4310(.0333) .4732(.0240) .5435(.0269) .5049(.0253) 31 50 20 .5190(.0424) .4477(.0295) .4870(.0433) .5124(.0429) .5830(.0171) 32 50 20 .4354(.0476) .5705(.0476) .4996(.0375) .4285(.0429) .5166(.0211) 33 50 20 .4427(.0266) .4426(.0838) .5165(.0366) .4525(.0279) .5206(.0234) 34 50 20 .4108(.0384) .3945(.0262) .4911(.0332) .6015(.0239) .4579(.0253) 35 50 20 .4421(.0349) .4138(.0222) .4140(.0165) .6610(.0189) .5049(.0253) 36 100 20 .6054(.0179) .59169(.0093) .6228(.0222) .5685(.0229) .4571(.0343) 37 100 20 .6584(.0142) .6678(.0101) .5857(.0152) .5735(.0182) .6989(.0188) 38 100 20 .6152(.0196) .6175(.0136) .5525(.0111) .6791(.0267) .6015(.0173) 39 100 20 .6495(.0185) .6515(.0191) .6695(.0103) .6305(.0229) .6885(.0188) 40 100 20 .6830(.0158) .6322(.0100) .6467(.0104) .6913(.0339) .6165(.0199) 6.4. RESULTS AND DISCUSSIONS 163 6.4. Table 6.4: The mean and standard deviation over the IGD values on the test instances of the compared algorithms. Analysis of dispatching rules ID #J #M GP-NSGA-III GP-PLS-I-s GP-PLS-I-r GP-PLS-II-U GP-PLS-II-A 1 15 15 .01897(.0008) .02748(.00011) .01479(.00012) .0119(.00017) .0299(.00043) 2 15 15 .0127(.0005) .0124(.0004) .0117(.0003) .0098(.00015) .0133(.00023) 3 15 15 .0173(.0008) .0126(.0006) .0137(.0002) .0115(.0005) .0112(.0026) 4 15 15 .0162(.0007) .0260(.0006) .0135(.0002) .0300(.0006) .0206(.0005) 5 15 15 .0193(.0005) .0200(.0003) .0085(.0007) .0119(.0001) .0115(.0005) 6 20 15 .0079(.0004) .0147(.0001) .0124(.0007) .0111(.0001) .0300(.0006) 7 20 15 .0250(.0013) .0109(.0012 .0121(.0014) .0010(.0009) .0093(.0002) 8 20 15 .0133(.0001) .0156(.0012) .0147(.0013) .0116(.0016) .0106(.0002) 9 20 15 .0201(.0001) .0108(.0005) .0142(.0003) .0078(.0008) .0085(.0001) 10 20 15 .0129(.0002) .0077(.0001) .0128(.0003) .0149(.0002) .0108(.0001) 11 20 20 .0094(.0006) .0090(.0001) .0091(.0003) .0221(.0007) .0119(.0014) 12 20 20 .0217(.0004) .0168(.0001) .0147(.0006) .0109(.0002) .0008(.0018) 13 20 20 .0155(.0001) .0112(.0005) .0100(.0004) .0009(.0001) .0008(.0002) 14 20 20 .0246(.0009) .0174(.0005) .0216(.0004) .0165(.0008) .0283(.0029) 15 20 20 .0191(.0008) .0315(.0003) .0289(.0004) .0083(.0001) .0024(.0008) 16 30 15 .0111(.0006) .0090(.0002) .0070(.0008) .0013(.0001) .0018(.0006) 17 30 15 .0061(.0003) .0047(.0008) .0077(.0007) .0057(.0001) .0058(.0003) 18 30 15 .0060(.0006) .0092(.0008) .0063(.0007) .0068(.0002) .0038(.0007) 19 30 15 .0065(.0004) .0059(.0008) .0054(.0007) .0064(.0002) .0063(.0007) 20 30 15 .0019(.0005) .0054(.0003) .0065(.0002) .0092(.0007) .0062(.0007) 21 30 15 .0094(.0022) .0097(.0013) .0125(.00012) .0122(.0003) .0079(.0009) 22 30 20 .0098(.0010) .0103(.0006) .0095(.00014) .0079(.00017) .0061(.0005) 23 30 20 .0077(.00004) .0074(.00006) .0063(.00024) .0082(.0015) .0127(.0019) 24 30 20 .0069(.00014) .0056(.00016) .0085(.00021) .0052(.0008) .0058(.0001) 25 30 20 .0056(.00012) .0071(.00005) .0074(.00014) .0057(.0001) .0067(.0005) 26 50 15 .0057(.00009) .0058(.00005) 0054(.00044) .0048(.00007) .0043(.0013) 27 50 15 .0038(.00029) 0054(.00013) .0042(.00041) .0045(.0006) .0048(.0007) 28 50 15 .0036(.00010) .0035(.00033) .0040(.00051) .0041(.0005) .0045(.0005) 29 50 15 0035(.00032) .0062(.00034) .0045(.00031) .0066(.00033) .0043(.0003) 30 50 15 .0043(.00010) .0045(.00044) .0042(.00011) .0049(.0008) .0053(.0008) 31 50 20 .0047(.00029) .0046(.00003) .0052(.00018) .0041(.0007) .0044(.0002) 32 50 20 .0040(.00009) .0054(.00006) .0041(.00008) .0062(.0005) .0051(.0005) 33 50 20 .0054(.00001) .0056(.00002) .0044(.00008) .0027(.0002) .0018(.0008) 34 50 20 .0037(.00005) .0035(.00008) .0033(.00009) .0036(.0004) .0049(.0006) 35 50 20 .0032(.00015) .0034(.00007) .0031(.00009) .0030(.0002) .0047(.0004) 36 100 20 .0027(.00020) .0032(.00001) .0021(.00067) .0039(.0006) .0023(.0004) 37 100 20 .0034(.00018) .0031(.00001) .0027(.00017) .0035(.0004) .0025(.0002) 38 100 20 .0044(.00028) 0037(.00004) 0040(.00008) .0028(.0004) .0032(.0003) 39 100 20 .0026(.00020) .0028(.00001) .0024(.00067) .0028(.0003) .0018(.0001) 40 100 20 0018(.00001) .0019(.00002) .0022(.00009) .0023(.0003) .0013(.0002) 164CHAPTER 6. GENETIC PROGRAMMING WITH PARETO LOCAL SEARCH 164CHAPTER 6. GENETIC PROGRAMMING WITH PARETO LOCAL SEARCH algorithm have W and PT terminals. According to the literature [70], MRT, PT, WKR, WINQ, NOINQ, FDD, NOPT are useful terminals (relevant) for optimizing ﬂowtime objectives. Speciﬁcally, PT, WINQ, and WKR are the most important three terminals for optimizing ﬂowtime objective. On the other hand, WINQ, NOINQ, NOPT, W, PT, MRT, DD are the useful termi- nals for optimizing tardiness objectives. Of these, PT, DD, and W are the most useful terminals for optimizing tardiness objectives. It can be seen from Figure 6.8 that the number of occurrences of W, WINQ, PT, DD, MRT, NOPT and FDD terminals are higher in the local search algorithms (GP-PLS-I-s, GP-PLS-I-r, GP-PLS-II, and GP-PLS-II-A) than in the GP-NSGA-III algorithms. This is because all the GP-PLS algo- rithms enhanced the exploitation ability and evolved signiﬁcantly better rules as compared to the other algorithms in terms of HV and IGD. There- fore, these algorithms selected the rules which are well-optimized. As a result, there are more chances of occurrences of useful terminals in GP-PLS algorithms. Further, it can also be seen from Figure 6.8 that GP-PLS-II-A has more useful terminals (W, PT, DD, FDD, and WINQ) than GP-PLS-I-s, GP-PLS-I-r, GP-PLS-II. This analysis shows the effectiveness of adaptive reference points with PLS. Figure. 6.9 shows the length of rules from each generation in GP- NSGA-III, GP-PLS-I-s, GP-PLS-I-r, GP-PLS-II. In most of the generations, GP-PLS-II produces shorter rules as compared to GP-PLS-I-r, GP-PLS-I- s, and GP-NSGA-III. In fact, all the hybridized algorithms with PLS have relatively short rules as compared to GP-NSGA-III. It is noted that the pro- gram lengths in Figure. 6.9 average length of the best rules of each inde- pendent run. This is another advantage of PLS as the short rules are easier to analyze and interpret. All PLS-II requires less time for evaluations as compared to GP-NSGA-III. This, in turn, may affect the computational time of the PLS-algorithm. 165 6.4. RESULTS AND DISCUSSIONS Figure 6.9: Length of rules from each generation in GP-NSGA-III, GP-PLS- I-s, GP-PLS-I-r, GP-PLS-II. Figure 6.9: Length of rules from each generation in GP-NSGA-III, GP-PLS- I-s, GP-PLS-I-r, GP-PLS-II. Further analysis To further investigate how PLS affects the GP search process, we plotted (a) the average HV and IGD of non-dominated solutions evolved by GP- PLS across multiple generations in Figures. 6.10 and 6.11, (b) parallel co- ordinate plots of non-dominated solutions evolved by GP-PLS and GP- NSGA-III algorithms on one problem instance in Figures 6.12 (a) to 6.16 (b), and (c) box-plots of the HV and IGD values on different test instances in Figures 6.19 (a) to 6.22 (b). Figures 6.10 and 6.11 reveal that GP-PLS-II-A has better convergence curves in terms of both HV and IGD than other compared algorithms. Fig- ures 6.10 and 6.11 also show that both algorithms of GP-PLS-II achieved better performance than the GP-PLS-I and GP-NSGA-III in terms of HV and IGD. These results reveal that the selection of the solutions based on FV (convergence and diversity) can improve the GP-PLS algorithm’s over- all performance. 6CHAPTER 6. GENETIC PROGRAMMING WITH PARETO LOCAL SEARCH 166CHAPTER 6. GENETIC PROGRAMMING WITH PARETO LOCAL SEARCH 166CHAPTER 6. GENETIC PROGRAMMING WITH PARETO LOCAL SEARC 0 5 10 15 20 25 30 35 40 45 50 Generation 0.2 0.3 0.4 0.5 0.6 0.7 0.8 HV values HV GP-NSGA-III GP-PLS-I-s GP-PLS-I-r GP-PLS-II-U GP-PLS-II-A Figure 6.10: The curves of the average number of HV value of the non- dominated solutions on the training set during the 30 independent GP runs. 0 5 10 15 20 25 30 35 40 45 50 Generation 0.2 0.3 0.4 0.5 0.6 0.7 0.8 HV values HV GP-NSGA-III GP-PLS-I-s GP-PLS-I-r GP-PLS-II-U GP-PLS-II-A Figure 6.10: The curves of the average number of HV value of the non- dominated solutions on the training set during the 30 independent GP runs. 0 5 10 15 20 25 30 35 40 45 50 Generation 0.2 0.3 0.4 0.5 0.6 0.7 0.8 HV values HV HV Figure 6.10: The curves of the average number of HV value of the non- dominated solutions on the training set during the 30 independent GP runs. Figure 6.10: The curves of the average number of HV value of the non- dominated solutions on the training set during the 30 independent GP runs. Further analysis 0 5 10 15 20 25 30 35 40 45 50 Generation -5 0 5 10 15 20 IGD values 10-4 IGD GP-NSGA-III GP-PLS-I-s GP-PLS-I-r GP-PLS-II-U GP-PLS-II-A Figure 6.11: The curves of the average number of IGD value of the non- dominated solutions on the training set during the 30 independent GP runs. IGD Figure 6.11: The curves of the average number of IGD value of the non- dominated solutions on the training set during the 30 independent GP runs. The parallel coordinate plots in Figures 6.12 (a) to 6.16 (b) depict the non-dominated set of dispatching rules obtained respectively by GP- NSGA-III. Figures 6.12 (a), 6.13 (a), 6.14 (a), 6.15 (a), and 6.16 (a) show that GP-PLS-I-r successfully evolved rules with better coverage for the objec- 167 RESULTS AND DISCUSSIONS 6.4. tives (on mean ﬂowtime and maximum weighted tardiness) in generation 10. On the other hand, it can be seen in Figures 6.12 (b), 6.13 (b), 6.14 (b), 6.15 (b), and 6.16(b) that rules evolved by GP-PLS-II-A are more diversi- ﬁed to cover a much wider range of the objective space in generation 50 than other compared algorithms. We can also observe that GP-PLS-II-U and GP-PLS-I-r are also well-diversiﬁed as compared to GP-NSGA-III and GP-PLS-I-s. Obji Obj2 Obj3 Obj4 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (a) Generation 10. Obji Obj2 Obj3 Obj4 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (b) Generation 50. Figure 6.12: Parallel coordinate plot of GP-NSGA-III. Obji Obj2 Obj3 Obj4 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (a) Generation 10. Obji Obj2 Obj3 Obj4 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (b) Generation 50. Figure 6.13: Parallel coordinate plot of GP-PLS-I-s. Obji Obj2 Obj3 Obj4 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (a) Generation 10. Obji Obj2 Obj3 Obj4 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (b) Generation 50. Figure 6.12: Parallel coordinate plot of GP-NSGA-III. Obji Obj2 Obj3 Obj4 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (a) Generation 10. Obji Obj2 Obj3 Obj4 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (b) Generation 50. (b) Generation 50. Figure 6.12: Parallel coordinate plot of GP-NSGA-III. Obji Obj2 Obj3 Obj4 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (a) Generation 10. Obji Obj2 Obj3 Obj4 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (b) Generation 50. Further analysis Figure 6.13: Parallel coordinate plot of GP-PLS-I-s. Obji Obj2 Obj3 Obj4 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values ( ) G i 10 Obji Obj2 Obj3 Obj4 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (b) Generation 50. (b) Generation 50. (a) Generation 10. Figure 6.13: Parallel coordinate plot of GP-PLS-I-s. 168CHAPTER 6. GENETIC PROGRAMMING WITH PARETO LOCAL SEARCH Obji Obj2 Obj3 Obj4 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (a) Generation 10. Obji Obj2 Obj3 Obj4 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (b) Generation 50. Figure 6.14: Parallel coordinate plot of GP-PLS-I-r. (a) Generation 10. (b) Generation 50. Figure 6.15: Parallel coordinate plot of GP-PLS-II-U. Obji Obj2 Obj3 Obj4 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (a) Generation 10. Obji Obj2 Obj3 Obj4 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (b) Generation 50. Figure 6.14: Parallel coordinate plot of GP-PLS-I-r. Obji Obj2 Obj3 Obj4 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (b) Generation 50. Obji Obj2 Obj3 Obj4 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (a) Generation 10. (a) Generation 10. (b) Generation 50. Figure 6.14: Parallel coordinate plot of GP-PLS-I-r. (a) Generation 10. (b) Generation 50. Figure 6.15: Parallel coordinate plot of GP-PLS-II-U. (b) Generation 50. (a) Generation 10. (a) Generation 10. (b) Generation 50. Figure 6.15: Parallel coordinate plot of GP-PLS-II-U. Figures 6.17 and 6.18 show the distribution of solutions of GP-PLS-II- A and GP-PLS-II-U to understand how GP-PLS-II-A performs better than GP-PLS-II-U. These ﬁgures show that GP-PLS-II-A has widely distributed optimal solutions than uniformly distributed decomposition-based algo- rithms. Figures 6.19 (a) to 6.22 (b) show the box plots of the HV and IGD values 6.4. RESULTS AND DISCUSSIONS 169 (b) Generation 50 (a) Generation 10. (b) Generation 50. Figure 6.16: Parallel coordinate plot of GP-PLS-II-A. (b) Generation 50. Figure 6.16: Parallel coordinate plot of GP-PLS-II-A. on different test instances. Figure 6.19 (a) shows a small instance where GP-PLS-II-A signiﬁcantly outperforms GP-PLS-II-U in terms of HV. The box plot of GP-PLS-II-A is comparatively taller than GP-PLS-II-U which shows that the area occupied by non-dominated solutions is higher than GP-PLS-II-U. Further, the long upper whisker reveals that most of the HV values are close to the positive quartile region in GP-PLS-II-A. On the other hand, GP-PLS-II-U also has long upper whiskers with outliers. Further analysis Figure 6.19 (b) shows the box plot on one of the complex instances with a larger number of machines where GP-PLS-II-A is signiﬁcantly bet- ter than GP-PLS-II-U in terms of HV. The box plot shows that GP-PLS-II- A produces higher HV values above the central region. Further, the box plot for GP-PLS-II-U is comparatively taller than GP-PLS-II-A. This result indicates that the area occupied by the non-dominated solutions is more signiﬁcant than GP-PLS-II-A. Figures 6.20 (a) and 6.20 (b) show the box plots of those instances where GP-PLS-II-U performs signiﬁcantly better than GP-PLS-A in terms of HV. Figures 6.20 (a) and 6.20 (b) depict the box plot, showing that most of the 170CHAPTER 6. GENETIC PROGRAMMING WITH PARETO LOCAL SEARCH Figure 6.17: Distribution of solutions of GP-PLS-II-U on instance 26. Figure 6.17: Distribution of solutions of GP-PLS-II-U on instance 26. HV values are close to the positive quartile region for GP-PLS-II-U. On the other hand, GP-PLS-II-A also has a long upper whisker with outliers. Figures 6.21 (a) to 6.21 (b) show that the box-plots of IGD values on smallest and complex test instances, respectively. Figure 6.21 (a) shows that GP-PLS-II-A signiﬁcantly outperforms GP-PLS-II-U in terms of IGD. The box plot of GP-PLS-II-A is comparatively lower than GP-PLS-II-U. Figure 6.21 (b) also indicates that the Pareto-optimal solutions are much closer to the Pareto-front as compared to Pareto-optimal solutions ob- tained by GP-PLS-II-U. Figures 6.22 (a) and 6.22 (b) show the box plots of those instances where GP-PLS-II-U performs signiﬁcantly better than GP-PLS-II-A in terms of 171 6.4. RESULTS AND DISCUSSIONS Figure 6.18: Distribution of solutions of GP-PLS-II-A on instance 26. Figure 6.18: Distribution of solutions of GP-PLS-II-A on instance 26. IGD. From the HV and IGD box plot, we found that generally evolved rules performed much better on the complex test problems. Further, HV and IGD values are more varied on simple problems than the complex problems. In this study, our experiment results showed that GP-PLS performs much better than the base algorithm, GP-NSGA-III, in terms of both HV and IGD. Moreover, GP-PLS-r performs signiﬁcantly better than GP-PLS- s. This indicates the effectiveness of using the dominance relation in the comparisons during a local search. Both GP-PLS-II versions performed signiﬁcantly better than GP-PLS-I-r and GP-PLS-I-s. This result reveals that the selection of solutions based on the FV (convergence and diver- sity) for neighborhood exploration will improve solutions’ quality. Further analysis GP- PLS-II-A performed signiﬁcantly better solutions than GP-PLS-II-U and has widely distributed optimal solutions than GP-PLS-II-U. 172CHAPTER 6. GENETIC PROGRAMMING WITH PARETO LOCAL SEARCH 172CHAPTER 6. GENETIC PROGRAMMING WITH PARETO LOCAL SEARCH (a) Instance 3. (b) Instance 39. Figure 6.19: Box-plots of the HV values. (a) Instance 3. (b) Instance 39. Figure 6.19: Box-plots of the HV values. (a) Instance 9. (b) Instance 38. Figure 6.20: Box-plots of the HV values. (a) Instance 9. (b) Instance 38. Figure 6.20: Box-plots of the HV values. 6.5. CHAPTER SUMMARY 173 (a) Instance 3. (b) Instance 39. Figure 6.21: Box-plots of the IGD values. (a) Instance 9. (b) Instance 38. Figure 6.22: Box-plots of the IGD values. 6.5 Chapter summary (a) Instance 3. (b) Instance 39. Figure 6.21: Box-plots of the IGD values. (b) Instance 39. (a) Instance 3. (b) Instance 39. (a) Instance 3. Figure 6.21: Box-plots of the IGD values. (a) Instance 9. (b) Instance 38. Figure 6.22: Box-plots of the IGD values. (a) Instance 9. (b) Instance 38. (b) Instance 38. (a) Instance 9. Figure 6.22: Box-plots of the IGD values. 6.5 Chapter summary In this chapter, we combine GP with a PLS for solving many-objective JSS problems. This approach’s key idea is to perform multiple local search 174CHAPTER 6. GENETIC PROGRAMMING WITH PARETO LOCAL SEARCH steps and effectively ﬁnd the neighborhood of non-dominated dispatch- ing rules. This local search mechanism helps create excellent exploitation abilities in GP-PLS. GP-PLS features the use of a newly designed restricted neighborhood structure and the partial acceptance mechanism for MaOPs. In this study, two common selection strategies, scalarization, and replace- ment are experimentally evaluated.From the experiments, we found that the total number of generations and the maximum number of local search steps are highly inﬂuential on GP-PLS-I performance. Further, we found that the dominance relation to compare two rules is signiﬁcantly better than the scalarization strategy. The second algorithm of GP-PLS-II followed the decomposition-based approach. In this approach, a set of reference points (either uniform or adaptive) decomposed the whole objective space into a number of small subspaces. A ﬁtness-based selection criterion was proposed for select- ing initial solutions for neighborhood exploration — the selection criteria based on convergence and diversity. Extensive experiments have been performed to understand the effec- tiveness of the proposed GP-PLS as compared to GP-NSGA-III by using the Taillard static job-shop benchmark set. Experiment results showed that GP-PLS performed much better than the GP-NSGA-III algorithm without the local search in terms of both HV and IGD. This study was further an- alyzed to reveal the different preferences over the use of terminals. This chapter is a ﬁrst step investigation of PLS in GP. The PLS used in GP im- proves the effectiveness of evolved rules in terms of HV and GP. The next chapter discusses the summary of the whole thesis, achieved research objectives, conclusions. It highlights the potential research direc- tions that could be carried out in the future related to the research investi- gations in this thesis. Chapter 7 This thesis’s overall goal has been successfully achieved, and several new genetic programming-based hyper-heuristics (GP-HH) methods have been developed. These developed methods evolved effective dispatch- ing rules for many-objective job shop scheduling (JSS) problems. The reusability and the effectiveness of the evolved dispatching rules have been demonstrated on the tested benchmark JSS instances. Further, we solved several issues of many-objective JSS problems, such as diversity. This thesis also used Pareto local search (PLS) to improve the quality of evolved dispatching rules. This chapter sets out the research goals that have been achieved, fol- lowed by the main conclusions. Finally, the chapter provides potential research areas of more general future works. 7.1 Achieved objectives The following research objectives have been fulﬁlled in this thesis: • The ﬁrst research objective developed a new many-objective GP- HH (GP-NSGA-III) to evolve a Pareto-front of non-dominated dis- patching rules for JSS. The GP-NSGA-III simultaneously evolved 175 175 176 CHAPTER 7. CONCLUSIONS non-dominated rules for conﬂicting objectives instead of aggregating many-objective problems into a single-objective optimization prob- lem. GP-NSGA-III seamlessly combined GP-HH for evolving dis- patching rules with the selection technique introduced in one of the state-of-the-art-algorithms: NSGA-III. This research objective in- vestigated the trade-offs among many commonly considered objec- tives in JSS problems, i.e., the mean ﬂowtime (mF) (see equation (2.1) of Chapter 2) , maximal ﬂowtime (maxF) (see equation (2.2) of Chapter 2), mean weighted tardiness (mWT) (see equation (2.6) of Chapter 2) and maximal weighted tardiness (maxWT) (see equa- tion (2.7) of Chapter 2). Furthermore, in this objective, we found that many-objective JSS problems suffered from the same issue of scalability when many-objective JSS problems used multi-objective optimization algorithms (NSGA-II and SPEA2). The results of this research objective in terms of HV and IGD also showed that the pro- posed algorithm, GP-NSGA-III, is signiﬁcantly better than the multi- objective optimization algorithms (NSGA-II and SPEA2). • The second research objective addressed the diversity maintenance issue of many-objective optimization problems. In this objective, we found that uniformly distributed reference points have many useless points [86]. These points fail to ﬁnd associated Pareto- optimal solutions on the objective space for combinatorial optimiza- tion problems. We successfully developed adaptive reference point approaches for generating the reference points according to the can- didate solutions’ distribution. The following two adaptive reference point approaches were developed: (i) model-free approach and (ii) model-based approach. • The second research objective addressed the diversity maintenance issue of many-objective optimization problems. In this objective, we found that uniformly distributed reference points have many useless points [86]. These points fail to ﬁnd associated Pareto- optimal solutions on the objective space for combinatorial optimiza- tion problems. We successfully developed adaptive reference point approaches for generating the reference points according to the can- didate solutions’ distribution. The following two adaptive reference point approaches were developed: (i) model-free approach and (ii) model-based approach. In the model-free approach (GP-A-NSGA-III(PSO)), particle swarm optimization (PSO) was incorporated into the NSGA-III. Essential changes to particle dynamics in PSO were also introduced in our GP- 7.1. • The third research objective combined GP with PLS. The two ver- 7.1 Achieved objectives ACHIEVED OBJECTIVES 177 A-NSGA-III(PSO) to prevent the majority of reference points from converging to small areas in the objective space. The experiment results showed that GP-A-NSGA-III(PSO) decreased useless refer- ence points during the evolutionary search. Further, the generation of reference points adaptively improved the performance of GP-A- NSGA-III(PSO) in terms of HV and IGD. The results also showed that evolved Pareto-front was more diversiﬁed with adaptive refer- ence points than uniform reference points. In the model-based approach, models were constructed explicitly. They were constructed according to the density of solutions from each deﬁned sub-location in a whole objective space. This density- based model learned the distribution of candidate solutions and ap- proximated the Pareto-front based on the evolved solutions. For generating reference points, we introduced two different methods. The ﬁrst method produced reference points close to vertices, and the second method generated reference points on intermediate locations. Both of these methods enhanced the match between reference points and Pareto-optimal solutions. As a result, useless reference points decreased throughout the evolutionary process. To further improve of the density model, we reduced the density-noise by predicting the mean value in the sub-location. The model-based approach reduced the useless reference points and provided a better distribution of Pareto-optimal solutions on the en- tire Pareto-front. Further, a better distribution of reference points improved the diversity of solutions that were observed visually and in terms of HV and IGD. Experimental results demonstrated that the adaptive reference point approach (model-based) performed signif- icantly better than the other compared algorithms on irregular, dis- connected, degenerated, and inverted Pareto-front shapes. • The third research objective combined GP with PLS. The two ver- 178 CHAPTER 7. CONCLUSIONS sions of GP-PLS algorithms were developed in this research objec- tive, GP-PLS-I and GP-PLS-II. A new ﬁtness-based selection mecha- nism was developed for the PLS. Both versions of GP-PLS features used a newly designed restricted neighborhood structure and the partial acceptance mechanism for many-objective optimization prob- lems (MaOPs). In this study, we further considered and examined the use of two common selection strategies for the selection of new solutions (scalarization and replacement). Experimental results demonstrated the effectiveness of the newly developed GP-PLS as compared with the baseline algorithm (GP- NSGA-III). Evolved rules of GP with PLS were more effective than rules evolved by GP without PLS because GP-PLS enhanced the ex- ploitation ability of the algorithm. 7.2 Main conclusions The main conclusions for the three research objectives drawn from the three contribution chapters (Chapter 4, Chapter 5, and Chapter 6) are dis- cussed in this section. 7.1 Achieved objectives As a result, rules obtained by GP- PLS had more useful terminals than GP-NSGA-III. 7.2.1 Many-Objective GP for JSS In Chapter 4, we investigated many-objective JSS and focused on evolving a set of trade-offs dispatching rules for many-objective JSS. For this, we combined GP with NSGA-III. In order to tackle many objectives, different aspects have to be considered. Pareto-front The experimental results showed that the GP-NSGA-III performed on training instances was signiﬁcantly better than GP-NSGA-II and GP- SPEA2 in terms of HV and IGD. The evolved rules of GP-NSGA-III are also performed signiﬁcantly better than GP-NSGA-II and GP-SPEA2 on the test instance, especially on large-size test instances. Thus rules evolved from GP-NSGA-III exhibit the generalization and re-usability abilities. New evolutionary search mechanisms In this thesis, the evolution process of GP and the selection scheme of NSGA-III were combined. GP is a commonly-used hyper-heuristic for 179 7.2. MAIN CONCLUSIONS evolving dispatching rules for JSS and has achieved great success [68, 87]. The combination of GP and NSGA-III was designed as a competitive al- gorithm for evolving a set of trade-offs rules in many-objective JSS. In this algorithm, GP-NSGA-III combined the initialization, evaluation, and evo- lutionary operators of GP with the selection scheme of NSGA-III. Effectiveness of dispatching rules The experimental results showed that the evolved rules from GP-NSGA- III were signiﬁcantly better than GP-NSGA-II and GP-SPEA2. Therefore, rules evolved from GP-NSGA-III were well optimized (well-converged and well-diversiﬁed). As a result, GP-NSGA-III rules effectively identify and use useful terminals than GP-NSGA-II and GP-SPEA2. 7.2.2 Non-uniform Pareto-front In NSGA-III, the diversity of solutions is encouraged by adopting a set of uniformly distributed reference points in the objective space. How- ever, evenly distributed reference points may not be effective for prob- lems with disconnected and non-uniform Pareto-front. These problems generate many useless reference points that are never associated with any of the Pareto-optimal solutions. The existence of these useless reference points in NSGA-III has signiﬁcant effects on its performance. Two adap- tive reference point generation approaches were developed in this thesis 180 CHAPTER 7. CONCLUSIONS to address the useless reference points issue. These two approaches are described below: to address the useless reference points issue. These two approaches are described below: (1) Model-free approach The ﬁrst adaptive reference point method inspired by PSO [153] is called ”Adaptive-NSGA-III(PSO)”. This approach is a simple and efﬁcient one than the model-based approach. We found that useless reference points kept decreased during the evolutionary process. Especially at the later stage of the search, the adaptive reference point scheme led to less useless reference points and a better reﬁnement of the population’s regions. The proposed reference point adaptation algorithm signiﬁcantly improved the performance of GP-NSGA-III in terms of both HV and IGD which indicate that the algorithm performs the best if the distribution of the reference point is consistent with the distribution of the Pareto-optimal solutions. Benchmarks problems The performance of the adaptive model-based approach is also veriﬁed on ten benchmark problems. These problems have various shapes of Pareto-front. We have used these problems as the testbed in the empirical studies problem, where the number of objectives was scaled from three to eight. The experimental results demonstrated that the model-based approach performed signiﬁcantly better than the other competing algo- rithms ( REVA, Two-ARCh, NSGA-III, IMMOEA, and A-NSGA-III) on the different types of Pareto-front which have irregular, disconnected, de- generate, and inverted shapes of Pareto-front. The model-based approach performed signiﬁcantly better as compared to the other MOEAs with var- ious types of Pareto-front. This is because the model-based approach gets the concrete knowledge of the objective space and provided better matches between the distribution of Pareto-optimal solutions and reference points. Further, a better distribution of reference points improved the diversity of the solutions in terms of HV and IGD. (2) Model-based approach In the model-based approach, the models are explicitly constructed and learn the distribution of the candidate solutions of the current genera- tion and generate reference points according to the distribution of the so- lutions. We also used Gaussian process-based modeling for smoothing the surface and calculated an area under the Gaussian process model’s mean function. The model-based approach estimated the approximately accurate number of reference points in each subspace. Furthermore, we found that the model-based approach was effective than the model-free approach. Further, the density-based model is more efﬁcient than Gaus- sian process-based model. However, Gaussian process-based model is highly effective than the model-free approach and density-based model. The result showed that the number of useless points during the GP search process is constantly decreasing in the model-based reference points adap- tation approach. 7.2. MAIN CONCLUSIONS 181 Sensitivity analysis In the experiment, a sensitivity analysis was performed and found that the total number of generations and the maximum number of local search steps together provide varied trade-offs between global and local searches in GP-PLS. 7.2.3 Pareto local search (PLS) In Chapter 6, we combined GP with PLS. The ﬁrst GP-PLS mechanism, GP-PLS-I investigated the inﬂuence of Pareto local search with GP. Based on the results of GP-PLS-I, a new algorithm, GP-PLS-II, was developed. These ﬁndings are as follows: (1) GP-PLS required a ﬁtness-based selection mechanism for selecting the initial solution for neighborhood exploration. (2) GP-PLS should perform an adequate number of local search steps for local searches to show its effectiveness, and (3) The total number of gen- erations and the maximum number of local search steps signiﬁcantly af- fected by the performance of GP-PLS. Both GP-PLS algorithms used the three major components of PLS, (1) selection of a solution, (2) neighbor- hood exploration, and (3) comparison. From the GP-PLS, we found that the ﬁtness-based selection mechanism based on the convergence and di- 182 CHAPTER 7. CONCLUSIONS versity criteria for selecting initial solutions performed signiﬁcantly better than the random selection. Further, a neighborhood structure for GP, the restricted mutation, can effectively prevent a new neighboring rule dis- covered during the local search process from being signiﬁcantly different from the original rule. For comparing the new rule with its immediate par- ent rule, the following two strategies were considered: (1) the scalarization strategy [79], and (2) the replacement strategy [26]. The result showed in terms of HV and IGD indicates the effectiveness of using the dominance relation in the comparisons during the local search. Extensive experiments were performed to understand the effectiveness of the proposed GP-PLS as compared with the base-line algorithm, GP- NSGA-III. Experiment results showed that GP-PLS performs much bet- ter than the current state-of-the-art method without local search in terms of both HV and IGD. Moreover, programs obtained by GP-PLS-II and GP-PLS-I were more effective than the GP-NSGA-III because GP-PLS en- hances the exploitation ability; as a result, GP-PLS algorithms have the majority of useful terminals. 7.3.1 Incorporate user preferences to many-objective JSS This thesis used reference points that decomposed the original many- objective space into subspace. However, we can further use these reference points to target the user’s interested in different sub-regions of the entire Pareto-front. The use of user preferences is important for MaOPs because the user may have a preferred subset of Pareto-optimal solutions that fall speciﬁc subregion of the objective space. Also, it has been shown in the literature that reference points are one of the efﬁcient and effective meth- ods to preference articulation [31]. This preference articulation is particu- larly crucial in many-objective where it is improbable to cover the whole Pareto-front [31] comprehensively. 7.3 Future work This section highlights key areas of future work related to the ﬁeld of many-objective JSS. 7.3. FUTURE WORK 183 7.3.2 Incorporate locality of search operators to Genetic Programming As we know, every optimization problem can be decomposed into a genotype-phenotype mapping and a phenotype-ﬁtness mapping. In EC, locality refers to how well neighboring genotypes correspond to neigh- boring phenotypes. The locality of a representation is high if all neigh- boring genotypes correspond to neighboring phenotypes. In contrast, the locality of a representation is low if some neighboring genotypes do not correspond to neighboring phenotypes. It is already mentioned that a rep- resentation of high locality is necessary for efﬁcient evolutionary search [52, 171]. In most local search methods, mutation use to explore the neigh- borhood of its immediate parents. If the mutation-based search algorithm does not have a high locality representation, it would jump randomly around the search space. There is no explicit genotype-phenotype mapping in the tree- structured GP, so we can say that there are no explicit phenotypes distinct from genotypes [52]. This is common in GP, therefore, to study instead of the behavior of the mapping from genotype to ﬁtness. However, in JSS, a 184 CHAPTER 7. CONCLUSIONS rule is represented as a GP tree and two dispatching rules with different tree structures can essentially have exactly the same behavior in making decisions [130]. For example, the dispatching rules with the priority func- tions of x, 2x will always make the same decision [66]. So, in JSS, rules are different in terms of phenotypic behavior instead of genotypic structure. So to compare the similarity of solutions during neighborhood exploration in PLS, we will adopt a phenotypic characterization approach [66] which can characterize the behavior of a dispatching rule as a ﬁxed-length nu- meric vector. This phenotypic characterization approach is based on a reference dispatching rule and applied to single objective JSS problems. However, it is a potential research direction for many-objective JSS. There- fore, we will explore this research direction in the future and ﬁnd refer- ence rules for many-objective JSS. Moreover, we will also integrate this approach with our PLS algorithm (see Chapter 6). 7.3.3 Incorporate effective crossover operator to many- objective JSS This thesis tackles several issues of many-objective JSS. One of the chal- lenges of many-objective JSS problems is designing an effective crossover operator. Sato et al. [174] revealed that in combinatorial optimization problems such as JSS problems, the recombination of two parents close to the Pareto-front might generate an offspring distant from the Pareto-front since a conventional crossover operator might be too disruptive for many- objective combinatorial optimization. This is because, in many-objective JSS problems, non-dominated solutions can have quite different building blocks, which are useful for different objectives. Randomly selecting par- ents based on the dominance relation may result in combining building blocks for different objectives together which may not lead to good off- spring. So, in this case, a good recombination operator is needed to dimin- ish crossover operators’ disruptive effect in many-objective JSS problems. 7.3. FUTURE WORK 185 7.3.4 Incorporate adaptive terminal selection to many- objective JSS This thesis developed several GP-HH approaches for many-objective JSS. Each approach used a GP system for evolving and effective rules. These rules utilized several potential terminals (and functions) that can be given as parameters to the GP system to discover dispatching rules automati- cally. The terminal set was formed by machine, job, and shop attribute [47, 150]. Although GP based approaches in this thesis outperformed the other compared algorithms, the terminal selection was not based on the preference objectives. The size of dispatching rules also increases with the number of objectives and rules may have redundant terminals that affect the interpretability of dispatching rules. The interpretability of rules can be achieved by eliminating the irrelevant terminals from the terminal set. 7.3.5 Dispatching rules for many-objective dynamic JSS In this thesis, we have only used the GP to evolve dispatching rules (prior- ity functions) for static JSS environments because many-objective research is comparatively newer than single objective JSS and multi-objective JSS. Furthermore, dynamic JSS environments are more complex than static JSS environments. This is because, in dynamic JSS environments, jobs contin- uously arrive on the shop ﬂoor at various instances of time with no prior process information, e.g., release date, due date, and processing time are not given before job arrival. Further, the existing GP-HH for dynamic JSS is not efﬁcient regarding ﬁtness evaluation. Therefore, this thesis evolved effective and efﬁcient dispatching rules for static JSS. This thesis assisted in the understanding of how optimization approaches can be adapted to improve many-objective JSS algorithms performances. Therefore, there is a need to investigate dynamic many-objective JSS and evolve effective dis- patching rules. CHAPTER 7. CONCLUSIONS 186 7.3.6 Dispatching rules for many-objective ﬂexible JSS Flexible job-shop scheduling (FJSS) problems [16] is an extension of clas- sical JSS problems that allows the processing of each operation on more than one machine. FJSS problems consist of two sub-problems: (1) the routing sub-problem and (2) the scheduling sub-problem. The routing sub-problem focuses on assigning each operation to a machine out of the set of capable machines. The scheduling sub-problem aims to sequence the assigned operation on all the machines to obtain a feasible schedule that optimizes one or many objectives. It is more complicated than JSS problems because it involves two strongly coupled decisions, the routing sub-problem and the scheduling sub-problem [16]. FJSS is an NP-hard problem[16]; therefore, it is essential to know how to efﬁciently use the GP system and evolve effective rules for many-objective FJSS problems. Furthermore, over the past decades, the single-objective FJSS problems mainly focused on minimizing the makespan that has been extensively studied in the JSS literature. Compared to the single-objective FJSS prob- lems, the multi-objective FJSS problems research is relatively limited to three objectives [202]. However, many real-world scheduling problems involve the simultaneous optimization of several objectives that conﬂict. Despite the requirement of many-objective research, there are only a few studies on many-objective FJSS [57] but in these studies, many-objective optimization was not explicitly investigated [57]. In line with this re- search trend, we believe that many-objective FJSS will attract increasing research attention in the near future. Therefore, research to construct a hyper-heuristic framework to generate effective rules for many-objective FJSS problem is a potentially signiﬁcant research direction that demands a new and specialized GP approach. In summary, many future research directions arise as a result of the work discussed in this thesis. 7.4. GENERAL CONSIDERATIONS 187 7.4 General Considerations This section ﬁrst describes the main component of the proposed algo- rithms with algorithmic parameters. Next, this section shows the appli- cation of the proposed algorithm to a cloud platform. 7.4.1 Main components of thr Proposed algorithm This section shows the main components and characteristics of the de- veloped algorithm. Figure 7.1 shows these three components of the hy- bridized algorithm developed in this thesis. Figure 7.1: Main Components of Algorithm Figure 7.1: Main Components of Algorithm Initialized the Population In this thesis, we used the GP system to evolve automated dispatching rules for JSS problems represented as trees and made up of function and terminal nodes. Terminals nodes are usually deﬁned by constants and variables from the speciﬁc problem domains. 188 CHAPTER 7. CONCLUSIONS The GP system usually requires genetic parameters such as the popu- lation size, termination parameter, crossover and mutation probabilities, and their associated parameters. These parameters are not algorithmic parameters but problem-speciﬁc parameters to generate a population of individuals. Although we evolved dispatching rules for static JSS problems, dis- patching rules can cope in dynamic environments, such as cloud schedul- ing [188] and vehicle routing problems [16]. Dispatching rules is one of the potential approaches to deal with dynamic changes because they are computationally efﬁcient and can react quickly to dynamic changes on the shop ﬂoor [141, 159]. So, one can also use GP based system for evolving dispatching rules for dynamic problems. Our algorithm is not limited to generate dispatching rules by using GP system but one can generate a population of individuals from any other benchmark functions (see Appendix A) and feed these individuals to the next component of algorithms to select the best solution. Selection scheme The algorithm in this thesis solves many-objective JSS problems by com- bining GP and selection schemes of MaOPs. The new ﬁtness-based se- lection mechanism based on convergence and diversity was introduced. This new selection scheme balance the convergence criteria and diversity criteria and does not require to set any new parameter. To improve the effectiveness of the solution, we further integrate global search with the PLS. Here, PLS requires two additional parameters: the archive’s size and the maximum number of local search steps. Reference point adaption method In real-world scheduling problems, the true Pareto-front is unknown and due to the discrete and combinatorial solution space, the Pareto- front of scheduling problems can be very irregular. It was shown in this thesis that the adoption of uniformly distributed reference points affects the perfor- mance of the algorithms adversely. Therefore, we developed three adap- tive reference points approaches (PSO-based adaptive reference points approach, density-based model adaptive reference points approach, and gaussian process model adaptive reference points approach). The PSO-based adaptive reference point approach is more efﬁcient than the other two approaches but it requires PSO parameters such as in- ertia weight and acceleration coefﬁcients [101]. The density-based model, adaptive reference points approach, is more effective than the PSO-based model adaptive reference and more efﬁcient 7.4. GENERAL CONSIDERATIONS 189 than the Gaussian process model adaptive reference points approach. This approach does not require any controlled parameters. The Gaussian process model, adaptive reference points approach, is more effective than the other two approaches. This approach requires the hyperparameters of the covariance functions [170]. These adaptive reference points approaches which are mentioned above require the size of reference points but the number of reference points is not an algorithmic parameter, as this is entirely at the disposal of the user. The population size is dependent on the size of reference points, as population size ≈number of reference points. Further, these adaptive reference point methods might be easily de- ployed with any other reference points-based evolutionary multi-objective optimization (EMO) algorithms to improve their performance on prob- lems with irregular Pareto-front. 7.4.2 Cloud task scheduling problem In this section, we apply our many-objective scheduling algorithm on the cloud task scheduling problem. However, we already applied our algorithm to the number of benchmark test problems (see Appendix A) 190 CHAPTER 7. CONCLUSIONS with three to eight objectives to show our many-objective scheduling al- gorithm’s general applicability and usefulness. In cloud computing, users may utilize hundreds or thousands of vir- tualized resources and it is impossible for everyone to allocate each task manually. Task scheduling problem refers to how to reasonably arrange many tasks provided by users in virtual machines, which is very impor- tant in cloud computing. In cloud computing, the cost, response time, resource utilization, and load balancing are often conﬂicting [54]. Hence cloud scheduling can be considered as a many-objective optimization problem. To achieve better scheduling performance, one should to minimize cost, minimize time, im- prove resource utilization, and load. In order to test the performance of the algorithm on the task schedul- ing problem, we can use discrete event simulation (DES) platforms [22] to support cloud computing environments such as the Cloudsim platform [22]. First of all, detailed parameter settings of the virtual machine and tasks should be provided. Also, initialize the population of size N. The GP system can randomly initialize the population, where each individual represents a dispatching rule. After objective values are calculated by ap- plying the rule to the training instances of cloud scheduling. The ﬁtness of dispatching rules is evaluated by a DES. Then the following steps can use for the selection of N best solutions 1. combine offspring with the parent population and obtain a com- bined population of size 2N. 2. generate reference points adaptively which split the objective space into a number of independent sub-regions. This partitions the pop- ulation into multiple subpopulations by associating each individual with its closest reference points. 3. Assign ﬁtness values using equation (7.1) (where d1 represents the distance from solution to the ideal point. Similarly, d2 represents the 7.4. GENERAL CONSIDERATIONS 7.4. GENERAL CONSIDERATIONS 19 ENERAL CONSIDERATIONS 191 7.4. GENERAL CONSIDERATIONS 191 inverse of the acute angle between solutions and reference points) to each solution and N solutions are selected based on their ﬁtness values. d FV = d1 + d2 θm . (7.1) (7.1) This method not only makes a greater contribution to the convergence performance of the entire population but also improves the diversity of the population. For evaluating the performance of the algorithm on the cloud scheduling problem, we perform simulation experiments using the Cloudsim platform. CHAPTER 7. CONCLUSIONS CHAPTER 7. CONCLUSIONS CHAPTER 7. CONCLUSIONS 192 A.1 Introduction Although the motivation of our research focuses on JSS problems, to show the general applicability, usefulness, and effectiveness of the MARP- NSGA-III (see Chapter 5). We also apply MARP-NSGA-III to the bench- mark test problems to further check the effectiveness of the MARP-NSGA- III. In the benchmark problem case, we have compared the performance of MARP-NSGA-III with NSGA-III [43] and A-NSGA-III [85] with three, ﬁve, and eight objectives. A-NSGA-III is an updated version of NSGA-III, where RVEA* is an extension of RVEA [31]. These extensions are mainly relocating reference points which were suggested to handle problems with irregular Pareto- front. IMMOEA [30] is a model-based method that constructs Gaussian process-based inverse models that map all the non-dominated solutions from the objective space to the decision space. In these experiments, we focus mainly on ten benchmark problems with irregular Pareto-front. These problems are DTLZ4, DTLZ5, and DTLZ7 [31], Inverted DTLZ1 (IDTLZ1), Inverted DTLZ2 (IDTLZ2) [76], WFG1, WFG2, WFG9 [31], MAF1,and MAF2 [92]. The characteristics of all the ten benchmark problems are summarized in Table A.1 of the next 193 194 APPENDIX A. FURTHER STUDIES Table A.1: The characteristics of benchmark problems Problems Characteristics IDTLZ1 inverted IDTLZ2 Convex, inverted DTLZ4 Concave,biased DTLZ5 Concave, degenerate DTLZ7 Mixed, disconnected, multi-model MAF1 Linear, multi-model, inverted MAF2 Disconnected WFG1 Sharp tails WFG2 Disconnected WFG9 Concave Table A.1: The characteristics of benchmark problems section. A.2 Benchmark functions on MaOPs Benchmark function are widely used in the literature to evaluate MaOPs algorithms’ performance on problems with irregular Pareto-front. In this thesis, we used Inverted DTLZ1 (IDTLZ1), Inverted DTLZ2 (IDTLZ2), DTLZ4, DTLZ5, DTLZ7,WFG1, WFG2, WFG9, MAF1, and MAF2 benchmark problems. The Pareto-front for these problems is also irregu- lar. MAF1 and MAF2 are taken from the CEC-2017 competition on evolu- tionary many-objective optimization [32]. The characteristics of all the 10 benchmark problems are summarized in Table A.1. It should be noted here that all the ﬁgures of benchmark problems are taken from MATLAB-based EMO platform called PlatEMO [187]. A.2.1 IDTLZ1 and IDTLZ2 problems IDTLZ1 and IDTLZ2 denote the problems of inverted DTLZ1 and DTLZ2, respectively, where the regular Pareto-front of DTLZ1 and DTLZ2 are in- 195 A.2. BENCHMARK FUNCTIONS ON MAOPS verted and thus become irregular [86]. This test problem is used to assess whether MOEA is capable of dealing with an inverted Pareto-front. Fig- ures A.1 (a) and A.1 (b) shows IDTLZ1 and IDTLZ2, respectively. (a) IDTLZ1. (b) IDTLZ2. Figure A.1: The Pareto-front with three objectives on IDTLZ problems. Figure A.1: The Pareto-front with three objectives on IDTLZ problems. A.2.3 MAF1 and MAF2 Problems MAF1 is a modiﬁed version of IDTLZ1. The feasible range of MAF1 is [0,1] on the objective space. This test problem is used to assess whether MOEAs are capable of dealing with inverted Pareto-front. MAF2 increases the difﬁculty of convergence in DTLZ2. In DTLZ2, all in the do not have to be optimized simultaneously in order to reach the true Pareto-front. In contrast, the MAF2 problem is used to assess whether the MOEAs can perform concurrent convergence on different objectives. Figures A.3 (a) and A.3 (b) show MAF1 and MAF2, respectively. A.2.2 DTLZ-4, DTLZ5, and IDTLZ-7 Problems The Pareto-front of DTLZ-4 has a badly-scaled Pareto-front (where each objective function is scaled to a substantially different range), especially when the ﬁtness landscape is highly multi-modal. DTLZ-5 is originally intended to be test problems with degenerated Pareto-front. The shape of the true Pareto-front is a 1-D curve which shows this problem independent of the number of objectives. The shape of the true Pareto-front of DTLZ-7 is disconnected. This problem is usually used to assess the capability of those MOEAs which have disconnected Pareto-front. Figures A.2 (a), A.2 (b), and A.2 (c) show DTLZ4, DTLZ5, and DTLZ7, respectively. APPENDIX A. FURTHER STUDIES 196 (a) DTLZ4. (b) DTLZ5. (c) DTLZ7. Figure A.2: The Pareto-front with three objectives on DTLZ problems. (b) DTLZ5. Figure A.2: The Pareto-front with three objectives on DTLZ problems. (a) MAF1. Figure A.3: The Pareto-front with three objectives on MAF problems. A.2.4 WFG1,WFG2 and WFG9 Problems The WFG test problems (WFG1, WFG2, and WFG9) have scaled true Pareto-front. These WFG test problems are widely used to test the per- formance of MOEAs on MaOPs [31]. WFG1 contains both convex and concave segments. Whereas the WFG2 test problem is capable of dealing with scaled disconnected Pareto-front. WFG9 has concave Pareto-front. Moreover, its ﬁtness landscape is highly multi-modal. Figures A.4 (a), A.4 (b), and A.4 (c) show WFG1, WFG2, and WFG9, respectively. A.3. EXPERIMENT DESIGN 197 (a) MAF1. (b) MAF2. Figure A.3: The Pareto-front with three objectives on MAF problems. Table A.3: Number of Reference Points and Population Size for WFG. No. of Obj(M) Ref. Points(H) Pop. Size(N) 3 105 108 5 175 176 8 217 220 and [31] respectively. The parameter setting of WFG is taken from [31]. No. of Obj(M) Ref. Points(H) Pop. Size(N) 3 105 108 5 175 176 8 217 220 and [31] respectively. The parameter setting of WFG is taken from [31]. A.3.1 Parameter setting of benchmark problems The number of decision variables for DTLZ and inverted DTLZ test prob- lems is set as recommended in [45]. The MAF problems, the solution di- mension is set according to [92], whereas, for WFG problems, we follow the recommendation of decision variables in [31]. The population sizes Table A.2: Number of Reference Points and Population Size for DTLZ and MAF. No. of Obj(M) Ref. Points(H) Pop. Size(N) 3 91 92 5 210 212 8 156 156 for DTLZ and MAF problems of all compared algorithms are shown in Table A.2. The settings for WFG are shown in Table A.3. For DTLZ and MAF problems, the maximum number of generations is adopted from [43] APPENDIX A. FURTHER STUDIES 198 (a) WFG1. (a) WFG1. (b) WFG2. (c) WFG9. Figure A.4: The Pareto-front with three objectives on WFG problems. (b) WFG2. (c) WFG9. Figure A.4: The Pareto-front with three objectives on WFG problems. Table A.3: Number of Reference Points and Population Size for WFG. Table A.3: Number of Reference Points and Population Size for WFG. A.4. RESULTS AND DISCUSSION A.4. RESULTS AND DISCUSSION 199 A.4 Results and discussion In the experiment, 30 independent runs are carried out for each algorithm. Then, the mean and standard deviations of HV and IGD values were re- ported. The best value for each problem was marked in boldface. The benchmark problems are substantially studied in the literature. The source codes of RVEA* [31], Two-ARCh [167], and IMMOEA [30] are available on different platforms. Therefore, we compare these algorithms with our pro- posed algorithms (MARP-NSGA-III) with three, ﬁve, and eight objectives. A.4.1 Performance of obtained solutions Tables A.4 and A.5 present the HV and IGD values scored by the proposed algorithm and other compared algorithms. In general, MARP-NSGA-III has achieved signiﬁcantly better performance on 15 of 30 instances in HV and 14 of 30 instances in IGD. On the other hand, RVEA* is a very com- petitive algorithm and has the best performance in nine instances. A.3.2 Algorithms parameter settings The widely used genetic operators, i.e., the simulated binary crossover (SBX) [43] and the polynomial mutation [42] are employed to create the offspring population, as in many other MOEAs [42, 43]. The evolution operator is kept identical for all approaches: (i) SBX probability (pc) is set to 1.0, (ii) polynomial mutation probability (pm) is set to 1/n (where n is the number of variables) [43]. The distribution index(ηc) is kept as 30 for SBX crossover and polynomial distribution index (ηm) are set to 20. For RVEA, the penalty parameter α is set to 2, and the frequency of reference point adaptation fr is set to 0.1. Performance on IDTLZ1 and IDTLZ2 IDTLZ1 has an inverted Pareto-front and many uniformly distributed ref- erence points created on the normalized hyper-plane will not have an asso- ciated Pareto-optimal point. Therefore, adaptive reference points play an important role in achieving a higher degree of population diversity than uniformly distributed reference points. Tables A.4 and A.5 show that A- NSGA-III, RVEA, and MARP-NSGA-III have better HV and IGD values because they can generate higher population diversity than the NSGA- III with predeﬁned uniformly distributed reference points. The statistical results in Tables A.4 and A.5 reveal that the MARP-NSGA-III performs signiﬁcantly better than the other methods in the ﬁve-objective instances. However, it has a competitive performance in eight-objective cases. In the APPENDIX A. FURTHER STUDIES 200 Table A.4: The mean and standard deviation (¯x±σ) over the average HV values on M-objectives on IDTLZ1, IDTLZ2, DTLZ4,DTLZ5, DTLZ7, MAF1, MAF2, WFG1, WFG2 problems. Table A.4: The mean and standard deviation (¯x±σ) over the average HV values on M-objectives on IDTLZ1, IDTLZ2, DTLZ4,DTLZ5, DTLZ7, MAF1, MAF2, WFG1, WFG2 problems. HV Mean(std) Function M NSGA-III A-NSGA-III RVEA* Two-ARCh IMMOEA MARP-NSGA3 Inv-DTLZ1 3 1.07e-1(4.0e-3) 1.30e-1(2.8e-3) 1.32e-1(3.9e-3) 1.19e-1(1.5e-3) 1.29e-1(3.00e-3) 1.33e-1(3.4e-3) 5 3.91e-3(4.2e-4) 4.2e-3(4.9e-4) 4.23e-3(3.4e-5) 4.32e-3(8.4e-6) 4.34e-3(1.5e-4) 5.3e-3(6.7e-4) 8 6.22e-4(3.8e-5) 7.05e-4(1.5e-3) 6.66e-4(8.8e-5) 5.81e-4(3.7e-3) 5.71e-4(1.9e-5) 6.67e-4(4.0e-4) Inv-DTLZ2 3 4.52e-1(3.1e-3) 4.72e-1(3.0e-3) 4.58e-1(9.5e-4) 4.857e-2(7.0e-3) 5.26e-1(8.8e-3) 4.62e-1(3.7e-3) 5 6.30e-2(2.4e-3) 6.89e-2(5.9e-3) 7.00e-2(2.9e-3) 9.94e-2(2.8e-3) 8.60e-2(5.5e-3) 1.07e-1(6.9e-3) 8 6.62e-3(6.8e-4) 9.37e-3(1.1e-4) 2.7e-3(4.3e-4) 1.47e-1(4.9e-4) 2.577e-3(3.7e-4) 1.54e-2(1.7e-4) DTLZ4 3 7.02e-1(1.2e-2) 7.47e-1(3.1e-3) 7.34e-1(1.1e-2) 7.19e-1(4.1e-2) 6.97e-1(5.1e-2) 7.66e-1(5.5e-4) 5 7.03e-1(2.6e-3) 7.17e-1(1.2e-4) 7.40e-1(4.2e-4) 7.32e-1(3.7e-3) 7.20e-(5.2e-4) 7.59e-1(3.7e-3) 8 8.45e-1(2.6e-4) 8.59e-1(5.5e-3) 8.71e-1(1.6e-4) 8.67e-1(9.7e-4) 8.57e-1(1.7e-4) 8.74e-1(9.4e-4) DTLZ5 3 8.19e-2(1.7e-2) 8.55e-2(6.3e-4) 9.19e-2(2.9e-3) 1.3e-1(6.1e-4) 8.23e-2(1.4e-3) 8.49e-2(7.2e-4) 5 5.28e-1(2.6e-2) 6.9e-1(3.7e-2) 7.48e-1(2.8e-4) 6.6e-3(1.6e-4) 7.87e-1(2.2e-4) 5.3e-1(4.4e-2) 8 6.03e-1(1.9e-2) 6.40e-1(2.3e-2) 5.62e-1(1.7e-4) 6.37e-1(3.8e-6 5.56e-1(4.99e-2)) 6.38e-1(2.1e-2) DTLZ7 3 3.25e-1(7.9e-3) 3.48e-1(9.4e-3) 3.44e-1(8.3e-3 3.53e-1(5.8e-2) 3.14e-1(9.6e-2) 3.58e-1(1.2e-3) 5 2.240e-1(3.8e-3) 3.23e-1(6.3e-3) 2.970e-1(2.1e-2) 1.35e-1(2.3e-2) 3.21e-1(1.8e-3) 3.5e-1(7.2e-2) 8 3.08e-1(4.8e-3) 3.25e-1(6.3e-3) 2.98e-1(1.7e-3) 1.56e-1(4.9e-4) 2.56e-1(4.99e-2) 3.31e-1(2.3e-2) MAF1 3 1.15e-1(2.0e-3) 1.34e-1(3.0e-4) 2.88e-1 (1.0e-3) 2.20e-1 (6.59e-3) 2.80e-1 (7.0e-4) 1.38e-2(3.4e-3) 5 2.04e-2(2.9e-4) 2.62-2(3.8e-4) 6.35e-2 (7.93e-4) 2.04e-2(3.01e-4) 7.23e-3 (5.50e-4) 7.18e-2(4.77e-3) 8 6.57e-4(1.0e-5) 6.24-4(6.9e-5) 6.69e-4(4.9e-5) 6.60e-4(4.3e-5) 6.44e-4(3.9e-4) 6.75e-4(4.7e-5) MAF2 3 3.30e-1(2.1e-2) 4.05e-1(3.9e-3) 4.19e-1(1.4e-3) 3.66e-1(3.0e-3) 3.48e-1(2.2e-3) 3.94e-1(2.3e-3) 5 1.22e-1(1.1e-2) 1.57e-1(1.3e-2) 1.61e-1(2.3e-2) 1.52e-1(2.1e-2) 1.38e-1(4.3e-2) 1.68e-1(5.3e-2) 8 1.77e-1(1.82e-3) 2.03e-1(3.2e-3) 1.60e-1(2.4e-3) 1.57e-1(9.7e-3) 1.32e-1(6.6e-3) 2.30e-1(4.2e-3) WFG1 3 8.48e-1(5.2e-2) 8.44e-1(5.0e-2) 8.70e-1 (3.3e-2) 8.56e-1 (3.9e-2) 8.11e-1 (2.74e-2) 8.22e-1(8.4e-2) 5 7.09e-1(1.6e-2) 8.39e-1(1.9e-2) 8.33e-1(3.4e-2) 8.32e-1(5.1e-2) 8.06e-1(2.6e-2) 8.29e-1(1.9e-2) 8 6.09e-1(4.4e-2) 8.35e-1(5.9e-2) 8.29e-1(7.4e-2) 8.34e-1(3.3e-2) 7.86e-1(4.6e-2) 8.38e-1(4.9e-2) WFG2 3 7.07e-1(4.99e-3) 8.44e-1(8.3e-2) 8.55e-1 (4.3e-2) 7.61e-1 (4.5e-2) 7.70e-1 (2.74e-2) 8.42e-1(7.0e-2) 5 9.89e-1(2.2e-3) 9.53e-1(6.4e-2) 9.86e-1(2.6e-2) 9.80e-1(5.6e-2) 9.36e-1(3.6e-2) 9.89e-1(3.3e-2) 8 9.22e-1(9.5e-2) 9.40e-1(7.5e-2) 9.52e-1(8.6e-2) 9.46e-1(7.6e-2) 9.17e-1(7.2e-2) 9.41e-1(6.0e-2) WFG9 3 6.20e-1(4.8e-2) 6.75e-1(4.9e-2) 6.64e-1(1.4e-2) 6.66e-1(3.0e-2) 6.48e-1(3.0e-2) 6.68e-1(3.9e-2) 5 6.22e-1(6.5e-2) 6.52e-1(5.6e-2) 7.0e-1(5.0e-2) 6.69e-1(2.1e-2) 6.38e-1(5.7e-2) 6.53e-1(5.5e-2) 8 7.19e-1(6.9e-2) 7.28e-1(3.2e-2) 8.28e-1(1.9e-2) 7.77e-1(1.7e-2) 7.07e-1(2.7e-2) 8.31e-1(4.9e-2) A.4. RESULTS AND DISCUSSION 201 Table A.5: The mean and standard deviation (¯x ± σ) over the average IGD values on M-objectives on IDTLZ1, IDTLZ2, DTLZ4, DTLZ5, and DTLZ7, MAF1, MAF2, WFG1, WFG2 problems. Table A.4: The mean and standard deviation (¯x±σ) over the average HV values on M-objectives on IDTLZ1, IDTLZ2, DTLZ4,DTLZ5, DTLZ7, MAF1, MAF2, WFG1, WFG2 problems. IGD Mean(std) Function M NSGAIII A-NSGA-III RVEA* Two-ARCh IMMOEA MARP-NSGAIII IDTLZ1 3 3.58E-2(1.0e-3) 2.47e-2(8.0e-3) 2.37-02(2.7e-2) 1.66e-1(1.6e-2) 4.22e-1(1.0e-3) 2.38e-2(2.1e-3) 5 8.62e-2(3.8e-3) 5.53e-2(8.3e-3) 1.47e-1(2.0e-2) 1.90e-1(3.2e-2) 6.75e-1 (4.36e-1) 3.13e-2(8.3e-3) 8 9.62e-2(8.8e-3) 5.13e-2(9.2e-3) 2.45e-1(2.5e-2) 1.96e-1(2.0e-2) 4.85e-1(3.0e-2) 5.49e-2(8.9e-3) IDTLZ2 3 6.73e-2(8.3e-3) 6.42e-2(5.6e-3) 8.07e-2(1.0e-3) 9.0e-2(7.0e-3) 6.20e-2(4.0e-3) 6.26e-2(2.5e-3) 5 2.33e-1(1.2e-2) 2.03e-1(1.2e-2) 2.73e-1(1.3e-3) 2.1e-1(5.8e-3) 3.36e-1 (1.2e-2) 1.78e-1(1.8e-2) 8 5.62e-1(3.8e-2) 5.16e-1(2.3e-2) 6.10e-1(6.5e-3) 4.1e-1(6.7e-3) 5.44e-1(1.5e-2) 3.11e-1(2.3e-2) DTLZ4 3 7.05e-2(6.01e-3) 6.26e-2(6.2e-3) 7.07e-2(8.4e-2) 9.06e-2(1.4e-2) 7.57e-2(2.6e-3) 6.2e-2(6.43e-3) 5 1.63e-02(4.3e-4) 1.64e-2(5.8e-4) 1.53e-2(5.4e-4) 1.71e-2(1.4e-4) 1.68e-2(3.9e-4) 1.39e-2(3.9e-4) 8 2.68e-02(1.0e-3) 4.07e-3(5.7e-4) 3.25e-2(2.63e-3) 2.28e-2(1.33e-3) 1.22e-1(1.03e-3) 2.16e-2(1.9e-3) DTLZ5 3 2.18e-2(2.2e-3) 1.20e-2(1.7e-3) 1.37e-2(1.0e-3) 1.22e-2(7.0e-3) 1.96e-2(4.0e-3) 1.75e-2(1.4e-3) 5 2.70e-1(5.16e-2) 1.99e-1(5.7e-2) 2.48e-1(2.8e-4) 2.6e-3(1.6e-4) 1.87e-1(2.2e-4) 2.26e-1(3.9e-2) 8 4.04e-1(8.8-e-2) 4.01e-1(7.2e-2) 3.62e-1(1.7e-4) 6.377e-1(3.8e-6) 5.56e-1(4.99e-2)) 4.3e-1(9.1e-2) DTLZ7 3 9.44e-2(4.99e-3) 6.79e-2(5.2e-3) 7.05e-2(1.56e-3) 7.15e-2(8.5e-2) 8.34e-2(3.9e-2) 6.75e-2(3.03e-3) 5 4.54e-1(2.2e-2) 3.44e-1(2.4e-2) 4.98e-1(9.0e-3) 5.3e-1(9.0e-2) 5.65e-1 (2.74e-2) 3.08e-1(4.1e-2) 8 7.89e-1(3.7e-2) 7.61e-1(2.7e-2) 7.82e-1(1.6e-1) 7.53e-1(8.6e-2) 8.14e-1(9.6e-2) 7.44e-1(8.6e-2) MAF1 3 6.44e-2(2.3e-3) 5.07e-2(1.1e-3) 4.59e-2 (9.4e-4) 4.37e-2 (4.3e-4) 1.09e-1 (9.9e-3) 4.67e-2(3.6e-3) 5 1.72e-1(9.4e-3) 1.75e-1(2.3e-3) 1.46e-1 (7.2e-3) 1.52e-1 (8.42e-4) 2.05e-1 (7.41e-3) 1.34e-1 (2.4e-3) 8 3.92e-1(1.1e-2) 3.28e-1(1.8e-2) 2.90e-1 (2.1e-2) 2.25e-1 (3.3e-2) 3.02e-1 (8.57e-3) 1.73e-1 (2.1e-2) MAF2 3 4.42e-2(1.6e-2) 3.24e-2(2.7e-3) 2.98e-2(5.4e-3) 3.15e-2(1.4e-3) 3.22e-2(1.0e-3) 3.27e-2(2.2e-3) 5 1.22e-1(1.1e-2) 1.07e-1(1.3e-2) 9.03e-2(6.4e-2) 1.05e-1(1.4e-2) 1.22e-1(1.0e-3) 8.6e-2(1.3e-3) 8 2.05e-01(1.3e-2) 1.63e-1(1.0e-2) 2.50e-1(1.4e-2) 1.69e-1(2.0e-3) 2.22e-1(2.4e-3) 1.44e-1(1.6e-3) WFG1 3 1.29e-1(5.16e-2) 1.18e-1(4.4e-2) 1.17e-1(4.43e-2) 1.25e-1(3.13e-2) 4.95e-1(2.03e-2) 1.16e-1(7.43e-2) 5 3.25e-1(2.2e-2) 2.01e-1(1.5e-2) 2.40e-1(6.6e-2) 3.75e-1(4.3e-2) 2.54e0(4.1e-2) 2.09e-1(1.4e-2) 8 1.86e-1(9.2e-2) 1.77e-1(4.43e-2) 1.73e-1(2.03e-2) 1.71e-1(7.43e-2) 1.79e-1(3.6e-1) 1.72e-1(8.6e-1) WFG2 3 8.02e-1(4.52e-2) 8.40e-1(3.9e-2) 8.66e-1(5.43e-3) 6.75e-2(4.43e-3) 6.95e-2(3.03e-3) 8.86e-1(6.43e-3) 5 6.54e-1(2.2e-2) 6.47e-1(3.8e-2) 6.37e-1(8.5e-2) 6.85e-1(9.9e-2) 6.50e-1(6.6e-2) 6.35e-1(3.9e-2) 8 1.42e0(9.1e-2) 1.26e0(7.9e-2) 2.04e0(8.9e-1) 1.08e0(4.9e-2) 1.50e0(4.6e-1) 1.15e0(3.6e-1) WFG9 3 2.75e-01(8.35e-2) 2.350e-1(3.9e-2) 1.86e-1(3.43e-2) 3.25e-1(1.43e-2) 3.35e-1(2.53e-3) 1.37e-1(3.43e-3) 5 9.75e-01(1.0e-2) 9.37e-1(8.3e-2) 9.65e-1(8.05e-2) 9.85e-1(1.0e-2) 9.88e-1(9.77e-2) 9.78e-1(9.05e-2) 8 1.55e0(1.7e-1) 1.49e0(1.48e-1) 1.48e0(5.43e-1) 1.88e0(3.09e-1) 1.92e-1(3.03e-1) 1.45e0(1.74e-1) 201 APPENDIX A. FURTHER STUDIES 202 case of three-objective, MARP-NSGA-III is slightly better in HV than A- NSGA-III, RVEA, and A-NSGA-III. Relevant results on problem instances with three-objective problems can also be seen in Figures A.5 (a) to A.5 (c). The approximate Pareto-front of each algorithm shows that A-NSGA- III, RVEA, and MARP-NSGA-III have a similar distribution of obtained points. 0 0.2 0.4 0.6 0.8 0 0.5 1 0 0.2 0.4 0.6 0.8 (a) ANSGA-III. 0 0.2 0.4 0.6 0.8 0 0.5 1 0 0.2 0.4 0.6 0.8 (b) RVEA. 0 0.2 0.4 0.6 0.8 0 0.5 1 0 0.2 0.4 0.6 0.8 (c) MARP-NSGA-III. Figure A.5: Approximate Pareto-front for 3-objective IDTLZ1 problem (c) MARP-NSGA-III. Figure A.5: Approximate Pareto-front for 3-objective IDTLZ1 problem For the IDTLZ2 problem, Tables A.4 and A.5 show that IMMOEA achieved a signiﬁcantly better result compared to all other competing al- gorithms in terms of HV and IGD with three-objective problems. Table A.4: The mean and standard deviation (¯x±σ) over the average HV values on M-objectives on IDTLZ1, IDTLZ2, DTLZ4,DTLZ5, DTLZ7, MAF1, MAF2, WFG1, WFG2 problems. How- ever, IMMOEA performed poorly on problems with more than three ob- jectives. In comparison, MARP-NSGA-III achieved the best performance among all the competing algorithms on ﬁve objectives in terms of HV. However, in terms of IGD, MARP-NSGA-III outperforms all the other competing algorithms on eight-objective problems. Performance on DTLZ4, DTLZ5, and DTLZ7 DTLZ4 has a biased density (the density of solutions is different from one location to another) targeted points on the true Pareto-front. This prob- lem is to verify whether many-objective algorithms can maintain a proper distribution of the solutions. Table A.4 conﬁrms that MARP-NSGA-III achieves overall the best performance in terms of HV. However, Table A.5 A.4. RESULTS AND DISCUSSION 203 reveals that MARP-NSGA-III is signiﬁcantly better than the other compet- ing algorithms on three-objective and ﬁve-objective. For three-objective, Figures A.6 (a) and A.6 (b) show that MARP-NSGA-III has widely dis- tributed the solution as compared to A-NSGA-III. In terms of HV, the MARP-NSGA-III performed signiﬁcantly better than the other compared algorithms on eight-objective problems. By contrast, A-NSGA-III has the best IGD value over eight-objective problems when compared to other competing algorithms. (a) A-NSGA-III. 0 0.5 1 1.5 0 0.5 1 1.5 0 0.5 1 1.5 (b) MARP-NSGA-III. Figure A.6: Approximate Pareto-front for 3-objective DTLZ4 problem 0 0.5 1 1.5 0 0.5 1 1.5 0 0.5 1 1.5 (b) MARP-NSGA-III. Figure A.6: Approximate Pareto-front for 3-objective DTLZ4 problem DTLZ5 has a degenerated Pareto-front, i.e., the Pareto-front is always a curve regardless of the objective space’s dimensionality. For the three ob- jective DTLZ5, Two-ARCh performed signiﬁcantly better than other com- peting algorithms in terms of HV. The result can be observed in Table A.4. This result is also illustrated in Figures A.7 (a) and A.7 (b) where Two- ARCh has well-diversiﬁed optimal solutions. It can be observed in Tables A.4 and A.5 that no single algorithm can perform consistently better than the rest algorithms on DTLZ5. DTLZ7 has a disconnected Pareto-front. Due to this feature, those al- gorithms that rely on uniformly distributed reference points performed poorly. Tables A.4 and A.5 reveal that the algorithms having adaptive reference points eventually have signiﬁcantly better performance. In the APPENDIX A. FURTHER STUDIES 204 0 0.5 1 0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8 1 1.2 1.4 0 0.5 1 0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8 1 1.2 1.4 (a) Two-ARCh. 0 0.5 1 0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8 1 1.2 1.4 (b) MARP-NSGA-III. Figure A.7: Approximate Pareto-front for 3-objective DTLZ5 problem 0 0.5 1 0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8 1 1.2 1.4 (b) MARP-NSGA-III. Performance on DTLZ4, DTLZ5, and DTLZ7 Figure A.7: Approximate Pareto-front for 3-objective DTLZ5 problem three-objective case, our proposed algorithm (MARP-NSGA-III) is slightly better than Two-ARCh in terms of HV. However, MARP-NSGA-III outper- forms RVEA which can also be seen in Figures A.8 (a) and A.8 (b). These ﬁgures show that MARP-NSGA-III has well-diversiﬁed solutions in three out of four regions of the Pareto-front. RVEA* apparently failed to cover some regions on the Pareto-front. In ﬁve-objective and eight-objective cases, MARP-NSGA-III also performs the best which can be seen in Ta- bles A.4 and A.5. Meanwhile, as evidenced in Figures A.9 (a) and A.9 (b), MARP-NSGA-III can evolve more diversiﬁed solutions than A-NSGA-III. Performance on MAF1 and MAF2 MAF1 and MAF2 are taken from the CEC-2017 competition on MaOPs [32]. Moreover, they have irregular Pareto-front. MAF1 is a modiﬁed inverted DTLZ1 which has an inverted Pareto-front. This problem has the same structure as IDTLZ1. Therefore, we can see similar behavior of IDTLZ1 in the result. This can be seen in Tables A.4 and A.5. The exper- imental results reveal that the MARP-NSGA-III performed signiﬁcantly better on problems with ﬁve and eight objectives than other competing A.4. RESULTS AND DISCUSSION 205 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 5 10 (a) RVEA. 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 5 10 (b) MARP-NSGA-III. 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 5 10 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 5 10 (a) RVEA. (b) MARP-NSGA-III. (a) RVEA. (b) MARP-NSGA-III. Figure A.8: Approximate Pareto-front for 3-objective DTLZ7 problem Figure A.8: Approximate Pareto-front for 3-objective DTLZ7 problem Figure A.8: Approximate Pareto-front for 3-objective DTLZ7 problem Obji Obj2 Obj3 Obj4 Obj5 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (a) A-NSGA-III. obj1 obj2 obj3 obj4 obj5 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (b) MARP-NSGA-III. Figure A.9: Parallel coordinate plot for the ﬁtness values of the population for 5-objective DTLZ7 problem obj1 obj2 obj3 obj4 obj5 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values Obji Obj2 Obj3 Obj4 Obj5 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (b) MARP-NSGA-III. Figure A.9: Parallel coordinate plot for the ﬁtness values of the population for 5-objective DTLZ7 problem algorithms. In Figures A.9 (a) and A.9 (b), the parallel plots show that optimal solutions obtained from MARP-NSGA-III are widely distributed. MAF2 is a modiﬁed form of DTLZ2, with the higher irregularity of the Pareto-front. Tables A.4 and A.5 show that MARP-NSGA-III has achieved APPENDIX A. FURTHER STUDIES 206 Obji Obj2 Obj3 Obj4 Obj5 Objectives 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Objectives Values (a) RVEA. Obji Obj2 Obj3 Obj4 Obj5 Objectives 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Objectives Values (b) MARP-NSGA-III. Performance on MAF1 and MAF2 Figure A.10: Parallel coordinate plot for the ﬁtness values of the popula- tion for 5-objective MAF1 problem Obji Obj2 Obj3 Obj4 Obj5 Objectives 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Objectives Values Obji Obj2 Obj3 Obj4 Obj5 Objectives 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Objectives Values Figure A.10: Parallel coordinate plot for the ﬁtness values of the popula- tion for 5-objective MAF1 problem signiﬁcantly better results on MAF2 as compared to the other algorithms on ﬁve-objective problems. We can see that in Figures A.11 (a) and A.11 (b), MARP-NSGA-III has better diversity than its competitor RVEA. In conclusion, MARP-NSGA-III outperforms the other algorithms on ﬁve-objective and eight-objective MAF problems. Performance on WFG1,WFG2, and WFG9 Similar observations of MARP-NSGA-III can be made from the results on WFG in Table A.4 where MARP-NSGA-III outperforms the other algo- rithms on the eight-objective instances. RVEA has also shown compet- itive performance on WFG problems. WFG1 is a mixed structure problem involving many transformation functions in the problem deﬁnition. These transformation functions make it hard to produce well-diversiﬁed solutions [92]. We can observe from A.4. RESULTS AND DISCUSSION 207 Obji Obj2 Obj3 Obj4 Obj5 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (a) RVEA* Obji Obj2 Obj3 Obj4 Obj5 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values (b) MARP-NSGA-III. Figure A.11: Parallel coordinate plot for the ﬁtness values of the popula- tion for 5-objective MAF2 problem Obji Obj2 Obj3 Obj4 Obj5 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values Obji Obj2 Obj3 Obj4 Obj5 0 0.2 0.4 0.6 0.8 1 Objectives Objectives Values Figure A.11: Parallel coordinate plot for the ﬁtness values of the popula- tion for 5-objective MAF2 problem Table A.4 and Table A.5 that RVEA* performs better than the other com- pared algorithms with three objectives problems, while A-NSGA-III out- performs the other algorithms on WFG1 with ﬁve-objective problems. In the case of eight objectives, our proposed algorithm MARP-NSGA-III achieved a higher HV value than other algorithms. WFG2 is the only WFG problem with disconnected Pareto-front. It can be observed from Tables A.4 and A.5 that the MARP-NSGA-III outper- forms the other algorithms with ﬁve-objective. This can be seen in Figures A.12 (a) to A.12 (c). While RVEA* has better performance than A-NSGA- III. It performed signiﬁcantly worse performance than MARP-NSGA-III. This is because MARP-NSGA-III can produce well distributed and well- diversiﬁed solutions. In the meantime, MARP-NSGA-III achieves compet- itive performance as RVEA* on problem instances with eight objectives. MARP-NSGA-III shows promising performance on some eight-objective. Moreover, RVEA* shows generally competitive performance. WFG9 has a scaled concave Pareto-front and its decision variables are non-separable. Tables A.4 and A.5 show that MARP-NSGA-III achieved APPENDIX A. FURTHER STUDIES 208 Obji Obj2 Obj3 Obj4 Obj5 Objectives 0 0.2 0.4 0.6 0.8 1 1.2 Objectives Values (a) A-NSGA-III. Obji Obj2 Obj3 Obj4 Obj5 Objectives 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Objectives Values (b) RVEA. Performance on WFG1,WFG2, and WFG9 Obji Obj2 Obj3 Obj4 Obj5 Objectives 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Objectives Values (c) MARP-NSGA-III Figure A.12: Parallel coordinate plot for the ﬁtness values of the popula- tion for 5-objective WFG2 problem Obji Obj2 Obj3 Obj4 Obj5 Objectives 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Objectives Values (c) MARP-NSGA-III Obji Obj2 Obj3 Obj4 Obj5 Objectives 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Objectives Values Obji Obj2 Obj3 Obj4 Obj5 Objectives 0 0.2 0.4 0.6 0.8 1 1.2 Objectives Values Figure A.12: Parallel coordinate plot for the ﬁtness values of the popula- tion for 5-objective WFG2 problem leading performance on WFG9 with eight objectives. Figures A.13 (a) and A.13 (b) show that MARP-NSGA-III can produce more diversiﬁed solu- tions than RVEA. Obji Obj2 Obj3 Obj4 Obj5 Obj6 Obj7 Obj8 0 2 4 6 8 10 12 14 16 18 Objectives Objectives Values (a) RVEA. Obji Obj2 Obj3 Obj4 Obj5 Obj6 Obj7 Obj8 0 2 4 6 8 10 12 14 16 18 Objectives Objectives Values (b) MARP-NSGA-III. Figure A.13: Parallel coordinate plot for the ﬁtness values of the popula- tion for eight-objective WFG9 problem Obji Obj2 Obj3 Obj4 Obj5 Obj6 Obj7 Obj8 0 2 4 6 8 10 12 14 16 18 Objectives Objectives Values Obji Obj2 Obj3 Obj4 Obj5 Obj6 Obj7 Obj8 0 2 4 6 8 10 12 14 16 18 Objectives Objectives Values Figure A.13: Parallel coordinate plot for the ﬁtness values of the popula- tion for eight-objective WFG9 problem As we can observe in Tables A.4 and A.5, MARP-NSGA-III shows sig- niﬁcantly better results on most of the ﬁve-objective and eight-objective A.4. RESULTS AND DISCUSSION 209 APPENDIX A. FURTHER STUDIES A.4. RESULTS AND DISCUSSION instances as compared to other algorithms. By contrast, RVEA shows promising performance on some of the eight-objective instances. A- NSGA-III and Two-ARCh show effectiveness on three objective problems. Based on the experimental results, we can conclude that MARP-NSGA-III shows better performance on many benchmark problems. This algorithm performs particularly well on many irregular Pareto-front problems, such as inverted, disconnected, and scaling problems. APPENDIX A. FURTHER STUDIES APPENDIX A. FURTHER STUDIES 210 Bibliography [1] AARTS, E., AND LENSTRA, J. K. Local search in combinatorial opti- mization. Wiley, Chichester, 1997. [2] ABDUL-RAZAQ, T. S., POTTS, C. N., AND VAN WASSENHOVE, L. N. A survey of algorithms for the single machine total weighted tardiness scheduling problem . Discrete Applied Mathematics 26, 2–3 (1990), 235–253. [3] ADAMS, J., BALAS, E., AND ZAWACK, D. The shifting bottleneck procedure for job shop scheduling. Management science 34, 3 (1988), 391–401. [4] ADIBI, M. A., ZANDIEH, M., AND AMIRI, M. Multi-objective scheduling of dynamic job shop using variable neighborhood search. Expert Syst. Appl. 37, 1 (2010), 282–287. [5] AITZAI, A., BENMEDJDOUB, B., AND BOUDHAR, M. Branch-and- bound and PSO algorithms for no-wait job shop scheduling. http: //dx.doi.org/10.1007/s10845-014-0906-7, 2014. [6] ALPAYDIN, E. Introduction to Machine Learning, 2 ed. MIT press, 2010. [7] APPLEGATE, D., AND COOK, W. A Computational Study of the Job- Shop Scheduling Problem. ORSA Journal on Computing 3, 2 (1991), 149–156. 211 212 BIBLIOGRAPHY [8] APPLEGATE, D., AND COOK, W. A computational study of the job- shop scheduling problem. ORSA Journal on computing 3, 2 (1991), 149–156. [8] APPLEGATE, D., AND COOK, W. A computational study of the job- shop scheduling problem. ORSA Journal on computing 3, 2 (1991), 149–156. [9] ATTAR, S., MOHAMMADI, M., AND TAVAKKOLI-MOGHADDAM, R. Hybrid ﬂexible ﬂowshop scheduling problem with unrelated paral- lel machines and limited waiting times. The International Journal of Advanced Manufacturing Technology 68, 5-8 (2013), 1583–1599. [10] AVANTHAY, C., HERTZ, A., AND ZUFFEREY, N. A variable neigh- borhood search for graph coloring. European Journal of Operational Research 151, 2 (2003), 379–388. [11] BAKER, K. R., AND TRIETSCH, D. Principles of sequencing and schedul- ing. John Wiley & Sons, 2013. [12] BENI, G., AND WANG, J. Swarm intelligence in cellular robotic systems. In Robots and Biological Systems: Towards a New Bionics? Springer, 1993, pp. 703–712. [13] BŁA ˙ZEWICZ, J., DOMSCHKE, W., AND PESCH, E. The job shop scheduling problem: Conventional and new solution techniques. European Journal of Operational Research 93, 1 (1996), 1–33. [14] BLOT, A., JOURDAN, L., AND KESSACI, M.- ´E. Automatic design of multi-objective local search algorithms: case study on a bi-objective permutation ﬂowshop scheduling problem. In Proceedings of the Ge- netic and Evolutionary Computation Conference (2017), ACM, pp. 227– 234. [15] BOSMAN, P. A., AND THIERENS, D. The naive MIDEA: A baseline multi–objective EA, x-fetchedfrom = Google Scholar, year = 2005. Bibliography In International Conference on Evolutionary Multi-Criterion Optimization, Springer, pp. 428–442. 213 BIBLIOGRAPHY [16] BRANDIMARTE, P. Routing and scheduling in a ﬂexible job shop by tabu search. Annals OR 41, 3 (1993), 157–183. [17] BRANKE, J., NGUYEN, S., PICKARDT, C., AND ZHANG, M. Auto- mated Design of Production Scheduling Heuristics: A Review. IEEE Trans. Evolutionary Computation 20, 1 (2016), 110–124. [18] BRINGMANN, K., AND FRIEDRICH, T. An Efﬁcient Algorithm for Computing Hypervolume Contributions**. Evol. Comput. 18, 3 (Sept. 2010), 383–402. [19] BRUCKER, P., JURISCH, B., AND SIEVERS, B. A branch and bound al- gorithm for the job-shop scheduling problem. Discrete applied math- ematics 49, 1 (1994), 107–127. [20] BURKE, E. K., HYDE, M. R., KENDALL, G., OCHOA, G., OZCAN, E., AND WOODWARD, J. R. Exploring hyper-heuristic method- ologies with genetic programming. In Computational intelligence. Springer, 2009, pp. 177–201. [21] BURKE, E. K., HYDE, M. R., KENDALL, G., AND WOODWARD, J. Automatic heuristic generation with genetic programming: evolv- ing a jack-of-all-trades or a master of one. In Proceedings of the 9th an- nual conference on Genetic and evolutionary computation (2007), ACM, pp. 1559–1565. [22] BYRNE, J., SVOROBEJ, S., GIANNOUTAKIS, K. M., TZOVARAS, D., BYRNE, P. J., ¨OSTBERG, P.-O., GOURINOVITCH, A., AND LYNN, T. A review of cloud computing simulation platforms and related envi- ronments. In International Conference on Cloud Computing and Services Science (2017), vol. 2, SCITEPRESS, pp. 679–691. [23] C¸ ALIS¸, B., AND BULKAN, S. A research survey: Review of AI solu- tion strategies of job shop scheduling problem. Journal of Intelligent Manufacturing 26, 5 (2015), 961–973. 214 BIBLIOGRAPHY [24] CARLIER, J. The One-Machine Sequencing Problem. European Jour- nal of Operational Research 11, 1 (1982), 42–47. [25] CARLIER, J., AND PINSON, E. An Algorithm for Solving the Job- shop Problem. Management Science 35, 2 (1989), 164–176. [26] CHEN, B., ZENG, W., LIN, Y., AND ZHANG, D. A new local search- based multiobjective optimization algorithm. IEEE Transactions on Evolutionary Computation 19, 1 (2015), 50–73. [27] CHEN, S.-W., AND CHIANG, T.-C. Evolutionary many-objective optimization by MO-NSGA-II with enhanced mating selection. In IEEE Congress on Evolutionary Computation (2014), IEEE, pp. 1397– 1404. [28] CHENG, R., GEN, M., AND TSUJIMURA, Y. A tutorial survey of job- shop scheduling problems using genetic algorithms, part II: hybrid genetic search strategies . Computers & Industrial Engineering 36, 2 (1999), 343–364. [29] CHENG, R., HE, C., JIN, Y., AND YAO, X. Model-based evolutionary algorithms: a short survey. Bibliography Complex & Intelligent Systems 4, 4 (2018), 283–292. [30] CHENG, R., JIN, Y., NARUKAWA, K., AND SENDHOFF, B. A multi- objective evolutionary algorithm using Gaussian process-based in- verse modeling. IEEE Transactions on Evolutionary Computation 19, 6 (2015), 838–856. [31] CHENG, R., JIN, Y., OLHOFER, M., AND SENDHOFF, B. A refer- ence vector guided evolutionary algorithm for many-objective opti- mization. IEEE Transactions on Evolutionary Computation 20, 5 (2016), 773–791. 215 BIBLIOGRAPHY [32] CHENG, R., LI, M., TIAN, Y., ZHANG, X., YANG, S., JIN, Y., AND YAO, X. A benchmark test suite for evolutionary many-objective optimization. Complex & Intelligent Systems 3, 1 (2017), 67–81. [32] CHENG, R., LI, M., TIAN, Y., ZHANG, X., YANG, S., JIN, Y., AND YAO, X. A benchmark test suite for evolutionary many-objective optimization. Complex & Intelligent Systems 3, 1 (2017), 67–81. [33] CHOI, I.-C., AND CHOI, D.-S. A local search algorithm for jobshop scheduling problems with alternative operations and sequence- dependent setups. Computers & Industrial Engineering 42, 1 (2002), 43–58. [34] COLORNI, A., DORIGO, M., MANIEZZO, V., AND TRUBIAN, M. Ant system for job-shop scheduling. Belgian Journal of Operations Research, Statistics and Computer Science 34, 1 (1994), 39–53. [35] CORNELL, J. A. Experiments with mixtures: designs, models, and the analysis of mixture data, vol. 403. John Wiley & Sons, 2011. [36] CORRIVEAU, G., GUILBAULT, R., TAHAN, A., AND SABOURIN, R. Bayesian network as an adaptive parameter setting approach for ge- netic algorithms. Complex & Intelligent Systems 2, 1 (2016), 1–22. [37] COWLING, P., KENDALL, G., AND SOUBEIGA, E. A Hyperheuristic Approach to Scheduling a Sales Summit. In Practice and Theory of Au- tomated Timetabling III, vol. 2079 of Lecture Notes in Computer Science. Springer Berlin Heidelberg, 2001, pp. 176–190. [38] CROES, A. A method for solving traveling salesman problems. Op- erations Research 5 (1958), 791—812. [39] DAS, I., AND DENNIS, J. E. Normal-boundary intersection: A new method for generating the Pareto surface in nonlinear multicriteria optimization problems. SIAM Journal on Optimization 8, 3 (1998), 631–657. [40] DAVIS, L. Evolutionary algorithms. The IMA volumes in mathematics and its applications. Springer, 1999. 216 BIBLIOGRAPHY [41] DE BONET, J. S., ISBELL JR, C. L., AND VIOLA, P. A. MIMIC: Find- ing optima by estimating probability densities. In Advances in neural information processing systems (1997), pp. 424–430. [42] DEB, K. Multi-Objective Optimization using Evolutionary Algorithms. John Wiley & Sons, 2001. [43] DEB, K., AND JAIN, H. Bibliography An evolutionary many-objective optimiza- tion algorithm using reference-point-based nondominated sorting approach, part I: Solving problems with box constraints. IEEE Trans- actions on Evolutionary Computation 18, 4 (2014), 577–601. [44] DEB, K., PRATAP, A., AGARWAL, S., AND MEYARIVAN, T. A fast and elitist multiobjective genetic algorithm: NSGA-II. IEEE Transac- tions on Evolutionary Computation 6 (2002), 182–197. [45] DEB, K., THIELE, L., AND ZITZLER, E. Scalable multi-objectove optimization test problems. In IEEE Congress on Evolutionary Com- putation (2002), IEEE, pp. 825–830. [46] DEMIRKOL, E., MEHTA, S., AND UZSOY, R. Benchmarks for shop scheduling problems. European Journal of Operational Research 109, 1 (1998), 137–141. [47] DICK, G. Sensitivity-like analysis for feature selection in genetic programming. In Proceedings of the Genetic and Evolutionary Compu- tation Conference (2017), pp. 401–408. [48] DRAGOMIR, S. S., CERONE, P., AND SOFO, A. Some remarks on the midpoint rule in numerical integration. RGMIA research report collection 1, 2 (1998). [49] DUBOIS-LACOSTE, J., L ´OPEZ-IB ´A ˜NEZ, M., AND ST ¨UTZLE, T. A hy- brid TP+PLS algorithm for bi-objective ﬂow-shop scheduling prob- lems. Computers & OR 38, 8 (2011), 1219–1236. 217 BIBLIOGRAPHY [50] DUBOIS-LACOSTE, J., L ´OPEZ-IB ´A ˜NEZ, M., AND ST ¨UTZLE, T. Any- time pareto local search. European Journal of Operational Research 243, 2 (2015), 369–385. [50] DUBOIS-LACOSTE, J., L ´OPEZ-IB ´A ˜NEZ, M., AND ST ¨UTZLE, T. Any- time pareto local search. European Journal of Operational Research 243, 2 (2015), 369–385. [51] FOWLER, L., PFUND, M., YU, L., FOWLER, J. W., AND CARLYLE, W. M. Development Of A Robust Scheduling Rule For A Printed Wiring Board Drilling Operation With Multiple Scheduling Objec- tives And Fixed Order Release/Pickup Times. In IIE Annual Confer- ence. Proceedings (2002), Citeseer, p. 1. [52] GALV ´AN-L ´OPEZ, E., MCDERMOTT, J., O’NEILL, M., AND BRABAZON, A. Deﬁning locality as a problem difﬁculty measure in genetic programming. Genetic Programming and Evolvable Machines 12, 4 (2011), 365–401. [53] GAREY, M. R., JOHNSON, D. S., AND SETHI, R. The complexity of ﬂowshop and jobshop scheduling. Mathematics of Operations Research 1, 2 (1976), 117–129. [54] GENG, S., WU, D., WANG, P., AND CAI, X. Many-objective cloud task scheduling. IEEE Access 8 (2020), 79079–79088. [55] GIFFLER, B., AND THOMPSON, G. L. Algorithms for solving production-scheduling problems. Operations research 8, 4 (1960), 487–503. [56] GONG, X., DE PESSEMIER, T., MARTENS, L., AND JOSEPH, W. Bibliography Energy-and labor-aware ﬂexible job shop scheduling under dy- namic electricity pricing: A many-objective optimization investiga- tion. Journal of Cleaner Production 209 (2019), 1078–1094. [57] GONG, X., DE PESSEMIER, T., MARTENS, L., AND JOSEPH, W. Energy-and labor-aware ﬂexible job shop scheduling under dy- namic electricity pricing: A many-objective optimization investiga- tion. Journal of cleaner production 209 (2019), 1078–1094. 218 BIBLIOGRAPHY [58] GONZ ´ALEZ RODR´IGUEZ, I., RODR´IGUEZ VELA, M. D. C., PUENTE PEINADOR, J., AND HERN ´ANDEZ ARAUZO, A. Improved local search for job shop scheduling with uncertain durations. In Proceed- ings of the Nineteenth International Conference on Automated Planning and Scheduling (2009), Association for the Advancement of Artiﬁcial Intelligence (AAAI). [59] GRIFFIN, J. L., SCHLOSSER, S. W., GANGER, G. R., AND NAGLE, D. Operation Management. Cengage Learning, 2015. [60] GROMICHO, J. A., VAN HOORN, J. J., SALDANHA-DA GAMA, F., AND TIMMER, G. T. Solving the job-shop scheduling problem opti- mally by dynamic programming . Computers & Operations Research 39, 12 (2012), 2968–2977. [61] GUPTA, A. K., AND SIVAKUMAR, A. I. Job shop scheduling tech- niques in semiconductor manufacturing. The International Journal of Advanced Manufacturing Technology 27, 11-12 (2006), 1163–1169. [62] HART, E., ROSS, P., AND CORNE, D. Evolutionary Scheduling: A Review. Genetic Programming and Evolvable Machines 6, 2 (2005), 191– 220. [63] HEGER, J., HILDEBRANDT, T., AND SCHOLZ-REITER, B. Dispatch- ing rule selection with gaussian processes. Central European Journal of Operations Research 23, 1 (2015), 235–249. [64] HELD, M., AND KARP, R. M. A Dynamic Programming Approach to Sequencing Problems. In Proceedings of the 16th ACM National Meeting (ACM 1961) (1961), pp. 71.201–71.204. [65] HERRERO, J. G., BERLANGA, A., AND LOPEZ, J. M. M. Effective evolutionary algorithms for many-speciﬁcations attainment: Appli- cation to air trafﬁc control tracking ﬁlters. Evolutionary Computation, IEEE Transactions on 13, 1 (2009), 151–168. 219 BIBLIOGRAPHY [66] HILDEBRANDT, T., AND BRANKE, J. On using surrogates with ge- netic programming. Evolutionary computation 23, 3 (2015), 343–367. [66] HILDEBRANDT, T., AND BRANKE, J. On using surrogates with ge- netic programming. Evolutionary computation 23, 3 (2015), 343–367. [67] HILDEBRANDT, T., HEGER, J., AND SCHOLZ-REITER, B. Towards improved dispatching rules for complex shop ﬂoor scenarios: A ge- netic programming approach. In Proceedings of Genetic and Evolution- ary Computation Conference (2010), ACM, pp. 257–264. [68] HO, N., AND TAY, J. Evolving dispatching rules for solving the ﬂex- ible job-shop problem. In IEEE Congress on Evolutionary Computation (2005), vol. 3, IEEE, pp. 2848–2855. [69] HOLLAND, J. Adaption in Natural and Artiﬁcial Systems. MIT Press, Cambridge, MA, 1992. 1st edition: 1975, The University of Michigan Press, Ann Arbor. [70] HOLTHAUS, O., AND RAJENDRAN, C. Efﬁcient jobshop dispatching rules: Further developments. Production Planning & Control 11, 2 (2000), 171–178. [71] HORST, R., AND ROMEIJN, H. E. Handbook of Global Optimization, vol. 2. BIBLIOGRAPHY Springer Science & Business Media, 2002. [72] HUGHES, E. J. Multiple single objective Pareto sampling. In Evolu- tionary Computation, 2003. CEC’03. The 2003 Congress on (2003), vol. 4, IEEE, pp. 2678–2684. [73] HUNT, R., JOHNSTON, M., AND ZHANG, M. Evolving ”less- myopic” scheduling rules for dynamic job shop scheduling with genetic programming. In GECCO (2014), D. V. Arnold, Ed., ACM, pp. 927–934. [74] HUNT, R., JOHNSTON, M., AND ZHANG, M. Evolving Dispatch- ing Rules with Greater Understandability for Dynamic Job Shop Scheduling. Tech. rep., Victoria University, June 2015. 220 BIBLIOGRAPHY BIBLIOGRAPHY [75] HUNT, R., JOHNSTON, M., AND ZHANG, M. Using Local Search to Evaluate Dispatching Rules in Dynamic Job Shop Scheduling. In EvoCOP (2015), G. Ochoa and F. Chicano, Eds., vol. 9026 of Lecture Notes in Computer Science, Springer, pp. 222–233. [76] IBRAHIM, A., RAHNAMAYAN, S., MARTIN, M. V., AND DEB, K. EliteNSGA-III: An improved evolutionary many-objective op- timization algorithm. In CEC (2016), IEEE, pp. 973–982. [77] IKEDA, K.-I., KITA, H., AND KOBAYASHI, S. Failure of Pareto-based MOEAs: does non-dominated really mean near to optimal? In Evo- lutionary Computation, 2001. Proceedings of the 2001 Congress on (2001), vol. 2, IEEE, pp. 957–962. [78] ISHIBUCHI, H., IMADA, R., SETOGUCHI, Y., AND NOJIMA, Y. Refer- ence point speciﬁcation in hypervolume calculation for fair compar- ison and efﬁcient search. In Proceedings of the Genetic and Evolutionary Computation Conference (2017), pp. 585–592. [79] ISHIBUCHI, H., AND MURATA, T. A multi-objective genetic local search algorithm and its application to ﬂowshop scheduling. IEEE Transactions on Systems, Man, and Cybernetics, Part C: Applications and Reviews, 28, 3 (1998), 392–403. [80] ISHIBUCHI, H., SETOGUCHI, Y., MASUDA, H., AND NOJIMA, Y. Performance of decomposition-based many-objective algorithms strongly depends on Pareto front shapes. IEEE Transactions on Evo- lutionary Computation 21, 2 (2016), 169–190. [81] ISHIBUCHI, H., AND YOSHIDA, T. Hybrid Evolutionary Multi- Objective Optimization Algorithms. In HIS (2002), A. Abraham, J. R. del Solar, and M. K¨oppen, Eds., vol. 87 of Frontiers in Artiﬁcial Intel- ligence and Applications, IOS Press, pp. 163–172. 221 BIBLIOGRAPHY [82] ISHIBUCHI, H., YOSHIDA, T., AND MURATA, T. Balance between genetic search and local search in memetic algorithms for multiob- jective permutation ﬂowshop scheduling. IEEE Trans. Evolutionary Computation 7, 2 (2003), 204–223. [83] JACKSON, J. An extension of Johnson’s result on job-lot scheduling. Naval Research Logistics Quarterly 3, 3 (1956), 201–204. [84] JAIMES, A. L., AND COELLO, C. A. C. Many-Objective Problems: Challenges and Methods. In Handbook of Computational Intelligence, J. Kacprzyk and W. Pedrycz, Eds. Springer, 2015, pp. 1033–1046. [85] JAIN, H., AND DEB, K. An Improved Adaptive Approach for Eli- tist Nondominated Sorting Genetic Algorithm for Many-Objective Optimization. In EMO (2013), R. C. Purshouse, P. J. Fleming, C. M. Fonseca, S. Greco, and J. Shaw, Eds., vol. 7811 of Lecture Notes in Computer Science, Springer, pp. 307–321. [86] JAIN, H., AND DEB, K. BIBLIOGRAPHY An Evolutionary Many-Objective Optimiza- tion Algorithm Using Reference-Point Based Nondominated Sort- ing Approach, Part II: Handling Constraints and Extending to an Adaptive Approach. IEEE Transactions. Evolutionary Computation 18, 4 (2014), 602–622. [87] JAKOBOVI ´C, D., AND MARASOVI ´C, K. Evolving priority scheduling heuristics with genetic programming. Applied Soft Computing 12, 9 (2012), 2781–2789. [88] JASZKIEWICZ, A. On the performance of multiple-objective genetic local search on the 0/1 knapsack problem - a comparative experi- ment. IEEE Transactions on Evolutionary Computation 6, 4 (Aug. 2002), 402–412. [89] JASZKIEWICZ, A. Many-Objective Pareto Local Search. European Journal of Operational Research 271, 3 (2018), 1001–1013. 222 BIBLIOGRAPHY [90] JAYAMOHAN, M., AND RAJENDRAN, C. New dispatching rules for shop scheduling: a step forward. International Journal of Production Research 38, 3 (2000), 563–586. [91] JAYAMOHAN, M., AND RAJENDRAN, C. Development and analysis of cost-based dispatching rules for job shop scheduling. European Journal of Operational Research 157, 2 (2004), 307–321. [92] JIANG, S., AND YANG, S. A strength Pareto evolutionary algorithm based on reference direction for multiobjective and many-objective optimization. IEEE Transactions on Evolutionary Computation 21, 3 (2017), 329–346. [93] JIN, Y. Surrogate-assisted evolutionary computation: Recent ad- vances and future challenges. Swarm and Evolutionary Computation 1, 2 (2011), 61–70. [94] JONES, A., RABELO, L. C., AND SHARAWI, A. T. Survey of job shop scheduling techniques. Wiley encyclopedia of electrical and electronics engineering (2001). [95] JONG, K. D. Evolutionary Computation: A Uniﬁed Approach. The MIT Press, 2006. [96] KACEM, I., HAMMADI, S., AND BORNE, P. Pareto-optimality ap- proach for ﬂexible job-shop scheduling problems: hybridization of evolutionary algorithms and fuzzy logic. Mathematics and computers in simulation 60, 3 (2002), 245–276. [97] KARUNAKARAN, D., CHEN, G., AND ZHANG, M. Parallel multi- objective job shop scheduling using genetic programming. In Aus- tralasian Conference on Artiﬁcial Life and Computational Intelligence (2016), Springer, pp. 234–245. 223 BIBLIOGRAPHY BIBLIOGRAPHY [98] KARUNAKARAN, D., MEI, Y., CHEN, G., AND ZHANG, M. Sam- pling heuristics for multi-objective dynamic job shop scheduling using island based parallel genetic programming. In International Conference on Parallel Problem Solving from Nature (2018), Springer, pp. 347–359. [99] KASHAN, A. H., KESHMIRY, M., DAHOOIE, J. H., AND ABBASI- POOYA, A. A simple yet effective grouping evolutionary strategy (ges) algorithm for scheduling parallel machines. Neural Computing and Applications 30, 6 (2018), 1925–1938. [100] KENNEDY, J., AND EBERHART, R. Particle swarm optimization. In Proceedings of IEEE international conference on neural networks (1995), vol. 4, Perth, Australia, pp. 1942–1948. [101] KENNEDY, J., AND EBERHART, R. C. Swarm Intelligence. Morgan Kaufmann, April 2001. [102] KNOWLES, J., AND CORNE, D. M-PAES: a memetic algorithm for multiobjective optimization. In Proceedings of the 2000 Congress on Evolutionary Computation. CEC00 (Cat. No.00TH8512) (July 2000), vol. 1, IEEE, pp. 325–332 vol.1. [103] KOLAHAN, F., AND KAYVANFAR, V. A heuristic algorithm approach for scheduling of multi-criteria unrelated parallel machines. World, Academy of Science, Engineering and Technology 59 (2009), 102. [104] KOZA, J. R. A Genetic Approach to the Truck Backer Upper Prob- lem and the Inter-Twined Spiral Problem. In Proceedings of IJCNN International Joint Conference on Neural Networks (1992), vol. IV, IEEE Press, pp. 310–318. [105] KOZA, J. R. Genetic Programming – On the Programming of Computers by Means of Natural Selection. MIT Press, 1992. 224 BIBLIOGRAPHY [106] KREIPL, S. A large step random walk for minimizing total weighted tardiness in a job shop. Journal of Scheduling 3, 3 (2000), 125–138. [107] LAND, A. H., AND DOIG, A. G. An automatic method for solving discrete programming problems. Econometrica (1960), 497–520. [108] LARA, A., SANCHEZ, G., COELLO, C. A. C., AND SCH ¨UTZE, O. HCS: A New Local Search Strategy for Memetic Multiobjective Evo- lutionary Algorithms. IEEE Trans. Evolutionary Computation 14, 1 (2010), 112–132. [109] LARRA ˜NAGA, P., AND LOZANO, J. A. Estimation of distribution algo- rithms: A new tool for evolutionary computation, vol. 2. Springer Sci- ence & Business Media, 2001. [110] LAUMANNS, M., AND OCENASEK, J. Bayesian optimization algo- rithms for multi-objective optimization. In International Conference on Parallel Problem Solving from Nature (2002), Springer, pp. 298–307. [111] LAWLER, E. L., LENSTRA, J. K., RINNOOY KAN, A. H. G., AND SHMOYS, D. B. Sequencing and Scheduling: Algorithms and Com- plexity . In Logistics of Production and Inventory, vol. 4 of Handbooks in Operations Research and Management Science . BIBLIOGRAPHY Elsevier, 1993, pp. 445– 522. [112] LAWLER, E. L., AND MOORE, J. M. A Functional Equation and Its Application to Resource Allocation and Sequencing Problems. Man- agement Science 16, 1 (1969), 77–84. [113] LAWRENCE, S. Resouce constrained project scheduling: An exper- imental investigation of heuristic scheduling techniques (Supple- ment). Graduate School of Industrial Administration, Carnegie-Mellon University (1984). 225 BIBLIOGRAPHY BIBLIOGRAPHY [114] LEE, Y. H., BHASKARAN, K., AND PINEDO, M. A heuristic to mini- mize the total weighted tardiness with sequence-dependent setups. IIE transactions 29, 1 (1997), 45–52. [115] LEUNG, J. Y.-T., Ed. Handbook of Scheduling - Algorithms, Models, and Performance Analysis. Chapman and Hall/CRC, 2004. [116] LI, B., LI, J., TANG, K., AND YAO, X. Many-objective evolutionary algorithms: A survey. ACM Computing Surveys 48, 1 (2015), 13. [117] LI, K., KWONG, S., CAO, J., LI, M., ZHENG, J., AND SHEN, R. Achieving balance between proximity and diversity in multi- objective evolutionary algorithm. Inf. Sci. 182, 1 (2012), 220–242. [118] LI, M. Pareto or non-Pareto: Bi-criterion evolution in multiobjective optimization. IEEE Transactions on Evolutionary Computation 20, 5 (2015), 645–665. [119] LI, M., YANG, S., AND LIU, X. Shift-Based Density Estimation for Pareto-Based Algorithms in Many-Objective Optimization. IEEE Trans. Evolutionary Computation 18, 3 (2014), 348–365. [120] LI, X., MABU, S., AND HIRASAWA, K. A novel graph-based estima- tion of the distribution algorithm and its extension using reinforce- ment learning. Evolutionary Computation, IEEE Transactions on 18, 1 (2014), 98–113. [121] LIEFOOGHE, A., HUMEAU, J., MESMOUDI, S., JOURDAN, L., AND TALBI, E.-G. On dominance-based multiobjective local search: de- sign, implementation and experimental analysis on scheduling and traveling salesman problems. Journal of Heuristics 18, 2 (2012), 317– 352. [122] LIN, S.-W., AND YING, K.-C. Minimizing makespan and total ﬂowtime in permutation ﬂowshops by a bi-objective multi-start 226 BIBLIOGRAPHY simulated-annealing algorithm. Computers & OR 40, 6 (2013), 1625– 1647. [123] LIU, Y., GONG, D., SUN, X., AND ZHANG, Y. A reference points- based evolutionary algorithm for many-objective optimization. In Proceedings of the Companion Publication of the 2014 Annual Conference on Genetic and Evolutionary Computation (2014), ACM, pp. 1053–1056. [124] LOURENC¸ O, H. R., MARTIN, O. C., AND ST ¨UTZLE, T. Iterated local search: Framework and applications. In Handbook of metaheuristics. Springer, 2019, pp. 129–168. [125] LUKE, S., ET AL. A Java-based Evolutionary Computation Research System . https://cs.gmu.edu/˜eclab/projects/ecj/. [126] LUO, Q., DENG, Q., GONG, G., ZHANG, L., HAN, W., AND LI, K. An efﬁcient memetic algorithm for distributed ﬂexible job shop scheduling problem with transfers. Expert Systems with Applications 160 (2020), 113721. [127] LUST, T., AND TEGHEM, J. The multiobjective multidimensional knapsack problem: a survey and a new approach, July 2010. [128] MASOOD, A., MEI, Y., CHEN, G., AND ZHANG, M. BIBLIOGRAPHY A PSO-Based Reference Point Adaption Method for Genetic Programming Hyper- Heuristic in Many-Objective Job Shop Scheduling. In ACALCI (2017), M. Wagner, X. Li, and T. Hendtlass, Eds., vol. 10142 of Lec- ture Notes in Computer Science, pp. 326–338. [129] MATTFELD, D. C., AND BIERWIRTH, C. An efﬁcient genetic algo- rithm for job shop scheduling with tardiness objectives. European Journal of Operational Research 155, 3 (2004), 616–630. [130] MEI, Y., NGUYEN, S., XUE, B., AND ZHANG, M. An efﬁcient fea- ture selection algorithm for evolving job shop scheduling rules with 227 BIBLIOGRAPHY genetic programming. IEEE Transactions on Emerging Topics in Com- putational Intelligence 1, 5 (2017), 339–353. [131] MEI, Y., ZHANG, M., AND NYUGEN, S. Feature Selection in Evolving Job Shop Dispatching Rules with Genetic Programming, in GECCO, ACM, 2016. [132] MENCHACA-MENDEZ, A., MONTERO, E., AND MART´INEZ, S. Z. An Improved S-Metric Selection Evolutionary Multi-Objective Al- gorithm With Adaptive Resource Allocation. IEEE Access 6 (2018), 63382–63401. [133] MICHALEWICZ, Z., AND FOGEL, D. B. How to Solve It: Modern Heuristics, 2 ed. Springer Science & Business Media, 2013. [134] MITCHELL, T. M. Machine learning. McGraw-Hill, New York, NY [u.a., 2010. [135] MIYASHITA, K. Job-Shop Scheduling with GP. In Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2000) (Las Vegas, Nevada, USA, July 2000), D. Whitley, D. Goldberg, E. Cantu- Paz, L. Spector, I. Parmee, and H.-G. Beyer, Eds., Morgan Kaufmann, pp. 505–512. [136] MOSLEHI, G., AND MAHNAM, M. A Pareto approach to multi- objective ﬂexible job-shop scheduling problem using particle swarm optimization and local search. International Journal of Production Eco- nomics 129, 1 (2011), 14–22. [137] MURATA, T., ISHIBUCHI, H., AND TANAKA, H. Multi-objective ge- netic algorithm and its applications to ﬂowshop scheduling. Com- puters & Industrial Engineering 30, 4 (1996), 957–968. [138] MUSSELMAN, K., AND TALAVAGE, J. A tradeoff cut approach to multiple objective optimization. Operations Research 28, 6 (1980), 1424–1435. 228 BIBLIOGRAPHY [139] MUTH, J. F., AND THOMPSON, G. L. Industrial Scheduling. Prentice- Hall, 1963. [140] NARUKAWA, K., AND RODEMANN, T. Examining the performance of evolutionary many-objective optimization algorithms on a real- world application. In 2012 Sixth International Conference on Genetic and Evolutionary Computing (2012), IEEE, pp. 316–319. [141] NGUYEN, S. Automatic Design of Dispatching Rules for Job Shop Scheduling with Genetic Programming. PhD thesis, 2013. [142] NGUYEN, S., MEI, Y., MA, H., CHEN, A., AND ZHANG, M. Evo- lutionary scheduling and combinatorial optimisation: Applications, challenges, and future directions. BIBLIOGRAPHY In 2016 IEEE Congress on Evolu- tionary Computation (CEC) (July 2016), IEEE, pp. 3053–3060. [143] NGUYEN, S., ZHANG, M., AND JOHNSTON, M. A genetic program- ming based hyper-heuristic approach for combinatorial optimisa- tion. In GECCO (2011), N. Krasnogor and P. L. Lanzi, Eds., ACM, pp. 1299–1306. [144] NGUYEN, S., ZHANG, M., AND JOHNSTON, M. A sequential ge- netic programming method to learn forward construction heuristics for order acceptance and scheduling. In 2014 IEEE congress on evolu- tionary computation (CEC) (2014), IEEE, pp. 1824–1831. [145] NGUYEN, S., ZHANG, M., JOHNSTON, M., AND TAN, K. C. A computational study of representations in genetic programming to evolve dispatching rules for the job shop scheduling problem. Evo- lutionary Computation, IEEE Transactions 17, 5 (2013), 621–639. [146] NGUYEN, S., ZHANG, M., JOHNSTON, M., AND TAN, K. C. Auto- matic design of scheduling policies for dynamic multi-objective job shop scheduling via cooperative coevolution genetic programming. IEEE Transactions on Evolutionary Computation 18, 2 (2013), 193–208. 229 BIBLIOGRAPHY [147] NGUYEN, S., ZHANG, M., JOHNSTON, M., AND TAN, K. C. Dy- namic multi-objective job shop scheduling: A genetic programming approach. In Automated Scheduling and Planning. Springer, 2013, pp. 251–282. [147] NGUYEN, S., ZHANG, M., JOHNSTON, M., AND TAN, K. C. Dy- namic multi-objective job shop scheduling: A genetic programming approach. In Automated Scheduling and Planning. Springer, 2013, pp. 251–282. [148] NGUYEN, S., ZHANG, M., JOHNSTON, M., AND TAN, K. C. Learn- ing iterative dispatching rules for job shop scheduling with genetic programming. The International Journal of Advanced Manufacturing Technology 67, 1-4 (2013), 85–100. [149] NGUYEN, S., ZHANG, M., JOHNSTON, M., AND TAN, K. C. Auto- matic programming via iterated local search for dynamic job shop scheduling. IEEE Transactions on Cybernetics 45, 1 (2015), 1–14. [150] OK, S., MIYASHITA, K., AND NISHIHARA, S. Improving perfor- mance of GP by adaptive terminal selection. In PRICAI 2000 Topics in Artiﬁcial Intelligence. Springer, 2000, pp. 435–445. [151] OKABE, T., JIN, Y., SENDOFF, B., AND OLHOFER, M. Voronoi-based estimation of distribution algorithm for multi-objective optimiza- tion. In Proceedings of the 2004 Congress on Evolutionary Computation (IEEE Cat. No. 04TH8753) (2004), vol. 2, IEEE, pp. 1594–1601. [152] OSMAN, I. H., AND KELLY, J. P. Meta-heuristics: An Overview. Springer, 1996. [153] OWEN, A., AND HARVEY, I. Adapting Particle Swarm Optimisa- tion for Fitness Landscapes with Neutralit. IEEE Swarm Intelligence Symposium, IEEE. [154] PANWALKAR, S. S., AND ISKANDER, W. A survey of scheduling rules. BIBLIOGRAPHY Operations research 25, 1 (1977), 45–61. [155] PAQUETE, L., CHIARANDINI, M., AND ST ¨UTZLE, T. Pareto local optimum sets in the biobjective traveling salesman problem: An 230 BIBLIOGRAPHY experimental study. In Metaheuristics for multiobjective optimisation. Springer, 2004, pp. 177–199. [156] PAQUETE, L., AND ST ¨UTZLE, T. A study of stochastic local search algorithms for the biobjective QAP with correlated ﬂow matrices. European Journal of Operational Research 169, 3 (2006), 943–959. [157] PARK, J., NGUYEN, S., ZHANG, M., AND JOHNSTON, M. Genetic programming for order acceptance and scheduling. In 2013 IEEE Congress on Evolutionary Computation (2013), IEEE, pp. 1005–1012. [158] PARK, J., NGUYEN, S., ZHANG, M., AND JOHNSTON, M. A Sin- gle Population Genetic Programming based Ensemble Learning Ap- proach to Job Shop Scheduling. In Proceedings of the Companion Pub- lication of the 2015 on Genetic and Evolutionary Computation Conference (2015), ACM, pp. 1451–1452. [159] PARK, J., NGUYEN, S., ZHANG, M., AND JOHNSTON, M. Evolving Ensembles of Dispatching Rules Using Genetic Programming for Job Shop Scheduling. In Genetic Programming. Springer, 2015, pp. 92– 104. [160] PELIKAN, M., GOLDBERG, D. E., AND CANTU-PAZ, E. Linkage problem, distribution estimation, and Bayesian networks. Evolution- ary computation 8, 3 (2000), 311–340. [161] PELIKAN, M., GOLDBERG, D. E., AND LOBO, F. G. A survey of opti- mization by building and using probabilistic models. Computational optimization and applications 21, 1 (2002), 5–20. [162] PENG, B., L ¨U, Z., AND CHENG, T. A tabu search/path relinking algorithm to solve the job shop scheduling problem. Computers & Operations Research 53 (2015), 154–164. 231 BIBLIOGRAPHY [163] PFUND, M., FOWLER, J. W., AND GUPTA, J. N. A survey of al- gorithms for single and multi-objective unrelated parallel-machine deterministic scheduling problems. Journal of the Chinese Institute of Industrial Engineers 21, 3 (2004), 230–241. [163] PFUND, M., FOWLER, J. W., AND GUPTA, J. N. A survey of al- gorithms for single and multi-objective unrelated parallel-machine deterministic scheduling problems. Journal of the Chinese Institute of Industrial Engineers 21, 3 (2004), 230–241. [164] PICKARDT, C. W., HILDEBRANDT, T., BRANKE, J., HEGER, J., AND SCHOLZ-REITER, B. Evolutionary generation of dispatching rule sets for complex dynamic scheduling problems. International Journal of Production Economics 145, 1 (2013), 67–77. [165] PINEDO, M. L. Scheduling: theory, algorithms, and systems. Springer Science & Business Media, 2012. [166] POTTS, C. N., AND STRUSEVICH, V. A. Fifty years of scheduling: a survey of milestones. JORS 60, S1 (2009). BIBLIOGRAPHY [167] PRADITWONG, K., AND YAO, X. A new multi-objective evolution- ary optimisation algorithm: the two-archive algorithm. In Computa- tional intelligence and security, 2006 international conference on (2006), vol. 1, IEEE, pp. 286–291. [168] RAGHU, T., AND RAJENDRAN, C. An efﬁcient dynamic dispatching rule for scheduling in a job shop . International Journal of Production Economics 32, 3 (1993), 301–313. [169] RAJENDRAN, C., AND HOLTHAUS, O. A comparative study of dis- patching rules in dynamic ﬂowshops and jobshops. European journal of operational research 116, 1 (1999), 156–170. [170] RASMUSSEN, C. E., AND WILLIAMS, C. K. I. Gaussian Processes for Machine Learning. MIT Press, 2006. [171] ROTHLAUF, F., AND OETZEL, M. On the locality of grammatical evolution. In European Conference on Genetic Programming (2006), Springer, pp. 320–330. 232 BIBLIOGRAPHY BIBLIOGRAPHY [172] RUSSELL, S. J., NORVIG, P., AND RUSSELL, S. J. Artiﬁcial Intelli- gence: A Modern Approach (Prentice Hall Series in Artiﬁcial Intelligence), 0002 ed. Prentice Hall, 2003. [173] SABAR, N. R., TURKY, A., AND SONG, A. A genetic program- ming based iterated local search for software project scheduling. In GECCO (2018), H. E. Aguirre and K. Takadama, Eds., ACM, pp. 1364–1370. [174] SATO, H., AGUIRRE, H. E., AND TANAKA, K. Genetic Diversity and Effective Crossover in Evolutionary Many-objective Optimization. In LION (2011), C. A. C. Coello, Ed., vol. 6683 of Lecture Notes in Computer Science, Springer, pp. 91–105. [175] SEADA, H. A., ABOUHAWWASH, M., AND DEB, K. Towards a Better Diversity of Evolutionary Multi-Criterion Optimization Algorithms Using Local Searches. In Proceedings of the 2016 on Genetic and Evo- lutionary Computation Conference Companion (New York, NY, USA, 2016), GECCO ’16 Companion, ACM, pp. 77–78. [176] SELS, V., GHEYSEN, N., AND VANHOUCKE, M. A comparison of priority rules for the job shop scheduling problem under different ﬂow time-and tardiness-related objective functions. International Journal of Production Research 50, 15 (2012), 4255–4270. [177] SHA, D., AND HSU, C.-Y. A hybrid particle swarm optimization for job shop scheduling problem . Computers & Industrial Engineering 51, 4 (2006), 791–808. [178] SHAO, L., LIU, L., AND LI, X. Feature learning for image classiﬁ- cation via multiobjective genetic programming. Neural Networks and Learning Systems, IEEE Transactions on 25, 7 (2014), 1359–1371. [179] SPRECHER, A., KOLISCH, R., AND DREXL, A. Semi-active, ac- tive, and non-delay schedules for the resource-constrained project 233 BIBLIOGRAPHY scheduling problem . European Journal of Operational Research 80, 1 (1995), 94–102. [180] S ¨ULFLOW, A., DRECHSLER, N., AND DRECHSLER, R. Robust multi- objective optimization in high dimensional spaces. In International conference on evolutionary multi-criterion optimization (2007), Springer, pp. 715–726. [181] SURESH, V., AND CHAUDHURI, D. Dynamic scheduling – a survey of research. International Journal of Production Economics 32, 1 (1993), 53–63. [182] TAILLARD, E. Benchmarks for basic scheduling problems. european journal of operational research 64, 2 (1993), 278–285. [183] TAILLARD, E. D. Parallel taboo search techniques for the job shop scheduling problem. ORSA journal on Computing 6, 2 (1994), 108–117. [184] TAVAKKOLI-MOGHADDAM, R., RAHIMI-VAHED, A., AND MIRZAEI, A. H. A hybrid multi-objective immune algorithm for a ﬂow shop scheduling problem with bi-objectives: Weighted mean completion time and weighted mean tardiness. Information Sciences 177, 22 (2007), 5072–5090. [185] TAY, J. C., AND HO, N. B. BIBLIOGRAPHY Evolving dispatching rules using genetic programming for solving multi-objective ﬂexible job-shop problems. Computers & Industrial Engineering 54, 3 (2008), 453–473. [186] THIERENS, D., AND BOSMAN, P. A. Multi-objective mixture-based iterated density estimation evolutionary algorithms. In Proceedings of the 3rd Annual Conference on Genetic and Evolutionary Computation (2001), Morgan Kaufmann Publishers Inc., pp. 663–670. [187] TIAN, Y., CHENG, R., ZHANG, X., AND JIN, Y. PlatEMO: A MAT- LAB Platform for Evolutionary Multi-Objective Optimization [Edu- cational Forum], Nov. 2017. 234 BIBLIOGRAPHY BIBLIOGRAPHY [188] TSAI, C.-W., AND RODRIGUES, J. J. P. C. Metaheuristic Scheduling for Cloud: A Survey. IEEE Systems Journal 8, 1 (2014), 279–291. [189] VAESSENS, R. J. M., AARTS, E. H., AND LENSTRA, J. K. Job shop scheduling by local search. INFORMS Journal on Computing 8, 3 (1996), 302–317. [190] VAN LAARHOVEN, P. J., AARTS, E. H., AND LENSTRA, J. K. Job shop scheduling by simulated annealing. Operations research 40, 1 (1992), 113–125. [191] VEPSALAINEN, A. P. J., AND MORTON, T. E. Priority Rules for Job Shops with Weighted Tardiness Costs. Management Science 33, 8 (1987), 1035–1047. [192] WANG, C., JI, Z., AND WANG, Y. Many-objective ﬂexible job shop scheduling using nsga-iii combined with multi-attribute decision making. Modern Physics Letters B 32, 34n36 (2018), 1840110. [193] WANG, H., JIAO, L., AND YAO, X. Two Arch2: An Improved Two- Archive Algorithm for Many-Objective Optimization. IEEE Transac- tions on Evolutionary Computation 19, 4 (Aug. 2015), 524–541. [194] WIDMER, G., AND KUBAT, M. Learning in the Presence of Concept Drift and Hidden Contexts. Machine Learning 23, 1 (1996), 69–101. [195] WILCOXON, F. Individual comparisons by ranking methods. Springer, 1992. [196] WONG, T. C., AND NGAN, S. C. A comparison of hybrid ge- netic algorithm and hybrid particle swarm optimization to minimize makespan for assembly job shop. Applied Soft Computing 13, 3 (2013), 1391–1399. 235 BIBLIOGRAPHY BIBLIOGRAPHY [197] XIA, W., AND WU, Z. An effective hybrid optimization approach for multi-objective ﬂexible job-shop scheduling problems. Computers & Industrial Engineering 48, 2 (2005), 409–425. [198] XU, W.-J., HE, L.-J., AND ZHU, G.-Y. Many-objective ﬂow shop scheduling optimisation with genetic algorithm based on fuzzy sets. International Journal of Production Research (2019), 1–25. [199] YANG, S., LI, M., LIU, X., AND ZHENG, J. A Grid-Based Evolu- tionary Algorithm for Many-Objective Optimization. IEEE Trans. Evolutionary Computation 17, 5 (2013), 721–736. [200] YAQIN, Z., BEIZHI, L., AND LV, W. Study on job-shop scheduling with multi-objectives based on genetic algorithms. In Proceedings of 2010 International Conference on Computer Application and System Modeling (2010), vol. 10, IEEE, pp. V10–294. [201] YENISEY, M. M., AND YAGMAHAN, B. Multi-objective permutation ﬂow shop scheduling problem: Literature review, classiﬁcation and current trends. Omega 45 (2014), 119–135. [202] YUAN, Y., AND XU, H. A memetic algorithm for the multi-objective ﬂexible job shop scheduling problem. In Proceedings of the 15th an- nual conference on Genetic and evolutionary computation (2013), ACM, pp. 559–566. [203] ZAHIRI, A., AND AZAMATHULLA, H. M. Comparison between lin- ear genetic programming and M5 tree models to predict ﬂow dis- charge in compound channels. Neural Computing and Applications 24, 2 (2014), 413–420. [204] ZHANG, G., SHAO, X., LI, P., AND GAO, L. An effective hybrid particle swarm optimization algorithm for multi-objective ﬂexible job-shop scheduling problem. Computers & Industrial Engineering 56, 4 (2009), 1309–1318. 236 BIBLIOGRAPHY BIBLIOGRAPHY [205] ZHANG, Q., AND LI, H. MOEA/D: A multiobjective evolutionary algorithm based on decomposition. Evolutionary Computation, IEEE Transactions on 11, 6 (2007), 712–731. [206] ZHANG, Q., ZHOU, A., ZHAO, S., SUGANTHAN, P. N., LIU, W., AND TIWARI, S. Multiobjective optimization test instances for the CEC 2009 special session and competition. University of Essex, Colchester, UK and Nanyang technological University, Singapore, special session on performance assessment of multi-objective optimization algo- rithms, technical report (2008), 1–30. [207] ZHANG, R., CHANG, P.-C., SONG, S., AND WU, C. Local search enhanced multi-objective pso algorithm for scheduling textile pro- duction processes with environmental considerations. Applied Soft Computing 61 (2017), 447–467. [208] ZHANG, X., TIAN, Y., AND JIN, Y. A Knee Point-Driven Evolu- tionary Algorithm for Many-Objective Optimization. IEEE Trans. Evolutionary Computation 19, 6 (2015), 761–776. [209] ZHAO, F., CHEN, Z., WANG, J., AND ZHANG, C. An improved moea/d for multi-objective job shop scheduling problem. Interna- tional Journal of Computer Integrated Manufacturing 30, 6 (2017), 616– 640. [210] ZHU, G.-Y., DING, C., AND ZHANG, W.-B. Optimal foraging al- gorithm that incorporates fuzzy relative entropy for solving many- objective permutation ﬂow shop scheduling problems. IEEE Trans- actions on Fuzzy Systems (2020). [211] ZITZLER, E., LAUMANNS, M., AND THIELE, L. SPEA2: Improv- ing the strength pareto evolutionary algorithm. In EUROGEN 2001. Evolutionary Methods for Design, Optimization and Control with Appli- cations to Industrial Problems (2002), pp. 95–100. 237 BIBLIOGRAPHY [212] ZITZLER, E., THIELE, L., LAUMANNS, M., FONSECA, C. M., AND DA FONSECA, V. G. Performance assessment of multiobjective op- timizers: an analysis and review. Evolutionary Computation, IEEE Transactions on 7, 2 (2003), 117–132.
https://openalex.org/W2980397916	https://www.jstatsoft.org/index.php/jss/article/download/v104i03/4383	English	null	AMR - An R Package for Working with Antimicrobial Resistance Data	bioRxiv (Cold Spring Harbor Laboratory)	2,019	cc-by	13,381	JSS JSS Journal of Statistical Software September 2022, Volume 104, Issue 3. doi: 10.18637/jss.v104.i03 Abstract Antimicrobial resistance is an increasing threat to global health. Evidence for this trend is generated in microbiological laboratories through testing microorganisms for re- sistance against antimicrobial agents. International standards and guidelines are in place for this process as well as for reporting data on (inter-)national levels. However, there is a gap in the availability of standardized and reproducible tools for working with laboratory data to produce the required reports. It is known that extensive efforts in data cleaning and validation are required when working with data from laboratory information systems. Furthermore, the global spread and relevance of antimicrobial resistance demands to in- corporate international reference data in the analysis process. p y p In this paper, we introduce the AMR package for R that aims at closing this gap by providing tools to simplify antimicrobial resistance data cleaning and analysis, while incorporating international guidelines and scientifically reliable reference data. The AMR package enables standardized and reproducible antimicrobial resistance analyses, includ- ing the application of evidence-based rules, determination of first isolates, translation of various codes for microorganisms and antimicrobial agents, determination of (multi-drug) resistant microorganisms, and calculation of antimicrobial resistance, prevalence and fu- ture trends. The AMR package works independently of any laboratory information system and provides several functions to integrate into international workflows (e.g., WHONET software provided by the World Health Organization). Keywords: antimicrobial resistance, data analysis, R, software, epidemiology. AMR: An R Package for Working with Antimicrobial Resistance Data Matthijs S. Berends University of Groningen Christian F. Luz University of Groningen Alexander W. Friedrich University of Groningen Bhanu N. M. Sinha University of Groningen Casper J. Albers University of Groningen Corinna Glasner University of Groningen AMR: Working with Antimicrobial Resistance Data in R AMR: Working with Antimicrobial Resistance Data in R 2 2 to both patients and health care systems. Current estimates show the immense dimensions we are already facing, such as claiming at least 50,000 lives due to antimicrobial resistance each year across Europe and the United States alone (O’Neill 2014). Although estimates for the burden through antimicrobial resistance and their predictions are disputed (De Kraker, Stewardson, and Harbarth 2016) the rising trend is undeniable (CDC 2019), thus calling for worldwide efforts on tackling this problem. Surveillance programs and reliable data are key for controlling and streamlining these efforts. Surveillance data of antimicrobial resistance at higher levels (national or international) usually comprise aggregated numbers. The basis of this information is generated and stored at local microbiological laboratories where isolated microorganisms are tested for their susceptibility to a whole range of antimicrobial agents. The efficacy of these agents against microorganisms is nowadays interpreted as follows (EUCAST 2019): • R: Resistant. There is a high likelihood of therapeutic failure. • R: Resistant. There is a high likelihood of therapeutic failure. • S: Susceptible, standard dosing regimen. There is a high likelihood of therapeutic success using a standard dosing regimen of an antimicrobial agent. • S: Susceptible, standard dosing regimen. There is a high likelihood of therapeutic success using a standard dosing regimen of an antimicrobial agent. • I: Susceptible, increased exposure. There is a high likelihood of therapeutic success, but only when exposure to an antimicrobial agent is increased by adjusting the dosing regimen or its concentration at the site of infection. • I: Susceptible, increased exposure. There is a high likelihood of therapeutic success, but only when exposure to an antimicrobial agent is increased by adjusting the dosing regimen or its concentration at the site of infection. Generally, antimicrobial resistance is defined as the proportion of resistant microorganisms (R) among all tested microorganisms of the same species (R + S + I). Today, the two ma- jor guideline institutes to define the international standards on antimicrobial resistance are the European Committee on Antimicrobial Susceptibility Testing (EUCAST, Leclercq et al. 2013) and the Clinical and Laboratory Standards Institute (CLSI, Clinical and Laboratory Standards Institute 2014). The guidelines from these two institutes are adopted by 94% of all countries reporting antimicrobial resistance to the WHO (World Health Organization 2018a). 1. Introduction Antimicrobial resistance is a global health problem and of great concern for human medicine, veterinary medicine, and the environment alike. It is associated with significant burdens AMR: Working with Antimicrobial Resistance Data in R Although these standardized guidelines are in place on the laboratory level for the data gen- eration process, stored data in laboratory information systems are often not yet suitable for data analysis. Laboratory information systems are often designed to fit billing purposes rather than epidemiological data analysis. Furthermore, (inter-)national surveillance is hin- dered by inadequate standardization of epidemiological definitions, different types of samples and data collection, settings included, microbiological testing methods (including suscepti- bility testing), and data sharing policies (Tacconelli et al. 2018). The necessity of accurate data analysis in the field of antimicrobial resistance has just recently been further underlined (Limmathurotsakul et al. 2019). Antimicrobial resistance analyses require a thorough under- standing of microbiological tests and their results, the biological taxonomy of microorganisms, the clinical and epidemiological relevance of the results, their pharmaceutical implications, and (inter-)national standards and guidelines for working with and reporting antimicrobial resistance. Here, we describe the AMR package (Berends, Luz, Friedrich, Sinha, Albers, and Glas- ner 2022) for R (R Core Team 2022), which has been developed to standardize clean and reproducible antimicrobial resistance data analyses using international standardized recom- mendations (Leclercq et al. 2013; Clinical and Laboratory Standards Institute 2014) while Journal of Statistical Software 3 patient date test_no specimen mo PEN AMC CIP 1 2019-03-08 100 blood esccol R I S 1 2019-03-09 101 blood esccol R I S 2 2019-03-08 102 blood staaur R S – 3 2019-03-08 103 urine pseaer R R R Table 1: Example of an antimicrobial resistance report. Abbreviations: R = resistant, S = sus- ceptible, I = susceptible, increased exposure, mo = microorganism, PEN = penicillin, AMC = amoxicillin/clavulanic acid, CIP = ciprofloxacin. Table 1: Example of an antimicrobial resistance report. Abbreviations: R = resistant, S = sus- ceptible, I = susceptible, increased exposure, mo = microorganism, PEN = penicillin, AMC = amoxicillin/clavulanic acid, CIP = ciprofloxacin. incorporating scientifically reliable reference data about valid laboratory outcome, antimi- crobial agents, and the complete biological taxonomy of microorganisms. The AMR pack- age provides solutions and support for these aspects while being independent of underly- ing laboratory information systems, thereby democratizing the analysis process. Devel- oped in R and available from the Comprehensive R Archive Network (CRAN) at https: //CRAN.R-project.org/package=AMR since February 22, 2018 (Berends et al. 2022), the AMR package enables reproducible workflows as described in other fields, such as environ- mental science (Lowndes et al. 2017). AMR: Working with Antimicrobial Resistance Data in R The AMR package provides a new technical instrument to aid in curbing the global threat of antimicrobial resistance. Furthermore, local and regional data in the laboratories can now become relevant in any setting for public health. While no other packages R package with the purpose of dealing with antimicrobial resis- tance data are available on CRAN or Bioconductor, the AMR package may be integrated in workflows of related packages. For example, the R Epidemics Consortium (RECON) pro- vides high-quality packages for data analysis in infectious disease outbreaks or epidemics (for example incidence and epicontacts, Jombart et al. 2020; Nagraj, Jombart, Randhawa, Sudre, Campbell, and Crellen 2021). In addition, on the laboratory side the antibioticR package provides approaches to work with disk diffusion zone diameter and minimum in- hibitory concentration data from environment samples (Petzoldt 2021). We aim at providing a comprehensive and standardized toolbox for antimicrobial resistance data processing and analysis, with a focus on microbiological, clinical, and epidemiological purposes that was yet missing. The following sections describe the functionality of the AMR package according to its core functionalities for transforming, enhancing, and analyzing antimicrobial resistance data using scientifically reliable reference data. 2. Antimicrobial resistance data Microbiological tests can be performed on different specimens, such as blood or urine samples or nasal swabs. After arrival at the microbiological laboratory, the specimens are traditionally cultured on specific media, such as blood agar. If a microorganism can be isolated from these media, it is tested against several antimicrobial agents. Based on the minimal inhibitory concentration (MIC) of the respective agent and interpretation guidelines, such as guidelines by EUCAST (Leclercq et al. 2013) and CLSI (Clinical and Laboratory Standards Institute 2014), test results are reported as “resistant” (R), “susceptible” (S) or “susceptible, increased exposure” (I). A typical data structure is illustrated in Table 1 (Leclercq et al. 2013). AMR: Working with Antimicrobial Resistance Data in R 4 4 patient date test_no specimen mo PEN AMC CIP 1 2019-03-08 100 blood esccol R I S 1 2019-03-09 101 blood esccol R I S 2 2019-03-08 102 blood StaAur > 8 (R)* < 0.01 (S)* . 3 2019-03-08 103 urine P. aeru. R S** S Table 2: Antimicrobial resistance report example – ambiguous formats. * = Mixed reporting of minimal inhibitory concentration (MIC) and susceptibility interpretation of MIC value. ** = False reporting; Pseudomonas aeruginosa (mo = P. aeru.) is intrinsically resistant to amox- icillin/clavulanic acid (AMC). Abbreviations: R = resistant, S = susceptible, I = susceptible, increased exposure, mo = microorganism, PEN = penicillin, AMC = amoxicillin/clavulanic acid, CIP = ciprofloxacin. Table 2: Antimicrobial resistance report example – ambiguous formats. * = Mixed reporting of minimal inhibitory concentration (MIC) and susceptibility interpretation of MIC value. ** = False reporting; Pseudomonas aeruginosa (mo = P. aeru.) is intrinsically resistant to amox- icillin/clavulanic acid (AMC). Abbreviations: R = resistant, S = susceptible, I = susceptible, increased exposure, mo = microorganism, PEN = penicillin, AMC = amoxicillin/clavulanic acid, CIP = ciprofloxacin. For the first two rows, the information should be read as: Escherichia coli (mo = esccol) was isolated from blood of patient 1 and was found to be resistant to penicillin, and susceptible to amoxicillin/clavulanic acid and ciprofloxacin. However, often (especially when merging sources) data is reported in ambiguous formats as exemplified in Table 2. It is crucial that source data can be analyzed in a reliable way, especially when the outcome will be used to evaluate patient treatment options. This requires reproducible and field-specific, specialized data cleaning and transforming. 2. Antimicrobial resistance data The AMR package aims at providing a standardized and automated way of cleaning, trans- forming, and enhancing these typical data structures (Table 1 and 2), independent of the underlying data source. Processed data would be similar to Table 3 that highlights several package functionalities in the sections below. 3.1. Working with taxonomically valid microorganism names Coercing is a computational process of forcing output based on an input. For microor- ganism names, coercing user input to taxonomically valid microorganism names is crucial to ensure correct interpretation and to enable grouping based on taxonomic properties. To this end, the AMR package includes all microbial entries from The Catalogue of Life (https://www.catalogueoflife.org/), the most comprehensive and authoritative global index of species currently available (Bánki et al. 2022). It holds essential information on the names, relationships, and distributions of more than 1.9 million species. The integration of it into the AMR package is described in the Appendix A. The as.mo() function makes use of this underlying data to transform a vector of characters to a new class ‘mo’ of taxonomically valid microorganism name. The resulting values are microbial IDs, which are human-readable for the trained eye and contain information about the taxonomic kingdom, genus, species, and subspecies (Figure 1). The as.mo() function compares the user input with taxonomically valid microorganism names, rates the matching with a score and returns results based on the highest score. This 5 5 Journal of Statistical Software patient date test_no specimen moa PENb AMCb CIPb first_isolatec named gram_staine 1 2019-03-08 100 blood B_ESCHR_COLI R I S TRUE Escherichia coli Gram-negative 1 2019-03-09 101 blood B_ESCHR_COLI R I S FALSE Escherichia coli Gram-negative 2 2019-03-08 102 blood B_STPHY_AURS R S NA TRUE Staphylococcus aureus Gram-positive 3 2019-03-08 103 urine B_PSDMN_AERG R R S TRUE Pseudomonas aeruginosa Gram-negative a) as.mo() function b) eucast_rules() function applied c) first_isolate() function d) mo_name() function e) mo_gramstain() function Table 3: Enhanced antimicrobial resistance report example. Abbreviations: R = resistant, S = susceptible, I = susceptible, increased exposure, mo = microorganism, PEN = penicillin, AMC = amoxicillin/clavulanic acid, CIP = ciprofloxacin. patient date test_no specimen moa PENb AMCb CIPb first_isolatec named gram_staine 1 2019-03-08 100 blood B_ESCHR_COLI R I S TRUE Escherichia coli Gram-negative 1 2019-03-09 101 blood B_ESCHR_COLI R I S FALSE Escherichia coli Gram-negative 2 2019-03-08 102 blood B_STPHY_AURS R S NA TRUE Staphylococcus aureus Gram-positive 3 2019-03-08 103 urine B_PSDMN_AERG R R S TRUE Pseudomonas aeruginosa Gram-negative a) as.mo() function b) eucast_rules() function applied c) first_isolate() function d) mo_name() function e) mo_gramstain() function Table 3: Enhanced antimicrobial resistance report example. Abbreviations: R = resistant, S = susceptible, I = susceptible, increased exposure, mo = microorganism, PEN = penicillin, AMC = amoxicillin/clavulanic acid, CIP = ciprofloxacin. 3.1. Working with taxonomically valid microorganism names Group 2 consists of all microorganisms where the taxonomic phylum is Proteobacteria, Firmicutes, Actinobacteria or Sarcomastigophora, or where the taxonomic genus is Absidia, Acremonium, Journal of Statistical Software 7 Actinotignum, Alternaria, Anaerosalibacter, Apophysomyces, Arachnia, Aspergillus, Aureobac- terium, Aureobasidium, Bacteroides, Basidiobolus, Beauveria, Blastocystis, Branhamella, Bor- relia, Calymmatobacterium, Candida, Capnocytophaga, Catabacter, Chaetomium, Chlamy- dia, Chlamydophila, Chryseobacterium, Chryseomonas, Chrysonilia, Cladophialophora, Cla- dosporium, Conidiobolus, Cryptococcus, Curvularia, Exophiala, Exserohilum, Flavobacterium, Fonsecaea, Fusarium, Fusobacterium, Hendersonula, Hypomyces, Koserella, Lelliottia, Lep- tosphaeria, Leptotrichia, Malassezia, Malbranchea, Mortierella, Mucor, Mycocentrospora, My- coplasma, Nectria, Ochroconis, Oidiodendron, Phoma, Piedraia, Pithomyces, Pityrosporum, Prevotella,Pseudallescheria, Rhizomucor, Rhizopus, Rhodotorula, Scolecobasidium, Scopular- iopsis, Scytalidium,Sporobolomyces, Stachybotrys, Stomatococcus, Treponema, Trichoderma, Trichophyton, Trichosporon, Tritirachium or Ureaplasma. Group 3 consists of all other mi- croorganisms. This will lead to the effect that e.g., "E. coli" will return the microbial ID of Escherichia coli (m = 0.688, a highly prevalent microorganism found in humans) and not Entamoeba coli (m = 0.079, a less prevalent microorganism in humans), although the latter would alphabetically come first. The matching score function is for users available as mo_matching_score(). If any coercion rules are applied, a warning is printed to the console and scores can be reviewed by calling mo_uncertainties(), that prints all other matches with their matching scores. Users can furthermore control the coercion rules by setting the allow_uncertain argument in the as.mo() function. The following values or levels can be used: • 0: no additional rules are applied; • 1: allow previously accepted (but now invalid) taxonomic names and minor spelling errors; • 2: allow all of 1, strip values between brackets, inverse the words of the input, strip off text elements from the end keeping at least two elements; • 3: allow all of level 1 and 2, strip off text elements from the end, allow any part of a taxonomic name; • TRUE (default): equivalent to 2; • FALSE: equivalent to 0. • FALSE: equivalent to 0. To support organization specific microbial IDs, users can specify a custom reference ‘data.frame’, by using as.mo(..., reference_df = ...). This process can also be automated by users with the set_mo_source() function. 3.1. Working with taxonomically valid microorganism names AMR: Working with Antimicrobial Resistance Data in R 6 subspecies, a 4-5 letter acronym species, a 4-5 letter acronym genus, a 5-7 letter acronym kingdom: A (Archaea), AN (Animalia), B (Bacteria), C (Chromista), F (Fungi), or P (Protozoa) Microbial ID Full name B_KLBSL Klebsiella B_KLBSL_PNMN Klebsiella pneumoniae B_KLBSL_PNMN_RHNS Klebsiella pneumoniae rhinoscleromatis Figure 1: The structure of a typical microbial ID as used in the AMR package. An ID consists of two to four elements, separated by an underscore. The first element is the abbre- viation of the taxonomic kingdom. The remaining elements consist of abbreviations of the lowest taxonomic levels of every microorganism: genus, species (if available) and subspecies (if available). Abbreviations used for the microbial IDs of microorganism names were created using the base R function abbreviate(). matching score (m), ranging from 0 to 1, is calculated using the following equation: m(x,n) = ln −0.5 · min{ln, lev(x, n)} ln · pn · kn where: • x is the user input; • x is the user input; • n is a taxonomic name (genus, species, and subspecies); • lev is the Levenshtein distance function (Levenshtein 1966), which counts any insertion, deletion and substitution as 1 that is needed to change x into n; • lev is the Levenshtein distance function (Levenshtein 1966), which counts any insertion, deletion and substitution as 1 that is needed to change x into n; • pn is the human pathogenic prevalence group of n, as described below; • kn is the taxonomic kingdom of n, set as Bacteria = 1, Fungi = 2, Protozoa = 3, Archaea = 4, others = 5. The grouping into human pathogenic prevalence (p) is based on experience from several microbiological laboratories in the Netherlands in conjunction with international reports on pathogen prevalence (De Greeff, Mouton, Schoffelen, and Verduin 2019; European Cen- tre for Disease Prevention and Control 2010; World Health Organization 2018a). Group 1 (most prevalent microorganisms) consists of all microorganisms where the taxonomic class is Gammaproteobacteria or where the taxonomic genus is Enterococcus, Staphylococcus or Streptococcus. This group consequently contains all common Gram-negative bacteria, such as Pseudomonas and Legionella and all species within the order Enterobacterales. 3.2. Working with antimicrobial names or codes The AMR package includes the antibiotics data set, which comprises common laboratory information system codes, official names, anatomical therapeutic chemical (ATC) codes, de- fined daily doses (DDD) and more than 5,000 trade names of 464 antimicrobial agents (see Appendix A). The ATC code system and the reference list for DDDs have been developed and made available by the World Health Organization Collaborating Centre for Drug Statistics Methodology (WHOCC) to standardize pharmaceutical classifications (WHO Collaborating Centre for Drug Statistics Methodology 2018). All agents in the antibiotics data set have a unique antimicrobial ID, which is based on abbreviations used by the European Antimi- crobial Resistance Surveillance Network (EARS-Net), the largest publicly funded system for antimicrobial resistance surveillance in Europe (European Centre for Disease Prevention and Control 2018). Furthermore, the AMR package includes the antivirals data set containing antiviral agents, which is also described in the Appendix A. Properties of microorganisms The package contains functions to return a specific (taxonomic) property of a microor- ganism from the microorganisms data set (see Appendix A). Functions that start with mo_* can be used to retrieve the most recently defined taxonomic properties of any mi- croorganism quickly and conveniently. These functions rely on the as.mo() function in- ternally: mo_name(), mo_fullname(), mo_shortname(), mo_subspecies(), mo_species(), mo_genus(), mo_family(), mo_order(), mo_class(), mo_phylum(), mo_kingdom(), mo_type(), mo_gramstain(), mo_ref(),mo_authors(), mo_year(), mo_rank(), AMR: Working with Antimicrobial Resistance Data in R 8 mo_taxonomy(), mo_synonyms(), mo_info() and mo_url(). Determination of the Gram stain, by using mo_gramstain(), is based on the taxonomic subkingdom and phylum. Ac- cording to Cavalier-Smith (2002), who defined the subkingdoms Negibacteria and Posibac- teria, only the following phyla are Posibacteria: Actinobacteria, Chloroflexi, Firmicutes and Tenericutes. Bacteria from these phyla are considered Gram-positive – all other bacteria are considered Gram-negative. Gram stains are only relevant for species within the kingdom of Bacteria. For species outside this kingdom, mo_gramstain() will return NA. Properties of antimicrobial agents It is a common task in microbiological data analyses (and other clinical or epidemiological fields) to work with different antimicrobial agents. The AMR package provides several func- tions to translate inputs such as ATC codes, abbreviations, or names in any direction. Using as.ab(), any input will be transformed to an antimicrobial ID of class ‘ab’. Helper functions are available to get specific properties of antimicrobial IDs, such as ab_name() for getting the official name, ab_atc() for the ATC code, or ab_cid() for the compound ID (CID) used by PubChem (Kim et al. 2019). Trade names can be also used as input. For example, the input values "Amoxil", "dispermox", "amox" and "J01CA04" all return the ID of amoxicillin (AMX): R> as.ab("Amoxicillin") Selecting and filtering data based on classes of antimicrobial agents The application of the ATC classification system also enables grouping of antimicrobial agents for data analyses. Data sets with microbial isolates can be filtered on isolates with spe- cific results for tested antimicrobial agents in a specific antimicrobial class. For example, carbapenems() can be used to select columns or filter rows based on any of the 14 available antimicrobial agents in the group of carbapenems according to the antibiotics data set. [1] "Amoxicillin" If more than one antimicrobial agent is found in the input string, a warning with the additional findings is printed to the console. R> ab_name("J01CA04") R> ab_name("J01CA04") Class <ab> [1] AMX R> as.ab(c("Amoxil", "dispermox", "amox", "J01CA04")) R> as.ab(c("Amoxil", "dispermox", "amox", "J01CA04")) R> ab_name("Amoxil") [1] "Amoxicillin" Journal of Statistical Software Class Minimum Maximum Unit ‘mic’ ≤0.001 ≥1024 µg/ml ‘disk’ ≤6 ≥50 mm Table 4: Antimicrobial suceptibility test classes. Journal of Statistical Software Class Minimum Maximum Unit ‘mic’ ≤0.001 ≥1024 µg/ml ‘disk’ ≤6 ≥50 mm Table 4: Antimicrobial suceptibility test classes. 9 Journal of Statistical Software 9 R> ab_atc("amox") Class <rsi> [1] R When using the as.rsi() function on existing antimicrobial interpretations, it tries to coerce the input to the values “R”, “S” or “I”. These values can in turn be used to calculate the proportion of antimicrobial resistance. 3.3. Working with antimicrobial susceptibility test results coli", "ampicillin", guideline = "EUCAST 2021") Class <rsi> [1] S 3.3. Working with antimicrobial susceptibility test results Minimal inhibitory concentrations (MIC) are susceptibility test results measured by microbi- ological laboratory equipment to determine at which minimum antimicrobial drug concentra- tion 99.9% of a microorganism is inhibited in growth. These concentrations are often capped at a minimum and maximum, for example ≤0.02 µg/ml and ≥32 µg/ml, respectively. The ‘mic’ class, an ordered ‘factor’ containing valid MIC values, keeps these operators while still ordering all possible outcomes correctly so that e.g., "<= 0.02" will be considered lower than "0.04".f Another susceptibility testing method is the use of drug diffusion disks, which are small tablets containing a specified concentration of an antimicrobial agent. These disks are applied onto a solid growth medium or a specific agar plate. After 24 hours of incubation, the diameter of the growth inhibition around a disk can be measured in millimeters with a ruler. The ‘disk’ class can be used to clean these kinds of measurements, since they should always be valid numeric values between 6 and 50. The supported minima and maxima of valid values for both classes, ‘mic’ and ‘disk’, are displayed in Table 4. The higher the MIC or the smaller the growth inhibition diameter, the more active substance of an antimicrobial agent is needed to inhibit cell growth, i.e., the higher the antimicrobial resistance against the tested antimicrobial agent. At high MICs and small diameters, guide- lines interpret the microorganism as “resistant” (R) to the tested antimicrobial agent. At low MICs and wide diameters, guidelines interpret the microorganism as “susceptible” (S) to the tested antimicrobial agent. In between, the microorganism is classified as “susceptible, AMR: Working with Antimicrobial Resistance Data in R 10 increased exposure” (I). For these three interpretations the ‘rsi’ class has been developed. When using as.rsi() on MIC values (of class ‘mic’) or disk diffusion diameters (of class ‘disk’), the values will be interpreted according to the guidelines from the CLSI or EUCAST (Clinical and Laboratory Standards Institute 2019; The European Committee on Antimicro- bial Susceptibility Testing 2020, all guidelines between 2011 and 2021 are included in the AMR package). Guidelines can be changed by setting the guidelines argument. R> # Low MIC value R> as.rsi(as.mic(2), "E. coli", "ampicillin", guideline = "EUCAST 2021") R> # Low MIC value R> as.rsi(as.mic(2), "E. coli", "ampicillin", guideline = "EUCAST 2021" Class <rsi> [1] S R> # High MIC value R> as.rsi(as.mic(32), "E. 3.4. Interpretative rules by EUCAST Next to supplying guidelines to interpret raw MIC values, the EUCAST has developed a set of expert rules to assist clinical microbiologists in the interpretation and reporting of antimicrobial susceptibility tests (Leclercq et al. 2013). The rules comprise assistance on in- trinsic resistance, exceptional phenotypes, and interpretive rules. The AMR package covers intrinsic resistant and interpretive rules for data transformation and standardization pur- poses. The first prevents false susceptibility reporting by providing a list of organisms with known intrinsic resistance to specific antimicrobial agents (e.g., cephalosporin resistance of all enterococci). Interpretative rules apply inference from established resistance mechanisms (Winstanley and Courvalin 2011; Courvalin 1992, 1996; Livermore, Winstanley, and Shannon 2001). Both groups of rules are based on classic IF THEN statements (e.g., IF Enterococcus spp. resistant to ampicillin THEN also report as resistant to imipenem). Some rules provide assistance for further actions when certain resistance has been detected, i.e., performing ad- ditional testing of the isolated microorganism. The AMR package function eucast_rules() can apply all EUCAST rules that do not rely on additional clinical information, such as addi- tional information on patients’ diagnoses. Table 2 and 3 highlight the transformation for the reporting of AMX = S in patient_id = 000003 to the correct report according to EUCAST rules of AMX = R. Of note, however, EUCAST rules overwrite original data to correct for the difference in how antimicrobial agents affect the tested microorganism in vitro (in the laboratory) and in vivo (in the human body). This requires users to closely collaborate with the data source provider to ensure correct versioning, backward compatibility, reproducibility, and taking into account specific local regulation for resistance reporting. Typical scenarios 11 Journal of Statistical Software where changes to the original data points apply include in vitro test results indicating sus- ceptibility when resistance in vivo is known. The changes are based on scientific consensus to ensure reliable high-quality reporting of antimicrobial susceptibility results. All changes to the data are printed to the console and can also be reviewed in detail by setting the argument eucast_rules(..., verbose = TRUE). EUCAST rules are subject to regular updates which are implemented into the AMR pack- age by the AMR maintenance team shortly after publication. The eucast_rules() func- tion supports versioning of the rules. The arguments version_breakpoints and version_ expertrules can be set to current or previous versions. By default, the eucast_rules() function uses the latest implemented version. [1] 1.5 R> atc_online_groups("amoxicillin") [1] "ANTIINFECTIVES FOR SYSTEMIC USE" [2] "ANTIBACTERIALS FOR SYSTEMIC USE" [3] "BETA-LACTAM ANTIBACTERIALS, PENICILLINS" [4] "Penicillins with extended spectrum" 3.5. Working with defined daily doses (DDD) DDDs are essential for standardizing antimicrobial consumption analysis, for inter-institutional or international comparison. The official DDDs are published by the WHOCC (WHO Collab- orating Center for Drug Statistics Methodology 2019). Updates to the official publication are monitored by the AMR maintenance team and implemented in the antibiotics data set in- cluded in the AMR package. Other metrics exist such as the recommended daily dose (RDD) or the prescribed daily dose (PDD). However, DDDs are the only metric that is independent of a patient’s disease and therapeutic choices and thus suitable for the AMR package. Functions from the atc_online_() family take any text as input that can be coerced with as.ab() (i.e., to class ‘ab’). Next, the functions access the WHOCC online registry (WHO Collaborating Center for Drug Statistics Methodology 2019, internet connection re- quired) and download the property defined in the arguments (e.g., administration = "O" or administration = "P" for oral or parenteral administration and property = "ddd" or property = "groups" to get DDD or the group of the selected antimicrobial defined by its ATC code). R> atc_online_ddd("amoxicillin", administration = "O") [1] 1.5 R> atc_online_groups("amoxicillin") [1] "ANTIINFECTIVES FOR SYSTEMIC USE" [2] "ANTIBACTERIALS FOR SYSTEMIC USE" [3] "BETA-LACTAM ANTIBACTERIALS, PENICILLINS" [4] "Penicillins with extended spectrum" atc_online_ddd("amoxicillin", administration = "O") R> atc_online_ddd("amoxicillin", administration = "O") AMR: Working with Antimicrobial Resistance Data in R 12 AMR: Working with Antimicrobial Resistance Data in R statistics is dependent on two prerequisites: the data should only comprise the first isolates and a minimum required number of 30 isolates should be met for every stratum in further analysis (Clinical and Laboratory Standards Institute 2014). An isolate is a microorganism strain cultivated on specified growth media in a laboratory, so its phenotype can be determined. First isolates are isolates of any species found first in a patient per episode, regardless of the body site or the type of specimen (such as blood or urine) (Clinical and Laboratory Standards Institute 2014). The selection on first isolates (using function first_isolate()) is important to prevent selection bias, as it would lead to overestimated or underestimated resistance to an antimicrobial agent. For example, if a patient is admitted with a multi-drug resistant microorganism and that microorganism is found in five different blood cultures the following week, it would overestimate resistance if all isolates were to be included in the analysis. The episode in days can be set with the argument episode_days, which defaults to 365 as suggested by the Clinical and Laboratory Standards Institute (2014) guideline. 4.1. Determining first isolates Determining antimicrobial resistance or susceptibility can be done for a single agent (mono- therapy) or multiple agents (combination therapy). The calculation of antimicrobial resistance Multi-drug resistant tuberculosis Tuberculosis is a major threat to global health caused by Mycobacterium tuberculosis (MTB) and is one of the top ten causes of death worldwide (World Health Organization 2018b). Exceptional antimicrobial resistance in MTB is therefore of special interest. To this end, the international WHO guideline for the classification of drug resistance in MTB (World Health Organization 2014) is included in the AMR package. The mdr_tb() function is a convenient wrapper around mdro(..., guideline = "TB"), which returns an other or- dered ‘factor’ than other mdro() functions. The output will contain the ‘factor’ levels Negative < Mono-resistant < Poly-resistant < Multi-drug-resistant < Extensive drugresistant, following the WHO guideline. 5.1. Calculation of antimicrobial resistance The AMR package contains several functions for fast and simple resistance calculations of bacterial or fungal isolates. A minimum number of available isolates is needed for the relia- bility of the outcome. The CLSI guideline suggests a minimum of 30 available first isolates irrespective of the type of statistical analysis (Clinical and Laboratory Standards Institute 2014). Therefore, this number is used as the default setting for any function in the pack- age that calculates antimicrobial resistance or susceptibility, which can be changed with the minimum argument in all applicable functions. 4.2. Determining multi-drug resistant organisms (MDRO) Abbreviations: mo = microorganism, AMC = amoxicilline/clavulanic acid, GEN = gentamicin, TOB = tobramycin, CIP = ciprofloxacin, MFX = moxifloxacin, MDRO = multi-drug resistant organism, B_ESCHR_COLI = microorganism code of Es- cherichia coli. Table 5: Example of a multi-drug resistant organism (MDRO) in a data set identified by apply- ing Dutch guidelines. Abbreviations: mo = microorganism, AMC = amoxicilline/clavulanic acid, GEN = gentamicin, TOB = tobramycin, CIP = ciprofloxacin, MFX = moxifloxacin, MDRO = multi-drug resistant organism, B_ESCHR_COLI = microorganism code of Es- cherichia coli. 4.2. Determining multi-drug resistant organisms (MDRO) Definitions of multi-drug resistant organisms (MDRO) are regulated by national and interna- tional expert groups and differ between nations. The AMR package provides the functionality to quickly identify MDROs in a data set using the mdro() function. Guidelines can be set with the argument guideline. At default, it applies the guideline as proposed by Magio- rakos et al. (2012). Their work describes the definitions for bacteria being “MDR” (multi- drug-resistant), “XDR” (extensively drug-resistant) or “PDR” (pan-drug-resistant). These definitions are widely adopted (Abat, Fournier, Jimeno, Rolain, and Raoult 2018) and known in the field of medical microbiology. Other guidelines currently supported are the international EUCAST guideline (guideline = "EUCAST", European Committee on Antimicrobial Susceptibility Testing 2016), the interna- tional WHO guideline on the management of drug-resistant tuberculosis, (guideline = "TB", World Health Organization 2014), and the national guidelines of The Netherlands (guideline = "NL", Werkgroep Infectiepreventie 2011), and Germany (guideline = "DE", Müller et al. 2015). Some guidelines require a minimum availability of tested antimicrobial agents per isolate. This is needed to prevent false-negatives, since no reliable determination can be performed on only a few test results. This required minimum defaults to 50%, but can be set by the user with the pct_minimum_classes. Isolates that do not meet this requirement will be skipped for determination and will return NA (not applicable), with an informative warning printed to the console. The rules are applied per row of the data. The mdro() function automatically identifies the variables containing the microorganism codes and antimicrobial agents based on the guess_ab_col() function. Following the guideline set by the user, it analyzes the specific antimicrobial resistance of a microorganism and flags that microorganism accordingly. The outcome is demonstrated in Table 5, where the first row is an MDRO according to the Dutch guidelines (Werkgroep Infectiepreventie 2011). The returned value is an ordered ‘factor’ with the levels Negative < Positive, unconfirmed < Positive. For some guideline rules that require additional testing (e.g., molecular confir- mation), the level Positive, unconfirmed is returned. Journal of Statistical Software 13 mo AMC GEN TOB CIP MFX MDRO B_ESCHR_COLI S R R R R Positive B_ESCHR_COLI R S R R S Negative B_ESCHR_COLI S S S R S Negative Table 5: Example of a multi-drug resistant organism (MDRO) in a data set identified by apply- ing Dutch guidelines. Proportions Calculation of antimicrobial resistance is carried out by counting the number of first resis- tant isolates (interpretation of “R”) and dividing it by the number of all first isolates, see Equation 1. This is implemented in the proportion_R() function. To calculate antimi- crobial susceptibility, the number of susceptible first isolates (interpretation of “S” and “I”) has to be counted and divided by the number of all first isolates, which is implemented in the proportion_SI() function. For convenience, the resistance() function is an alias of the proportion_R() function, and the susceptibility() function is an alias of the proportion_SI() function. The functions proportion_R(), proportion_IR(), proportion_I(), proportion_SI(), and proportion_S() follow the same logic as the count_() functions and all return a vector of class ‘double’ with a value between 0 and 1. The argument minimum defines the minimal allowed number of available (tested) isolates (default: minimum = 30). Any number below the set minimum will return NA with a warning. For calculating the proportion (P) of antimicrobial resistance or susceptibility to one antimi- crobial agent, the following equation is used: P(x,o) = Pk i=1[xi ∈o] Pk i=1[xi ∈{R, S, I}] , (1) (1) where P is the proportion of outcome o (that is either “R”, “S”, “I”, or a combination of two of them), where x is a character vector of length k only consisting of values “R”, “S”, or “I” and [xi ∈o] is the indicator function, returning 1 if the indicator function is true and 0 otherwise. The denominator must include the collection {R, S, I} so that “wrong” elements in x (i.e., elements not being “R”, “S”, or “I”) will not be counted. Thus, the theoretical antimicrobial susceptibility of the vector x = {S, S, I, R, R} is: P(x,o={S,I}) = 3 5 = 0.6 P(x,o={S,I}) = 3 5 = 0.6 For the proportion of empiric susceptibility (s) for more than one antimicrobial agent, the calculation can be carried out in two ways (Table 6). The first method is to count the total number of first isolates where at least one agent was tested as “S” or “I” and divide it by the number of first isolates tested where any of the agents was tested (Equation 2). Counts The AMR package relies on the concept of tidy data (Wickham 2014), although not strictly following its rules (one row per test rather than one row per observation). Function names to calculate the number of available isolates follow these general resistance interpretation standards with count_S(), count_I(), and count_R() respectively. Combinations of an- timicrobial resistance interpretations can be counted with count_SI() and count_IR(). All these functions take a vector of interpretations of the class ‘rsi’ (as discussed above) or are internally transformed with as.rsi(). The returned value is the sum of the respective inter- pretation in the selected test column. All count_() functions support quasi-quotation with pipes, grouped variables, and can be used with dplyr::summarize() (Wickham, François, Henry, and Müller 2022). 14 AMR: Working with Antimicrobial Resistance Data in R AMR: Working with Antimicrobial Resistance Data in R Proportions This method will be used when setting only_all_tested = FALSE in the susceptibility() function: s(x,y) = Pk i=1[xi ∈{S, I} ∨yi ∈{S, I}] Pk i=1[xi ∈{R, S, I} ∨yi ∈{R, S, I}] , (2) (2) where x is a character vector only consisting of values “R”, “S”, or “I” (i.e., “agent A”) and y is another character vector only consisting of values “R”, “S”, or “I” (i.e., “agent B”).i The second method is to count the total number of first isolates where at least one agent was tested as “S” or “I” and where all agents were tested divided by the number of first isolates tested where all of the agents were tested (Equation 3). This method will be used when setting only_all_tested = TRUE in the susceptibility() function: s′ (x,y) = Pk i=1[(xi ∈{S, I} ∨yi ∈{S, I}) ∧xi ∈{R, S, I} ∧yi ∈{R, S, I}] Pk i=1[xi ∈{R, S, I} ∧yi ∈{R, S, I}] s′ (x,y) = Pk i=1[(xi ∈{S, I} ∨yi ∈{S, I}) ∧xi ∈{R, S, I} ∧yi ∈{R, S, I}] Pk i=1[xi ∈{R, S, I} ∧yi ∈{R, S, I}] (3) Journal of Statistical Software 15 Antimicrobial agent All isolates Only isolates tested for both agents (only_all_tested = FALSE) (only_all_tested = TRUE) Agent A Agent B Include as Include as Include as Include as numerator denominator numerator denominator S or I S or I ✓ ✓ ✓ ✓ R S or I ✓ ✓ ✓ ✓ N/A S or I ✓ ✓ S or I R ✓ ✓ ✓ ✓ R R ✓ ✓ N/A R S or I N/A ✓ ✓ R N/A N/A N/A Table 6: Example calculation for determining empiric susceptibility (%SI) for more than one antimicrobial agent. Abbreviations: R = resistant, S = susceptible, I = susceptible/increased exposure, N/A = not tested/missing. Antimicrobial agent All isolates Only isolates tested for both agents (only_all_tested = FALSE) (only_all_tested = TRUE) Agent A Agent B Include as Include as Include as Include as numerator denominator numerator denominator S or I S or I ✓ ✓ ✓ ✓ R S or I ✓ ✓ ✓ ✓ N/A S or I ✓ ✓ S or I R ✓ ✓ ✓ ✓ R R ✓ ✓ N/A R S or I N/A ✓ ✓ R N/A N/A N/A Table 6: Example calculation for determining empiric susceptibility (%SI) for more than one antimicrobial agent. R> example_isolates %>% summarize( + si_gen_amx = proportion_SI(GEN, AMX), n_gen_amx = n_rsi(GEN, AMX), + n_total = n()) Proportions Abbreviations: R = resistant, S = susceptible, I = susceptible/increased exposure, N/A = not tested/missing. Table 6: Example calculation for determining empiric susceptibility (%SI) for more than one antimicrobial agent. Abbreviations: R = resistant, S = susceptible, I = susceptible/increased exposure, N/A = not tested/missing. Table 6: Example calculation for determining empiric susceptibility (%SI) for more than one antimicrobial agent. Abbreviations: R = resistant, S = susceptible, I = susceptible/increased exposure, N/A = not tested/missing. Based on Equation 1, the overall resistance and susceptibility of antimicrobial agents like gentamicin (GEN) and amoxicillin (AMX) can be calculated using the following syntax. The example_isolates is an example data set included in the AMR package, see Appendix A. The n_rsi() function is analogous to the n() function of the dplyr package. It counts the number of available isolates, but only includes observations with valid antimicrobial results (i.e., “R”, “S”, or “I”). R> example_isolates %>% summarize( p _ + r_gen = proportion_R(GEN), r_amx = proportion_R(AMX), + n_gen = n_rsi(GEN), n_amx = n_rsi(AMX), n_total = n()) + r_gen = proportion_R(GEN), r_amx = proportion_R(AMX), + n_gen = n_rsi(GEN), n_amx = n_rsi(AMX), n_total = n()) r_gen r_amx n_gen n_amx n_total 1 0.2458221 0.5955556 1855 1350 2000 This output reads: the resistance to gentamicin of all isolates in the example_isolates data set is P(x = GEN, o = {R}) = 24.6%, based on 1,855 out of 2,000 available isolates. This means that the susceptibility is P(x = GEN, o = {S, I}) = 75.4%. The susceptibility to amoxicillin is P(x = AMX, o = {S, I}) = 40.4% based on 1,350 isolates.f To calculate the effect of combination therapy, i.e., treating patients with multiple agents at the same time, all proportion_() functions can handle multiple variables as arguments as defined in Equation 2 and 3. 6. Design decisions The AMR package follows the rationale of tidyverse packages as authored by Wickham et al. (2019). Most functions take a ‘data.frame’ or ‘tibble’ as input, support piping (%>%) operations, can work with quasi-quotations, and can be integrated into dplyr workflows, such as mutate() to create new variables and group_by() to group by variables. Although the AMR package integrates well into tidyverse workflows, it can also be used with base R only. To this extent, the AMR package was developed to be independent of any other R package to ensure and maintain sustainability. The AMR package supports multiple languages. Currently supported languages are Dan- ish, Dutch, English, French, German, Italian, Portuguese, Russian, Spanish and Swedish. The system language will be used if the language is supported but can be overwritten with options(AMR_locale = ...). Multi-language support affects language-dependent output of functions such as mo_name(), mo_gramstain(), mo_type(), and ab_name().f The AMR package uses S3 classes, object oriented systems available in R. They allow different types of output based on the user input. The AMR package introduces five S3 classes (‘mo’, ‘ab’, ‘rsi’, ‘mic’, and ‘disk’) to increase the convenience when working with antimicrobial susceptibility data. Proportions For example, to calculate the empiric susceptibility of a combination therapy comprising gentamicin (GEN) and amoxicillin (AMX): R> example_isolates %>% summarize( + si_gen_amx = proportion_SI(GEN, AMX), n_gen_amx = n_rsi(GEN, AMX), + n_total = n()) R> example_isolates %>% summarize( + si_gen_amx = proportion_SI(GEN, AMX), n_gen_amx = n_rsi(GEN, AMX), + n_total = n()) AMR: Working with Antimicrobial Resistance Data in R 16 si_gen_amx n_gen_amx n_total 1 0.931843 1921 2000 1 This leads to the conclusion that combining gentamicin with amoxicillin would cover s(x = GEN, y = AMX) = 93.2% based on 1,921 out of 2,000 available isolates, which is 17.8% more than when treating with gentamicin alone (P(x = GEN, o = {S, I}) = 75.4%). With these functions, exact calculations can be done to evaluate the empiric success of treat- ing infections with one or more antimicrobial agents. 7. Reproducible example We consider the problem of working with antimicrobial resistance data from three different hospitals between 2011-01-01 and 2020-01-01. After loading the AMR package and additional tidyverse packages to allow transformation and plotting, we load the example_isolates_ unclean example data from the AMR package into the global environment and assign it a new name. R> library("dplyr") R> library("tidyr") R> library("AMR") R> options(AMR_locale = "en") R> data <- example_isolates_unclean R> glimpse(data) Rows: 3,000 Columns: 8 R> library("dplyr") R> library("tidyr") R> library("AMR") R> options(AMR_locale = "en") R> data <- example_isolates_unclean R> glimpse(data) Rows: 3,000 Columns: 8 R> library("dplyr") R> library("tidyr") R> library("AMR") R> options(AMR_locale = "en") R> data <- example_isolates_unclean R> glimpse(data) 17 Journal of Statistical Software 17 $ patient_id <chr> "J3", "R7", "P3", "P10", "B7", "W3", "J8", "M3", ... $ hospital <chr> "A", "A", "A", "A", "A", "A", "A", "A", "A", "A",... $ date <date> 2012-11-21, 2018-04-03, 2014-09-19, 2015-12-10, ... $ bacteria <chr> "E. coli", "K. pneumoniae", "E. coli", "E. coli",... $ AMX <chr> "R", "R", "R", "S", "S", "R", "R", "R", "S", "S",... $ AMC <chr> "I", "I", "S", "I", "S", "S", "S", "S", "S", "S",... $ CIP <chr> "S", "S", "S", "S", "S", "R", "S", "S", "S", "S",... $ GEN <chr> "S", "S", "S", "S", "S", "S", "S", "S", "S", "S",... $ patient_id <chr> "J3", "R7", "P3", "P10", "B7", "W3", "J8", "M3", ... $ hospital <chr> "A", "A", "A", "A", "A", "A", "A", "A", "A", "A",... $ date <date> 2012-11-21, 2018-04-03, 2014-09-19, 2015-12-10, ... $ bacteria <chr> "E. coli", "K. pneumoniae", "E. coli", "E. coli",... $ AMX <chr> "R", "R", "R", "S", "S", "R", "R", "R", "S", "S",... $ AMC <chr> "I", "I", "S", "I", "S", "S", "S", "S", "S", "S",... $ CIP <chr> "S", "S", "S", "S", "S", "R", "S", "S", "S", "S",... $ GEN <chr> "S", "S", "S", "S", "S", "S", "S", "S", "S", "S",... [1] "E. coli" "K. pneumoniae" [3] "S. aureus" "S. pneumoniae" [5] "klepne" "strpne" [7] "esccol" "staaur" [9] "Escherichia coli" "Staphylococcus aureus" [11] "Streptococcus pneumoniae" "Klebsiella pneumoniae" bacteria n 1 E. coli 494 2 esccol 508 3 Escherichia coli 516 4 K. pneumoniae 108 5 Klebsiella pneumoniae 102 6 klepne 116 7 S. aureus 247 8 S. pneumoniae 151 9 staaur 240 10 Staphylococcus aureus 243 11 Streptococcus pneumoniae 139 12 strpne 136 The data contains 3,000 observations of 8 variables from 3 hospitals. The bacteria variable comprises 12 unique elements. 7. Reproducible example However, they appear to encode the same information in dif- ferent formats ("E. coli", "K. pneumoniae", "S. aureus", "S. pneumoniae", "klepne", "strpne", "esccol", "staaur", "Escherichia coli", "Staphylococcus aureus", "Streptococcus pneumoniae", "Klebsiella pneumoniae"). We can use the as.mo() func- tion to standardize the bacterial codes and add a variable with the official scientific name. The correct transformation of the bacterial codes can be reviewed by calling the mo_uncertainties() function. The data contains 3,000 observations of 8 variables from 3 hospitals. The bacteria variable comprises 12 unique elements. However, they appear to encode the same information in dif- ferent formats ("E. coli", "K. pneumoniae", "S. aureus", "S. pneumoniae", "klepne", "strpne", "esccol", "staaur", "Escherichia coli", "Staphylococcus aureus", "Streptococcus pneumoniae", "Klebsiella pneumoniae"). We can use the as.mo() func- tion to standardize the bacterial codes and add a variable with the official scientific name. The correct transformation of the bacterial codes can be reviewed by calling the mo_uncertainties() function. AMR: Working with Antimicrobial Resistance Data in R 18 R> data < data %>% mutate( + bacteria = as.mo(bacteria), bacteria_name = mo_name(bacteria)) R> mo_uncertainties() Matching scores are based on pathogenicity in humans and the resemblance between the input and the full taxonomic name. See `?mo_matching_score`. "E. coli" -> Escherichia coli (B_ESCHR_COLI, 0.688) "E. coli" -> Escherichia coli (B_ESCHR_COLI, 0.688) Also matched: Entamoeba coli (0.079) "K. pneumoniae" -> Klebsiella pneumoniae (B_KLBSL_PNMN, 0.786) Also matched: Klebsiella pneumoniae ozaenae (0.707), Klebsiella pneumoniae pneumoniae (0.688) and Klebsiella pneumoniae rhinoscleromatis (0.658) "S. aureus" -> Staphylococcus aureus (B_STPHY_AURS, 0.690) Also matched: Staphylococcus aureus aureus (0.643), Streptomyces "K. pneumoniae" -> Klebsiella pneumoniae (B_KLBSL_PNMN, 0.786) p p _ _ , Also matched: Klebsiella pneumoniae ozaenae (0.707), Klebsiella pneumoniae pneumoniae (0.688) and Klebsiella pneumoniae rhinoscleromatis (0.658) "S. aureus" -> Staphylococcus aureus (B_STPHY_AURS, 0.690) Also matched: Staphylococcus aureus aureus (0.643), Streptomyces aureus (0.355) and Stentor aureus (0.052) "S. pneumoniae" -> Streptococcus pneumoniae (B_STRPT_PNMN, 0.750) Also matched: Spirabiliibacterium pneumoniae (0.700) R> data %>% count(bacteria, bacteria_name) bacteria bacteria_name n 1 B_ESCHR_COLI Escherichia coli 1518 2 B_KLBSL_PNMN Klebsiella pneumoniae 326 3 B_STPHY_AURS Staphylococcus aureus 730 4 B_STRPT_PNMN Streptococcus pneumoniae 426 In a next step, we can further enrich the data with additional microbial taxonomic data based on the bacteria variable, such as Gram-stain and microorganism family. 7. Reproducible example R> data <- data %>% mutate( + gram_stain = mo_gramstain(bacteria), family = mo_family(bacteria)) R> data %>% count(gram_stain) R> data <- data %>% mutate( + gram_stain = mo_gramstain(bacteria), family = mo_family(bacteria)) R> data %>% count(gram_stain) gram_stain n 1 Gram-negative 1844 2 Gram-positive 1156 R> data %>% count(family) family n 1 Enterobacteriaceae 1844 2 Staphylococcaceae 730 3 Streptococcaceae 426 gram_stain n 1 Gram-negative 1844 2 Gram-positive 1156 family n 1 Enterobacteriaceae 1844 2 Staphylococcaceae 730 3 Streptococcaceae 426 19 Journal of Statistical Software Journal of Statistical Software The variables AMX, AMC, CIP, and GEN contain antimicrobial susceptibility test results. The abbreviations stand for the tested antimicrobial agent. The official names and additional information about the antimicrobial agents can be checked with the ab_info() function from the AMR package. The variables AMX, AMC, CIP, and GEN contain antimicrobial susceptibility test results. The abbreviations stand for the tested antimicrobial agent. The official names and additional information about the antimicrobial agents can be checked with the ab_info() function from the AMR package. AMR: Working with Antimicrobial Resistance Data in R R> ab_info("AMX") R> data <- data %>% + mutate_at(vars(AMX:GEN), as.rsi) R> data %>% + select(AMX:GEN) %>% + pivot_longer(everything(), names_to = "antimicrobials", + values_to = "interpretation") %>% + count(interpretation) R> data <- data %>% + mutate_at(vars(AMX:GEN), as.rsi) R> data %>% + select(AMX:GEN) %>% + pivot_longer(everything(), names_to = "antimicrobials", + values_to = "interpretation") %>% + count(interpretation) # A tibble: 3 × 2 interpretation n <rsi> <int> 1 S 6288 2 I 1105 3 R 4607 # A tibble: 3 × 2 interpretation n <rsi> <int> 1 S 6288 2 I 1105 3 R 4607 After this transformation, the eucast_rules() function can be applied to apply the latest resistance reporting guidelines. R> data <- data %>% eucast_rules() R> data <- data %>% eucast_rules() The output to the console lists the changes made to data: The rules affected 508 out of 3,000 rows, making a total of 657 edits => added 0 test results The rules affected 508 out of 3,000 rows, making a total of 657 edits => added 0 test results => changed 657 test results - 11 test results changed from "S" to "I" - 473 test results changed from "S" to "R" - 85 test results changed from "I" to "R" - 19 test results changed from "I" to "S" - 33 test results changed from "R" to "I" - 36 test results changed from "R" to "S" => changed 657 test results - 11 test results changed from "S" to "I" - 473 test results changed from "S" to "R" - 85 test results changed from "I" to "R" - 19 test results changed from "I" to "S" - 33 test results changed from "R" to "I" - 36 test results changed from "R" to "S" The data is now clean and ready for further analysis, for example, the identification of multi- drug resistant microorganisms. In this example, we use the Dutch guideline to determine multi-drug resistance (Werkgroep Infectiepreventie 2011). R> ab_info("AMX") $tradenames [1] "actimoxi" "amoclen" "amolin" [4] "amopen" "amopenixin" "amoxibiotic" [7] "amoxicaps" "amoxicilina" "amoxicillin" [10] "amoxicilline" "amoxicillinum" "amoxiden" [13] "amoxil" "amoxivet" "amoxy" [16] "amoxycillin" "anemolin" "aspenil" [19] "biomox" "bristamox" "cemoxin" [22] "clamoxyl" "delacillin" "dispermox" [25] "efpenix" "flemoxin" "hiconcil" [28] "histocillin" "hydroxyampicillin" "ibiamox" [31] "imacillin" "lamoxy" "metafarma capsules [34] "metifarma capsules" "moxacin" "moxatag" [37] "ospamox" "pamoxicillin" "piramox" [40] "robamox" "sawamox pm" "tolodina" [43] "unicillin" "utimox" "vetramox" $loinc [1] "16365-9" "25274-2" "3344-9" "80133-2" "metafarma capsules" "moxatag" "piramox" "tolodina" "vetramox" AMR: Working with Antimicrobial Resistance Data in R 20 $ddd $ddd$oral $ddd$oral$amount [1] 1.5 $ddd$oral$units [1] "g" $ddd$iv $ddd$iv$amount [1] 3 $ddd$iv$units [1] "g" $ddd$oral$units [1] "g" In a data set containing antimicrobial names or codes (e.g., antimicrobial prescription data), the as.ab() function can be used to transform all values to valid antimicrobial codes. Extra columns with the official name and the defined daily dose (DDD) for intravenous administra- tion could be added using ab_name() and ab_ddd(). R> ab_example <- data.frame(agents = c("AMX", "Ceftriaxon", "Cipro")) R> ab_example <- data.frame(agents = c("AMX", "Ceftriaxon", "Cipro")) R> ab_example %>% mutate( + agents = as.ab(agents), agent_names = ab_name(agents), + ddd_iv = ab_ddd(agents, administration = "iv")) agents agent_names ddd_iv 1 AMX Amoxicillin 3.0 2 CRO Ceftriaxone 2.0 3 CIP Ciprofloxacin 0.8 Coming back to the cleaning of the data, the columns for the antimicrobial susceptibility test results (“AMX”, “AMC”, “CIP”, “GEN”) need to be checked to contain only standard values (“R”, “S”, “I”). R> data %>% + select(AMX:GEN) %>% + pivot_longer(everything(), names_to = "antimicrobials", + values_to = "interpretation") %>% + count(interpretation) # A tibble: 4 × 2 interpretation n <chr> <int> 1 < 0.5 S 143 2 I 1105 3 R 4607 4 S 6145 R> data %>% + select(AMX:GEN) %>% + pivot_longer(everything(), names_to = "antimicrobials", + values_to = "interpretation") %>% + count(interpretation) 21 Journal of Statistical Software The values contain some mixed values. The as.rsi() function can be used to clean these values and to assign a new class (‘rsi’) for further use of AMR functions. R> ab_info("AMX") R> data <- data %>% mutate(mdro = mdro(., guideline = "nl")) R> data %>% count(bacteria_name, mdro) bacteria_name mdro n 1 Escherichia coli Negative 1123 2 Escherichia coli Positive 395 3 Klebsiella pneumoniae Negative 237 4 Klebsiella pneumoniae Positive 89 5 Staphylococcus aureus Negative 730 6 Streptococcus pneumoniae Negative 426 bacteria_name mdro n 1 Escherichia coli Negative 1123 2 Escherichia coli Positive 395 3 Klebsiella pneumoniae Negative 237 4 Klebsiella pneumoniae Positive 89 5 Staphylococcus aureus Negative 730 6 Streptococcus pneumoniae Negative 426 22 AMR: Working with Antimicrobial Resistance Data in R According to the Dutch guideline, 484 multi-drug resistant microorganisms were found in 3,000 tested isolates. No multi-drug resistance was found in Staphylococcus aureus and Strep- tococcus pneumoniae. According to the Dutch guideline, 484 multi-drug resistant microorganisms were found in 3,000 tested isolates. No multi-drug resistance was found in Staphylococcus aureus and Strep- tococcus pneumoniae. As described in Section 4.1, the identification of first isolates is essential for the reporting of resistance patterns. Using the filter_first_isolate() function and proportion_df() in combination with group_by(), we get a complete resistance analysis per hospital, bacteria, first isolate, and tested antimicrobial agent in one call: R> resistance_proportion <- data %>% + filter_first_isolate() %>% + group_by(hospital) %>% + proportion_df() R> head(resistance_proportion) hospital antibiotic interpretation value 1 A Amoxicillin SI 0.5796253 2 A Amoxicillin R 0.4203747 3 A Amoxicillin/clavulanic acid SI 0.8103044 4 A Amoxicillin/clavulanic acid R 0.1896956 5 A Ciprofloxacin SI 0.7974239 6 A Ciprofloxacin R 0.2025761ii From the console we get the information how many first isolates were identified and used in the filter. From here on, the data is ready for further analysis with functions for plotting (e.g., the ggplot2 package Wickham 2016), AMR extension functions for base R (e.g., summary(), plot()), or AMR helper functions for plotting and basic modelling (e.g., ggplot_rsi(), geom_rsi(), resistance_predict()). Journal of Statistical Software All institutions admitted that there was no standard operating procedure to maintain their taxonomic reference data. Implementing and maintaining the taxonomic data for these and other institutions has been challenging, since no common machine-readable, re- liable and up-to-date resource for the microbial taxonomy was publicly available. For reliable reference data about antimicrobial agents, this also holds true. The AMR package provides machine-readable reference data files for the complete microbial taxonomy and for more than 500 antimicrobial agents. Using functions starting with mo_* and ab_, names of microorgan- isms and antimicrobial agents can be translated between different LIS codes or other forms of text codes for microorganisms and consequently allows to merge data sets from different sites with little effort. Secondly, antimicrobial resistance interpretation guidelines (Leclercq et al. 2013; Clinical and Laboratory Standards Institute 2014) and taxonomic definitions of microorganisms are under constant change and are continually published in dedicated peer-reviewed journals. This is further complicated by differences between local, regional, and national guidelines. Yet, comparability and reproducibility across setting and time are key in research and clinics. The AMR package functions eucast_rules() (to apply guidelines to data), mdro() (to check for multi-drug resistance according to guidelines), or first_isolate() (to determine first isolates according to guidelines) address the needs to standardize comparability, and empower data analysts beyond the capabilities of their local LIS. The AMR package can be used as an extra layer of data validation when retrieving raw data from a LIS. Overall, the functionality of the AMR package has the potential to improve data validity in clinical settings, to ease multi- center study workflows, and to foster research reporting practices. The inherent global nature of antimicrobial resistances requires researchers, clinicians, and policy makers to reach beyond the borders of their local laboratory. The AMR package can build the bridge to link these sources and further encourages open science principles through its open source approach. The AMR package also has limitations. It does not introduce novel statistical tests or mod- els, nor does it add additional analytical approaches for AMR research. The calculation of the proportion of susceptibility for more than one antimicrobial agent simultaneously (see Section 5.1) seems simple but is subject to unclear reporting in clinical practice (Schechner, Temkin, Harbarth, Carmeli, and Schwaber 2013; Ma et al. 2017). Journal of Statistical Software Journal of Statistical Software Firstly, simple codes for microorganisms show substantial differences between different LIS and presumably correct taxonomic names are often misspelled or outdated. We analyzed the taxonomic names of bacteria used in reports from seven different public health institutions that perform microbiological diagnostics in the Netherlands and compared them with an of- ficial scientific up-to-date source for microbial taxonomy, the Catalogue of Life (Bánki et al. 2022). These institutions cover microbiological diagnostics for hospitals and primary care for 15% of the total Dutch population. All institutions reported outdated taxonomic names with a maximum lag ranging between 34 and 41 years. Given that antimicrobial resistance guidelines are strongly based on the microbial taxonomy (some rules only apply to a specific genus, other rules apply to a specific family), it is crucial that this information is correct and timely updated. All institutions admitted that there was no standard operating procedure to maintain their taxonomic reference data. Implementing and maintaining the taxonomic data for these and other institutions has been challenging, since no common machine-readable, re- liable and up-to-date resource for the microbial taxonomy was publicly available. For reliable reference data about antimicrobial agents, this also holds true. The AMR package provides machine-readable reference data files for the complete microbial taxonomy and for more than 500 antimicrobial agents. Using functions starting with mo_ and ab_*, names of microorgan- isms and antimicrobial agents can be translated between different LIS codes or other forms of text codes for microorganisms and consequently allows to merge data sets from different sites with little effort. Firstly, simple codes for microorganisms show substantial differences between different LIS and presumably correct taxonomic names are often misspelled or outdated. We analyzed the taxonomic names of bacteria used in reports from seven different public health institutions that perform microbiological diagnostics in the Netherlands and compared them with an of- ficial scientific up-to-date source for microbial taxonomy, the Catalogue of Life (Bánki et al. 2022). These institutions cover microbiological diagnostics for hospitals and primary care for 15% of the total Dutch population. All institutions reported outdated taxonomic names with a maximum lag ranging between 34 and 41 years. Given that antimicrobial resistance guidelines are strongly based on the microbial taxonomy (some rules only apply to a specific genus, other rules apply to a specific family), it is crucial that this information is correct and timely updated. 8. Discussion For the first time, a free and open source software solution is available to cover all aspects of working with antimicrobial resistance data. The AMR package provides functionalities that enable standardized and reproducible workflows from raw laboratory data to publish- able results, for research and clinical workflows alike. In the field of clinical microbiology and infectious diseases, research and clinical workflows are closely linked. For example, a per- formed research study on the prevalence of antimicrobial-resistant bacteria can have direct implications on the choice of antimicrobial agents for the treatment of patients. The AMR package was developed to be used in any research or clinical setting where the data analysis on microorganisms, antimicrobial resistance, antimicrobial agents is required. Both, researchers and clinicians rely on the data from electronic laboratory information sys- tems (LIS) where laboratory test results are processed, stored, and archived. Although some commercial solutions exist to conduct medical microbiological data analysis, these solutions are not comprehensive enough to apply antimicrobial resistance analysis for any clinical or research setting. Costs of these tools are a further constraint in resource-limited settings. Moreover, researchers and clinicians that require data from multiple LIS sources to be used in multi-center studies experience major barriers which cannot be solved by available commercial solutions. 23 Journal of Statistical Software The lack of clearly defined algorithms can lead to the effect that co-resistance rates for more than one antimicrobial agent are dropped altogether (Baur et al. 2017). The inclusion of isolates that are tested for some agents (only_all_tested = FALSE) or only isolates tested for all agents (only_all_tested = TRUE) can have an imminent clinical impact on patient care, if one combination of an- timicrobial agents is preferred over another. Therefore, the AMR package provides different algorithms to standardize this crucial calculation. Unfortunately, unambiguous methodol- AMR: Working with Antimicrobial Resistance Data in R 24 ogy for determining the right algorithm is lacking in scientific literature. An analysis on the algorithms used in the AMR package and their clinical impact is in preparation. Reliable information about antimicrobial resistance is vital for clinical decision-making in infectious diseases, since the outcome of local antimicrobial resistance analyses support med- ical professionals/clinicians in the treatment choices for their patients. Moreover, when this information can be reliably stratified by, for example, year, hospital, and type of patients, new information can lead to new insights for choosing the best antimicrobial therapy for patients suffering from infections. The AMR package enables this by providing all required analysis tools and can therefore empower decision-making in infectious management. The AMR package is already being applied to this end in six hospitals in the Netherlands. The choice of empirical antimicrobial treatment (meaning; choosing the initial therapy at a time of not knowing the infection-causing pathogen) for septic non-post-surgical patients has been altered in at least one Dutch hospital, by analyzing antimicrobial resistance data with the AMR package. The clinical effect of this adjustment is being studied at the moment. To improve the quality of such analyses, planned future developments comprise the implementa- tions of an imputation algorithm specifically for antimicrobial agents, and method guidance for applying prediction modelling in a health care setting based on patient-specific properties. Since the first package release, users from different public and private settings have been sug- gesting additional functionalities, in particular, the incorporation of country- or time-specific guidelines (e.g., Magiorakos et al. 2012). This community-centered development will be con- tinued and maintained by researchers at the University Medical Center Groningen and data scientists at Certe Medical Diagnostics and Advice, both non-profit public health organiza- tions located in Groningen, the Netherlands. Journal of Statistical Software Moreover, a group of contributors from five different Dutch health care institutions has been formed at the Dutch Association for Med- ical Microbiology (Nederlandse Vereniging voor Medische Microbiologie - NVMM) that also peer-review major changes to the package, including the implementation of guideline updates. This way, updates required for scientific developments as well as maintaining consistent repro- ducibility are ensured. Updates to databases and guidelines included in the AMR package are incorporated on a regular and automated basis, while preserving version control. Any function making use of guidelines (e.g., eucast_rules()) refers to the latest implemented version of the guideline by default. The aim of the AMR package is to provide a comprehensive toolbox of solutions for antimi- crobial resistance data processing and analysis on an institution- and country-independent scale for clinical practice and research that are required according to international standards, but were not available to date. Journal of Statistical Software Journal of Statistical Software While the burden of antimicrobial resistance is increasing worldwide, reliable data and data analyses are needed to better understand current and future developments. Open source approaches, such as the AMR package for R, have the potential to help democratizing the required tools in the field for researchers, clinicians, and policy makers alike. In organizations or countries with very limited resources, this free and open-source package could also over- come a financial limitation that would otherwise hinder antimicrobial resistance analysis in these settings. Across settings, we believe the AMR package can be used to support clinical decision-making in infection management by providing improved insight into current local and regional resistance levels. Furthermore, data analysis approaches based on individual patient or microbiological data, which the AMR package enables, fosters empowerment of laboratory staff, infection control practitioners, and public health services. Computational details The results in this paper were obtained using R 4.2.0 in RStudio 2022.07.1 (RStudio, PBC 2022) with the AMR package 1.8.2, running under macOS Monterey 12.5. R itself and all packages used are available from the CRAN at https://CRAN.R-project. The results in this paper were obtained using R 4.2.0 in RStudio 2022.07.1 (RStudio, PBC 2022) with the AMR package 1.8.2, running under macOS Monterey 12.5. R itself and all packages used are available from the CRAN at https://CRAN.R-project. org/. All development versions of the AMR package are available at https://github.com/ msberends/AMR/. R itself and all packages used are available from the CRAN at https://CRAN.R-project. org/. All development versions of the AMR package are available at https://github.com/ msberends/AMR/. Acknowledgments The authors Matthijs S. Berends and Christian F. Luz contributed equally to this publication. For their contributions to the development of the AMR package, we would like to thank (in alphabetical order) Judith M. Fonville, Erwin E.A. Hassing, Eric H.L.C.M. Hazenberg, Gwen Knight, Annick Lenglet, Bart C. Meijer, Sofia Ny, Rogier P. Schade, Dennis Souverein, and Anthony Underwood. The development of the AMR package was partly supported by the INTERREG V A (202085) funded project EurHealth-1Health (http://www.eurhealth1health.eu), part of a Dutch- German cross-border network supported by the European Commission, the Dutch Ministry of Health, Welfare and Sport, the Ministry of Economy, Innovation, Digitalization and Energy of the German Federal State of North Rhine-Westphalia and the Ministry for National and European Affairs and Regional Development of Lower Saxony. Furthermore, the AMR package was developed as part of a project funded by the Euro- pean Commission Horizon 2020 Framework Marie Skłodowska-Curie Actions (grant agree- ment number: 713660-PRONKJEWAIL-H2020-MSCA-COFUND-2015). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. 9. Summary This paper demonstrates the AMR package and its use for working with antimicrobial re- sistance data. It can be used to clean, enhance, and analyze such data according to (in- ter)national recommendations and guidelines while incorporating scientifically reliable refer- ence data on microbiological laboratory test results, antimicrobial agents, and the biological taxonomy of microorganisms. Consequently, it allows for reproducible analyses, regardless of the many possible ways in which raw and uncleaned data are stored in laboratory information systems. 25 EUCAST (2019). “EUCAST New Definitions of S, I and R from 2019.” Accessed: 2019-12-02, URL https://www.eucast.org/newsiandr/. References Abat C, Fournier PE, Jimeno MT, Rolain JM, Raoult D (2018). “Extremely and Pandrug- Resistant Bacteria Extra-Deaths: Myth Or Reality?” European Journal of Clinical Micro- biology & Infectious Diseases, 37(9), 1687–1697. doi:10.1007/s10096-018-3300-0. AMR: Working with Antimicrobial Resistance Data in R 26 Bánki O, Roskov Y, Döring M, Ower G, Vandepitte L, Hobern D, Remsen D, Schalk P, DeWalt R, Keping M, Miller J, Orrell T, et al. (2022). Catalogue of Life Checklist. Catalogue of Life, Leiden. doi:10.48580/dfpz. Baur D, Gladstone BP, Burkert F, Carrara E, Foschi F, Döbele S, Tacconelli E (2017). “Effect of Antibiotic Stewardship On the Incidence of Infection and Colonisation with Antibiotic-Resistant Bacteria and Clostridium Difficile Infection: A Systematic Review and Meta-Analysis.” The Lancet Infectious Diseases, 17(9), 990–1001. doi:10.1016/ s1473-3099(17)30325-0. Berends MS, Luz CF, Friedrich AW, Sinha BNM, Albers CJ, Glasner C (2022). AMR: An- timicrobial Resistance Analysis. R package version 1.8.2, URL https://CRAN.R-project. org/package=AMR. Cavalier-Smith T (2002). “The Neomuran Origin of Archaebacteria, the Negibacterial Root of the Universal Tree and Bacterial Megaclassification.” International Journal of Systematic and Evolutionary Microbiology, 52(Pt 1), 7–76. doi:10.1099/00207713-52-1-7. CDC (2019). “AR Threats Report: Antibiotic Resistance Threats in the United States, 2019.” https://www.cdc.gov/drugresistance/pdf/threats-report/ 2019-ar-threats-report-508.pdf. Clinical and Laboratory Standards Institute (2014). “M39-A4, Analysis and Presentation of Cumulative Antimicrobial Susceptibility Test Data.” 4th edition. Clinical and Laboratory Standards Institute (2019). “Susceptibility Testing of Infectious Agents and Evaluation of Performance of Antimicrobial Susceptibility Test Devices – Part 1.” ISO 20776-1Ed2:2019, 2nd edition. Courvalin P (1992). “Interpretive Reading of Antimicrobial Susceptibility Tests. Molecular Analysis and Therapeutic Interpretation of In Vitro Tests to Improve Antibiotic Ther- apy.” ASM American Society for Microbiology News, 58(7), 368–375. doi:10.1111/j. 1469-0691.1996.tb00872.x. Courvalin P (1996). “Interpretive Reading of In Vitro Antibiotic Susceptibility Tests (the Antibiogramme).” Clinical Microbiology and Infection, 2, S26–S34. doi:10.1111/j. 1469-0691.1996.tb00872.x. De Greeff SC, Mouton JW, Schoffelen AF, Verduin CM (2019). “NethMap 2019: Consumption of Antimicrobial Agents and Antimicrobial Resistance among Medically Important Bacteria in the Netherlands / maran 2019: Monitoring of Antimicrobial Resistance and Antibiotic Usage in Animals in the Netherlands in 2018.” RIVM. De Kraker MEA, Stewardson AJ, Harbarth S (2016). “Will 10 Million People Die A Year Due to Antimicrobial Resistance By 2050?” PLOS Medicine, 13(11), e1002184. doi: 10.1371/journal.pmed.1002184. 27 Journal of Statistical Software European Centre for Disease Prevention and Control (2010). “European Antimicrobial Resistance Surveillance Network (EARS-Net).” https://ecdc.europa.eu/en/about-us/ partnerships-and-networks/disease-and-laboratory-networks/ears-net. Ac- cessed: 2019-04-09. Ac- European Centre for Disease Prevention and Control (2018). References “Antimicrobial Resistance (AMR) Reporting Protocol 2018. European Antimicrobial Resistance Surveillance Network (EARS-Net) Surveillance Data for 2017.” European Committee on Antimicrobial Susceptibility Testing (2016). “EUCAST Ex- pert Rules, Intrinsic Resistance and Exceptional Phenotypes Tables. Version 3.1.” http://www.eucast.org/fileadmin/src/media/PDFs/EUCAST_files/Expert_Rules/ Expert_rules_intrinsic_exceptional_V3.1.pdf. Jombart T, Kamvar ZN, FitzJohn R, Cai J, Bhatia S, Schumacher J, Pulliam JRC (2020). incidence: Compute, Handle, Plot and Model Incidence of Dated Events. R package ver- sion 1.7.3, URL https://CRAN.R-project.org/package=incidence. Kim S, Chen J, Cheng T, Gindulyte A, He J, He S, Li Q, Shoemaker BA, Thiessen PA, Yu B, Zaslavsky L, Zhang J, Bolton EE (2019). “PubChem 2019 Update: Improved Access to Chemical Data.” Nucleic Acids Research, 47(D1), D1102–D1109. doi:10.1093/nar/ gky1033. Leclercq R, Cantón R, Brown DFJ, Giske CG, Heisig P, MacGowan AP, Mouton JW, Nord- mann P, Rodloff AC, Rossolini GM, Soussy CJ, Steinbakk M, Winstanley TG, Kahlmeter G (2013). “EUCAST Expert Rules in Antimicrobial Susceptibility Testing.” Clinical Mi- crobiology and Infection, 19(2), 141–160. Levenshtein VI (1966). “Binary Codes Capable of Correcting Deletions, Insertions, and Re- versals.” Soviet Physics Doklady, 10(8), 707–710. doi:10.1109/isit.1994.394907. Limmathurotsakul D, Dunachie S, Fukuda K, Feasey NA, Okeke IN, Holmes AH, Moore CE, Dolecek C, Van Doorn HR, Shetty N, Lopez AD, Peacock SJ, Surveillance and Epidemiology of Drug Resistant Infections Consortium (2019). “Improving the Estimation of the Global Burden of Antimicrobial Resistant Infections.” Lancet Infectious Diseases. doi:10.1016/ s1473-3099(19)30276-2. Livermore DM, Winstanley TG, Shannon KP (2001). “Interpretative Reading: Recognizing the Unusual and Inferring Resistance Mechanisms From Resistance Phenotypes.” Journal of Antimicrobial Chemotherap, 48 Suppl 1, 87–102. doi:10.1093/jac/48.suppl_1.87. Lowndes JSS, Best BD, Scarborough C, Afflerbach JC, Frazier MR, O’Hara CC, Jiang N, Halpern BS (2017). “Our Path to Better Science in Less Time Using Open Data Science Tools.” Nature Ecology and Evolution, 1(6), 160. doi:10.1038/s41559-017-0160. Ma J, Li N, Liu Y, Wang C, Liu X, Chen S, Xie X, Gan S, Wang M, Cao W, Wang F, Liu Y, Wan D, Sun L, Sun H (2017). “Antimicrobial Resistance Patterns, Clini- cal Features, and Risk Factors for Septic Shock and Death of Nosocomial E Coli Bac- teremia in Adult Patients with Hematological Disease.” Medicine, 96(21), e6959. doi: 10.1097/md.0000000000006959. AMR: Working with Antimicrobial Resistance Data in R 28 Magiorakos AP, Srinivasan A, Carey RB, Carmeli Y, Falagas ME, Giske CG, Harbarth S, Hindler JF, Kahlmeter G, Olsson-Liljequist B, Paterson DL, Rice LB, Stelling J, Struelens MJ, Vatopoulos A, Weber JT, Monnet DL (2012). References “Multidrug-Resistant, Extensively Drug- Resistant and Pandrug-Resistant Bacteria: An International Expert Proposal for Interim Standard Definitions for Acquired Resistance.” Clinical Microbiology and Infection, 18(3), 268–81. doi:10.1111/j.1469-0691.2011.03570.x. Müller J, Voss A, Köck R, Sinha B, Rossen JW, Kaase M, Mielke M, Daniels-Haardt I, Jurke A, Hendrix R, Kluytmans JA, Kluytmans-van den Bergh MF, Pulz M, Herrmann J, Kern WV, Wendt C, Friedrich AW (2015). “Cross-Border Comparison of the Dutch and Ger- man Guidelines On Multidrug-Resistant Gram-Negative Microorganisms.” Antimicrobial Resistance and Infection Control, 4, 7. doi:10.1186/s13756-015-0047-6. Nagraj V, Jombart T, Randhawa N, Sudre B, Campbell F, Crellen T (2021). epicontacts: Handling, Visualisation and Analysis of Epidemiological Contacts. R package version 1.1.2, URL https://CRAN.R-project.org/package=epicontacts. O’Neill J (2014). Antimicrobial Resistance: Tackling a Crisis for the Health and Wealth of Nations. Review on Antimicrobial Resistance. Petzoldt T (2021). antbioticR: Analysis of Antibiotic Resistance Data. R package version 0.3.3, URL https://github.com/tpetzoldt/antibioticR. R Core Team (2022). R: A Language and Environment for Statistical Computing. R Founda- tion for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/. RStudio, PBC (2022). RStudio: Integrated Development Environment for R. Boston. URL https://www.RStudio.com/. Schechner V, Temkin E, Harbarth S, Carmeli Y, Schwaber MJ (2013). “Epidemiological Interpretation of Studies Examining the Effect of Antibiotic Usage On Resistance.” Clinical Microbiology Reviews, 26(2), 289–307. doi:10.1128/cmr.00001-13. Tacconelli E, Sifakis F, Harbarth S, Schrijver R, Van Mourik M, Voss A, Sharland M, Ra- jendran NB, Rodríguez-Baño J, EPI-Net COMBACTE-MAGNET Group (2018). “Surveil- lance for Control of Antimicrobial Resistance.” Lancet Infectious Diseases, 18(3), e99–e106. doi:10.1016/s1473-3099(17)30485-1. The European Committee on Antimicrobial Susceptibility Testing (2020). “Breakpoint Tables for Interpretation of mics and Zone Diameters, Version 10.0.” http://www.eucast.org/ clinical_breakpoints/. Werkgroep Infectiepreventie (2011). “Bijzonder Resistente Micro-Organismen (BRMO).” WHO Collaborating Center for Drug Statistics Methodology (2019). “ATC/DDD Index.” https://www.whocc.no/atc_ddd_index/. Accessed: 2019-04-09. WHO Collaborating Centre for Drug Statistics Methodology (2018). “Guidelines for atc Classification and ddd Assignment.” Oslo, Norway. WHO Collaborating Centre for Surveillance of Antimicrobial Resistance (2019). “WHONET.” http://www.whonet.org/. Accessed: 2019-07-16. 29 Journal of Statistical Software Wickham H (2014). “Tidy Data.” Journal of Statistical Software, 59(10), 1–23. doi:10. 18637/jss.v059.i10. Wickham H (2016). ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag. URL https://ggplot2.tidyverse.org/. Wickham H, Averick M, Bryan J, Chang W, McGowan LD, François R, Grolemund G, Hayes A, Henry L, Hester J, Kuhn M, Pedersen TL, Miller E, Bache SM, Müller K, Ooms J, Robinson D, Seidel DP, Spinu V, Takahashi K, Vaughan D, Wilke C, Woo K, Yutani H (2019). References “Welcome to the Tidyverse.” Journal of Open Source Software, 4(43), 1686. doi:10.21105/joss.01686. Wickham H, François R, Henry L, Müller K (2022). dplyr: A Grammar of Data Manipulation. R package version 1.0.9, URL https://CRAN.R-project.org/package=dplyr. Winstanley T, Courvalin P (2011). “Expert Systems in Clinical Microbiology.” Clinical Microbiology Reviews, 24(3), 515–556. doi:10.1128/cmr.00061-10. World Health Organization (2014). “Companion Handbook to the WHO Guidelines for the Programmatic Management of Drug-Resistant Tuberculosis.” World Health Organization (2018a). “Global Antimicrobial Resistance Surveillance System (GLASS) Report: Early Implementation 2017–2018.” World Health Organization (2018b). “Global Tuberculosis Report 2018.” AMR: Working with Antimicrobial Resistance Data in R 30 • antibiotics • antibiotics A ‘data.frame’ containing 464 antibiotic agents with 14 columns. All entries in this data set have three different identifiers: a human readable EARS-Net code (as used by ECDC (European Centre for Disease Prevention and Control 2010) and WHONET (WHO Collaborating Centre for Surveillance of Antimicrobial Resistance 2019) and primarily used by this package), an ATC code (as used by the WHO, WHO Collaborating Centre for Drug Statistics Methodology 2018), and a CID code (Compound ID, as used by PubChem, Kim et al. 2019). The data set contains more than 5,000 official brand names from many different countries, as found in PubChem. Other properties in this data set are derived from one or more of these codes, such as official names of pharmacological and chemical subgroups, and defined daily doses (DDD). A. Included data sets • microorganisms • antivirals • antivirals A ‘data.frame’ containing 102 antiviral agents with 9 columns. Like the antibiotics data set, it contains ATC codes (as used by the WHO, WHO Collaborating Centre for Drug Statistics Methodology 2018), and a CID code (Compound ID, as used by PubChem, Kim et al. 2019), as well as the official name and defined daily dose (DDD) for each antiviral agent. • example_isolates A ‘data.frame’ containing test results of 2,000 microbial isolates. The data set reflects real patient data and can be used to practice AMR analysis. It is structured in the typical format of laboratory information systems with one row per isolate and one column per tested antimicrobial agent (i.e., an antibiogram). • microorganisms A ‘data.frame’ containing 70,760 (sub)species with 16 columns comprising their com- plete microbial taxonomy according to the Catalogue of Life (Bánki et al. 2022). In- cluded microorganisms and their complete taxonomic tree of all included (sub)species from kingdom to subspecies with year of scientific publication and responsible author(s): – All 59,024 (sub)species from the kingdoms of Archaea, Bacteria, Chromista and Protozoa. – All 9,582 (sub)species from these orders of the kingdom of Fungi: Eurotiales, Ony- genales, Pneumocystales, Saccharomycetales, Schizosaccharomycetales and Tremel- lales. – All 2,153 (sub)species from 47 other relevant genera from the kingdom of Animalia (like Strongyloides and Taenia). – All 14,338 previously accepted names of included (sub)species that have been taxo- nomically renamed. The kingdom of Fungi is a very large taxon with almost 300,000 different (sub)species, of which most are not microbial (but rather macroscopic, such as mushrooms). Therefore, not all fungi fit the scope of the AMR package. By only including the aforementioned taxonomic orders, the most relevant fungi are covered (such as all species of Aspergillus, Candida, Cryptococcus, Histoplasma, Pneumocystis, Saccharomyces and Trichophyton). • example_isolates Journal of Statistical Software 31 • example_isolates_unclean • example_isolates_unclean A ‘data.frame’ containing test results of 3,000 microbial isolates that require cleaning up before they can be used for analysis. This data set can be used to practice AMR analysis and is featured in Section 7. p _ _ A ‘data.frame’ containing test results of 3,000 microbial isolates that require cleaning up before they can be used for analysis. This data set can be used to practice AMR analysis and is featured in Section 7. • WHONET A ‘data.frame’ containing 500 observations and 53 columns, with the exact same struc- ture as an export file from WHONET 2019 software (WHO Collaborating Centre for Surveillance of Antimicrobial Resistance 2019). Such files can be used with the AMR package, as this example data set demonstrates. The antibiotic test results are from the example_isolates data set. All patient names are created using online surname generators and are only in place for practice purposes. A ‘data.frame’ containing 500 observations and 53 columns, with the exact same struc- ture as an export file from WHONET 2019 software (WHO Collaborating Centre for Surveillance of Antimicrobial Resistance 2019). Such files can be used with the AMR package, as this example data set demonstrates. The antibiotic test results are from the example_isolates data set. All patient names are created using online surname generators and are only in place for practice purposes. Affiliation: Affiliation: Matthijs S. Berends Certe Medical Diagnostics and Advice Foundation Van Swietenlaan 2 9728 NZ Groningen, The Netherlands and University of Groningen University Medical Center Groningen Department of Medical Microbiology and Infection Prevention Hanzeplein 1 9713 GZ Groningen, The Netherlands E-mail: m.berends@certe.nl, m.s.berends@umcg.nl Christian F. Luz, Alexander W. Friedrich, Bhanu N.M. Sinha, Corinna Glasner University of Groningen University Medical Center Groningen Department of Medical Microbiology and Infection Prevention Hanzeplein 1 9713 GZ Groningen, The Netherlands Casper J. Albers University of Groningen Heymans Institute for Psychological Research Grote Kruisstraat 2/1 9712 TS Groningen, The Netherlands Journal of Statistical Software https://www.jstatsoft.org/ published by the Foundation for Open Access Statistics https://www.foastat.org/ September 2022, Volume 104, Issue 3 Submitted: 2019-12-18 doi:10.18637/jss.v104.i03 Accepted: 2022-02-05 Affiliation: Matthijs S. Berends Certe Medical Diagnostics and Advice Foundation Van Swietenlaan 2 9728 NZ Groningen, The Netherlands and University of Groningen University Medical Center Groningen Department of Medical Microbiology and Infection Prevention Hanzeplein 1 9713 GZ Groningen, The Netherlands E-mail: m.berends@certe.nl, m.s.berends@umcg.nl Christian F. Luz, Alexander W. Friedrich, Bhanu N.M. Sinha, Corinna Glasner University of Groningen University Medical Center Groningen Department of Medical Microbiology and Infection Prevention Hanzeplein 1 9713 GZ Groningen, The Netherlands Casper J. Albers University of Groningen Heymans Institute for Psychological Research Grote Kruisstraat 2/1 9712 TS Groningen, The Netherlands Journal of Statistical Software https://www.jstatsoft.org/ published by the Foundation for Open Access Statistics https://www.foastat.org/ September 2022, Volume 104, Issue 3 Submitted: 2019-12-18 doi:10.18637/jss.v104.i03 Accepted: 2022-02-05 Affiliation: Matthijs S. Berends Certe Medical Diagnostics and Advice Foundation Van Swietenlaan 2 9728 NZ Groningen, The Netherlands and University of Groningen University Medical Center Groningen Department of Medical Microbiology and Infection Prevention Hanzeplein 1 9713 GZ Groningen, The Netherlands E-mail: m.berends@certe.nl, m.s.berends@umcg.nl Christian F. Luz, Alexander W. Friedrich, Bhanu N.M. Sinha, Corinna Glasner University of Groningen University Medical Center Groningen Department of Medical Microbiology and Infection Prevention Hanzeplein 1 9713 GZ Groningen, The Netherlands Casper J. Albers University of Groningen Heymans Institute for Psychological Research Grote Kruisstraat 2/1 9712 TS Groningen, The Netherlands Journal of Statistical Software https://www.jstatsoft.org/ published by the Foundation for Open Access Statistics https://www.foastat.org/ September 2022, Volume 104, Issue 3 Submitted: 2019-12-18 doi:10.18637/jss.v104.i03 Accepted: 2022-02-05 Casper J. Albers University of Groningen Heymans Institute for Psychological Research Grote Kruisstraat 2/1 9712 TS Groningen, The Netherlands
https://openalex.org/W3131154467	https://scipost.org/10.21468/SciPostPhys.10.2.046/pdf	English	null	Casting a graph net to catch dark showers	SciPost physics	2,021	cc-by	15,814	Casting a graph net to catch dark showers Elias Bernreuther⋆, Thorben Finke†, Felix Kahlhoefer‡, Michael Krämer◦and Alexander Mück× Institute for Theoretical Particle Physics and Cosmology (TTK), RWTH Aachen University, D-52056 Aachen, Germany ⋆ebernreuther@physik.rwth-aachen.de, † ﬁnke@physik.rwth-aachen.de, ‡ kahlhoefer@physik.rwth-aachen.de, ◦mkraemer@physik.rwth-aachen.de, × mueck@physik.rwth-aachen.de Abstract Strongly interacting dark sectors predict novel LHC signatures such as semi-visible jets resulting from dark showers that contain both stable and unstable dark mesons. Distin- guishing such semi-visible jets from large QCD backgrounds is difﬁcult and constitutes an exciting challenge for jet classiﬁcation. In this article we explore the potential of su- pervised deep neural networks to identify semi-visible jets. We show that dynamic graph convolutional neural networks operating on so-called particle clouds outperform convo- lutional neural networks analysing jet images as well as other neural networks based on Lorentz vectors. We investigate how the performance depends on the properties of the dark shower and discuss training on mixed samples as a strategy to reduce model dependence. By modifying an existing mono-jet analysis we show that LHC sensitivity to dark sectors can be enhanced by more than an order of magnitude by using the dynamic graph network as a dark shower tagger. Copyright E. Bernreuther et al. This work is licensed under the Creative Commons Attribution 4.0 International License. Published by the SciPost Foundation. Received 23-06-2020 Accepted 10-02-2021 Published 23-02-2021 Check for updates doi:10.21468/SciPostPhys.10.2.046 Copyright E. Bernreuther et al. This work is licensed under the Creative Commons Attribution 4.0 International License. Published by the SciPost Foundation. Received 23-06-2020 Accepted 10-02-2021 Published 23-02-2021 Check for updates doi:10.21468/SciPostPhys.10.2.046 Received 23-06-2020 Accepted 10-02-2021 Published 23-02-2021 Check for updates doi:10.21468/SciPostPhys.10.2.046 SciPost Phys. 10, 046 (2021) 1 Introduction The huge wealth of data taken at the LHC offers a unique opportunity to explore the prop- erties of dark sectors and uncover the nature of dark matter (DM). At the same time, such an amount of data poses an unprecedented challenge to precisely determine and efﬁciently suppress backgrounds in order to identify potential signals of new physics. As the complexity of experimental analyses increases, there has been rapidly growing interest in using machine learning techniques to distinguish signal from background. For example, deep neural networks are powerful tools for the classiﬁcation of jets, which can signiﬁcantly improve the sensitivity to new physics signals hidden in QCD background, see e.g. the reviews [1, 2]. A particularly interesting and well-motivated case are so-called dark showers [3–11], which may resemble QCD jets even though they result from new interactions and contain exotic particles. Dark showers arise in extensions of the Standard Model (SM) that contain new strong dynamics, i.e. exotic fermions charged under a new gauge group that conﬁnes at low ener- gies [8–33]. If these fermions are produced at the LHC (for example via a heavy mediator), they will undergo fragmentation and hadronisation similar to SM quarks. These processes then lead to a shower of composite states that are neutral under the new gauge group, which in analogy to the SM we will refer to as dark mesons and dark baryons. The detailed spec- trum of such a dark sector is difﬁcult to predict from ﬁrst principles, but a common feature of many models is the existence of dark pions πd, which are the Pseudo-Goldstone bosons from the spontaneous breaking of chiral symmetry, and dark vector mesons ρd, which have a mass similar to the conﬁnement scale Λd. What makes dark showers exciting to study at the LHC is the fact that some of the dark mesons (e.g. the dark pions) may be stable on cosmological scales, thus providing a potential explanation for DM, while other dark mesons (e.g. the dark vector mesons) may decay on collider scales into SM particles, and in particular into SM quarks. This combination of visible and invisible particles in the same shower then leads to so-called semi-visible jets [4,5,7,9–11]. The fraction of invisible particles in a dark shower will ﬂuctuate around the expectation value rinv, leading to sizeable amounts of missing energy even if two dark showers are produced back-to-back. 1For other examples of graph networks in the context of LHC physics see e.g. refs. [40–45]. Contents Co e s 1 Introduction 1 2 Dark sector models and semi-visible jets 3 3 Deep neural networks for semi-visible jets 4 3.1 Classiﬁcation performance 6 3.2 Model dependence of semi-visible jet classiﬁcation 7 3.3 Mitigating model dependence with mixed samples 9 4 Mono-jet analysis with machine learning 10 5 Conclusions 13 A Jet and event generation 15 B Neural network architectures 16 References 21 1 1 SciPost Phys. 10, 046 (2021) 1 Introduction Moreover, there is a ﬁnite probability that a dark shower will remain completely invisible, in which case the resulting signature is a mono-jet signature with a single semi-visible jet pointing in the direction opposite to the missing energy vector. Although LHC searches for jets and missing energy are not optimised for semi-visible jets, their sensitivity can be enhanced signiﬁcantly by suppressing QCD backgrounds with improved jet classiﬁcation. Traditional jet tagging algorithms rely on hand-crafted high-level features such as N-subjettiness [34]. Basic machine learning algorithms like boosted decision trees can then learn decision boundaries along those high-level features for classiﬁcation. Deep learning algorithms on the other hand are able to work on low-level quantities, such as particle four- momenta, and extract complex features relevant for the classiﬁcation. For an overview of deep learning jet taggers and their performance on separating hadronic top jets from light QCD jets we refer to ref. [35]. In this article we explore the potential of supervised deep neural networks to identify semi- visible jets. For this purpose we consider three different architectures: a convolutional neural network (CNN) working on jet images [36], a Lorentz layer (LoLa) network [37] working on an ordered set of four-momenta of jet constituents and the dynamic graph convolutional neural network (DGCNN) [38] of ref. [39] working on an unordered set of particles, a so-called particle cloud, similar to the concept of point clouds in computer vision.1 While the different techniques show similar performance on the task of top classiﬁcation, we show that their performances differ more signiﬁcantly for the classiﬁcation of semi-visible jets. In particular 2 SciPost Phys. 10, 046 (2021) dynamic graph neural networks are shown to be powerful tools for tagging semi-visible jets. We ﬁnd that it may be possible to enhance the sensitivity of LHC searches for signatures with semi-visible jets by an order of magnitude through jet classiﬁcation with a DGCNN. This paper is structured as follows. In section 2 we discuss the properties of dark showers and in particular semi-visible jet signatures. The neural networks used in our analysis are introduced in section 3. We compare the classiﬁcation performance of the different neural network architectures and study the dependence of the jet identiﬁcation on the parameters of the dark sector model. 2 Dark sector models and semi-visible jets The structure of a semi-visible jet is mostly characterised by three parameters: The fraction of invisible particles rinv, the mass mmeson of the unstable dark mesons and the conﬁnement scale Λd. Indeed, even for rinv = 0 dark showers may differ substantially from ordinary QCD jets, because of the different running of the dark gauge coupling, the absence of heavy quarks in the shower and the presence of substructure corresponding to the decays of individual dark mesons. With increasing rinv the semi-visible jet becomes increasingly different from QCD jets, but also harder to study because of the smaller number of visible constituents. In the following we will investigate how the properties of semi-visible jets enable us to distinguish them from SM backgrounds. Unless explicitly stated otherwise, we will assume that the dark mesons have a mass close to the conﬁnement scale: mmeson ≈Λd. Furthermore, we will limit ourselves to the case of a dark SU(3) gauge group with two ﬂavours of dark quarks qd. This choice is based on a recent study on strongly interacting dark sectors [10], which identiﬁed this scenario as particularly interesting, because it allows for a viable DM candidate consistent with all cosmological and laboratory constraints. In the following we brieﬂy review the key aspects of this model, which will serve as a benchmark scenario for the present study. The two dark quarks qd are assumed to be in the fundamental representation of the dark SU(3) gauge group and carry opposite charges with respect to an additional U(1)′ gauge sym- metry. Below the conﬁnement scale the dark quarks form mesons, which we denote by π0 d,π± d and ρ0 d,ρ± d in analogy to their QCD counterparts, with the superscripts denoting the U(1)′ charges. The dark sector is therefore characterised by the meson masses mπd and mρd and the coupling strength gd of their interactions. We furthermore assume that the dark baryons are sufﬁciently heavy that they are irrelevant for phenomenology. It can be shown that in this set-up all dark pions are stable on cosmological scales and therefore constitute a potential DM candidate. The interactions of the dark sector with the SM are mediated by the massive U(1)′ gauge boson Z′ with vector couplings to both dark and SM quarks, denoted ed and gq, respectively. Couplings to leptons, as well as mixing between the Z′ and SM gauge bosons, are assumed to be suppressed. 1 Introduction To demonstrate the potential of a semi-visible jet classiﬁer, we adapt an existing mono-jet search to use a jet classiﬁer based on a dynamic graph neural network in section 4. Our conclusions are presented in section 5. In appendix A we describe the generation of signal and background events or jets. The architectures of the neural networks employed in our analysis are presented in detail in appendix B. 2 Dark sector models and semi-visible jets 10, 046 (2021) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 1 1 1 1 1 1 0 0 0 0 0 0 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 0 0 1 1 0 0 0 0 0 0 1 1 1 1 1 1 0 0 0 0 0 0 1 1 1 1 1 1 0 0 1 1 0 0 1 1 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 QCD jet Figure 1: Schematic illustration of a dark shower from the decay of a Z′ produced in association with a gluon. Figure taken from ref. [10]. Figure 1: Schematic illustration of a dark shower from the decay of a Z′ produced in association with a gluon. Figure taken from ref. [10]. 2 Dark sector models and semi-visible jets In analogy to γ-ρ0 mixing in the SM, the Z′ mixes with the ρ0 d, which induces small couplings between the ρ0 d and SM quarks and renders the ρ0 d unstable. For mρd < 2mπd the ρ± d mesons can only decay into three-body ﬁnal states via an off-shell Z′, which makes them stable with respect to collider phenomenology. We assume that each mesonic degree of freedom is produced with the same probability during the dark hadronisation process while the production of dark baryons in the shower is negligible, and that the ρ0 d mesons decay 3 SciPost Phys. 2We note that for small Z′ couplings the ρ0 d can be long-lived and lead to displaced vertices at the LHC. The corresponding production cross sections can nevertheless be sufﬁciently large that thousands of such events have already gone unnoticed at ATLAS and CMS. Ongoing detector upgrades as well as new analysis strategies make these signatures a promising target for future LHC runs. Exploring the sensitivity of searches for displaced vertices for dark sector models is subject of separate work in progress. 2 Dark sector models and semi-visible jets promptly.2 The invisible energy fraction in a dark shower is then given by rinv = 0.75, which we will use as the benchmark value in the following. Furthermore, the relevant mass for characterising the dark shower is the mass of the dark vector mesons: mmeson = mρd. We note in passing that the assumption mρd < 2mπd can be motivated from cosmology, because the relic density of dark pions is determined by the rate of the annihilation process πdπd →ρdρd, which becomes Boltzmann suppressed at low temperatures. Provided mπd and mρd are sufﬁciently close, the observed relic abundance can be reproduced even for weak portal interactions and/or heavy Z′ bosons, which makes it possible to satisfy constraints from direct detection experiments. For example, for mπd = 4 GeV and gd = 1 one requires mρd ≈5 GeV, while the Z′ mediator can be in the TeV range [10]. LHC phenomenology for this model is then dominated by the on-shell production of the Z′ (possibly in association with SM particles) and its subsequent decays into either SM or dark quarks. While the former case leads to di-jet resonances that can be easily reconstructed, the latter case gives rise to more challenging semi-visible jets, see ﬁgure 1. Although existing LHC searches for missing energy have some sensitivity to this set-up, they are not optimised for the case of dark showers, where the missing energy tends to be aligned with a visible jet. The reason is that such a conﬁguration is difﬁcult to disentangle from QCD backgrounds resulting from mis-reconstructed jets [46,47]. A detailed reinterpretation of existing exclusion limits from a search for di-jet resonances [48] and searches for missing energy [46, 49, 50] was performed in ref. [10]. It was found that for the benchmark values mqd = 500 MeV, mπd = 4 GeV, mρd = Λd = 5 GeV and mZ′ = 1TeV couplings of the order of 0.1 are still consistent with all constraints, even though the production cross section for dark showers is of the order of picobarn. In order to enhance experimental sensitivity to dark showers it is essential to improve background suppression, which potentially allows for other selection cuts to be relaxed. The most promising strategy for doing so is to develop techniques for distinguishing semi-visible jets from QCD jets and extend existing analyses by a dedicated tagger for semi-visible jets. 3 Deep neural networks for semi-visible jets Deep neural networks have been applied to a wide range of jet classiﬁcation problems, includ- ing the identiﬁcation of jet substructures. It is not clear a priori how the information encoded in a jet should be mapped to a particular data structure. The representation of jets as images is motivated by jet reconstruction with calorimeters [51–53]. Convolutional neural networks (CNNs) are a powerful tool to analyse jet images. CNNs apply convolution ﬁlters that oper- ate on small windows of an image array and allow for an efﬁcient identiﬁcation of features in the image. Convolutional networks have been very successful in the identiﬁcation of jet substructures, for example in the case of top jets, and serve as a benchmark for assessing jet classiﬁcation tools [35,36,54,55]. To illustrate the challenge of identifying semi-visible jets based on images, we show average jet images for light QCD di-jets, semi-visible jets and hadronically decaying boosted top jets in ﬁgure 2. The event generation and the preprocessing steps for the generation of the images are described in appendices A and B, respectively. The average top jet has a clearly visible substructure originating from the hadronic top decay and thus differs substantially from light QCD jets. The semi-visible jets from the dark shower, on the other hand, are very similar to QCD. Instead of an image, a jet may be represented as a collection of particle constituents. A strong top-jet identiﬁcation performance can be achieved with so-called Lorentz-layer (LoLa) networks. These architectures map constituent four-vectors to quantities more directly related to physical observables, such as invariant masses, transverse momenta or linear combinations of constituent energies [37, 56–58]. The input to the Lorentz layer typically consists of the original particle four-momenta complemented by various learned linear combinations of those. Providing learned linear combinations allows the network to identify jet substructures more efﬁciently. The constituent four-momenta together with the learned linear combinations are then transformed into invariant masses and other physical observables by the Lorentz layer, before classiﬁcation by a fully connected neural network. Dynamic graph convolutional neural networks (DGCNNs) [38] are another class of pow- erful classiﬁers which apply so-called edge convolutions to particle constituents, or particle clouds, characterised by features such as energies, transverse momenta, or angular separa- tions [39]. The edge convolution differs from a convolution over an image in the deﬁnition of the local patch that the convolution kernel observes. 2 Dark sector models and semi-visible jets In the following section we will study how to achieve this goal with a neural network trained to identify dark showers. 4 SciPost Phys. 10, 046 (2021) SciPost Phys. 10, 046 (2021 0.8 0.4 0.0 0.4 0.8 0.8 0.4 0.0 0.4 0.8 Mean QCD jet image 0.8 0.4 0.0 0.4 0.8 0.8 0.4 0.0 0.4 0.8 Mean semi-visible jet image 0.8 0.4 0.0 0.4 0.8 0.8 0.4 0.0 0.4 0.8 Mean top jet image Figure 2: Average jet images in the plane of pseudo-rapidity η and azimuthal angle φ for light QCD (left), semi-visible dark jets (middle), and boosted top jets (right). See appendices A and B for details on the event generation and the preprocessing steps for the generation of the images. 0.8 0.4 0.0 0.4 0.8 0.8 0.4 0.0 0.4 0.8 Mean semi-visible jet image 0.8 0.4 0.0 0.4 0.8 0.8 0.4 0.0 0.4 0.8 Mean QCD jet image 0.8 0.4 0.0 0.4 0.8 0.8 0.4 0.0 0.4 0.8 Mean top jet image Figure 2: Average jet images in the plane of pseudo-rapidity η and azimuthal angle φ for light QCD (left), semi-visible dark jets (middle), and boosted top jets (right). See appendices A and B for details on the event generation and the preprocessing steps for the generation of the images. 3.1 Classiﬁcation performance Our neural networks are trained on 200k background and 200k signal jets, with a validation split of 10%. For the signal generation we use the benchmark parameters introduced above, i.e. rinv = 0.75, mmeson = 5GeV and mZ′ = 1TeV. The network performance results presented below are based on an independent test set of 100k background and 100k signal jets. Note that for the moment we use the same dark sector parameters for training and testing, even though these parameters would be unknown in a realistic setting. We will return to the issue of model dependence in section 3.2 and present a mitigation strategy in section 3.3. The networks output two numbers for each jet, which can be interpreted as the prob- abilities to belong to the background or the signal class, respectively. Deﬁning a threshold probability necessary for a jet to be labelled as signal and scanning over this threshold, one obtains the receiver operating characteristic (ROC) curve, i.e. the inverse of the fraction of background jets passing the threshold (the background rejection 1/εB) as a function of the fraction of signal events passing the threshold (the signal efﬁciency εS). Figure 3 shows the ROC curve for semi-visible jet identiﬁcation (left panel) and for top-jet identiﬁcation (right panel). To estimate the stability and reproducibility of the network performances, ﬁve net- works with independent random weight initialisations are trained on the same training data and tested on the same testing data. The small spread in performance indicated by the shaded band around the ROC curves in ﬁgure 3 shows that the training convergence of the networks is stable. For a comparison of the top-tagging performance of our networks with the results of ref. [35], see ﬁgure 10 in appendix B. Various network performance measures are collected in table 1. We display the accuracy, i.e. the ratio of the number of correctly classiﬁed jets over the total number of jets, the area under the ROC curve, AUC = R dεB εS(εB), and the background rejection at a signal efﬁciency of 30%. Error estimates correspond to the spread obtained from the ﬁve independent network trainings mentioned above. The results presented in ﬁgure 3 and table 1 ﬁrst of all conﬁrm that the classiﬁcation of semi-visible jets is more challenging than that of top jets. 3 Deep neural networks for semi-visible jets In an image, the local patch corresponds to some neighbourhood of pixels. For an edge convolution, a local graph is constructed for 5 SciPost Phys. 10, 046 (2021) each point in the particle cloud from its nearest neighbours using a distance measure in the space of input features. Calculating new nearest neighbours dynamically from the output of the previous edge convolution allows for particles that are initially far apart to become close in feature space already for the next convolution. In this way, long range correlations can be accessed efﬁciently with few convolutional layers and the network is potentially able to learn the graph structure that offers most information. For a cloud of particles representing a jet, it appears natural that the correlation of particles which are not close in the initial features, can be important for the classiﬁcation of the jet. The dynamic update enables the network to link those initially distant particles. In the following, we will analyse and compare the classiﬁcation performance of a CNN, a LoLa network and a DGCNN for semi-visible jets against light QCD background jets. For com- parison, we also show results for the well-established benchmark task of top-jet identiﬁcation. The architectures of the CNN, the LoLa network and the DGCNN are described in detail in appendix B. 3.2 Model dependence of semi-visible jet classiﬁcation In this section we explore the model dependence of the semi-visible jet classiﬁcation with the DGCNN, i.e. we study how the performance changes as we vary the parameters of the strongly interacting dark sector. This not only sheds light on how much a speciﬁc network generalises to other dark shower scenarios, but it also provides some indication of which signal features the network may learn. As a crucial parameter, the invisible fraction rinv represents the average percentage of miss- ing energy and characterises the composition of the dark showers. The model described in section 2 predicts rinv = 0.75. However, for the purpose of this section we treat rinv as a phe- nomenological parameter which can assume any value between zero and one. To this end, we decay all the dark mesons in PYTHIA with branching ratio rinv into invisible particles and branching ratio 1 −rinv into Standard Model quarks, respectively. Training and testing the DGCNN classiﬁer architecture on dark shower samples with different values of rinv we obtain the ROC curves shown in the left panel of ﬁgure 4. We ﬁnd that dark showers with larger rinv are in general easier to distinguish from QCD. For 0.1 < εS < 0.3, which is the most inter- esting range for improving an analysis with the jet tagger, the background suppression varies by roughly an order of magnitude. Note that for very small rinv the classiﬁcation performance increases again as almost all the energy from the Z′ resonance ends up in visible jets leading to a harder jet pT distribution which is more different from QCD. As another dark sector parameter we vary the dark meson mass mmeson = mπd = mρd, together with the dark conﬁnement scale Λd = mmeson. Note that the small mass splitting between the πd and ρd motivated by cosmology has no impact on the LHC phenomenology. Larger values of Λd lead to a stronger running of the dark sector coupling αd at the energy scale of the semi-visible jet. Among other effects, the running of αd changes jet substructure observables such as the distribution of the two-point energy correlation function discussed in ref. [11]. Moreover, as the jet constituents arise from dark meson decays they encode the dark meson mass scale mmeson. 3.1 Classiﬁcation performance Comparing the CNN, LoLa and DGCNN architectures, we ﬁnd that the DGCNN performs best for both top and semi-visible jet identiﬁcation. While the difference between the CNN, the LoLa network and the DGCNN is moderate for top identiﬁcation, the strength of the DGCNN is particularly signiﬁcant for the classiﬁcation of semi-visible jets. As shown in ﬁgure 3, the background rejection at a given signal efﬁciency, which is most relevant for an experimental analysis, is signiﬁcantly improved by a DGCNN for a wide range of signal efﬁciencies. Speciﬁcally, at a signal efﬁciency of 30%, the background rejection of the DGCNN is almost a factor of ﬁve stronger than that of the 6 SciPost Phys. 10, 046 (2021 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 S 100 101 102 103 104 1/ B Semi-visible jets, 150 GeV < pTj < 350 GeV DGCNN CNN LoLa 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 S 100 101 102 103 104 1/ B Top jets, 550 GeV < pTj < 650 GeV DGCNN CNN LoLa Figure 3: Comparison of the ROC curves in background rejection 1/εB and signal efﬁciency εS for semi-visible jet identiﬁcation (left panel) and for boosted top jet identiﬁcation (right panel) as obtained by the CNN, LoLa and DGCNN architectures, respectively. The error bands correspond to the spread obtained from ﬁve indepen- dent initialisations of the network. SciPost Phys. 10, 046 (2021) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 S 100 101 102 103 104 1/ B Top jets, 550 GeV < pTj < 650 GeV DGCNN CNN LoLa 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 S 100 101 102 103 104 1/ B Semi-visible jets, 150 GeV < pTj < 350 GeV DGCNN CNN LoLa Figure 3: Comparison of the ROC curves in background rejection 1/εB and signal efﬁciency εS for semi-visible jet identiﬁcation (left panel) and for boosted top jet identiﬁcation (right panel) as obtained by the CNN, LoLa and DGCNN architectures, respectively. The error bands correspond to the spread obtained from ﬁve indepen- dent initialisations of the network. 3.2 Model dependence of semi-visible jet classiﬁcation As shown in the right panel of ﬁgure 4, changing the conﬁnement and meson mass scale between 5 GeV and 20 GeV has no signiﬁcant effect on the classiﬁcation performance. 7 SciPost Phys. 10, 046 (2021) Table 1: Performance measures for classifying semi-visible jets and top jets by the CNN, LoLa and DGCNN architectures, respectively, corresponding to the ROC curves presented in ﬁgure 3. We show the accuracy, the AUC value and the background rejection at a signal efﬁciency of 30%. The central value is the mean of the ﬁve independent training runs of the network, while the error estimate corresponds to the spread in the performance. Acc [%] AUC 1/εB (εS = 0.3) semi-visible jets CNN 79.88+0.21 −0.22 0.8790+0.0019 −0.0019 137+6 −4 LoLa 83.26+0.14 −0.13 0.9118+0.0008 −0.0010 220+11 −17 DGCNN 85.04+0.12 −0.08 0.9258+0.0007 −0.0007 608+36 −40 top jets CNN 92.98+0.05 −0.09 0.9802+0.0002 −0.0005 785+40 −29 LoLa 92.83+0.11 −0.11 0.9791+0.0007 −0.0008 540+77 −53 DGCNN 93.47+0.06 −0.06 0.9831+0.0001 −0.0002 1073+47 −75 Since the values of the dark sector parameters are not known a priori, it is a relevant question to which extent the classiﬁcation is model-dependent. Therefore, we next consider samples with different dark sector parameters for training and testing. We train the network on dark showers with our benchmark parameters rinv = 0.75 and mmeson = 5 GeV, and evaluate the performance on a range of samples with different choices for rinv and mmeson, respectively. The corresponding ROC curves are displayed in ﬁgure 5. ROC curves for training and testing with identical parameter values (see ﬁgure 4) are shown for comparison. As expected, we ﬁnd a drop in performance as the difference between the model parameter settings in the training and test samples increases. Varying rinv, the decrease in performance is modest. Only for rinv = 0.1, the drop is larger since the tagger cannot beneﬁt any more from the harder jet pT distribution. The model dependence is signiﬁcantly more pronounced for the dark meson mass. The background suppression is reduced by about an order of magnitude for 0.1 < εS < 0.3, indicating that the network learns to reconstruct the dark meson mass to some extent from the constituents. 3We note that the DGCNN signiﬁcantly outperforms both the CNN and the LoLa architecture also for trainin on mixed samples. 3.3 Mitigating model dependence with mixed samples 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 S 100 101 102 103 104 1/ B mZ′ = 1 TeV, 150 GeV < pTj < 300 GeV rinv = 0.9 rinv = 0.75 rinv = 0.5 rinv = 0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 S 100 101 102 103 104 1/ B mZ′ = 1 TeV, rinv = 0.75, 150 GeV < pTj < 300 GeV mmeson = 5 GeV mmeson = 10 GeV mmeson = 15 GeV mmeson = 20 GeV Figure 4: DGCNN ROC curves for the discrimination of dark showers from QCD for different values of rinv (left panel) and mmeson (right panel). The error bands cor- respond to the spread obtained from ﬁve independent initialisations of the network. 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 S 100 101 102 103 104 1/ B mZ′ = 1 TeV, 150 GeV < pTj < 300 GeV rinv = 0.9 rinv = 0.75 rinv = 0.5 rinv = 0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 S 100 101 102 103 104 1/ B mZ′ = 1 TeV, rinv = 0.75, 150 GeV < pTj < 300 GeV mmeson = 5 GeV mmeson = 10 GeV mmeson = 15 GeV mmeson = 20 GeV Figure 5: ROC curves (dotted lines) for the DGCNN trained on dark showers with the benchmark values (rinv = 0.75 and mmeson = Λd = 5 GeV) and tested on different values of rinv (left panel) and mmeson (right panel). ROC curves for training and testing on samples with identical parameters are shown for comparison (solid lines, as in ﬁgure 4). 3.3 Mitigating model dependence with mixed samples A simple way to mitigate this behaviour and provide a more model-independent semi-visible jet classiﬁer, is to train the network on a mixed sample, which contains a range of different rinv values or dark meson masses. Here we consider a mixed rinv sample containing an equal num- ber of jets with rinv = 0.1, 0.5, 0.75, and 0.9, as well as a mixed meson mass sample consisting of an equal number of jets with mmeson = 5 GeV, 10 GeV, and 20 GeV. This way the network is forced to learn features common to the different samples instead of learning to reconstruct, for example, one speciﬁc dark meson mass. The performance of such a more general classiﬁer is signiﬁcantly better than that of a classiﬁer trained on speciﬁc values of rinv and mmeson when both are applied to a wider range of model parameters, see ﬁgure 6 and table 2.3 A signiﬁ- cant improvement is also present for dark meson masses that were not included in the mixed training sample, as the results for mmeson = 15 GeV show. Note that it may be possible to use 8 SciPost Phys. 10, 046 (2021) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 S 100 101 102 103 104 1/ B mZ′ = 1 TeV, 150 GeV < pTj < 300 GeV rinv = 0.9 rinv = 0.75 rinv = 0.5 rinv = 0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 S 100 101 102 103 104 1/ B mZ′ = 1 TeV, rinv = 0.75, 150 GeV < pTj < 300 GeV mmeson = 5 GeV mmeson = 10 GeV mmeson = 15 GeV mmeson = 20 GeV Figure 4: DGCNN ROC curves for the discrimination of dark showers from QCD for different values of rinv (left panel) and mmeson (right panel). The error bands cor- respond to the spread obtained from ﬁve independent initialisations of the network. 3.3 Mitigating model dependence with mixed samples 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 S 100 101 102 103 104 1/ B mZ′ = 1 TeV, 150 GeV < pTj < 300 GeV rinv = 0.9 rinv = 0.75 rinv = 0.5 rinv = 0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 S 100 101 102 103 104 1/ B mZ′ = 1 TeV, rinv = 0.75, 150 GeV < pTj < 300 GeV mmeson = 5 GeV mmeson = 10 GeV mmeson = 15 GeV mmeson = 20 GeV Figure 5: ROC curves (dotted lines) for the DGCNN trained on dark showers with the benchmark values (rinv = 0.75 and mmeson = Λd = 5 GeV) and tested on different values of rinv (left panel) and mmeson (right panel). ROC curves for training and testing on samples with identical parameters are shown for comparison (solid lines, as in ﬁgure 4). networks that each have learned a speciﬁc dark meson mass to reconstruct this mass from a possible dark shower signal, e.g. with a parametrised network [59]. We have also studied the dependence of the network training and performance on the Z′ mediator mass. We ﬁnd only small differences in the ROC curves when varying the Z′ mass between 1 TeV and 2 TeV. Moreover, training the network on a Z′ mass different from the mass used in the test sample only has a small effect on the network performance. 4 Mono-jet analysis with machine learning In this section, we study the sensitivity improvement for a dark shower search with the help of a DGCNN classiﬁer. As an example we consider the ATLAS mono-jet analysis with 36.1 fb−1 [46] applied to a dark shower signal with the benchmark parameters from ref. [10], i.e. rinv = 0.75, mπd = 4 GeV and mρd = Λd = 5 GeV. The mono-jet search is sensitive to dark shower events 9 SciPost Phys. 10, 046 (2021) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 S 100 101 102 103 104 1/ B mZ′ = 1 TeV, 150 GeV < pTj < 300 GeV rinv = 0.9 rinv = 0.75 rinv = 0.5 rinv = 0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 S 100 101 102 103 104 1/ B mZ′ = 1 TeV, rinv = 0.75, 150 GeV < pTj < 300 GeV mmeson = 5 GeV mmeson = 10 GeV mmeson = 15 GeV mmeson = 20 GeV Figure 6: ROC curves (dashed lines) for the DGCNN trained on mixed samples of dark showers with different values of rinv and mmeson, and tested on pure samples each containing a speciﬁc value for rinv (left panel) and mmeson (right panel). ROC curves for training and testing on samples with identical parameters are shown for comparison (solid lines, as in ﬁgure 4). 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 S 100 101 102 103 104 1/ B mZ′ = 1 TeV, 150 GeV < pTj < 300 GeV rinv = 0.9 rinv = 0.75 rinv = 0.5 rinv = 0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 S 100 101 102 103 104 1/ B mZ′ = 1 TeV, rinv = 0.75, 150 GeV < pTj < 300 GeV mmeson = 5 GeV mmeson = 10 GeV mmeson = 15 GeV mmeson = 20 GeV Figure 6: ROC curves (dashed lines) for the DGCNN trained on mixed samples of dark showers with different values of rinv and mmeson, and tested on pure samples each containing a speciﬁc value for rinv (left panel) and mmeson (right panel). ROC curves for training and testing on samples with identical parameters are shown for comparison (solid lines, as in ﬁgure 4). where one of the dark showers stays invisible leading to a mono-jet topology, i.e. 4 Mono-jet analysis with machine learning a large angular distance ∆φ ≈π between missing energy and the semi-visible jet.4 To identify these jets as originating from a dark shower, we integrate our graph network into the analysis as a dark shower tagger. To generate signal and background events, we use the tools described in appendix A. Note that we focus on the dominant Z+jets background. The signal jets for training the tagger are extracted from a dark shower signal with mZ′ = 1 TeV. The sample of a given signal region consists of all fat jets from the corresponding signal events, where we require a truth-level dark quark within the jet cone. We emphasise that we apply two different jet deﬁnitions to each event. While the signal events are deﬁned using the ATLAS jet deﬁnition of ref. [46], the fat jets for the tagger training are anti-kT jets with a minimal transverse momentum of 100 GeV and a jet radius R = 0.8 in order to contain all radiation from an underlying dark quark. The background jet samples consist of all fat jets from the corresponding Z+jets events. We train on 200k signal and 200k background jets. In the analysis we ﬁrst apply the cuts from ref. [46]. We then sort the remaining events into signal regions and apply the DGCNN dark shower tagger, trained on the appropriate signal region, to all fat jets in each event. If at least one of the jets in an event is tagged as a dark shower jet the event is accepted. Otherwise the event is rejected. By varying the tagging threshold we control the signal event efﬁciency and Z+jets background rejection rate. The corresponding ROC curve is shown in the left panel of ﬁgure 7 for the signal region EM4, which corresponds to 400 GeV< /ET < 500 GeV. EM4 is the signal region most sensitive to the dark shower signal with our benchmark parameters. The efﬁciencies εS and εB shown in ﬁgure 7 are relative to the event numbers after the ordinary mono-jet cuts. Hence, the existing analysis without a dark shower tagger is equivalent to the point εS = εB = 1 in the lower right-hand corner of the plot. To estimate the inﬂuence of detector effects on the DGCNN tagger, we also show the anal- ogous ROC curve for a tagger based on detector level quantities, i.e. 4Other event topologies, where both dark showers are visible and recoil against an ISR jet to obtain a sizeable ∆φ, have been shown to be sub-leading in ref. [10]. 4 Mono-jet analysis with machine learning towers and tracks from DELPHES [60], instead of particles as input in the training and in the analysis. We ﬁnd that detector effects lead to a slightly reduced background rejection compared to the case with 10 SciPost Phys. 10, 046 (2021) Table 2: Performance measures for the DGCNN for different dark meson masses. We show the performance for networks trained and tested on the same mmeson, trained on mmeson = 5 GeV, and trained on a mixed sample, corresponding to the ROC curves in the right panels of Figs. 4, 5, and 6. We show the accuracy, the AUC value and the background rejection at a signal efﬁciency of 30%. The central value is the mean of ﬁve independent training runs, while the error estimate corresponds to the spread in the performance. mmeson [GeV] test mmeson [GeV] training Acc [%] AUC 1/εB (εS = 0.3) 5 5 85.14+0.04 −0.06 0.9267+0.0002 −0.0005 589+47 −46 mixed 83.61+0.09 −0.09 0.9148+0.0009 −0.0008 224+20 −15 10 10 86.04+0.05 −0.05 0.9333+0.0004 −0.0004 774+67 −59 5 81.2+0.3 −0.2 0.8965+0.0015 −0.0012 106+12 −6 mixed 84.03+0.05 −0.03 0.9180+0.0004 −0.0003 304+6 −7 15 15 86.24+0.03 −0.03 0.9336+0.0002 −0.0002 720+43 −53 5 81.00+0.17 −0.18 0.8950+0.0005 −0.0012 91+6 −3 mixed 84.38+0.11 −0.12 0.9198+0.0007 −0.0007 330+16 −15 20 20 86.03+0.09 −0.06 0.9316+0.0006 −0.0004 682+43 −33 5 79.2+0.2 −0.3 0.883+0.001 −0.002 65+2 −2 mixed 83.96+0.08 −0.08 0.9161+0.0011 −0.0009 270+15 −16 particles as DGCNN input. particles as DGCNN input. particles as DGCNN input. Using the improved background suppression due to the DGCNN tagger in the mono-jet search, we derive an expected limit on the dark shower cross section. The background event numbers B and systematic uncertainties ∆B from the ATLAS analysis [46] are divided by the background rejection obtained from our simulation of the dominant Z+jets background. We apply the same additional rejection rate for the sub-leading background of W+jets events. Fur- thermore we assume that contributions from other backgrounds, in particular from di-bosons as well as t¯t and single tops, are still negligible in the analysis with the tagger. This assump- tion is based on the fact that dark showers are easier to distinguish from top jets than from QCD jets, which should result in a tagger rejection rate of the top background that is at least comparable to the rate for the V+jets background. 4 Mono-jet analysis with machine learning The improvement in the expected limit σ95 exp achieved by the dark shower tagger is shown in the right panel of ﬁgure 7 for the signal region EM4. Note that the cross section limit also takes the additional signal rejection caused by the tagger into account. The event numbers and the expected limit for the optimal tagging threshold (corresponding to εS = 0.13) are compiled in table 3. We can then translate σ95 exp into a limit on the dark sector model coupling gq, i.e. the coupling between the Z′ mediator and the SM quarks. We use that σ(pp →qd¯qd) ∝g2 q BR(Z′ →qd¯qd), which holds as long as the Z′ resonance is sufﬁciently narrow. For each Z′ mass we determine the limit based on the signal region that is most sen- sitive without a dark shower tagger. In the mass range considered here these are EM4 for 1 TeV ≲mZ′ ≲1.3 TeV and EM2 for smaller mZ′. Exploring larger Z′ masses would be com- putationally expensive, as it would require training the network on different signal regions. The expected limit is shown together with the existing LHC limits from ref. [10] in ﬁgure 8. We conclude that the use of a DGCNN classiﬁer for semi-visible jets from dark showers has the potential to signiﬁcantly improve the sensitivity of the search. For our benchmark model, in particular, the DGCNN tagger may allow to probe dark sector model couplings in a region of parameter space where conventional searches without neural network classiﬁers fail and where searches for displaced vertices are not yet sensitive. We emphasise that the analysis presented above is based on the selection cuts of an exist- ing mono-jet analysis, which is not optimised for dark showers. Hence, even greater improve- ments in sensitivity can be expected when combining the background reduction achieved by the tagger for semi-visible jets with relaxed cuts on the overall event topology. Of particular importance in the context of semi-visible jets is the cut on the separation angle ∆φ in the azimuthal plane between the missing energy vector and any of the leading jets. In events where two semi-visible jets are produced back-to-back one typically obtains ∆φ ≈0, unless one of the jets remains fully invisible. However, conventional mono-jet analyses require ∆φ to be sufﬁciently large to suppress backgrounds from mismeasured jets. 4 Mono-jet analysis with machine learning Moreover these backgrounds would still have little bearing on the ﬁnal limits even if the rejection were signiﬁcantly worse than for V+jets. Hence, we apply the same universal rejection factor to all background contributions and their respective systematic uncertainties. We derive the expected 95% C.L. limit on the number of signal events assuming that the number of observed events is equal to the back- ground prediction. Hence, we construct the proﬁle likelihood [61] L(µ) = 1 B! µS + B 1 + ∆B B θB B e−(µS+B(1+ ∆B B θB)) e−θ 2 B/2 , (1) (1) with the value of the nuisance parameter θB chosen such that it maximises the likelihood for a given signal strength µ. We obtain the limit by excluding points for which the log-likelihood 11 SciPost Phys. 10, 046 (2021) qµ = −2(logL(µ = 1) −logL(µ = 0)) (2) (2) is larger than 3.84, which corresponds to a p-value of 0.05 for a χ2 distribution with 1 degree of freedom. Thus we arrive at a limit on the number of signal events for a given signal region. If backgrounds can be estimated directly from data in the same way as for existing monojet searches (e.g. using Z(→µµ) + jets as a control region for Z(→νν) + jets), using a dark shower tagger should not signiﬁcantly increase the relative systematic uncertainties of the background estimates (apart from decreasing the number of events in the control region, hence increasing the corresponding statistical uncertainty). As indicated by the background numbers and uncertainties in table 3, applying the dark shower tagger then takes the search from a systematics dominated to a strongly statistics dominated regime. In other words, the number of signal events that the search is sensitive to depends dominantly on the number of expected background events rather than on the uncertainty of the background estimate. The translation of the number of signal events into a dark shower production cross sec- tion is potentially subject to large systematic uncertainties, which are difﬁcult to estimate from Monte Carlo simulations alone. Here we estimate the expected limit on the dark shower production cross section σ95 exp using the nominal performance of the dark shower tagger, em- phasizing that the tagger may show a different performance when applied to real data (see discussion below). 4 Mono-jet analysis with machine learning If such mismeasured jets can be reliably distinguished from semi-visible jets using deep neural networks, the cut on ∆φ could be relaxed, which would signiﬁcantly enhance the acceptance for semi-visible jets. Accurate simulations of this particular background are however very challenging, and we leave a study of the potential sensitivity of such a search to future work. 12 SciPost Phys. 10, 046 (2021) 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 S 100 101 102 103 1/ B mZ′ = 1 TeV, rinv = 0.75, signal region EM4 particle level detector level 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 S 5 10 15 20 25 95 exp, w/oNN / 95 exp, w/NN Improvement expected limit EM4 particle level detector level Figure 7: Left: Event-level ROC curves for a mono-jet search including a DGCNN dark shower tagger. Shown are the additional background event rejection and signal event efﬁciency relative to the search without a tagger. Right: Corresponding im- provement of the expected limit on the dark shower production cross section. The jet constituents used as input for the DGCNN are either at the particle (solid lines) or detector level (dash-dotted lines). 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 S 5 10 15 20 25 95 exp, w/oNN / 95 exp, w/NN Improvement expected limit EM4 particle level detector level 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 S 100 101 102 103 1/ B mZ′ = 1 TeV, rinv = 0.75, signal region EM4 particle level detector level Figure 7: Left: Event-level ROC curves for a mono-jet search including a DGCNN dark shower tagger. Shown are the additional background event rejection and signal event efﬁciency relative to the search without a tagger. Right: Corresponding im- provement of the expected limit on the dark shower production cross section. The jet constituents used as input for the DGCNN are either at the particle (solid lines) or detector level (dash-dotted lines). Figure 7: Left: Event-level ROC curves for a mono-jet search including a DGCNN g j g dark shower tagger. Shown are the additional background event rejection and signal event efﬁciency relative to the search without a tagger. Right: Corresponding im- provement of the expected limit on the dark shower production cross section. The jet constituents used as input for the DGCNN are either at the particle (solid lines) or detector level (dash-dotted lines). 5 Conclusions 10, 046 (2021) Table 3: Expected number of background events with systematic errors and expected limit on the number of signal events in the signal region EM4 at an integrated lumi- nosity of 36.1 fb−1. Listed are the event numbers for the mono-jet search without a dark shower tagger, with a DGCNN tagger operating on particles, and with a DGCNN tagger operating on detector level objects. For each case we also state the correspond- ing improvement of the expected limit on the dark shower production cross section. The DGCNN was trained and tested on a dark shower signal with our benchmark parameters (see main text). B S95 exp (σ95 exp)w/oNN/σ95 exp without DGCNN tagger 27640 ± 610 1239 1 with DGCNN tagger (particle level) 12.1 ± 0.3 8.2 19.7 with DGCNN tagger (detector level) 27.7 ± 0.6 11.7 13.8 However, the values of the dark sector parameters are not known a priori, and we ﬁnd that the performance of the network signiﬁcantly deteriorates when different parameters are used for training and testing. To mitigate this model dependence, we have trained the network on a mixed sample, which contains semi-visible jets with varying properties. This approach yields a more general classiﬁer, which performs signiﬁcantly better when applied to a range of model parameters. However, the values of the dark sector parameters are not known a priori, and we ﬁnd that the performance of the network signiﬁcantly deteriorates when different parameters are used for training and testing. To mitigate this model dependence, we have trained the network on a mixed sample, which contains semi-visible jets with varying properties. This approach yields a more general classiﬁer, which performs signiﬁcantly better when applied to a range of model parameters. Finally, we have shown how the sensitivity of the LHC to dark showers can be substan- tially enhanced by applying a jet classiﬁer based on a dynamic graph neural network. For this purpose we have considered an existing ATLAS search for mono-jets, which is sensitive to events with one fully invisible dark shower and one semi-visible jet. We have then estimated the sensitivity that can be achieved by integrating our graph network into the analysis as a dark shower tagger. For our benchmark scenario we ﬁnd an improvement in the sensitivity of more than an order of magnitude, leading to a signiﬁcantly improved expected limit on the couplings of the model. 5 Conclusions Dark sectors with new strong dynamics may reveal themselves at the LHC in the form of dark showers resulting from the fragmentation and hadronisation of dark quarks. If some of the dark mesons in the shower are stable on cosmological scales (potentially explaining dark mat- ter), while other dark mesons decay on collider scales, such dark showers lead to semi-visible jets. Although semi-visible jets depend on a number of dark sector parameters, such as the fraction of invisible particles and the mass of the dark mesons, in practice they often resemble QCD jets and are challenging to distinguish from backgrounds. Novel jet classiﬁcation tech- niques are hence essential to enhance the sensitivity of the LHC to strongly interacting dark sectors. In this paper we have explored the potential of supervised deep neural networks to identify semi-visible jets. As speciﬁc benchmark we have considered a scenario with GeV-scale dark mesons produced via a heavy vector mediator with mass in the TeV range; such a scenario is motivated by cosmological and astrophysical considerations and at the same time leads to a sizeable cross section for events with semi-visible jets at the LHC. We have compared three different types of neural network architectures: a convolutional neural network working on jet images, a Lorentz layer network based on an ordered set of four-momenta of jet constituents, and a dynamic graph convolutional neural network operating on particle clouds, i.e. an un- ordered set of jet constituents. While these three different neural network techniques deliver comparable results for the classiﬁcation of top jets, we ﬁnd that their performance differs no- tably in the more challenging classiﬁcation of semi-visible jets. In particular, by dynamically updating the relation between jet constituents the graph neural network is able to learn more abstract features of a jet and outperforms the image-based convolutional and the Lorentz layer networks that we have considered. We have then studied how the performance of the dynamic graph network changes as we vary the parameters of the strongly interacting dark sector and the corresponding semi-visible jets. As long as the same parameters are used for training and testing, the dark meson and mediator masses have no strong effect on the classiﬁcation performance, while semi-visible jets with a larger fraction of invisible particles are in general easier to distinguish from QCD. 13 SciPost Phys. Acknowledgements We thank Martin Erdmann, Silvia Manconi, Tilman Plehn, Jennifer Thompson and Patrick Tunney for discussions. This work is funded by the Deutsche Forschungsgemeinschaft (DFG) through the Collaborative Research Center TRR 257 “Particle Physics Phenomenology after the Higgs Discovery” and the Emmy Noether Grant No. KA 4662/1-1. Simulations were performed with computing resources granted by RWTH Aachen University under project thes0678. 5 Conclusions The couplings of the Z′ mediator to dark quarks and SM quarks are denoted by ed and gq, respectively. Other LHC limits are taken from ref. [10]. Figure 8: Expected limit on the benchmark model considered in this work from a mono-jet search [46] including a dynamic graph convolutional neural network dark shower tagger (labelled “monojet DGCNN"). The couplings of the Z′ mediator to dark quarks and SM quarks are denoted by ed and gq, respectively. Other LHC limits are taken from ref. [10]. information and structure than the QCD background jets, an autoencoder trained to recon- struct QCD may also be able to reconstruct semi-visible jets and thus may not detect semi- visible jets as an anomaly. Adapting the autoencoder approach for the detection of simple jet structures, and exploring alternative unsupervised and semi-supervised deep learning tech- niques [72,77–80] for the identiﬁcation of dark shower events, is left for future work. 5 Conclusions The background reduction from tagging semi-visible jets may allow to relax cuts on the overall event topology and thereby further improve the sensitivity. Various directions for future research on detecting dark showers with deep learning meth- ods should be pursued. While we could not identify a particular observable that would control the classiﬁcation performance of the neural network, more work is needed to explore what the network actually learns and how the choice of input features may further enhance the network performance. We refer to refs. [62–68] for examples of deep learning architectures that incorporate speciﬁc physics features to guide event classiﬁcation. In supervised learning, we rely on Monte Carlo events for the training, and it is crucial to avoid that the classiﬁcation performance is biased by Monte Carlo artefacts. One should thus try to incorporate systematic uncertainties that account for the approximate modelling of a semi-visible jets, see e.g. [69–72], or further improve the Monte Carlo predictions for observables that drive the event classiﬁcation of such subtle signatures. Likewise it will be important to asses the effect of pile-up removal on the performance of the neural network. We leave this question to future work. Last but not least, one would like to use semi-supervised or unsupervised learning meth- ods for the identiﬁcation of dark shower events. For example, unsupervised machine learning algorithms based on autoencoders have successfully been used to search for anomalous jet substructure, see e.g. [73–76]. However, we ﬁnd that it is not straightforward to apply this technique to the detection of semi-visible jets. Since the semi-visible jets often contain less 14 SciPost Phys. 10, 046 (2021) 700 1000 2000 mZ′ [GeV] 0.00 0.05 0.10 0.15 0.20 0.25 0.30 gq di-jet monojet SUSY monojet DGCNN ed = 0.4 Figure 8: Expected limit on the benchmark model considered in this work from a mono-jet search [46] including a dynamic graph convolutional neural network dark shower tagger (labelled “monojet DGCNN"). The couplings of the Z′ mediator to dark quarks and SM quarks are denoted by ed and gq, respectively. Other LHC limits are taken from ref. [10]. 700 1000 2000 mZ′ [GeV] 0.00 0.05 0.10 0.15 0.20 0.25 0.30 gq di-jet monojet SUSY monojet DGCNN ed = 0.4 Figure 8: Expected limit on the benchmark model considered in this work from a mono-jet search [46] including a dynamic graph convolutional neural network dark shower tagger (labelled “monojet DGCNN"). A Jet and event generation In this appendix, we ﬁrst describe the generation of the signal and background jets used to train and test the networks in section 3. We then give details on the event generation for the search in section 4. To simulate the dark shower jets, we generate leading-order parton-level events for the dark quark production process pp → qdqd at a collider energy of 14 TeV with MADGRAPH5_AMC@NLO [81] using the NN23LO1 PDF set [82] and a UFO ﬁle for the model introduced in section 2 implemented with FEYNRULES [83]. Renormalisation and factorisa- tion scales are set to the default dynamic scale choice of MADGRAPH5_AMC@NLO. The sam- ples are MLM-matched [84] with up to one additional hard jet, setting the matching scale to xqcut = 100 GeV. Shower and hadronisation are performed with PYTHIA 8 [85]. We employ PYTHIA’s Hidden Valley module [14], which is adapted to the simulation of the dark shower 15 SciPost Phys. 10, 046 (2021) and of dark meson production as detailed in [10]. The running of the dark coupling αd is determined by the conﬁnement scale Λd. If not explicitly stated otherwise we use the default parameters mZ′ = 1 TeV, mπ = mρ = Λd = 5 GeV. The couplings of the Z′ mediator to dark quarks and SM quarks are set to ed = 0.4 and gq = 0.1, respectively. These couplings enter the Z′ width, but their value is not relevant for the production of the training and test sets of the dark shower jets. The light QCD background jets are obtained from di-jet events generated at leading or- der at a collider energy of 14 TeV with MADGRAPH5_AMC@NLO using the NN23LO1 PDF set. Renormalisation and factorisation scales are set to the default dynamic scale choice of MAD- GRAPH5_AMC@NLO. Shower and hadronisation are performed with PYTHIA 8. In both the dark shower signal and the background event samples, we employ the Fast- Jet [86] implementation of DELPHES 3 [60] to cluster fat jets using the anti-kT algorithm [87] with jet radius R = 0.8. No detector simulation is performed unless explicitly stated. When a detector simulation is included, it is performed with DELPHES 3 using the ATLAS detector card. A Jet and event generation To select jets originating from dark showers for the signal sample, we additionally require ∆R < 0.8 for the angular distance ∆R = p ∆η2 + ∆φ2 between the jet axis and a truth-level dark quark. Otherwise jets from QCD initial state radiation would enter the signal samples of semi-visible jets. For the network comparisons in section 3 we use fat jets within the trans- verse momentum range pT,jet = 150...350 GeV. To simulate samples with larger transverse momenta would be computationally more expensive since there is no generator cut which can be used to signiﬁcantly enhance event generation in the high-pT tail. For the classiﬁcation of top-quark jets we use the benchmark dataset from ref. [37] to be able to compare the network performance for top-tagging with the results quoted in [37]. The dataset consists of jets from hadronically decaying tops and light QCD di-jets at a collision energy of 14 TeV, simulated with PYTHIA 8. Jets in the pT interval [550 GeV, 650 GeV] are clustered according to the anti-kT jet algorithm with a jet radius of 0.8, after a fast detector simulation with DELPHES 3. Jets are required to fulﬁl \|η\|jet < 2. For the top jets, a parton level top is required to fall within ∆R = 0.8 of the ﬁnal jet. Additionally, the three quarks from the hadronic top decay at tree-level are required to obey ∆R < 0.8 with respect to the top. For all jet samples, the four-momenta of the 200 constituents with highest pT are stored in descending pT order. For jets with fewer constituents, zeros are added to obtain the same array size for each jet. For the search discussed in section 4, the samples for training the jet tagger are produced with the tool chain described above. Signal events are generated at a collision energy of 13 TeV for the dark sector benchmark model with rinv = 0.75, mπd = 4 GeV and mρd = Λd = 5 GeV. The QCD jets are extracted from Z+jets events produced at 13 TeV including MLM matching with up to two hard jets. We generate two different samples which populate the ﬁducial volume of the signal regions EM2 and EM4 deﬁned in ref. [46], respectively. The samples consist of 200k signal and 200k background jets. A Jet and event generation To test exclusion for different parameter points, we generate events as described above to determine the cut and tagging efﬁciencies. Note that much fewer events have to be generated compared to the large event number needed to extract the jet sample for training. B Neural network architectures In this appendix we present the architectures for the convolutional neural network (CNN), the Lorentz-layer neural network (LoLa), and the dynamic graph convolutional neural network (DGCNN) used for the classiﬁcation of semi-visible and top jets in the main part of the paper. We use KERAS 2.3.1 [88] with TENSORFLOW 1.13.1 [89] as backend for the implemen- 16 SciPost Phys. 10, 046 (2021) tation, training and evaluation of our networks. If not stated differently, we use the ADAM optimiser [90] in its default conﬁguration to optimise the categorical cross entropy loss. The categorical cross entropy for one-hot encoded labels is given by CE(ytrue, ypred) = − 2 X i=1 ytrue,i ln(ypred,i) = −ln(ypred,true). Here, ytrue and ypred correspond to the true labels and the predicted labels, respectively. Since labels are one-hot encoded, ytrue,i is equal to zero for the wrong class and equal to one for the correct class. Convolutional neural networks We use a CNN consisting of several convolutions for feature extraction and maximum pooling layers for dimensionality reduction, followed by a fully connected part for classiﬁcation. The convolutional layers use 128 ﬁlters with kernels covering 3x3 pixels, the pooling layers apply the max-function on 2x2 pixel windows with stride two, reducing the dimension of the image along both axes by a factor of two. Using the max function for pooling implicitly assumes that pixels with higher intensity are more important for the classiﬁcation. The activation function of choice throughout the network is ReLu, only the last layer applies softmax activation. The output consists of two nodes, one for each class. A sketch of the architecture is shown in the left panel of ﬁgure 9. We have conﬁrmed that varying the network architecture does not improve the performance for semi-visible jets. To obtain a jet image from the four-momenta of its constituents, we ﬁrst calculate the pseudo-rapidity η, azimuthal angle φ and the transverse momentum pT of each constituent. The following preprocessing steps are applied [36]: (1) The pT-weighted centroid of the jet is shifted to the origin in the η-φ-plane. (2) All constituents are rotated such that the pT- weighted principal axis points in the η-direction. (3) The axes is ﬂipped such that the maxi- mum intensity (sum of pT) is in the upper right quadrant. (4) The jet image is generated as the pT weighted histogram in η and φ and normalised by dividing by the total pT. We use 40 bins within the interval [-0.8, 0.8] for both η and φ. For the results shown in this appendix, we train the network on the training set for top tagging provided in [37] to be able to compare with the results presented in [37]. The set consists of 600k jets for each class. The maximum number of epochs is set to 100. The learning rate is reduced, if the validation loss does not improve for three epochs, and the training is stopped after ﬁve more epochs without improvement. The network performance results in this section are based on the independent test of 200k background and 200k signal jets provided along the benchmark data set. Lorentz-layer neural networks Following ref. [37] we construct a network based on a so-called combination layer (CoLa) followed by a Lorentz layer (LoLa). The CoLa receives a list of particle four-momenta, ordered in pT, and calculates a number of linear combinations of those vectors. The coefﬁcients in these linear combinations are trainable, and the output of this layer consists of the original list of momenta appended by the learned linear combinations. The LoLa then transforms every four-vector into ˜kj →kj =      m2(˜kj) pT(˜kj) w(E) jm E(˜km) w(d) jmd2 jm     . (3) (3) 17 SciPost Phys. 10, 046 (2021) Convolution (3x3) (40, 40, 128) MaxPooling (20, 20, 128) Convolution (3x3) (20, 20, 128) MaxPooling (2x2) (10, 10, 128) Convolution (3x3) (10, 10, 128) Flatten (12800) Input (40, 40, 1) Dense (32) Dense (6) Dense (2) CoLa (4, N+M) LoLa (7, N+M) Flatten (7 * (N+M)) Dense (100) Input (4, N) Dense (10) Dense (2) EdgeConv k = 16, C = (64, 64, 64) (N, 64) EdgeConv k = 16, C = (128, 128, 128) (N, 128) EdgeConv k = 16, C = (256, 256, 256) (N, 256) Concatenate (N, 455) Input (N, 7) GlobalAveragePooling (455) Dense (2) Dense (256) Dense (128) Figure 9: Sketch of the CNN architecture for jet image classiﬁcation (left panel), the LoLa architecture for classiﬁcation on four-vectors (middle panel) and for the DGCNN architecture for classiﬁcation on particle clouds (right panel). Each block corresponds to one layer in the network. The ﬁrst line of each block describes the kind of layer and the kernel size, if applicable. The ﬁrst line in the blocks for EdgeConv layers give the number of nearest neighbours k and the number of ﬁlters used in the three convolutions C. The second line states the output dimension of each layer, with N the number of jet constituents used, and M the number of added linear combinations in the LoLa network. Figure 9: Sketch of the CNN architecture for jet image classiﬁcation (left panel), the LoLa architecture for classiﬁcation on four-vectors (middle panel) and for the DGCNN architecture for classiﬁcation on particle clouds (right panel). Each block corresponds to one layer in the network. The ﬁrst line of each block describes the kind of layer and the kernel size, if applicable. Lorentz-layer neural networks The ﬁrst line in the blocks for EdgeConv layers give the number of nearest neighbours k and the number of ﬁlters used in the three convolutions C. The second line states the output dimension of each layer, with N the number of jet constituents used, and M the number of added linear combinations in the LoLa network. The ﬁrst entry corresponds to the invariant mass and the second entry to the transverse mo- mentum of the particle. The third entry is a linear combination of the energies of all particles weighted by trainable parameters. The last entry is a trainable combination of Minkowski dis- tances between particle four-momenta. In practice four distance combinations are added to the vector. For two of the added entries, we sum over the index m, while we take the minimum for the other two entries. To obtain a classiﬁcation, the output of the LoLa is ﬂattened and passed on to a fully connected network. We use ReLu as activation for the fully connected lay- ers, except for the classiﬁcation output, where we apply softmax. CoLa and LoLa do not apply activation functions. A sketch of the architecture is displayed in the central panel of ﬁgure 9. We have conﬁrmed that varying the network architecture does not improve the performance for semi-visible jets. Training is performed in the same way as for the CNN, including the learning rate schedule and early stopping. The performance of the CoLa/LoLa architecture depends on the number of jet constituents that is used as input for the network and on the number of linear combinations added by the CoLa. Ordering the particles in descending pT, we ﬁnd that the best performance is achieved with about 40 jet constituents. The network performance is not particularly sensitive to the number of linear combinations in the CoLa. We have chosen 40 constituents and 15 linear combinations, consistent also with ref. [37]. 18 SciPost Phys. 10, 046 (2021) Dynamic graph neural networks Dynamic graph convolutional neural networks (DGCNNs) have been introduced in ref. [38] and applied to jet tagging in ref. [39]. These network architectures operate on point clouds with so-called edge convolution (EdgeConv) layers. For jet-tagging the point cloud consists of particles, i.e. the jet constituents. The edge convolution differs from a convolution over an image in the deﬁnition of the local patch that the kernel observes. In an image, the local patch corresponds to some neigh- bourhood of pixels. For an edge convolution, a local graph is constructed for each point in the cloud by ﬁnding its k nearest neighbours with respect to some metric which has to be speci- ﬁed. The corresponding graph is called a k-nearest-neighbour (knn) graph. For each particle the convolution is then performed over its nearest neighbours, i.e. x′ i = Ωk j=1hΘ(xi, x j). Here, xi corresponds to the i-th point in the cloud and x′ i to the output of the convolution at this point. The kernel hΘ(xi, x j) is implemented as a fully connected layer and calculates edge features for a point and each of its k neighbours. Those k edge features are reduced to one output feature vector x′ by the aggregation function Ω. This function should not depend on the order of inputs. We use the mean in this work. The same h is then used on all points and their neighbours, just like the kernel in a regular convolution. We follow ref. [39] and use hΘ(xi,∆x j), where ∆x j is the difference of the features of xi and x j. j j j Since the EdgeConv operation produces as output again a point cloud with the same num- ber of points as the input, one can stack EdgeConvs onto each other. Note that the number of output features for the particles is variable and changes from layer to layer. Calculating new nearest neighbours from the output of the previous EdgeConv allows for points that are initially far apart to be grouped close in feature space. We follow ref. [39] in selecting the following input features for the DGCNN: 1. ∆η = η −ηjet where η (ηjet) is the rapidity of the constituent (the jet), 2. ∆φ = φ −φjet where φ (φjet) is the azimuthal angle of the constituent (the jet), 3. log(pT) - constituent’s transverse momentum in GeV, 4. Dynamic graph neural networks log(pT/pTjet) - constituent’s pT relative to the jet pT, 5. log(E) - constituent’s energy in GeV, 6. log(E/Ejet) - constituent’s energy relative to the jet energy, 7. ∆R = p ∆η2 + ∆φ2. 1. ∆η = η −ηjet where η (ηjet) is the rapidity of the constituent (the jet), 2. ∆φ = φ −φjet where φ (φjet) is the azimuthal angle of the constituent (the jet), 3. log(pT) - constituent’s transverse momentum in GeV, 4. log(pT/pTjet) - constituent’s pT relative to the jet pT, 5. log(E) - constituent’s energy in GeV, 5. log(E) - constituent’s energy in GeV, 6. log(E/Ejet) - constituent’s energy relative to the jet energy, We use a combination of EdgeConv layers followed by fully connected layers for the classiﬁ- cation of particle clouds. First, we construct three EdgeConv blocks. At the beginning of each block a new k-nearest-neighbours graph is generated. We set k = 16 for all blocks. In the ﬁrst block, the distance between particles is calculated only in η and φ. In the later blocks, the distance is calculated as the euclidean distance of the complete feature vector. Each EdgeConv block consists of three convolutions on the constructed graph with the same number of ﬁlters. The number of convolution ﬁlters corresponds to the number of features for each particle in the next layer. We use 64 ﬁlters in the ﬁrst block, 128 in the second and 256 in the third. The increasing number of ﬁlters allows the network to extract more and more detailed features. We concatenate the input features and the features from each EdgeConv block for each particle, so that we end up with 7 + 64 + 128 + 256 = 455 features for each jet constituent after the EdgeConv layers of the network. Since we want to keep the network independent of the ordering of the particles, we need to aggregate the constituents in a way that is invariant under permutation. We choose to use the average feature vector, since it shows better results 19 SciPost Phys. 10, 046 (2021) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 S 100 101 102 103 104 1/ B Top jets, 550 GeV < pTj < 650 GeV DGCNN CNN LoLa Figure 10: ROC curves for top tagging with the different neural networks architec- tures described in this appendix. Dynamic graph neural networks The dashed lines denote the ROC curves of the corresponding architectures as presented in ref. [35]. Figure 10: ROC curves for top tagging with the different neural networks architec- tures described in this appendix. The dashed lines denote the ROC curves of the corresponding architectures as presented in ref. [35]. than, for example, the max aggregation. This results in one feature vector with 455 features. These 455 features are the input to the fully connected part of the network. We use three fully connected layers with 256, 128 and 2 nodes respectively and adopt ReLu activation for each layer except for the classiﬁcation output, where we apply softmax. To prevent the fully connected part of the network from overﬁtting, we use dropout layers in front of the ﬁrst two fully connected layers for regularisation and update only 90% of the weights. A sketch of the architecture is shown in the right panel of ﬁgure 9. As before, we use the training and test set for top tagging provided in ref. [35]. We use a learning rate schedule during training. The initial learning rate is set to 3×10−4. We increase it linearly for 8 epochs to 3×10−3 and decrease it to its initial value within another 8 epochs. The next 4 epochs we reduce the learning rate further to 5 × 10−7. Such a learning rate schedule is supposed to lead to faster convergence [91]. Training ﬁnishes after 20 epochs. We do not perform a dedicated hyperparameter optimisation in this work. The parameters we use for this network are comparable to those in ref. [39], except for the number of jet constituents which we ﬁx to 40 for comparability with the LoLa network. We note that different hyperparameter settings could be optimal for top tagging or for tagging semi-visible dark jets. This optimisation is left for future work. Network performances This is due to the number of calculations needed speciﬁcally to compute the pairwise distance for all particles and to construct the k-nearest-neighbours graph. the different networks in table 4. The inference time of the networks may, for example, be crucial when using them as event triggers and also affects the computational effort needed for training. To obtain the inference time, we predict the output of 400k jets with a batch size of 512. We test the networks on a NVIDIA Tesla V100-SXM2-16GB GPU and display the average time needed per image. While the CoLa/LoLa consists mostly of hand crafted, hard coded features and thus involves comparably few trainable parameters, the large number of ﬁlters in the convolutions results in a much larger number of trainable parameters for the CNN. Also the DGCNN has many ﬁlters in the convolutions performed in the EdgeConv blocks, and thus signiﬁcantly more parameters than the CoLa/LoLa network. The inference time is largest for the DGCNN, even though it has fewer parameters than the CNN. This is due to the number of calculations needed speciﬁcally to compute the pairwise distance for all particles and to construct the k-nearest-neighbours graph. Network performances Figure 10 compares the ROC curves (see section 3.1) for top tagging with the CNN, the LoLa and the DGCNN, respectively, to the results presented in ref. [35]. As in section 3.1 the stability and reproducibility of the network performance is evaluated by training ﬁve networks with independent random weight initialisation. The small spread in performance indicated by the shaded band around the ROC curves shows that the training convergence of the network is stable. We ﬁnd very good agreement between our results and the results presented in ref. [35] for the CNN and DGCNN, and a reasonable agreement for the LoLa network. Since the DGCNN provides the best performance, we focus on this architecture in this work and do not attempt to further optimise the performance of the LoLa network. We compare the number of parameters, the inference time and the required storage for 20 SciPost Phys. 10, 046 (2021) Table 4: Comparison of the number of parameters, the inference time and the storage needed for the different architectures introduced in this section. Table 4: Comparison of the number of parameters, the inference time and the storage needed for the different architectures introduced in this section. Table 4: Comparison of the number of parameters, the inference time and the storage needed for the different architectures introduced in this section. Network Parameters Inference time [µs] Needed storage [MB] CNN 706,292 19.1 8.2 CoLa/LoLa 48,031 3.59 0.61 DGCNN 411,458 141.1 4.8 Network Parameters Inference time [µs] Needed storage [MB] the different networks in table 4. The inference time of the networks may, for example, be crucial when using them as event triggers and also affects the computational effort needed for training. To obtain the inference time, we predict the output of 400k jets with a batch size of 512. We test the networks on a NVIDIA Tesla V100-SXM2-16GB GPU and display the average time needed per image. While the CoLa/LoLa consists mostly of hand crafted, hard coded features and thus involves comparably few trainable parameters, the large number of ﬁlters in the convolutions results in a much larger number of trainable parameters for the CNN. Also the DGCNN has many ﬁlters in the convolutions performed in the EdgeConv blocks, and thus signiﬁcantly more parameters than the CoLa/LoLa network. The inference time is largest for the DGCNN, even though it has fewer parameters than the CNN. References [1] A. J. Larkoski, I. Moult and B. Nachman, Jet substructure at the Large Hadron Collider: A review of recent advances in theory and machine learning, Phys. Rep. 841, 1 (2020), doi:10.1016/j.physrep.2019.11.001, [arXiv:1709.04464]. [2] D. Guest, K. Cranmer and D. Whiteson, Deep learning and its application to LHC physics, Annu. Rev. Nucl. Part. Sci. 68, 161 (2018), doi:10.1146/annurev-nucl-101917-021019, [arXiv:1806.11484]. [3] P. Schwaller, D. Stolarski and A. Weiler, Emerging jets, J. High Energy Phys. 05, 059 (2015), doi:10.1007/JHEP05(2015)059, [arXiv:1502.05409]. [4] T. Cohen, M. Lisanti, and H. K. Lou, Semivisible jets: Dark Matter undercover at the LHC, Phys. Rev. Lett. 115, 171804 (2015), doi:10.1103/PhysRevLett.115.171804, [arXiv:1503.00009]. [5] T. Cohen, M. Lisanti, H. Keong Lou and S. Mishra-Sharma, LHC searches for dark sec- tor showers, J. High Energy Phys. 11, 196 (2017), doi:10.1007/JHEP11(2017)196, [arXiv:1707.05326]. [6] A. Pierce, B. Shakya, Y. Tsai and Y. Zhao, Searching for conﬁning hidden valleys at LHCb, ATLAS, and CMS, Phys. Rev. D 97, 095033 (2018), doi:10.1103/PhysRevD.97.095033, [arXiv:1708.05389]. [7] M. Park and M. Zhang, Tagging a jet from a dark sector with jet substructures at colliders, Phys. Rev. D 100, 115009 (2019), doi:10.1103/PhysRevD.100.115009, [arXiv:1712.09279]. 21 SciPost Phys. 10, 046 (2021) [8] S. Renner and P. Schwaller, A ﬂavoured dark sector, J. High Energy Phys. 08, 052 (2018), doi:10.1007/JHEP08(2018)052, [arXiv:1803.08080]. [9] H. Beauchesne, E. Bertuzzo and G. Grilli di Cortona, Dark matter in Hidden Valley mod- els with stable and unstable light dark mesons, J. High Energy Phys. 04, 118 (2019), doi:10.1007/JHEP04(2019)118, [arXiv:1809.10152]. [10] E. Bernreuther, F. Kahlhoefer, M. Krämer and P. Tunney, Strongly interacting dark sectors in the early Universe and at the LHC through a simpliﬁed portal, J. High Energy Phys. 01, 162 (2020), doi:10.1007/JHEP01(2020)162, [arXiv:1907.04346]. [11] T. Cohen, J. Doss and M. Freytsis, Jet substructure from dark sector showers, J. High Energy Phys. 09, 118 (2020), doi:10.1007/JHEP09(2020)118, [arXiv:2004.00631]. [12] M. J. Strassler and K. M. Zurek, Echoes of a hidden valley at hadron colliders, Phys. Lett. B 651, 374 (2007), doi:10.1016/j.physletb.2007.06.055, [arXiv:hep-ph/0604261]. [13] T. Han, Z. Si, K. M. Zurek, and M. J. Strassler, Phenomenology of hidden valleys at hadron colliders, J. High Energy Phys. 07, 008 (2008), doi:10.1088/1126-6708/2008/07/008, [arXiv:0712.2041]. [14] L. Carloni and T. Sjöstrand, Visible effects of invisible Hidden Valley radiation, J. High Energy Phys. 09, 105 (2010), doi:10.1007/JHEP09(2010)105, [arXiv:1006.2911]. [15] L. Carloni, J. Rathsman, and T. Sjöstrand, Discerning secluded sector gauge structures, J, High Energy Phys. 04, 091 (2011), doi:10.1007/JHEP04(2011)091, [arXiv:1102.3795]. [16] Y. Hochberg, E. References Kuﬂik, T. Volansky and J. G. Wacker, Mechanism for thermal relic dark matter of strongly interacting massive particles, Phys. Rev. Lett. 113, 171301 (2014), doi:10.1103/PhysRevLett.113.171301, [arXiv:1402.5143]. [17] Y. Hochberg, E. Kuﬂik, H. Murayama, T. Volansky and J. G. Wacker, Model for thermal relic dark matter of strongly interacting massive particles, Phys. Rev. Lett. 115, 021301 (2015), doi:10.1103/PhysRevLett.115.021301, [arXiv:1411.3727]. [18] M. Hansen, K. Langæble, and F. Sannino, SIMP model at NNLO in chiral perturba- tion theory, Phys. Rev. D 92, 075036 (2015), doi:10.1103/PhysRevD.92.075036, [arXiv:1507.01590]. [19] Y. Hochberg, E. Kuﬂik, and H. Murayama, SIMP Spectroscopy, J. High Energy Phys. 05, 090 (2016), doi:10.1007/JHEP05(2016)090, [arXiv:1512.07917]. [20] G. D. Kribs and E. T. Neil, Review of strongly-coupled composite dark mat- ter models and lattice simulations, Int. J. Mod. Phys. A 31, 1643004 (2016), doi:10.1142/S0217751X16430041, [arXiv:1604.04627]. [21] C. Englert, K. Nordström and M. Spannowsky, Towards resolving strongly-interacting dark sectors at colliders, Phys. Rev. D 94, 055028 (2016), doi:10.1103/PhysRevD.94.055028, [arXiv:1606.05359]. [22] K. R. Dienes, F. Huang, S. Su, and B. Thomas, Dynamical dark mat- ter from strongly-coupled dark sectors, Phys. Rev. D 95, 043526 (2017), doi:10.1103/PhysRevD.95.043526, [arXiv:1610.04112]. [23] S.-M. Choi, H. Min Lee and M.-S. Seo, Cosmic abundances of SIMP dark matter, J. High Energy Phys. 04, 154 (2017), doi:10.1007/JHEP04(2017)154, [arXiv:1702.07860]. 22 SciPost Phys. 10, 046 (2021) [24] A. Berlin, N. Blinov, S. Gori, P. Schuster and N. Toro, Cosmology and acceler- ator tests of strongly interacting dark matter, Phys. Rev. D 97, 055033 (2018), doi:10.1103/PhysRevD.97.055033, [arXiv:1801.05805]. [25] S.-M. Choi, H. Min Lee, P. Ko and A. Natale, Resolving phenomenological problems with strongly-interacting-massive-particle models with dark vector resonances, Phys. Rev. D 98, 015034 (2018), doi:10.1103/PhysRevD.98.015034. [26] M. Heikinheimo, K. Tuominen, and K. Langæble, Hidden strongly interacting mas- sive particles, Phys. Rev. D 97, 095040 (2018), doi:10.1103/PhysRevD.97.095040, [arXiv:1803.07518]. [27] Y. Hochberg, E. Kuﬂik, R. McGehee, H. Murayama and K. Schutz, Strongly inter- acting massive particles through the axion portal, Phys. Rev. D 98, 115031 (2018), doi:10.1103/PhysRevD.98.115031, [arXiv:1806.10139]. [28] G. D. Kribs, A. Martin, and T. Tong, Effective theories of dark mesons with custodial symmetry, J. High Energy Phys. 08, 020 (2019), doi:10.1007/JHEP08(2019)020, [arXiv:1809.10183]. [29] G. D. Kribs, A. Martin, B. Ostdiek, and T. Tong, Dark mesons at the LHC, J. High Energy Phys. 07, 133 (2019), doi:10.1007/JHEP07(2019)133, [arXiv:1809.10184]. [30] S.-M. Choi, H. M. Lee, Y. Mambrini, and M. Pierre, Vector SIMP dark mat- ter with approximate custodial symmetry, J. High Energy Phys. 07, 049 (2019), doi:10.1007/JHEP07(2019)049, [arXiv:1904.04109]. References [31] P. Brax, S. Fichet and P. Tanedo, The warped dark sector, Phys. Lett. B 798, 135012 (2019), doi:10.1016/j.physletb.2019.135012, [arXiv:1906.02199]. [32] H.-C. Cheng, L. Li, E. Salvioni and C. B. Verhaaren, Light hidden mesons through the Z portal, J. High Energy Phys. 11, 031 (2019), doi:10.1007/JHEP11(2019)031, [arXiv:1906.02198]. [33] H. Beauchesne and G. Grilli di Cortona, Classiﬁcation of dark pion multiplets as dark matter candidates and collider phenomenology, J. High Energy Phys. 02, 196 (2020), doi:10.1007/JHEP02(2020)196, [arXiv:1910.10724]. [34] J. Thaler and K. Van Tilburg, Identifying boosted objects with N-subjettiness, J. High Energy Phys. 03, 015 (2011), doi:10.1007/JHEP03(2011)015, [arXiv:1011.2268]. [35] G. Kasieczka et al., The Machine Learning landscape of top taggers, SciPost Phys. 7, 014 (2019), doi:10.21468/SciPostPhys.7.1.014, [arXiv:1902.09914]. [36] S. Macaluso and D. Shih, Pulling out all the tops with computer vision and deep learning, J. High Energy Phys. 10, 121 (2018), doi:10.1007/JHEP10(2018)121, [arXiv:1803.00107]. [37] A. Butter, G. Kasieczka, T. Plehn and M. Russell, Deep-learned top tagging with a Lorentz layer, SciPost Phys. 5, 028 (2018), doi:10.21468/SciPostPhys.5.3.028, [arXiv:1707.08966]. [38] Y. Wang et al., Dynamic Graph CNN for learning on Point Clouds, ACM Transactions on Graphics (TOG) 38, 1 (2019). 23 23 SciPost Phys. 10, 046 (2021) [39] H. Qu and L. Gouskos, Jet tagging via particle clouds, Phys. Rev. D 101, 056019 (2020), doi:10.1103/PhysRevD.101.056019, [arXiv:1902.08570]. [40] J. Arjona Martínez, O. Cerri, M. Spiropulu, J. R. Vlimant and M. Pierini, Pileup mitigation at the Large Hadron Collider with graph neural networks, Eur. Phys. J. Plus 134, 333 (2019), doi:10.1140/epjp/i2019-12710-3, [arXiv:1810.07988]. [41] S. R. Qasim, J. Kieseler, Y. Iiyama and M. Pierini, Learning representations of irregular particle-detector geometry with distance-weighted graph networks, Eur. Phys. J. C 79, 608 (2019), doi:10.1140/epjc/s10052-019-7113-9, [arXiv:1902.07987]. [42] E. A. Moreno et al., JEDI-net: a jet identiﬁcation algorithm based on interac- tion networks, Eur. Phys. J. C 80, 58 (2020), doi:10.1140/epjc/s10052-020-7608-4, [arXiv:1908.05318]. [43] E. A. Moreno, T. Q. Nguyen, J.-R. Vlimant, O. Cerri, H. B. Newman, A. Periwal, M. Spirop- ulu, J. M. Duarte and M. Pierini, Interaction networks for the identiﬁcation of boosted H →bb decays, Phys. Rev. D 102, 012010 (2020), doi:10.1103/PhysRevD.102.012010, [arXiv:1909.12285]. [44] X. Ju et al., Graph neural networks for particle reconstruction in high energy physics detec- tors (2020), arXiv:2003.11603. [45] M. Abdughani, D. Wang, L. Wu, J. M. Yang and J. Zhao, Probing triple Higgs coupling with machine learning at the LHC (2020), arXiv:2005.11086. [46] M. References Aaboud et al., Search for dark matter and other new phenomena in events with an energetic jet and large missing transverse momentum using the ATLAS detector, J. High Energy Phys. 01, 126 (2018), doi:10.1007/JHEP01(2018)126, [arXiv:1711.03301]. [47] A. M. Sirunyan et al., Search for new physics in ﬁnal states with an energetic jet or a hadron- ically decaying W or Z boson and transverse momentum imbalance at ps = 13 TeV, Phys. Rev. D 97, 092005 (2018), doi:10.1103/PhysRevD.97.092005, [arXiv:1712.02345]. [48] ATLAS Collaboration, Search for new phenomena in Dijet events using 139 fb−1 of pp col- lisions at ps = 13 TeV collected with the ATLAS Detector, ATLAS-CONF-2019-007 (2019). [49] ATLAS Collaboration, Search for dark matter and other new phenomena in events with an energetic jet and large missing transverse momentum using the ATLAS detector, ATLAS- CONF-2017-060 (2017). [50] A. M. Sirunyan et al., Search for supersymmetry in multijet events with missing trans- verse momentum in proton-proton collisions at 13 TeV, Phys. Rev. D 96, 032003 (2017), doi:10.1103/PhysRevD.96.032003, [arXiv:1704.07781]. [51] J. Cogan, M. Kagan, E. Strauss and A. Schwarztman, Jet-images: computer vi- sion inspired techniques for jet tagging, J. High Energy Phys. 02, 118 (2015), doi:10.1007/JHEP02(2015)118, [arXiv:1407.5675]. [52] L. de Oliveira, M. Kagan, L. Mackey, B. Nachman and A. Schwartzman, Jet-images — deep learning edition, J. High Energy Phys. 07, 069 (2016), doi:10.1007/JHEP07(2016)069, [arXiv:1511.05190]. [53] P. Baldi, K. Bauer, C. Eng, P. Sadowski and D. Whiteson, Jet substructure classiﬁca- tion in high-energy physics with deep neural networks, Phys. Rev. D 93, 094034 (2016), doi:10.1103/PhysRevD.93.094034, [arXiv:1603.09349]. 24 SciPost Phys. 10, 046 (2021) [54] G. Kasieczka, T. Plehn, M. Russell and T. Schell, Deep-learning top taggers or the end of QCD?, J. High Energy Phys. 05, 006 (2017), doi:10.1007/JHEP05(2017)006, [arXiv:1701.08784]. [55] J. Duarte et al., FPGA-accelerated machine learning inference as a service for particle physics computing, Comput. Softw. Big Sci. 3, 13 (2019), doi:10.1007/s41781-019-0027- 2, [arXiv:1904.08986]. [56] L. G. Almeida, M. Backovi´c, M. Cliche, S. J. Lee and M. Perelstein, Playing tag with ANN: boosted top identiﬁcation with pattern recognition, J. High Energy Phys. 07, 086 (2015), doi:10.1007/JHEP07(2015)086, [arXiv:1501.05968]. [57] G. Louppe, K. Cho, C. Becot and K. Cranmer, QCD-aware recursive neural networks for jet physics, J. High Energy Phys. 01, 057 (2019), doi:10.1007/JHEP01(2019)057, [arXiv:1702.00748]. [58] M. Erdmann, E. Geiser, Y. Rath and M. Rieger, Lorentz Boost Networks: autonomous physics-inspired feature engineering, J. Inst. 14, P06006 (2019), doi:10.1088/1748- 0221/14/06/P06006, [arXiv:1812.09722]. [59] P. Baldi, K. References Cranmer, T. Faucett, P. Sadowski and D. Whiteson, Parameterized neural net- works for high-energy physics, Eur. Phys. J. C 76, 235 (2016), doi:10.1140/epjc/s10052- 016-4099-4, [arXiv:1601.07913]. [60] J. de Favereau et al., DELPHES 3: a modular framework for fast simulation of a generic col- lider experiment, J. High Energy Phys. 02, 057 (2014), doi:10.1007/JHEP02(2014)057, [arXiv:1307.6346]. [61] D. Dercks, N. Desai, J. Soo Kim, K. Rolbiecki, J. Tattersall and T. Weber, Check- MATE 2: From the model to the limit, Comput. Phys. Commun. 221, 383 (2017), doi:10.1016/j.cpc.2017.08.021, [arXiv:1611.09856]. [62] P. T. Komiske, E. M. Metodiev and J. Thaler, Energy ﬂow polynomials: a com- plete linear basis for jet substructure, J. High Energy Phys. 04, 013 (2018), doi:10.1007/JHEP04(2018)013, [arXiv:1712.07124]. [63] A. Butter, G. Kasieczka, T. Plehn and M. Russell, Deep-learned top tagging with a Lorentz layer, SciPost Phys. 5, 028 (2018), doi:10.21468/SciPostPhys.5.3.028, [arXiv:1707.08966]. [64] P. T. Komiske, E. M. Metodiev and J. Thaler, Energy ﬂow networks: deep sets for particle jets, J. High Energy Phys. 01, 121 (2019), doi:10.1007/JHEP01(2019)121, [arXiv:1810.05165]. [65] L. Moore, K. Nordström, S. Varma and M. Fairbairn, Reports of my demise are greatly exaggerated: N-subjettiness taggers take on jet images, SciPost Phys. 7, 036 (2019), doi:10.21468/SciPostPhys.7.3.036, [arXiv:1807.04769]. [66] Y.-C. J. Chen, C.-W. Chiang, G. Cottin and D. Shih, Boosted W and Z tag- ging with jet charge and deep learning, Phys. Rev. D 101, 053001 (2020), doi:10.1103/PhysRevD.101.053001, [arXiv:1908.08256]. [67] A. Andreassen, I. Feige, C. Frye and M. D. Schwartz, Binary JUNIPR: An inter- pretable probabilistic model for discrimination, Phys. Rev. Lett. 123, 182001 (2019), doi:10.1103/PhysRevLett.123.182001, [arXiv:1906.10137]. 25 SciPost Phys. 10, 046 (2021) [68] A. Chakraborty, S. Hak Lim, M. M. Nojiri and M. Takeuchi, Neural network-based top tagger with two-point energy correlations and geometry of soft emissions, J. High Energy Phys. 07, 111 (2020), doi:10.1007/JHEP07(2020)111, [arXiv:2003.11787]. [69] S. Bollweg, M. Haussmann, G. Kasieczka, M. Luchmann, T. Plehn and J. Thomp- son, Deep-learning jets with uncertainties and more, SciPost Phys. 8, 006 (2020), doi:10.21468/SciPostPhys.8.1.006, [arXiv:1904.10004]. [70] B. Nachman, A guide for deploying Deep Learning in LHC searches: How to achieve optimality and account for uncertainty, SciPost Phys. 8, 090 (2020), doi:10.21468/SciPostPhys.8.6.090, [arXiv:1909.03081]. [71] B. Nachman and C. Shimmin, AI safety for high energy physics (2019), arXiv:1910.08606. [72] O. Amram and C. Mantilla Suarez, Tag N’ Train: a technique to train improved classiﬁers on unlabeled data, J. High Energy Phys. 01, 153 (2021), doi:10.1007/JHEP01(2021)153, [arXiv:2002.12376]. [73] J. Hajer, Y.-Y. Li, T. References Liu and H. Wang, Novelty detection meets collider physics, Phys. Rev. D 101, 076015 (2020), doi:10.1103/PhysRevD.101.076015, [arXiv:1807.10261]. [74] T. Heimel, G. Kasieczka, T. Plehn and J. Thompson, QCD or what?, SciPost Phys. 6, 030 (2019), doi:10.21468/SciPostPhys.6.3.030, [arXiv:1808.08979]. [75] M. Farina, Y. Nakai and D. Shih, Searching for new physics with deep autoencoders, Phys. Rev. D 101, 075021 (2020), doi:10.1103/PhysRevD.101.075021, [arXiv:1808.08992]. [76] O. Cerri, T. Q. Nguyen, M. Pierini, M. Spiropulu and J.-R. Vlimant, Variational autoen- coders for new physics mining at the Large Hadron Collider, J. High Energy Phys. 05, 036 (2019), doi:10.1007/JHEP05(2019)036, [arXiv:1811.10276]. [77] E. M. Metodiev, B. Nachman and J. Thaler, Classiﬁcation without labels: learning from mixed samples in high energy physics, J. High Energy Phys. 10, 174 (2017), doi:10.1007/JHEP10(2017)174, [arXiv:1708.02949]. [78] J. Collins, K. Howe and B. Nachman, Anomaly detection for resonant new physics with machine learning, Phys. Rev. Lett. 121, 241803 (2018), doi:10.1103/PhysRevLett.121.241803, [arXiv:1805.02664]. [79] B. Nachman and D. Shih, Anomaly detection with density estimation, Phys. Rev. D 101, 075042 (2020), doi:10.1103/PhysRevD.101.075042, [arXiv:2001.04990]. [80] A. Andreassen, B. Nachman and D. Shih, Simulation assisted likelihood-free anomaly detection, Phys. Rev. D 101, 095004 (2020), doi:10.1103/PhysRevD.101.095004, [arXiv:2001.05001]. [81] J. Alwall et al., The automated computation of tree-level and next-to-leading order differen- tial cross sections, and their matching to parton shower simulations, J. High Energy Phys. 07, 079 (2014), doi:10.1007/JHEP07(2014)079, [arXiv:1405.0301]. [82] R. D. Ball, V. Bertone, S. Carrazza, L. Del Debbio, S. Forte, A. Guffanti, N. P. Hartland and J. Rojo, Parton distributions with QED corrections, Nucl. Phys. B 877, 290 (2013), doi:10.1016/j.nuclphysb.2013.10.010, [arXiv:1308.0598]. [83] A. Alloul, N. D. Christensen, C. Degrande, C. Duhr and B. Fuks, FeynRules 2.0 — A com- plete toolbox for tree-level phenomenology, Comput. Phys. Commun. 185, 2250 (2014), doi:10.1016/j.cpc.2014.04.012, [arXiv:1310.1921]. 26 SciPost Phys. 10, 046 (2021) [84] S. Hoeche et al., Matching parton showers and matrix elements, in HERA and the LHC: A Workshop on the Implications of HERA for LHC Physics: CERN - DESY Workshop 2004/2005, 288 (2005), doi:10.5170/CERN-2005-014.288, [arXiv:hep-ph/0602031]. [85] T. Sjöstrand et al., An introduction to PYTHIA 8.2, Comput. Phys. Commun. 191, 159 (2015), doi:10.1016/j.cpc.2015.01.024, [arXiv:1410.3012]. [86] M. Cacciari, G. P. Salam and G. Soyez, FastJet user manual, Eur. Phys. J. C 72, 1896 (2012), doi:10.1140/epjc/s10052-012-1896-2, [arXiv:1111.6097]. [87] M. Cacciari, G. P. Salam and G. Soyez, The anti-kt jet clustering algorithm, J. High Energy Phys. 04, 063 (2008), doi:10.1088/1126-6708/2008/04/063, [arXiv:0802.1189]. [88] F. Chollet et al., Keras (2015), https://keras.io. [89] M. References Abadi et al., TensorFlow: Large-scale machine learning on heterogeneous systems (2015), Software available from tensorﬂow.org. [90] D. P. Kingma and J. Ba, Adam: A Method for Stochastic Optimization (2014), arXiv:1412.6980. [91] L. N. Smith, A disciplined approach to neural network hyper-parameters: Part 1 – learning rate, batch size, momentum, and weight decay (2018), arXiv:1803.09820. 27
https://openalex.org/W4390640740	https://www.estudosinstitucionais.com/REI/article/download/669/915	Portuguese	null	PLURALIDADE E MUDANÇA INSTITUCIONAL	Revista de Estudos Institucionais	2,024	cc-by	29,206	1 Doutor em Direito pela Universidade Federal de Pernambuco (UFPE). Mestre em Direito Público pela Universidade Federal de Alagoas (Ufal). Pesquisado do Grupo de Pesquisa JUSPOLÍTICA - Diálogos, Historicidades e Judicialização de Políticas (Direito/UPE). Procurador da República. 2 Mestre e Doutora em Direito Público pela Universidade Federal de Pernambuco (UFPE). Professora da Faculdade de Direito da Universidade de Pernambuco (UPE/Recife). Líder do Grupo de Pesquisa JUSPOLÍTICA - Diálogos, Historicidades e Judicialização de Políticas (Direito/UPE). Professora Visitante da Universitat de València/ Espanha. Advogada da União. SUBMETIDO: 19 OUT. 2021 ACEITO: 12 AGO. 2023 SUBMETIDO: 19 OUT. 2021 ACEITO: 12 AGO. 2023 doi: 10.21783/rei.v10i1.669 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PLURALITY AND INSTITUTIONAL CHANGE: IS THERE EVIDENCE TO SUPPORT THE SHIFTING PRIORITIES HYPOTHESIS IN THE BRAZILIAN PUBLIC PROSECUTOR’S OFFICE INSTITUTIONAL PERFORMANCE? BRUNO LAMENHA 1 FLÁVIA DANIELLE SANTIAGO LIMA 2 RESUMO: O trabalho apresenta duas perguntas de pesquisa: i) os efeitos do modelo federativo brasileiro no desenho do MP brasileiro (pluralidade institucional) permite pensá-lo como uma organização nacional? ii) há suporte empírico para a hipótese de deslocamento das prioridades de atuação do MP do campo da tutela coletiva para a repressão penal anticorrupção a partir de meados dos anos 2000 (giro de prioridades), considerando essa pluralidade institucional? O marco teórico- metodológico articula o institucionalismo histórico e o mapeamento de processos como um modelo analítico adequado para a construção de explicações causais sobre processos históricos-temporais com base em evidências qualitativas. Três eixos de análise são percorridos no texto: i) investigam-se regularidades compartilhadas pelos órgãos do MP brasileiro e até que ponto isso permitiria seu estudo como organização nacional; ii) mapeiam-se elementos relevantes da sequência temporal relacionada ao desenvolvimento de uma agenda nacional anticorrupção a partir de 2003; iii) analisa-se o desempenho do MP brasileiro, a partir dos dados de produtividade publicados pelo CNMP desde 2014. Como resultado, propõe-se que, embora imponha limites, a pluralidade institucional não 37 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 ORCID 0000-0002-6655-2735 ORCID 0000-0002-6995-0982 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA impede o estudo do MP como organização nacional; e que a hipótese do giro de prioridades tem suporte empírico para ser pensada como uma tendência nacional. impede o estudo do MP como organização nacional; e que a hipótese do giro de prioridades tem suporte empírico para ser pensada como uma tendência nacional. PALAVRAS-CHAVE: Ministério Público; Pluralidade Institucional; Mudança Institucional; Giro de Prioridades PALAVRAS-CHAVE: Ministério Público; Pluralidade Institucional; Mudança Institucional; Giro de Prioridades ABSTRACT: The papers presents two research questions: i) is it possible to think the Public Prosecutor’s Office (PPO) as a national organization, despite the effects of Brazilian federative model on its design (institutional plurality)?; ii) is there empirical evidence supporting the shifting priorities hypothesis on PPO agency from an collective redress approach to an anticorruption criminal repression focus since the mid-2000s, considering the institutional plurality? The theoretical- methodological framework articulates historical institutionalism and process tracing as an adequate analytical model for the construction of causal explanations about historical-temporal processes based on qualitative evidence. 3 A expressão “Ministério Público brasileiro” ou “MP brasileiro” será utilizada para se referir à reunião de todos 30 (trinta) órgãos ministeriais do sistema de justiça: os quatro ramos do Ministério Público da União: Ministério Público Federal (MPF), Ministério Público do Trabalho (MPT), Ministério Público Militar (MPM) e Ministério Público do Distrito Federal e Territórios (MPDFT); e todos os 26 (vinte e seis) Ministérios Públicos estaduais. Exclui-se dessa contagem os chamados Ministérios Públicos de Contas, uma vez que, apesar da nomenclatura, são órgãos inseridos na estrutura dos Tribunais de Contas e não organizações ministeriais autônomas. PLURALITY AND INSTITUTIONAL CHANGE: IS THERE EVIDENCE TO SUPPORT THE SHIFTING PRIORITIES HYPOTHESIS IN THE BRAZILIAN PUBLIC PROSECUTOR’S OFFICE INSTITUTIONAL PERFORMANCE? Three sections are covered in the text: i) an investigation about possible regularities shared by the bodies of Brazilian PPO, focusing in the possibility of its study as a national organizational; ii) the identification of relevant elements of the temporal sequence related to the development of a national anti-corruption agenda from 2003; iii) the analysis of PPO performance based on productivity data published by the CNMP since 2014. Two main results are achieved. First, it is proposed that, although it imposes limits, institutional plurality does not prevent the study of Brazilian PPO as a national organization. Secondly, there are empirical support to think the priority shifting hypothesis as a national trend on Brazilian PPO. 38 KEYWORDS: Public Prosecutor’s Office; Institutional Plurality; Institutional Change; State Pluralism; Priorities Shift. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? dos direitos difusos, coletivos e individuais homogêneos para privilegiar o campo da repressão penal 4 anticorrupção (AVRITZER; MARONA, 2017; SILVA et al., 2017; SAMPAIO; VIEGAS, 2019). A esse fenômeno cuja procedência ainda será discutida neste trabalho, dá-se o nome de giro de prioridades. dos direitos difusos, coletivos e individuais homogêneos para privilegiar o campo da repressão penal 4 anticorrupção (AVRITZER; MARONA, 2017; SILVA et al., 2017; SAMPAIO; VIEGAS, 2019). A esse fenômeno cuja procedência ainda será discutida neste trabalho, dá-se o nome de giro de prioridades. Um problema de partida dessa hipótese reside no fato de que, de maneira geral, a literatura especializada sobre o MP tem dado pouca atenção às implicações do modelo federativo brasileiro nesta discussão, na chave do que Lamenha e Lima (2021, p. 87) denominam como pluralidade institucional, isto é, o fato de que não há um único Ministério Público no País, mas sim 30 (trinta) órgãos ministeriais distintos, sem que haja subordinação finalística entre eles. q j O presente texto pretende oferecer alguns elementos para a reflexão dessa potencial tensão que se estabelece entre a hipótese do giro de prioridades e o fenômeno da pluralidade institucional acima delineada. Nessa perspectiva, duas questões principais são propostas: i) o fenômeno da pluralidade institucional permite pensar analiticamente o Ministério Público brasileiro como uma organização nacional?; ii) há suporte empírico para afirmar que o giro de prioridades é uma tendência verificável no MP brasileiro como um todo? Fixados os problemas de pesquisa, o texto será organizado em duas seções, além desta introdução e das considerações finais. Primeiro, discutir-se-á mais extensamente a questão da pluralidade institucional, apontando se e em que condições o MP brasileiro pode ser pensado como uma organização nacional, já cotejando o tema com a exposição do argumento do giro de prioridades na literatura especializada. Na segunda seção, será debatido se o argumento do giro de prioridades tem base empírica considerando o fenômeno da pluralidade institucional, aqui assumido como premissa analítica para o estudo do MP brasileiro. 39 39 p p O marco teórico-metodológico principal articula o institucionalismo histórico (IH) e o mapeamento de processos (process tracing) como um modelo analítico adequado para a construção de explicações causais sobre processos históricos- temporais com base em evidências qualitativas, além de permitir a consideração da capacidade de agência de organizações estatais, como o MP, como elemento distinto da atuação de seus membros. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 INTRODUÇÃO Este artigo se propõe a investigar uma hipótese cada vez mais mobilizada por parte importante da literatura especializada sobre o Ministério Público brasileiro3 (MP), nomeadamente aquela voltada para estudos da trajetória e dos arranjos institucionais associados à organização. Essa leitura propõe que, a partir de meados dos anos 2000, o MP teria deixado de priorizar a sua atuação no campo da defesa 3 A expressão “Ministério Público brasileiro” ou “MP brasileiro” será utilizada para se referir à reunião de todos 30 (trinta) órgãos ministeriais do sistema de justiça: os quatro ramos do Ministério Público da União: Ministério Público Federal (MPF), Ministério Público do Trabalho (MPT), Ministério Público Militar (MPM) e Ministério Público do Distrito Federal e Territórios (MPDFT); e todos os 26 (vinte e seis) Ministérios Públicos estaduais. Exclui-se dessa contagem os chamados Ministérios Públicos de Contas, uma vez que, apesar da nomenclatura, são órgãos inseridos na estrutura dos Tribunais de Contas e não organizações ministeriais autônomas. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 5 Observe-se que, no período entre a ruptura de 1964 até a promulgação da Emenda Constitucional n. 1/1969, o Ministério Público não era institucionalmente vinculado ao Poder Executivo nos textos constitucionais do período (Cartas de 1946 e 1967). 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? Aqui, interessa pensar, entre outros pontos, a hipótese do giro de prioridades a partir de categorias clássicas do IH, como a dependência da trajetória a mudança institucional informal ou gradual (MAHONEY; THELEN, 2010, p. 4-14; HACKER et al., 2013, p. 203-204; ONOMA, 2016, p. 68). Além disso, a análise deste fenômeno em particular, e também da questão da pluralidade institucional, exigem a análise dos elementos mobilizados 4 Enfatiza-se aqui a atuação no âmbito criminal, uma vez que um mesmo ato ilícito associado à prática de corrupção pode gerar repercussão no âmbito administrativo, cível e penal. Isso decorre da chamada independência entre instâncias de responsabilização que caracteriza o direito brasileiro. No caso da atuação do MP, a atuação neste campo abrange, além do âmbito criminal, também a dimensão cível, nomeadamente através da responsabilização por ato de improbidade administrativa, no marco da Lei n. 8.429/92. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA ORCID 0000-0002-6655-2735 ORCID 0000-0002-6995-0982 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA pela literatura especializada para dar suporte a tal hipótese, inseridos em uma abordagem contexto-dependente. A fim de oferecer uma perspectiva adicional à testagem da hipótese, o recorte proposto neste trabalho se voltará para o desempenho institucional dos órgãos ministeriais brasileiros, a partir dos dados de produtividade anualmente publicados pelo CNMP, compreendendo o período de 2014-2021. 2. O DESENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO DE 1988 E O FENÔMENO DA PLURALIDADE INSTITUCIONAL 2.1. A TENSÃO ENTRE O GIRO DE PRIORIDADES PROPOSTO PELA LITERATURA E A PLURALIDADE INSTITUCIONAL: O RISCO DA GENERALIZAÇÃO DE EXPERIÊNCIAS PARTICULARES 2. O DESENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO DE 1988 E O FENÔMENO DA PLURALIDADE INSTITUCIONAL 2.1. A TENSÃO ENTRE O GIRO DE PRIORIDADES PROPOSTO PELA LITERATURA E A PLURALIDADE INSTITUCIONAL: O RISCO DA GENERALIZAÇÃO DE EXPERIÊNCIAS PARTICULARES As mudanças institucionais formais ocorridas nos anos 1970 e 1980, consistentes em uma certa reengenharia da estrutura normativo-constitucional do Ministério Público, são reconhecidas como um dos principais legados da Assembleia Nacional Constituinte de 1987-1988 (ANC 1987-1988). Este é um ponto recorrentemente explorado pela literatura especializada e o ponto de partida de uma parte importante das reflexões sobre a organização nas últimas décadas (SADEK, 2000; ARANTES, 2002; KERCHE, 2009; CARVALHO e LEITÃO, 2010; LEMGRUBER et. al., 2016; RIBEIRO, 2017; LIMA e LAMENHA, 2021). De uma organização vinculada ao Executivo na maior parte do regime de 1964- 19855, — e que, no âmbito federal, detinha uma natureza bifronte: além de titular da ação penal, funcionava também como órgão de advocacia de governo (acumulando as funções atualmente desempenhadas pela Advocacia-Geral da União) —, as modificações havidas no perfil do Ministério Público brasileiro durante a transição política para a Nova República foram numerosas. Foi-lhe atribuído, além de sua clássica atribuição de persecução criminal, um amplo papel de defesa judicial e extrajudicial dos direitos difusos, coletivos e individuais homogêneos e uma peculiar tarefa de defesa da ordem jurídica e do regime democrático. 40 Ao lado dessas alterações significativas na atuação finalística da organização, também foi consolidado um extenso plexo de prerrogativas e garantias: de autonomia administrativa e financeira para a organização; e de vitaliciedade, inamovibilidade e, sobretudo, independência funcional para os seus membros. Além dessa profunda mudança institucional culminada no texto da Constituição de 1988, o particular interesse pelo novo desenho normativo do MP parece associado, na literatura, a uma certa e amplamente difundida ideia de 5 Observe-se que, no período entre a ruptura de 1964 até a promulgação da Emenda Constitucional n. 1/1969, o Ministério Público não era institucionalmente vinculado ao Poder Executivo nos textos constitucionais do período (Cartas de 1946 e 1967). 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PLURALIDADE E MUDANÇA INSTITUCIONAL: 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? singularidade (LIMA; MACHADO, 2022, p. 5) por ele revelada e que se desdobra em, ao menos, dois sentidos. Um primeiro sentido, que se pode denominar externo, afirma o perfil “sem paralelo no mundo” do MP de 19886 (LEITE, 2020, p. 168), nomeadamente em razão do seu amplo plexo de atribuições e que tem sido mobilizado, ao longo dos anos, como uma evidência do caráter avançado do perfil institucional do MP de 19887. O segundo sentido, que se pode denominar interno, associado à ideia de que o desenho institucional legado pela CRFB/88 atribuiu ao MP a capacidade, as atribuições e as prerrogativas que o converteram em um ator capaz de interferir, virtualmente, em qualquer questão de relevo público na sociedade brasileira 8 (SAMPAIO; VIEGAS, 2019). Tais características e singularidades, mobilizadas no debate público e espelhadas na literatura especializada, parecem ter obnubilado alguns elementos fundamentais para a análise das possibilidades de estudo do Ministério Público brasileiro como uma organização nacional. Um desses aspectos é, justamente, o fenômeno da pluralidade institucional, isto é, as implicações do modelo federativo brasileiro na análise do desempenho institucional ou funcional do MP, uma terminologia mobilizada já em trabalho anterior (LAMENHA; LIMA, 2021, p. 76- 77). A esse respeito, rememore-se o alerta de Maria Tereza Sadek (2000, p. 17), uma referência nos estudos sobre o MP de 1988, apontando que “a rigor, o que se refere ao Ministério Público dos estados, não se pode falar de um único modelo no qual estejam contidas todas as instituições estaduais. As variações regionais e estaduais, como observamos, são notáveis”. 41 41 6 No mesmo sentido, ver Nóbrega (2007, p. 62) e Macedo Júnior (1995, p. 38). Este último autor, por sinal, enfatiza que a Constituição de 1988 “deu à luz um novo Ministério Público, sem paralelo no mundo, forte, independente e incumbido da defesa da ordem jurídica, do regime democrático e dos interesses sociais e individuais indisponíveis”. 6 No mesmo sentido, ver Nóbrega (2007, p. 62) e Macedo Júnior (1995, p. 38). Este último autor, por sinal, enfatiza que a Constituição de 1988 “deu à luz um novo Ministério Público, sem paralelo no mundo, forte, independente e incumbido da defesa da ordem jurídica, do regime democrático e dos interesses sociais e individuais indisponíveis”. 6 No mesmo sentido, ver Nóbrega (2007, p. 62) e Macedo Júnior (1995, p. 38). 6 No mesmo sentido, ver Nóbrega (2007, p. 62) e Macedo Júnior (1995, p. 38). Este último autor, por sinal, enfatiza que a Constituição de 1988 “deu à luz um novo Ministério Público, sem paralelo no mundo, forte, independente e incumbido da defesa da ordem jurídica, do regime democrático e dos interesses sociais e individuais indisponíveis”. BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA ORCID 0000-0002-6655-2735 ORCID 0000-0002-6995-0982 BRUNO LAMENHA BRUNO LAMENHA Abrucio e outros (2021), em reflexão específica sobre este tema, apontam essa lacuna nas análises sobre o Ministério Público como “[...] a agenda esquecida do federalismo brasileiro”. É que, no quadro institucional da federação na CRFB/88, há um sistema de justiça em cada estado-membro, além de uma estrutura semelhante no âmbito federal e em matérias específicas — trabalhista, eleitoral e militar. Do ponto de vista orgânico, portanto, há 30 (trinta) Ministérios Públicos distintos no Estado brasileiro, isto é, trinta organizações formais diferentes, em que pese a existência de um marco normativo constitucional comum. Não há uma hierarquia entre os diferentes órgãos ministeriais e o Procurador-Geral da República não é o chefe de todo MP brasileiro9. Apesar dessa complexa paisagem institucional, alguns trabalhos de relevo na literatura tomam por referência a experiência de alguns ou poucos órgãos ministeriais, especialmente o Ministério Público Federal, para propor uma análise geral do desempenho institucional de todo o MP brasileiro. Este cenário se acentuou, mais recentemente, com a grande repercussão e impacto político- institucional da chamada Operação Lava Jato (ABRUCIO et al., 2021). O ponto aqui, ressalte-se, não é propor que haja um desconhecimento acerca do fato da pluralidade institucional na literatura especializada, mas problematizar a consideração e o peso desse elemento em algumas reflexões recentes sobre o Ministério Público brasileiro como organização e, sobretudo, sobre o seu desempenho institucional. O tema do giro de prioridades, por sinal, é um exemplo privilegiado da extensão analítica aqui referida. 42 Veja-se, a título de exemplo, o trabalho de Avritzer e Marona (2017) intitulado “A tensão entre soberania e instituições de controle na democracia brasileira”. Estabelecendo uma hipótese de trabalho segundo a qual houve uma guinada nas prioridades de atuação do MP brasileiro, os autores propõem que, [...] a partir dos anos 2000 a via criminal começou a parecer mais atrativa, particularmente depois da ‘reabilitação do prestígio do inquérito policial’, que veio a reboque do enorme incremento institucional que a polícia judiciária experimentou, particularmente a Polícia Federal e de inúmeras inovações legais e jurisprudenciais no campo do processo penal brasileiro. (AVRITZER; MARONA, 2017, p. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? Este último autor, por sinal, enfatiza que a Constituição de 1988 “deu à luz um novo Ministério Público, sem paralelo no mundo, forte, independente e incumbido da defesa da ordem jurídica, do regime democrático e dos interesses sociais e individuais indisponíveis”. 6 No mesmo sentido, ver Nóbrega (2007, p. 62) e Macedo Júnior (1995, p. 38). Este último autor, por sinal, enfatiza que a Constituição de 1988 “deu à luz um novo Ministério Público, sem paralelo no mundo, forte, independente e incumbido da defesa da ordem jurídica, do regime democrático e dos interesses sociais e individuais indisponíveis”. 7 José Paulo Sepúlveda Pertence, um dos pais fundadores do Ministério Público de 1988 e o primeiro Procurador-Geral da República do período democrático (1985/1987; 1987/1989) enfatiza o caráter singular do desenho institucional do MP brasileiro e a amplitude de suas atribuições: “[…] a Constituição de 1988 cria, a meu ver, um Ministério Público não só incomparável com o que existia no Brasil, sobretudo na área da União, mas também, dificilmente, com algum paralelo significativo no mundo. Tão seja pela armadura de independência e de autonomia que lhe foram conferidos, seja pelo plexo de atribuições confiadas ao Ministério Público. A partir de 88, que eu digo, creio que não haja nenhuma instituição do gênero no mundo que tenha este conjunto tão amplo de atribuições, de possibilidades de ação como tem hoje o Ministério Público Brasileiro. (PERTENCE, 2005, s.p.) 8 É nesta chave que um dos principais analistas do Ministério Público de 1988 na literatura nacional, Rogério Arantes (2002, p. 15) já propunha, vinte anos atrás, que o Ministério Público, no curso da institucionalização da democracia na Nova República, se tornou o principal agente da interação entre o sistema de justiça e o sistema político. 8 É nesta chave que um dos principais analistas do Ministério Público de 1988 na literatura nacional, Rogério Arantes (2002, p. 15) já propunha, vinte anos atrás, que o Ministério Público, no curso da institucionalização da democracia na Nova República, se tornou o principal agente da interação entre o sistema de justiça e o sistema político. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? A discussão sobre a pertinência da hipótese, contudo, é desenvolvida apenas com base em dados relativos à atuação da Polícia Federal (PF) e do MPF, tais como o quantitativo de operações policiais levadas a efeito pela PF e o número de ações de improbidade administrativa ajuizadas pelo Ministério Público Federal. Não há uma única consideração sobre a pluralidade institucional do MP brasileiro, tampouco sobre os limites e as possibilidades de extensão da análise referente a essa significativa alteração nas prioridades institucionais para outros órgãos ministeriais do sistema de justiça nacional. Além disso, mesmo quando a diversidade de órgãos ministeriais é apontada, inclusive com referência a dados específicos, por vezes, as conclusões havidas da reflexão não alertam para a importância da consideração de tais elementos na análise do desempenho institucional do MP brasileiro. É o caso, por exemplo, da versão preliminar do texto “O Ministério Público Brasileiro: de defensor de direitos à combatente da corrupção”, publicado por Kerche e Viegas (2020), no 12º Encontro da Associação Brasileira de Ciência Política (ABCP). Na sua apresentação, o artigo se propõe a [...] mostrar que o combate à corrupção não era a atribuição mais importante pensada para esse novo Ministério Público. Na verdade, o que o constituinte buscava primordialmente era uma agência, com autonomia em relação aos governos, encarregada da defesa de direitos sociais e coletivos. Ministério Público que se projeta como um combatente da corrupção é fruto de mudanças exógenas e endógenas que se construíram especialmente após 2003 (KERCHE; VIEGAS, 2020, s.p.). 43 43 O fato da pluralidade institucional é um elemento que claramente está no radar dos autores, seja porque há expressa menção a dados nacionais acerca dos órgãos ministeriais brasileiros, ou ainda porque, ao menos um deles tem tratado especificamente do tema em outras publicações recentes (SAMPAIO; VIEGAS, 2019; ABRUCIO et al., 2021; VIEGAS et al, 2022, p. 432-433). Apesar disso, esse trabalho em particular analisa dados do CNMP sobre produtividade judicial e extrajudicial exclusivamente do Ministério Público Federal – mesmo quando tais dados sobre os outros MPs também estão disponíveis. BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA 367) 9 Há 26 (vinte e seis) órgãos estaduais (MPEs), sem qualquer relação administrativa entre si; e os 4 (quatro) ramos do Ministério Público da União (MPU): Ministério Público Federal (MPF) o MPDFT (Ministério Público do Distrito Federal e Territórios), Ministério Público do Trabalho (MPT) e o Ministério Público Militar (MPM). No caso do MPU, embora o Procurador-Geral da República (PGR) seja o chefe administrativo dos quatro órgãos, somente atua de forma finalística no âmbito do MPF, de maneira que os outros três ramos também possuem Procuradores-Gerais específicos. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA comparados com o combate à corrupção. Tão importante quanto isso, o Ministério Público Federal não apenas prioriza o combate à corrupção como faz, de uma maneira geral, em detrimento de outros temas sensíveis para a população. É possível identificar como foi intensificado o combate à corrupção em momentos específicos, como as vésperas das eleições em 2013, em um dos momentos mais críticos da história política do país e que culminou com o impeachment de Dilma Rousseff (KERCHE; VIEGAS, 2020, s.p.). Outro recentíssimo texto, publicado por Milena Brentini Santiago (2022), na Revista de Ciências Sociais, incorre em imprecisão semelhante, que parece derivar da mesma raiz. O texto intitulado “O Ministério Público Federal e sua ‘vocação’ de combate à corrupção: conflitos sociais e ideologia institucional” utiliza os termos “Ministério Público” e “Ministério Público Federal” indistintamente, o que — em si — não é exatamente um problema, por haver uma delimitação no título. No entanto, em uma curta nota de rodapé, a autora consigna que “[...] o Ministério Público brasileiro é regido por um princípio de unidade que integra os diferentes ramos — o Ministério Público da União (MPU) e os Ministérios Públicos dos Estados (MPs) — em uma mesma e única instituição” (SANTIAGO, 2022, p. 109) — assumindo, portanto, que se trata de uma única organização 10 11. 44 p q Mesmo análises mais específicas, como o recente livro publicado por Kerche e Marona (2022), intitulado “A política no banco dos réus: a Operação Lava Jato e a erosão da democracia no Brasil” e que dedica uma seção inteira ao Ministério Público, não tematiza a questão da pluralidade institucional. Apesar da dimensão e do impacto do caso na conjuntura político-institucional brasileira, é sabido que o 10 A mesma imprecisão é reproduzida por Lemgruber et al. (2016, p. 66): “O princípio da unidade tampouco parece harmonizar-se com a estrutura federativa do país, bastando uma visita aos 27 sites dos MPEs e do MPDFT para verificar o quanto se diferencia a atuação dos membros do órgão nas várias UFs, em que pesem medidas recentes voltadas a produzir alguma homogeneidade na geração e divulgação de informações. A assimetria na distribuição dos recursos humanos estaduais – com taxas de variando de 3,5 a 13 promotores e procuradores por 100 mil habitantes – é mais um indício da falta de padrões básicos que justificariam considerar-se o MP uma instituição nacional una e indivisível”. 10 A mesma imprecisão é reproduzida por Lemgruber et al. (2016, p. 66): “O princípio da unidade tampouco parece harmonizar-se com a estrutura federativa do país, bastando uma visita aos 27 sites dos MPEs e do MPDFT para verificar o quanto se diferencia a atuação dos membros do órgão nas várias UFs, em que pesem medidas recentes voltadas a produzir alguma homogeneidade na geração e divulgação de informações. A assimetria na distribuição dos recursos humanos estaduais – com taxas de variando de 3,5 a 13 promotores e procuradores por 100 mil habitantes – é mais um indício da falta de padrões básicos que justificariam considerar-se o MP uma instituição nacional una e indivisível”. PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? Nas considerações finais, em que pese de tratar de um artigo pretensamente sobre o MP brasileiro (e não especificamente sobre o MPF), os pesquisadores se limitam a discutir os achados sobre o MP Federal: O que os relatórios mais atuais de atividades do Ministério Público Federal registram, contudo, é que os temas primordiais para os constituintes se tornaram, aparentemente, secundários quando 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 ORCID 0000-0002-6655-2735 ORCID 0000-0002-6995-0982 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA 11 É uníssono na literatura jurídica e na própria arquitetura constitucional do MP brasileiro que o princípio da unidade denota que os “[...] membros do Ministério Público integram um só órgão sob a direção de um só chefe (...) Nesse sentido, não há unidade (...) entre os membros de Ministério Público diversos, só dentro de cada Ministério Público e, assim mesmo, apenas nos limites da lei” (MAZZILLI, 2015, p. 72). Isso não se confunde com o caráter nacional do Ministério Público, que é justamente o fato dessa pluralidade de órgãos se submeter a um marco normativo comum, a despeito da inexistência de vínculo institucional. 11 É uníssono na literatura jurídica e na própria arquitetura constitucional do MP brasileiro que o princípio da unidade denota que os “[...] membros do Ministério Público integram um só órgão sob a direção de um só chefe (...) Nesse sentido, não há unidade (...) entre os membros de Ministério Público diversos, só dentro de cada Ministério Público e, assim mesmo, apenas nos limites da lei” (MAZZILLI, 2015, p. 72). Isso não se confunde com o caráter nacional do Ministério Público, que é justamente o fato dessa pluralidade de órgãos se submeter a um marco normativo comum, a despeito da inexistência de vínculo institucional. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? órgão ministerial a quem coube a persecução criminal dos acusados foi o Ministério Público Federal (MPF). Era de se esperar, portanto, que a análise do desempenho institucional do MP em uma reflexão como essa estivesse circunscrita, portanto, ao MP Federal. Em várias passagens, contudo, os autores parecem estender a sua reflexão a todo o complexo de órgãos do MP brasileiro, embora a atuação submetida a escrutínio tenha sido especificamente a do MPF no referido caso. Veja- se um dos trechos: Em resumo: o Ministério Público desenhado pelos constituintes e que foi o modelo que existiu até 2003 era autônomo externamente em relação aos outros Poderes e atores do Estado e, também, internamente, ostentando uma hierarquia frouxa e pouco efetiva. Sua discricionariedade era ampla em relação às matérias cíveis, especialmente em relação à preservação de direitos, e relativamente menor quando se tratava de ações penais. Neste tipo de atuação, seu desempenho era limitado pelo modelo “competitivo” da justiça criminal brasileira, que impunha o estabelecimento de relações com a polícia e o Poder Judiciário. Se a adoção daquele modelo “independente” de Ministério Público já transformava o órgão em um poderoso ator político, especialmente por sua capacidade de intervenção no campo das políticas públicas – ambientais, de saúde, de educação etc. –, as mudanças institucionais que ocorreram após 2003 colocaram os procuradores no epicentro da política nacional, facilitando a mudança de patamar da atuação, especialmente do MPF. (KERCHE; MARONA, 2022, s.p.) 45 45 O caráter nacional do MP Federal, sua maior visibilidade e, no caso específico da agenda anticorrupção, o grande impacto dos desdobramentos da Operação Lava Jato (VIEGAS et al., 2020) são alguns dos elementos que provavelmente explicam o porquê dessa extensão automática de achados/análises relacionados ao MPF para o MP brasileiro como um todo. Há ainda prováveis razões de fundo metodológico, nomeadamente aquelas relacionadas à assimetria na transparência e na disponibilização de dados específicos sobre cada órgão ministerial e também quanto ao peso político dos diferentes atores ministeriais no contexto institucional brasileiro. Em consonância com a preocupação aqui mobilizada, levantamentos já realizados, a exemplo do Diagnóstico do Ministério Público dos Estados, publicado em 2006, pelo Conselho Nacional de Procuradores-Gerais dos Ministérios Públicos dos Estados e da União (CNPG) 12, evidenciam que há “grandes diferenças na 12 Fundado em 1981, apesar do nome, o CNPG é uma associação privada integrada pelos Chefes dos Ministérios Públicos dos Estados e da União. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 12 Fundado em 1981, apesar do nome, o CNPG é uma associação privada integrada pelos Chefes dos Ministérios Públicos dos Estados e da União. De acordo com o seu sítio virtual, funcionando 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? De acordo com o seu sítio virtual, funcionando 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA ORCID 0000-0002-6655-2735 ORCID 0000-0002-6995-0982 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA estrutura desses órgãos e as disparidades entre os Estados no número de funcionários, promotores e no orçamento” (AZEVEDO, 2008, p. 241). Os dados compilados anualmente no CNMP no relatório anual “MP: um retrato”, também demonstram haver diferenças importantes entre os órgãos do MP brasileiro quanto à natureza e ao volume de trabalho relacionado à atividade finalística (BRASIL, 2021a). Sustenta-se, no entanto, que o pouco destaque dado à pluralidade institucional dos órgãos ministeriais, contudo, não invalida, em si, o estudo do Ministério Público brasileiro como organização nacional, tampouco a hipótese do giro de prioridades como fenômeno extensível a todo o Ministério Público brasileiro. Na esteira de Lamenha e Lima (2021, p. 77), propõe-se que o reconhecimento da ausência de vínculos institucionais formais não afasta, integralmente, a possibilidade de traçar uma trajetória institucional comum ao MP brasileiro baseada em alguns pontos de contato que permitem propor a existência de um certo núcleo partilhado entre a visão e os sentidos dos diferentes órgãos ministeriais acerca de sua identidade e suas prioridades de atuação. O fato do pluralismo institucional, no entanto, há de se impor, necessariamente, como pressuposto do modelo de análise, delimitando os limites e as possibilidades da extensão da experiência de um ou mais órgãos ministeriais para todo o complexo institucional que integra o MP brasileiro. 46 13 De 1971 a 1978, a entidade utilizou o nome Confederação das Associações Estaduais do Ministério Público (CAEMP). Em agosto de 1978, como ingresso dos ramos do MPU, o nome da Associação se alterou para Confederação Nacional do Ministério Público (CONAMP). Em junho de 2000, como parte de uma estratégia para ser reconhecida como ente legitimado a deflagrar a jurisdição constitucional, o nome foi alterado uma terceira vez para Associação Nacional do Ministério Público, mas manteve a sigla CONAMP por já ser amplamente difundida no MP brasileiro (CONAMP, 2022). 2.2. O MP PODE SER ESTUDADO COMO UMA ORGANIZAÇÃO NACIONAL? 2.2. O MP PODE SER ESTUDADO COMO UMA ORGANIZAÇÃO NACIONAL? Apesar de pouco numerosos, alguns trabalhos recentes demonstram a plena viabilidade de um empreendimento de pesquisa com foco na trajetória institucional do Ministério Público de 1988. Citar-se-ão dois, a título exemplificativo. Daiane Londero (2021) publicou tese de doutorado no Programa de Pós-Graduação em Políticas Públicas da Universidade Federal do Rio Grande do Sul sobre o desenvolvimento das capacidades institucionais do Ministério Público Federal (MPF) na área do combate à corrupção cobrindo o período de 1988 e 2021. Viegas e outros (2022), em publicação em dossiê específico sobre o Ministério Público na Revista Direito Público, investigam os efeitos da mudança do espaço formal do MP, nomeadamente quanto às alterações legislativas a ele associados, sobre a organização e seus membros no período entre 1989 e 2016. Esses esforços de pesquisa sugerem, portanto, o acerto da proposição do final do tópico anterior: o fato do pluralismo institucional, embora estabeleça cautelas e limites analíticos, não inviabiliza a possibilidade de estudo do MP brasileiro como na forma de um Colegiado, o Conselho foi criado com o objetivo de defender os princípios, prerrogativas e funções institucionais do Ministério Público. É sua função integrar os Ministérios Públicos de todos os Estados brasileiros; promover intercâmbio de experiências funcionais e administrativas; e trabalhar pelo aperfeiçoamento da instituição, traçando políticas e planos de atuação uniformes ou integrados que respeitem as peculiaridades regionais” (CNPG, 2022). 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? organização nacional e de sua trajetória institucional. Propõe-se aqui quatro elementos que evidenciam a existência de pontos de contato entre os diferentes órgãos ministeriais da paisagem institucional brasileira, uma espécie de tronco simbólico comum, a despeito da ausência de vinculação funcional entre eles. Primeiro, apesar das implicações decorrentes do modelo federativo estabelecido na Constituição de 1988, há um marco normativo único para o MP brasileiro: o próprio texto constitucional, que estabelece um só perfil institucional para todos os órgãos ministeriais, e a legislação infraconstitucional que, embora se desdobre em uma lei orgânica específica para o MP da União e seus quatro ramos (LC 75/93) e outra para os MPs estaduais (Lei n. 8.625/93), consegue manter uma relativa unidade organizacional e procedimental para os 30 (trinta) órgãos ministeriais existentes no País (LAMENHA; LIMA, 2021, p.77). ( p ) Segundo, há um aspecto relacionado à dimensão corporativa. Apesar de existirem associações vinculadas a cada um dos órgãos do MP brasileiro, também há organizações associativas de alcance nacional e que recorrentemente mobilizam pautas institucionais únicas — é o caso, principalmente, da Associação Nacional do Ministério Público (CONAMP)13, que reúne associações de todos os órgãos do MP brasileiro, e do Conselho Nacional de Procuradores-Gerais do Ministério Público dos Estados e da União (CNPG)14, que é composto pelos chefes de todos as trinta organizações ministeriais existentes no País. Tais espaços não se reduzem à articulação de uma agenda corporativa e/ou de um lobby ministerial frente a outros órgãos do quadro institucional da República, mas também operam como fóruns de troca de informações e boas práticas, e até mesmo de promoção de uma certa educação corporativa e formação de quadros, através da realização de eventos, como, por exemplo, o Congresso Nacional do MP, realizado anualmente pela CONAMP desde 1986 (PENNA, 2021, p. 234-238). De mais a mais, os processos institucionais que tornaram possível a construção do desenho institucional do MP de 1988 contaram, em grande medida, com a decisiva influência da atuação de tais 47 47 13 De 1971 a 1978, a entidade utilizou o nome Confederação das Associações Estaduais do Ministério Público (CAEMP). Em agosto de 1978, como ingresso dos ramos do MPU, o nome da Associação se alterou para Confederação Nacional do Ministério Público (CONAMP). 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? Em junho de 2000, como parte de uma estratégia para ser reconhecida como ente legitimado a deflagrar a jurisdição constitucional, o nome foi alterado uma terceira vez para Associação Nacional do Ministério Público, mas manteve a sigla CONAMP por já ser amplamente difundida no MP brasileiro (CONAMP, 2022). 14 Fundado em 1981, apesar do nome, o CNPG é uma associação privada integrada pelos Chefes dos Ministérios Públicos dos Estados e da União. De acordo com o seu sítio virtual, funcionando na forma de um Colegiado, o Conselho foi criado com o objetivo de defender os princípios, prerrogativas e funções institucionais do Ministério Público. É sua função integrar os Ministérios Públicos de todos os Estados brasileiros; promover intercâmbio de experiências funcionais e administrativas; e trabalhar pelo aperfeiçoamento da instituição, traçando políticas e planos de atuação uniformes ou integrados que respeitem as peculiaridades regionais” (CNPG, 2022) 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA ORCID 0000-0002-6655-2735 ORCID 0000-0002-6995-0982 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA entidades corporativas desde meados dos anos 1970, com especial destaque para a própria CONAMP (SABELLA et al., 2013, p. 36-37; LIMA; LAMENHA, 2021, p. 94). entidades corporativas desde meados dos anos 1970, com especial destaque para a própria CONAMP (SABELLA et al., 2013, p. 36-37; LIMA; LAMENHA, 2021, p. 94). p p p p Terceiro, merece menção o papel do Conselho Nacional do Ministério Público (CNMP), criado pela EC 45/04, como órgão de controle externo do MP brasileiro e dotado, entre outras atribuições, a expedição de atos regulamentares e recomendações relacionadas à preservação da autonomia funcional e administrativa do Ministério Público (CRFB, art. 130-A, §2º, I). Na prática, nos últimos anos, esta função particular do CNMP tem se dirigido à promoção de uma padronização institucional no âmbito dos órgãos ministeriais (OLIVEIRA et al., 2020, p. 190). Ver-se-á mais adiante, inclusive, como essa atividade regulamentadora e indutora do CNMP tem se dirigido, por exemplo, para a difusão de práticas e rotinas de trabalho no campo anticorrupção. Quarto, como destacam Faveri e outros (2014, p. 15 Em linhas gerais e no sentido particular que se quer traduzir aqui, a noção de isomorfismo está associada à ideia de que organizações submetidas a ambientes sociais, políticos e normativos semelhantes tendem se a assumir configurações, identidades e estratégias de atuação similares. Embora tal princípio opere em diferentes níveis e atores sociais, as organizações constituem um campo privilegiado para a sua observação, dado que vários elementos, como a busca por legitimidade social, a necessidade de lidar com as incertezas da vida social, as pressões por profissionalização e a aspiração à perenidade estimulam a assunção de uma postura de homogeneização frente a outras organizações e práticas. (DIMAGGIO; POWELL, 1983, p. 149-154; SCOTT, 2014, p. 185). 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? anos optou pela adoção de um modelo diferente, baseado em Forças-Tarefas temporárias, passou a adotar, a partir de 2020, a estrutura de GAECO em diversos estados (PALUDO, 2011, p. 28; BRASIL, 2021b). anos optou pela adoção de um modelo diferente, baseado em Forças-Tarefas temporárias, passou a adotar, a partir de 2020, a estrutura de GAECO em diversos estados (PALUDO, 2011, p. 28; BRASIL, 2021b). ( p ) A despeito da pluralidade institucional, portanto, esses quatro pontos de contato acima descritos evidenciam haver espaço para uma análise de trajetória comum ao MP brasileiro e para o seu estudo como uma organização nacional, desde que estabelecida uma lente baseada em elementos comuns à experiência desses diferentes órgãos. Esse esforço de pesquisa, e a própria identificação de um padrão de isomorfismo entre tais organizações, não significam, contudo, equivalência de infraestrutura, recursos humanos, volume de trabalho ou o mesmo padrão de interação com organizações e agentes individuais próximos, como o Judiciário (AZEVEDO, 2008, p. 241; FAVERI e et al., 2014, p. 2). Assim, essa linha analítica não exclui a necessidade de abordagens sobre a experiência particular dos Ministérios Públicos isoladamente considerados. Ao contrário, apenas a reforça. Definir um terreno de regularidades partilhadas por todo o MP brasileiro pode oferecer condições analíticas fecundas para a identificação e o desenvolvimento de estudos sobre o que é específico na trajetória de cada órgão ministerial. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? 2), há um isomorfismo 15 institucional significativo entre os diferentes órgãos do MP brasileiro, ou seja, considerando que se trata de organizações submetidas a um ambiente institucional semelhante, os diferentes MPs existentes na federação brasileira tendem a assumir configurações, identidades e estratégias de atuação semelhantes (DIMAGGIO; POWELL, 1983, p. 151-154; SCOTT, 2014, p. 185)16. Isso explica, por exemplo, o porquê da replicação de estruturas operacionais e de práticas reputadas como exitosas antes mesmo da existência do CNMP e de sua agenda de padronização institucional. 48 Um bom exemplo aqui são os Grupos de Atuação Especial de Combate ao Crime Organizado (GAECOs), estruturas especializadas para o enfrentamento da macro criminalidade e em coordenação com outros órgãos de controle, cuja primeira experiência foi desenvolvida no MPSP, no ano de 1995. A partir desses marcos, esse saber-fazer passou a ser difundido e adaptado às diferentes realidades dos MPs e, hoje, quase todos os órgãos ministeriais estaduais, hoje, contam com uma estrutura especializada semelhante. Até mesmo o MP Federal que, durante 16 Lamenha e Lima (2021) discutem como até mesmo organizações à primeira vista bastante distintas, como o MP e a Defensoria Pública, acabam emulando elementos uma da outra. A competição institucional, portanto, é uma chave analítica importante para pensar a questão do isomorfismo. Sobre o tema, ver DiMaggio e Powell (1983). 17Neste período, portanto, as mudanças institucionais e o novo perfil institucional da organização ganharam “expressão no debate público mediante a seguinte formulação: o Ministério Público deixou de ser defensor do Estado para ser defensor da sociedade” (ARANTES, 1999, p. 90). A esse respeito, Maria Tereza Sadek (2010, pos. 327) consignou que “o novo estatuto legal do Ministério Público representou, a rigor, um ganho para os setores mais progressistas da instituição. Assim, estes têm procurado dar publicidade às suas novas funções e à importância de seu papel como guardiães de uma ordem democrática e justa”. 3. UM BREVE PANORAMA DA CONJUNTURA EM QUE TERIA SE DESENVOLVIDO O FENÔMENO DO GIRO DE PRIORIDADES O que se denominou acima como giro de prioridades é uma espécie de mudança institucional bastante distinta de fenômenos como a própria construção do marco normativo que culminou na disciplina do Ministério Público na Constituição de 1988 ou a criação de um órgão de controle externo do MP. Estes dois últimos processos de mudança institucional, como é sabido, foram associados a reformas constitucionais e/ou legislativas no âmbito do Parlamento. No caso do giro de prioridades, por sua vez, trata-se de um fenômeno apontado por alguns setores da literatura especializada e ainda de forma um tanto assistemática. 49 49 Como se expôs na introdução deste trabalho e nos excertos colacionados no item 2.1 acima, esse fenômeno apontaria para a existência de certos elementos indicativos de uma particular alteração dos discursos, das atividades finalísticas e, de certa forma, da identidade institucional da organização tomada como entidade nacional, direcionando-os de uma perspectiva mais voltada para a defesa de direitos transindividuais para o campo da repressão penal anticorrupção. É interessante observar que, ainda em 2011, Rogério Arantes, em um texto sobre a construção institucional da Polícia Federal (PF) no pós-1988, é o primeiro a apontar para o fenômeno, embora não na abrangência posteriormente proposta por outros pesquisadores. Neste ponto em particular, o autor argumenta que a expansão do fenômeno da repressão penal anticorrupção, verificado a partir do ano de 2003 e que teve na PF um dos seus protagonistas, é explicável a partir de alguns deslocamentos havidos no “[...] arranjo institucional que congrega ações de combate à corrupção e ao crime organizado”. Um deles seria um deslocamento, 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA ORCID 0000-0002-6655-2735 ORCID 0000-0002-6995-0982 BRUNO LAMENHA capitaneado pelo Ministério Público, de uma atuação temática no campo da repressão anticorrupção que privilegiaria a esfera cível e o manejo da ação de improbidade administrativa e que passara, justamente, para o âmbito criminal, que seria mais efetivo e disporia de mais recursos investigativos à disposição (ARANTES, 2011, p. 99, 122-123). Em que condições, todavia, se pode afirmar que, até meados do ano 2000, os órgãos ministeriais brasileiros privilegiavam, de fato, a atuação no campo dos direitos transindividuais? 3. UM BREVE PANORAMA DA CONJUNTURA EM QUE TERIA SE DESENVOLVIDO O FENÔMENO DO GIRO DE PRIORIDADES Ainda que não estejam disponíveis dados nacionais sistematizados sobre a atuação judicial e extrajudicial dos órgãos ministeriais naquele período, propõe-se que há elementos de conjuntura que indicam, no período imediatamente após 1988, uma efetiva predominância de discursos e a construção de uma identidade institucional profundamente orientada para a atuação no campo da tutela coletiva no MP brasileiro (SADEK, 2000, p. 30). Isso pode ser creditado, entre algumas razões, à novidade em si associada às significativas mudanças institucionais advindas da constitucionalização do desenho institucional do chamado novo MP e ainda à grande visibilidade conferida ao tema da defesa dos direitos transindividuais no período (SILVA, 2001, p. 163)17. Em que condições, todavia, se pode afirmar que, até meados do ano 2000, os órgãos ministeriais brasileiros privilegiavam, de fato, a atuação no campo dos direitos transindividuais? Ainda que não estejam disponíveis dados nacionais sistematizados sobre a atuação judicial e extrajudicial dos órgãos ministeriais naquele período, propõe-se que há elementos de conjuntura que indicam, no período imediatamente após 1988, uma efetiva predominância de discursos e a construção de uma identidade institucional profundamente orientada para a atuação no campo da tutela coletiva no MP brasileiro (SADEK, 2000, p. 30). Isso pode ser creditado, entre algumas razões, à novidade em si associada às significativas mudanças institucionais advindas da constitucionalização do desenho institucional do chamado novo MP e ainda à grande visibilidade conferida ao tema da defesa dos direitos transindividuais no período (SILVA, 2001, p. 163)17. Uma evidência particularmente interessante advém do survey “O Ministério Público e a Justiça no Brasil”, coordenado pela própria Sadek e por Rogério Arantes, e cuja base de dados se esteia em entrevistas realizadas com membros do MP de sete estados da Federação no ano de 199618. Uma das questões apresentadas aos inquiridos solicitou que eles indicassem a sua avaliação sobre as prioridades temáticas na atuação do órgão em que atuam nos últimos 2 (dois) anos e as expectativas sobre as prioridades temáticas que prevaleceram nos próximos 2 (dois) anos. Na esteira do que propõe Arantes (2002, p. 116-119), a análise global desses dois grupos de dados oferece um panorama amplo da proeminência Uma evidência particularmente interessante advém do survey “O Ministério Público e a Justiça no Brasil”, coordenado pela própria Sadek e por Rogério Arantes, e cuja base de dados se esteia em entrevistas realizadas com membros do MP de sete estados da Federação no ano de 199618. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 3. UM BREVE PANORAMA DA CONJUNTURA EM QUE TERIA SE DESENVOLVIDO O FENÔMENO DO GIRO DE PRIORIDADES Uma das questões apresentadas aos inquiridos solicitou que eles indicassem a sua avaliação sobre as prioridades temáticas na atuação do órgão em que atuam nos últimos 2 (dois) anos e as expectativas sobre as prioridades temáticas que prevaleceram nos próximos 2 (dois) anos. Na esteira do que propõe Arantes (2002, p. 116-119), a análise global desses dois grupos de dados oferece um panorama amplo da proeminência 50 17Neste período, portanto, as mudanças institucionais e o novo perfil institucional da organização ganharam “expressão no debate público mediante a seguinte formulação: o Ministério Público deixou de ser defensor do Estado para ser defensor da sociedade” (ARANTES, 1999, p. 90). A esse respeito, Maria Tereza Sadek (2010, pos. 327) consignou que “o novo estatuto legal do Ministério Público representou, a rigor, um ganho para os setores mais progressistas da instituição. Assim, estes têm procurado dar publicidade às suas novas funções e à importância de seu papel como guardiães de uma ordem democrática e justa”. 18 “A pesquisa ‘O Ministério Público e a Justiça no Brasil’ faz parte de um amplo programa de estudos sobre a justiça que o IDESP iniciou em 1993, com uma pesquisa junto aos juízes, já publicada. (...) O questionário foi elaborado tendo em vista responder a cinco indagações básicas: qual é o perfil dos integrantes do Ministério Público, como eles avaliam a distribuição de justiça no país, como percebem as mudanças introduzidas pela Constituição de 1988 e, por fim, como poderiam ser caracterizados em termos de posição ideológica (...) fazem parte dessa pesquisa os estados de Goiás, São Paulo, Rio de Janeiro, Paraná Rio Grande do Sul, Sergipe e Bahia” (SADEK, 1996, p. 29). 18 “A pesquisa ‘O Ministério Público e a Justiça no Brasil’ faz parte de um amplo programa de estudos sobre a justiça que o IDESP iniciou em 1993, com uma pesquisa junto aos juízes, já publicada. (...) O questionário foi elaborado tendo em vista responder a cinco indagações básicas: qual é o perfil dos integrantes do Ministério Público, como eles avaliam a distribuição de justiça no país, como percebem as mudanças introduzidas pela Constituição de 1988 e, por fim, como poderiam ser caracterizados em termos de posição ideológica (...) fazem parte dessa pesquisa os estados de Goiás, São Paulo, Rio de Janeiro, Paraná Rio Grande do Sul, Sergipe e Bahia” (SADEK, 1996, p. 29). 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? assumida pelo tema da defesa dos direitos transindividuais na visão dos próprios promotores e procuradores, a partir dos seguintes achados: i) a identificação do foco prioritário na matéria criminal decresceu 11% (onze por cento) da avaliação da atuação passada para a expectativa futura; ii) no mesmo comparativo, todos os campos associados à tutela de direitos transindividuais verificaram incremento, com destaque para o meio ambiente (+11%), minorias étnicas (+13%), pessoas com deficiência (+13%), patrimônio histórico e cultural (14%), idosos (+15%) serviços de relevância pública (+17%) e a improbidade administrativa 19 (+22%) 20. Veja-se a tabela abaixo: 19 O survey de 1996 trata o controle da administração pública no campo dos direitos difusos, coletivos e individuais homogêneos, não o inserindo no âmbito criminal. Na época, a visibilidade da atuação anticorrupção no âmbito do MP se dava na sua projeção cível, com aplicação da recente Lei n. 8.429/92 (Lei de Improbidade Administrativa – LIA), havendo alguns entraves na repressão penal anticorrupção – embora ela também existisse, em menor proporção, à época (LAMENHA; LIMA, 2021, p. 91-92). Voltar-se-á ao tema mais adiante. 20 É curioso observar, ademais, que essa visão da priorização da atuação no campo da tutela coletiva nos anos imediatamente seguintes à CFRB/88 não era compartilhada apenas no âmbito interno ou de interlocutores simpáticos a essa guinada na atuação finalística do MP. Mesmo os que empreendiam uma leitura crítica da atuação ministerial na área dos direitos coletivos reconheciam, ainda que de forma tácita, o destaque que esse tipo de atividade vinha assumindo comparativamente às atribuições ministeriais clássicas no processo penal. Veja-se, nesse sentido, manifestação da professora Ada Pellegrini Grinover, no ano de 1996, em evento sobre o papel do MP no sistema de justiça brasileiro: “[...] Eu não tenho dúvida de que as funções atribuídas ao Ministério Público pela Constituição de 1998 são funções da mais elevada importância, e que por isso mesmo o órgão adquiriu uma relevância que não tinha tido até então, mas sem dúvida nenhuma, pelo menos no início da vigência da Constituição de 1988, eu pude observar, externamente, como consumidora da justiça, um certo deslumbramento, digamos assim, do Ministério Público em relação a essa titularidade das ações coletivas, com relação a essas funções de promoção de direitos sociais, de interesses sociais etc... 3. UM BREVE PANORAMA DA CONJUNTURA EM QUE TERIA SE DESENVOLVIDO O FENÔMENO DO GIRO DE PRIORIDADES 10, n. 1, p. 37 - 86, jan./abr. 2024 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 21 Levantamento semelhante também foi realizado, no ano de 1997, no Ministério Público Federal, alcançando resultados muito parecidos. e todas as atribuições apresentadas aos membros entrevistados, somente houve decréscimo entre a avaliação de prioridade passada e expectativa de prioridade futura nas funções clássicas do velho Ministério Público de 1964-1985: crime (- 11,5%) e atuação como fiscal da lei (-6,5%). Em todas as áreas relacionadas à defesa de direitos difusos, coletivos e individuais homogêneos, foi verificado algum incremento, com destaque para as áreas de controle da administração pública (improbidade administrativa) e defesa do consumidor, ambas com aumento de 27%, serviços de relevância pública (saúde, educação, lazer, segurança, etc.), com acréscimo de 25%, idosos (+22,5%), pessoas com deficiência (+21,5%) e minorias étnicas (20%). (CASTILHO; SADEK, 1998, p. 29). 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? Eu diria que houve, sobretudo no início, pelo menos aqui em São Paulo, no Ministério Público que eu mais conheço, uma certa tendência à socialização da demanda, que é relevantíssima, sem dúvida nenhuma, mas que preponderou em excesso com relação àquela que continua sendo a função essencial do Ministério Público: a condução da ação penal. E digo que a considero essencial por uma razão muito simples. Para a ação penal, o Ministério Público tem um monopólio; ele tem a titularidade exclusiva da ação penal pública e, portanto, a sua responsabilidade avulta em relação a esse ponto, enquanto no exercício das demais atribuições, a própria Constituição e as leis com relação às ações civis públicas, estabelecem uma legitimidade concorrente. Mas não é só isso. Eu diria que principalmente nesse início de novas atividades do Ministério Público, houve sim um exagero. Houve um excesso que significou, a meu ver, uma banalização da atuação do Ministério Público nesse tipo de ação. E isso foi muito grave, foi muito sério. (...) (LAMOUNIER et al., 1996, p. 13). Os destaques são dos autores. 51 51 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 ORCID 0000-0002-6655-2735 ORCID 0000-0002-6995-0982 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA ORCID 0000-0002-6995-0982 Tabela 1 – Áreas de prioridade pretérita e futura - MPEs21 Atribuição Prioridade – último 2 anos Prioridade – próximos 2 anos Diferença Crime 72% 61% -11% Criança e adolescente 41% 42% +1% Improbidade administrativa 38% 60% +22% Serviços de relevância pública (saúde, educação etc.) 36% 53% +17% Patrimônio público e social 35% 45% +10% Controle externo da atividade policial 34% 44% +10% Meio ambiente 32% 43% +11% Interesses individuais 31% 30% -1% Consumidor 29% 38% +9% Pessoas com deficiência 20% 33% +13% Patrimônio histórico e cultural 16% 30% +14% Idosos 16% 31% +15% Minorias étnicas 8% 21% +13% Fonte: Autor (2023) com base nos dados de Sadek (1996, p. 47-59) 52 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? Pode-se assumir, portanto, que – apesar do expressivo volume de trabalho decorrente da condição de titular da ação penal pública, nos anos imediatamente posteriores a promulgação da CRFB/88 –, o Ministério Público angariou uma extensa visibilidade e um significativo capital político a partir de sua atuação no campo da tutela coletiva. O que resta indagar a essa altura é, justamente, o que teria tornado possível que, a partir de meados dos anos 2000, um processo de mudança institucional precipitasse um deslocamento da priorização do campo dos direitos coletivos para a repressão penal anticorrupção? Dentre vários elementos apontados pela literatura especializada e dado o espaço limitado de discussão neste texto, discutir-se-ão dois pontos, um de caráter externo e outro de caráter interno ao Ministério Público: i) o desenvolvimento de uma agenda de mobilização do direito penal para o enfrentamento da corrupção política no âmbito do sistema de justiça brasileiro (ARANTES, 2011, p. 99; AVRITZER e MARONA, 2017, p. 366; ENGELMANN, 2021, p. 54-55); ii) o papel regulamentador do CNMP na atuação anticorrupção e seus efeitos no isoformismo de modelos institucionais 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 3.1. A EMERGÊNCIA DE UMA AGENDA E UMA POLÍTICA INTERNA ANTICORRUPÇÃO A PARTIR DE MEADOS DOS ANOS 2000 Um marco aqui é a criação da Controladoria-Geral da União, pela Medida Provisória n. 103, de 1º de janeiro de 200324. O efeito isomórfico desse evento no âmbito dos estados-membros Corrupção (1996), internalizada pelo Brasil em 2002, a Convenção sobre Corrupção de Funcionários Públicos em Transações Comerciais Internacionais, aprovada pela OCDE em 1997 e internalizada no Brasil em 2000 (Decreto n. 3.678), e a a Convenção da ONU contra a Corrupção, adotada em 2003, e internalizada no Brasil em 2006 (Decreto n. 5.687). Não se propõe que a conjuntura nacional pode ser explicada como mera replicação dos discursos e práticas internacionais sobre a temática. Sustenta-se, no entanto, a possibilidade de convergência entre as agendas externa e interna sobre a questão da corrupção política. No plano interno, por exemplo, essa pauta global encontrara ressonância, por exemplo, junto ao esforço de modernização da Administração Pública levado a efeito desde a década de 1980 e que passa, entre outros pontos, pela compreensão da necessidade de aperfeiçoamento de mecanismos de controle de fiscalização e controle da Administração (ABRUCIO, 2007, p. 69, 79; FILGUEIRAS; ARANHA, 2011, p. 362). Para uma reflexão específica sobre o tema, ver o recente trabalho de Engelmann (2021). Corrupção (1996), internalizada pelo Brasil em 2002, a Convenção sobre Corrupção de Funcionários Públicos em Transações Comerciais Internacionais, aprovada pela OCDE em 1997 e internalizada no Brasil em 2000 (Decreto n. 3.678), e a a Convenção da ONU contra a Corrupção, adotada em 2003, e internalizada no Brasil em 2006 (Decreto n. 5.687). Não se propõe que a conjuntura nacional pode ser explicada como mera replicação dos discursos e práticas internacionais sobre a temática. Sustenta-se, no entanto, a possibilidade de convergência entre as agendas externa e interna sobre a questão da corrupção política. No plano interno, por exemplo, essa pauta global encontrara ressonância, por exemplo, junto ao esforço de modernização da Administração Pública levado a efeito desde a década de 1980 e que passa, entre outros pontos, pela compreensão da necessidade de aperfeiçoamento de mecanismos de controle de fiscalização e controle da Administração (ABRUCIO, 2007, p. 69, 79; FILGUEIRAS; ARANHA, 2011, p. 362). Para uma reflexão específica sobre o tema, ver o recente trabalho de Engelmann (2021). 54 23 A esse propósito, relatando as “Origens da ENCCLA”, Antenor Madruga (ENCCLA, 2012, p. 3.1. A EMERGÊNCIA DE UMA AGENDA E UMA POLÍTICA INTERNA ANTICORRUPÇÃO A PARTIR DE MEADOS DOS ANOS 2000 34), diretor do DRCI entre 2004 e 2007, relata que: ö desafio era criar uma política pública que tornasse efetivo o combate à lavagem de dinheiro no Brasil e, consequentemente, o sequestro e perdimento dos ativos instrumentos e produtos de atividade criminosa. O objetivo principal era dar meios ao Estado para combater o crime praticado por organizações, cuja existência e atuação ultrapassavam e independiam dos indivíduos que as integravam. O ano era 2003 e o recém-empossado Ministro da Justiça, Márcio Thomaz Bastos, estava decidido a aplicar no Brasil o consenso que já se formava em vários países e organizações internacionais, no sentido de que a “asfixia patrimonial” pelo desapossamento e expropriação dos recursos materiais que constituem o “fundo de comércio” da empresa criminosa é medida tão ou mais importante que a restrição à liberdade ou outras sanções aplicáveis aos seus recursos humanos”. 24 Posteriormente convertida na Lei n. 10.683/03. Em seu sítio virtual, o órgão se define como um 24 Posteriormente convertida na Lei n. 10.683/03. Em seu sítio virtual, o órgão se define como uma “[...] típica agência anticorrupção do país, é o órgão encarregado de assistir direta e imediatamente ao Presidente da República no desempenho de suas atribuições quanto aos assuntos que, no âmbito do Poder Executivo, sejam relativos à defesa do patrimônio público e ao incremento da transparência da gestão, por meio das atividades de controle interno, auditoria pública, correição, prevenção e combate à corrupção, e ouvidoria. A CGU é ainda órgão central do Sistema de Controle Interno e do Sistema de Correição, ambos do Poder Executivo Federal” (BRASIL, 2023a). Sua vocação de atuar “em rede”, portanto, advém de seu desenho institucional, visto que o órgão opera como um centro gravitacional do sistema de controle interno da Administração Pública da União, fazia parte de suas atribuições principais a interação constante com outras agências de accountability no âmbito federal (ARANHA, 2015, p. 137). g “[...] típica agência anticorrupção do país, é o órgão encarregado de assistir direta e imediatamente ao Presidente da República no desempenho de suas atribuições quanto aos assuntos que, no âmbito do Poder Executivo, sejam relativos à defesa do patrimônio público e ao incremento da transparência da gestão, por meio das atividades de controle interno, auditoria pública, correição, prevenção e combate à corrupção, e ouvidoria. 3.1. A EMERGÊNCIA DE UMA AGENDA E UMA POLÍTICA INTERNA ANTICORRUPÇÃO A PARTIR DE MEADOS DOS ANOS 2000 Redes de accountability derivam da paisagem institucional de uma dada comunidade política, reunindo organizações estatais cuja atribuição é a fiscalização e o controle do desempenho funcional de outras agências e funcionários públicos (MAINWARING, 2003, p. 15). Assim, por exemplo, no âmbito federal, é possível se pensar em uma rede envolvendo Polícia Federal, Receita Federal, Tribunal de Contas da União, Advocacia-Geral da União, Ministério Público Federal, entre outros órgãos. Essas entidades atuam em rede porque, independentemente da sua disposição, com maior ou menor frequência, elas irão manter relações interinstitucionais — de cooperação, competição e competição. É possível se deduzir, portanto, que há dois tipos de interação nessas redes de accountability: formais, através do estabelecimento de algum instrumento que regule a relação entre duas ou mais organizações, como um convênio ou termo de cooperação técnica; e, sobretudo, informais, estabelecida nos contatos profissionais entre agentes que atuam em áreas ou tópicos similares (ARANHA, 2015, p. 189; LONDERO, 2021, p. 172). 53 53 A partir de 200322, com a chegada do Partido dos Trabalhadores (PT) à chefia do Executivo Nacional, verificam-se sensíveis mudanças nesse contexto, 22 Ressalte-se que há elementos que sugerem, justamente no início dos anos 2000, a consolidação de uma certa agenda global anticorrupção, referente à difusão de discursos e modelos institucionais construídos, sobretudo, a partir de atores e mecanismos transnacionais, que pode ser sintetizado na ideia que compreende o estabelecimento de uma política pública de prevenção e repressão anticorrupção como um sinônimo ou pré-requisito para a boa governança de Estados e corporações privadas (MCCOY; HECKEL, p. 85-86). Nessa chave, importantes documentos internacionais sobre o tema foram firmados, como a Convenção Interamericana Contra a 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA ORCID 0000-0002-6655-2735 ORCID 0000-0002-6995-0982 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA possivelmente correlacionadas com a ambiência internacional favorável, a difusão de um saber-fazer relativamente homogêneo e a assunção, pelo Brasil, de compromissos internacionais na temática anticorrupção 23. Um marco aqui é a criação da Controladoria-Geral da União, pela Medida Provisória n. 103, de 1º de janeiro de 200324. O efeito isomórfico desse evento no âmbito dos estados-membros possivelmente correlacionadas com a ambiência internacional favorável, a difusão de um saber-fazer relativamente homogêneo e a assunção, pelo Brasil, de compromissos internacionais na temática anticorrupção 23. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 25 Analisando uma série histórica de 1995 a 2008, Suzart e outros (2011, p. 49) verificaram que, final do período, havia 14 (quatorze) órgãos com funções típicas de controladoria na União e nos estados-membros da Federação, incluída a própria CGU. Somente três preexistiam à criação da CGU (Rondônia, Sergipe e Rio Grande do Norte). As outras nove, portanto, foram criadas após a CGU, e 5 (cinco) delas no próprio ano de 2003. Esse cenário sugere a hipótese de que a temática anticorrupção poderia reproduzir um padrão de espelhamento semelhante em outras áreas, e que decorre do caráter centralizador do federalismo brasileiro (LIZIERO; CARVALHO, 2018, p. 1501): a União, por deter mais recursos materiais e humanos, acabaria sendo modelo para a institucionalização de determinadas políticas públicas. PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? é notável, com a replicação do modelo por vários estados da Federação nos anos seguintes25. é notável, com a replicação do modelo por vários estados da Federação nos anos seguintes25. Outras iniciativas exitosas e até hoje vigentes datam do período, com destaque para duas que se estabeleceram na estrutura do Ministério da Justiça. A primeira é a criação do Departamento de Recuperação de Ativos de Cooperação Jurídica Internacional (DRCI), criado em 2004, especializado em cooperação jurídica internacional em matéria penal e autoridade central no País de vários tratados internacionais e que o Brasil é parte. A segunda é a Estratégia Nacional de Combate à Corrupção e Lavagem de Dinheiro (ENCCLA), criada em 2003, sob a coordenação da Secretaria Nacional de Justiça, funcionam há 20 (vinte) anos, como um espaço articulador de inúmeras redes formais e informais entre organizações estatais, dos três Poderes e dos três níveis da Federação, e entidades da sociedade civil, para o “[...] arranjo, discussões, formulação e concretização de políticas públicas e soluções de enfrentamento à corrupção e à lavagem de dinheiro” (ENCCLA, 2023). Conta, atualmente, com mais de 80 (oitenta) participantes e estabelece metas e ações de vigência avaliadas anualmente em sua Plenária, na qual podem participar todas as organizações ligadas ao coletivo. Mantém iniciativas regulares de capacitação de agentes públicos 26, inclusive com ampla internacionalização, e troca não só de expertise, mas de tecnologia na atuação na área da lavagem de dinheiro e repressão anticorrupção27. Marona e Kerche (2021, p. 8) propõem que a ENCCLA é um fator- 55 55 25 Analisando uma série histórica de 1995 a 2008, Suzart e outros (2011, p. 49) verificaram que, final do período, havia 14 (quatorze) órgãos com funções típicas de controladoria na União e nos estados-membros da Federação, incluída a própria CGU. Somente três preexistiam à criação da CGU (Rondônia, Sergipe e Rio Grande do Norte). As outras nove, portanto, foram criadas após a CGU, e 5 (cinco) delas no próprio ano de 2003. Esse cenário sugere a hipótese de que a temática anticorrupção poderia reproduzir um padrão de espelhamento semelhante em outras áreas, e que decorre do caráter centralizador do federalismo brasileiro (LIZIERO; CARVALHO, 2018, p. 1501): a União, por deter mais recursos materiais e humanos, acabaria sendo modelo para a institucionalização de determinadas políticas públicas. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 3.1. A EMERGÊNCIA DE UMA AGENDA E UMA POLÍTICA INTERNA ANTICORRUPÇÃO A PARTIR DE MEADOS DOS ANOS 2000 A CGU é ainda órgão central do Sistema de Controle Interno e do Sistema de Correição, ambos do Poder Executivo Federal” (BRASIL, 2023a). Sua vocação de atuar “em rede”, portanto, advém de seu desenho institucional, visto que o órgão opera como um centro gravitacional do sistema de controle interno da Administração Pública da União, fazia parte de suas atribuições principais a interação constante com outras agências de accountability no âmbito federal (ARANHA, 2015, p. 137). 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? 26 Veja-se, por exemplo, o Programa Nacional de Capacitação e Treinamento e Combate à Corrupção e à Lavagem de Dinheiro (PLND), o Programa Nacional de Difusão da Cooperação Jurídica Internacional (GROTIUS-Brasil) — replicando o modelo europeu de cooperação jurídica internacional —, responsáveis por capacitar centenas de agentes públicos. Segundo Engelmann (2021, p. 62), “além de cursos de capacitação de agentes públicos voltados para as técnicas e operações, o foco direciona-se para fomentar a publicação de obras e estudos e incrementar o ensino e a pesquisa acadêmica na área”. Outra iniciativa digna de registro é a Rede Nacional de Laboratórios de Tecnologia contra a Lavagem de Dinheiro (Rede-Lab), que se destina ao desenvolvimento e troca de tecnologia em matéria de soluções de análise de grandes volumes de dados, além da difusão de outros conhecimentos e de boas práticas. 27 Londero (2021, p. 163) cita também, à semelhança da ENCCLA, a criação da Rede de Controle de Gestão Pública no ano de 2009 e que tem abrangência nacional. Em seu sítio virtual, a Rede se define como um “[...] é um centro decisório interorganizacional, de atuação nacional, que busca 27 Londero (2021, p. 163) cita também, à semelhança da ENCCLA, a criação da Rede de Controle de Gestão Pública no ano de 2009 e que tem abrangência nacional. Em seu sítio virtual, a Rede se define como um “[...] é um centro decisório interorganizacional, de atuação nacional, que busca 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA BRUNO LAMENHA BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA chave para compreender a formulação e a implementação de uma agenda anticorrupção que desenvolve uma nova institucionalidade na qual o tratamento da corrupção é mediado, sobretudo, por atores integrantes de organizações do sistema de justiça. De mais a mais, é nesse período, em meados da década de 2000, que os órgãos de controle nacionais passaram a integrar redes internacionais de cooperação entre burocracias de diferentes países, criando “[...] um espaço internacional que incrementa, para além das prescrições, métodos de controle e técnicas no ‘combate a corrupção’, inspiradas nas iniciativas de ‘guerra às drogas’”28 (ENGELMANN, 2021, p. 62) — caso do acordo firmado pela CGU e o CNMP com a Internacional Anti-Corruption Academy (IACA), criada em 2011, entre outros órgãos, pelo Escritório das Nações Unidas contra Drogas e Crime (UNODC); e da Rede de Formação de Ministérios Públicos Ibero-americanos (RECAMPI), criada em Madri, no ano de 2010. Ressalte-se ainda que, no âmbito interno, um ponto divisor de águas foi o processo de reconstrução de imagem e de identidade experimentado pela Polícia Federal (PF) nesse período, o que tem uma relação umbilical com o “encaixe institucional” 29 construído pela organização no campo da repressão penal anticorrupção. Desenvolvendo uma expertise própria e multidisciplinar, o que — evidentemente — não veio sem uma política de reestruturação administrativa e investimento30 (ARANTES, 2011, p. 114-115; MARONA; BARBOSA, 2018, p. 133- 56 aprimorar a efetividade da função de controle do Estado sobre a gestão pública (...) O trabalho segue a lógica da cooperação e do uso e adaptação de experiências bem-sucedidas. O único órgão formal da Rede é a Secretaria-Executiva, constituída em 2019, a qual está encarregada de organizar o encontro anual de coordenadores, de manter a página na internet, de ser o ponto focal na relação com órgãos e entidades nacionais e internacionais e, principalmente, de promover a difusão de boas práticas” (BRASIL, 2023b). aprimorar a efetividade da função de controle do Estado sobre a gestão pública (...) O trabalho segue a lógica da cooperação e do uso e adaptação de experiências bem-sucedidas. O único órgão formal da Rede é a Secretaria-Executiva, constituída em 2019, a qual está encarregada de organizar o encontro anual de coordenadores, de manter a página na internet, de ser o ponto focal na relação com órgãos e entidades nacionais e internacionais e, principalmente, de promover a difusão de boas práticas” (BRASIL, 2023b). BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA ORCID 0000-0002-6655-2735 ORCID 0000-0002-6995-0982 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA aprimorar a efetividade da função de controle do Estado sobre a gestão pública (...) O trabalho segue a lógica da cooperação e do uso e adaptação de experiências bem-sucedidas. O único órgão formal da Rede é a Secretaria-Executiva, constituída em 2019, a qual está encarregada de organizar o encontro anual de coordenadores, de manter a página na internet, de ser o ponto focal na relação com órgãos e entidades nacionais e internacionais e, principalmente, de promover a difusão de boas práticas” (BRASIL, 2023b). 28 Mais que uma figura de linguagem, a referência a “guerra às drogas” espelha o fato de que um dos principais atores internacionais na agenda anticorrupção tem sido o Escritório das Nações Unidas sobre Drogas e Crimes que, desde 1997, promove uma série de projetos nessa temática em diferentes países. Um desses programas é justamente o Global Judicial Integrity Network, criado mais recentemente, em 2018, aberto a integrantes do sistema de justiça de todo o mundo, com o objetivo de estabelecer pontos de contato e, sobretudo, prover formação e a difusão de conhecimentos e princípios comuns em matéria anticorrupção (UNODC, 2023). 29 Arantes e Moreira (2019, p. 106) explicam que a sua definição de encaixe institucional “[...] é inspirada na ideia seminal de Skocpol e praticada pelo institucionalismo histórico para se referir ao ponto de junção que atores sociais mobilizados logram encontrar ou mesmo construir junto ao Estado”. 30 Segundo Arantes (2011, p. 114-115), entre 2002 e 2008, o orçamento da PF praticamente dobrou, de 1,8 bilhão para 3,4 bilhão. De mais a mais, também houve uma expansão no quantitativo de pessoal, inclusive com a criação de um quadro administrativo próprio, com mais de 1600 vagas 30 Segundo Arantes (2011, p. 114-115), entre 2002 e 2008, o orçamento da PF praticamente dobrou, de 1,8 bilhão para 3,4 bilhão. De mais a mais, também houve uma expansão no quantitativo de pessoal, inclusive com a criação de um quadro administrativo próprio, com mais de 1600 vagas 30 Segundo Arantes (2011, p. 114-115), entre 2002 e 2008, o orçamento da PF praticamente dobrou, de 1,8 bilhão para 3,4 bilhão. De mais a mais, também houve uma expansão no quantitativo de pessoal, inclusive com a criação de um quadro administrativo próprio, com mais de 1600 vagas 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 31 Arantes e Moreira (2019, p. 121) advertem, contudo, que o baixo grau de coesão interna decorrente do fato de existirem três carreiras independentes dentro da organização (delegados, escrivães e peritos) tem atrasado o processo de afirmação institucional da organização e, consequentemente, pleitos de natureza corporativa. em diferentes áreas, como medicina, engenharia civil, psicologia, pedagogia e até sociólogos e filósofos. PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? 135), a PF soube se fazer essencial nas investigações sobretudo nos grandes casos envolvendo a temática da corrupção política, em relação da “produção e qualidade da prova, elemento central par ao bom funcionamento do sistema de justiça criminal” (ARANTES; MOREIRA, 2019, p. 115). p Assim, embora autorizadas pelo Judiciário e contando com a atuação do Ministério Público Federal e, por vezes, da participação e expertise de outros órgãos, como a Receita Federal, as grandes operações envolvendo delitos associados à prática de corrupção e outros crimes correlatos se notabilizaram, tendo conhecidas, na mídia e na sociedade, como “operações da Polícia Federal”. Isso gerou capital político e visibilidade para a organização, colocando-a em posição de se posicionar, como um ator relevante, na dinâmica do pluralismo estatal, buscando benefícios corporativos sob o discurso de afirmação da sua essencialidade para a defesa da moralidade e do patrimônio público federal 31. Segundo Avritzer e Marona (2017, p. 367), o número de operações anticorrupção com a participação da PF saltou de 18, em 2004, para 556, no ano de 2015 — um número trinta vezes maior. Nessa esteira, vários casos que se notabilizaram nos meios de comunicação, como as Operações Anaconda (2003), Sanguessuga (2006), Navalha (2007) Satiagraha (2008) datam da primeira década dos anos 2000. Nestes anos iniciais, o principal ponto de inflexão nessa trajetória, sem dúvida, foi o chamado “escândalo do Mensalão” (2005), que afastou e responsabilizou agentes políticos de projeção nacional que serviam em altas posições do governo federal incumbente. De outro giro, a chamada Operação Lava Jato, com seu início já em 2014, em números superlativos pouco comparáveis a qualquer outro caso anticorrupção na história do País32, representou o ápice dessa crescente, colocando órgãos do sistema de justiça, como o MP, no centro das atenções da vida política e da opinião pública, fortalecendo narrativas dirigidas para uma certa refundação da República por meio de um esforço vigoroso de luta contra a corrupção através do Judiciário 33 57 57 em diferentes áreas, como medicina, engenharia civil, psicologia, pedagogia e até sociólogos e filósofos. 31 Arantes e Moreira (2019, p. 121) advertem, contudo, que o baixo grau de coesão interna decorrente do fato de existirem três carreiras independentes dentro da organização (delegados, escrivães e peritos) tem atrasado o processo de afirmação institucional da organização e, consequentemente, pleitos de natureza corporativa. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 (CHALOUB; LYNCH, 2018, p. 270-272; ENGELMANN, 2021, p. 5; DIAS et al., 2020, p. 266-267). Pode-se objetar aqui que o histórico acima traçado privilegia elementos e a trajetória de arranjos institucionais no âmbito federal. No entanto, como tantos processos de mudança institucional recente no Estado brasileiro, esse parece ter sido mesmo o fluxo das sequências temporais: um movimento isomórfico do plano da União para os demais níveis federativos, a exemplo do ilustrado acima com o caso da CGU. Esse, por sinal, é mais elemento que ajuda a explicar a tendência da literatura especializada em reduzir a explicação do fenômeno do giro de prioridades ao desempenho institucional do MPF. Nessa perspectiva, está-se ciente que a realidade das polícias judiciárias estaduais está distante do quadro geral da PF (COSTA; OLIVEIRA JÚNIOR, 2016, p. 161-162). Comparativamente ao Ministério Público, o cenário da pluralidade institucional em relação aos órgãos policiais nos estados da Federação é mais precário. Embora haja atribuições comuns entre as polícias civis estaduais, decorrentes da legislação penal e processual em vigor, não há um marco normativo partilhado entre as organizações policiais para além da atribuição para funcionar como polícia judiciária prevista na CRFB-88. Falta, portanto, um estatuto jurídico mínimo, a exemplo da Lei n. 8.625/93 em relação aos MPs estaduais. Além disso, há um baixo grau de coesão interna, decorrente de disputas corporativas que se dão entre as carreiras da própria corporação. 58 De todo modo, a literatura tem sugerido que a especialização de estruturas de repressão penal anticorrupção no âmbito dos MPs estaduais – nomeadamente os já citados GAECOS – tem criado um espaço para que as polícias civis se integrem a essa expertise, ainda que através de um grupo pequeno de agentes policiais e que, por vezes, não atuam propriamente em nome da Polícia, mas recrutados e requisitados pelo próprio MP (SILVA, 2020, p. 11-12). Na prática, portanto, essa cooperação institucional informal entre os órgãos ministeriais e as polícias estaduais parece ter “compensado” ou mitigado os prejuízos decorrentes da infraestrutura material e humana de algumas das polícias judiciárias estaduais e refundação do Brasil”, no qual conclui que “[...] a corrupção favorece os piores. É a prevalência dos desonestos sobre os íntegros. Esse modelo não se sustenta indefinidamente. Só se o mal pudesse mais do que o bem. Mas, se fosse assim, nada valeria a pena. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? 32 A Lava Jato contou, em seu auge, com 4 (quatro) forças-tarefa em diferentes unidades da Federação (PR, RJ, DF e SP) e, entre 2014 e julho de 2018, “obteve 134 condenações somente em seu braço paranaense e denunciou ilícitos associados à corrupção que envolvem, somente na Petrobrás, uma malversação de recursos públicos da ordem de 6,2 bilhões de reais” (LAMENHA, 2019). 32 A Lava Jato contou, em seu auge, com 4 (quatro) forças-tarefa em diferentes unidades da Federação (PR, RJ, DF e SP) e, entre 2014 e julho de 2018, “obteve 134 condenações somente em seu braço paranaense e denunciou ilícitos associados à corrupção que envolvem, somente na Petrobrás, uma malversação de recursos públicos da ordem de 6,2 bilhões de reais” (LAMENHA, 2019). 33 Um exemplo deste tipo de narrativa pode ser extraído de prefácio escrito pelo ministro do STF, Luís Roberto Barroso (2018), a uma obra coletivo sobre a temática da corrupção. O título do texto do jurista é sugestivo: “Empurrando a história: combate à corrupção, mudança de paradigmas e 33 Um exemplo deste tipo de narrativa pode ser extraído de prefácio escrito pelo ministro do STF, Luís Roberto Barroso (2018), a uma obra coletivo sobre a temática da corrupção. O título do texto do jurista é sugestivo: “Empurrando a história: combate à corrupção, mudança de paradigmas e 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 BRUNO LAMENHA ORCID 0000-0002-6655-2735 ORCID 0000-0002-6995-0982 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA FLÁVIA DANIELLE SANTIAGO LIMA (CHALOUB; LYNCH, 2018, p. 270-272; ENGELMANN, 2021, p. 5; DIAS et al., 2020, p. 266-267). 34 Veja-se, por exemplo, algumas notícias recentes divulgadas na imprensa sobre operações conjuntas entre o MPE via GAECO e as Polícias Civis: i) Piauí (2023): Operação Cerrados, atuação conjunta do MPPI através do GAECO, Polícia Civil do Piauí, Polícia Militar do Piauí, Tribuna de Contas do Piauí e MPMA; ii) Santa Catarina (2022): “Gaeco, Geac e Polícia Civil cumprem mandados de busca e apreensão em operação contra corrupção em Urussanga”; iii) São Paulo (2022): “GAECO e Corregedoria [da Polícia Civil] miram corrupção policial em Osasco com Operação Haalapenz (SÃO PAULO, 2022); iv) Paraíba (2023): “Polícia e Gaeco deflagraram operação para investigar os suspeitos de corrupção, peculato e extorsão por parte de advogada e colegas” (METRÓPOLES, 2023). PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? permitido o desenvolvimento e a difusão de uma agenda e de uma expertise voltas para a repressão penal anticorrupção também no plano subnacional34. permitido o desenvolvimento e a difusão de uma agenda e de uma expertise voltas para a repressão penal anticorrupção também no plano subnacional34. (CHALOUB; LYNCH, 2018, p. 270-272; ENGELMANN, 2021, p. 5; DIAS et al., 2020, p. 266-267). A maneira desassombrada como a sociedade brasileira — e parte das suas instituições — vem enfrentando a corrupção e a impunidade, dentro do estado de direito, produzirá, logo ali na esquina do tempo, uma transformação cultural importante: a revalorização dos bons em lugar dos espertos. Quem tiver talento para produzir uma inovação relevante, ou for capaz de baixar custos de uma obra pública, será mais importante do que quem conhece a autoridade administrativa que paga qualquer preço, desde que receba uma vantagem por fora. Essa talvez seja uma das maiores conquistas que poderá vir de um novo padrão de decência e seriedade”. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 3.2. O PAPEL REGULAMENTADOR DO CNMP NA ATUAÇÃO ANTICORRUPÇÃO E O ISOMORFISMO DE MODELOS INSTITUCIONAIS A repressão anticorrupção, hoje definida como “defesa da probidade administrativa” 35 , é um dos principais temas a receber atenção do Conselho Nacional do Ministério Público em seu esforço de padronização de rotinas e de modelos institucionais para os órgãos ministeriais, na esteira do que apontado mais acima, a propósito da possibilidade de estudo do MP brasileiro como organização nacional. Discutir-se-á aqui os dois principais atos associados a essa temática: a Recomendação CNMP n. 42/2016 e a Resolução CNMP n. 185/2018. O primeiro documento, ressalte-se, não tem caráter mandatório: o objetivo de uma Recomendação é expor as razões fáticas e jurídicas sobre uma determinada questão, tencionando persuadir o destinatário a atender, voluntariamente, o comando recomendado. No caso do CNMP, o conteúdo desse tipo de instrumento tem sido mobilizado prioritariamente na atividade finalística do MP, em face da compreensão mais clássica do princípio da independência funcional36. A Recomendação CNMP n. 42, de 5 de setembro de 2016, recomenda aos órgãos do Ministério Público brasileiro que criem estruturas especializadas no Ministério Público para a otimização do enfrentamento à corrupção, com atribuição cível e criminal. Nos fundamentos de sua edição, o expediente destaca 5 (cinco) 59 59 34 Veja-se, por exemplo, algumas notícias recentes divulgadas na imprensa sobre operações conjuntas entre o MPE via GAECO e as Polícias Civis: i) Piauí (2023): Operação Cerrados, atuação conjunta do MPPI através do GAECO, Polícia Civil do Piauí, Polícia Militar do Piauí, Tribuna de Contas do Piauí e MPMA; ii) Santa Catarina (2022): “Gaeco, Geac e Polícia Civil cumprem mandados de busca e apreensão em operação contra corrupção em Urussanga”; iii) São Paulo (2022): “GAECO e Corregedoria [da Polícia Civil] miram corrupção policial em Osasco com Operação Haalapenz (SÃO PAULO, 2022); iv) Paraíba (2023): “Polícia e Gaeco deflagraram operação para investigar os suspeitos de corrupção, peculato e extorsão por parte de advogada e colegas” (METRÓPOLES, 2023). 35 É interessante observar que, em outubro de 2022, o CNMP alterou a nomenclatura de uma de suas comissões permanentes, a Comissão de Enfrentamento da Corrupção, que havia sido incluída em seu Regimento Interno em 2018, para “Comissão de Defesa da Probidade Administrativa”, através da Emenda Regimental n. 46/2022. Esse movimento, embora sutil, pode sinalizar uma nova tendência de possível alteração da política do Conselho para a temática, como um dos efeitos do rescaldo pós Lava Jato (BRASIL, 2022). 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 3.2. O PAPEL REGULAMENTADOR DO CNMP NA ATUAÇÃO ANTICORRUPÇÃO E O ISOMORFISMO DE MODELOS INSTITUCIONAIS 36 Essa feição mais clássica da independência funcional deve ser entendida como a liberdade de atuação de que gozam os membros do MP no exercício de sua atividade-fim sem a necessidade de subordinação ou aquiescência de membros ou instâncias superiores do órgão (MAZZILLI, 2015, p. 73; LAMENHA, 2019). Está para o promotor e o procurador, assim como a independência fundada no livre convencimento motivado está para os juízes na decisão dos processos judiciais (CARVALHO e LEITÃO, 2019, p. 52-53) 36 Essa feição mais clássica da independência funcional deve ser entendida como a liberdade de atuação de que gozam os membros do MP no exercício de sua atividade-fim sem a necessidade de subordinação ou aquiescência de membros ou instâncias superiores do órgão (MAZZILLI, 2015, p. 73; LAMENHA, 2019). Está para o promotor e o procurador, assim como a independência fundada no livre convencimento motivado está para os juízes na decisão dos processos judiciais (CARVALHO e LEITÃO, 2019, p. 52-53) 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 ORCID 0000-0002-6655-2735 ORCID 0000-0002-6995-0982 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA experiências de órgãos ministeriais específicas que o Conselho tenciona replicar através dessa política institucional específica: i) o Grupo Especial de Defesa do Patrimônio Público (GEPP), criado em 2007 no âmbito do MP-MG; ii) o Grupo de Atuação Regional de Defesa do Patrimônio Público (GARRP), criado em 2013 no âmbito do MP-RN; iii) o Grupo Especial Anticorrupção (GEAC), criado em 2015 no âmbito do MPSC; iv) a criação da 5ª Câmara de Coordenação e Revisão na estrutura do MPF, com objetivo específico de atuar em feitos cíveis e criminais associados à repressão anticorrupção; v) a criação, também no âmbito do MPF, de Núcleos de Combate à Corrupção (NCCs) em quase todos os Estados da Federação, com atribuição para atuar tanto no âmbito cível quanto criminal (BRASIL, 2016a). 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 38 No passado, as questões cíveis eram submetidas a uma Câmara específica (a de Patrimônio Público, antiga nomenclatura da 5ª Câmara) e as criminais para outra (a 2a Câmara), especializada em matéria penal. 37 De acordo com o levantamento de Londero (2022, p. 202-210), em 2008, por exemplo, foram ofertados 42 cursos. Em 2017, registrou-se 156 eventos. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? efetivamente ocorria quanto a repressão anticorrupção era coordenada por duas instâncias superiores distintas. efetivamente ocorria quanto a repressão anticorrupção era coordenada por duas instâncias superiores distintas. A alta especialização também promove a expansão de um novo conjunto de conhecimentos em persecução que aponta para uma certa “modernização da investigação criminal” (VIECELI, 2017), como o desenvolvimento de modelos institucionais específicos para o enfrentamento da grande criminalidade, como as forças-tarefas39 e as operações. Esse know how, evidentemente, não fica restrito aos domínios do MPF, sendo prudente considerar em que medida os efeitos de sua difusão “tem o potencial de gerar efeitos na cultura institucional do MP mobilizando um maior interesse dos membros na atuação anticorrupção, especialmente focada na repressão penal” (LAMENHA; LIMA, 2021, p. 96-97). Evidentemente, não se quer propor que esse movimento somente se dá do MPF para os demais órgãos ministeriais. As trocas de conhecimentos e boas práticas são duais. Por exemplo, durante anos, o MPF utilizou o modelo de força-tarefa, que priorizava a montagem de uma equipe temporária para o enfrentamento de um caso em específico e seus desdobramentos. Findo o caso, o grupo se desmobilizava. Em que pese também haja vantagens, um dos grandes problemas do modelo era a dispersão do conhecimento acumulado com os trabalhos. Nessa linha, até mesmo o MPF, a partir e 2020, adotou o modelo dos Grupos de Atuação Especial de Combate ao Crime Organizado (GAECOs), estruturas especializadas para o enfrentamento da macro criminalidade e em coordenação com outros órgãos de controle, cuja primeira experiência foi desenvolvida no MPSP, no ano de 1995. 61 61 j p p Não à toa, a principal providência recomendada pelo CNMP no bojo da multicitada Recomendação é, justamente, a criação de “[...] grupos de atuação especial para o enfrentamento à corrupção, com atuação preventiva e repressiva, e com atribuição extrajudicial e judicial, cível e criminal” (BRASIL, 2016a). É adotado, portanto, o modelo de GAECO. Em que pese não se tratar de um ato mandatório, mas de uma sugestão de práticas reconhecidas como eficientes, vários órgãos ministeriais que ainda não tinham estruturas específicas para uma atualização especializada na área anticorrupção estruturaram órgãos com esse objetivo. 39 Segundo o Manual de Atuação da Escola do Ministério Público da União, força-tarefa é “uma equipe de especialistas dotada de meios materiais necessários à consecução de um objetivo específico, de reconhecida complexidade, e que recomende, por certo período de tempo, a coordenação de esforços de um ou mais órgãos, nacionais ou estrangeiros.” (PALUDO et al., 2011, p. 28). Trata-se, contudo, de um modelo em desuso, nomeadamente porque até o próprio MPF, atualmente, não mais se utiliza desse tipo de estrutura para o enfrentamento dos grandes casos anticorrupção. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 3.2. O PAPEL REGULAMENTADOR DO CNMP NA ATUAÇÃO ANTICORRUPÇÃO E O ISOMORFISMO DE MODELOS INSTITUCIONAIS As duas referências ao MPF —tanto quanto à criação dos NCCs como de um órgão de coordenação e controle revisional — indicam que, ao tempo da expedição do citado ato pelo CNMP, o órgão ministerial federal era o principal paradigma de estruturação e especialização para a atuação no campo da repressão ainda corrupção, nomeadamente, na área penal – reforçando a hipótese explicativa da tendência da literatura especializada na explicação do fenômeno do giro de prioridades apenas a partir da atuação do MP Federal. Nesse sentido, em seu estudo sobre o desenvolvimento dessas capacidades institucionais em particular no âmbito do MPF, Londero (2021, p. 202-238) identificou alguns elementos endógenos centrais que explicam algumas das razões que tornaram esse órgão ministerial em específico em uma referência na temática em discussão, dentre os quais: i) a mobilização da Escola Superior do Ministério Público da União (ESMPU), com um aumento expressivo dos cursos de aperfeiçoamento na área anticorrupção ao longo dos anos37; ii) a criação da 5ª Câmara de Coordenação de Revisão em 2014, que unificou o tratamento da matéria anticorrupção no âmbito cível e penal em uma mesma instância de coordenação e revisão 38; iii) o estabelecimento de uma lógica crescente de especialização, que permitiu que as unidades do MPF que ainda não tinham um órgão especializado na repressão anticorrupção passassem a contar com ele. Em 2018, por exemplo, apenas a Procuradoria da República no estado do Piauí não contava com um Núcleo de Combate a Corrupção organizado em sua estrutura. 60 Um quarto aspecto é uma consequência dos três primeiros: “o aumento da expertise em assuntos de corrupção, no compartilhamento de provas e em uma visão mais global dos fatos, além de evitar a cegueira interna” (LONDERO, 2021, p. 215), ou seja, que um mesmo fato seja investigado em duplicidade — o que 38 No passado, as questões cíveis eram submetidas a uma Câmara específica (a de Patrimônio Público, antiga nomenclatura da 5ª Câmara) e as criminais para outra (a 2a Câmara), especializada em matéria penal. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? Sem pretensão exaustiva, é o caso, por exemplo, do: i) Ministério Público do Acre, que criou o Grupo de Atuação Especial de Combate à Corrupção (GAECC) através do Ato PGJ n. 019/2019 (ACRE, 2019); ii) os Ministérios Públicos do Ceará e do Mato Grosso do Sul, que criaram o Grupo Especial de Combate à Corrupção (GECOC), 39 Segundo o Manual de Atuação da Escola do Ministério Público da União, força-tarefa é “uma equipe de especialistas dotada de meios materiais necessários à consecução de um objetivo específico, de reconhecida complexidade, e que recomende, por certo período de tempo, a coordenação de esforços de um ou mais órgãos, nacionais ou estrangeiros.” (PALUDO et al., 2011, p. 28). Trata-se, contudo, de um modelo em desuso, nomeadamente porque até o próprio MPF, atualmente, não mais se utiliza desse tipo de estrutura para o enfrentamento dos grandes casos anticorrupção. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA ORCID 0000-0002-6655-2735 ORCID 0000-0002-6995-0982 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA através respectivamente do Provimento n. 093/2018 e da Resolução n. 008/2016-CPJ (CEARÁ, 2018; MATO GROSSO DO SUL, 2016). Nestes casos e nos referidos mais acima pela própria Recomendação, a função precípua desses coletivos não é se substituir ao promotor/procurador natural na apuração dos casos em particular, mas prover auxílio e inteligência especializada, qualificando a capacidade de investigação e, consequentemente, de obtenção de êxito na repressão de ilícitos anticorrupção. Essa é justamente a grande vantagem do modelo de GAECO em relação ao de força-tarefa: a perenidade estrutural facilita o acúmulo de expertise. p p O segundo ato publicado pelo CNMP a que se fez referência é a Resolução CNMP n. 185, de 2 de março de 2018, que cria a Comissão Especial de Enfrentamento à Corrupção, a qual tem como uma de suas funções “[...] fomentar a integração entre os ramos e as unidades do Ministério Público e entre esses e outros órgãos públicos e entidades da sociedade civil essenciais ao enfrentamento da corrupção” (BRASIL, 2018). Recentemente, vale ressaltar, a sua nomenclatura mudou, passando-se a chamar “Comissão de Defesa da Probidade Administrativa” (BRASIL, 2022). Atualmente são desenvolvidas, entre outras iniciativas, como grupos de trabalho sobre inovação no enfrentamento da corrupção por meio da tecnologia da informação e roteirização e fluxograma de atuação das diversas tipologias de corrupção e improbidade administrativa. PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? Ainda a respeito do CNMP, Penna (2012, p. 238-245) sublinha o seu importante papel no estabelecimento de uma política comum de educação corporativa para o MP brasileiro, através de iniciativas como a Unidade Nacional de Capacitação, criada pela Resolução CNMP n. 146/2016, cuja atribuição principal é regulamentar os cursos de ingresso, formação e aperfeiçoamento de membros e servidores do MP brasileiro, tendo por diretrizes centrais, entre outras, a cooperação interinstitucional e o alinhamento de objetivos estratégicos (arts. 2º e 3º). Destacam- se, ainda, por exemplo, eventos como o Congresso Brasileiro de Gestão do Ministério Público, que, desde 2010, discute anualmente temas relacionados a um certo repertório gerencial a ser difundido, como boas práticas, em todo o território nacional, mesclando essa abordagem com outras questões diretamente relacionadas à atividade finalística, como o combate à corrupção e a mobilização da persecução penal como instrumento de garantia dos direitos humanos. 62 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PLURALIDADE E MUDANÇA INSTITUCIONAL HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO Tabela 2 – Atos do CNMP associados à difusão de conhecimento e boas práticas em matéria de enfrentamento à corrupção. Ato do CNMP Finalidade Recomendação CNMP n. 42/2016 Recomenda a criação de estruturas especializadas no Ministério Público para otimização do enfrentamento à corrupção , com atribuição cível ec riminal Resolução CNMP n. 185, de 2 de março de 2018 Cria um órgão interno do CNMP, hoje nomeado como “Comissão de Defesa da Probidade (originalmente, chama-se Comissão de Enfrentamento da Corrupção). Resolução CNMP n. 146/2016 Cria a Unidade Nacional de Capacitação, que regulamenta e subsidia os cursos de ingresso, formação e aperfeiçoamento de membros e servidores do MP brasileiro. Fonte: Autor (2023 Fonte: Autor (2023) Fonte: Autor (2023) Outras experiências de estruturas especializadas e permanentes, inspiradas no próprio GAECO, são conhecidas, caso, por exemplo, do GAEMA – Grupo de Atuação Especial de Defesa do Meio Ambiente do MP-SP, que também existe em outros órgãos ministeriais, como o Paraná e Rondônia. É sintomático, todavia, e favorece a tese da força de uma agenda nacional anticorrupção, que os Grupos de Atuação Especial anticorrupção são, até os dias de hoje, o único tipo de coletivo especial objeto de recomendação pelo CNMP. Não se quer dizer com isso que a política institucional adotada pelo CNMP nessa temática concorreu ou causou o fenômeno do giro de prioridades, caso ele seja comprovado. No entanto, é possível afirmar peremptoriamente que há incentivos institucionais claros, vindos do Conselho, na direção da priorização, pelos MPs brasileiros, da temática anticorrupção — ainda que não exclusivamente na atuação penal. 63 63 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 4. HÁ BASE EMPÍRICA PARA O FENÔMENO DESCRITO COMO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DOS ÓRGÃOS MINISTERIAIS BRASILEIROS? Neste último tópico, concluir-se-á o percurso estabelecido neste artigo, enfrentando as questões que ficaram pendentes. Primeiro, conforme demonstrado mais acima, a atuação do Ministério Público Federal vem sendo mobilizada como paradigma do giro de prioridades na literatura especializada. Assim, convém discutir se há evidências que demonstrem a hipótese de giro de prioridades no próprio MPF. Após, acaso confirmada a primeira hipótese, a questão é saber se há evidências que demonstrem haver efeitos desse giro de prioridades também no âmbito dos MPs estaduais. Em trabalho anterior (LAMENHA; LIMA, 2021, p. 91-96), foram discutidos em conjunto os achados do survey referido na Tabela 1 acima e um outro 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA ORCID 0000-0002-6655-2735 ORCID 0000-0002-6995-0982 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA levantamento 40 , produzido duas décadas depois (LEMGRUBER et al., 2016), abordando questões muito semelhantes ao primeiro acerca da visão dos membros em relação ao desempenho da instituição, suas prioridades e sua relação com a sociedade. Na análise, verificou-se que, de fato, é possível identificar uma tendência nacional no sentido de uma mudança institucional informal caracterizada pela passagem de “uma dominância discursiva e política da agenda da tutela coletiva para o campo anticorrupção, notadamente no seu aspecto de repressão penal” (LAMENHA; LIMA, 2021, p. 95), mesmo levando em consideração o fato de pluralidade institucional. Este é um achado de particular interesse, visto que, consideradas as principais características institucionais e funcionais do MP brasileiro – notadamente o amplo rol de atribuições e a independência funcional assegurada aos membros –, a visão de mundo e as preferências do dos promotores e procuradores têm uma influência inegável na agenda, no discurso e no desempenho institucional da própria organização41. Não se pode esquecer, contudo, que essa regressão da predominância do tema da tutela coletiva no discurso e no etos profissional dos membros do Ministério Público não se traduz necessariamente em números quantitativos. A atribuição criminal sempre fora e sempre será, quantitativamente, a mais volumosa da atividade ministerial, diante do fato de que o MP detém o monopólio da ação penal. 40Essa segunda pesquisa, intitulada “Ministério Público: guardião da democracia brasileira?” (LEMGRUBER et al, 2016), coordenada, entre outros, por Julita Lemgruber e Ludmila Ribeiro, e também realizada junto a membros do MP, no período de um ano (2015-2016), abarcou 899 promotores e procuradores de todas as 30 (trinta) instituições ministeriais, o que representava 7% dos membros da ativa na época 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 4. HÁ BASE EMPÍRICA PARA O FENÔMENO DESCRITO COMO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DOS ÓRGÃOS MINISTERIAIS BRASILEIROS? Assim, se é possível propor, do ponto de vista discursivo e do etos profissional, a existência de um giro de prioridades, é desejável que se promova uma perquirição acerca de eventuais reflexos desse cenário de mudança institucional informal também no desempenho institucional do Ministério Público brasileiro. Para tanto, utilizar-se-ão os dados extraídos dos relatórios anuais do Conselho Nacional do 64 40Essa segunda pesquisa, intitulada “Ministério Público: guardião da democracia brasileira?” (LEMGRUBER et al, 2016), coordenada, entre outros, por Julita Lemgruber e Ludmila Ribeiro, e também realizada junto a membros do MP, no período de um ano (2015-2016), abarcou 899 promotores e procuradores de todas as 30 (trinta) instituições ministeriais, o que representava 7% dos membros da ativa na época 41Sobre a dimensão da discricionariedade na priorização temática, de casos e na rotina de trabalho dos promotores e procuradores, uma das primeiras reflexões a chamar atenção sobre a questão foi proposta por Maria Tereza Sadek, em texto do ano 2000, ao observar que a autonomia institucional e a independência funcional, livre de balizamentos mínimos, apresentava um espectro de atuação que variava desde “[...] um Ministério Público mais voltado para a persecução penal1até um Ministério Público que se define, sobretudo, como agente possuidor de responsabilidade social” (SADEK, 2000, p. 17). No ano seguinte, ganhou notoriedade o trabalho Cátia Aída Silva (2001) sobre o etos profissional dos membros do MP-SP. Nele, a autora faz a conhecida distinção entre promotor de fatos e promotor de gabinete, por “[...] meio da ênfase em certas características e elementos encontrados na atuação concreta dos promotores selecionados”, com o fito de “[...] demarcar diferenças significativas em relação a duas formas distintas de atuação” (SILVA, 2001, p. 90). 41Sobre a dimensão da discricionariedade na priorização temática, de casos e na rotina de trabalho dos promotores e procuradores, uma das primeiras reflexões a chamar atenção sobre a questão foi proposta por Maria Tereza Sadek, em texto do ano 2000, ao observar que a autonomia institucional e a independência funcional, livre de balizamentos mínimos, apresentava um espectro de atuação que variava desde “[...] um Ministério Público mais voltado para a persecução penal1até um Ministério Público que se define, sobretudo, como agente possuidor de responsabilidade social” (SADEK, 2000, p. 17). No ano seguinte, ganhou notoriedade o trabalho Cátia Aída Silva (2001) sobre o etos profissional dos membros do MP-SP. 4. HÁ BASE EMPÍRICA PARA O FENÔMENO DESCRITO COMO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DOS ÓRGÃOS MINISTERIAIS BRASILEIROS? Nele, a autora faz a conhecida distinção entre promotor de fatos e promotor de gabinete, por “[...] meio da ênfase em certas características e elementos encontrados na atuação concreta dos promotores selecionados”, com o fito de “[...] demarcar diferenças significativas em relação a duas formas distintas de atuação” (SILVA, 2001, p. 90). 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? Ministério Público no período entre 2014 e 2021, analisando globalmente os números da atuação funcional dos órgãos ministeriais brasileiros, com objetivo de identificar se há elementos que permitam identificar um giro de prioridades também na atuação cotidiana dos promotores e procuradores. Cabe aqui uma advertência inicial quanto os limites da análise aqui proposta. A hipótese do giro de prioridades, como visto, sinaliza uma tendência de reordenação de recursos institucionais do campo da tutela coletiva para a repressão penal anticorrupção. O mapeamento desenvolvido por Londero (2021), acerca do desenvolvimento das capacidades institucionais na área anticorrupção no âmbito do MPF, evidencia a viabilidade de estudos específicos sobre o cenário particular dos diferentes órgãos ministeriais brasileiros. Assim, idealmente, a testagem da hipótese em cada um dos 30 (trinta) entes que integram o MP brasileiro seria a medida analítica desejável para um juízo mais completo sobre a questão, nomeadamente porque “é esperado que órgãos diferentes possam espelhar infraestruturas, regramentos internos e culturas institucionais variadas.” (LAMENHA; LIMA, 2021, p. 77). p As pretensões deste trabalho são, evidentemente, mais modestas. A proposição quanto à viabilidade e, sobretudo, aos limites do estudo do MP como organização nacional permite, portanto, avaliar se há evidências que permitam verificar a existência de uma tendência nacional que confirme ou refute essa hipótese. Em um ou outro caso, a conclusão alcançada permitirá obter, no máximo, uma espécie de panorama geral, que não substitui a utilidade e, sobretudo, a necessidade de análise da situação de cada órgão ministerial em particular. Essa estratégia também se relaciona com a limitação da base de dados mobilizada para a análise: assim como no caso dos surveys comparados por Lamenha e Lima (2021), o CNMP disponibiliza os dados relativos à atuação funcional dos MPs estaduais e distrital de forma agregada, não sendo possível analisá-los de forma individualizada. 65 65 4.1. A HIPÓTESE DO GIRO DE PRIORIDADES E O DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO A Resolução CNMP n. 74/2011, várias vezes alterada desde a sua publicação, dispõe sobre a “[...] aplicação do controle da atuação da gestão de pessoas, da Tecnologia da Informação, da gestão estrutural, da gestão orçamentária do Ministério Público, bem como da atuação funcional dos seus membros” (BRASIL, 2011). Este ato normativo é o responsável por viabilizar uma das ferramentas mais úteis para a análise do desempenho institucional brasileiro, inclusive no campo da pesquisa acadêmica, e mais um indicativo do papel central do CNMP como mecanismo prospectivo (e não só retrospectivo-punitivo) de accountability). Isso porque é estabelecida a obrigação anual a todos os órgãos do MP brasileiro a fornecerem dados organizados conforme as Tabelas Unificadas do Ministério Público, também padronizadas pelo Conselho, contemplando vários dados de 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 ORCID 0000-0002-6655-2735 ORCID 0000-0002-6995-0982 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA interesse à transparência, à análise e à formulação de políticas institucionais para o Ministério Público, tais como “[...] estrutura de pessoal, tecnologia da informação, orçamentária, financeira (...) e sobre o desempenho funcional” (art. 1º). Tais dados, portanto, são produzidos pelos próprios órgãos ministeriais para esta finalidade e fornecidos ao CNMP. Posteriormente são tornados públicos. Atualmente, há uma ferramenta de Business Inteligence provida pelo CNMP que permite a extração e comparação de muitos dados relevantes sobre o funcionamento do Ministério Público brasileiro, dando continuidade ao projeto “MP: um retrato”, que, até 2018, era disponibilizado em um documento convencional, com muitos menos recursos que a solução atual interesse à transparência, à análise e à formulação de políticas institucionais para o Ministério Público, tais como “[...] estrutura de pessoal, tecnologia da informação, orçamentária, financeira (...) e sobre o desempenho funcional” (art. 1º). Tais dados, portanto, são produzidos pelos próprios órgãos ministeriais para esta finalidade e fornecidos ao CNMP. Posteriormente são tornados públicos. 4.1. A HIPÓTESE DO GIRO DE PRIORIDADES E O DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO Atualmente, há uma ferramenta de Business Inteligence provida pelo CNMP que permite a extração e comparação de muitos dados relevantes sobre o funcionamento do Ministério Público brasileiro, dando continuidade ao projeto “MP: um retrato”, que, até 2018, era disponibilizado em um documento convencional, com muitos menos recursos que a solução atual Esta informação é importante, especialmente quanto ao desempenho institucional ou funcional, visto que os sistemas internos de cada MP dependem, fundamentalmente, da alimentação externa para delinearam um cenário fidedigno da atuação do órgão. Havendo um marco temporal e um formato específico exigido pelo CNMP, também há possibilidade de saneamento e revisão dos dados a serem unificados, corrigindo imprecisões, lacunas e outros tipos de tratamentos necessários para que as informações reflitam, com a maior precisão possível, o cenário da atividade finalística do órgão. g Deve-se ter em mente, todavia, algumas limitações: i) mesmo com a possibilidade de revisão e saneamento, os dados continuam passíveis de problemas de alimentação e portanto, sempre devem ser interpretados levando isso em consideração; ii) embora a Resolução CNMP n. 4.1. A HIPÓTESE DO GIRO DE PRIORIDADES E O DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO 74 seja do ano de 2011, a disponibilidade e higidez dos dados cobre séries diferentes no caso de MPEs e MPDFT e do MPF; iii) os dados disponíveis não se estendem ao período no qual teria começado a se verificar o fenômeno do giro de prioridades, isto é, meados da primeira década dos anos 2000; iv) em 2019, houve uma mudança na classificação dos temas de atuação no âmbito extrajudicial que é de interesse para análise a ser empreendida: até 2018, havia duas categorias associadas à temática anticorrupção — “improbidade administrativa” e “patrimônio público”; a partir de 2019, tudo ficou concentrado na temática “improbidade administrativa”; v) ainda na temática anticorrupção, o tema “improbidade administrativa” tem natureza dúplice, como amplamente discutido mais acima, de maneira que é necessária uma análise da atuação judicial nessa área para verificar se há uma diferença de padrão de atuação na área cível e criminal; vi) na alimentação dos dados, conforme alertado pelo próprio CNMP (BRASIL, 2021a), um mesmo procedimento pode ser vinculado a mais de um tema; vii) até 2018, não havia disponibilização de dados sobre as notícias de fato, apenas dos procedimentos preparatórios, inquéritos civis, Deve-se ter em mente, todavia, algumas limitações: i) mesmo com a possibilidade de revisão e saneamento, os dados continuam passíveis de problemas de alimentação e portanto, sempre devem ser interpretados levando isso em consideração; ii) embora a Resolução CNMP n. PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? procedimentos administrativos de acompanhamento 42 e procedimentos investigatórios criminais43. procedimentos administrativos de acompanhamento 42 e procedimentos investigatórios criminais43. Por fim, um registro de ordem metodológica na análise dos dados: no campo extrajudicial, não se discutiram dados criminais, visto que os inquéritos policiais, embora sob controle externo do MP, são conduzidos por agentes de outra organização, o que prejudica a sua utilização como vetor de análise de priorização da área criminal. Também não se utilizará os PICs, por não se tratar de um tipo de investigação exclusiva do MP, como o é o inquérito civil, e por seu quantitativo diminuto o universo da investigação criminal. Assim, ainda que o MP possa investigar delitos por conta própria, o inquérito policial ainda é, por larga margem, a modalidade de instrução pré-processual criminal mais utilizadas. Em 2021, por exemplo, foram registrados 740.901 inquéritos policiais no país, enquanto foram instaurados no mesmo período apenas 18.153 procedimentos investigatórios criminais, uma razão de quase quarenta e uma vezes (BRASIL, 2021c). 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 4.1. A HIPÓTESE DO GIRO DE PRIORIDADES E O DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO 74 seja do ano de 2011, a disponibilidade e higidez dos dados cobre séries diferentes no caso de MPEs e MPDFT e do MPF; iii) os dados disponíveis não se estendem ao período no qual teria começado a se verificar o fenômeno do giro de prioridades, isto é, meados da primeira década dos anos 2000; iv) em 2019, houve uma mudança na classificação dos temas de atuação no âmbito extrajudicial que é de interesse para análise a ser empreendida: até 2018, havia duas categorias associadas à temática anticorrupção 66 — “improbidade administrativa” e “patrimônio público”; a partir de 2019, tudo ficou concentrado na temática “improbidade administrativa”; v) ainda na temática anticorrupção, o tema “improbidade administrativa” tem natureza dúplice, como amplamente discutido mais acima, de maneira que é necessária uma análise da atuação judicial nessa área para verificar se há uma diferença de padrão de atuação na área cível e criminal; vi) na alimentação dos dados, conforme alertado pelo próprio CNMP (BRASIL, 2021a), um mesmo procedimento pode ser vinculado a mais de um tema; vii) até 2018, não havia disponibilização de dados sobre as notícias de fato, apenas dos procedimentos preparatórios, inquéritos civis, — “improbidade administrativa” e “patrimônio público”; a partir de 2019, tudo ficou concentrado na temática “improbidade administrativa”; v) ainda na temática anticorrupção, o tema “improbidade administrativa” tem natureza dúplice, como amplamente discutido mais acima, de maneira que é necessária uma análise da atuação judicial nessa área para verificar se há uma diferença de padrão de atuação na área cível e criminal; vi) na alimentação dos dados, conforme alertado pelo próprio CNMP (BRASIL, 2021a), um mesmo procedimento pode ser vinculado a mais de um tema; vii) até 2018, não havia disponibilização de dados sobre as notícias de fato, apenas dos procedimentos preparatórios, inquéritos civis, 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 FLÁVIA DANIELLE SANTIAGO LIMA FLÁVIA DANIELLE SANTIAGO LIMA bens jurídicos e não apenas a atuação anticorrupção. Na série seguinte (2019-2021), como explicado, o CNMP passou a utilizar apenas a terminologia “improbidade administrativa”. Aqui, a prevalência também foi inconteste: metade das instaurações de PPs e ICs (50%) instaurados no período foram classificados como “improbidade administrativa”. Figuras 1 e 2 - Número de instaurações de PPs e ICs na temática anticorrupção (2015-2018 e 2019-2021) - MPF Fonte: Autor (2023) Essa ampla prevalência da temática anticorrupção nos feitos extrajudiciais não se explica apenas pela hipótese do giro de prioridades. Existem especificidades na atribuição do MPF que devem ser consideradas: i) a apuração federal, em regra, é residual em relação à esfera estadual, compreendendo — salvo algumas atribuições específicas como os direitos coletivos indígenas — questões que importem em lesão ou ameaça de lesão da bem, interesse ou serviço da União Federal (CRFB, art. 109, I). Assim, matérias que geram muito volume no âmbito estadual, como o campo da infância de juventude e do direito urbanístico, são muito episódicas na atuação do MPF; ii) o caráter centralizador do federalismo brasileiro (LIZIERO; CARVALHO, 2018, p. 1501) implica uma grande dependência da maior parte dos municípios brasileiros de repasse de recursos federais. Muitos desses repasses são empreendidos via transferências voluntárias ou, mesmo em se tratando de transferências constitucionais, implicam em interesse finalístico da União na execução de determinadas políticas públicas (v.g. FUNDEB e insumos e serviços relacionados à média e alta complexidade no SUS). Esse cenário implica uma altíssima demanda para o MPF, nomeadamente na área da tutela preventiva e repressiva do patrimônio público, o que engloba a temática anticorrupção. Diante deste último elemento buscou-se uma alternativa a partir dos dados Figuras 1 e 2 - Número de instaurações de PPs e ICs na temática anticorrupção (2015-2018 e 2019-2021) - MPF Figuras 1 e 2 - Número de instaurações de PPs e ICs na temática anticorrupção (2015-2018 e 2019-2021) - MPF Figuras 1 e 2 - Número de instaurações de PPs e ICs na temática anticorrupção (2015-2018 e 2019-2021) - MPF Fonte: Autor (2023) 68 Essa ampla prevalência da temática anticorrupção nos feitos extrajudiciais não se explica apenas pela hipótese do giro de prioridades. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 4.2. O DESEMPENHO FUNCIONAL DO MINISTÉRIO PÚBLICO FEDERAL NA TEMÁTICA ANTICORRUPÇÃO No caso do Ministério Público Federal, a análise dos dados dos procedimentos extrajudiciais nos períodos compreendidos entre 2015-2018 e 2019-2021 demonstra amplamente a prevalência da agenda anticorrupção na atividade ministerial. No primeiro período, temas associados ao assunto representaram 62% das instaurações de procedimentos preparatórios (PPs) e inquéritos civis (ICs): 45% classificada como improbidade administrativa e 17%, como patrimônio público. Preferiu-se não unificar tais dados na figura abaixo, visto que a terminologia genérica “patrimônio público” pode englobar também a tutela preventiva desses 67 67 42 No marco das Resoluções CNMP 23/2007 e n. 174/2017, uma investigação de natureza cível pode compreender três tipos de procedimento extrajudicial: i) a notícia de fato, que é qualquer demanda encaminhada ao Ministério Público, independentemente da forma (presencial ou não; através de representação, notícia, requerimento, etc). Presta-se a avaliar se há ilícito a apurar e a viabilidade da investigação. Seu prazo de vigência é de 30 (trinta) dias, prorrogáveis por mais 90 (noventa) uma única vez; ii) procedimento preparatório, que é uma instrução preliminar que pode anteceder o inquérito civil, com vistas ao aprofundamento da análise de viabilidade ou não da investigação. Seu prazo de vigência é de 90 (noventa) dias, prorrogáveis por igual prazo uma única vez; iii) inquérito civil, que o procedimento por meio do qual uma investigação de ilícito civil é realizada, após a análise da sua viabilidade. Seu prazo de vigência de 1 (um) ano, prorrogáveis quantas vezes forem necessárias até o seu encerramento. Há ainda uma quarta espécie, o procedimento administrativo de acompanhamento, o que se presta ao monitoramento de políticas públicas ou de compromissos extrajudiciais firmados pelo MP, a exemplo de TACs. 43 Inicialmente regulamentado pela Resolução CNMP n. 13/2006, o procedimento investigatório criminal (PIC), hoje disciplinado pela Resolução CNMP n. 181/2017, é o instrumento mobilizado pelo Ministério Público para a realização de investigações de natureza criminal na fase pré- processual, com objetivo de subsidiar, ou não, a propositura de uma denúncia. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA ORCID 0000-0002-6655-2735 ORCID 0000-0002-6995-0982 BRUNO LAMENHA 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? medir esse cenário é verificar a proporção de conversão de novas notícias de fato em PPs e ICs na temática anticorrupção comparativamente ao universo total de procedimentos iniciais no mesmo assunto. Rememorando: notícia de fato (NF) é toda representação ou denúncia que chega ao Ministério Público para ser analisada a viabilidade de abertura de uma apuração. Podem-se empreender diligências complementares nesse feito original, com o fito exclusivo de avaliar se há viabilidade persecutória. Demonstrada a pertinência da investigação, o promotor/procurador responsável decide pela instauração de PP ou IC, estes, sim, procedimentos formais de investigação. Assim, a ideia é comparar qual a proporção de instauração de investigação a partir do universo de NFs que relatem eventual prática relacionada à corrupção. medir esse cenário é verificar a proporção de conversão de novas notícias de fato em PPs e ICs na temática anticorrupção comparativamente ao universo total de procedimentos iniciais no mesmo assunto. Rememorando: notícia de fato (NF) é toda representação ou denúncia que chega ao Ministério Público para ser analisada a viabilidade de abertura de uma apuração. Podem-se empreender diligências complementares nesse feito original, com o fito exclusivo de avaliar se há viabilidade persecutória. Demonstrada a pertinência da investigação, o promotor/procurador responsável decide pela instauração de PP ou IC, estes, sim, procedimentos formais de investigação. Assim, a ideia é comparar qual a proporção de instauração de investigação a partir do universo de NFs que relatem eventual prática relacionada à corrupção. Como já esclarecido acima, há somente dados sobre instauração de notícias de fato no banco de dados do CNMP a partir de 2019. Os achados do período de 2019- 2021 indicam foram instauradas 288.600 notícias de fato (NFs) no período. Dessas, 25232 foram classificadas como na área de improbidade administrativa, ou seja, quase 9%. Do universo de NFs associadas à improbidade no período, 39,57% foram instauradas como PPs ou ICs. Isso significa, em tese, duas coisas: i) é feita uma filtragem efetiva das NFs que serão objeto de investigação pré-processual em um PP ou IC — 6 em 10 casos são resolvidos no curto prazo de vigência das NFs, que é de, no máximo, 120 dias44; ii) o fato de representarem apenas 9% do universo total de NFs instauradas e, apesar da filtragem critériosa, no mesmo período, 50% dos PPs e ICs instaurados o foram à temática da improbidade administrativa. 44 Isso não significa que as NFs que não são convertidas em PPs e ICs são necessariamente arquivadas. Dada a atribuição residual federal, é possível que muitos desses casos sejam objeto de declínio de atribuição a outros MPs, embora esse dado não esteja disponível nos relatórios do CNMP. Também é possível a adoção de soluções judiciais ou extrajudiciais imediatas, se o caso o requerer, mas isso está longe de ser a regra. FLÁVIA DANIELLE SANTIAGO LIMA Existem especificidades na atribuição do MPF que devem ser consideradas: i) a apuração federal, em regra, é residual em relação à esfera estadual, compreendendo — salvo algumas atribuições específicas como os direitos coletivos indígenas — questões que importem em lesão ou ameaça de lesão da bem, interesse ou serviço da União Federal (CRFB, art. 109, I). Assim, matérias que geram muito volume no âmbito estadual, como o campo da infância de juventude e do direito urbanístico, são muito episódicas na atuação do MPF; ii) o caráter centralizador do federalismo brasileiro (LIZIERO; CARVALHO, 2018, p. 1501) implica uma grande dependência da maior parte dos municípios brasileiros de repasse de recursos federais. Muitos desses repasses são empreendidos via transferências voluntárias ou, mesmo em se tratando de transferências constitucionais, implicam em interesse finalístico da União na execução de determinadas políticas públicas (v.g. FUNDEB e insumos e serviços relacionados à média e alta complexidade no SUS). Esse cenário implica uma altíssima demanda para o MPF, nomeadamente na área da tutela preventiva e repressiva do patrimônio público, o que engloba a temática anticorrupção. Diante deste último elemento, buscou-se uma alternativa, a partir dos dados disponíveis, para verificar se o número de PPs e ICs na área de improbidade administrativa é alto porque a demanda de notícias de fato é alta. Uma forma de 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 45 Compilou-se para essa finalidade os dados relativos aos seguintes delitos: i) crimes contra a administração pública; ii) crimes com a administração da justiça; iii) crimes contra as finanças públicas; iv) crimes contra o sistema financeiro nacional; v) crimes de responsabilidade; vi) crimes da lei da lavagem de capitais; vii) crimes da lei de licitações. Foram contabilizadas as denúncias- crimes, transações penais e ANPPs relativos a inquéritos policiais, PICs e Termos Circunstanciados. PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? Isso indica que, efetivamente, há uma certa priorização da área no âmbito do MPF. 69 69 44 Isso não significa que as NFs que não são convertidas em PPs e ICs são necessariamente arquivadas. Dada a atribuição residual federal, é possível que muitos desses casos sejam objeto de declínio de atribuição a outros MPs, embora esse dado não esteja disponível nos relatórios do CNMP. Também é possível a adoção de soluções judiciais ou extrajudiciais imediatas, se o caso o requerer, mas isso está longe de ser a regra. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 ORCID 0000-0002-6655-2735 ORCID 0000-0002-6995-0982 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA Figura 3 - Notícias de fato: comparativo entre instaurações e proporção de conversão no tema da improbidade administrativa Figura 3 - Notícias de fato: comparativo entre instaurações e proporção de conversão no tema da improbidade administrativa Fonte: Autor (2023) Fonte: Autor (2023) Fonte: Autor (2023) A temática da improbidade administrativa, como já dito algumas vezes, é de natureza dúplice, ou seja, pode resultar em atuações cíveis e criminais. Verificar a hipótese do giro de prioridades depende de um esforço de isolar a atuação criminal da cível. A partir dos dados à disposição nos relatórios do CNMP, é possível comparar o número de ações de improbidade e TACs e o de ações penais e de acordos de não persecução (ANPPs) associados à temática anticorrupção 45. Os achados comprovam consistentemente a hipótese de Arantes (2011, p. 122-123), indicando um deslocamento da atuação na área da improbidade para o campo criminal. (2016-2019)46. Veja-se: 70 45 Compilou-se para essa finalidade os dados relativos aos seguintes delitos: i) crimes contra a administração pública; ii) crimes com a administração da justiça; iii) crimes contra as finanças públicas; iv) crimes contra o sistema financeiro nacional; v) crimes de responsabilidade; vi) crimes da lei da lavagem de capitais; vii) crimes da lei de licitações. Foram contabilizadas as denúncias- crimes, transações penais e ANPPs relativos a inquéritos policiais, PICs e Termos Circunstanciados. 45 Compilou-se para essa finalidade os dados relativos aos seguintes delitos: i) crimes contra a administração pública; ii) crimes com a administração da justiça; iii) crimes contra as finanças públicas; iv) crimes contra o sistema financeiro nacional; v) crimes de responsabilidade; vi) crimes da lei da lavagem de capitais; vii) crimes da lei de licitações. PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? Foram contabilizadas as denúncias- crimes, transações penais e ANPPs relativos a inquéritos policiais, PICs e Termos Circunstanciados. 46 Ao discutir o deslocamento proposto por Arantes (2011, p. 122-123) da repressão anticorrupção da área cível para criminal, Londero (2012, p. 73) apresenta um comparativo entre a propositura de ações civis públicas por improbidade administrativa e ações penais associadas à prática de corrupção no MPF de 1993 a 2018. Os achados e conclusões da autora são diferentes dos aqui encontrados, registrando um “ movimento não linear de flutuação entre as prioridades civis e penais no enfrentamento da corrupção por parte do MPF” (LONDERO, 2021, p. 40, 71-72): i) as ações penais foram maioria entre 2003 e 2008; ii) a partir de 2008, e em especial de 2010, as ações de improbidade prevaleceram, mesmo no período do auge da operação Lava Jato. Os dados por ela utilizados foram extraídos dos sistemas internos do MPF pela própria pesquisadora, que obteve autorização do órgão para tanto. Há, no entanto, algumas discrepâncias nos números quando cruzados com os do CNMP (no período disponível). Por exemplo, no ano de 2018, último da série de Londero (2021, p. 73), seu registro apresenta a propositura de 1328 improbidades e 46 Ao discutir o deslocamento proposto por Arantes (2011, p. 122-123) da repressão anticorrupção da área cível para criminal, Londero (2012, p. 73) apresenta um comparativo entre a propositura de ações civis públicas por improbidade administrativa e ações penais associadas à prática de corrupção no MPF de 1993 a 2018. Os achados e conclusões da autora são diferentes dos aqui encontrados, registrando um “ movimento não linear de flutuação entre as prioridades civis e penais no enfrentamento da corrupção por parte do MPF” (LONDERO, 2021, p. 40, 71-72): i) as ações penais foram maioria entre 2003 e 2008; ii) a partir de 2008, e em especial de 2010, as ações de improbidade prevaleceram, mesmo no período do auge da operação Lava Jato. Os dados por ela utilizados foram extraídos dos sistemas internos do MPF pela própria pesquisadora, que obteve autorização do órgão para tanto. Há, no entanto, algumas discrepâncias nos números quando cruzados com os do CNMP (no período disponível). Por exemplo, no ano de 2018, último da série de Londero (2021, p. 73), seu registro apresenta a propositura de 1328 improbidades e 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? 37 - 86, jan./abr. 2024 PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? Figura 4 - Comparativo de atuações de finalização de apuratórios na área cível e criminal (exceto arquivamentos e declínios) – MPF (2016-2019) Fonte: Autor (2023) Associando-se os achados aqui discutidos com as mudanças institucionais identificadas por Londero (2021) em sua pesquisa e expostas mais acima, é possível afirmar que há suporte empírico para sustentar a hipótese do giro de prioridades no âmbito do Ministério Público Federal. Isso já é assumido largamente, como visto, pela literatura especializada. É preciso avaliar, no entanto, em que medida esse padrão é verificável no âmbito dos MPEs e do MPDFTs. Fonte: Autor (2023) Associando-se os achados aqui discutidos com as mudanças institucionais identificadas por Londero (2021) em sua pesquisa e expostas mais acima, é possível afirmar que há suporte empírico para sustentar a hipótese do giro de prioridades no âmbito do Ministério Público Federal. Isso já é assumido largamente, como visto, pela literatura especializada. É preciso avaliar, no entanto, em que medida esse padrão é verificável no âmbito dos MPEs e do MPDFTs. Associando-se os achados aqui discutidos com as mudanças institucionais identificadas por Londero (2021) em sua pesquisa e expostas mais acima, é possível afirmar que há suporte empírico para sustentar a hipótese do giro de prioridades no âmbito do Ministério Público Federal. Isso já é assumido largamente, como visto, pela literatura especializada. É preciso avaliar, no entanto, em que medida esse padrão é verificável no âmbito dos MPEs e do MPDFTs. 71 71 1203 ações penais anticorrupção. Os dados extraídos da plataforma do CNMP, todavia, indicam um número 23% (vinte e três por cento) menor em relação às improbidades (1012) Em relação às ações penais, também no mesmo período, a divergência é mais significativa: os dados do CNMP são quase 4x (quatro vezes) maiores que os apresentados por Londero (2021, p 73), totalizando 4.637 ações penais propostas. Esta diferença entre os dados especialmente no âmbito criminal também é identificado em outros períodos. No interregno 2016-2017, compreendendo o intervalo de um ano no mandato de Rodrigo Janot, por exemplo, em pleno auge da Operação Lava Jato, teriam sido propostas apenas 725 ações penais de acordo com os dados apresentados por Londero (2021, p. 73). 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? O compilado do CNMP, por sua vez, indica um número oito vezes maior, totalizando 5.867 ações penais ajuizadas em todo o ano de 2016. Acrescente-se que os critérios de definição da tipologia delitiva associada à corrupção utilizada nesta pesquisa e em Londero (2021) é bastante semelhante. Considerando que, no marco da Resolução CNMP 74/2011, os dados enviados ao CNMP o são para a finalidade específica de medição da atividade funcional e que foram extraídos diretamente pelos próprios autores, optou-se por utilizar, para a análise nesta pesquisa, apenas os dados do CNMP. O contraste entre os dois conjuntos de dados aqui cotejados, todavia, pode ensejar discussões posteriores e mais aprofundadas. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA ORCID 0000-0002-6655-2735 ORCID 0000-0002-6995-0982 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? ensejando a instauração de procedimentos de investigação formais (PPs e ICs), postergando a sua finalização ainda na fase de notícia de fato. É que, enquanto apenas 7% das 1.390.694 notícias de fato instauradas entre 2019 e 2021 – isto é, 103.111 casos – foram convertidos em PPs ou ICs, o percentual de conversão quando observada apenas a temática anticorrupção é significativamente maior: 57,36% das notícias de fato classificadas como “improbidade administrativa” conduziram à instauração de PPs ou ICs. ensejando a instauração de procedimentos de investigação formais (PPs e ICs), postergando a sua finalização ainda na fase de notícia de fato. É que, enquanto apenas 7% das 1.390.694 notícias de fato instauradas entre 2019 e 2021 – isto é, 103.111 casos – foram convertidos em PPs ou ICs, o percentual de conversão quando observada apenas a temática anticorrupção é significativamente maior: 57,36% das notícias de fato classificadas como “improbidade administrativa” conduziram à instauração de PPs ou ICs. Figura 7 e 8 - Notícias de fato: comparativo entre instaurações e proporção de conversão no tema da improbidade administrativa – MPE e MPDFT (2019-2023) Fonte: Autor (2023) A hipótese de Arantes (2011, p. 122-123) quanto ao deslocamento da repressão anticorrupção da área cível para a criminal também é demonstrada de forma consistente no âmbito dos MPs Estaduais, sobretudo nos últimos três anos da série analisada. Apesar dos percentuais mais elásticos no MPF indicando a predominância da atividade criminal, o que era esperado, a menor diferença da série foi de 14%, em 2015. Nos últimos três anos observados, há uma diferença significativa, com percentuais de 75% de predominância de atuação de finalização de investigação. Ressalte-se quem, no caso dos MPEs e MPDFT, a série analisada é mais ampla em razão da maior disponibilidade de dados pelo CNMP sobre o desempenho institucional desses órgãos no âmbito judicial. Fonte: Autor (2023) Fonte: Autor (2023) 73 73 Fonte: Autor (2023) A hipótese de Arantes (2011, p. 122-123) quanto ao deslocamento da repressão anticorrupção da área cível para a criminal também é demonstrada de forma consistente no âmbito dos MPs Estaduais, sobretudo nos últimos três anos da série analisada. Apesar dos percentuais mais elásticos no MPF indicando a predominância da atividade criminal, o que era esperado, a menor diferença da série foi de 14%, em 2015. Nos últimos três anos observados, há uma diferença significativa, com percentuais de 75% de predominância de atuação de finalização de investigação. 4.3. O DESEMPENHO FUNCIONAL DOS MINISTÉRIOS PÚBLICOS ESTADUAIS E DO MPDFT NA TEMÁTICA ANTICORRUPÇÃO No caso dos órgãos ministeriais estaduais e do MP distrital, a análise dos dados dos procedimentos extrajudiciais nos períodos compreendidos entre 2015-2018 e 2019-2021 apresentam um cenário distinto em relação ao MPF. Nos dois períodos, os apuratórios classificados como “improbidade administrativa” são a maioria, embora com uma margem estreita em relação ao tema subsequente, o qual é o meio ambiente. Na primeira série (2015-2018), somando-se as classificações “improbidade administrativa” e “patrimônio público”, a prevalência é mais alta: 42% dos apuratórios. Na série seguinte (2019-2021), esse número cai (35%), sugerindo a acerto da leitura proposta mais acima de que a classificação “patrimônio público” não englobava apenas a tutela repressiva, mas principalmente a abordagem preventiva anticorrupção. No último período, também aumenta bastante a demanda no âmbito do meio ambiente (25% na primeira série, 34% na última). Figura 5 e 6 - Número de instaurações de PPs e ICs na temática anticorrupção (2015-2018 e 2019-2021) - MPEs Fonte: Autor (2023) 72 Fonte: Autor (2023) Tais achados, isoladamente, não significam não se possa falar em giro de prioridades no âmbito do MPE. É preciso analisar os demais dados, justamente pelo motivo já explicado quanto ao MPF, aqui entendido na razão inversa: a atribuição estadual é comum, ou seja, abrange uma quantidade muito maior de temas e, nos temas compartilhados com a área federal, uma abrangência muito maior de casos. Nessa perspectiva, os dados relativos às notícias de fato confirmam a mesma tendência do MPF, com um diferencial inesperado. Proporcionalmente, pode-se propor que, na seara extrajudicial, os MPEs e o MPDFT dedicam maior prioridade à temática anticorrupção que o próprio MPF – ou, no mínimo, que a análise dos casos envolvendo a matéria é mais criteriosa, 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? Ressalte-se quem, no caso dos MPEs e MPDFT, a série analisada é mais ampla em razão da maior disponibilidade de dados pelo CNMP sobre o desempenho institucional desses órgãos no âmbito judicial. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 ORCID 0000-0002-6655-2735 ORCID 0000-0002-6995-0982 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA Figura 9 - Comparativo de atuações de finalização de apuratórios na área cível e criminal (exceto arquivamentos e declínios) – MPEs e MPDFT (2016-2019) Fonte: Autor (2023) Figura 9 - Comparativo de atuações de finalização de apuratórios na área cível e criminal (exceto arquivamentos e declínios) – MPEs e MPDFT (2016-2019) Fonte: Autor (2023) Do ponto de vista do desempenho funcional, portanto, pode-se sustentar que os MPEs e o MPDFT também empreenderam um giro em suas prioridades, privilegiado a atuação no campo da repressão penal anticorrupção. Considerando que se está abordando a realidade de 27 (vinte e sete) órgãos, é prudente encarar esse achado como uma tendência geral de mudança institucional gradual, explicável por fenômenos como: i) o isomorfismo existente entre os órgãos ministeriais, reproduzindo o movimento do âmbito federal para o estadual, a exemplo do que ocorrera no caso da CGU; ii) uma possível influência da atuação regulamentar do CNMP, sendo um elemento a ser aprofundado por pesquisas futuras a mensuração do papel do processo de padronização institucional promovido pelo órgão de controle na difusão de uma tendência geral à mudança institucional gradual aqui discutida. 74 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? Poder-se-ia objetar se a criação de estruturas internas especializadas, como a criação e a mobilização dos GAECOs e órgãos assemelhados para essa finalidade, não seriam mudanças institucionais formais. Propõe-se, no entanto, que essas modificações formais parecem advir em um momento posterior, quando já há uma tendência de dedicar maior atenção à repressão anticorrupção no âmbito criminal. Perceba-se, por exemplo, que a Recomendação CNMP n. 42/2016 (BRASIL, 2016ª), enquanto o fenômeno aqui descrito é mapeado pela literatura desde a primeira década dos anos 2000. A maior parte das estruturas estaduais também, e a própria 5ª CCR no âmbito do MPF, datam de meados nos anos 2010. Uma hipótese explicativa aqui é que a gênese dessas regras informais se baseia na baixa efetividade das ações de improbidade e na maior versatilidade de instrumentos de investigação à disposição na área criminal, justamente como propôs Rogério Arantes, ainda em 2011. Um fenômeno digno de nota, e que merece uma agenda de pesquisa específica, é que - também a partir da segunda metade da década de 2010 – levantamentos referentes à confiança da população nas organizações, começaram a apontar um declínio em relação ao MP, que sempre fora uma das entidades mais bem colocadas. Veja-se, por exemplo, o Índice de Confiança na Justiça, medido pela Fundação Getúlio Vargas continuamente no período entre 2011 e 2017, e o Ministério Público, que era a terceira organização mais confiável aos olhos dos inquiridos em 2011, com percentual de 52%, passou a ser a sétima mais confiável, com percentual de apenas 28% em 2017 (CUNHA et al., 2011, 2017). Não haveria como se propor que há uma relação de causalidade entre o giro de prioridades e essa queda na confiança, no entanto, essa constitui uma agenda de pesquisa fundamental para o campo de estudos e que merece ser explorada. 75 75 p p q p De mais a mais, apesar de ser possível identificar essa tendência geral de mudança institucional gradual relativa ao giro de prioridades, essa constatação não dispensa a necessidade de análises particularistas, com o objetivo de averiguar como essa dinâmica de giro de prioridades opera em cada um dos órgãos ministeriais e como as regras informais que permitiram esse movimento se construíram em contexto local. 5. CONSIDERAÇÕES FINAIS A discussão empreendida neste artigo, portanto, indica que a hipótese do giro de prioridades proposta pela literatura tem suporte empírico no desempenho funcional do MP brasileiro, inclusive no âmbito dos órgãos ministeriais estaduais. No marco da tipologia proposta pela literatura sobre mudanças institucionais, trata-se de um tipo de mudança institucional informal, gradual e transformativa do tipo deslocamento (STREECK; THELEN, 2005, p. 119-29; MAHONEY; THELEN, 2010, p. 15-22), que é, aliás a nomenclatura utilizada por Arantes (2011, p. 122-123). É dizer: já havia uma regra informal, porque não estabelecida em nenhum regramento legal ou infralegal relativo ao MP, indicando uma priorização da tutela coletiva e teria havido uma substituição por um comando, também informal, orientado para a priorização da repressão penal anticorrupção. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? É possível, inclusive, que algumas dessas corporações destoem da hipótese geral e, ainda, é muito provável que a intensidade e o tempo de eventuais mudanças associadas a essa mudança institucional se apresentem de forma variada. Este esforço, todavia, reclama frentes de pesquisa específicas e, certamente, um maior detalhamento dos dados de atuação finalística (como os produzidos pelo CNMP), especialmente no âmbito criminal e com uma abordagem não só em aspectos quantitativos, além de um levantamento de recursos materiais e humanos alocados pelos diferentes MPs em uma ou outra temática ao longo de uma mesma sequência temporal. De mais a mais, não há nenhuma evidência de que o giro de prioridades da tutela coletiva em direção à repressão criminal anticorrupção constitua uma mudança institucional definitiva no âmbito do MP brasileiro. Aliás, neste ponto reside a 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA ORCID 0000-0002-6655-2735 ORCID 0000-0002-6995-0982 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA vantagem na abordagem analítica esteada no novo institucionalismo histórico. A mera constatação de que o MP brasileiro, hoje, se encontra mais voltado para um discurso e uma atuação voltada à aplicação do direito penal na repressão anticorrupção, embora relevante, apenas apresenta a questão de forma parcial. vantagem na abordagem analítica esteada no novo institucionalismo histórico. A mera constatação de que o MP brasileiro, hoje, se encontra mais voltado para um discurso e uma atuação voltada à aplicação do direito penal na repressão anticorrupção, embora relevante, apenas apresenta a questão de forma parcial. No fundo, a principal contribuição do conjunto de achados discutido neste trabalho é oferecer subsídios fáticos para propor que o perfil institucional do MP brasileiro lhe provê uma dimensão potencialmente mimética, isto é, uma singular plasticidade para um constante reposicionamento de agenda a depender de fatores externos e internos mobilizados, sobretudo pelos seus próprios membros. Com o recente arrefecimento da agenda anticorrupção no debate político nacional, a significativa perda de capital político pelo MP, o advento de modificações legislativas no microssistema de prevenção e repressão a ilícitos anticorrupção, entre outros elementos, pode-se esperar, brevemente, a afirmação de novos elementos no seio da agenda de prioridades do MP brasileiro. ARANTES, Rogério Bastos. Direito e política: o Ministério Público e a defesa dos direitos coletivos. Revista Brasileira de Ciências Sociais, v. 14, n. 18, 1999, p. 83- 102. ARANTES, Rogério Bastos. Direito e política: o Ministério Público e a defesa dos direitos coletivos. Revista Brasileira de Ciências Sociais, v. 14, n. 18, 1999, p. 83- 102. ARANTES, Rogério B. Ministério Público e Política no Brasil. São Paulo: Sumaré, 2002. PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? ARANTES, Rogério Bastos. Direito e política: o Ministério Público e a defesa dos direitos coletivos. Revista Brasileira de Ciências Sociais, v. 14, n. 18, 1999, p. 83- 102. REFERÊNCIAS ABRUCIO, Fernando Luiz. Trajetória recente da gestão pública brasileira: um balanço crítico e a renovação da agenda de reformas. Revista de Administração Pública, v. 41., n.spe, 2007. 76 ABRUCIO, Fernando Luiz; VIEGAS, Rafael Rodrigues; RODRIGUES, Rayane Vieira. A agenda esquecida do federalismo brasileiro: assimetria, heterogeneidade e diversidade dos Ministérios Públicos. Preprint. Scielo, 2021. Disponível em https://doi.org/10.1590/SciELOPreprints.2669 Acesso em 8 jul. 2022. ACRE. Ato n. 019, de 26 de agosto de 2019. Institui o Grupo de Atuação Especial de Combate à Corrupção do Ministério Público do Estado do Acre e disciplina a sua forma de funcionamento. Disponível em <https://www.mpac.mp.br/wp- content/uploads/ATO-PGJ-019-2019_-_Institui_o_GAECC.pdf>. Acesso em 18 dez 2022. ARANHA, Ana Luiza Melo. A rede brasileira de instituições de accountability: um mapa do enfrentamento da corrupção na esfera local. Tese (Doutorado em Ciência Política). Programa de Pós-Graduação em Ciência Política – Universidade Federal de Minas Gerais. Belo Horizonte, 2015. ARANHA, Ana Luiza; FILGUEIRAS, Fernando. Instituições de accountability no Brasil: mudança institucional, incrementalismo e ecologia processual. Cadernos, 44. Brasília: Enap, 2016. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 ARANTES, Rogério B. Ministério Público e Política no Brasil. São Paulo: Sumaré, 2002. ARANTES, Rogério . B. Polícia Federal e construção institucional. In: AVRITZER, Leonardo; FILGUEIRAS, F. (Ed.). Corrupção e Sistema Político no Brasil. Rio de Janeiro: Civilização Brasileira, 2011. ARANTES, Rogério Bastos; MOREIRA, Thiago de Miranda Queiroz. Democracia, instituições de controle e justiça sob a ótica do pluralismo estatal. Opinião Pública [online]. v. 25, n. 1, p. 97-135, 2019. AVRITZER, Leonardo; MARONA, Marjorie. A tensão entre soberania e instituições de controle na democracia brasileira. Dados, v. 60, n. 2, p. 359-393, 2017. AZEVEDO, Rodrigo Ghiringhelli de. O Ministério Público no Brasil. In: DIAS, João Paulo; AZEVEDO, Rodrigo Ghiringhelli de (org). O papel do Ministério Público: estudo comparado dos países latino-americanos. Coimbra: Almedina, 2008, p. 223-249. 77 77 BARROSO, Luís Roberto. Empurrando a história: combate à corrupção, mudança de paradigmas e refundação do Brasil (prefácio). In: PINOTTI, Maria Cristina. Corrupção: Lava Jato e Mãos Limpas (org. Maria Cristina Pinotti). São Paulo: Portfolio-Penguin, 2019. BARROSO, Luís Roberto. Empurrando a história: combate à corrupção, mudança de paradigmas e refundação do Brasil (prefácio). In: PINOTTI, Maria Cristina. Corrupção: Lava Jato e Mãos Limpas (org. Maria Cristina Pinotti). São Paulo: Portfolio-Penguin, 2019. BRASIL. Conselho Nacional do Ministério Público. Resolução CNMP n. 84/2011, de 19 de julho de 2011. Dispõe sobre a aplicação do controle da atuação da gestão de pessoas, da Tecnologia da Informação, da gestão estrutural, da gestão orçamentária do Ministério Público, bem como da atuação funcional de seus Membros. Disponível em https://www.cnmp.mp.br/portal/images/Resolucoes/Resoluo-0742.pdf Acesso em 07 jan 2021. BRASIL. Conselho Nacional do Ministério Público. Resolução n. 146, de 21 de junho de 2016a. Dispõe sobre as diretrizes administrativas para a formação de Membros e servidores do Ministério Público. Cria no âmbito do Conselho Nacional do Ministério Público, a Unidade Nacional de Capacitação do Ministério 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 ORCID 0000-0002-6655-2735 ORCID 0000-0002-6995-0982 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA Público. Brasília, DF. Disponível em https://www.cnmp.mp.br/portal/images/Normas/Resolucoes/RESOLUO_146_assi nado.pdf Acesso em 28 fev 2023. BRASIL. Conselho Nacional do Ministério Público. Recomendação n. 42, de 5 de setembro de 2016b. Recomenda a criação de estruturas especializadas no Ministério . Conselho Público para a otimização enfrentamento à corrupção, com atribuição cível e criminal. Brasília: CNMP, 2016. ç p Ministério . Conselho Público para a otimização enfrentamento à corrupção, com atribuição cível e criminal. Brasília: CNMP, 2016. ARANTES, Rogério B. Ministério Público e Política no Brasil. São Paulo: Sumaré, 2002. BRASIL Nacional do Ministério Público. Resolução n. 185, de 2 de março de 2018 Dispõe sobre a criação da Comissão Especial de. Enfrentamento à Corrupção. . Brasília: CNMP, 2018. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 Brasília: CNMP, 2018. BRASIL. Conselho Nacional do Ministério Público. Ministério Público: um retrato 2014-2021. Brasília, DF, 2021a. Disponível em https://www.cnmp.mp.br/portal/relatoriosbi/mp-um-retrato-2021 Acesso em 08 mai 2022. BRASIL. Ministério Público Federal. GAECO. 2021b. Disponível em: http://www.mpf.mp.br/atuacao-tematica/gaeco. Acesso em: 27 maio 2021. BRASIL. Conselho Nacional de Justiça. Estatísticas do Poder Judiciário. 2023a. Disponível em https://painel-estatistica.stg.cloud.cnj.jus.br/estatisticas.html Acesso em 08 mar 2023. BRASIL. Controladoria-Geral da União. Histórico. 2023a. Disponível em https://www.gov.br/cgu/pt-br/acesso-a- informacao/institucional/historico/historico- 2#:~:text=Em%202003%2C%20a%20lei%20de,a%20defesa%20do%20patrim%C3% B4nio%20p%C3%BAblico Acesso em 18 abr 2023. BRASIL. Rede de Controle da Gestão Pública. Sobre a Rede. 2023b. Disponível em https://www.rededecontrole.gov.br/quem-somos/ Acesso em 18 abr 2023. CARVALHO, Ernani.; LEITÃO, Natália. O novo desenho institucional do Ministério Público e o processo de judicialização da política. Revista Direito GV, v. 6, n. 2, p. 399-422, dez. 2010. CARVALHO, Ernani; LEITÃO, Natália. Operationalizing and Measuring Prosecutorial . Independence: the Brazilian Case. In: INGRAM, Matthew C.; BRASIL. Conselho Nacional do Ministério Público. Ministério Público: um retrato 2014-2021. Brasília, DF, 2021a. Disponível em https://www.cnmp.mp.br/portal/relatoriosbi/mp-um-retrato-2021 Acesso em 08 mai 2022. BRASIL. Ministério Público Federal. GAECO. 2021b. Disponível em: http://www.mpf.mp.br/atuacao-tematica/gaeco. Acesso em: 27 maio 2021. 78 BRASIL. Conselho Nacional de Justiça. Estatísticas do Poder Judiciário. 2023a. Disponível em https://painel-estatistica.stg.cloud.cnj.jus.br/estatisticas.html Acesso em 08 mar 2023. BRASIL. Controladoria-Geral da União. Histórico. 2023a. Disponível em https://www.gov.br/cgu/pt-br/acesso-a- informacao/institucional/historico/historico- 2#:~:text=Em%202003%2C%20a%20lei%20de,a%20defesa%20do%20patrim%C3% B4nio%20p%C3%BAblico Acesso em 18 abr 2023. CARVALHO, Ernani.; LEITÃO, Natália. O novo desenho institucional do Ministério Público e o processo de judicialização da política. Revista Direito GV, v. 6, n. 2, p. 399-422, dez. 2010. CARVALHO, Ernani.; LEITÃO, Natália. O novo desenho institucional do Ministério Público e o processo de judicialização da política. Revista Direito GV, v. 6, n. 2, p. 399-422, dez. 2010. CARVALHO, Ernani; LEITÃO, Natália. Operationalizing and Measuring Prosecutorial . Independence: the Brazilian Case. In: INGRAM, Matthew C.; CARVALHO, Ernani; LEITÃO, Natália. Operationalizing and Measuring Prosecutorial . Independence: the Brazilian Case. In: INGRAM, Matthew C.; 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? KAPISZEWSKI, Diana (org.). Beyond High Courts: The Justice Complex in Latin American. Notre Dame: University of Notre Dame, 2019, p. 51-68. CASTILHO, Ela Wiecko V. de; SADEK, Maria Tereza. O Ministério Público e a administração da Justiça no Brasil. São Paulo: Idesp, Sumaré, 1998. CEARÁ. Provimento n. 093, de 19 de novembro de 2018. Cria, na estrutura da Procuradoria-Geral de Justiça do Estado do Ceará, o Grupo Especial de Combate à Corrupção - GECOC. Disponível em <http://www.mpce.mp.br/wp- content/uploads/2018/10/Provimento-n%C2%BA-093-2018-Grupo-Especial-de- Combate-%C3%A0-Corrup%C3%A7%C3%A3o.pdf> Acesso em 18 dez 2022. CHALOUB, Jorge; LYNCH, Christian Edward Cyril. O pensamento político- constitucional da República de 1988: um balanço preliminar. In: HOLANDA, C. B. de; VEIGA et al (Org.). A Constituição de 88 trinta anos depois. Curitiba: Editora da UFPR, 2018. v. 1. p. 251-280. CNPG. Conselho Nacional de Procuradores-Gerais. Quem somos. 2022. Disponível em https://www.cnpg.org.br/quem-somos/o-que-e-o-cnpg.html Acesso em 29 ago 2022. Disponível em https://www.cnpg.org.br/quem-somos/o-que-e-o-cnpg.html Acesso em 29 ago 2022. 79 79 CONAMP. Histórico. 2022. Disponível em. https://www.conamp.org.br/institucional/historico.html. Acesso 18 fev 2023. COSTA, Arthur Trindade Maranhão; OLIVEIRA JÚNIOR, Almir de. Novos padrões de investigação policial no Brasil. Sociedade e Estado, v. 31, n. 1, p. 147– 164, jan. 2016. CUNHA, Luciana Gross; BUENO, Rodrigo de Losso da Silveira; OLIVEIRA, Fabiana Luci de Sampaio; RAMOS, Joelson Oliveira; KLINK, Luciana de Oliveira; CAMPOS, Yuri. Relatório com os dados da pesquisa Índice de Confiança na Justiça (ICJBrasil) referente ao 4º trimestre de 2012. São Paulo: FGV, 2013. Disponível em < https://bibliotecadigital.fgv.br/dspace/handle/10438/10754> Acesso em 04 mar 2019. DIAS, Paulo Thiago Fernandes; CALLEGARI, André Luiz; ZAGHLOUT, Sara Alacoque Guerra. As operações de combate à corrupção no Brasil e o impacto nas ciências criminais. Direito Público, [S.l.], v. 17, n. 93, jul. 2020. DIAS, Paulo Thiago Fernandes; CALLEGARI, André Luiz; ZAGHLOUT, Sara Alacoque Guerra. As operações de combate à corrupção no Brasil e o impacto nas ciências criminais. Direito Público, [S.l.], v. 17, n. 93, jul. 2020. DIMAGGIO, Paul J.; POWELL, Walter W. The Iron Cage Revisited: Institutional Isomorphism and Collective Rationality in Organizational Field. American Sociological Review, v. 48, n. 2, abr./1983, pp. 147-160. DIMAGGIO, Paul J.; POWELL, Walter W. The Iron Cage Revisited: Institutional Isomorphism and Collective Rationality in Organizational Field. American Sociological Review, v. 48, n. 2, abr./1983, pp. 147-160. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? 2024 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA ORCID 0000-0002-6655-2735 ORCID 0000-0002-6995-0982 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA ENCCLA. ENCCLA: 10 anos de organização do Estado brasileiro contra o Crime organizado. Brasília: Ministério da Justiça, 2012. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 ENCCLA. ENCCLA: 10 anos de organização do Estado brasileiro contra o Crime organizado. Brasília: Ministério da Justiça, 2012. In: Anais do Encontro PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? Internacional Participação, Democracia e Políticas Públicas, Porto Alegre, 2019. Disponível em: https://bit.ly/3xJ604j. Acesso em: 18 out. 2019. Internacional Participação, Democracia e Políticas Públicas, Porto Alegre, 2019. Disponível em: https://bit.ly/3xJ604j. Acesso em: 18 out. 2019. LAMENHA, Bruno.; LIMA, Flávia Danielle Santiago. Quem defenderá a sociedade? Trajetórias e competição institucional em torno da tutela coletiva entre Ministério Público e Defensoria no pós-1988. Espaço Jurídico Journal of Law, v. 22, n. 1, p. 73-104, jun./2021. LAMOUNIER, Bolívar; MARREY, Luiz Antônio; GRINOVER, Ada Pellegrini; BENEDTI, Sidnei; PAZZAGLINI FILHO, Marino. Primeira Sessão: o papel do Ministério Público entre as Instituições que compõem o Sistema Brasileiro de Justiça. In: SADEK, Maria Tereza (org.) O Ministério Público e a Justiça no Brasil. São Paulo: IDESP, Sumaré, 1996. LEMGRUBER, Julita; RIBEIRO, Ludmila; MUSUMECI, Leonarda; DUARTE, Thais. ENCCLA. ENCCLA: 10 anos de organização do Estado brasileiro contra o Crime organizado. Brasília: Ministério da Justiça, 2012. ENCCLA. Estrutura. 2023. Disponível em http://enccla.camara.leg.br/quem- somos/estrutura Acesso em 18 abr 2023. ENGELMANN, F. Movimentos internacionais anticorrupção, lutas políticas e judiciais no espaço nacional. In: ENGELMANN, F.; PILAU, L. S. B. Justiça e poder político: elites jurídicas, internacionalização e lutas anticorrupção. Porto Alegre: UFRGS/CEGOV, 2021. ENGELMANN, F. Movimentos internacionais anticorrupção, lutas políticas e judiciais no espaço nacional. In: ENGELMANN, F.; PILAU, L. S. B. Justiça e poder político: elites jurídicas, internacionalização e lutas anticorrupção. Porto Alegre: UFRGS/CEGOV, 2021. FAVERI, Diego de; PEREIRA, Carlos, GOLDSZMIDT, Rafael. Cachorro Grande Precisa de Coleira Forte: Determinantes do Ativismo dos Ministérios Públicos Estaduais no Brasil. XXXVIII Encontro da ANPAD. Anais Digitais. Rio de Janeiro, 2014. Disponível em < http://anpad.com.br/pt_br/article_search/archive?search%5Bq%5D=Cachorro%20 Grande&search%5Bsubmit%5D= > Acesso em 01 dez 2021. FILGUEIRAS, Fernando; ARANHA, Ana L. M. Controle da corrupção e burocracia da linha de frente: regras, discricionariedade e reformas no Brasil. Dados, v. 54, n. 2, p. 349-387, 2011. 80 HACKER, Jacob S.; PIERSON, Paul; THELEN, Kathleen. Drift and Conversion: Hidden faces of institucional change. In THELEN, Kathleen; MAHONEY, James (org). Advances in Comparative-Historical Analysis: Cambridge University Press, p. 180-208. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 , p KERCHE, Fábio. Virtude e limites: autonomia e atribuições do Ministério Público no Brasil. São Paulo: EdUSP, 2009. KERCHE, Fábio; MARONA, Marjorie. A política no banco dos réus: a Operação Lava Jato e a erosão da democracia no Brasil. Belo Horizonte: Autêntica, 2022. Edição Kindle. KERCHE, Fábio; VIEGAS, Rafael R. O Ministério Público: de defensor de direitos a combate da corrupção. In: Anais do 12º Encontro da ABCP, 2020. LAMENHA, Bruno. A tarefa de defesa do regime democrático confiada ao Ministério Público brasileiro na Nova República: perspectivas a partir da noção de “contrademocracia”de Pierre Rosanvallon. In: Anais do Encontro KERCHE, Fábio. Virtude e limites: autonomia e atribuições do Ministério Público no Brasil. São Paulo: EdUSP, 2009. KERCHE, Fábio; MARONA, Marjorie. A política no banco dos réus: a Operação Lava Jato e a erosão da democracia no Brasil. Belo Horizonte: Autêntica, 2022. Edição Kindle. KERCHE, Fábio; VIEGAS, Rafael R. O Ministério Público: de defensor de direitos a combate da corrupção. In: Anais do 12º Encontro da ABCP, 2020. LAMENHA, Bruno. A tarefa de defesa do regime democrático confiada ao Ministério Público brasileiro na Nova República: perspectivas a partir da noção de “contrademocracia”de Pierre Rosanvallon. Ministério Público: guardião da democracia brasileira? Rio de Janeiro: CESeC, 2016. Ministério Público: guardião da democracia brasileira? Rio de Janeiro: CESeC, 2016. Disponível em: https://cesecseguranca.com.br/livro/ministerio-publico-guardiao- da- democracia-brasileira. Acesso em: 16 abr. 2020. 81 81 LEITE, Ruano Fernando da Silva. O que move o Ministério Público? Análise do processo de desenvolvimento de uma identidade institucional no enfrentamento da corrupção. Dissertação (Mestrado em Segurança Pública, Justiça e Cidadania). Salvador: Universidade Federal da Bahia, 2020. LIMA, Flávia Danielle Santiago; LAMENHA, Bruno. Mobilizando a agenda dos direitos coletivos, assegurando espaço institucional: Ministério Público e Defensoria Pública na transição democrática. Revista de Informação Legislativa: RIL, Brasília, DF, v. 58, n. 231, p. 87-108, jul./set. 2021. LIMA, Flávia Danielle Santiago; MACHADO, Igor Suzano. Carta da apresentação do dossiê. Revista de Direito Público, 19, n. 101, jan./mar 2022, p. 5-19. LIZIERO, Leonam Baesso da Silva; CARVALHO, Fabrício. Federalismo e centralização no Brasil: contrastes na construção da Federação Brasileira / Federalism and centralization in Brazil: contrasts in the construction of the Brazilian Federation. Revista de Direito da Cidade, [S.l.], v. 10, n. 3, p. 1483-1503, ago. 2018 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA ORCID 0000-0002-6655-2735 ORCID 0000-0002-6995-0982 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA LONDERO. Daiane. O desenvolvimento de capacidades institucionais do Ministério Público Federal no combate à corrupção. Tese (Doutorado em Políticas Públicas). Programa de Pós-Graduação em Políticas Públicas, Universidade Federal do Rio Grande do Sul, 2021. LONDERO. Daiane. O desenvolvimento de capacidades institucionais do Ministério Público Federal no combate à corrupção. Tese (Doutorado em Políticas Públicas). Programa de Pós-Graduação em Políticas Públicas, Universidade Federal do Rio Grande do Sul, 2021. MACEDO JÚNIOR, Ronaldo Porto. A evolução institucional do ministério público brasileiro. In: SADEK, Maria Tereza (org.) Uma introdução ao estudo da Justiça. São Paulo: Sumaré, 1995. MAHONEY, James; THELEN, Kathleen. A theory of gradual institutional change. In: MAHONEY, J.; THELEN, K. (Ed.). Explaining institutional change: ambiguity, agency and power. New York: Cambridge University Press, 2010. p. 1-37. MAINWARING, Scott. Introduction: Democratic Accountability in Latin America. In MAINWARING, Scott; WELNA, Christopher. Democratic Accountability in Latin America. Oxford: Oxford University Press, p. 2003. MCCOY, Jennifer L.; HECKEL, Heather. The emergence of global anti-corruption norm. International Politics, v. 38, p. 65-90, mar. 2001. 82 MARONA, Marjorie; BARBOSA, Leon Queiroz. (2018), “Protagonismo judicial no Brasil: MARONA, Marjorie; BARBOSA, Leon Queiroz. (2018), “Protagonismo judicial no Brasil: do que estamos falando?”. Ministério Público: guardião da democracia brasileira? Rio de Janeiro: CESeC, 2016. In: MARONA, M.; DEL RIO, A. (orgs.), Justiça no Brasil: às margens da democracia. Belo Horizonte: Arraes Editores, pp. 128-15 g da democracia. Belo Horizonte: Arraes Editores, pp. 128-15 MARONA, Marjorie; KERCHE, Fábio. From the Banestado Case to Operation Car Wash: Building an Anti-Corruption Institutional Framework in Brazil. Dados, v. 64, n. 3, p. 1-37, 2021. MATO GROSSO DO SUL. Resolução n. 08, de 26 de outubro de 2016. Dispõe sobre a criação do Grupo Especial de Combate à Corrupção (GECOC) e dá outras providências.Disponível em <https://www.mpms.mp.br/atos-e- normas/download/32195> Acesso em 18 dez 2022. MAZZILLI, Hugo Nigro. Introdução ao Ministério Público. 9 ed. São Paulo: Saraiva, 2015. PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? tps://www.metropoles.com/brasil/advogada-e-policiais-civis-sao-suspeitos-de- b A 07 b 2023 PLURALIDADE E MUDANÇA INSTITUCIONAL: <https://www.metropoles.com/brasil/advogada-e-policiais-civis-sao-suspeitos-de- corrupcao-e-extorsao-na-pb> Acesso em 07 abr 2023. NÓBREGA, Flavianne Fernanda Bitencourt. Entre o Brasil formal e o Brasil real: Ministério Público no Brasil, instituição para o fortalecimento do Estado de direito? Dissertação (Mestrado em Ciência Política) – Departamento de Ciência Polícia, Universidade Federal de Pernambuco, Recife, 2007. NÓBREGA, Flavianne Fernanda Bitencourt. Entre o Brasil formal e o Brasil real: Ministério Público no Brasil, instituição para o fortalecimento do Estado de direito? Dissertação (Mestrado em Ciência Política) – Departamento de Ciência Polícia, Universidade Federal de Pernambuco, Recife, 2007. OLIVEIRA, Vanessa Elias de; LOTTA, Gabriela; VASCONCELOS, Natália Pires de. Ministério Público, autonomia funcional e discricionariedade: ampla atuação em políticas públicas, baixa accountability. Revista de Estudos Empíricos em Direito: REED, São Paulo, SP, v.7., n. 1., p. 181-195, abr. 2020. OLIVEIRA, Vanessa Elias de; LOTTA, Gabriela; VASCONCELOS, Natália Pires de. Ministério Público, autonomia funcional e discricionariedade: ampla atuação em políticas públicas, baixa accountability. Revista de Estudos Empíricos em Direito: REED, São Paulo, SP, v.7., n. 1., p. 181-195, abr. 2020. ONOMA, Ato Kwamena. The Contradictory Potential of Institutions: The Rise andDecline of Land Documentation in Kenya. In: MAHONEY, James; THELEN, Kathleen. (ed.). Explaining institutional change: ambiguity, agency, and power. New York: Cambridge University Press, 2010. p. 64-93. MAZZILLI, Hugo Nigro. Introdução ao Ministério Público. 9 ed. São Paulo: Saraiva, 2015. METRÓPOLES. Advogada e policiais civis são suspeitos de corrupção e extorsão na PB. 10 de março de 2023. Disponível em METRÓPOLES. Advogada e policiais civis são suspeitos de corrupção e extorsão na PB. 10 de março de 2023. Disponível em 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 SABELLA, Walter Paulo; DAL POZZO, Antônio Araldo Ferraz; BURLE FILHO, J é SABELLA, Walter Paulo; DAL POZZO, Antônio Araldo Ferraz; BURLE FILHO, José SABELLA, Walter Paulo; DAL POZZO, Antônio Araldo Ferraz; BURLE FILHO, José Emmanuel. Memórias das lutas pela obtenção do texto do Ministério Público na Constituição de 1988. In: SABELLA, Walter Paulo; DAL POZZO, Antônio Araldo Ferraz; BURLE FILHO, José Emmanuel (coord.). Ministério Público: vinte e cinco anos do novo perfil constitucional. São Paulo: Malheiros, 2013. p. 21-181. Emmanuel. Memórias das lutas pela obtenção do texto do Ministério Público na Constituição de 1988. In: SABELLA, Walter Paulo; DAL POZZO, Antônio Araldo Ferraz; BURLE FILHO, José Emmanuel (coord.). Ministério Público: vinte e cinco anos do novo perfil constitucional. São Paulo: Malheiros, 2013. p. 21-181. SADEK, Maria Tereza. O Ministério Público e a Justiça no Brasil. São Paulo: Idesp, Sumaré, 1996. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PALUDO, Januário (coord.). Forças-tarefas: direito comparado e legislação aplicável. Brasília: ESMPU, 2011. PALUDO, Januário (coord.). Forças-tarefas: direito comparado e legislação aplicável. Brasília: ESMPU, 2011. 83 83 PENNA, Luciana Rodrigues. Novos padrões na formação das elites de Estado: educação jurídica corporativa e o caso do Ministério Público. In: ENGELMANN, F. et al. Justiça e poder político: elites jurídicas, internacionalização e lutas anticorrupção. Porto Alegre: UFRGS/CEGOV, 2021. PERTENCE, José Paulo Sepúlveda. Projeto História Oral: Memória do Ministério Público Federal [Entrevista concedida a José Walter Nunes]. Brasília: 2005. PIAUÍ. Ministério Público do Piauí. Operação Cerrados: GAECO/MPPI realiza operação contra desvios de recursos públicos no município de Uruçuí. 20 de março de 2023. Disponível em https://www.mppi.mp.br/internet/2023/03/operacao-cerrados-gaeco-mppi-realiza- operacao-contra-desvios-de-recursos-publicos-no-municipio-de-urucui/ Acesso em 04 abr 2023. PIAUÍ. Ministério Público do Piauí. Operação Cerrados: GAECO/MPPI realiza operação contra desvios de recursos públicos no município de Uruçuí. 20 de março de 2023. Disponível em https://www.mppi.mp.br/internet/2023/03/operacao-cerrados-gaeco-mppi-realiza- operacao-contra-desvios-de-recursos-publicos-no-municipio-de-urucui/ Acesso em 04 abr 2023. RAMOS, Luciana Oliveira; CUNHA, Luciana Gross; BUENO, Rodrigo de Losso da Silveira; OLIVEIRA, Fabiana Luci de Sampaio; RAMOS, Joelson Oliveira; ÚBIDA, Giovanna. Relatório com os dados da pesquisa Índice de Confiança na Justiça (ICJBrasil) referente ao 1º semestre de 2017. São Paulo: FGV, 2013. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 ORCID 0000-0002-6655-2735 ORCID 0000-0002-6995-0982 BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA Disponível em https://bibliotecadigital.fgv.br/dspace/handle/10438/19034 Acesso em 04 mar 2019. RIBEIRO, Ludmila Mendonça Lopes. Ministério Público: velha instituição com novas funções? Revista Crítica de Ciências Sociais, [Coimbra], n. 113, p. 51-81, set. 2017. Disponível em: https://journals.openedition.org/rccs/6654. Acesso em: 16 abr. 2021. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? PLURALIDADE E MUDANÇA INSTITUCIONAL: HÁ EVIDÊNCIAS PARA SUSTENTAR A HIPÓTESE DO GIRO DE PRIORIDADES NO DESEMPENHO INSTITUCIONAL DO MINISTÉRIO PÚBLICO BRASILEIRO? SÃO PAULO. Ministério Público de São Paulo. Gaeco e Corregedoria miram corrupção policial em Osasco com Operação Haalapenz. 03 jun 2022. Disponível em https://www.mpsp.mp.br/w/com-opera%C3%A7%C3%A3o-haalapenz-gaeco- e-corregedoria-miram-corrup%C3%A7%C3%A3o-policial-em-osasco. Acesso em 08 jan 2023. SCOTT, W. Richard. Institutions and Organizations: Ideas, Interests, and Identities. Fourth Edition. Thousand Oaks: SAGE, 2014. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 SADEK, Maria Tereza. O Ministério Público e a Justiça no Brasil. São Paulo: Idesp, Sumaré, 1996. SADEK, Maria Tereza. Cidadania e Ministério Público. In SADEK, Maria Tereza (org.). Justiça e cidadania no Brasil. São Paulo: Idesp: Sumaré, 2000. Edição Kindle. 84 SADEK, Maria Tereza. O sistema de justiça. In: SADEK, Maria Tereza (org.). O sistema de justiça [online]. Rio de Janeiro: Centro Edelstein de Pesquisas Sociais, 2010. Disponível em SciELO Books http://books.scielo.org. Acesso em 8 dez 2020. SAMPAIO, Marianna; VIEGAS, Rafael Rodrigues. Ministério Público: de fiscal a elaborador de políticas públicas. In: Encontro da ANPOCS, 43., Caxambu. Anais [...] Caxambu, 2019. SANTA CATARINA. Ministério Público do estado de Santa Catarina. Gaeco, Geac e Polícia Civil cumprem mandados de busca e apreensão em operação contra corrupção em Urussanga. 02 de dezembro de 2022. Disponível em <https://www.mpsc.mp.br/noticias/gaeco-geac-e-policia-civil-cumprem- mandados-de-busca-e-apreensao-em-operacao-contra-corrupcao-em-urussanga> Acesso em 07 abr 2023. SANTA CATARINA. Ministério Público do estado de Santa Catarina. Gaeco, Geac e Polícia Civil cumprem mandados de busca e apreensão em operação contra corrupção em Urussanga. 02 de dezembro de 2022. Disponível em <https://www.mpsc.mp.br/noticias/gaeco-geac-e-policia-civil-cumprem- mandados-de-busca-e-apreensao-em-operacao-contra-corrupcao-em-urussanga> Acesso em 07 abr 2023. SANTIAGO, Milena Brentini. O Ministério Público Federal e sua “vocação” de combate à corrupção: conflitos sociais e ideologia institucional. Revista de Ciências Sociais – Fortaleza, v. 53, n. 2, jul./out.,2022, p. 107-152. SANTIAGO, Milena Brentini. O Ministério Público Federal e sua “vocação” de combate à corrupção: conflitos sociais e ideologia institucional. Revista de Ciências Sociais – Fortaleza, v. 53, n. 2, jul./out.,2022, p. 107-152. 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024 SILVA, Cátia Aida. Justiça em Jogo. São Paulo: Edusp, 2001. ILVA, Cátia Aida. Justiça em Jogo. São Paulo: Edusp, 2001. SILVA, Leonardo Assis. Ministério Público estadual e combate à corrupção: padrão Lava Jato? Anais. 44º Encontro Anual da ANPOCS. 2020. SILVA, Philippe; ARAÚJO, Mateus Morais; RIBEIRO, Ludmila; MARONA, Marjorie. O Ministério Público no contexto atual: de protagonista a ombudsman? In: ENCONTRO ANUAL DA ANPOCS, 41., 2017, Caxambu. Anais [...] Caxambu, 2017. Disponível em: https://bit.ly/2AgtOlj; Acesso em: 30 set. 2018. ILVA, Philippe; ARAÚJO, Mateus Morais; RIBEIRO, Ludmila; MARONA, Marjorie. O Ministério Público no contexto atual: de protagonista a ombudsman? j p g n: ENCONTRO ANUAL DA ANPOCS, 41., 2017, Caxambu. Anais [...] axambu, 2017. Disponível em: https://bit.ly/2AgtOlj; Acesso em: 30 set. 2018. STREECK, Wolfgang; THELEN, Kathleen. Introduction: Institutional Change in Advanced Political Economies. In: STREECK, Wolfgang; THELEN, Kathleen (org.). Institutional Change in Advanced Political Economies. Oxford: Oxford University Press, 2005, p. 1-39. 85 85 SUZART, Janilson Antonio da Silva A; MARCELINO, Carolina Venturini; ROCHA, Joseílton Silveira da. da. As Instituições Brasileiras de Controladoria Pública – Teoria versus Prática. Contabilidade Gestão e Governança, Brasília-DF, v. 14, n. 1, 2011. Disponível em: https://revistacgg.org/index.php/contabil/article/view/265. Acesso em: 5 maio. 2023. UNODC. Group Judicial Integrity Network. About. 2023. Disponível em https://www.unodc.org/ji/en/about.html Acesso em 18 abr 2023. VIECILI, Jerusa B. Lava Jato e modernização da investigação criminal no Brasil. Brasília, DF: Associação Nacional de Procuradores da República. 2017. Disponível em: https://bit.ly/2r2HD4A. Acesso em: 29 out. 2019. VIECILI, Jerusa B. Lava Jato e modernização da investigação criminal no Brasil. Brasília, DF: Associação Nacional de Procuradores da República. 2017. Disponível em: https://bit.ly/2r2HD4A. Acesso em: 29 out. 2019. VIEGAS, Rafael Rodrigues; LOUREIRO, Maria Rita. A Lava Jato como ponto de inflexão sobre o Estado de Direito e a democracia no Brasil. Congreso Internacional de Ciencia Política de la Asociación Mexicana de Ciencias Políticas, ORCID 0000-0002-6655-2735 ORCID 0000-0002-6995-0982 BRUNO LAMENHA BRUNO LAMENHA BRUNO LAMENHA BRUNO LAMENHA FLÁVIA DANIELLE SANTIAGO LIMA FLÁVIA DANIELLE SANTIAGO LIMA 7. Anais [...] Guadalajara, 2020. Disponível em https://bit.ly/2Ur0Wnb. Acesso em 8 abr. 2021. VIEGAS, Rafael Rodrigues; BERLATTO, Fabia; MONGELÓS, Silvia Avelina Arias; LIMA, Débora Dossiati de. O espaço formal de ação do Ministério Público entre 1989 e 2016: mudanças incrementais e ativação estratégica. Revista de Direito Público, 19, n. 101, p. 428-449, jan./mar 2022. 7. Anais [...] Guadalajara, 2020. Disponível em https://bit.ly/2Ur0Wnb. Acesso em 8 abr. 2021. VIEGAS, Rafael Rodrigues; BERLATTO, Fabia; MONGELÓS, Silvia Avelina Arias; LIMA, Débora Dossiati de. O espaço formal de ação do Ministério Público entre 1989 e 2016: mudanças incrementais e ativação estratégica. Revista de Direito Público, 19, n. 101, p. 428-449, jan./mar 2022. 86 10 JOURNAL OF INSTITUTIONAL STUDIES 1 (2024) Revista Estudos Institucionais, v. 10, n. 1, p. 37 - 86, jan./abr. 2024
https://openalex.org/W2324494414	https://europepmc.org/articles/pmc6444734?pdf=render	English	null	A fully electronic microfabricated gas chromatograph with complementary capacitive detectors for indoor pollutants	Microsystems & nanoengineering	2,016	cc-by	10,859	INTRODUCTION systems have been demonstrated using microfabricated gas pumps and conventional vapor injectors15. Volatile organic compounds (VOCs) are common contributors to indoor air pollution. The species of interest include aromatic hydrocarbons, such as benzene, toluene and xylenes (BTX), as well as alkanes, halogenated compounds, alcohols, aldehydes, ketones and terpenes. These VOCs are typically found in higher concentrations in indoor environments than outdoors1,2. The exposure limits of these VOCs are typically in the range of 1–100 ppm3. Certain species constitute severe health hazards—for example, benzene is a known carcinogen4. In addition, each of the microscale components used in these μGCs has been fabricated by separate processes. The cost and complexity of pursuing multiple microfabrication processes has constrained the implementation of fully microfabricated μGC systems. Our group has previously demonstrated an approach for fabricating all the ﬂuidic components of a μGC using the same three-mask microfabrication process16,17. This approach included the integration of bi-directional gas pumps that allowed quantitative analysis without the use of the valves that are typically necessary to reverse the ﬂow direction between sampling and separation intervals. However, our prior work used gas detectors based on microdischarge emission spectroscopy. The use of spectroscopy favors the use of a spectrometer and the use of an inert carrier gas instead of room air. A detector with direct electrical readout that can be fabricated by the fabrication process common to all other components could substantially reduce the cost and complexity of the overall system. Although VOCs can be collectively detected by a variety of standalone gas detectors, these devices typically lack the selectivity to discriminate between constituents of complex mixtures5. Simple mixtures can be analyzed by mass spectrometry and infrared spectroscopy systems6,7, whereas complex mixtures typically require the use of gas chromatography systems8. For pervasive deployment in households, ofﬁces and public facilities, micro- fabrication offers a pathway to reducing system cost. A system deployed for long-term indoor VOC monitoring should ideally be capable of performing automated sampling and analyses without requiring consumables or frequent replacement of components. Power consumption is not a critical concern because power sockets are widely available in indoor environments. A number of detectors have been reported for use with μGC systems. Some have used an external optical interface18–21. OPEN Microsystems & Nanoengineering (2016) 2, 15049; doi:10.1038/micronano.2015.49 www.nature.com/micronano Center for Wireless Integrated MicroSensing and Systems (WIMS2), University of Michigan, Ann Arbor, MI 48109, USA. Correspondence: Yutao Qin (yutaoqin@umich.edu) Received: 10 September 2015; revised: 30 November 2015; accepted: 2 December 2015 ARTICLE A fully electronic microfabricated gas chromatograph with complementary capacitive detectors for indoor pollutants Yutao Qin and Yogesh B Gianchandani This paper reports a complete micro gas chromatography (μGC) system in which all the components are lithographically microfabricated and electronically interfaced. The components include a bi-directional Knudsen pump, a preconcentrator, separation columns and a pair of capacitive gas detectors; together, these form the iGC3.c2 system. All the ﬂuidic components of the system are fabricated by a common three-mask lithographic process. The Knudsen pump is a thermomolecular pump that provides air ﬂow to the μGC without any moving parts. The ﬁlm heaters embedded in the separation columns permit temperature programming. The capacitive detectors provide complementary response patterns, enhancing vapor recognition and resolving co- eluting peaks. With the components assembled on printed circuit boards, the system has a footprint of 8 × 10 cm2. Using room air as the carrier gas, the system is used to experimentally demonstrate the analysis of 19 chemicals with concentration levels on the order of parts per million (p.p.m.) and parts per billion (p.p.b.). The tested chemicals include alkanes, aromatic hydrocarbons, aldehydes, halogenated hydrocarbons and terpenes. This set of chemicals represents a variety of common indoor air pollutants, among which benzene, toluene and xylenes (BTX) are of particular interest. Keywords: Knudsen pump; preconcentrator; separation column; gas sensor; chemical sensing; air monitoring; BTX Microsystems & Nanoengineering (2016) 2, 15049; doi:10.1038/micronano.2015.49; Published online: 29 February 2016 Knudsen pump; preconcentrator; separation column; gas sensor; chemical sensing; air monitoring; BTX Keywords: Knudsen pump; preconcentrator; separation column; gas sensor; chemical sensing; air monitoring; BTX Microsystems & Nanoengineering (2016) 2, 15049; doi:10.1038/micronano.2015.49; Published online: 29 February 2016 s & Nanoengineering (2016) 2, 15049; doi:10.1038/micronano.2015.49; Published online: 29 February 2016 INTRODUCTION Others that rely on an electronic detection mechanism have used a thermal conductivity detector22, a discharge photoionization detector23, a heterodyne graphene detector24, a surface acoustic wave (SAW) detector11, a metal oxide semiconductor (MOX) detector12, a chemiresistor13 or capacitive detectors25–27. Whereas many detectors are compatible with the use of air as carrier gas, the ﬁrst three in the list of electronic detectors rely on a supply of helium as either a carrier gas or auxiliary gas. Within the context of the stated goals of micro gas chromatograph systems, capacitive detectors beneﬁt from a commercially available, low-cost Since the late 1970s, a variety of micro gas chromatography (μGC) systems have been reported, in which a subset of the necessary components have been microfabricated litho- graphically9. Most commonly, the separation column has been micromachined. Micromachined preconcentrators have been reported for sampling and injection of low-concentration VOCs into the columns10. However, almost all μGCs reported to date have used conventional gas valves and pumps11–14. Partial Fully electronic microfabricated gas chromatograph Y Qin and YB Gianchandani 2 Figure 1 iGC3.c2 system architecture and component design. Figure 1 iGC3.c2 system architecture and component design. chip-scale electronic interface and compatibility with a variety of simple fabrication processes. The design of the system and various components is described in Introduction, the fabrication is described in Materials and Methods, the experimental results are presented in Results, and a discussion and conclusions are presented in Discussion. Of the capacitive detectors that have been reported in the past, one approach has involved the use of a parallel plate structure. One implementation25,26 features a polymer ﬁlm sandwiched between a ﬂat bottom electrode plate and a perforated upper electrode plate that allows vapor to reach the polymer. A more common approach has been to use interdigitated electrodes on a substrate covered by a polymer ﬁlm27–30. The exposed polymer undergoes both swelling and changes in electrical permittivity. Both responses affect the capacitance. However, as elucidated in this work, these responses can be utilized in ways that are complementary and consequently can be used to separate certain co-eluting species. In contrast to parallel plate electrodes, interdigitated planar electrodes provide a greater surface area over which the sampled vapor can interact with the polymer, which generally leads to a faster response. In addition, the thickness of the polymer can be varied in the ﬁnal step, after lithographic fabrication is completed. Microsystems & Nanoengineering MATERIALS AND METHODS Design The gas ﬂow inlet and outlet are located at the upper end of the ‘U’ shape, which is cantilevered above a glass spacer, thermally isolating it from the substrate. This design relieves thermal stress and reduces heat loss. Two segments of sorbents, Carbopack-B and Carbopack-X (Sigma Aldrich, St Louis, MO, USA), are packed in series into the ‘U’-shaped channel to target a wide range of vapor species. Other types of sorbents can also be used, depending on the speciﬁc applications. pillar-like structures and glass beads that are packed against the pillars. A thin-ﬁlm metal heater is located in this region to provide localized heating for vapor desorption. The gas ﬂow inlet and outlet are located at the upper end of the ‘U’ shape, which is cantilevered above a glass spacer, thermally isolating it from the substrate. This design relieves thermal stress and reduces heat loss. Two segments of sorbents, Carbopack-B and Carbopack-X (Sigma Aldrich, St Louis, MO, USA), are packed in series into the ‘U’-shaped channel to target a wide range of vapor species. Other types of sorbents can also be used, depending on the speciﬁc applications. The separation columns are serpentine channels with a hydraulic diameter of 230 μm and a length of 30 cm per chip. The interior of these channels is coated with a 0.2-μm-thick OV-1 stationary phase (that is, 100% polydimethylsiloxane or PDMS, Ohio Valley Specialty, OH, USA). In this work, two identical column chips were used, providing a total length of 60 cm. The microfabricated columns are typically implemented in one of two shapes: a serpentine pattern and a spiral pattern. Whereas it has been reported that the serpentine pattern provides slightly higher performance than the spiral pattern34, in this work the serpentine pattern was used mainly because it provides a more uniform pressure distribution and a better immunity to leakage than does the spiral pattern (Figure 1). With a given pressure (ΔP) applied between the two ends of a serpentine column, the pressure is evenly distributed from one end of the chip to the other end. Therefore, the maximum pressure between any two adjacent channels is only a fraction of ΔP. In the spiral column, however, the pressure between two adjacent channels varies from zero in the chip center to ΔP near the chip perimeter. Fabrication We previously reported a three-mask process for fabricating various μGC components16. In that process, a glass wafer was ﬁrst metallized and subsequently micromachined. In the current study, metallization and micromachining were performed on different glass substrates that were later bonded together. Consequently, the micromachining and metallization steps were not performed sequentially. Although the new approach cannot be used for stacked architectures16,17 that require through holes to provide vertical gas ﬂow paths, it is feasible for the planar arrangement of components described in this study. The basic elements of each of the system components are fabricated from glass wafers by a common three-mask process. One mask is used to pattern a thin ﬁlm of Ti/Pt (25/100 nm) for the heaters, thermistors and capacitor electrodes. The other two masks are used for sandblasting (available as a service from Bullen Ultrasonics, Inc., OH, USA). One mask deﬁnes gas ﬂow channels that are sandblasted to a depth of 250 μm on one side of a glass wafer; the other mask deﬁnes through perforations that are sandblasted from the other side of the wafer to make chip-to-chip ﬂuidic connections. The sandblasting process typically creates a 22° sidewall taper angle and a 125 μm corner radius16. The capacitive detectors are formed using closely spaced interdigitated electrodes (Figure 1). The electrodes are covered by a layer of OV-1, which is also used as the stationary phase in the separation columns. Upon exposure to VOCs, this polymer undergoes changes in both thickness and permittivity. In principle, detector performance can be improved by downscaling various feature dimensions. Because the detector output signal is capacitance change (ΔC), a smaller interelectrode gap provides a higher ΔC. A smaller electrode width permits a denser integration of electrodes, which also provides a higher ΔC per unit area. In addition, a thinner polymer layer provides a faster response because the detector response time is mainly constrained by the speed of vapor absorption and desorption in the polymer layer. The interdigitated electrodes have widths and gaps of 1 μm because of the lithographic limits of the equipment available for this project. Two detectors with different polymer thicknesses are connected in series along the ﬂuidic path, downstream of the separation column. The ﬁrst detector, CapDet1, uses an OV-1 layer with a thickness of 0.25 μm. MATERIALS AND METHODS Design The iGC3.c2 system is composed of a chromatography module and a pump module (Figure 1). The chromatography module features microfabricated chips that include a two-stage preconcentrator, two separation columns and two complementary capacitive detectors. The pump module includes only a two-stage, bi- directional Knudsen pump. Gas is pumped against a temperature gradient established across a mesoporous polymer material within which transport is limited to free-molecular or transitional ﬂow regimes by the small pores31,33. The modules can be readily reconﬁgured with more separation columns and more Knudsen pump stages. g p p This paper reports a full μGC system, the iGC3.c2 system, that integrates a microfabricated pump, a preconcentrator, separation columns and a pair of capacitive detectors that provide distinguishable responses to permittivity changes and swelling. The integrated pump is a Knudsen pump that operates by thermal transpiration31; it requires no moving parts, thereby providing high reliability. The preconcentrator uses a cantilever structure that alleviates thermal stress. A pair of separation columns use a non-polar stationary phase to separate BTX and other VOCs. All the components are fabricated using a common three-mask process that is a simpliﬁed version of a process described in our prior work16. The resulting system is a complete, easily manu- facturable μGC that does not rely on external actuators, optical components, or carrier gas supply—the type of system well suited for long-term, continuous, automated monitoring of indoor pollutants. Preliminary results have been previously reported in the form of a conference abstract32. During vapor sampling, the Knudsen pump provides negative pressure to the chromatography module, drawing the vapor analytes into the preconcentrator through the capacitive detec- tors and separation columns. During analytical separation, the Knudsen pump provides positive pressure, reversing the ﬂow through the chromatography module. The preconcentrator is then used to thermally inject the analytes, which are separated by the columns and detected by the capacitive detectors. The columns can be selectively heated to accelerate the separation. Embedded thin-ﬁlm metal heaters are used to heat the Knudsen pump, preconcentrator and separation columns, whereas embedded thin-ﬁlm metal thermistors are used to monitor the temperatures of these components. In the preconcentrator, the sorbent particles, which absorb VOCs, are held in a channel with an elongated ‘U’ shape, of 1350 μm width and 300 μm height (Figure 1). MATERIALS AND METHODS Design Therefore, any potential leak site near the chip perimeter could directly shunt the column (that is, causes gas ﬂow to pass through the leak site but not the rest of the column), greatly compromising the performance of the separation column. In the present manifestation of the overall system, all the ﬂuidic components are arranged in a planar manner in both the chromatography module and the Knudsen pump module. Whereas the ﬂuidic connection between the two modules uses conventional tubing to facilitate modular reconﬁgurability, within each module, the components are connected by co-fabricated gas ﬂow connectors. The short length of these connectors (7 mm) permits compact arrangement of the components, minimizing the overall footprint. MATERIALS AND METHODS Design The sorbent particles are conﬁned to the lower half of the ‘U’ shape by Microsystems & Nanoengineering doi:10.1038/micronano.2015.49 Fully electronic microfabricated gas chromatograph Y Qin and YB Gianchandani 3 be achieved in the gas ﬂow. Each stage uses a stack of mesoporous mixed-cellulose-ester (MCE) membranes (Millipore, MA, USA) as the medium for thermal transpiration33. Each membrane has an average pore diameter of 25 nm, a porosity of 70%, and a thickness of 100 μm. The stack has a total thickness of 500 μm and a footprint of 16 × 16 mm2. The membrane stack is sandwiched by three glass dies on each side. The two inner glass dies contain ﬂow channels, and the two middle glass dies contain embedded thin-ﬁlm metal heaters and thermistors, whereas the two outer glass dies provide common anchor planes (Figure 1). Commercially available ﬁn-shaped heat sinks are attached to the external surfaces of the outer glass dies. Because the ﬂow is directed from the unheated side to the heated side, having the heaters embedded on both sides of the membrane stack allows the ﬂow direction to be reversed when switching from sampling to separation. be achieved in the gas ﬂow. Each stage uses a stack of mesoporous mixed-cellulose-ester (MCE) membranes (Millipore, MA, USA) as the medium for thermal transpiration33. Each membrane has an average pore diameter of 25 nm, a porosity of 70%, and a thickness of 100 μm. The stack has a total thickness of 500 μm and a footprint of 16 × 16 mm2. The membrane stack is sandwiched by three glass dies on each side. The two inner glass dies contain ﬂow channels, and the two middle glass dies contain embedded thin-ﬁlm metal heaters and thermistors, whereas the two outer glass dies provide common anchor planes (Figure 1). Commercially available ﬁn-shaped heat sinks are attached to the external surfaces of the outer glass dies. Because the ﬂow is directed from the unheated side to the heated side, having the heaters embedded on both sides of the membrane stack allows the ﬂow direction to be reversed when switching from sampling to separation. pillar-like structures and glass beads that are packed against the pillars. A thin-ﬁlm metal heater is located in this region to provide localized heating for vapor desorption. Fabrication Because this dimension is substan- tially smaller than the interelectrode gap, the electric ﬁeld lines extend beyond the thickness of the polymer. Its capacitance change, ΔC1, is always positive, dominated by the swelling response of OV-1. The second detector, CapDet2, uses an OV-1 layer with a thickness of 1.6 μm. Its capacitance change, ΔC2, can be either positive or negative, depending on the dielectric constant of the vapor species relative to that of OV-1. Whereas the swelling response of the OV-1 layer does not contribute to ΔC2, the change in permittivity caused by the penetration of the gas molecules has a strong effect because the electric ﬁeld lines are mostly contained within the polymer30. In each detector, the electrodes occupy a footprint of 1 mm2, which is enclosed in a gas ﬂow channel with a width of 800 μm and a height of 300 μm. μ The Knudsen pump is assembled by sandwiching the MCE membrane stack within the microfabricated glass dies and sealed with an epoxy (Stycast2850FT, Henkel, Germany). The MCE membrane and the epoxy can potentially introduce mild out- gassing issues that interfere with the chromatographic separation; these issues can be addressed by proper system operation, as discussed in the experimental section. The preconcentrator is assembled by bonding the glass dies with another epoxy (Epotek-377, Epoxy Technology Inc., MA, USA), followed by baking at 300 °C for 41 h to minimize outgassing during operation. (Additional purges can be performed at lower temperatures even after the assembly of the complete system.) Two types of sorbent particles, that is, Carbopack-X and Carbopack-B, as well as glass beads, are sequentially packed by a gentle vacuum into the ‘U’ channel from a loading port that is subsequently sealed with Stycast2850FT epoxy. Fabrication The separation column is assembled by bonding the glass dies with a spin-coated layer of adhesion material (SU-8, MicroChem, MA, USA), followed by the application The Knudsen pump module is implemented in two stages connected in series, which allows for a higher output pressure to Microsystems & Nanoengineering doi:10.1038/micronano.2015.49 Fully electronic microfabricated gas chromatograph Y Qin and YB Gianchandani 4 Table 1 List of chemicals and test conditions Run ID: 1 2 3 4 5 1a 1b, 1c 5a 5b Sampling time (min) 30 120 Chemicals Concentration (p.p.m.) a1 1 0.5 20 10 10 a2–a6 1 0.5 0.1 0.05 0.01 b1–b5 1 0.5 0.1 0.05 0.01 c1–c5 0.2 0.1 0.02 0.01 0.002 d1–d3 1 0.5 0.1 0.05 0.01 a1–a6: n-alkanes C5–C10 (nonpolar). b1–b5: benzene, toluene, m- and o-xylene, mesitylene (nonpolar). c1–c5: hexanal, chlorobenzene, chlorohex- ane, 4-chlorotoluene, 1,3-dichlorobenzene (mildly polar). d1–d3: cyclohep- tane, α-pinene, 3-carene (nonpolar). Relative humidity: 10–20%. Column temperature ramped during separation for Runs 1b, 1c. Figure 2 Photograph of the iGC3.c2 system. One of the Knudsen pump heat sinks has been removed to reveal the inner structure. Table 1 List of chemicals and test conditions a1–a6: n-alkanes C5–C10 (nonpolar). b1–b5: benzene, toluene, m- and o-xylene, mesitylene (nonpolar). c1–c5: hexanal, chlorobenzene, chlorohex- ane, 4-chlorotoluene, 1,3-dichlorobenzene (mildly polar). d1–d3: cyclohep- tane, α-pinene, 3-carene (nonpolar). Relative humidity: 10–20%. Column temperature ramped during separation for Runs 1b, 1c. a1–a6: n-alkanes C5–C10 (nonpolar). b1–b5: benzene, toluene, m- and o-xylene, mesitylene (nonpolar). c1–c5: hexanal, chlorobenzene, chlorohex- ane, 4-chlorotoluene, 1,3-dichlorobenzene (mildly polar). d1–d3: cyclohep- tane, α-pinene, 3-carene (nonpolar). Relative humidity: 10–20%. *Column temperature ramped during separation for Runs 1b, 1c. 4-chlorotoluene and 1,3-dichlorobenzene; and the d-group contained three other non-polar chemicals (cycloalkane and terpenes): cycloheptane, α-pinene, 3-carene. Most of these chemicals are common indoor air pollutants2. Figure 2 Photograph of the iGC3.c2 system. One of the Knudsen pump heat sinks has been removed to reveal the inner structure. To prepare the test vapor, the 19 chemicals in liquid form were ﬁrst formulated into a mixture. The volume of each chemical was adjusted (based on density and molar mass) such that the concentrations of the non-polar chemicals were equal in the resulting vapor and the concentration of each mildly polar chemical was 1/5 of that concentration. Fabrication For mixtures that presented relatively high levels of concentration (in which each non-polar chemical had a concentration of 0.5–1 p.p.m.), liquid mixtures with volumes of 85–170 nl were injected into a 2-l dilution bottle preﬁlled with room air. For mixtures that presented lower concentrations, the injected liquid volume became difﬁcult to control accurately. Therefore, these liquid mixtures were diluted with pentane before being injected into the 2-l bottle. In each case, the mixture injected into the bottle evaporated completely and formed the test vapor. Table 1 summarizes the vapor concentrations that were tested. In this paper, the units for vapor concentration, i.e., ‘p.p.m.’ and ‘p.p.b.’, are deﬁned as the number of molecules of each chemical divided by the number of molecules of air in the prepared vapor, multiplied by 106 and 109, respectively. of an OV-1 stationary phase coating to the interior of the column using the established static coating method35. For the capacitive detectors, the metallized substrate is coated with a 10-nm-thick Al2O3 layer using atomic layer deposition. This is an optional, precautionary step to prevent potential shorting caused by dust; that is, the step can be excluded if the assembly is performed in a clean environment. Next, an OV-1 solution is spin- coated onto the metallized substrate. The OV-1 solution is prepared by dissolving OV-1 in nonane. A small quantity of dicumyl peroxide (o10% of OV-1 in weight) is also added to the solution to induce cross-linking35. Other common non-polar solvents such as pentane and hexane have also been evaluated for preparing the solution. However, these solvents result in randomly located cavities in the OV-1 layer, possibly because these solvents are highly volatile and form bubbles during spin-coating. In contrast, the less volatile nonane provides a smooth and uniform OV-1 ﬁlm after spin coating. Subsequently, the OV-1-coated substrate is bonded to the sandblasted substrate with epoxy. The ﬂuidic components of the chromatography and pump modules are mounted on separate printed circuit boards (PCBs). The microfabricated gas ﬂow connectors are connected between the ﬂuidic components. In the current manifestation of the system, commercial pins and headers are directly soldered to the components to provide electrical lead transfer between the components. To prevent on-board electromagnetic crosstalk that can affect the detector readout, CapDet1 is fenced in by an aluminum foil perimeter that is grounded. A photograph of the system is shown in Figure 2. Fabrication The overall footprint of the system, including both modules, is 8 × 10 cm2. Test method During the vapor sampling stage of each experiment, when the sample VOCs were accumulated in the preconcentrator, the bottle was connected to the port upstream of the capacitive detectors (Figure 3a). The vapor was drawn from the bottle through the detectors and the separation columns into the preconcentrator by the Knudsen pump module, which was operated at 0.2 sccm for 10–120 min. The temperature in the pump was controlled by the on-chip heaters and temperature sensors operated by a custo- mized closed-loop control algorithm in LabVIEW, which used pulse-width modulation (PWM) to control the heater power. As previously reported16, the output ﬂow rate of the Knudsen pump is roughly proportional to the temperature difference across the MCE membranes (ΔT) and it is also dependent upon the average temperature of the MCE membranes (Tavg). For the experiments reported in this study, the ﬂow rate could be determined by setting either ΔT or the ratio ΔT/Tavg; both provided stable control of the ﬂow rate. If Tavg varied widely, then a more complex function consisting of ΔT and Tavg could be used in the PWM algorithm. The pump consumed 3.2 W during vapor sampling. The power estimate accounts for the experimentally measured Microsystems & Nanoengineering Test vapor preparation For separation columns of relatively short length, the selected set of VOC indoor pollutants provided modest elution times, reducing the need for temperature programming. In experiments that were performed without deliberately heating the separation columns, Column 1 experienced a temporary temperature increase of 2 °C due to crosstalk from the thermal pulse delivered to the preconcentrator, whereas Column 2 experienced an increase of about 0.5 °C (Figure 4). average voltage and current generated by the PWM controller during this period. A commercial ﬂow meter (Model # MW-20SCCM-D, Alicat Scientiﬁc, Inc., Tuscon, AZ, USA) was optionally connected downstream of the pump to provide readings of the ﬂow rate. The other end of the ﬂow meter was open to room air. The control of the system did not rely on the ﬂow meter. The chromatography module was maintained at room temperature. After vapor sampling, the Knudsen pump was turned off and allowed to cool down for 15–25 min before initiating analytical separation. average voltage and current generated by the PWM controller during this period. A commercial ﬂow meter (Model # MW-20SCCM-D, Alicat Scientiﬁc, Inc., Tuscon, AZ, USA) was optionally connected downstream of the pump to provide readings of the ﬂow rate. The other end of the ﬂow meter was open to room air. The control of the system did not rely on the ﬂow meter. The chromatography module was maintained at room temperature. After vapor sampling, the Knudsen pump was turned off and allowed to cool down for 15–25 min before initiating analytical separation. During either sampling or separation, the 0.2-sccm air ﬂow through the system resulted from the overall pressure load of the ﬂuidic path presented to the Knudsen pump module (Figure 5). The ﬂow properties of the two modules were measured in separate experiments. The ﬂow resistance of the chromatography module was measured by using an external vacuum pump to create an air ﬂow through the module, whereas the ﬂow rate and pressure difference along the module were measured by a ﬂow meter (MW-20SCCM-D, Alicat Scientiﬁc, Inc.,) and pressure sensor (MPX5010DP, Freescale Semiconductor Inc., Austin, TX, USA), respectively. The ﬂow resistance of the Knudsen pump module was measured in the same manner. The output pressure head and ﬂow rate characteristics of the Knudsen pump module were measured at 3.2 W. The pump provided a maximum ﬂow rate of 0.5 sccm and a blocking pressure of 1 kPa. Test vapor preparation To evaluate its performance, the iGC3.c2 system was used to sample, separate and detect mixtures of 19 chemicals. The chemical species were labeled in groups based on their molecular structures or properties. The a-group contained six non-polar n-alkanes ranging from pentane (C5) to decane (C10); the b-group contained ﬁve non-polar aromatic hydrocarbons: benzene, toluene, m-xylene, o-xylene and mesitylene; the c-group contained ﬁve mildly polar chemicals (aldehyde and halo- genated hydrocarbons): hexanal, chlorobenzene, chlorohexane, Microsystems & Nanoengineering doi:10.1038/micronano.2015.49 Fully electronic microfabricated gas chromatograph Y Qin and YB Gianchandani 5 Figure 4 Typical temperature control of the system during analytical separations without active temperature programming of the separation columns. (a) The temperatures at the Knudsen pump. Continuous temperature difference between the hot side and cold side provides continuous ﬂow. (b) The preconcentrator temperature during a typical thermal pulse to perform vapor injection. (c) The temperature rise of the separation columns resulting from thermal crosstalk with the preconcentrator. Figure 3 Test setup for emulating two scenarios of system operation. Scenario A: a conventional scenario in which vapor samples are manually collected and fed to the system; the carrier gas is room air without the analytes. Scenario B: an emerging scenario in which the system automatically monitors the ambient pollutants; the carrier gas contains the analytes. (a) Test setup for vapor sampling. (b) Test setup for emulating analytical separation in Scenario A. (c) Test setup for emulating analytical separation in Scenario B. VOC, Volatile organic compound. Figure 4 Typical temperature control of the system during analytical separations without active temperature programming of the separation columns. (a) The temperatures at the Knudsen pump. Continuous temperature difference between the hot side and cold side provides continuous ﬂow. (b) The preconcentrator temperature during a typical thermal pulse to perform vapor injection. (c) The temperature rise of the separation columns resulting from thermal crosstalk with the preconcentrator. Figure 3 Test setup for emulating two scenarios of system operation. Scenario A: a conventional scenario in which vapor samples are manually collected and fed to the system; the carrier gas is room air without the analytes. Scenario B: an emerging scenario in which the system automatically monitors the ambient pollutants; the carrier gas contains the analytes. (a) Test setup for vapor sampling. (b) Test setup for emulating analytical separation in Scenario A. (c) Test setup for emulating analytical separation in Scenario B. VOC, Volatile organic compound. heated. Microsystems & Nanoengineering Test vapor preparation The intercept of the pump characteristic line and the chromatography module resistance line appeared at 0.2 sccm, as was experimentally conﬁrmed by the ﬂow meter. During the analytical separation stage of each experiment, the dilution bottle was removed. As a result, room air served as the carrier gas in the separation stage (Figure 3b). This arrangement emulated the conventional application scenario in which vapor samples are manually collected and injected to the system. An alternative scenario was also emulated: autonomous operation in which polluted air containing the analytes serves as the carrier gas during separation (Figure 3c); this scenario is described in a separate subsection. The moment at which the software sequence for control and readout was initiated was deﬁned as time t = 0. Starting at time t≈20 s, the pump provided a reversed 0.2-sccm ﬂow (using 3.2 W), pushing room air without analytes into the chromatography module. After the ﬂow rate reached the appropriate level, starting at time t≈30 s, the preconcentrator was heated to 250 °C in 5 s and maintained at this temperature for 10 s (Figure 4) to thermally inject the sampled vapors into the closest separation column (Column 1). Like the Knudsen pumps, the preconcentrator was heated using the embedded thin-ﬁlm metal heaters, and its temperature was monitored using embedded thermistors. Even though temperature programming of the separation columns was evaluated as described in the next section, for most experiments the columns were not actively The two capacitive detectors were simultaneously measured by two commercial 24-bit capacitive-to-digital converter (CDC) chips (AD7746, Analog Devices, Norwood, MA, USA). Available CDC options offering lower resolution (for example, 16 bit) were not in order to prevent electronic noise from contributing to the observed VOC detection limits. Each CDC provided a square- wave excitation signal to one set of electrodes in the capacitive detector and measured the voltage from the opposite set of electrodes. Because the excitation signal for one detector could affect the voltage readout from the other detector through doi:10.1038/micronano.2015.49 Microsystems & Nanoengineering Fully electronic microfabricated gas chromatograph Y Qin and YB Gianchandani were identiﬁed in the chromatogram; this was performed manually in the present work, but in principle can be performed Figure 6 iGC3.c2 system results for 19 chemicals with varying concentrations. Separations were performed at room temperature. (The chromatograms are shifted vertically for visual clarity). (a) Chromatograms provided by CapDet1. (b) Chromatograms provided by CapDet2. Quantitative analysis To demonstrate quantitative analysis, the system was used to evaluate the 19-chemical mixture (Table 1). The chromatograms obtained from CapDet1 and CapDet2 are shown in Figures 6a and b, respectively. A sampling time of 30 min was sufﬁcient to demonstrate the effective separation and detection of chemicals with concentrations ranging from 0.05 to 1 p.p.m. for non-polar species and from 0.01 to 0.2 p.p.m. for mildly polar species (for example, in Runs 1–4, Figures 6a and b). For lower concentrations (that is, 10 p.p.b. for non-polar species and 2 p.p.b. for mildly polar species), the 30-min sample did not provide distinguishable peaks (for example, in Run 5a). However, a sampling time of 2 h did provide distinguishable peaks (for example, in Run 5b, insets of Figures 6a and b). Figure 6 iGC3.c2 system results for 19 chemicals with varying concentrations. Separations were performed at room temperature. (The chromatograms are shifted vertically for visual clarity). (a) Chromatograms provided by CapDet1. (b) Chromatograms provided by CapDet2. (c) Measured peak heights (ΔC1 and ΔC2) showing proportionality to prepared vapor concentrations. p.p.m., parts per million. Figure 6 iGC3.c2 system results for 19 chemicals with varying concentrations. Separations were performed at room temperature. (The chromatograms are shifted vertically for visual clarity). (a) Chromatograms provided by CapDet1. (b) Chromatograms provided by CapDet2. (c) Measured peak heights (ΔC1 and ΔC2) showing proportionality to prepared vapor concentrations. p.p.m., parts per million. Figure 6 iGC3.c2 system results for 19 chemicals with varying concentrations. Separations were performed at room temperature. (The chromatograms are shifted vertically for visual clarity). (a) Chromatograms provided by CapDet1. (b) Chromatograms provided by CapDet2. (c) Measured peak heights (ΔC1 and ΔC2) showing proportionality to prepared vapor concentrations. p.p.m., parts per million. were identiﬁed in the chromatogram; this was performed manually in the present work, but in principle can be performed automatically by software36. Next, a polynomial was used to ﬁt the baseline segments and was then subtracted from the raw chromatogram. Test vapor preparation (c) Measured peak heights (ΔC1 and ΔC2) showing proportionality to prepared vapor concentrations. p.p.m., parts per million. 6 Figure 5 Pump characteristics and ﬂow resistance of the Knudsen pump module and chromatography module. Figure 5 Pump characteristics and ﬂow resistance of the Knudsen pump module and chromatography module. electromagnetic coupling, an electromagnetic shield, made of custom-cut aluminum sheets and connected to the CDC ground, was placed on the PCB, as noted in ‘Materials and Methods’. The CDCs used a detection rate of 6–7 Hz. When the CDC was not connected to any capacitor, its rated value and measured value of root-mean-square (RMS) noise were both 4 aF. When the CDC was connected to the capacitive detector in the current setup, the measured RMS noise was 40 aF. The noise increase could potentially be reduced with improved PCB design. For all the experiments performed in this work, the capacitive detectors were operated at an ambient temperature of 22 °C. A number of experiments were performed on test samples, as explained in the next section. Before each analytical cycle, the preconcentrator was ﬂushed for 35 s by providing the type of thermal pulse used for the vapor injection, while the pump provided 0.2-sccm ﬂow in the direction used for vapor sampling. This procedure prevents the accumulation of contaminants in the preconcentrator from run to run, essentially regenerating the system; it could be particularly helpful when using polluted air as the carrier gas. Resolving co-eluting peaks In some examples of μGC systems, arrays of detectors, each having a unique response pattern to various chemicals, have been used to enhance the discrimination of eluted peaks13,26,37. In contrast to the reported chemiresistor arrays37, which provide only positive peaks to almost all chemicals, capacitive detectors can provide both positive and negative peaks27,30, thereby providing higher discrimination capabilities for certain chemical species. In this study, the a- and d-group chemicals in particular could be easily differentiated from the c-group chemicals using CapDet1 and CapDet2. In addition, these two detectors could be used to resolve co-eluting peaks. To resolve these peaks the detector response ratio, i.e., r = ΔC2/ΔC1, was ﬁrst measured for each chemical (Figure 7a). These measurements were performed in two sets, with each set including chemicals that did not produce co- eluting peaks. (Additional sets may be necessary in other cases.) In a chemical mixture that contains two co-eluting species, A and B, the following equation matrix can be formulated: ΔC1A þ ΔC1B ¼ ΔC1 rAUΔC1A þ rBUΔC1B ¼ ΔC2 ð1Þ ð1Þ where ΔC1A and ΔC1B are the contributions of chemicals A and B to the peak obtained from CapDet1, and rA and rB are the respective response ratios. By solving this equation matrix for every data point in the chromatogram segments containing the two co-eluting peaks, each co-eluting peak can be separately reconstructed. As shown in Figures 7b and c, this method successfully resolves partially co-eluting peaks c3–b3 and d2–c4, as well as fully co-eluting peaks c1–d1. Figure 7 Co-eluting peaks resolved by the two capacitive detector responses. (a) Capacitive detector response ratios (ΔC2/ΔC1) for various chemicals. (b) Co-eluting peaks resolved in CapDet1 chromatogram. (c) Co-eluting peaks resolved in CapDet2 chromatogram. The chromatograms were de-noised by a progressive moving average method. Moving averages are typically used to process transient data streams: a ﬁxed-length window moves along the stream, and raw data within the window are averaged to generate a new data point. The window must be long enough to minimize the noise in the data yet small enough to preserve the signal intensity. In the case of GC chromatograms, later retention peaks are typically wider than earlier ones because of diffusion in the separation columns8. Therefore, the GC chromatograms are amenable to the calculation of a moving average with an increasing window length. Temperature programming For microfabricated separation columns, each column chip can often be individually programmed13. Two options were evaluated: with both separation columns actively heated independently (Figure 8a) and with only the downstream column heated (Figure 8b). In the former case, in Run 1c (Figure 8a), the temperature of Column 1, which was upstream of Column 2 during separation, was ramped linearly from room temperature to 40 °C from t = 100 s to t = 200 s and then allowed to cool naturally. Further heating of Column 1 was unnecessary because all the tested chemicals had eluted Column 1. The temperature of Column 2 was linearly ramped to 52 °C from t = 100 s to t = 320 s and then allowed to cool naturally. Compared with room- temperature separation, this temperature programming run accelerated the separation by 50% and increased the peak heights of certain chemicals by 50–100%. The columns consumed a total average power of 0.35 W from t = 100 s to t = 320 s. The power consumption can be slightly reduced by heating only the downstream column. In the latter case, in which only the downstream column was heated, in Run 1d (Figure 8b), the Resolving co-eluting peaks For the chromatograms of Runs 3–5 (in the insets of Figures 6a and b), the window length for the chromatograms was linearly ramped from 5 points (before t≈30 s) to 101 points (at the end of the chromatograms). This method clearly reduces the noise without signiﬁcantly distorting the retention peaks. Data processing Baseline variations in raw chromatograms were digitally compen- sated by a customized MATLAB program. First, baseline segments Microsystems & Nanoengineering doi:10.1038/micronano.2015.49 doi:10.1038/micronano.2015.49 Fully electronic microfabricated gas chromatograph Y Qin and YB Gianchandani 7 Figure 7 Co-eluting peaks resolved by the two capacitive detector responses. (a) Capacitive detector response ratios (ΔC2/ΔC1) for various chemicals. (b) Co-eluting peaks resolved in CapDet1 chromatogram. (c) Co-eluting peaks resolved in CapDet2 chromatogram. sampled (Figure 6c). This proportionality is expected for capacitive detectors. Compared with detectors that provide linear response in log or log–log scales12,24, the directly proportional response of the capacitive detectors can be more readily interpreted. Furthermore, as indicated by the slopes in Figure 6c, both CapDet1 and CapDet2 showed higher sensitivity to the c-group (mildly polar) chemicals than to the a-, b- and d-group (non-polar) chemicals. Compared with CapDet1, which showed positive peaks for all the tested chemicals, CapDet2 showed positive peaks for the c-group, negative peaks for the a- and d-group, and nearly zero response to the b-group. In addition, both detectors showed a common trend for the sensitivity to alkanes: the absolute value of sensitivity increased from pentane (a1) to nonane (a5) but decreased from nonane to decane (a6). These observations are discussed in the next section. Linearity and sensitivity trends The peak heights extracted from the chromatograms varied proportionally with the corresponding vapor concentrations being Microsystems & Nanoengineering doi:10.1038/micronano.2015.49 Fully electronic microfabricated gas chromatograph Y Qin and YB Gianchandani Figure 8 Accelerated separation provided by temperature program- ming of the separation columns. (a) Run 1b: similar to Run 1a in Figure 6 but with both columns heated for fast separation. (b) Run 1c: heating only Column2 to accelerate separation with lower power consumption. 8 Figure 9 Chromatograms of repeated runs using both ambient and polluted air as the carrier gas. Run 6a used the (relatively clean) ambient air as the carrier gas. Run 6b-c used polluted air with 10–25% ambient relative humidity as the carrier gas for bench- marking. Run 6d-g used polluted air with 50% relative humidity as the carrier gas. (a) CapDet1 response. (b) CapDet2 response. Figure 9 Chromatograms of repeated runs using both ambient and polluted air as the carrier gas. Run 6a used the (relatively clean) ambient air as the carrier gas. Run 6b-c used polluted air with 10–25% ambient relative humidity as the carrier gas for bench- marking. Run 6d-g used polluted air with 50% relative humidity as the carrier gas. (a) CapDet1 response. (b) CapDet2 response. Figure 8 Accelerated separation provided by temperature program- ming of the separation columns. (a) Run 1b: similar to Run 1a in Figure 6 but with both columns heated for fast separation. (b) Run 1c: heating only Column2 to accelerate separation with lower power consumption. temperature of Column 2 was linearly ramped to 50 °C between 200 and 400 s. Column 1 was not actively heated but may have been passively heated by a few degrees because of thermal crosstalk. Compared with room-temperature separation, this temperature programming run accelerated the separation by 30%, while the columns consumed only 0.32 W. The energy consumption could be further reduced by stopping the heating of Column 2 at the end of the ramp (Figure 8b). overestimation of the VOC concentrations. Therefore, before each vapor sampling cycle, the preconcentrator was ﬂushed by ﬂow in the direction of sampling for a short duration. In experimental evaluation, the system was tested by performing consecutive runs. As demonstrated by the experimental results in Figure 9, the chromatograms for the 19 chemicals (1 p.p.m. in the a-, b- and d- groups, 0.2 p.p.m. Linearity and sensitivity trends in the c- group) in the carrier gas (Run 6b–g) overlap with the chromatogram obtained using the relatively clean ambient air as the carrier gas (Run 6a). These results indicate that the system can perform repeatable analyses in a polluted environment. Microsystems & Nanoengineering Ambient VOC in carrier gas The aforementioned experiments used the ambient room air as the carrier gas, which was relatively clean and not contaminated by the analyte samples. These results demonstrated the operation of the system in a conventional test scenario, in which vapor samples are manually collected and fed to the system for analyses. However, if the system is operated in a different scenario in which the ambient pollutants are being monitored automatically, then the carrier gas used for analytical separation would be polluted by the same vapor being sampled. Effect of humidity The relative humidity in the laboratory where the tests were conducted varied between 10 and 20%, as measured by a commercial humidity sensor (#42280, Extech Instruments Cor- poration, Nashua, NH, USA). To test the system response to moisture, a volume of water was injected into the 2-l dilution bottle together with the 19 VOCs such that the relative humidity in the sampled vapor was 50%, which is equivalent to a concentra- tion of 12 000 p.p.m.. The increased moisture caused baseline variations. However, after baseline compensation, the system response to the moisturized vapor (Run 6d–g) overlapped with that for the non-moisturized vapor (Run 6a–c, Figure 9). To mimic this latter scenario, another set of experiments was performed, in which the dilution bottle ﬁlled with the tested vapor mixture was connected upstream of the Knudsen pump during analytical separation to provide the carrier gas (Figure 3c). In principle, the preconcentrator, after thermal desorption, traps the VOCs in the carrier gas, preventing the VOCs from entering the separation columns. Therefore, in a single run, the results obtained using polluted air as the carrier gas should be similar to those obtained using ambient air. In consecutive runs, however, the VOCs trapped by the preconcentrator during previous analytical separation could be injected, together with the new sample, into the separation column in the following run, causing hysteresis and DISCUSSION The partition coefﬁcient multiplied by the phase ratio (β), that is, volume ratio of the air to the polymer in the detector, can be used to determine the fraction of chemicals (η) adsorbed by the polymer and useful for sensing8: The sensitivity of the capacitive detectors also depends on the partition coefﬁcients (KD) of chemicals between the polymer and air. The partition coefﬁcient is deﬁned as the ratio—at equilibrium In addition, in this study, considering 4-chlorotoluene as an example, the peak width (measured at half peak height) was 17 s. The gas ﬂow rate was 0.2 sccm during separation and hence this species occupied a volume of 0.06 ml. Therefore, the maximum concentration of the 4-chlorotoluene peak was 0.8 p.p.m. because 24 ml of the original 2 p.p.b. vapor was concentrated in the 0.06 ml peak. The KD of 4-chlorotoluene, as calculated based on a previously reported formula40, is 5590. Thus, only 15% of the 4-chlorotoluene molecules remained in the gas phase during the time that the peak passed through CapDet1. If the detector had been tested in a non-transient steady-state gas environment with a large head space, a concentration of 120 p.p.b. of 4-chlorotoluene would be necessary to produce the same reading. This type of correction is necessary when comparing the iGC3.c2 results with those in which chemical sensors are characterized at steady state. p q —of the chemical concentration in the polymer to that in air. This measure is used to describe the extent of vapor adsorption in the polymer. The partition coefﬁcient multiplied by the phase ratio (β), that is, volume ratio of the air to the polymer in the detector, can be used to determine the fraction of chemicals (η) adsorbed by the polymer and useful for sensing8: η ¼ KD=β KD=β þ 1 ð2Þ η ¼ KD=β KD=β þ 1 ð2Þ On one hand, large KD values produce large η values and hence increase the magnitude of ΔC. On the other hand, because the separation columns use the same polymer as the detectors in the iGC3.c2 system, larger KD values cause longer retention and broader peaks, thereby reducing the magnitude of ΔC. The β values for CapDet1 and CapDet2 are 1000 and 160, respectively. The partition coefﬁcients of the a-group chemicals between PDMS and air at room temperature40, as well as the calculated values of η, are listed in Table 2. DISCUSSION that exists in a μGC detector, the actual fractions of chemicals adsorbed by the polymer are possibly smaller than the calculated values shown in Table 2. A more detailed understanding of the dynamic distribution of the chemical molecules can be obtained with computational modeling27. that exists in a μGC detector, the actual fractions of chemicals adsorbed by the polymer are possibly smaller than the calculated values shown in Table 2. A more detailed understanding of the dynamic distribution of the chemical molecules can be obtained with computational modeling27. g g g In CapDet1, the polymer thickness of 0.25 μm is substantially smaller than the electrode width and gap of 1 μm. Therefore, polymer swelling does affect ΔC1 because the entire polymer thickness is within the distribution of electric ﬁeld lines for sensing. Polymer swelling always produces a positive ΔC1 because an increasing number of electric ﬁeld lines traverse the polymer. The VOC molecules that cause the swelling of the polymer layer also contribute dipole moments that add to the capacitive response. As demonstrated by the results (Figure 6a), ΔC1 is positive for all the tested chemicals, even for the a-group chemicals that provide negative εch–εp. In addition, ΔC1 is strongly dependent upon the value of εch–εp. The c-group chemicals provide values of εch–εp, that are several times higher than those provided by non-polar chemicals, and hence the values of ΔC1 are also several times higher. Further calculations can be performed to elucidate the chemical quantities being analyzed. For example, in Run 5b, the 120-min sample taken at a ﬂow rate of 0.2 sccm provided a sampling volume of 24 ml. Hence, the mass of chemicals being sampled and subsequently injected by the preconcentrator was 1 ng for the non-polar chemicals and 0.2 ng for the mildly polar chemicals. (The effective mass actually sampled into the preconcentrator was slightly smaller because of the threshold volume, as discussed in the following text; nevertheless, this difference was negligible for the 120-min sample.) These values represent the lowest mass levels that were experimentally detected by the capacitive detectors. g The sensitivity of the capacitive detectors also depends on the partition coefﬁcients (KD) of chemicals between the polymer and air. The partition coefﬁcient is deﬁned as the ratio—at equilibrium —of the chemical concentration in the polymer to that in air. This measure is used to describe the extent of vapor adsorption in the polymer. Microsystems & Nanoengineering DISCUSSION The results of the two capacitive detectors warrant discussion. As noted previously, the ΔC depends on changes in both thickness and permittivity that are caused by the eluted peaks of the VOCs in the OV-1 polymer ﬁlm covering the electrodes. The permittivity doi:10.1038/micronano.2015.49 Fully electronic microfabricated gas chromatograph Y Qin and YB Gianchandani 9 Table 2 Calculation of chemicals partitioned between polymer and air in the detectors VOC chemicals Partition coefﬁcient (KD)40 Fraction of chemical in polymer in CapDet1 (η1) (%) Fraction of chemical in polymer in CapDet2 (η2) (%) Pentane 82 8 34 Hexane 215 18 58 Heptane 565 36 78 Octane 1486 60 90 Nonane 3908 80 96 Decane 10 280 91 99 Benzene 296 23 65 Toluene 815 45 84 m-Xylene 2190 69 93 o-Xylene 2710 73 95 Mesitylene 6150 86 98 Table 2 Calculation of chemicals partitioned between polymer and air in the detectors response can be expressed as the difference in the dielectric constants of the polymer and chemical, i.e., εch–εp, in the affected region of OV-1. If this term is positive, then the average dielectric constant of the polymer–chemical matrix is higher than that of the polymer, and the resulting ΔC will also be positive. Conversely, if this term is negative, then its contribution to the ΔC will be negative. g In CapDet2, where the polymer thickness of 1.6 μm is greater than the electrode width and gap of 1 μm, polymer swelling does not affect ΔC because most electric ﬁeld lines for sensing are distributed only in the lower portion of the polymer. Hence, ΔC2 depends almost entirely on εch–εp within the ﬁxed volume traversed by the ﬁeld lines. The dielectric constant of OV-1 (εp), which is composed of 100% polydimethylsiloxane or PDMS, at room temperature is 2.3–2.8 (ref. 38), whereas the dielectric constants of the a-, b-, c- and d- group chemicals (εch) are 1.8–2.0, 2.2–2.6, 5–10 and 2.0–2.2, respectively39. It should be noted that the dielectric constants of chemicals under consideration are the values for the liquid phase rather than for the gas form because the vapor molecules adsorbed in the polymer are condensed. Therefore, the c- group contributes to a positive ΔC2 value, the b group contributes to a nearly zero ΔC2 value and the a and d groups contribute to a negative ΔC2 value (Figure 6b). REFERENCES Kirk-Othmer Encyclopedia of Chemical Technology. John Wiley & Sons, Inc., Hoboken, NJ, USA, 2000. 8 Barry EF. Columns: Packed and capillary; column selection in gas chromatography. In:Grob RL, Barry EF, editors. Modern Practice of Gas Chromatography. John Wiley & Sons, Inc.: Hoboken, NJ, USA; 2004: 65–191. 9 Terry SC, Jerman JH, Angell JB. A gas chromatographic air analyzer fabricated on a silicon wafer. IEEE Transactions on Electron Device. 1979; 26: 1880–1886. 10 Voiculescu I, Zaghloul M, Narasimhan N. Microfabricated chemical pre- concentrators for gas-phase microanalytical detection systems. Trends in Analy- tical Chemistry 2008; 27: 327–343. 11 Lewis PR, Manginell RP, Adkins DR et al. Recent advancements in the gas-phase MicroChemLab. IEEE Sensors Journal 2006; 6: 784–795. 12 Zampolli S, Elmi I, Mancarella F et al. Real-time monitoring of sub-ppb con- centrations of aromatic volatiles with a MEMS-enabled miniaturized gas-chro- matograph. Sensors and Actuators B: Chemical 2009; 141: 322–328. p y pp y In conclusion, the results show the viability of a full μGC system in which all the components are microfabricated and electro- nically interfaced. The system has no moving parts and uses room air as the carrier gas, and all the ﬂuidic components can be co- fabricated by a common low-cost process. These features can potentially increase the overall applicability of μGC technology. The current manifestation of the system provided a detection limit of 2 ppb for certain chemicals, which is low enough for use in applications such as indoor pollutant monitoring. Compared with other reported μGCs that provide comparable performance12,13, the results of this study are signiﬁcant because they are provided by a fully microfabricated and integrated system. With further reﬁnement of the electronic hardware and software, the system is expected to be suitable for continuous, unmanned monitoring. 13 Kim SK, Chang H, Zellers ET. Microfabricated gas chromatograph for the selective determination of trichloroethylene vapor at sub-parts-per-billion concentrations in complex mixtures. Analytical Chemistry 2011; 83: 7198–7206. 14 Garg A, Akbar M, Vejerano E et al. Zebra GC: A mini gas chromatography system for trace-level determination of hazardous air pollutants. Sensors and Actuators B: Chemical 2015; 212: 145–154. 15 Kim H, Steinecker WH, Reidy S et al. A micropump-driven high-speed MEMS gas chromatography system. IEEE International Conference on Solid-State Sensors, Actuators and Microsystems Conference (Transducers); 10–14 June 2007; Lyon, France; 2007: 1505–1508. 16 Qin Y, Gianchandani YB. ACKNOWLEDGEMENTS requiring a threshold sampling duration to be exceeded for the preconcentrator to receive sampled vapor. This threshold sampling duration can be measured at any vapor concentration by plotting the variation in the output with sampling duration. For the iGC3.c2 system with a ﬂow rate of 0.2 sccm during sampling, the threshold sampling durations for m-xylene, 4-chlorotoluene and decane are 4, 7 and 10 min, respectively. The study was supported in part by the Global Challenges for a Third Century (GCTC) project at University of Michigan. The facilities used for this research included the Lurie Nanofabrication Facility (LNF) at University of Michigan. The authors thank Professor K. Wise, and R. Gordenker for access to the test facilities. DISCUSSION From pentane to nonane, the η value substantially increases (Table 2), causing the magnitude of ΔC to increase in both detectors (Figure 6). Compared with nonane, decane yields only marginally larger η values but much broader peaks, hence providing slightly smaller ΔC magnitudes in both detectors (Figure 6). Similar trends are also evident for the b-group chemicals in the CapDet1 responses (Figure 6). It is worth noting that the partition coefﬁcient describes the chemical distribution only when equilibrium is reached. Because of the continuous ﬂow In this study, the capacitive detectors used OV-1 as the polymer ﬁlm. Being non-polar, this material is most responsive to non-polar and mildly polar chemicals. For highly polar chemicals, a polar polymer ﬁlm is recommended. Polar chemicals inherently provide higher values of εch–εp than chemicals with lower polarities and are likely to provide even lower detection limits25,26. It is notable that the detection principle of the interdigitated capacitive detectors can be modeled27,30, and hence a library of predictable detector responses can be constructed using compu- tational methods. Such a library can reduce the burden of evaluating the technology for a previously untested application. In the iGC3.c2 system, the sampled vapor must pass through the detectors and columns before reaching the preconcentrator, doi:10.1038/micronano.2015.49 Microsystems & Nanoengineering Fully electronic microfabricated gas chromatograph Y Qin and YB Gianchandani 10 Microsystems & Nanoengineering COMPETING INTERESTS The use of ambient air as the carrier gas eliminates the need for a bottled carrier gas, which is a consumable resource and an additional hardware item. The use of air is attractive for applications that require extended autonomous operation. For a given separation column, the highest achievable separation efﬁciency provided by air as a carrier gas is similar to that provided by the more commonly used nitrogen or helium41. However, to achieve the highest efﬁciency, the required ﬂow rate of air is less than that of helium41. For high-speed separations in which the ﬂow rate often exceeds the value required to achieve the highest efﬁciency, helium provides higher efﬁciency than air41 because helium enables faster mass transfer of VOCs between the carrier gas and the stationary phase. In contrast, for analyses in which speed is not a primary concern, the ﬂow rate can be set to the value required to achieve the highest efﬁciency The authors declare no conﬂict of interest. The authors declare no conﬂict of interest. REFERENCES 1 Brown SK, Sim MR, Abramson MJ et al. Concentrations of volatile organic com- pounds in indoor air—a review. Indoor Air 1994; 4: 123–134. 1 Brown SK, Sim MR, Abramson MJ et al. Concentrations of volatile organic com- pounds in indoor air—a review. Indoor Air 1994; 4: 123–134. 2 Kostiainen R. Volatile organic compounds in the indoor air of normal and sick houses. Atmospheric Environment 1995; 29: 693–702. 2 Kostiainen R. Volatile organic compounds in the indoor air of normal and sick houses. Atmospheric Environment 1995; 29: 693–702. 3 Occupational Safety & Health Administration (OSHA). Permissible exposure limits —annotated tables. Available at https://www.osha.gov/dsg/annotated-pels/index. html. Accessed on 16 December 2015. 3 Occupational Safety & Health Administration (OSHA). Permissible exposure limits —annotated tables. Available at https://www.osha.gov/dsg/annotated-pels/index. html. Accessed on 16 December 2015. 4 Huff J. Benzene-induced cancers: Abridged history and occupational health impact. International Journal of Occupational and Environmental Health 2007; 13: 213–221. 5 Bonne U. Gas sensors. In: Gianchandani YB, Tabata O, Zappe Heds. Comprehen- sive Microsystems. Elsevier B.V., Amsterdam, The Netherlands; 2008: 375–432. 5 Bonne U. Gas sensors. In: Gianchandani YB, Tabata O, Zappe Heds. Comprehen- sive Microsystems. Elsevier B.V., Amsterdam, The Netherlands; 2008: 375–432. y 6 Westman-Brinkmalm A, Brinkmalm GA mass spectrometer's building blocksMass Spectrometry. John Wiley & Sons, Inc., Hoboken, NJ, USA; 2008: 15–87. 6 Westman-Brinkmalm A, Brinkmalm GA mass spectrometer's building blocksMass Spectrometry. John Wiley & Sons, Inc., Hoboken, NJ, USA; 2008: 15–87. q g y The use of air as the carrier gas can potentially cause PDMS degradation (manifested as weight loss) at elevated temperatures. Because this polymer is used both as the stationary phase in the separation columns and the sensing layer in the capacitive detectors, this degradation can be a signiﬁcant concern. However, the literature shows that thermal oxidative degradation starts to become signiﬁcant only above 290 °C42. In experiments, a typical iGC3.c2 system was operated at room temperature (22 °C) intermittently for 450 runs over a period of 4 months without showing a decrease in the VOC retention times or the capacitances of the detectors, both indicating no PDMS weight loss. In temperature programming experiments that were performed in this study, the maximum temperature in the separation columns was only approximately 50 °C. y y 7 Stuart B. Infrared spectroscopy. Kirk-Othmer Encyclopedia of Chemical Technology. John Wiley & Sons, Inc., Hoboken, NJ, USA, 2000. 7 Stuart B. Infrared spectroscopy. Fully electronic microfabricated gas chromatograph Y Qin and YB Gianchandani 11 24 Kulkarni GS, Reddy K, Zhong Z et al. Graphene nanoelectronic heterodyne sensor for rapid and sensitive vapour detection. Nature Communications 2014; 5: 4376. 35 Reidy S, Lambertus G, Reece J et al. High-performance, static-coated silicon microfabricated columns for gas chromatography. Analytical Chemistry 2006; 78: 2623–2630. 35 Reidy S, Lambertus G, Reece J et al. High-performance, static-coated silicon microfabricated columns for gas chromatography. Analytical Chemistry 2006; 78: 2623–2630. 25 Patel SV, Mlsna TE, Fruhberger B et al. Chemicapacitive microsensors for volatile organic compound detection. Sensors and Actuators B: Chemical 2003; 96: 541–553. 36 Tzeng T-H, Kuo C-Y, Wang S-Y et al. A portable micro gas chromatography system for volatile compounds detection with 15 ppb of sensitivity. IEEE International Solid-State Circuits Conference (ISSCC); 22–26 Feb 2015; San Francisco, CA, USA; 2015: 1–3. 26 Mlsna TE, Cemalovic S, Warburton M et al. Chemicapacitive microsensors for chemical warfare agent and toxic industrial chemical detection. Sensors and Actuators B: Chemical 2006; 116: 192–201. 37 Bryant-Genevier J, Scholten K, Kim SK et al. Multivariate curve resolution of co- eluting vapors from a gas chromatograph with microsensor array detector. Sen- sors and Actuators B: Chemical 2014; 202: 167–176. 27 Morimoto K, Qin Y, Gianchandani YB. Modeling and characterization of the transient performance of a gas detector based on fringe-ﬁeld capacitance. IEEE SENSORS; 2–5 November 2014. Valencia, Spain; 2014. 38 Mark JE. Polymer Data Handbook. Oxford University Press, Inc.: Oxford, UK. 1999. 39 Wohlfarth C. Static Dielectric Constants of Pure Liquids and Binary Liquid Mix- tures: Supplement to IV/6. Springer: Berlin, Germany. 2008. 28 Zaretsky MC, Melcher JR, Cooke CM. Moisture sensing in transformer oil using thin-ﬁlm microdielectrometry. IEEE Transactions on Electrical Insulation 1989; 24: 1167–1176. 40 Martos PA, Saraullo A, Pawliszyn J. Estimation of air/coating distribution coefﬁ- cients for solid phase microextraction using retention indexes from linear temperature-programmed capillary gas chromatography. Application to the sampling and analysis of total petroleum hydrocarbons in air. Analytical Chemistry 1997; 69: 402–408. 29 Hagleitner C, Hierlemann A, Lange D et al. Smart single-chip gas sensor micro- system. Nature 2001; 414: 293–296. y 30 Igreja R, Dias CJ. Dielectric response of interdigital chemocapacitors: The role of the sensitive layer thickness. Sensors and Actuators B: Chemical 2006; 115: 69–78. 41 Sacks RD. High-Speed Gas Chromatography. Modern Practice of Gas Chromato- graphy. John Wiley & Sons, Inc.: Hoboken, NJ, USA; 2004: 229–274. 31 Knudsen M. Eine revision der gleichgewichtsbedingung der gase. REFERENCES iGC1: An integrated ﬂuidic system for gas chromato- graphy including Knudsen pump, preconcentrator, column, and detector micro- fabricated by a three-mask process. Journal of Microelectromechanical Systems. 2014; 23: 980–990. 17 Qin Y, Gianchandani YB. iGC2: An architecture for micro gas chromatographs utilizing integrated bi-directional pumps and multi-stage preconcentrators. Journal of Micromechanics and Microengineering. 2014; 24: 065011. p g In the future, the system platform can be reconﬁgured for other applications. For example, detectors with different coatings of polymer species can be easily integrated to enhance vapor recognition. More preconcentrators or separation columns can be added to enhance the chromatography performance. Whereas in the current system a two-stage Knudsen pump is used to demonstrate a scalable architecture, a single-stage Knudsen pump with improved design and implementation is also expected to be capable of driving the system, thereby reducing the system complexity and footprint. Other micro gas pumps can also be integrated in place of the Knudsen pump module. Planned future work also includes the comprehensive characterization of all components—particularly the Knudsen pump, separation column, and capacitive detectors—operating at various ambient tempera- tures, which will establish a database for ﬁeld use. 18 Reddy K, Guo Y, Liu J et al. Rapid, sensitive, and multiplexed on-chip optical sensors for micro-gas chromatography. Lab on a Chip. 2012; 12: 901–905. 19 Zhu H, Nidetz R, Zhou M et al. Flow-through microﬂuidic photoionization detectors for rapid and highly sensitive vapor detection. Lab on a Chip 2015; 15: 3021–3029. 20 Scholten K, Fan X, Zellers ET. A microfabricated optoﬂuidic ring resonator for sensitive, high-speed detection of volatile organic compounds. Lab on a Chip 2014; 14: 3873–3880. 21 Scholten K, Collin WR, Fan X et al. Nanoparticle-coated micro-optoﬂuidic ring resonator as a detector for microscale gas chromatographic vapor analysis. Nanoscale 2015; 7: 9282–9289. 22 Kaanta BC, Chen H, Zhang X. A monolithically fabricated gas chromatography separation column with an integrated high sensitivity thermal conductivity detector. Journal of Micromechanics and Microengineering. 2010; 20: 055016. 23 Akbar M, Shakeel H, Agah M. GC-on-chip: Integrated column and photoionization detector. Lab on a Chip. 2015; 15: 1748–1758. Microsystems & Nanoengineering doi:10.1038/micronano.2015.49 Fully electronic microfabricated gas chromatograph Y Qin and YB Gianchandani Fully electronic microfabricated gas chromatograph Y Qin and YB Gianchandani Thermische Molekularstromung. Annalen der Physik Leipzig 1909; 336: 205–229. 42 Camino G, Lomakin SM, Lazzari M. Polydimethylsiloxane thermal degradation Part 1. Kinetic aspects. Polymer 2001; 42: 2395–2402. 32 Qin Y, Gianchandani YB An all electronic, fully microfabricated micro gas chromatograph. IEEE International Conference on Solid-State Sensors, Actuators and Microsystems (Transducers); 21–25 June 2015; Anchorage, Alaska; 2015: 626–629. This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/ by/4.0/ 33 Gupta NK, Gianchandani YB. Thermal transpiration in mixed cellulose ester membranes: Enabling miniature, motionless gas pumps. Microporous and Meso- porous Materials. 2011; 142: 535–541. 34 Radadia AD, Salehi-Khojin A, Masel RI et al. The effect of microcolumn geometry on the performance of micro-gas chromatography columns for chip scale gas analyzers. Sensors and Actuators B: Chemical 2010; 150: 456–464. doi:10.1038/micronano.2015.49 Microsystems & Nanoengineering Microsystems & Nanoengineering
https://openalex.org/W4200057727	https://zenodo.org/records/5751507/files/B82971210220.pdf	English	null	PM 2.5 Concentration Prediction By Data Mining Method	International journal of innovative technology and exploring engineering	2,021	cc-by	5,257	International Journal of Innovative Technology and Exploring Engineering (IJITEE) ISSN: 2278-3075 (Online), Volume-11 Issue-1, November, 2021 Hung Thuan Nguyen, Chi Quynh Nguyen These algorithms gave positive results in terms of predictive accuracy. However, these methods were conducted in datasets collected at different times and locations, so it is difficult to choose a predictive method from the above studies that is suitable for the data about the air in Hanoi. Therefore, we have surveyed various studies related to pollution level of PM2.5 indicators to get an overview on prediction methods in part 2. On that basis, in part 3, we analyze the collected data, propose a new way to extract features and choose a suitable model training method for PM2.5 prediction in Hanoi in one hour later. Meteorological indicators are necessary for the prediction, in addition to other pollution indicators (particulate matter with a diameter of 10 µm - PM10, concentration of CO2, total volatile organic compounds - TVOC) and time factor is also considered to influence the predictions. By this extraction method, we make a comparison between the old extraction method and tests on different prediction models: SVM, RF, MLP (Multi-layer Perceptron) and XGBoost (Extreme Gradient Boosting) in part 4. Finally, we draw the conclusions and discuss future development in part 5. Abstracts: The global air pollution is constantly increasing and causing negative effects on human health such as respiratory, cardiovascular diseases and cancers. Recently, pollution in Hanoi has become increasingly worse, especially when PM2.5 concentration is always at high level. Thus, PM2.5 prediction is of more urgency to issue early forecasts. Depending on air data including meteorological indicators and air pollution indicators collected in Hanoi, we have proposed a new characteristic extraction method that gave better results when uing the same algorithm compared to those of old methods. XGBoost algorithm was applied to predict the concentration of PM2.5 and the test showed that the accuracy of this algorithm is higher than that of other data mining algorithms while the training time is significantly lower. Keywords: air quality forecasting, data mining, PM2.5 prediction, XGBoost. I. INTRODUCTION In Increasing air polution is raising many problems concerning human health. According to World Health Organization, air pollution has impact on everbody in all countries. This resulted in 4.2 million premature deaths globally in 2016, 91% of which is from South East Asian and West Pacific countries. The main cause derives from particulate matter sized 2.5 µm or smaller in the air pollution, which are responsible for cardiovascular, respiratory and cancer diseases. The problem of air pollution is more serious in big cities due to the high density of the population causing the increase in air emissions. Besides, the construction of buildings, roads also increases the amount of dust in the air in big cities. Hanoi is facing the increase in air pollution. In September 2019, Hanoi was ranked as one of the most air-polluted cities in the world. The key factor to this situation is the surge of PM2.5 concentration in the air. This dust type has a negative impact on human health, therefore, predicting the PM2.5 dust pollution level is increasingly necessary. PM2.5 prediction methods have been studied in developed countries over the years. The algorithms applied included hybrid system combined with fuzzy inference, Random Forest (RF), Support Vector Machine (SVM) and neural network. PM2.5 Concentration Prediction By Data Mining Method Hung Thuan Nguyen, Chi Quynh Nguyen II. RELATED WORKS In this part, we conducted a survey of related studies. First of all, some of the studies applied Adaptive Neuro Fuzzy Inference System (ANFIS) for prediction. The use of ANFIS made progress only when Fuzzy Inductive Reasoning (FIR) was applied; however, it did not make much difference. This inference derived from the study about PM2.5 concentration in the center of Mexico city. However, this study did not exploit many meteorological factors to predict the PM2.5 concentration. Another study on PM10 concentration prediction in Konya city also employed ANFIS, including some meteorological factors: temperature, humidity, pressure and wind speed. In terms of data processing, Output-dependent data scaling (ODDS) was recommeded, and it gave a promising result. However, the historical values of PM10 concentration were not taken into prediction [2][3]. In addition, there are studies that applied other algorithms such as SVM, RF in air quality forecasting. These studies all used meteorological factors and historical values of pollutants as inputs for their algorithms. As for studies using SVM method, the results are positive, but each pollutant was only consistent with a certain kernel function. According to the test results, RBF kernel function produced the best result for SO2 indicator, while linnear function worked the best on NO2 indicator. Besides SVM, RF is also an algorithm applied by a number of studies in air quality forecasting. A study conducted in Shenyang city (China) built an RF-based RAQ algorithm for air quality forecasting in the city. In addition, there are studies that applied other algorithms such as SVM, RF in air quality forecasting. These studies all used meteorological factors and historical values of pollutants as inputs for their algorithms. As for studies using SVM method, the results are positive, but each pollutant was only consistent with a certain kernel function. According to the test results, RBF kernel function produced the best result for SO2 indicator, while linnear function worked the best on NO2 indicator. Besides SVM, RF is also an algorithm applied by a number of studies in air quality forecasting. A study conducted in Shenyang city (China) built an RF-based RAQ algorithm for air quality forecasting in the city. Manuscript received on November 29, 2021. Revised Manuscript received on November 27, 2021. Manuscript published on November 30, 2021. * C d A h Hung Thuan Nguyen, Department of Bachelor of Science, Posts and Telecommunications Institute of Technology, Hanoi, Vietnam. PM2.5 Concentration Prediction By Data Mining Method PM2.5 Concentration Prediction By Data Mining Method They built and tested on data sets collected from 10 monitoring stations that included many factors: meteorological data, data of air pollution indicators, traffic data and geography. The prediction method with RAQ algorithm gave outstanding results, with accuracy up to 81.5%, while those of artificial neural network (ANN) and Decision Tree was respectively only 71.8% and 77.4%. Another study applying RF in their prediction method was conducted with data set collected in Warsaw city to predict the average level of contamination of substances over the next day. The method performed consisted of 2 main phases: feature selection and predictive application. In the phase of feature selection, they implemented two methods: Genetic Algorithm (GA) and Stepwise Fit (SF) to remove the features from the original feature set. In predictive application phase, they built two models, one with features through neural networks and other machine learning algorithms (MLP, RBF, SVM) and then RF to synthesize the predicted results of mentioned networks, the other model had the feature as a direct input to the RF. The pollution indicators tested for prediction in this study included: PM10, SO2, NO2, O3. The results showed that the selection of features influenced the predicted results, the SF method often gave results higher than GA up to 2.88%.[4] [5][4][6] [7] humidity, and light, as these are indicators reflecting weather and environmental conditions. They are also key factors in PM2.5 prediction model. Figure I – The data distribution of indicators in the data set In addition to SVM and RF, neural networks are also applied in PM2.5 prediction. The study on the data set collected in Hefei (China) gave high accuracy when conducting PM2.5 concentration prediction in the next day by artificial neural network (ANN). Their data included PM2.5 concentration and meteorological data. The design model had vector input including: PM2.5 concentration and meteorological factors (temperature, wind speed, wind direction, humidity). The study gave predictions high accuracy with the following measurements: Mean Absolute Error (MAE) [μg/m3] 0.92472; Root-mean-square Error (RMSE) [μg/m3]: 1.2756; Coefficient of Determination (R2-score): 0.9188; R: 0.9315 [8]. In recent years, Extreme Gradient Boosting (XGBoost) algorithm has emerged in solving this problem. Some studies applying this algorithm have superior accuracy compared to RF, MLP with shorter training time [9] [10]. PM2.5 Concentration Prediction By Data Mining Method Figure I – The data distribution of indicators in the data set Next, we removed the other unnecessary indicators by evaluating the correlation level with the PM2.5 indicator and its value. From and Figure I and Figure II, it can be infered that the presence of CO imposed no impact on the prediction, as its all values are equal to zero. Besides, the indicators TVOC, SO2, NH3, O3 were also omitted because they showed no correlation with PM2.5 indicator as shown in Figure 2. III. METHOD 5 Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP) © Copyright: All rights reserved. Figure 2 – Correlation chart of pollution indicators with PM2.5 indicator In this part, we present the implementation method including the steps: analyzing the collected data set, proposing feature selection and building predictive model. Figure 2 – Correlation chart of pollution indicators with PM2.5 indicator Figure 2 – Correlation chart of pollution indicators with PM2.5 indicator II. RELATED WORKS Email: hungnt61h@gmail.com Chi Quynh Nguyen*, Department of Computer Science, Posts and Telecommunications Institute of Technology, Hanoi, Vietnam. Email: chinq@ptit.edu.vn Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP) © Copyright: All rights reserved. Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP) © Copyright: All rights reserved. © The Authors. Published by Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/) Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP) © Copyright: All rights reserved. Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP) © Copyright: All rights reserved. Retrieval Number: 100.1/ijitee.B82971210220 DOI: 10.35940/ijitee.B8297.1111121 Journal Website: www.ijitee.org Retrieval Number: 100.1/ijitee.B82971210220 DOI: 10.35940/ijitee.B8297.1111121 Journal Website: www.ijitee.org 64 Retrieval Number: 100.1/ijitee.B82971210220 DOI: 10.35940/ijitee.B8297.1111121 Journal Website: www.ijitee.org A. Data description 329455 329455 329455 Average 17.927 3.535 14.405 Standard bias 8.972 1.605 4.604 Global minimum 0 1 0 25% 11 2 12.9 50% 17 4 14.3 75% 24 5 14.3 Global maximum 306 6 137.9 International Journal of Innovative Technology and Exploring Engineering (IJITEE) ISSN: 2278-3075 (Online), Volume-11 Issue-1, November, 2021 International Journal of Innovative Technology and Exploring Engineering (IJITEE) ISSN: 2278-3075 (Online), Volume-11 Issue-1, November, 2021 International Journal of Innovative Technology and Exploring Engineering (IJITEE) ISSN: 2278-3075 (Online), Volume-11 Issue-1, November, 2021 International Journal of Innovative Technology and Exploring Engineering (IJITEE) ISSN: 2278-3075 (Online), Volume-11 Issue-1, November, 2021 Table 1 – Value description of CO, NH3, O3, PM10 indicators Table 1 – Value description of CO, NH3, O3, PM10 indicators Table 1 – Value description of CO, NH3, O3, PM10 indicators CO NH3 O3 PM10 Record No. 329455 329455 329455 329455 Average 0.000 5.052 17.991 28.401 Standard deviation 0.000 2.374 2.456 15.641 Global minimum 0 0.5 6.5 0 25% 0 4.6 17.6 15.7 50% 0 5 18.5 25 75% 0 5.3 18.5 41.7 Global maximum 0 99.6 104.5 86.6 ▪ Meteorological features in the previous 24 hours: f36–38 - maximum, minimum and average PM2.5 indicator values in the previous 24 hours; f39–41 - maximum, minimum and average temperature values in the previous 24 hours; f42–44 - maximum, minimum and average values of humidity in the previous 24 hours. These features aim at showing the degree of environmental volatility within 24 hours, which affects the change in the PM2.5 indicator in the next hour. The predicted value is the average value of the PM2.5 indicator in the next hour. After feature extraction process, we executed data normalization by z-score normalization formula: 𝑧= 𝑥−𝜇 𝜎 (1) (1) Table 2 – Value description of PM0.1, TVOC, SO2 indicators µ of which is the average of the elements, σ is the standard deviation, x is the value to be normalized. PM0.1 TVOC SO2 Record No. 329455 329455 329455 Average 17.927 3.535 14.405 Standard bias 8.972 1.605 4.604 Global minimum 0 1 0 25% 11 2 12.9 50% 17 4 14.3 75% 24 5 14.3 Global maximum 306 6 137.9 Retrieval Number: 100.1/ijitee.B82971210220 DOI: 10.35940/ijitee.B8297.1111121 Journal Website: www.ijitee.org C. Predictive model description The predictive model we proposed includes the training process and the predictive process shown in Figure . p p p In the training process, from the input data which are meteorological indicators and pollution indicators, we extracted the 44-dimensional feature vector as described in the previous section. This vector was normalized, and the algorithm we applied was XGBoost built on Gradient Boost [11]. This algorithm has won numerous contests on Kaggle (the Machine Learning and Data Science community that regularly hosts competitions in the field) in recent years. Unlike RF, this algorithm uses boosting method to solve the problem. More specifically, new trees are generated sequentially with the aim of minimizing the error from the previous tree by partially re-learning the error from the previous tree, updating the bug to get a better tree. From that on, in the previous step, the wrongly assigned points will have more chances of being corrected in the future [12]. It can be seen that PM10 và PM2.5 indicators have the strongest relationship among the above indicators, so they are kept. Finally, the indicators needed for prediction are temperature, humidity, light, CO2, PM10 and distribution value are described in Error! Reference source not found. In the next part, we present the method to extract feature from the remaining indicators after data pre-processing. Figure 3 – Predictive model A. Data description Our data set was collected at an observation station in Hanoi from August 17, 2018 to July 22, 2019. Each record in the data set contains columns: time, SO2, NH3, O3, PM2.5, PM10, CO2, PM0.1, TVOC, CO, temperature, humidity, light. The sampling period was approximately 40 seconds/time. However, the data set had several nulls and noisy records. The distribution of feature values (columns) is shown in Figure 1. The existence of noisy records drove the chart of value distribution of most indicators to incline greatly towards the left. Next we filtered out noisy records and executed feature extraction. Figure 2 – Correlation chart of pollution indicators with PM2.5 indicator Figure 2 – Correlation chart of pollution indicators with PM2.5 indicator Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP) © Copyright: All rights reserved. Figure 2 Correlation chart of pollution indicators with PM2.5 indicator First, we removed the defective and noisy records that are outside the permissible range (for example, PM0.1 indicator exists negative value or measured temperature is greater than 50 C degrees). Through surveying many studies, meteorological factors: we stored the data of temperature, Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP) © Copyright: All rights reserved. Retrieval Number: 100.1/ijitee.B82971210220 DOI: 10.35940/ijitee.B8297.1111121 Journal Website: www.ijitee.org Retrieval Number: 100.1/ijitee.B82971210220 DOI: 10.35940/ijitee.B8297.1111121 Journal Website: www.ijitee.org 65 International Journal of IS Table 1 – Value description of CO, NH3, O3, PM10 indicators CO NH3 O3 PM10 Record No. 329455 329455 329455 329455 Average 0.000 5.052 17.991 28.401 Standard deviation 0.000 2.374 2.456 15.641 Global minimum 0 0.5 6.5 0 25% 0 4.6 17.6 15.7 50% 0 5 18.5 25 75% 0 5.3 18.5 41.7 Global maximum 0 99.6 104.5 86.6 Table 2 – Value description of PM0.1, TVOC, SO2 indicators PM0.1 TVOC SO2 Record No. PM2.5 Concentration Prediction By Data Mining Method randomly and divided into 2 sets: the training set accounts for 75% of the initial data and the remaining 25% of the data is the test set. 𝑦̂𝑖= 𝜙(𝒙𝑖) = ∑𝑓𝑘(𝒙𝑖) 𝐾 𝑘=1 , 𝑓𝑘∈ℱ (2) (2) The measures we used for evaluation include formulas: R2 - score (9), MAE (10) and RMSE (11) as follows: In this formula, ℱ= {𝑓(𝒙) = 𝑤𝑞(𝒙)}(𝑞: ℝ𝑚) →𝑇, 𝑤∈ℝ𝑇, 𝑞 is the tree to map the vector to the predictive value at leaf node, 𝑇 is the number of leaf node, 𝐾 is the number of tree, 𝑓𝑘 is the independent kth tree in the model, 𝑤𝑖 is the weight of ith leaf node and 𝑦̂𝑖 is the predictive value of the target function: ( ) ( ) 𝑅2 = 1 − ∑ (𝑦𝑖−𝑦̂𝑖)2 𝑛 𝑖=1 ∑ (𝑦𝑖−𝑦̅)2 𝑛 𝑖=1 (9) 𝑀𝐴𝐸= 1 𝑛∑\|𝑦𝑖−𝑦̂𝑖\| 𝑛 𝑖=1 (10) 𝑅𝑀𝑆𝐸= √1 𝑛∑(𝑦𝑖−𝑦̂𝑖)2 𝑛 𝑖=1 (11) 𝑅2 = 1 − ∑ (𝑦𝑖−𝑦̂𝑖)2 𝑛 𝑖=1 ∑ (𝑦𝑖−𝑦̅)2 𝑛 𝑖=1 (9) (9) 𝑓𝑘 is the independent kth tree in the model, 𝑤𝑖 is the weight of ith leaf node and 𝑦̂𝑖 is the predictive value of the target function: (10) ℒ(𝜙) = ∑𝑙(𝑦𝑖, 𝑦̂𝑖) 𝑛 𝑖=1 + ∑Ω(𝑓𝑘) 𝐾 𝑘 (3) (3) (11) in which, 𝑛 is the number of data point, Ω(𝑓) = 𝛾𝑇+ 1 2 𝜆‖𝑤‖2 is the regularization. As the target function can not be optimized by Stochastic Gradient Descent (SGD) method, learning process is described as follow: With 𝑦̂𝑖 (𝑡) = 𝑦̂𝑖 (𝑡−1) + 𝑓𝑡(𝒙𝑖) and 𝑦̂𝑖 (0) = 0 , 𝑦̂𝑖 (𝑡) as the predictive value of ith instance at tth iteration, the current target function is: in which, 𝑛 is the number of data point, Ω(𝑓) = 𝛾𝑇+ 1 2 𝜆‖𝑤‖2 is the regularization. As the target function can not be optimized by Stochastic Gradient Descent (SGD) method, learning process is described as follow: In which, n is the number of elements, yi is the real value, 𝑦̂𝑖 is the predicted value, and 𝑦̅ is the mean of the number of elements. These metrics are employed because they clearly represent the difference between the real value and the predicted value. This is suitable for regression problems because the predicted value is on the continuous domain instead of labels like the classification problem. B. Feature extraction method The features we extracted are based on the selection of features of the studies we have investigated previously, among which, SF and GA methods are applied to detect the best from the initial set of features. In order to solve PM2.5 prediction problem in Hanoi, we took the potential features from the results of the study in Warsaw city [7]. Those recommended features include:[7] [8] ▪ Current features: f1 - current PM2.5 indicator value; f2 - current PM10 indicator value; f3 - present temperature value; f4 - current moisture value; f5 - current light value; f6 - current CO2 indicator value. These are the values that describes the current air quality and aims at supporting prediction in the next hour. ▪ Time-based features: f7-8 - season (2-bit representation: 00 - spring, 01 - summer, 10 - autumn, 11 - winter); f9 - day off (1 - days off, 0 – working days); f10 - hour. Seasonal features are essential because the climate in Hanoi is tropical monsoon; as a result, there are four seasons in a year despite Hanoi’s tropical location. In addition, the day off and daytime features are also considered because air pollution is mainly triggered by human activities. Figure 3 – Predictive model The data set consists of (𝒙𝑖, 𝑦𝑖) pairs of which are the 44-dimensional feature vector and the corresponding predictive value. Learning model is described as follow: Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP) © Copyright: All rights reserved. ▪ Features of the previous 24 hours: f11–35 - PM2.5 indicator values for the previous 1 to 24 hours. These features provide hourly variation tracking to predict the next hour. Retrieval Number: 100.1/ijitee.B82971210220 DOI: 10.35940/ijitee.B8297.1111121 Journal Website: www.ijitee.org 66 Table 3 – The comparison results between two methods 𝑗 Error function in the entire tree: 𝑇 Table 3 – The comparison results between two methods Published By: Blue Eyes Intelligence Engineering Table 3 – The comparison results between two methods R2 – score MAE RMSE Method 1 0.9508 0.1387 0.2266 Method 2 0.9368 0.1515 0.2521 Figure 4 – The predictive results of two methods ℒ̃ (𝑡) = −1 2 ∑ 𝐺𝑗 2 𝐻𝑗+ 𝜆 𝑇 𝑗=1 + 𝛾𝑇 (8) (8) The training process ends after some iterations or when the target function value is less than a certain threshold. After the training process, the model is used to predict the mean value of the PM2.5 indicator over the next hour. With the data on meteorological and pollution indicators within 24 hours as input, the data was extracted into a 44-dimensional vector and then normalized. This vector is fed into the training model to provide a predictive value. In the next section, we test the extraction method and the predictive model that has been presented. Figure 4 – The predictive results of two methods Retrieval Number: 100.1/ijitee.B82971210220 DOI: 10.35940/ijitee.B8297.1111121 Journal Website: www.ijitee.org PM2.5 Concentration Prediction By Data Mining Method In terms of R2 – score, the higher the value, the stronger the model (which represents relevance to the data set) and preferably 1.00, with MAE and RMSE as small as possible (these two measures represent difference between predicted value and real value) g p With 𝑦̂𝑖 (𝑡) = 𝑦̂𝑖 (𝑡−1) + 𝑓𝑡(𝒙𝑖) and 𝑦̂𝑖 (0) = 0 , 𝑦̂𝑖 (𝑡) as the predictive value of ith instance at tth iteration, the current target function is: g p With 𝑦̂𝑖 (𝑡) = 𝑦̂𝑖 (𝑡−1) + 𝑓𝑡(𝒙𝑖) and 𝑦̂𝑖 (0) = 0 , 𝑦̂𝑖 (𝑡) as the predictive value of ith instance at tth iteration, the current target function is: ℒ(𝑡) = ∑𝑙(𝑦𝑖, 𝑦̂𝑖 (𝑡−1) + 𝑓𝑡(𝒙𝑖)) 𝑛 𝑖=1 + Ω(𝑓𝑡) (4) (4) (4) Approximation formula is: 𝑛 ℒ(𝑡) ≃∑[𝑙(𝑦𝑖, 𝑦̂(𝑡−1)) + 𝑔𝑖𝑓𝑡(𝒙𝑖) + 1 2 ℎ𝑖𝑓𝑡 2(𝒙𝑖)] 𝑛 𝑖=1 + Ω(𝑓𝑡) (5) With 𝑔𝑖= 𝜕𝑦̂(𝑡−1)𝑙(𝑦𝑖, 𝑦̂(𝑡−1)), ℎ𝑖= 𝜕 𝑦̂(𝑡−1) 2 𝑙(𝑦𝑖, 𝑦̂(𝑡−1)), if the constant is obmitted, the target function can be simplied as below: 𝑛 ℒ(𝑡) ≃∑[𝑙(𝑦𝑖, 𝑦̂(𝑡−1)) + 𝑔𝑖𝑓𝑡(𝒙𝑖) + 1 2 ℎ𝑖𝑓𝑡 2(𝒙𝑖)] 𝑛 𝑖=1 + Ω(𝑓𝑡) (5) With 𝑔𝑖= 𝜕𝑦̂(𝑡−1)𝑙(𝑦𝑖, 𝑦̂(𝑡−1)), ℎ𝑖= 𝜕 𝑦̂(𝑡−1) 2 𝑙(𝑦𝑖, 𝑦̂(𝑡−1)), if the constant is obmitted, the target function can be simplied as below: Next, we compared the results our extraction method presented in part 3 (Method 1) and another extraction method that includes only features extracted from meteorological factors (Method 2) [8]. Specifically, our method takes into account the time of day and year, in companion with the PM10 indicator and the input data of the indicators in the previous 24 hours. Method 2 only concerned the meteorological factors in the current range. Let’s compare the performance results with the measurements presented in Table III. The predicted results of the two methods are shown in Figure 5: the left side of which is the comparation between the real value with the predicted value when method 1 was applied, on the right is that of method 2. the constant is obmitted, the target function can be simplied as below: ℒ̃ (𝑡) = ∑[𝑔𝑖𝑓𝑡(𝒙𝑖) + 1 2 ℎ𝑖𝑓𝑡 2(𝒙𝑖)] 𝑛 𝑖=1 + Ω(𝑓𝑡) (6) Set 𝐺𝑗= ∑ 𝑔𝑖 𝑖∈𝐼𝑗 , 𝐻𝑗= ∑ ℎ𝑖 𝑖∈𝐼𝑗 , với 𝐼𝑗= {𝑖\|𝑞(𝒙𝑖) = 𝑗} to be the value set at leaf node 𝑗. PM2.5 Concentration Prediction By Data Mining Method ℒ̃ (𝑡) = ∑[𝑔𝑖𝑓𝑡(𝒙𝑖) + 1 2 ℎ𝑖𝑓𝑡 2(𝒙𝑖)] 𝑛 𝑖=1 + Ω(𝑓𝑡) (6) (6) 𝑖1 Set 𝐺𝑗= ∑ 𝑔𝑖 𝑖∈𝐼𝑗 , 𝐻𝑗= ∑ ℎ𝑖 𝑖∈𝐼𝑗 , với 𝐼𝑗= {𝑖\|𝑞(𝒙𝑖) = 𝑗} to be the value set at leaf node 𝑗. Set 𝐺𝑗= ∑ 𝑔𝑖 𝑖∈𝐼𝑗 , 𝐻𝑗= ∑ ℎ𝑖 𝑖∈𝐼𝑗 , với 𝐼𝑗= {𝑖\|𝑞(𝒙𝑖) = 𝑗} to be the value set at leaf node 𝑗. 𝑗 the value set at leaf node 𝑗. Optimized weight at each leaf node: 𝑤𝑗 ∗= − 𝐺𝑗 𝐻𝑗+ 𝜆 (7) 𝑤𝑗 ∗= − 𝐺𝑗 𝐻𝑗+ 𝜆 (7) 𝑗 Error function in the entire tree: 𝑇 REFERENCES 1. R. Yu, Y. Yang, L. Yang and G. Han, "RAQ–A Random Forest Approach for Predicting Air Quality in Urban Sensing Systems," Sensors, vol. 16, p. 86, 11 January 2016. Table V – Result comparison among algorithms R2 – score MAE RMSE Training time SVM 0.9553 0.1154 0.2101 27.0608 Random Forest 0.9587 0.1115 0.2020 35.5577 MLP 0.9562 0.1276 0.2078 8.2011 XGBoost 0.9595 0.1126 0.1999 4.8872 Table V – Result comparison among algorithms p y 2. WHO, "Air pollution," 2 May 2018. [Online]. Available: https://www.who.int/en/news-room/fact-sheets/detail/ambient-(outdo or)-air-quality-and-health. 3. C.-M. Vong, W.-F. Ip, P.-k. Wong and J.-y. Yang, "Short-Term Prediction of Air Pollution in Macau Using Support Vector Machines," Journal of Control Science and Engineering, vol. 2012, 2012. 4. K. Siwek and S. Osowski, "DATA MINING METHODS FOR PREDICTION OF AIR POLLUTION," Int. J. Appl. Math. Comput. Sci, vol. 26, 2016. Through the measure R2 - score, it can be seen that the XGBoost algorithm gave the highest fitting rate with data set (95.95%). Looking at the results of RMSE, we can see that the difference between the predicted value and the actual value is the smallest, or in other words, the accuracy of the prediction is the highest compared to the remaining values. Although the difference in accuracy among the algorithms is not too much, in terms of training time, XGBoost has the shortest. This shows the potential of this model in training and its predictive accuracy over time. 5. K. Polat and S. S. Durduran, "Usage of output-dependent data scaling in modeling and prediction of air pollution daily concentration values (PM10) in the city of Konya," Neural Computing and Applications, p. 21, 2011. À 6. À. Nebot and F. Mugica, "Small-particle pollution modeling using fuzzy approaches," Advances in Intelligent Systems and Computing, pp. 239-252, 2014. pp 7. NandigalaVenkatAnurag, YagnavalkBurra and S.Sharanya, "Air Quality Index Prediction with Meteorological Data Using Feature Based Weighted Xgboost," International Journal of Recent Technology and Engineering (IJRTE), vol. 8, no. 1, pp. 1355-1358, May 2019. y 8. A. Li và X. Xu, “A New PM2.5 Air Pollution Forecasting Model Based on Data Mining and BP Neural Network Model,” Advances in Computer Science Rese, tập 65, 2018. human activities in different time framework in a day or in a week. The experiments proved that our extraction method gave better results of PM2.5 prediction in Hanoi than the old method, which only concerned the meteorological factors. It can be seen that our extraction method gives ~ 2% higher results than the old method when we conduct the test on the same model. Therefore, PM10 indicator and time factor are proved to have an impact on the PM2.5 prediction in the next hour besides the basic meteorological factors such as temperature, humidity, light. The study also shows that the XGBoost algorithm is a good algorithm with high accuracy and short training time compared to other machine learning algorithms. Turning to our problem, this algorithm is suitable because of its accurate prediction ability and low model training cost. However, this algorithm's nature of trying to best match the data makes it susceptible to overfitting. So in the future, we will look at some methods to limit overfitting and do experiments on other deep learning algorithms to predict time series data problems. Next we compared the predictive model with other models: SVM, RF, MLP and XGBoost. The hyper-parameter of each algorithm is shown in Table IV. Table I – Hyper-parameter of each algorithm Hyper-parameter SVM gamma='auto' kernel='rbf' C=100 epsilon=0.0001 Random Forest n_estimators=150 max_features='auto' MLP hidden_layer_sizes=(192,128,96) max_iter=1000 learning_rate_init=0.01 tol=1e-6 batch_size=192 XGBoost n_estimators=200 max_depth=8 gamma=0.7 objective='reg:squarederror' In terms of the current data, we also lack some meteorological factors such as wind direction, wind speed. These are also factors that can affect the prediction of air pollution because wind can diffuse or concentrate dust density in an area. With the climate in Hanoi, wind also has different seasonal features such as wind direction, speed, and humidity. In addition, traffic data also needs to be considered due to the large number of personal vehicles in Hanoi. In the future, we will collect more data to observe the correlation between them and the level of air pollution in Hanoi and experiment with another model to improve the accuracy and predictive range of space and time. The criteria for similar comparison include the measures: R2 - score, MAE, RMSE and training time (in second). The results are presented in Table V. Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP) © Copyright: All rights reserved. IV. EXPERIMENTS AND RESULTS Figure 4 – The predictive results of two methods Since the sampling period is about 40 seconds, therefore, in order to conduct the test, they averaged those records by the hour. The results obtained were 6433 records of the hourly atmospheric readings. Next, we performed the preprocessing, extraction and normalization of this data. To perform the training and evaluation process, the records were taken Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP) © Copyright: All rights reserved. Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP) © Copyright: All rights reserved. Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP) © Copyright: All rights reserved. Retrieval Number: 100.1/ijitee.B82971210220 DOI: 10.35940/ijitee.B8297.1111121 Journal Website: www.ijitee.org 67 International Journal of Innovative Technology and Exploring Engineering (IJITEE) ISSN: 2278-3075 (Online), Volume-11 Issue-1, November, 2021 International Journal of Innovative Technology and Exploring Engineering (IJITEE) ISSN: 2278-3075 (Online), Volume-11 Issue-1, November, 2021 Retrieval Number: 100.1/ijitee.B82971210220 DOI: 10.35940/ijitee.B8297.1111121 Journal Website: www.ijitee.org V. CONCLUSION AND DEVELOPMENT TREND 9. M. Z. Joharestani, C. Cao, X. Ni, B. Bashir and S. Talebiesfandarani, "PM2.5 Prediction Based on Random Forest, XGBoost, and Deep Learning Using Multisource Remote Sensing Data," Atmosphere, 2019. By looking at the data we collected in Hanoi including meteorological factors and pollution indicators, we found that the PM10 indicator in Hanoi is correlated with the PM2.5 indicator. Since then, along with other surveys, we proposed a new feature extraction method. The new extraction method includes not only current meteorological and pollution factors, but also those of the past (many hours earlier). This provides better prediction because historical values help show the changing trend of the PM2.5 indicator in the next hour. In addition, time also plays a role in influencing the predicted results due to climate change, seasonal environmental changes in the year in Hanoi and different Learning Using Multisource Remote Sensing Data," Atmosphere, 2019. 10. W.-F. Ip, C.-M. Vong, J. Y. Yang and P. K. Wong, "Least squares support vector prediction for daily atmospheric pollutant level," Proc. 2010 IEEE/ACIS 9th International Conference on Computer and Information Science (ICIS), pp. 23-28, August 2010. 11. T. Chen and C. Guestrin, "XGBoost: A Scalable Tree Boosting System," 2016. 12. L. Breiman, "Random Forests," Machine Learning, vol. 45, pp. 5-32, 2001. g ) Machine Learning, vol. 45, pp. 5 32, Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP) © Copyright: All rights reserved. Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP) © Copyright: All rights reserved. Retrieval Number: 100.1/ijitee.B82971210220 DOI: 10.35940/ijitee.B8297.1111121 Journal Website: www.ijitee.org 68 PM2.5 Concentration Prediction By Data Mining Method Retrieval Number: 100.1/ijitee.B82971210220 DOI: 10.35940/ijitee.B8297.1111121 Journal Website: www.ijitee.org AUTHOR PROFILE o Hung Thuan Nguyen, graduated B.Sc in Information System at Posts and Telecommunications Institue of Technology, Hanoi, Vietnam in 2020, main research in data warehouse, mining time-series data and detect fraud in logistic industry. Chi Quynh Nguyen, graduated B.Sc in Information Technology at Hanoi University of Technology, Vietnam in 1999, M.Sc and Ph.D Candidate in Computer Science at University of California, Davis, USQ in 2004 and 2006. Her main research focuses on datawarehousing, data mining and bioinformatics. 69 Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP) © Copyright: All rights reserved. ber: 100.1/ijitee.B82971210220 /ijitee.B8297.1111121 e: www.ijitee.org 69 Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP) © Copyright: All rights reserved. 69 Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP) © Copyright: All rights reserved. Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP) © Copyright: All rights reserved. Retrieval Number: 100.1/ijitee.B82971210220 DOI: 10.35940/ijitee.B8297.1111121 Journal Website: www.ijitee.org Retrieval Number: 100.1/ijitee.B82971210220 DOI: 10.35940/ijitee.B8297.1111121 Journal Website: www.ijitee.org 69
https://openalex.org/W2143771453	https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0048976&type=printable	English	null	Segmentation and Visual Analysis of Whole-Body Mouse Skeleton microSPECT	PloS one	2,012	cc-by	6,451	Abstract s funded by Medical Delta (http://www.medicaldelta.nl/research/large-research-programs-and-initiatives/molecular-image-processing) in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: b.p.f.lelieveldt@lumc.nl Segmentation and Visual Analysis of Whole-Body Mouse Skeleton microSPECT Artem Khmelinskii1, Harald C. Groen2, Martin Baiker1, Marion de Jong2,3, Boudewijn P. F. Lelieveldt1,4* 1 Division of Image Processing, Department of Radiology, Leiden University Medical Center, Leiden, The Netherlands, 2 Department of Nuclear Medicine, Erasmus MC, Rotterdam, The Netherlands, 3 Department of Radiology, Erasmus MC, Rotterdam, The Netherlands, 4 Department of Intelligent Systems, Delft University of Technology, Delft, The Netherlands 1 Division of Image Processing, Department of Radiology, Leiden University Medical Center, Leiden, The Netherlands, 2 Department of Nuclear Medicine, Erasmus MC, Rotterdam, The Netherlands, 3 Department of Radiology, Erasmus MC, Rotterdam, The Netherlands, 4 Department of Intelligent Systems, Delft University of Technology, Delft, The Netherlands November 2012 \| Volume 7 \| Issue 11 \| e48976 Abstract Whole-body SPECT small animal imaging is used to study cancer, and plays an important role in the development of new drugs. Comparing and exploring whole-body datasets can be a difficult and time-consuming task due to the inherent heterogeneity of the data (high volume/throughput, multi-modality, postural and positioning variability). The goal of this study was to provide a method to align and compare side-by-side multiple whole-body skeleton SPECT datasets in a common reference, thus eliminating acquisition variability that exists between the subjects in cross-sectional and multi- modal studies. Six whole-body SPECT/CT datasets of BALB/c mice injected with bone targeting tracers 99mTc-methylene diphosphonate (99mTc-MDP) and 99mTc-hydroxymethane diphosphonate (99mTc-HDP) were used to evaluate the proposed method. An articulated version of the MOBY whole-body mouse atlas was used as a common reference. Its individual bones were registered one-by-one to the skeleton extracted from the acquired SPECT data following an anatomical hierarchical tree. Sequential registration was used while constraining the local degrees of freedom (DoFs) of each bone in accordance to the type of joint and its range of motion. The Articulated Planar Reformation (APR) algorithm was applied to the segmented data for side-by-side change visualization and comparison of data. To quantitatively evaluate the proposed algorithm, bone segmentations of extracted skeletons from the correspondent CT datasets were used. Euclidean point to surface distances between each dataset and the MOBY atlas were calculated. The obtained results indicate that after registration, the mean Euclidean distance decreased from 11.5612.1 to 2.662.1 voxels. The proposed approach yielded satisfactory segmentation results with minimal user intervention. It proved to be robust for ‘‘incomplete’’ data (large chunks of skeleton missing) and for an intuitive exploration and comparison of multi-modal SPECT/CT cross-sectional mouse data. Groen HC, Baiker M, de Jong M, Lelieveldt BPF (2012) Segmentation and Visual Analysis of Whole-Body Mouse Skeleton microSPECT. PLoS 0.1371/journal.pone.0048976 Received July 6, 2012; Accepted October 1, 2012; Published November 12, 2012 Copyright: 2012 Khmelinskii et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. nding: This work was funded by Medical Delta (http://www.medicaldelta.nl/research/large-research-programs-and-initiatives/molecu e funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. PLOS ONE \| www.plosone.org Introduction structures (organs) [1,2]. Also, there are no standardized protocols for animal positioning: if a subject is imaged using different imaging modalities and protocols during follow-up studies or if different animals are used, the subject is positioned in different ways and postural variations occur (e.g., of the head, back and front limbs, etc.). One way to cope with this variability is to use multimodal animal holders between different scanners or use combined SPECT/CT, PET/CT, PET/MRI, or PET/SPECT/ CT scanners that are becoming increasingly available. However, multimodal holders are not widely used or compatible and when they are, there are still significant differences in animal posture between different time points [3]. Whole-body small animal imaging is widely used for the in vivo visualization of functional and anatomical information to study cancer, and for evaluation of drugs in pre-clinical research. An efficient combination of functional and structural information enables the visualization of cellular function and the follow-up of molecular processes in the living animals in their anatomical context. Functional information is provided by modalities such as Positron Emission Tomography (PET), Single Photon Emission Computed Tomography (SPECT), Magnetic Resonance Imaging (MRI) and Optical Imaging (OI), while anatomical information is usually obtained using Computed Tomography (CT) and ultra- sound. Various approaches were proposed to handle heterogeneous multi-modality data: Joshi et al. [4] proposed a method for fitting an elastically deformable mouse atlas to surface topographic range data acquired by an optical system; this method does not incorporate the extremities. Savinaud et al. [5] proposed a novel model-based approach to track animals in 3D from monocular video which allows the quantification of bioluminescence (BLI) signal on freely moving animals. Wildeman et al. [6] proposed a 2D/3D registration of mCT data to multiview photographs based on a 3D distance map combining optical/BLI data with CT. Suh The data heterogeneity and volume created by whole-body multimodality imaging presents a complex problem with respect to combining, analyzing and quantifying this data with low inter- observer and intra-observer variability and minimal human input. This is caused in part by a high degree of shape and postural variability present in follow-up and cross-sectional animal studies. Whole-body SPECT/CT Mouse Data Acquisition y / q In this study, we aim to demonstrate the robustness of the atlas- based segmentation with respect to the different whole-body SPECT scan settings and high posture variability that exists between those scans. Therefore, six male, BALB/c mice were retrospectively collected from several different imaging studies to represent that variability. One half was intravenously injected with 5066 MBq 99mTc-MDP, and the other one with 5066 MBq 99mTc-HDP. All mice were scanned 4 hours later using the Bioscan NanoSPECT/CT TM device (Washington DC, USA), equipped with four gamma cameras and pinhole apertures. With the combined scanner, the SPECT and CT were acquired one after the other without movement of the animal, so both imaging modalities are registered by hardware calibration. SPECT images were reconstructed using the ordered subset expectation maximi- zation (OSEM) and CT images using the filtered back projection (FBP) algorithms. 99mTc-MDP and 99mTc-HDP are gamma- emitting radionuclide substances, where the metastable techne- tium (99mTc) is tagged onto a phosphonate compound (MDP, HDP) to generate 99mTc-MDP and 99mTc-HDP respectively, which selectively concentrate in the bone and are the primary imaging agents used to image changes in bone vascularity and osteoblastic activity [20]. Both tracers are used in translational research. For all mice, part of the tracer is cleared by the liver and as such, this organ is visible as well. Between these six datasets, the resolution of the scanner varies, ranging from SPECT voxel size of 0.6060.6060.60 mm3 to 0.2060.2060.20 mm3. The highest resolution CT dataset has a voxel size of 0.1060.1060.10 mm3 and the lowest 0.9960.9961.00 mm3 (see Table 1 for further detail). All procedures involving animals were approved by the Animal Experimental Committee (DEC) of the Erasmus MC and performed in agreement with The Netherlands Experiments on The main technical contributions of this paper are twofold: N We present a semi-automated atlas-based skeleton segmenta- tion method for whole-body SPECT mouse data that requires minimal user input; N Using the Articulated Planar Reformation (APR) algorithm [12,13], we provide the user with an intuitive side-by-side comparison and exploration platform for multi-modal (SPECT/CT), cross-sectional and follow-up data in a standardized layout, independent on the position of the animal during acquisition. Introduction This variability is due to the fact that an animal body is a highly deformable system with many rigid parts (bones) and non-rigid November 2012 \| Volume 7 \| Issue 11 \| e48976 1 PLOS ONE \| www.plosone.org Mouse SPECT Segmentation Mouse SPECT Seg PLOS ONE \| www.plosone.org 2 November 2012 \| Volume 7 \| Issue 11 \| e48976 November 2012 \| Volume 7 \| Issue 11 \| e48976 PLOS ONE \| www.plosone.org PLOS ONE \| www.plosone.org 2 Mouse SPECT Segmentation Mouse SPECT Segmentation Figure 1. Overview of the proposed segmentation method. Given a SPECT dataset (a), the skeleton is extracted from the SPECT dataset (b). Next, the atlas skeleton (c) and the extracted skeleton (b) are registered to each other (d, e) using an anatomically realistic kinematic model. After the registration, the segmented data is reformatted into segments corresponding to the mouse atlas and thus mapping the data to a standardized atlas space (f, g). The data is now ready for an easy, fast and intuitive side-by-side exploration (multi-modal, follow-up or cross-sectional data) (h). doi:10.1371/journal.pone.0048976.g001 et al. [7] published a serial registration method to serial mCT/ SPECT mouse lower extremities images. et al. [7] published a serial registration method to serial mCT/ SPECT mouse lower extremities images. mCT skeleton surface. The results are compared to the results present in the literature for low-contrast mCT whole-body mouse data [1]. For a visual overview of the proposed method see Figure 1. In [2,3], the authors suggested the use of articulated whole-body small animal atlases as a standard geometric reference to tackle the problem of segmenting and organizing heterogeneous whole-body multi-modality small animal data. Using the articulated whole- body MOBY atlas, Baiker et al. presented a fully-automated skeleton registration and organ approximation method in low- contrast mCT mouse data [1]. This method exploits the high contrast of bone to automate the registration process of the skeleton model and the subsequent organ approximation. How- ever, performing an anatomical CT scan together with a functional one is not always desired in longer term follow-up studies, where prolonged radiation exposure may become a confounding factor in cancer research, or may cause adverse radiation effects [8]. Articulated MOBY Atlas A realistic 4D digital mouse phantom was generated by Segars et al. [15] based on high-resolution 3D MRI data of a C57BL/6, 15 week old mouse from Duke University. The skeleton in this atlas did not distinguish between single bones and joints. To allow the registration to perform independent of the data acquisition protocol and large postural variations due to postural heteroge- neity between scans, we presented a segmentation of the skeleton into individual bones and added anatomically realistic kinematic constraints and DoFs to each joint in [1,3,16]. Using the AmiraTM V3.1 software [17] and guided by anatomical text books [18,19] the following bones/bone groups were labeled: scapulae, humeri (upper front limbs), ulnae/radii (lower front limbs), manus (front paws), femora (upper hind limbs), tibiae/fibulae (lower hind limbs), pedes (hind paws), caput (skull), columna vertebralis (spine), costae (ribs), sternum (chest bone), and pelves. Each joint position was identified and the corresponding DoFs and kinematic constraints were specified. Two types of joints were distinguished: ball joints and hinge joints. The resulting articulated version of the MOBY skeleton can be seen in Figure 2. The goal of this study is to provide a segmentation and exploration tool for whole-body skeleton SPECT mouse data that eliminates any postural variability between the study subjects with minimal user intervention. Whole-body skeletal SPECT imaging with bone targeting tracers is of great interest for arthritis studies [9], development of bone pain palliation agents [10] as well in the field of bone metastases imaging in animal models. Since the location of metastatic appearance is unknown, whole-body scans, including follow-up, are essential to assess the growth and/or metastatic response to treatment [11]. This is a challenging task due to the nature of whole-body skeleton SPECT data: usually noisy, due to the relative short acquisition time and low resolution with an incomplete skeleton image (several portions missing in limbs, skull, etc.). SPECT/CT Data Pre-processing Animals Act (1977) and the European Convention for Protection of Vertebrate Animals Used for Experimental Purposes (Stras- bourg, 18 March 1986). To reduce the noise, small objects and other artifacts present in the SPECT data, in the first step of the algorithm, a threshold combined with a connected components filtering and morpholog- ical operations (erosion and dilation) was applied to the SPECT data to estimate the skeleton. Due to the variation of the tracer distribution, the extraction of the best possible skeleton requires minimal user input to adjust the threshold and morphological operators parameter settings (more specifically in the extraction of the spine centerline step, see section below). This results in a coarse estimation of the major accumulations of the radioactive tracer: bladder, kidneys, part of the liver and the skeleton. In Figure 3, one can see that due to the differences in nature between the SPECT and CT data, the resultant skeleton in the case of the SPECT data is incomplete, with several parts missing (especially the front limbs, hind limbs and the skull, which is incomplete with large holes). For the CT datasets on the other hand a simple threshold returns the full, complete skeleton (Figure 3). The SPECT datasets were used for testing the proposed approach, and the correspondent mCT datasets to quantitatively validate the performance of the method. Table 1. Resolution of each SPECT and correspondent CT dataset. SPECT CT Resolution (voxel size in mm3) Mouse 1 0.6060.6060.60 0.8060.8060.80 Mouse 2 0.3060.3060.30 0.9960.9961.00 Mouse 3 0.3060.3060.30 0.2060.2060.20 Mouse 4 0.3060.3060.30 0.1060.1060.10 Mouse 5 0.2060.2060.20 0.2060.2060.20 Mouse 6 0.2060.20660.20 0.1060.1060.10 doi:10.1371/journal.pone.0048976.t001 Table 1. Resolution of each SPECT and correspondent CT dataset. Articulated Atlas-based SPECT Skeleton Data Segmentation Method Overview The first step of the proposed approach is to extract the skeleton from the SPECT data. Subsequently, the articulated mouse atlas is registered to the data following a hierarchical anatomical tree: first, the atlas is coarsely registered to the entire skeleton. Then, starting with the skull, each atlas bone is accurately registered to the correspondent bone in the data using the Iterative Closest Point (ICP) approach [14]. After the atlas is registered to the data, we apply the APR algorithm [12,13] to reformat the segmented data into segments corresponding to a mouse atlas and thus mapping the data to a standardized atlas space. The presented method is validated using 6 BALB/c mice, and the quantitative performance of the method is assessed calculating the Euclidian point to surface distance between the atlas and the correspondent November 2012 \| Volume 7 \| Issue 11 \| e48976 PLOS ONE \| www.plosone.org 3 Figure 2. The MOBY mouse atlas skeleton. As originally included in the atlas (top), after segmenting the individual bones (middle), and a detail of the kinematic constraints and the DoFs of the femur/tibia-fibula bone complex (bottom). The colors indicate the different labels of each bone. doi:10.1371/journal.pone.0048976.g002 Mouse SPECT Segmentation Mouse SPECT Segmentation Figure 2. The MOBY mouse atlas skeleton. As originally included in the atlas (top), after segmenting the individual bones (middle), and a detail of the kinematic constraints and the DoFs of the femur/tibia-fibula bone complex (bottom). The colors indicate the different labels of each bone. doi:10.1371/journal.pone.0048976.g002 Articulated Atlas-based SPECT Skeleton Data Segmentation In this step the articulated MOBY atlas is fitted to the skeleton extracted from the SPECT data. For this purpose a modified version of the fully automated approach presented in [1] was used. November 2012 \| Volume 7 \| Issue 11 \| e48976 4 Mouse SPECT Segmentation Figure 3. Examples of SPECT skeleton isosurfaces with the corresponding CT skeleton isosurfaces after the pre-processi figure shows the positioning differences of the mouse in the scanner, SPECT (left) and the correspondent CT (right). The SPECT incomplete, with several parts missing: especially in the case of front, hind limbs and the skull with large holes (blue arrows); also rem relevant objects such as lungs, kidneys and bladder are present (red arrows). In the bottom dataset the right femur and part of the spin (green arrows) due to incomplete acquisition. The CT skeletons are complete and clean after the pre-processing step and are used in the the proposed approach to calculate the Euclidean point to surface distance between the registered atlas and the skeleton surface. indicate examples of regions where over and underestimation of the bone thickness occurred during the skeleton estimation in processing step. d i 10 1371/j l 0048976 003 Mouse SPECT S Figure 3. Examples of SPECT skeleton isosurfaces with the corresponding CT skeleton isosurfaces after the pre-processing step. The figure shows the positioning differences of the mouse in the scanner, SPECT (left) and the correspondent CT (right). The SPECT skeletons are incomplete, with several parts missing: especially in the case of front, hind limbs and the skull with large holes (blue arrows); also remnants of non- relevant objects such as lungs, kidneys and bladder are present (red arrows). In the bottom dataset the right femur and part of the spine are missing (green arrows) due to incomplete acquisition. The CT skeletons are complete and clean after the pre-processing step and are used in the validation of the proposed approach to calculate the Euclidean point to surface distance between the registered atlas and the skeleton surface. Black arrows indicate examples of regions where over and underestimation of the bone thickness occurred during the skeleton estimation in the data pre- processing step. g doi:10.1371/journal.pone.0048976.g003 To deal with the large articulations between bones and/or bone groups, the registration of the atlas is employed following a hierarchical model tree (see Figure 4). Articulated Atlas-based SPECT Skeleton Data Segmentation It is initialized with a coarse alignment of the atlas and the entire target SPECT skeleton, where a similarity transformation model is applied to accommo- date for the animal pose in the scanner and for size differences between animals (7 DoFs are taken into account: three for translation, three for rotation and one for isotropic scaling). After the coarse alignment of the entire skeleton, the individual bones are registered stepwise using the ICP algorithm [14], which is a method for point-based registration (skeleton surfaces in this case). We start at the skull, then, if necessary, the user locates the spine November 2012 \| Volume 7 \| Issue 11 \| e48976 November 2012 \| Volume 7 \| Issue 11 \| e48976 PLOS ONE \| www.plosone.org PLOS ONE \| www.plosone.org 5 Mouse SPECT Segmentation Figure 4. Hierarchical anatomical tree followed during the registration process. * indicates where user input is necessary: to pin-point the spine location where the vertebra connects the spine to the pelvis. doi:10.1371/journal.pone.0048976.g004 Mouse SPECT Segmentation Figure 4. Hierarchical anatomical tree followed during the registration process. * indicates where user input is necessary: to pin-point the spine location where the vertebra connects the spine to the pelvis. doi:10.1371/journal.pone.0048976.g004 location where the vertebra connects the spine to the pelvis, the spinal centerline is extracted using three dimensional region growing until the pelvis is reached and registered and finally proceed to the back limbs, sternum and front limbs separately. The transformation models for the individual bones are dependent on the joint type (ball or hinge) and for each type a realistic motion model was defined ensuring that the bones remain in anatomically realistic shapes. See [1] and Movie S1 for more details. The result is a visualization that consists of a global whole-body view at the top, with a number of focus views of longitudinal, cross-sectional or multimodal data side-by-side at the bottom (Figure 1(h) and Figure 5, Movies S3, S4). This standardized layout facilitates the comparison between subjects, eliminating large differences in animal posture. It allows the user to quickly identify regions/volumes of interest in the global whole-body view and then study the differences or changes in synchronized local per-segment focus views. Quantitative Evaluation Indices (b) - side-by-side visualization of the CT femur bone of 3 subjects, (c) - side-by-side visualization of the SPECT pelvic bone of 3 subjects, (d) - side-by-side visualization of the CT skull data fused with the correspondent SPECT data for 3 subjects, (e) - side-by-side visualization of the skull data of one particular subject: CT, SPECT and a combination of both. Follow-up data visualization was demonstrated in [10] for longitudinal CT mouse data. doi:10.1371/journal.pone.0048976.g005 Figure 5. APR layout of the segmented mouse data. (a) - global articulated planar reformatted visualization of the atlas. (b), (c), (d) and (e) show the different data visualization options after applying the proposed approach. One can choose to visualize simultaneously and side-by-side particular region of interest in cross-sectional studies for CT, SPECT or the combination of both. (b) - side-by-side visualization of the CT femur bone o 3 subjects, (c) - side-by-side visualization of the SPECT pelvic bone of 3 subjects, (d) - side-by-side visualization of the CT skull data fused with th correspondent SPECT data for 3 subjects, (e) - side-by-side visualization of the skull data of one particular subject: CT, SPECT and a combination o Figure 5. APR layout of the segmented mouse data. (a) - global articulated planar reformatted visualization of the atlas. (b), (c), (d) and (e) show the different data visualization options after applying the proposed approach. One can choose to visualize simultaneously and side-by-side a particular region of interest in cross-sectional studies for CT, SPECT or the combination of both. (b) - side-by-side visualization of the CT femur bone of 3 subjects, (c) - side-by-side visualization of the SPECT pelvic bone of 3 subjects, (d) - side-by-side visualization of the CT skull data fused with the correspondent SPECT data for 3 subjects, (e) - side-by-side visualization of the skull data of one particular subject: CT, SPECT and a combination of both. Follow-up data visualization was demonstrated in [10] for longitudinal CT mouse data. doi:10.1371/journal.pone.0048976.g005 point to surface distance between the estimated SPECT and its correspondent CT skeletons (Table 2). positioning of the mouse in the scanner: the hind and front limbs were pulled towards the belly resulting in a lower alignment/ overlay between the atlas and the data surfaces during the coarse alignment. Quantitative Evaluation Indices The fitted atlas yields a completely segmented SPECT skeleton where each bone has its own unique label. In this step we use the Articulated Planar Reformation (APR) algorithm [12, 13, Movies S2, S3, S4], that uses the segmented bones and the correspondent transformation models to reformat the data into segments corresponding to the atlas and thus maps the data to a standardized atlas space. To quantitatively validate the registration accuracy and enable comparison with the registration error achieved in mCT data as reported in [1], we used the same error metric to evaluate the SPECT segmentation: the mean Euclidean point to surface distance in voxels, i.e., the shortest distance between objects in space. For each SPECT dataset, we calculated this distance between the registered MOBY atlas skeleton and the correspon- dent co-registered CT skeleton, before (i.e.: after the coarse alignment) and after articulated registration. The results were compared to the mean Euclidean point to surface distance published in [1] for mCT. For each bone in the atlas, based on the surface representation of the bone and the corresponding linear transform that was determined during registration a bounding box is automatically determined. Using the bounding box, the volume SPECT data is resampled for each bone with the aim of obtaining the volume in a standard coordinate frame, which facilitates comparison (Figures 1(f, g), Movies S2, S3, S4). To investigate quantitatively intrinsic differences between SPECT and CT (Figure 3), we calculated the mean Euclidean November 2012 \| Volume 7 \| Issue 11 \| e48976 PLOS ONE \| www.plosone.org November 2012 \| Volume 7 \| Issue 11 \| e48976 6 Mouse SPECT Segmentation point to surface distance between the estimated SPECT and its positioning of the mouse in the scanner: the hind and front limbs Figure 5. APR layout of the segmented mouse data. (a) - global articulated planar reformatted visualization of the atlas. (b), (c), (d) and (e) show the different data visualization options after applying the proposed approach. One can choose to visualize simultaneously and side-by-side a particular region of interest in cross-sectional studies for CT, SPECT or the combination of both. Quantitative Evaluation Indices The second reason is the fact that the SPECT skeleton after the pre-processing step in case of mouse 4 is underestimated (only remains of the limbs are visible, and a very small portion of the skull and the spine are present). Since the amount of total bone content of the skeleton has an influence on the coarse alignment step [1], in this case, during the coarse alignment, the atlas successfully accommodates for the animal position (prone/supine) and orientation of the animal, but the main overlap between the atlas and the animal after this coarse step happens between the skulls. Figure 3 shows three data examples with variations in November 2012 \| Volume 7 \| Issue 11 \| e48976 Results Table 3 presents the Euclidian point to surface distance before and after articulated registration for all the subjects. For all the mice, after the registration the Euclidean point to surface distance between the MOBY atlas and the SPECT skeleton decreased: the calculated mean of the distance decreased from 11.5612.1 to 2.662.1 voxels. In case of mouse 4 (Figure 3, bottom mouse), the Euclidean point to surface distance before registration is higher than for the other subjects for two reasons. The first reason is the November 2012 \| Volume 7 \| Issue 11 \| e48976 November 2012 \| Volume 7 \| Issue 11 \| e48976 PLOS ONE \| www.plosone.org 7 Mouse SPECT Segmentation Table 2. Mean Euclidian point to surface distance between the SPECT and CT skeletons after the pre-processing step. Mean Euclidian point to surface distance between the SPECT and correspondent CT skeletons and standard deviation (in voxels) Mouse 1 6.4611.1 Mouse 2 5.967.9 Mouse 3 4.565.7 Mouse 4 2.764.1 Mouse 5 9.6613.4 Mouse 6 7.768.1 Mean 6.168.4 doi:10.1371/journal.pone.0048976.t002 Mean Euclidian point to surface distance between the SPECT and correspondent CT skeletons and standard deviation (in voxels) Mean Euclidian point to surface distance between the SPECT and correspondent CT skeletons and standard deviation (in voxels) doi:10.1371/journal.pone.0048976.t002 posture with which the proposed method successfully coped. An example of a segmented SPECT dataset is presented in Figure 6 and Movie S1. articulated atlas-based skeleton segmentation approach combined with the articulated planar reformation algorithm for whole-body mouse bone imaging using SPECT. After applying the APR algorithm to both mCT and SPECT data, one can use a range of visualization techniques that enable exploration of both datasets, the result of which is depicted in Figure 5 and Movies S3, S4. The articulated layout visualization is shown, where all segments of the atlas have been spread out into a plane - Figure 5(a), Movies S3, S4. In Figure 5 and Movies S3, S4 the different visualization options are generated using the proposed approach. The femur, pelvis and skull were selected and are shown in the correspondent focus views. Results All focus views show an outline and an image slice visualization and one can visualize CT, SPECT, or a combination of both (where the SPECT data is shown as an overlay with a color map applied to it) in either cross-sectional analysis - Figures 5(b, c, d), Movie S3 or for multi-modality complementarity - Figure 5(e), Movie S4. Quantitative evaluation was performed by calculating the Euclidian point to surface distance between the registered atlas and the correspondent CT dataset. The obtained mean distance of 2.662.2 voxels, showed that the registration accuracy for the SPECT data is of the same order as the previously published results for mCT, 1.860.1 voxels [1]. The large difference in the standard deviation between the mCT fittings results and the ones presented in this paper might be due to the variable nature of the SPECT data (tracer uptake and distribution, where the tracer targets the bone growth and not the entire bone and partial volume effect) versus the more robust bone contrast in CT. Due to these factors the pre-processing step (extraction of the skeleton out of the data while removing the noise, small objects and other artifacts) may result in either a partial or a much thinner (Figure 3) or thicker skeleton than as seen in the mCT. This explains the difference in distance measures between the SPECT and CT skeleton surfaces (Table 2). As mentioned above, by collecting data from several different imaging studies, one of the goals of this study was to demonstrate the robustness of the atlas-based segmentation with respect to the different whole-body SPECT scan settings. Depending on the research question, the amount of injected tracer, the pinhole size and scan time a trade-off has to be chosen between resolution and signal. However, as long as a skeleton estimation is possible the approach presented here holds. The entire articulated registration process was implemented in MATLAB R2008b TM and took approximately 2 minutes of runtime (including minimal user interventions to assist the spine centerline extraction) on a standard desktop PC (2.40 GHz Intel Quad Core TM with 3 GB of RAM, Windows TM). Discussion and Conclusion In vivo visualization of functional and anatomical information produces heterogeneous, high throughput data. Efficiently com- bining, analyzing and quantifying whole-body small animal cross- sectional, longitudinal and multi-modal data is a complex problem. In this paper, we demonstrated the feasibility of the In [1], it was demonstrated that the proposed atlas-based segmentation method is robust with respect to osteolytic bone Table 3. Quantitative results of the MOBY atlas-to-skeleton registration for 6 mouse SPECT datasets and [1]. Table 3. Quantitative results of the MOBY atlas-to-skeleton registration for 6 mouse SPECT datasets an Quantitative results of the MOBY atlas-to-skeleton registration for 6 mouse SPECT datasets and [1]. Table 3. Quantitative results of the MOBY atlas-to-skeleton registration for 6 mouse SPECT datasets and [1]. Mean Euclidian point to surface distance and standard deviation (in voxels) Before registration After registration Mouse 1 10.3610.1 2.462.4 Mouse 2 6.667.8 2.562.1 Mouse 3 8.268.2 2.061.7 Mouse 4 25.8632.1 3.362.8 Mouse 5 10.468.7 2.661.9 Mouse 6 7.565.9 2.962.0 Mean 11.5612.1 2.662.1 Baiker et al. [1] (mCT) 8.861.9 1.860.1 doi:10.1371/journal.pone.0048976.t003 PLOS ONE \| www.plosone.org 8 November 2012 \| Volume 7 \| Issue 11 \| e48976 November 2012 \| Volume 7 \| Issue 11 \| e48976 Figure 6. Top and side views of the segmented SPECT skeleton initially presented in Figure 1a). The registered MOBY atlas is represented in green. doi:10.1371/journal.pone.0048976.g006 Mouse SPECT Segmentation Mouse SPECT Segmentation Figure 6. Top and side views of the segmented SPECT skeleton initially presented in Figure 1a). The registered MOBY atlas is represented in green. doi:10.1371/journal.pone.0048976.g006 defects. Here, it was demonstrated, that the use of the articulated mouse atlas, with defined DoFs and size restrictions for each bone, proved to be robust for ‘‘incomplete’’ data (i.e.: images where large bits of limbs are missing), like exemplified in Figures 3 (bottom mouse) and 6. If a lower or an upper part of a limb is completely missing, than the proposed approach will only segment the part that is present in the dataset, i.e., the part where there was significant/enough tracer uptake. evaluation of the method demonstrated it to be sufficiently accurate and robust for intuitive exploration of whole-body, cross- sectional multi-modal small animal imaging data. The approach presented here can be applied to other animals, provided there is an adequate atlas. (WMV) Movie S2 APR algorithm applied to mouse data. Demonstration of the APR algorithm as described in the section APR of combined SPECT/CT mouse data. (WMV) Movie S3 Side-by-side SPECT data exploration of 3 mice. Exploration and visualization of cross-sectional SPECT mouse data of three different subjects after applying the APR algorithm as described in the APR of combined SPECT/CT mouse data and Results sections. (WMV) algorithm as described in the APR of combined SPECT/CT mouse data and Results sections. (WMV) The segmentation approach presented here was developed to cope with a scenario when a combined whole-body SPECT/CT bone scan is not always desired or available. Thus, one of the limitations of the proposed approach is the fact that the skeleton should exhibit sufficient image contrast, i.e., direct application of the atlas fitting to SPECT data requires tracer uptake in the skeleton. When that is not the case, the limitation can be overcome by applying the fitting directly to the provided whole-body anatomical CT scan and then propagating it to the SPECT data. Furthermore, though very minimal, this method requires user input during the extraction of the possible skeleton out of the data. This only stands true when the method is applied to SPECT directly and correspondent CT data is not available. When CT whole-body data is available, due to its robust and consistent nature, this kind of user input is not required anymore, as shown and extensively validated in [1]. Movie S4 Side-by-side SPECT/CT mouse data explora- tion. Exploration and visualization of multi-modal, complemen- tary SPECT/CT and the fusion of both mouse data after applying the APR algorithm as described in the APR of combined SPECT/CT mouse data and Results sections. (WMV) Movie S4 Side-by-side SPECT/CT mouse data explora- tion. Exploration and visualization of multi-modal, complemen- tary SPECT/CT and the fusion of both mouse data after applying the APR algorithm as described in the APR of combined the APR algorithm as described in the APR of combined SPECT/CT mouse data and Results sections. (WMV) Supporting Information Movie S1 Atlas-based SPECT skeleton segmentation. Demonstration of the articulated atlas-based small animal SPECT skeleton segmentation algorithm as described in the section Articulated atlas-based SPECT skeleton data segmenta- tion. It also proved to be relatively insensitive to non-relevant objects still present in the image after threshold-based segmentation, like kidneys, bladder, some lung and liver. The proposed approach effectively compensated for the large variations in posture that existed within the data and yielded segmentation results requiring minimal user input. These were of satisfactory quality for the ensuing mapping of the data to the standard reference and side- by-side visualization. Applying the APR algorithm to multi-modal cross-sectional data proved to be useful to provide proper referencing and visualization for an intuitive exploration and comparison of mCT, SPECT data (Figure 5, Movies S3, S4). The authors are currently working on further extending the approach presented here to combine automatic segmentation of the different bones with tracer quantification. References 1. Baiker M, Milles J, Dijkstra J, Henning TD, Weber AW, et al. (2010) Atlas-based whole-body segmentation of mice from low-contrast micro-CT data. Med Image Anal 14(6): 723–737. 10. Ma´the´ D, Balogh L, Polya´k A, Kira´ly R, Ma´ria´n T, et al. (2010) Multispecies animal investigation on biodistribution, pharmacokinetics and toxicity of 177Lu- EDTMP, a potential bone pain palliation agent. Nucl Med Biol 37(2): 215–26. ( ) 2. Khmelinskii A, Baiker M, Kok P, de Swart J, Reiber JHC, et al. (2011) Atlas- based articulated skeleton segmentation of mSPECT mouse data. Proc IEEE Intl Symp Biomed Imaging 437–440. p p p g ( ) 11. Franc BL, Acton PD, Mari C, Hasegawa BH (2008) Small-animal SPECT and 11. Franc BL, Acton PD, Mari C, Hasegawa BH (2008) Small-animal SPECT and SPECT/CT: Important tools for preclinical investigation J Nucl Med 49(10): 11. Franc BL, Acton PD, Mari C, Hasegawa BH (2008) Small-animal SPECT and SPECT/CT: Important tools for preclinical investigation. J Nucl Med 49(10): 1651–1663. , , , g ( ) SPECT/CT: Important tools for preclinical investigation. J Nucl Med 49(10): 1651–1663. y p g g 3. Khmelinskii A, Baiker M, Chen XJ, Reiber JHC, Henkelman RM, et al. (2011) Articulated whole-body atlases for small animal image analysis: Construction and applications. Mol Imaging Biol 13(5): 898–910. 12. Kok P, Baiker M, Hendriks E, Post FH, Dijkstra J, et al. (2010) Articulated planar reformation for change visualization in small animal imaging. IEEE T Vis Comput Gr 16(6): 1396–1404. 4. Joshi AA, Chaudhari AJ, Li C, Shattuck DW, Dutta J, et al. (2009) Posture matching and elastic registration of a mouse atlas to surface topography range data. Proc IEEE Intl Symp Biomed Imaging 366–269. 13. CVP website. Available: http://graphics.tudelft.nl/, pkok/CVP/. Accessed 2012 Oct 13. 14. Besl PJ, McKay ND (1992) A method for registration of 3D shapes. IEEE T Pattern Anal 14(2): 239–256. 5. Savinaud M, de La Gorce M, Maitrejean S, Paragios N (2010) Model-based multi-view fusion of cinematic flow and optical imaging. Med Image Comput Comput Assist Interv 13(Pt2): 668–675. 15. Segars WP, Tsui BMW, Frey EC, Johnson GA, Berr SS (2004) Development of a 4-D digital mouse phantom for molecular imaging research. Mol Imaging Biol 6(3): 149–159. p ( ) 6. Wildeman MH, Baiker M, Reiber JHC, Lo¨wik C, Reinders MJT, et al. (2009) 2D/3D registration of micro-CT data to multi-view photographs based on a 3D distance map. Author Contributions Conceived and designed the experiments: AK MB MDJ BPFL. Performed the experiments: AK MB HCG. Analyzed the data: AK MB HCG. Contributed reagents/materials/analysis tools: AK MB HCG. Wrote the paper: AK. Performed the experiments with animals and acquired the data: HCG. Developed the registration/segmentation tool: AK MB. Discussed the results and commented on the manuscript: AK MB HCG MDJ BPFL. Conceived and designed the experiments: AK MB MDJ BPFL. Performed the experiments: AK MB HCG. Analyzed the data: AK MB HCG. Contributed reagents/materials/analysis tools: AK MB HCG. Wrote the paper: AK. Performed the experiments with animals and acquired the data: HCG. Developed the registration/segmentation tool: AK MB. Discussed the results and commented on the manuscript: AK MB HCG MDJ BPFL. An articulated atlas-base skeleton segmentation method for SPECT whole-body small animal data was presented. The PLOS ONE \| www.plosone.org PLOS ONE \| www.plosone.org November 2012 \| Volume 7 \| Issue 11 \| e48976 9 Mouse SPECT Segmentation Mouse SPECT Segmentation Mouse SPECT Segmentation References Proc IEEE Intl Symp Biomed Imaging 987–990. 16. LKEB website. Articulated Atlases Download section. Available: http://www. lkeb.nl. Accessed 2012 Oct 13. 7. Suh JW, Scheinost D, Dione DP, Dobrucki LW, Sinusas AJ, et al. (2011) A non- rigid registration method for serial lower extremity hybrid SPECT/CT imaging. Med Image Anal 15(1): 96–111. 17. Amira website. Available: http://www.amira.com. Accessed 2012 Oct 13. 18. Bab I, Hajbi-Yonissi C, Gabet Y, Mu¨ller R (200 18. Bab I, Hajbi-Yonissi C, Gabet Y, Mu¨ller R (2007) Micro-Tomographic Atlas of the Mouse Skeleton, Springer. 18. Bab I, Hajbi-Yonissi C, Gabet Y the Mouse Skeleton, Springer. j the Mouse Skeleton, Springer. 8. Hindorf C, Rodrigues J, Boutaleb S, Rosseau J, Govignon N, et al. (2010) Total absorbed dose to a mouse during microPET/CT imaging. Eur J Nucl Med Mol Imaging 37: S274. k MJ (1965) Anatomy of the laboratory mouse New York, Academic. 19. Cook MJ (1965) Anatomy of the laboratory mouse New Yo 99 19. Cook MJ (1965) Anatomy of the labo 20. 99mTc-MDP website. MICAD - Molecular Imaging Probes and Contrast Agents Dataset. Available: http://www.ncbi.nlm.nih.gov/books/NBK24575/. Ac- cessed 2012 Oct 13. 9. Ostendorf B, Scherer A, Wirrwar A, Hoppin JW, Lackas C, et al. (2006) High- resolution multipinhole single-photon-emission computed tomography in experimental and human arthritis. Arthritis Rheum 54(4): 1096–104. PLOS ONE \| www.plosone.org November 2012 \| Volume 7 \| Issue 11 \| e48976 PLOS ONE \| www.plosone.org 10
https://openalex.org/W2035139011	https://europepmc.org/articles/pmc4110024?pdf=render	English	null	Kathu Townlands: A High Density Earlier Stone Age Locality in the Interior of South Africa	PloS one	2,014	cc-by	6,665	Steven J. H. Walker1, Vasa Lukich2, Michael Chazan3* 1 Department of Archaeology, University of Cape Town, Rondebosch, South Africa, 2 Department of Earth Sciences, University of Toronto, Toronto, Ontario, Canada, 3 Department of Anthropology, University of Toronto, Toronto, Ontario, Canada Introduction Early to Middle Pleistocene localities that incorporate extremely high numbers of lithic artefacts present a challenge to our understanding of early hominin behaviour. In some cases, such as Ma’ayan Baruch, Israel [1], the deposits are dominated by very large numbers of bifaces while other localities show characteristics consistent with quarrying or primary production [2,3,4,5]. There are also localities with a more ambiguous character such as the site of Canteen Kopje, South Africa [6,7] which includes high densities of both production debris and finished artefacts. Both the quarry sites and sites dominated by finished artefacts provide evidence of stone tool transport but these sites also raise questions about group size and organization of activity among early hominin groups. Here we present the first detailed description of a site located in the town of Kathu, Northern Cape Province, South Africa that shows high intensity of lithic production during the Earlier Stone Age (ESA). Kathu Townlands is a site situated between the Kuruman Hills to the east and the Langberge mountains to the west on a low hill and is covered with a dense surface of lithics interspersed with exposures of bedrock, calcrete, and sand (Figure 1–2). National Heritage site in 2013 however the threat to deposits beyond the declared area remain acute. National Heritage site in 2013 however the threat to deposits beyond the declared area remain acute. Abstract * Email: mchazan@chass.utoronto.ca National Heritage site in 2013 however the threat to deposits beyond the declared area remain acute. Abstract Kathu Townlands is a high density Earlier Stone Age locality in the Northern Cape Province, South Africa. Here we present the first detailed information on this locality based on analysis of a sample of lithic material from excavations by P. Beaumont and field observations made in the course of fieldwork in 2013. The results confirm the remarkably high artefact density at Kathu Townlands and do not provide evidence consistent with high energy transport as a mechanism of site formation, suggesting that Kathu Townlands was the site of intensive exploitation of highly siliceous outcroppings of banded iron formation. The results presented here provide a first step towards understanding this complex locality and point to the need for further research and the importance of preserving this locality in the face of intensive and rapid development. Citation: Walker SJH, Lukich V, Chazan M (2014) Kathu Townlands: A High Density Earlier Stone Age Locality in the Interior of South Africa. PLoS ONE 9(7): e103436. doi:10.1371/journal.pone.0103436 Editor: Michael D. Petraglia, University of Oxford, United Kingdom Editor: Michael D. Petraglia, University of Oxford, United Kingdom Received April 9, 2014; Accepted July 2, 2014; Published July 24, 2014 Received April 9, 2014; Accepted July 2, 2014; Published July 24, 2014 Copyright: 2014 Walker et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability: The authors confirm that all data underlying the findings are fully available without restriction. McGregor Museum, Kimberley is the repository for all collections described here as well as our data (inventory number 6512). Analysis can be arranged by permission from the McGregor Museum. Funding: PZK Beleggings, under the direction of Jaco Minnie of Leon Frank and Partners, funded field work. Support was also provided by Canadian Social Sciences and Humanities Research Council Grant 410-2010-722. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The 2013 field work was carried out under the auspices of the Archaeology Department, McGregor Museum, for PZK Beleggings, under the direction of Jaco Minnie of Leon Frank and Partners. Although some of this project was funded commercially, this does not alter the authors’ adherence to PLOS ONE policies on sharing data and materials. July 2014 \| Volume 9 \| Issue 7 \| e103436 Archaeological Setting g g Kathu Townlands is a component of a grouping of ESA localities designated as the Kathu Complex. This complex also includes the excavated sites of Kathu Pan1 (KP1) and Bestwood 1 (BW 1, Figure 1c). At Kathu Pan, evidence of early hominin occupation has been observed at multiple locations within the pan, but ESA deposits have only been excavated at KP 1 [13]. Stratum 4a at KP1 is dated by a combination of OSL and ESR/U-series to ca. 500 k BP [13]. The lithic assemblage from St. 4a is characterized by a prepared core technology that produced both blades and points, and has been attributed to the Fauresmith industry [14,15]. The lithic assemblage of the underlying St. 4b at Kathu Pan 1 is characterized by well-made handaxes. At BW 1, located to the east of Kathu Townlands in a valley between two small hills, mining of sand has revealed a horizon at the interface of gravels and the overlying sands that contains abundant lithic artefacts [16]. These are characterized by bifaces, blades, and prepared cores and are dispersed over a very large area. Excavation at BW1 in 2012 exposed a surface of 36 m2 with over 1000 piece plotted artefacts recovered. All artefacts were observed lying flat on at the interface between the sands and gravels. The extremely fresh condition of the lithic artefacts at Bestwood 1 argues against this accumulation being a palimpsest deposited over a long period, or as the result of deflation. The possibility that the concentration of artefacts in a single horizon is the result of Calcretes are a common feature in the area around Kathu and figure significantly in the deposits in the vicinity of Kathu Townlands. Calcretes develop in arid or semi-arid environments as the result of lateral and vertical movement of carbonates in solution but the diagenetic process causing the development of these deposits can vary depending on local conditions [10]. At the Mamatwan Mine near Hotazel, a calcrete horizon approximately 2 meters thick produced optically stimulated luminescence ages of 113,000 and 108,000 years ago [11]. Without detailed analysis it is not possible to determine the age of the calcrete deposits at Kathu Townlands. A sand sheet, derived from the Kalahari, is found across the surface at Kathu Townlands and comprises the matrix within which most of the artefacts are found. Kathu Townlands Earlier Stone Age Locality similar deposits are known from other localities in the region, although no such deposits are known at Kathu Townlands. similar deposits are known from other localities in the region, although no such deposits are known at Kathu Townlands. Geological Setting g g The bedrock lithology is Precambrian, with exposures of banded iron formation (BIF), which belong to the Kuruman Formation within the late Archean to earliest Paleoproterozoic Transvaal Supergroup [8,9]. There is a wide degree of variability within the BIF in the Kuruman Formation both in terms of the scale of banding and the percentage of chert (SiO2) relative to iron-rich minerals. The outcroppings at Kathu Townlands (sometimes designated as jaspellite) are dominated by chert and show no fine-scale banding. As a result the structure of these rocks is ideal for stone tool manufacture and it is likely that the availability of high quality raw material is a major reason for repeated exploitation of the resource and the high density of stone tool and knapping debris at Kathu Townlands. Circular white to grey fossil traces are characteristic of the raw material found at Kathu Townlands. Outcroppings of raw material with similar fossil traces have not been identified in the surrounding region. The Kuruman Hills are today drained by a series of ephemeral streams that flow northwest (Figure 1). None of these streams pass through the research area, the closest drainage is the Vermul- sleegte, to the north of the site. There is evidence for far more substantial drainage systems at some point in the geological past in the area around Kathu. At the Bestwood site the archaeological horizon is underlain by at least ten meters of river gravels and Dense and broadly distributed archaeological deposits pose methodological and management challenges. The town of Kathu is rapidly expanding and this development is directly threatening Kathu Townlands (Figure 1c). The site was designated a Grade 1 PLOS ONE \| www.plosone.org PLOS ONE \| www.plosone.org July 2014 \| Volume 9 \| Issue 7 \| e103436 1 Kathu Townlands Earlier Stone Age Locality 2 July 2014 \| Volume 9 \| Issue 7 \| e103436 Previous Work at Kathu Townlands The archaeological deposit at Kathu Townlands was brought to the attention of archaeologists in 1980 by Naas Viljoen, the manager of the property [20,21]. The site was first described in a permit report submitted to the National Monuments Council outlining the results of a large scale survey for archaeological resources in the region [22]. Shortly afterwards, the outcropping of ironstone (and the tremendous amount of artefacts) was used as a source of road gravel. Mr Viljoen notified Mr Beaumont (who was excavating at KP1 at the time) that he had observed workmen using gravel that was composed primarily of artefacts to repair roads [20]. Excavations at the site were conducted by Beaumont in 1982, and then again in 1990 [20,21]. Unfortunately, the precise location of these excavations was not reported. Archaeological Setting Image dated 5/27/2011 predates recent development south of Frikkie Meyer St. Image source Terraserver. C. View of Towlands site north of Frikkie Meyer St. D. Detail showing scatter of artefacts on the surface of the site north of Frikkie Meyer St. doi:10.1371/journal.pone.0103436.g002 Frikkie Meyer Street. Since at least 1990 the deposit has been known to exist south of this road. The total extent of the deposit, particularly beneath the surface sands, remains to be determined. bioturbation also seems unlikely [17]. The excavated area is apparently representative of an extensive occupation covering several hectares. Kathu Townlands is located approximate 56 km. west of Wonderwerk Cave where a sequence of ESA occupation has produced early evidence of fire [18,19]. At Wonderwerk the ESA occupation is characterized by low artefact density in marked contrast to Kathu Townlands. Archaeological Setting As with calcretes, there were likely multiple cycles of sand accumulation in the Kathu area. Research at Wonderwerk Cave has demonstrated that Kalahari sands were blowing into this region by 2 million years ago [12] and it is likely that the sands at the base of Kathu Pan 1 are of such an early age. At the Bestwood 1 site, sands overlie the archaeological horizon and are thus of a younger age. At the Mamatwan Mine the Kalahari sands produced optically stimulated luminescence ages ranging from 62–44,000 years ago. e 1. Location of Kathu Townlands. A. Location map within South Africa. B. Location Map with relation to the regional topog erwerk Cave. C. Topographic context of sites of the Kathu Complex discussed in this article. Grey shading indicates developed area going development. Note that the boundaries of Kathu Pan are approximate and do not indicate the limit of areas of archaeologica 1371/journal.pone.0103436.g001 Figure 1. Location of Kathu Townlands. A. Location map within South Africa. B. Location Map with relation to the regional topography and Wonderwerk Cave. C. Topographic context of sites of the Kathu Complex discussed in this article. Grey shading indicates developed areas and areas undergoing development. Note that the boundaries of Kathu Pan are approximate and do not indicate the limit of areas of archaeological potential. doi:10.1371/journal.pone.0103436.g001 July 2014 \| Volume 9 \| Issue 7 \| e103436 PLOS ONE \| www.plosone.org PLOS ONE \| www.plosone.org 2 Kathu Townlands Earlier Stone Age Locality Figure 2. Views of Kathu Townlads. A. Digital Elevation Model (DEM) showing the topographic context of Kathu Townlands. Square shows the approximate area shown in Figure 2b. (DEM courtesy of Stephen Wessels, The Zimani Project). B. Aerial photo of Kathu Townlands. White shading indicates approximate limits of the declared locality. Asterisk indicates location of 2013 excavations. Image dated 5/27/2011 predates recent development south of Frikkie Meyer St. Image source Terraserver. C. View of Towlands site north of Frikkie Meyer St. D. Detail showing scatter of artefacts on the surface of the site north of Frikkie Meyer St. doi:10.1371/journal.pone.0103436.g002 Figure 2. Views of Kathu Townlads. A. Digital Elevation Model (DEM) showing the topographic context of Kathu Townlands. Square shows the approximate area shown in Figure 2b. (DEM courtesy of Stephen Wessels, The Zimani Project). B. Aerial photo of Kathu Townlands. White shading indicates approximate limits of the declared locality. Asterisk indicates location of 2013 excavations. Methods Field work was conducted under the South African Heritage Resources Agency permit #577 (CaseID: 2992). McGregor Museum, Kimberley is the repository for all collections described here as well as our data (inventory number 6512). Analysis can be arranged by permission from the McGregor Museum. July 2014 \| Volume 9 \| Issue 7 \| e103436 Lithic Analysis The complete collection from one ten centimetre deep spit from Beaumont’s 1 m by 1 m excavation (Square K6 20–30 cm.) was chosen for complete analysis to provide a baseline for future studies. The McGregor Museum, Kimberley is the repository for all collections (inventory number 6512) and future analysis of these collections can be arranged by permission from the McGregor Museum. A sample of 20 handaxes from Beaumont’s surface collections was added to allow for analysis of handaxe morphol- ogy. The size of the assemblage from this context is representative of all excavated contexts below covering sands at Kathu Town- lands. Measurements taken on flakes were limited to maximum dimension, degree of fragmentation (proximal, distal, medial, complete), condition (fresh, edge damage, abraded), raw material, There has also been a general lack of clarity regarding of the extent of the deposit. The initial reports identified an area north of PLOS ONE \| www.plosone.org July 2014 \| Volume 9 \| Issue 7 \| e103436 3 Kathu Townlands Earlier Stone Age Locality 1. Breakdown of the sample of flakes and flake fragments from the Beaumont excavation (n = 1283). Table 1. Breakdown of the sample of flakes and flake fragments from the Beaumont excavation (n = 1283). Fragment n % Complete 714 55.65% Proximal 223 17.38% Distal 191 14.89% Medial 155 12.08% Condition n % Fresh 297 23.15% Edge damage 879 68.51% Abraded 107 8.34% Type n % Primary 72 5.61% Secondary 222 17.30% Non-cortical 989 77.08% doi:10.1371/journal.pone.0103436.t001 doi:10.1371/journal.pone.0103436.t001 weight, and flake type (primary: cortex .80%, secondary: cortex ,80%, non-cortical). For cores, measurements taken were limited to description and weight. For bifaces, B1 (breadth 1/5 from the tip) and B2 (breadth 1/5 from the base) were measured along with length, maximum thickness, and weight. A subsample was selected for photography. The sample from this single spit includes 1283 flakes, 10 cores, and 6 bifaces (See Table 1). July 2014 \| Volume 9 \| Issue 7 \| e103436 2013 Excavations In August 2013 excavations were undertaken as part of a data recovery project to mitigate the destruction being caused by a mall being built on a small portion of the known deposit [23] (see Figure 2b, 3). The primary objective of the project was to recover samples from a controlled stratigraphic context distributed across the area slated for destruction in the limited time available. We aimed to obtain data useful for understanding both the horizontal Figure 3. Plan of trenches and excavation units from the 2013 excavations. The topographic map was produced by a surveyor for the mall construction before any surface modifications were made to the project area. Note that the topography dips towards the west showing that the site occurs along a N-S ridge. This ridge is visible in the DEM in Figure 2a. doi:10.1371/journal.pone.0103436.g003 Figure 3. Plan of trenches and excavation units from the 2013 excavations. The topographic map was produced by a surveyor for the mall construction before any surface modifications were made to the project area. Note that the topography dips towards the west showing that the site occurs along a N-S ridge. This ridge is visible in the DEM in Figure 2a. doi:10.1371/journal.pone.0103436.g003 July 2014 \| Volume 9 \| Issue 7 \| e103436 July 2014 \| Volume 9 \| Issue 7 \| e103436 PLOS ONE \| www.plosone.org 4 Kathu Townlands Earlier Stone Age Locality Table 2. Weight of lithic artefacts recovered in 2013 excavation by excavation unit broken down by arbitrary 10 cm. levels. Table 2. Weight of lithic artefacts recovered in 2013 excavation by excavation unit broken down by a Table 2. Weight of lithic artefacts recovered in 2013 excavation by excavation unit broken down by arbitrary 10 cm. levels. Trench A (square 1) Trench C (square 2) Trench E (square 3) Trench G (Square 4) Trench I (Square 5) Opening Elevation 1232.9 m. asl 1233.1 m. asl 1232.1 m. asl 1233.3 m. asl 1231.9 m. asl Cm. 2013 Excavations Below Surface Weight of artifacts (kg.) 0–10 13.4 0.1 0.1 0.1 10–20 20.4 20.2* 0.1 0.2 0.2 20–30 17.2 17.4 0.6 4.4 0.2 30–40 21 17 0.6 15.2 0.1 40–50 27.2 11.4 3.6 8.2 4.8 50–60 14.8 12.6 18.4 4.4 9.4 60–70 14.8 10.4 4 7.4 7.2 70–80 22.8 15.2 7.6 3.6 0.8 80–90 14.4 4.2 5.8 5.2 0.6 90–100 22 BEDROCK 2.8 10.8 0.1 100–110 9.4 3.2 7.4 BEDROCK 110–120 13.4 0.1 2.4 12– = 130 13.8 BEDROCK 0.1 130–140 7 BEDROCK 140–150 5.8 150–160 12 160–170 16.8 170–180 15.8 BEDROCK Closing Elevation 1231.1 1232.2 1231.1 1232.0 1230.9 Area Excavated 100650 cm. 100650 cm. 100650 cm. 50650 cm. 50650 cm. *Square 2: Levels 1 & 2 were dug as a single spit. doi:10.1371/journal.pone.0103436.t002 artefacts collected by level. All excavated material is curated in the McGregor Museum, Kimberley. and vertical distribution of the deposits over as wide an area as possible. The methods employed combined mechanical trenching to obtain data regarding the geomorphology, followed by hand excavations to obtain archaeological samples (Figure 3). Horizon- tal control across the project was maintained with averaged GPS waypoints. Vertical associations were able to be correlated using a high resolution topographic map of the project area (20 cm intervals; Figure 3). July 2014 \| Volume 9 \| Issue 7 \| e103436 Lithic Analysis The breakdown of the analyzed lithic assemblage is presented in Table 1. The flakes include 714 (55.65%) complete pieces, 223 (17.38%) proximal fragments, 191 distal fragments (14.89%), and 155 medial fragments (12.08%) (Figure 4). Only 297 pieces are fresh (23.15%, sharp edges with little microfracture), 879 have edge damage (68.51%, considerable microfracture, crushing, and/ or large scar damage along the edge), and 107 pieces are abraded (8.34%, smoothing due to abrasion, with or without edge damage). Maximum dimension for the large flakes has a mean of 4.6 cm (SD 2.28), and the mean for weight is 31.6 g (SD 55.9). The sample from the Beaumont excavation shows high fragmentation and a high degree of edge damage indicating significant mechanical damage. Flakes smaller than 2-cm. in maximum dimension are underrepresented, possibly due to the mesh size employed in sieving. However, there is no evidence for winnowing of flakes greater than 2 cm. as would be expected under high energy transport. Abrasion, although present, is not frequent. Under archaeological supervision, ten trenches were mechan- ically excavated across the property. The location of trenches was chosen with the intent of providing a cross section of the topography of the impacted area. Each of the ten trenches was cleaned to bedrock by hand and profiles were cleaned, measured and photographed. Additionally, a long trench dug by the developer was documented in an area that had been considered to be outside the site boundary. From these ten trenches, five were selected for further archaeological excavation to provide a horizontally distributed sample from trenches that were archaeologically rich. The first three units were 100 cm650 cm, however, due to the overwhelm- ing amount of material being recovered, the last two were reduced to 50650 cm (see Table 2). These units were dug in 10 cm levels as measured from modern ground surface (MGS). Vertical control was maintained utilizing a datum and line level. Each unit was dug to bedrock, all material was screened through a 3 mm sieve, and The representation of all size classes in the flake assemblage suggests that knapping took place on site and is consistent with the identification of the site as a quarry locality (Figure 5: a–b). Lithic Analysis However, the low frequency of primary flakes (n = 72, 5.61%) and PLOS ONE \| www.plosone.org July 2014 \| Volume 9 \| Issue 7 \| e103436 July 2014 \| Volume 9 \| Issue 7 \| e103436 5 Kathu Townlands Earlier Stone Age Locality Figure 4. Flakes and cores from Kathu Townlands, Beaumont Excavation. A. Large flake off of the edge of the core consistent with biface shaping removal. B. Large flake with centripedal dorsal scars. C. Blade, note that there is some cortex (indicated by C in the sketch) and that scars are not parallel. D–F. Small flakes, note that F is off the edge of the core. G. Discoidal core with removals off both faces. Break on one edge (upper edge in right view). H. Discoidal core with one large flake removal. Note that on the right hand face the working is unclear and it is possible that this is a natural surface. doi:10.1371/journal.pone.0103436.g004 Figure 4. Flakes and cores from Kathu Townlands, Beaumont Excavation. A. Large flake off of the edge of the core consistent with biface shaping removal. B. Large flake with centripedal dorsal scars. C. Blade, note that there is some cortex (indicated by C in the sketch) and that scars are not parallel. D–F. Small flakes, note that F is off the edge of the core. G. Discoidal core with removals off both faces. Break on one edge (upper edge in right view). H. Discoidal core with one large flake removal. Note that on the right hand face the working is unclear and it is possible that this is a natural surface. d i 10 1371/j l 0103436 004 doi:10.1371/journal.pone.0103436.g004 handaxes show scars from arched invasive biface thinning flakes. Raw material for handaxes is heavily dominated by local BIF, however rare examples on quartzite are also found (Figure 6c). secondary flakes (n = 222, 17.3%) and the high frequency of noncortical flakes (n = 989, 77.08%) does not match expectation for a quarry site where the initial stages of roughing out a block would take place. Due to the frequency of edge damage, clearly intentional retouch is difficult to identify in the sample. Examples of clear retouch are rare (n = 10) consisting of side scrapers, retouched pieces, and one notch. Lithic Analysis Cores are extremely variable in morphol- ogy and include an amorphous core lacking any evident organization of flaking surfaces (n = 1) centripedal Levallois (n = 1), unidirectional (n = 1), cores on flakes (n = 2), and possible bifacial preforms (n = 2), and discoidal (n = 3), (Figure 4: g–h). Flakes that are typologically Levallois, and blades are found in the assemblage but are rare (Figure 4b–c). Biface thinning flakes are absent but many of the flakes could be derived from stages of biface manufacture (Figure 4a). The raw material for the entire flake assemblage is the local BIF. July 2014 \| Volume 9 \| Issue 7 \| e103436 Excavation The excavated trenches exposed dense artefact deposits mixed with BIF rubble and sand reaching a maximum depth of 2.2 m in Trench B. A massive quantity of artefacts was recovered conforming to Beaumont’s observations. No stratigraphy was evident within the artefact horizon and artefacts and natural slabs appear randomly oriented within the profiles (Figure 7: a–b). The artefact horizon sits directly on bedrock in all units where bedrock was reached. There is thus an unconformity of over 2 billion years between the bedrock and the overlying rubble. In most units the archaeological horizon is overlain by sand mixed with very small (1–2 cm.), subangular pebbles. The contact between the archae- ological horizon and the overlying sands is abrupt. The biface sample from the excavation unit is highly variable in shape (Figure 5: c–e, Figure 6). By comparison the selected sample from the surface collection conforms more regularly to the morphology of pointed handaxes. Handaxes show both large shaping scars and smaller removals along the edges. None of the The bedrock topography does not conform to the modern land surface along the main NE-SW transect (Figure 8). Modern topography along this transect rises gently from NE to SW but the bedrock rises sharply between trenches E and D (ca. 1 m rise over PLOS ONE \| www.plosone.org July 2014 \| Volume 9 \| Issue 7 \| e103436 6 Kathu Townlands Earlier Stone Age Locality Figure 5. Metric attributes of flakes from Beaumont Excavation and bifaces from Beaumont Excavation and Surface collection. A. Frequency of maximum dimension for flakes. B. Frequency of weight for flakes. C–E. Metric attributes of bifaces (based on [25]). Beaumont excavation indicated by solid dot, surface collection by outline dot. doi:10.1371/journal.pone.0103436.g005 Figure 5. Metric attributes of flakes from Beaumont Excavation and bifaces from Beaumont Excavation and Surface collection. A. Frequency of maximum dimension for flakes. B. Frequency of weight for flakes. C–E. Metric attributes of bifaces (based on [25]). Beaumont excavation indicated by solid dot, surface collection by outline dot. doi:10.1371/journal.pone.0103436.g005 10 m). The bedrock topography then dips again into a channel centred on Trench B before rising again in Trenches A and G. The archaeological horizon appears to fill the irregularities in the bedrock topography of the channel. Moving to the SE, where the current topography drops rapidly, the archaeological horizon thins. Excavation artefacts recovered are consistent with an Acheulean age and are similar to those found in Beaumont’s excavation. Included are a small number of well-made handaxes (Figure 9a). In Square 2 the pieces of a refitting broken handaxe were recovered; the tip from Level 5 and the base from Level 7 (Figure 9b). This observation suggests redeposition of the sediments that make up the archaeological horizon. During excavation of Square 1 a prevalence of larger artefacts was noticed in the upper levels (to a depth of ca. 70 cm), and an excess of smaller debitage in the lower levels. In Square 2 a large slab of bedrock was encountered within the artefact horizon, roughly 40 cm above actual bedrock. Calcium carbonate concretions (calcretes) develop within the archaeological horizon at lower elevations (Trenches E and I; Figure 7: c–d). In trenches J/K and the developer’s trench calcretes were prevalent although not consolidated into a continuous horizon as has been documented in other contexts in the region. The calcretes in these trenches develop within sands however it appears that aspects of the archaeological horizon are also incorporated into the calcretes. PLOS ONE \| www.plosone.org Discussion The Kathu Townlands site is, as previously reported, a massive deposit of Acheulean artefacts likely the product of the exploitation of local BIF. Preliminary analysis of lithic material and fieldwork in Analysis of the archaeological assemblage recovered has not yet been carried out however a number of field observations provide insight into the formation of the archaeological horizon. All July 2014 \| Volume 9 \| Issue 7 \| e103436 PLOS ONE \| www.plosone.org 7 Kathu Townlands Earlier Stone Age Locality Figure 6. Handaxes from surface collection: A–B. Banded Ironstone. C. Quartzite. doi:10.1371/journal.pone.0103436.g006 Kathu Townlands Earlier Stone Figure 6. Handaxes from surface collection: A–B. Banded Ironstone. C. Quartzite. doi:10.1371/journal.pone.0103436.g006 possible that these are Late Pleistocene to Holocene in age as is the case for dated sands and other calcretes in the region [11]. a mitigation context provide some initial observations about site formation. The archaeological horizon is a dense rubble of artefacts and unworked BIF in a sand matrix. It is possible that the sands are derived from the overlying deposits rather than an initial component of the archaeological horizon. The archaeological horizon sits directly on the bedrock and appears to fill in irregularities in the bedrock topography, including a channel apparent in the NE-SW transect. Analogues to this situation can be found on the ironstone hills on the Bestwood farm about 1 km to the east where artefacts sit directly on the exposed bedrock. Kathu Townlands represents a complex and massive archaeo- logical context that requires further research. The occurrence of a low density Acheulean occupation at Wonderwerk Cave suggests significant differences in the intensity of hominin activity between the two flanks of the Kuruman Hills. It is likely that the density of archaeological remains within the Kathu Complex is related to local availability of water. It is intriguing to consider the possibility that the hominin occupation of Wonderwerk Cave during the Acheulean is the result of seasonal mobility of small groups of hominins dispersing from the core occupation area on the western flanks of the Kuruman Hills. Both lithic analysis and field observations failed to find compelling evidence of high energy transport. Polishing of artefacts is rare, clasts are angular, and there is little evidence of sorting. There is however considerable evidence of mechanical damage to artefacts (randomly organized microfracture on edges). July 2014 \| Volume 9 \| Issue 7 \| e103436 Discussion This observation is supported by the recovery of a broken handaxe found in levels separated by 20 cm within Square 2 (Figure 9b). Given the rapid development of the town of Kathu, long term protection of the deposits at Townlands is imperative. However, simply preserving isolated artefact-dense patches of land will not preserve the information necessary to interpret the Kathu Complex. Preservation cannot be limited to the presence and absence of surface archaeological material. In Figure 2b the area of the site is intentionally indicated by shading rather than a distinct boundary. The concept of site boundaries that play such a critical role in heritage resource management is inherently problematic [24]. A broader landscape-based effort of subsurface testing including palaeo-landcape and palaeo-environmental reconstruction is essential to our understanding of this extraordi- nary record. Sources of this information must be protected along with archaeological remains. Together with the other components of the Kathu Complex, this site represents a high density of hominin occupation that presents a challenge to reconstructions of hominin adaptations during the Early-Middle Pleistocene. Hope- As a working hypothesis, these observations suggest that at the time of occupation Kathu Townlands was a small exposed hill of chert-rich BIF. This locality was the site of ongoing intensive occupation and exploitation for stone tool manufacture. While one function of the site might have been as a quarry, rough-outs and primary flakes are rare, and there is a small component of finished tools (including rare handaxes made on non-local quartzite) suggesting that the site might have had a more diversified function. The artefact horizon was subject to mechanical forces and redeposition. The mechanisms responsible for these processes are currently unclear but high energy water transport does not appear to have played a role. The age of the covering sands and the development of calcretes remains to be explored, however it is PLOS ONE \| www.plosone.org July 2014 \| Volume 9 \| Issue 7 \| e103436 July 2014 \| Volume 9 \| Issue 7 \| e103436 8 Kathu Townlands Earlier Stone Age Locality Figure 7. Profiles from 2013 excavation. A. Trench A: Square 1. Massive deposit of Banded Irontone rubble and artefacts overlying bedrock in a sandy matrix. Note lack of bedding or sorting. B. Trench I: Square 5. Shallow massive deposit of Banded Ironstone rubble and artefacts overlying bedrock with overlying deposits of sand. C. Trench E: Square 3. Discussion Discrete calcrete nodule that developed near the interface of the rubble/artefact deposit and underlying bedrock. Note parallel bedding of the Ironstone within the calcrete nodule. Approximate width of image 50 cm. D. Trench J/ K. Discrete nodular calcrete developing in the sand and into the underlying Banded Ironstone rubble. Does not exhibit parallel Ironstone bedding found in (c). Approximate width of images 50 cm. doi:10.1371/journal.pone.0103436.g007 Figure 7. Profiles from 2013 excavation. A. Trench A: Square 1. Massive deposit of Banded Irontone rubble and artefacts overlying bedrock in a sandy matrix. Note lack of bedding or sorting. B. Trench I: Square 5. Shallow massive deposit of Banded Ironstone rubble and artefacts overlying bedrock with overlying deposits of sand. C. Trench E: Square 3. Discrete calcrete nodule that developed near the interface of the rubble/artefact deposit and underlying bedrock. Note parallel bedding of the Ironstone within the calcrete nodule. Approximate width of image 50 cm. D. Trench J/ K. Discrete nodular calcrete developing in the sand and into the underlying Banded Ironstone rubble. Does not exhibit parallel Ironstone bedding found in (c). Approximate width of images 50 cm. doi:10.1371/journal.pone.0103436.g007 Figure 8. Composite profile along 2013 excavation trenches. doi:10.1371/journal.pone.0103436.g008 PLOS ONE \| www.plosone.org 9 July 2014 \| Volume 9 \| Issue 7 \| e103436 Figure 8. Composite profile along 2013 excavation trenches. doi:10.1371/journal.pone.0103436.g008 July 2014 \| Volume 9 \| Issue 7 \| e103436 PLOS ONE \| www.plosone.org Kathu Townlands Earlier Stone Age Locality om the 2013 excavation. A. Biface in situ in Square 4 (Trench G). B. Broken biface found in Square 2 (Trench C). Tip found in w surface), base found in Level 7 (70–80 cm. below surface). e.0103436.g009 Kathu Townlands Earlier Stone Age Locality Figure 9. Artefacts from the 2013 excavation. A. Biface in situ in Square 4 (Trench G). B. Broken biface found in Square 2 (Trench C). Tip found in Level 5 (50–60 cm below surface), base found in Level 7 (70–80 cm. below surface). doi:10.1371/journal.pone.0103436.g009 Figure 9. Artefacts from the 2013 excavation. A. Biface in situ in Square 4 (Trench G). B. Broken biface found in Square 2 (Trench C). Tip found in Level 5 (50–60 cm below surface), base found in Level 7 (70–80 cm. below surface). doi:10.1371/journal.pone.0103436.g009 permit (#577, CaseID 2992) from the South African Heritage Resource Agency (SAHRA), and was greatly aided by the efforts of Colette Scheermeyer, Mariagrazia Galimberti, and Katie Smuts of SAHRA. Acknowledgments This research would not have been possible without support from the Canadian Social Science and Humanities Research Council, and the assistance of Dr. David Morris and the staff of the McGregor Museum. The paper was significantly improved by comments from Professor Francis Thackeray, Dr. Kathleen Kuman, and an anonymous third reviewer. Remaining errors are our own. Analysis of collections was under agreement with the McGregor Museum. Fieldwork was conducted under Discussion We are grateful to Abel Dichakane, Adam Thibeletsh, Peter Ntulini, Jacob Maine, Sam Tmapo, and Mahloko Moses of the National Museum, Bloemfontein for assistance with the excavations. We are grateful to Stephen Wessels and the Zamani Project for providing the DEM. fully, enough of this record will be preserved to allow for ongoing reconstruction and investigation of this important scientific resource. References 1. Lister AM, Dirks W, Assaf A, Chazan M, Goldberg P, et al (2013) New fossil remains of Elephas from the southern Levant: Implications for the evolutionary history of the Asian elephant. Palaeogeogr Palaeoclimatol Palaeoecol 386: 119– 130. 14. Wilkins J, Chazan M (2012) Blade production ,500 thousand years ago at Kathu Pan 1, South Africa: support for a multiple origins hypothesis for early Middle Pleistocene blade technologies. J Archaeol Sci 39: 1883–1900. g J 15. Wilkins J, Schoville BJ, Brown KS, Chazan M (2012) Evidence for Early Hafted Hunting Technology Science 338: 942–946. 2. Stiles D, Hay R, O’Neil J (1974) The MNK chert factory site, Olduvai Gorge, T i W ld A h l 5 285 308 2. Stiles D, Hay R, O’Neil J (1974) The MNK chert factory site, Olduvai Gorge, Tanzania. World Archaeol 5: 285–308. 16. Chazan M, Wilkins J, Morris D, Berna F (2012) Bestwood 1: a newly discovered Earlier Stone Age living surface near Kathu, Northern Cape Province, South Africa. Antiquity 86: 331. 3. Barkai R, Gopher A, La Porta PC (2002) Palaeolithic landscape of extraction: Flint surface quarries and workshops at Mt Pua, Israel. Antiquity 76: 672–680. 4. Petraglia M, LaPorta P, Paddayya K (1999) The first Acheulian quarry in India: stone tool manufacture, biface morphology, and behaviors. J Anthropol Res 55: 39–70. 17. Chazan M, Porat N, Sumner TA, Horwitz LK (2013) The use of OSL dating in unstructured sands: The archaeology and chronology of the Hutton Sands at Canteen Kopje (Northern Cape Province, South Africa). Archaeol Anthropol Sci 5(4): 351–363. 5. Kuman K (2001) An Acheulean Factory Site with Prepared Core Technology near Taung, South Africa. S Afr Archaeol Bull 56, 8–22. 18. Chazan M, Avery M, Bamford MK, Berna F, Brink J, et al (2012) The Oldowan horizon in Wonderwerk Cave (South Africa): Archaeological, geological, paleontological and paleoclimatic evidence. J Hum Evol 63 (6): 859–866. 6. McNabb J, Beaumont P (2011a) A Report on the Archaeological Assemblages from Excavations by Peter Beaumont at Canteen Koppie, Northern Cape, South Africa. In: University of Southampton Series in Archaeology 4. Oxford: Archaeopress. 86. paleontological and paleoclimatic evidence. J Hum Evol 63 (6): 859 19. Berna F, Goldberg P, Horwitz LK, Brink J, Holt S, et al (2012) Microstrati- graphic evidence of in situ fire in the Acheulean strata of Wonderwerk Cave, Northern Cape province, South Africa. Author Contributions Conceived and designed the experiments: SJHW VL MC. Performed the experiments: SJHW VL MC. Analyzed the data: SJHW VL MC. Contributed to the writing of the manuscript: SJHW VL MC. July 2014 \| Volume 9 \| Issue 7 \| e103436 PLOS ONE \| www.plosone.org 10 Kathu Townlands Earlier Stone Age Locality References Proc Natl Acad Sci 109 (20): E1215– E1220. p 7. McNabb J, Beaumont P (2011b). Excavations in the Acheulean Levels of the Earlier Stone Age Site of Canteen Koppie, Northern Cape Province, South Africa. Proc. Prehist. Soc. 78: 51–71. 20. Beaumont PB (1990) Kathu Townlands 1. In: Beaumont PB, Morris D, editors. Guide to archaeological sites in the Northern Cape. Kimberley: McGregor Museum. 96–97. 8. Knoll AH, Beukes NJ (2009) Introduction: Initial investigations of a Neoarchean shelf margin-basin transition (Transvaal Supergroup, South Africa). Precambri- an Res 169: 1–14. 21. Beaumont PB (2004) Kathu Pan and Kathu Townlands/Uitkoms. In: Morris D, Beaumont PB, editors. Archaeology in the Northern Cape: Some Key Sites. Kimberley: McGregor Museum. 50–53. 9. Altermann W, Siegfried HP (1997) Sedimentology and facies development of an Archaean shelf: carbonate platform transition in the Kaapvaal Craton, as deduced from a deep borehole at Kathu, South Africa. J Afr Earth Sci 24(3): 391–410. 22. Thackeray F, Thackeray A (1980) National Monuments Council Permit Per/1/ 231: Report on an archaeological survey of the central northern Cape Province. Permit Report on file at the South African Heritage Resources Agency. 10. Nash DJ, McLaren SJ (2003) Kalahari valley calcretes: their nature, origins, and environmental significance. Quat Int 111: 3–22. 23. Walker SJH, Chazan M, Lukich V, Morris D (2013) A second Phase 2 archaeological data recovery at the site of Kathu Townlands for Erf 5116: Kathu, Northern Cape Province. A Permit report by the Archaeology Department, McGregor Museum. Available at: http://www.sahra.org.za/ heritage-reports/kathu-townlands-erf-5116-second-phase-2-report. 11. Bateman MD, Thomas DSG, Singhvi AK (2003) Extending the aridity record of the Southwest Kalahari: current problems and future perspectives. Quat Int 111: 37–49. 12. Matmon A, Ron H, Chazan M, Porat N, Horwitz LK (2012) Reconstructing the history of sediment deposition in caves: A case study from Wonderwerk Cave, South Africa. Geol Soc Am Bull 124: 611–625. 24. Dunnell RC (1992) The Notion Site, in: Rossignol, J. & Wandsnider, L (Eds.), Space, time, and archaeological landscapes: Interdisciplinary contributions to archaeology. Plenum Press, New York, 21–41. 13. Porat N, Chazan M, Gru¨n R, Aubert M, Eisenmann V, et al (2010) New radiometric ages for the Fauresmith industry from Kathu Pan, southern Africa: Implications for the Earlier to Middle Stone Age transition. J Archaeol Sci 37: 269–283. 25. Roe DA (1964) The British Lower and Middle Palaeolithic: Some Problems, Methods of Study and Preliminary Results. Proc. Prehist. Soc. 30: 245–267. References PLOS ONE \| www.plosone.org PLOS ONE \| www.plosone.org July 2014 \| Volume 9 \| Issue 7 \| e103436 11
https://openalex.org/W4293588877	https://codrulcosminului.usv.ro/wp-content/uploads/2022/11/Article.3.Vol_.28.pdf	English	null	German Colonists from Bessarabia in Russian Historical Literature of the 19th – Early 20th Centuries	Codrul Cosminului/Analele ştiinţifice de istorie "Codrul Cosminului"	2,022	cc-by	8,966	ERROR: type should be string, got "https://doi.org/10.4316/CC.2022.01.03 https://doi.org/10.4316/CC.2022.01.03 Copyright © 2022 “Codrul Cosminului”, XXVIII, 2022, No. 1, p. 55–72. GERMAN COLONISTS FROM BESSARABIA \nIN RUSSIAN HISTORICAL LITERATURE \nOF THE 19TH – EARLY 20TH CENTURIES Vyacheslav STEPANOV \n \nInstitute of Slavic Studies of the Russian Academy of Sciences \nE-mail: vpstepanovpochta@gmail.com Abstract. The article analyses the Russian historiographical heritage of German \nsettlers in Bessarabia in the 19th century and the first decades of the 20th century, the least \nresearched period in Russian history. The historical Bessarabia territory is currently part of \nthe Republic of Moldova and Ukraine. Despite using the same documents, the studies on the \nGermans of Bessarabia reach different conclusions because they depended on the political \nconjuncture in which the papers were written. Russian authors of the 19th - the first decades \nof the 20th centuries viewed German settlers’ ethnic, confessional and agricultural \ndevelopment in Bessarabia from different angles, often guided by their subjective \nimpressions. The researchers covering this period relied exclusively on the recollections of \ncontemporaries of German ancestors who had been living in Bessarabia since the beginning \nof the 19th century. Often, the authors used unverified propaganda information. In this paper, \nprinted sources from the nineteenth and early twentieth centuries were used to create a \ncoherent image of the German population in Bessarabia and reflect the attitudes of Russian \nscientists and authorities regarding this ethnical group of immigrants who endured great \nhardship. 2 Pompei N. Batyushkov, Bessarabiya. Istoricheskoe opisanie [Bessarabia. Historical de-\nscription], Sankt-Peterburg, 1892, s. 144; I. F. Lagorio, Statisticheskie svedeniya o bessa-\nrabskih nemetskih koloniyah v 1861-1862 gg. [Statistical information about the \nBessarabian German colonies in 1861-1862], in “Zapiski Bessarabskogo oblastnogo \nstatisticheskogo komiteta”, t. I, Kishinev, 1864, p. 73. 1 Spiski naselennyh mest Rossijskoj imperii, sostavlennye i izdavaemye central'nym \nstatisticheskim komitetom ministerstva vnutrennih del. III Bessarabskoj oblast' [Lists of \nthe Russian Empire localities, compiled and published by the Central Statistical \nCommittee of the Ministry of the Interior. III Bessarabia region], Sankt-Peterburg, \n1861, p. XXIV. 3 Valentin S. Zelenchuk, Naselenie Bessarabii i Podnestrov'ya v XIX v. (Etnicheskie i social'no-\ndemograficheskie processy) [The Population of the Bessarabia and Transnistria in the \n19th century (Ethnic and Social-Demographical processes)], Kishinev, 1979, p. 212. Keywords: Germans, historiography, Bessarabia, population, authorities, migration Keywords: Germans, historiography, Bessarabia, population, authorities, migration Rezumat. Coloniștii germani din Basarabia în lucrările de istorie rusești din \nsecolul al XIX-lea și începutul secolului XX. Articolul analizează patrimoniul istoriografic \nde limbă rusă privitor la coloniștii germani din Basarabia în secolul al XIX-lea și primele \ndecenii ale secolului al XX-lea, în perioada cel mai slab cercetată din istoria Rusiei. În prezent, \nteritoriul Basarabiei istorice face parte din Republica Moldova și Ucraina. În ciuda faptului \ncă folosesc aceleași izvoare, studiile privind germanii din Basarabia ajung la concluzii \ndiferite, deoarece au depins de conjunctura politică în care s-au redactat lucrările. Autorii German Colonists from Bessarabia in Russian Historical Literature 56 ruși din secolul al XIX-lea și din primele decenii ale secolului XX au privit evoluția etnică, \nconfesională și agrară a coloniștilor germani din Basarabia din unghiuri diferite, ghidându-\nse adesea de impresiile lor subiective. Cercetătorii acestei perioade s-au bazat exclusiv pe \nmemoriile contemporanilor strămoșilor germani care s-au așezat în Basarabia încă de la \nînceputul secolului al XIX-lea. Adeseori, autorii au folosit informații de propagandă \nneverificate. În studiul de față, sursele tipărite din secolul al XIX-lea și începutul secolului XX \nau fost folosite pentru a schița o imagine coerentă a populației germane din Basarabia, \nprecum și pentru a reflecta atitudinile oamenilor de știință și ale autorităților ruse față de \nacest grup etnic de imigranți care a avut de îndurat mari greutăți. ruși din secolul al XIX-lea și din primele decenii ale secolului XX au privit evoluția etnică, \nconfesională și agrară a coloniștilor germani din Basarabia din unghiuri diferite, ghidându-\nse adesea de impresiile lor subiective. Cercetătorii acestei perioade s-au bazat exclusiv pe \nmemoriile contemporanilor strămoșilor germani care s-au așezat în Basarabia încă de la \nînceputul secolului al XIX-lea. Adeseori, autorii au folosit informații de propagandă \nneverificate. În studiul de față, sursele tipărite din secolul al XIX-lea și începutul secolului XX \nau fost folosite pentru a schița o imagine coerentă a populației germane din Basarabia, \nprecum și pentru a reflecta atitudinile oamenilor de știință și ale autorităților ruse față de \nacest grup etnic de imigranți care a avut de îndurat mari greutăți. INTRODUCTION Bessarabia was, in the 19th century, a gateway to the Balkans. It was evident \nwhen this territory became a space of confrontation between empires. The \nregion’s peculiarity is that all life activities closely interact between the \nautochthonous population and immigrants. One of the understudied ethnic \ngroups was the German colonists. They arrived in Bessarabia in the 19th century \nand nearly vanished entirely because of a decision made by the Stalinist \nadministration in 1940. German colonists represented a different governmental \ninitiative for recovering vacant areas, and a small portion of the Swiss and French \njoined them. They were also called “Warsaw settlers”, although there were very \nfew Poles among them. They left Poland and went to Bessarabia. By the mid-\nnineteenth century, they were around 26 thousand people.1 They all lived mainly \nin the region’s south in the Akkerman uyezd2. By the beginning of the 20th century, \nthe German population had already been living in 75 villages in Bessarabia3. Many Vyacheslav Stepanov 57 German colonies created subsidiary settlements to preserve land capital. German colonies created subsidiary settlements to preserve land capital. 5 Georg Leibbrandt, Die deutschen Kolonien in Cherson und Bessarabien: Berichte der \nGemeindeämter über Entstehung und Entwicklung der lutherischen Kolonien in der 1. \nHälfte des 19. Jahrhunderts, Stuttgart, Ausland und Heimat-Verl.-Aktienges., 1926. 4 Аnton Golopenția, Românii de la est de Bug [Romanians from the East of the Bug River], \nVol. 1, București, Editura Enciclopedică, 2006, р. 33. 6 A. Reimarus, Mennonity (kratkiy ocherk) [Mennonites (Brief Essay)], Moskva, \nBezbozhnik, 1930. p\n]\nş\n12 I. A. Nestorescu, Transnistria noastră [Our Transnistria], Buzău, 1942. 7 Viktor Schirmunski, Die deutschen Kolonisten in der Ukraine. Geschichte, Mundarten, \nVolkslied, Volkskunde, Charkow, Zentral-Völker-Verlag, 1928. 9 Hermann Aubin, Die Jahrtausendfeier der Rheinlande und die Ostmark, in Festschrift zur \nJahrtausendfeier der Rheinländer in Danzig, Danzig, 1925, S. 9–15. HISTORIOGRAPHICAL REVIEW German Colonists from Bessarabia in Russian Historical Literature 58 The following period focuses on Soviet Moldova. The scholars practically \ndid not study the German population. Separate studies in the field of ethnic \ndemographics only partially touched upon certain aspects of the history and \nculture of the German settlers’ everyday life (Valentin S. Zelenchuk13 and the book \nof Vladimir M. Kabuzan14). The data on the German population of the region’s \ncities is partly reflected in the works of Nikolai V. Babilunga15 and Stepanida F. Kustryabova16. The monographs of well-known historians of the time, Yakim S. Grosul17, Il’ya G. Budak18 and Ivan I. Meshcheryuk19 present aspects of the socio-\neconomic life of the German population in the region. The studies of Elena I. Druzhinina, who investigated the general situation in southern Ukraine in the 19th \ncentury, including the condition of the German colonies20, are worth mentioning. 14 Vladimir M. Kabuzan, Narodonaselenie Bessarabskoi oblasti i levoberezhnyh raionov \nPridnestrov'ya (konec XVIII – pervaia polovina XIX vv.) [The population of the \nBessarabian region and the left-bank regions of Transnistria (the end of the 18th – the \nfirst half of the 19th centuries)], Kishinev, “Shtiintsa”, 1974. 15 Nikolai V. Babilunga, Natsional'nyj sostav gorodskogo naseleniya Bessarabii v epohu \nimperializma [The national composition of the urban population of Bessarabia in the \nera of imperialism], in “Izvestiya AN MSSR: Seriya obshchestvennyh nauk”, 1979, № 3, \np. 24-33; Idem, Dinamika gorodskogo naseleniya Bessarabii v epohu imperializma \n[Dynamics of the urban population of Bessarabia in the era of imperialism], in \n“Izvestiya AN MSSR: Seriya obshchestvennyh nauk”, 1981, № 2, p. 53-61. 16 Stepanida F. Kustryabova, Polozhenie trudyashchihsya i demograficheskie protsessy v \ngorodah Bessarabii (1918–1940) [The situation of workers and demographic processes \nin the cities of Bessarabia (1918–1940)], Kishinev, “Shtiintsa”, 1977. 17 Yаkim S. Grosul, Krest'yane Bessarabii (1812-1861) [Peasants of Bessarabia (1812-\n1861)], Kishinev, Gosud. Izdatel’stvo Moldavii, 1956. 18 Yakim S. Grosul, Il'ya G. Budak, Krest'yanskaya reforma 60-70 gg. XIX veka v Bessarabii \n[Peasant reform 60-70s of 19th century in Bessarabia], Kishinev, Gosud. Izdatel’stvo \nMoldavii, 1956; Yakim S. Grosul, Il'ya G. Budak, Ocherki istorii narodnogo hozyaistva \nBessarabii (1812-1861 gg.) [Essays on the history of the national economy of \nBessarabia (1812-1861)], Kishinev, Kartia moldoveniaske, 1967. 19 Ivan I. Meshcheryuk, Sotsial'no-ekonomicheskoe razvitie bolgarskih i gagauzskih siol v \nyuzhnoi Bessarabii (1808–1856) [Socio-economic development of the Bulgarian and \nGagauz villages in southern Bessarabia (1808-1856)], Kishinev, Redaktsionno-\nizdatel’skij otdel AN MSSR, 1970. 20 Elena I. HISTORIOGRAPHICAL REVIEW The article focuses on German migration to Bessarabia and their impact and \ndevelopment of the Bessarabian land. During the interwar period, the Romanian \nsociologist Anton Golopenţia, a Romanian sociologist, was interested in the \nGerman population of Bessarabia. He dedicated his research in the Bug4 Region to \nstatistics, particularly those affecting the German people. Before the Second World \nWar, the theme of German settlers comes up with renewed vigour in Germany5 \nand the USSR. Moreover, in the Soviet Union, a steadily negative attitude towards \nthe German population began to take shape in the late 20s (A. Reimarus = David \nPenner6, Viktor Schirmunski7, Aleksandr Klibanov8). Hermann Aubin, one of the authors of the theory of “cultural province,” \nattempted to explain how the German people’s cultural area took shape, regardless \nof state and temporary borders9. According to the theory, the colonists played an \nessential role in expanding the ecumene of the Germanic ethnos. Many experts \nsupported this theory, including Heinrich Brunner, Joachim Schulz10, and others. The period of 1940-1944 was a period of active spread of Romanian values. This process also affected German settlers who survived after repatriation in \n1940. Particular attention was paid to the situation in the Transnistria of the \nDniester region. It is worth mentioning the works of Ion Nistor11, I. A. Nestorescu12, and others. 5 Georg Leibbrandt, Die deutschen Kolonien in Cherson und Bessarabien: Berichte der \nGemeindeämter über Entstehung und Entwicklung der lutherischen Kolonien in der 1. Hälfte des 19. Jahrhunderts, Stuttgart, Ausland und Heimat-Verl.-Aktienges., 1926. f\nJ\ng\ng\n6 A. Reimarus, Mennonity (kratkiy ocherk) [Mennonites (Brief Essay)], Moskva, \nBezbozhnik, 1930. 7 Viktor Schirmunski, Die deutschen Kolonisten in der Ukraine. Geschichte, Mundarten, \nVolkslied, Volkskunde, Charkow, Zentral-Völker-Verlag, 1928. 8 A. Klibanov, Menonity [Mennonites], Мoskva–Leningrad, 1931. y [\n]\ng\n9 Hermann Aubin, Die Jahrtausendfeier der Rheinlande und die Ostmark, in Festschrift zur \nJahrtausendfeier der Rheinländer in Danzig, Danzig, 1925, S. 9–15. 10 Joachim Schulz, Die deutschen Volksgruppen in Südosteuropa, Leipzig – Berlin, Teubner, 1940. 11 Ion Nistor, Aspecte geopolitice şi culturale din Transnistria [Geopolitical and cultural 10 Joachim Schulz, Die deutschen Volksgruppen in Südosteuropa, Leipzig – Berlin, Teubner, 1940. 11 Ion Nistor, Aspecte geopolitice şi culturale din Transnistria [Geopolitical and cultural \naspects of Transnistria], Bucureşti, 1942. 11 Ion Nistor, Aspecte geopolitice şi culturale din Transnistria [Geopolitical and cultural \naspects of Transnistria], Bucureşti, 1942. p\n]\nş\n12 I. A. Nestorescu, Transnistria noastră [Our Transnistria], Buzău, 1942. 14 Vladimir M. Kabuzan, Narodonaselenie Bessarabskoi oblasti i levoberezhnyh raionov \nPridnestrov'ya (konec XVIII – pervaia polovina XIX vv.) [The population of the \nBessarabian region and the left-bank regions of Transnistria (the end of the 18th – the \nfirst half of the 19th centuries)], Kishinev, “Shtiintsa”, 1974. 13 Valentin S. Zelenchuk, op. cit. 21 Ol'ga V. Shcherbinina, Kritika sovremennoi zapadnogeranskoi burzhuaznoi istoriografii \nistorii Moldavii [Criticism of modern West German bourgeois historiography of the \nhistory of Moldova], Kishinev, “Shtiintsa”, 1984. 22 Alfred Ziеbart, Arzis, Bessarabien. Blick auf 150 Jahre Kolonistenleben und –schicksal \n1816–1966, Ludwigsburg, Eigenverlag des Verfassers, 1971. 31 Sergei V. Kapustin, Etnografiya Bessarabii glazami issledovatelej nemetskogo proiskhozh-\ndeniya [Ethnography of Bessarabia through the eyes of researchers of German origin] \nin the «Uchenye zapiski Orlovskogo gosudarstvennogo universiteta”, 2017, no. 3(76), \np. 23–28; Idem, Coverage of the history and culture of life support of the German settlers \nof Bessarabia and the left bank of the Dniester in the late 20th - early decades of the 21st \ncentury in the printed scientific literature of Germany, in “ABYSS”, 2019, № 1(07), \nhttp://abyss.su/_media/abyss_issue/7/kapustin_statja_na_anglijskom.pdf (Accessed \non 22.01.2022). HISTORIOGRAPHICAL REVIEW Druzhinina, Yuzhnaya Ukraina 1800-1825 gg [Southern Ukraine 1800-1825], \nMoskva, “Nauka”, 1970; Eadem, Yuzhnaya Ukraina v period krizisa feodalizma. 1825–\n1860 gg. [Southern Ukraine in the period of feudalism crisis. 1825–1860], Moskva, \n“Nauka”, 1981. Vyacheslav Stepanov 59 As for the Soviet period, Ol’ga V. Shcherbinina21 wrote about bourgeois \nfalsifications, including the theme of German settlers. In the second half of the 20th \ncentury, German scholars published works devoted to various aspects of the past \nof the German colonists in Bessarabia (Alfred Ziеbart22, Jakob Becker23, etc.). Some \nauthors who wrote about the German colonists themselves were their \ndescendants (Alexander Suga24). The study of German issues became increasingly important in the post-\nSoviet period, in both the Republic of Moldova and elsewhere. Valentina \nChirtoagă’s work stands out among Moldovan specialists. She wrote the first \ndissertation research devoted to the study of various aspects of the lives of \nGerman colonists in Bessarabia25. Two researchers from Germany, descendants of \nthe Bessarabian Germans, Luminița Fassel26 and Ute Schmidt27, and a native of \nRussia, Viktor Disendorf,28 attempted to summarise information about the \nGermans of Bessarabia. 22 Alfred Ziеbart, Arzis, Bessarabien. Blick auf 150 Jahre Kolonistenleben und –schicksal \n1816–1966, Ludwigsburg, Eigenverlag des Verfassers, 1971. kob Becker, Bessarabien und sein Deutschtum, Bietigheim – Württemberg, Krug, 1966. 24 Alexander Suga, Die völkerrechtliche Lage Bessarabiens in der geschichtlichen \nEntwicklung des Landes, Bonn, 1958. 25 Valentina Chirtoagă, Etnicii germani din Basarabia: evoluție demografică, social-\neconomică și cultural-spirituală (1814-1917). Teză de doctor în istorie [The German \nethnic groups in Bessarabia: demographic, socio-economic and cultural-spiritual \nevolution (1814-1917). The Doctoral Thesis in History], Chișinău, 2018. 26 Lumința Fassel, Das deutsche Schulwesen in Bessarabien (1812-1940). Eine komparativ-\nhistorische und sozio-kulturelle Untersuchung, München, Verlag Südostdeutsches \nKulturwerk, 2000. 27 Ute Schmidt, Die Deutschen aus Bessarabien. Eine Minderheit aus Südosteuropa (1814 bis \nheute), 2 durchgesehene Auflage. Köln – Weimar – Wien, Böhlau Verlag 2004; Eadem, \nBessarabien. Deutsche Kolonisten am Schwarzen Meer, Potsdam, Deutsches \nKulturforum Östliches Europa, 2008; Eadem, Bessarabiya. Nemeckie kolonisty \nPrichernomor'ya: ocherki [Bessarabia. German colonists of the Black Sea region: \nessays], Odessa, Optimum, 2015. 28 Viktor Disendorf, Istoricheskaya demografiya nemetskogo naseleniya Rossiii SSSR (XVIII \n– nachalo XXI vv.) [Historical demography of the German population of Russia and the \nUSSR (18th – the beginning of the 21st centuries)], Rottenburg, 2010, \nhttp://wolgadeutsche.net/diesendorf/Ist_demograf.pdf (Accessed on 25.01.2022). HISTORIOGRAPHICAL REVIEW German Colonists from Bessarabia in Russian Historical Literature 60 A series of doctoral thesis from Ukraine demonstrates the surge of interest in \nthe history of German settlements in modern Ukrainian historiography. These \nworks touch upon various aspects of the history and culture of the German colonists \nin Ukraine, and some cover the border areas of the Republic of Moldova.29 Sergei Kretinin30, a modern researcher from Russia, made several advances \nin examining German colonists’ history. In his articles, Sergei V. Kapustin31 \naddressed ethnography and the history of the Bessarabian Germans. Although the German presence in Bessarabia has been explored, issues still \nneed to be addressed. One of them is connected with the history of personalities \nof German colonists, which is insufficiently covered. The traditional daily lifestyle \nof this ethnic-social segment of the Bessarabian population remains understudied. y\np\nof German colonists, which is insufficiently covered. The traditional daily lifestyle \nof this ethnic-social segment of the Bessarabian population remains understudied. The purpose of this study has been to focus briefly on the examination of \n19th-century Russian works about German colonists in Bessarabia. The purpose of this study has been to focus briefly on the examination of \n19th-century Russian works about German colonists in Bessarabia. 29 Kateryna S. Lyah, Nіmets'komovnі kolonіsty Pіvdnya Ukrayny v mul'tinatsіonal'nomu \notochennі: problema vzaemodіy kul'tur (XIX – pochatok XX st.) [German-speaking \ncolonists of Southern Ukraine in a multinational environment: the problem of cultural \ninteraction (19th – early 20th centuries)], Donetsk, 2005. 30 Sergei V. Kretinin, Bessarabskie nemtsy: poslednie kolonisty Rossiiskoj imperii [Bessa-\nrabian Germans: the Last Colonists of the Russian Empire], in “Rossiya – Germaniya”, \n2013, No 2(6), p. 65–69. 32 Vyacheslav P. Stepanov, Trudy po etnografii naseleniya Bessarabii XIX – nachala XX vv. \nOcherki istorii etnograficheskogo izucheniya bessarabskih ukraintsev [Works on the eth-\nnography of the population of Bessarabia in the 19th – early 20th centuries. Essays on \nthe history of the ethnographic study of Bessarabian Ukrainians], Kishinev, 2001, p. 69. DYNAMICS OF SCIENTIFIC INTEREST IN THE HISTORICAL FATE \nOF THE GERMAN COLONISTS IN THE 19TH – EARLY 20TH CENTURIES The historiography of the 19th century and early 20th century, devoted to the \nGerman settlers’ historical destiny and everyday culture in Bessarabia, can be \ndivided into several stages. Stage I – the first half of the 19th century until the 1860s was \ncharacterised by episodic information about the Germans in this region. Territorial features, such as the positioning of German settlers on the periphery, \nalso impacted this. However, it was precisely because of this fact that the Vyacheslav Stepanov 61 Bessarabian outskirts of the Russian Empire were recorded and explored in the \n19th century on behalf of various government departments.32 Pavel P. Svinyin’s description of Bessarabia was one of the first published \nmaterials approached. It mentioned that “in addition to several Austrian subjects, \nscattered in the cities and villages of Bessarabian for various trades, 1743 families \nwere resettled here from the Duchy of Warsaw in 1814; they were all of the \nLutheran confession, except 121 Catholic families”33. The privileges granted to the Warsaw colonists encouraged the influx of \nGerman settlers for nearly the entire 19th century. If the region had 33,501 \nGermans in 1861, there were 42,681 in 1891 and 59,998 people in 189734. The \nGerman colonists’ special status, long-term benefits, and the ban against dividing \nland plots contributed to the accumulation of land property35. Subsequently, \nGerman colonists were labelled as the prosperous but closed and \nuncommunicative representatives of a foreign ethnic environment.36 Stepan Kornilovich, who wrote about the region in the twenties of the 19th \ncentury, confirmed the words of Svinyin and noted that “several families from \nforeign lands were resettled to the region for agriculture and cultivation of \nvineyards; they were mostly from the Duchy of Warsaw, the Bavarian and \nWertemberg kingdoms and Switzerland”37. Kornilovich named several families of 33 Opisanie Bessarabskoi oblasti. Sostavleno vedomstva gosudarstvennoi kollegii inostran-\nnyh del nadvornym sovetnikom Pavlom Svinyinym, 1816 goda [Description of the \nBessarabian region. Compiled by the department of the State Collegium of Foreign \nAffairs by court adviser Pavel Svinyin, 1816], in “Zapiski Odesskogo obshchestva istorii \ni drevnostei”, vol. VI, 1867, p. 206. 34 Valentin S. Zelenchuk, op. cit., p. 211. More accurate is the figure given by Sеrgei Sulyak \n– 60,206 inhabitants of German origin. See: Sergei G. 36 Nikolai D. Zabotkin, Mirnoe zavoevanie nemtsami Yuzhnoi Bessarabii [Peaceful conquest \nof southern Bessarabia by the Germans], Akkerman, 1910, p. 3. epan Kornilovich, Statisticheskoe opisanie Bessarabii sobstvenno tak nazyvaemoi, i 35 Valentin S. Zelenchuk, op. cit., p. 211–212. 35 Valentin S. Zelenchuk, op. cit., p. 211–212. Budzhaka, s prilozheniem general’nogo plana ego kraya, sostavlennoe pri grazhdanskoi \ns’emke Bessarabii, proizvodivshei po vysochajshemu poveleniyu razmezhevanie zemel' \nonoi na uchastki s 1822 po 1828-i god [Statistical description of the so-called Bessarabia \nproper, or Budjak, with the application of the general plan of its region, compiled \nduring the civil survey of Bessarabia, which, by the highest command, delimited its \nlands into sections from 1822 to 1828], Akkerman, 1899, p. 23. 40 Aleksandr I. Zashchuk, Materialy dlya geografii i statistiki Rossii, sobrannye ofitserami \nGeneral’nogo shtaba. Bessarabskaya oblast' [Materials for the geography and statistics \nof Russia, collected by officers of the General Staff. Bessarabia region], Sankt-\nPeterburg, 1862, ch. 1-2; Idem, Etnografiya Bessarabskoi oblasti [Ethnography of the \nBessarabian region], in “Zapiski Odesskogo obshchestva istorii i drevnostei”, T. 5, \nOdessa, 1863, p. 491–589. DYNAMICS OF SCIENTIFIC INTEREST IN THE HISTORICAL FATE \nOF THE GERMAN COLONISTS IN THE 19TH – EARLY 20TH CENTURIES Sulyak, Naselenie Bessarabii v XIX \nv.: chislennost', natsional'nyi sostav, religioznaya prinadlezhnost' (Po materialam \ndorevolyutsionnyh issledovanii) [The population of Bessarabia in the 19th century: size, \nethnic composition, religious affiliation (based on pre-revolutionary studies)], in Trudy \nOb’edinennogo nauchnogo soveta po gumanitarnym problemam i istoriko-kul’turnomu \nnaslediyu. 2009, Sankt-Peterburg, 2010, p. 61. 35 Valentin S. Zelenchuk, op. cit., p. 211–212. 36 Nikolai D. Zabotkin, Mirnoe zavoevanie nemtsami Yuzhnoi Bessarabii [Peaceful conquest \nof southern Bessarabia by the Germans], Akkerman, 1910, p. 3. German Colonists from Bessarabia in Russian Historical Literature 62 German settlers in 1971: there were 4848 men and 4522 women.38 The author \nalso focused on the German occupations, listing all of the German colonies and \npresenting their economic characteristics39. There are only positive reviews on \nthe German farming tradition of the Germans near Daragan. Stage II, the 1860s - 1880s. In the second half of the 19th century, a \nconsiderable amount of statistical literature included information about the \nGerman settlers. Outlying territories were specially studied, which explains the \nauthorities’ attention to their development. The comprehensive works of General \nStaff officer Aleksandr Zashchuk, which contain information about the German \ncolonists, are evidence of this.40 In 1862, historian E. Busch published a study41 that provided information \non the number of German settlers in Russia. According to V. Disendorf, E. Busch’s \ninformation was underestimated for Novorossia (which included Bessarabia) by \n16.6%. Inaccuracies, in his opinion, are explained by the fact that Busch used the \narchives of the Evangelical Church, which did not consider the Germans of other \nfaiths.42 Friedrich Matthäi’s book was published a few years later. A separate section \nis devoted to the German colonists of Bessarabia43. A sombre work written in 1875 40 Aleksandr I. Zashchuk, Materialy dlya geografii i statistiki Rossii, sobrannye ofitserami \nGeneral’nogo shtaba. Bessarabskaya oblast' [Materials for the geography and statistics \nof Russia, collected by officers of the General Staff. Bessarabia region], Sankt-\nPeterburg, 1862, ch. 1-2; Idem, Etnografiya Bessarabskoi oblasti [Ethnography of the \nBessarabian region], in “Zapiski Odesskogo obshchestva istorii i drevnostei”, T. 5, \nOdessa, 1863, p. 491–589. 41 E. H. Busch, Materialien zur Geschichte und Statistik des Kirchen- und Schulwesens der Ev.-\nLuth. Gemeinden in Russland, St. Petersburg, Commissionsverlag von G. Haessel, 1862. 41 E. H. Busch, Materialien zur Geschichte und Statistik des Kirchen- und Schulwesens der Ev.-\nLuth. Gemeinden in Russland, St. Petersburg, Commissionsverlag von G. Haessel, 1862. 42 Viktor Disendorf, op. cit. 39 Ibid., p. 188–220. 43 Friedrich Matthäi, Die deutschen Ansiedlungen in Rußland. Ihre Geschichte und \nihrevolkswirtschaftliche Bedeutungfür die Vergangenheit und Zukunft. Studien über das \nrussische Kolonisationswesen und über die Herbeiziehung fremder Kulturkräfte nach \nRußland, Leipzig, Hermann Fries, 1866, p. 37–67. ]\np\n38 Ibid., p. 62. 38 Ibid., p. 62. p. 188–220. 44 Alekandr F. Rittikh, Plemennoi sostav kontingentov russkoi armii i muzhskago naseleniya \nEvropeiskoi Rossii [The tribal composition of the contingents of the Russian army and \nthe male population of European Russia], St. Petersburg, 1875; Idem, Etnografiches-\nkaya karta Evropeiskoi Rossii [Ethnographic map of European Russia], Sankt-\nPeterburg, 1875. 45 I. F. Lagorio, Statisticheskie svedeniya o Bessarabskih nemetskih koloniyah v 1861-1862 \ngg. [Statistical information about the Bessarabian German colonies in 1861-1862], in \n“Zapiski Bessarabskogo oblastnogo statisticheskogo komiteta”, T. I, Kishinev, 1864, \np. 73–77. 46 Peter I. Köppen, Hronologicheskii ukazatel’ materialov dlya istorii inorodtsev Evropeiskoi \nRossii [Chronological index of materials for the history of foreigners in European \nRussia], Sankt-Peterburg, 1861. 47 Dmitrii I. Bagalei, Kolonizatsiya Novorossiiskogo kraya i pervye shagi ego po puti kul’tury. \nIstoricheskii etyud [Colonization of the Novorossiysk District territory and its first steps \nalong the path of culture. Historical sketch], Kiev, 1889, p. 77–99. DYNAMICS OF SCIENTIFIC INTEREST IN THE HISTORICAL FATE \nOF THE GERMAN COLONISTS IN THE 19TH – EARLY 20TH CENTURIES 43 Friedrich Matthäi, Die deutschen Ansiedlungen in Rußland. Ihre Geschichte und \nihrevolkswirtschaftliche Bedeutungfür die Vergangenheit und Zukunft. Studien über das \nrussische Kolonisationswesen und über die Herbeiziehung fremder Kulturkräfte nach \nRußland, Leipzig, Hermann Fries, 1866, p. 37–67. Vyacheslav Stepanov 63 by General Staff Officer, a member of the Russian Geographical Society, Aleksandr \nF. Rittikh,44 was a reaction of the military department to universal military service, \nintroduced in the state in 1874. I. Logario45 presented certain statistical \ninformation about the German settlers of the region. Academician Peter von \nKöppen, a well-known Russian statistician, made substantial contributions to the \nstudy of southern ethnography, detailing the areas of settlement, the number, and \noccupation of the Germans.46 When discussing the later decades of the twentieth century, it is essential to \nmention Dmitrii I. Bagalei’s historical study, dedicated to the colonisation of \nNovorossia. The whole section was devoted to foreign settlers, and German \ncolonists were separately studied. The author emphasised that they differed from \neach other by their religious identity47. Stage III, between the 1890s and 1918, was the time of generalisation of \nthe accumulated knowledge, the publication of several comprehensive works on \nthe region’s history and ethnography, and the creation of local scientific centres. Pompei Batyushkov’s book Bessarabia... was published towards the end of the \n19th century; it was an example of Russian patriotic work.48 Batyushkov’s attitude \ntoward German presence in Russia reflects the views of the Slavophile intelligentsia, \nwhich advocated depriving the German population of special privileges in the post- 47 Dmitrii I. Bagalei, Kolonizatsiya Novorossiiskogo kraya i pervye shagi ego po puti kul’tury. Istoricheskii etyud [Colonization of the Novorossiysk District territory and its first steps \nalong the path of culture. Historical sketch], Kiev, 1889, p. 77–99. 47 Dmitrii I. Bagalei, Kolonizatsiya Novorossiiskogo kraya i pervye shagi ego po puti kul’tury. Istoricheskii etyud [Colonization of the Novorossiysk District territory and its first steps \nalong the path of culture. Historical sketch], Kiev, 1889, p. 77–99. 48 Batyushkov’s book was perceived differently in scientific circles. A. I. Yatsimirsky, the \nmodern Moldavian researcher Olga S. Lukyanets, criticized it for tendentiousness. See: \nOlga S. Lukyanets, Russkie issledovateli i moldavskaya etnograficheskaya nauka v XIX – \nnachale XX v. [Russian researchers and Moldovan ethnographic science in the 19th – \nearly 20th centuries], Kishinev, 1986, p. 70. German Colonists from Bessarabia in Russian Historical Literature 64 reform period (late 70-80s of the 19th century). V. S. 52 Bessarabiya. Graficheskii, istoricheskii, statisticheskii, ekonomicheskii, etnograficheskii, \nliteraturnyi i spravochnyi sbornik. Izdanie gazety “Bessarabets” pod red. P. A. Krushevana \n[Bessarabia. Graphic, historical, statistical, economic, ethnographic, literary and refer-\nence collection. Edition of the newspaper “Bessarabets” published by P. A. Krushevan], \nMoskva, 1903, p. 183. y\ny\np\n51 Irina Izhboldina, “Zhivaya starina” v proizvedeniyah russkih literatorov Bessarabii kontsa \nXIX - nachala XX vv.: issledovaniya i materialy [“Zhivaya starina” in the works of Russian \nwriters of Bessarabia in the late 19th – early 20th centuries: research and materials], \nKishinev, 2012, p. 13. 49 Valentin S. Zelenchuk, op. cit., с. 212. With reference to Nikolai Zabotkin, Mirnoe zavoe-\nvanie nemtsami Yuzhnoi Bessarabii [Peaceful conquest of southern Bessarabia by the \nGermans,], Akkerman, Ekon. tip. I.A. Kara-Fedorova, 1911, p. 2. 50 Igor P. Shornikov, Obshchestvenno-politicheskaya i literaturnaya deyatel'nost' P. A. Krushevana \n[Socio-political and literary activities of P. A. Krushevan]. Avtoreferat dissertatsii na \nsoiskanie uchenoi stepeni kandidata istoricheskih nauk. Special'nost' 07.00.02 – \nOtechestvennaya istoriya. Moskva, 2011, p. 33. 49 Valentin S. Zelenchuk, op. cit., с. 212. With reference to Nikolai Zabotkin, Mirnoe zavoe-\nvanie nemtsami Yuzhnoi Bessarabii [Peaceful conquest of southern Bessarabia by the \nGermans,], Akkerman, Ekon. tip. I.A. Kara-Fedorova, 1911, p. 2. \n50 Igor P. Shornikov, Obshchestvenno-politicheskaya i literaturnaya deyatel'nost' P. A. Krushevana \n[Socio-political and literary activities of P. A. Krushevan]. Avtoreferat dissertatsii na \nsoiskanie uchenoi stepeni kandidata istoricheskih nauk. Special'nost' 07.00.02 – \nOtechestvennaya istoriya. Moskva, 2011, p. 33. \n51 Irina Izhboldina, “Zhivaya starina” v proizvedeniyah russkih literatorov Bessarabii kontsa \nXIX - nachala XX vv.: issledovaniya i materialy [“Zhivaya starina” in the works of Russian \nwriters of Bessarabia in the late 19th – early 20th centuries: research and materials], \nKishinev, 2012, p. 13. \n52 Bessarabiya. Graficheskii, istoricheskii, statisticheskii, ekonomicheskii, etnograficheskii, \nliteraturnyi i spravochnyi sbornik. Izdanie gazety “Bessarabets” pod red. P. A. Krushevana \n[Bessarabia. Graphic, historical, statistical, economic, ethnographic, literary and refer-\nence collection. Edition of the newspaper “Bessarabets” published by P. A. Krushevan], \nMoskva, 1903, p. 183. DYNAMICS OF SCIENTIFIC INTEREST IN THE HISTORICAL FATE \nOF THE GERMAN COLONISTS IN THE 19TH – EARLY 20TH CENTURIES Zelenchuk gave another \nexample of such an approach, which intensified in connection with Russia’s \naccumulated socio-economic contradictions before World War I. For instance, \nNikolai D. Zabotkin wrote in 1910 that “German colonisation in the Akkerman uyezd \nthreatens with a complete collapse of the economy of other agricultural elements \nand is aimed at their displacement and enslavement.”49 Contemporary Moldavian historians have ambivalent attitudes toward \nPavel A. Krushevan’s personality. He is associated with Russian “imperial” \nnationalists, strong Russian state supporters50, and bright, creative characters.51 \nHowever, P. Krushevan was undoubtedly a very talented person, endowed with \nnumerous virtues and a man of his time – an odious monarchist. The detailed \ndescription of the German population of the region given in the almanack \n“Bessarabets”, in a separate paper “Bessarabia”, edited by P. A. Krushevan, is very \nindicative: “Hardworking workers, excellent agronomists and cattle breeders, the \nGermans live clean, even in wealth. Their long-standing resettlement on the \nsteppes of Budjak did not in the least change their patriarchal Germanic habits at \nthe end of the 18th century. They do not have a broad Russian prowess, capable of \nadopting everything and applying it to everything. The German peasant rarely \nlooks beyond the boundary of his plot and knows only himself and the society of \nhis village. Given this, their skill and diligence remain with them and are not \ninstilled in the native population.”52 Vladimir N. Butovich is connected to the German settlers’ topic, \nemphasising aspects and potential opportunities for cohabitation of numerous Vyacheslav Stepanov 65 ethnonational colonies on the territory of Bessarabia53. These studies can also \ninclude the later work of Jakob Stach54. Among the works of contemporaries of \nthat period, which covered the history and culture of everyday life of the German \ncolonists in southern Russia, reference must be made to the work of Jacob Prinz, \npublished in German in Moscow in 1898.55 Not only the Germans were interested in their compatriots – the works of \nIvan Sergeev,56 Petr Tkachev,57 Andrei Fadeev,58 and Grigorii Pisarevskii59, which \nchronicled the history and culture of the Germans in Russia, were published in the \nlate 19th and early 20th centuries. 53 Vladimir Butovich, Materialy dlya etnograficheskoi karty Bessarabskoi gubernii \n[Materials for the ethnographic map of the Bessarabian province], Kiev, 1916. 54 Jаkob Stach, Ocherki iz istorii i sovremennoi zhizni yuzhnorusskih kolonii [Essays on the \nhistory and modern life of the South Russian colonies], Moskva, 1916. 55 Jakob Prinz, Die Kolonien der Brüdergemeinde. Ein Beitrag zur Geschichte der Deutschen \nKolonien Südrusslands, Moskva, 1898. 56 Ivan I. Sergeev, Mirnoe zavoevanie Rossii nemtsami [Peaceful conquest of Russia by the \nGermans], Peterburg, 1915. 57 Petr N. Tkachev, Nemetskie idealisty i filistery [German idealists and philistines], in \n“Delo”, Sankt-Peterburg, 1867, No 12, p. 1–27. 60 Andrej M. Fadeev, op. cit., Part I, p. 47. \n61 Ibid., p. 60. 58 Andrei M. Fadeev, Vospominaniya Andreya Mihajlovicha Fadeeva. 1790-1867 gg. \n[Memoirs of Andrei Mikhailovich Fadeev. 1790-1867], Odessa, 1897, in 2 parts. 59 Grigorii G. Pisarevskii, Iz istorii inostrannoi kolonizatsii v Rossii v XVIII veke (po neiz-\nvestnym arhivnym materialam) [From the history of foreign colonization in Russia in \nthe 18th century (According to unknown archival materials)], Moskva, 1909. 60 Andrej M. Fadeev, op. cit., Part I, p. 47. \n61 Ibid., p. 60. 62 Natal'ya V. Savinova, Rossiiskii natsionalizm i nemetskie pogromy v Rossii v gody Pervoi \nmirovoi voiny: 1914-1917 gg. [Russian nationalism and German pogroms in Russia \nduring the First World War: 1914-1917]. Avtoreferat dissertacii na soiskanie uchenoj \nstepeni kandidata istoricheskih nauk. Special'nost': 07.00.02 – Otechestvennaya \nistoriya; \nhttp://www.dissercat.com/content/rossiiskii-natsionalizm-i-nemetskie-\npogromy-v-rossii-v-gody-pervoi-mirovoi-voiny-1914-1917-g [Accessed: 25.12.2021]; \nSvetlana V. Obolenskaya, “Nemeckii vopros” i predstavleniya v Rossii o nemtsah v gody \nPervoi mirovoi vojny [The “German question” and ideas in Russia about the Germans \nduring the First World War], in “Rossiya i Germaniya”, Moskva, vyp. 2, 2001, p. 189. DYNAMICS OF SCIENTIFIC INTEREST IN THE HISTORICAL FATE \nOF THE GERMAN COLONISTS IN THE 19TH – EARLY 20TH CENTURIES The Memoirs of Andrei Fadeev presents a \ndescription of the German colonies, located, among other things, in Transnistria \nand further in Novorossia.60 An interesting observation found in his memoirs \nleads to the conclusion that the influx of German colonists to Russia and, in \nparticular, to Bessarabia was associated with the mystical hobbies of Emperor \nAlexander I: “The reason for this was the petition of Mrs Krüdener and other \nmystics addressed to Emperor Alexander, who often travelled abroad. They \ngreatly influenced the Tsar and requested him to allow many inhabitants of all \ncountries of Germany, consisting mainly of pietists, to move to Russia. They also \nasked him to provide them with special protection.”61 The information presented \nabove cannot be considered an unequivocal description of the reasons for the \nmigration of the German colonists to Russia. Nevertheless, it makes one wonder \nhow much the fate of an entire ethnic-social community may depend on a \nsubjective factor. 55 Jakob Prinz, Die Kolonien der Brüdergemeinde. Ein Beitrag zur Geschichte der Deutschen \nKolonien Südrusslands, Moskva, 1898. 56 Ivan I. Sergeev, Mirnoe zavoevanie Rossii nemtsami [Peaceful conquest of Russia by the \nGermans], Peterburg, 1915. 57 Petr N. Tkachev, Nemetskie idealisty i filistery [German idealists and philistines], in \n“Delo”, Sankt-Peterburg, 1867, No 12, p. 1–27. 58 Andrei M. Fadeev, Vospominaniya Andreya Mihajlovicha Fadeeva. 1790-1867 gg. [Memoirs of Andrei Mikhailovich Fadeev. 1790-1867], Odessa, 1897, in 2 parts. 59 Grigorii G. Pisarevskii, Iz istorii inostrannoi kolonizatsii v Rossii v XVIII veke (po neiz-\nvestnym arhivnym materialam) [From the history of foreign colonization in Russia in \nthe 18th century (According to unknown archival materials)], Moskva, 1909. 60 Andrej M. Fadeev, op. cit., Part I, p. 47. 61 Ibid., p. 60. German Colonists from Bessarabia in Russian Historical Literature 66 On top of the “anti-German wave” during the First World War, the question \nof the loyalty of the Bessarabian Germans and the Germans in Russia became \nparticularly relevant. Then the restrictions on their rights signed by Nicholas II \ncame out.62 A series of anti-German sentiments swept through the ministries and \ndepartments and among the adherents of the country’s regime.63 On February 2, 1915, the Emperor signed the Law “On the Restriction of \nLand Ownership and Land Use of Enemy Subjects”, which also applied to the \nBessarabian Germans. During the war, radical approaches prevailed. 63 Nikolai Varadinov, Istoriya Ministerstva vnutrennih del [History of the Ministry of the \nInterior], Sankt-Peterburg, 1858, P. III, T. 1, p. 501 – 503; Ivan Sobolev, Bor'ba s \n“nemetskim zasil'em” v Rossii v gody Pervoi mirovoi vojny [The fight against “German \ndominance” in Russia during the First World War]. Avtoreferat dissertacii na soiskanie \nuchenoj stepeni kandidata istoricheskih nauk. Special'nost': Special'nost': 07.00.02 – \nOtechestvennaya istoriya. Sankt-Peterburg, 1998; Sergei P. Sheluhin, Nemetskaya \nkolonizatsiya na yuge Rossii [German colonization in southern Russia], Odessa, 1915. DYNAMICS OF SCIENTIFIC INTEREST IN THE HISTORICAL FATE \nOF THE GERMAN COLONISTS IN THE 19TH – EARLY 20TH CENTURIES The \ngovernment focused on eradicating the German ethnic community in Russia, \nalthough at the expense of the agricultural economy.64 This liquidation policy was \nespecially active in the front-line zones, although it had little impact. Therefore, \nKarl Lindemann65 emphasised that the Bessarabian Germans never thought their \nland could be taken from them. It would be incorrect to claim that only extreme opinions against the \nGerman settlers in Russia were expressed in the literature of the time. There were 63 Nikolai Varadinov, Istoriya Ministerstva vnutrennih del [History of the Ministry of the \nInterior], Sankt-Peterburg, 1858, P. III, T. 1, p. 501 – 503; Ivan Sobolev, Bor'ba s \n“nemetskim zasil'em” v Rossii v gody Pervoi mirovoi vojny [The fight against “German \ndominance” in Russia during the First World War]. Avtoreferat dissertacii na soiskanie \nuchenoj stepeni kandidata istoricheskih nauk. Special'nost': Special'nost': 07.00.02 – \nOtechestvennaya istoriya. Sankt-Peterburg, 1998; Sergei P. Sheluhin, Nemetskaya \nkolonizatsiya na yuge Rossii [German colonization in southern Russia], Odessa, 1915. 64 Aleksandra Yu. Bahturina, “Luchshe pust' nemtsy razoryatsya, chem budut shpionit'”: \nnemtsy-kolonisty i rossiiskoe obshchestvo v gody “germanskoi” vojny [“Better let the \nGermans go bankrupt than spy”: German colonists and Russian society during the \n“German” war], р. 10, in https://cyberleninka.ru/article/n/luchshe-pust-nemtsy-\nrazoryatsya-chem-budut-shpionit-nemtsy-kolonisty-i-rossiyskoe-obschestvo-v-gody-\ngermanskoy-voyny/viewer [Accessed on 04.01.2022]. 65 Karl E. Lindemann, Prekrashchenie zemlevladeniya i zemlepol'zovaniya poselyan-sob-\nstvennikov [Termination of land ownership and land use of village owners], Moskva, \nTipografіya K. L. Men'shova, 1917. 65 Karl E. Lindemann, Prekrashchenie zemlevladeniya i zemlepol'zovaniya poselyan-sob-\nstvennikov [Termination of land ownership and land use of village owners], Moskva, \nTipografіya K. L. Men'shova, 1917. 65 Karl E. Lindemann, Prekrashchenie zemlevladeniya i zemlepol'zovaniya poselyan-sob-\nstvennikov [Termination of land ownership and land use of village owners], Moskva, \nTipografіya K. L. Men'shova, 1917. Vyacheslav Stepanov 67 additional publications that took a more honest approach to the phenomena of \nGerman colonial presence in Russia. The works of Grigorii Pisarevskii66, Semion \nBondar’67 and Karl E. Lindemann68 are among them. However, because of the \npolitical climate and the previously prevailing public attitude, their voice was \nweak and had little impact on the overall situation. 66 Grigorii G. Pisarevskii, op. cit. \n67 Semion D. Bondar', Sekta menonitov Rossii [Russian Mennonite sect], Sankt-Peterburg, \nTipografiya V. D. Smirnova, 1916. \n68 Karl E. Lindemann, op. cit., p. 65-66. \n69 Jаkob Stach, op. cit., p. 97. \n70 Karl E. Lindemann, op. cit. \n71 Rossiya. Polnoe geograficheskoe opisanie nashego Otechestva: nastol'naya i dorozhnaya \nkniga dlya russkih lyudei. Novorossiya i Krym (Bessarabskaya, Hersonskaya, Tavriches-\nkaya i Ekaterinoslavskaya gubernii, oblasti Vojska Donskogo i Stavropol'skoi gubernii) \n[Russia. A complete geographical description of our Fatherland: a desktop and travel \nbook for Russian people. Novorossia and Crimea (Bessarabian, Kherson, Taurida and \nYekaterinoslav provinces, regions of the Don Army and Stavropol province)], Sankt- 66 Grigorii G. Pisarevskii, op. cit. \n67 Semion D. Bondar', Sekta menonitov Rossii [Russian Mennonite sect], Sankt-Peterburg, \nTipografiya V. D. Smirnova, 1916. \n68 Karl E. Lindemann, op. cit., p. 65-66. \n69 Jаkob Stach, op. cit., p. 97. \n70 Karl E. Lindemann, op. cit. 71 Rossiya. Polnoe geograficheskoe opisanie nashego Otechestva: nastol'naya i dorozhnaya \nkniga dlya russkih lyudei. Novorossiya i Krym (Bessarabskaya, Hersonskaya, Tavriches-\nkaya i Ekaterinoslavskaya gubernii, oblasti Vojska Donskogo i Stavropol'skoi gubernii) \n[Russia. A complete geographical description of our Fatherland: a desktop and travel \nbook for Russian people. Novorossia and Crimea (Bessarabian, Kherson, Taurida and \nYekaterinoslav provinces, regions of the Don Army and Stavropol province)], Sankt- DYNAMICS OF SCIENTIFIC INTEREST IN THE HISTORICAL FATE \nOF THE GERMAN COLONISTS IN THE 19TH – EARLY 20TH CENTURIES p\nAmong balanced studies, the extract from Jakob Stach’s book reveals the \nauthor’s attempt to uncover the cause for the German colonists’ poor mastery of \nthe Russian language and avoidance of interaction with Russian peasants: “until \nthe time of the Crimean War, and even until the sixties of the last century, there \ncan be no question of the attitude of the colonists towards the Russian population, \nbecause in South Russia there was almost no Russian population, and the sparsely \npopulated Russian cities very rarely communicated with the colonies, because at \nthat time there was no means of communication. The only exceptions were the \ncolonies located at the nearest distance from the cities: Odessa, Yekaterinoslav, \nand Elizavetgrad.69 When studying the issue concerning the situation of German settlers in the \nregion, one cannot disregard the policy of the official authorities about the German \nsettlement, which became more complicated during the First World War. In this \nregard, Karl E. Lindemann’s study, which reflected the situation of the German \npopulation by the decrees of February 2 and December 13, 1915, and July 10, 15 \nand August 19, 1916, is eloquent. The regulations had a considerable impact on \nthe economic status of German settlers before the Bolshevik Revolution70, as well \nas the development of an unfavourable attitude concerning colonists. The second half of the 19th century, especially the end and beginning of the \nnext century, is marked by the publication of several encyclopaedic works \ncontaining information about the Bessarabian Germans.71 German Colonists from Bessarabia in Russian Historical Literature 68 72 Olga S. Lukyanets, Kollektivnye predstavleniya narodov Bessarabii drug o druge (pervaya \npolovina ХХ v.) [Collective ideas of the peoples of Bessarabia about each other (the first \nhalf of the 20th century)], in Etnograficheskie issledovaniya v Respublike Moldova, \nKishinev, 2006, р. 181. Peterburg, 1910, vol. 14; Lev S. Berg, Bessarabiya. Strana. Lyudi. Hozyaistvo \n[Bessarabia. Country. People. Economy], Kishinev, Universitas, 1993, p. 116. Peterburg, 1910, vol. 14; Lev S. Berg, Bessarabiya. Strana. Lyudi. Hozyaistvo \n[Bessarabia. Country. People. Economy], Kishinev, Universitas, 1993, p. 116. \n72 Olga S. Lukyanets, Kollektivnye predstavleniya narodov Bessarabii drug o druge (pervaya Peterburg, 1910, vol. 14; Lev S. Berg, Bessarabiya. Strana. Lyudi. Hozyaistvo \n[Bessarabia. Country. People. Economy], Kishinev, Universitas, 1993, p. 116. \n72 Olga S. Lukyanets, Kollektivnye predstavleniya narodov Bessarabii drug o druge (pervaya \npolovina ХХ v.) [Collective ideas of the peoples of Bessarabia about each other (the first \nhalf of the 20th century)], in Etnograficheskie issledovaniya v Respublike Moldova, \nKishinev, 2006, р. 181. CONCLUSIONS Bessarabia in the 19th century never formed a single ethnocultural \ncommunity. There is no reason to assume that this was done on purpose. While \ncoexisting, the multicultural communities also kept their distance from one \nanother. Having their privileges, the Warsaw and Transdanubian colonists lived \nrelatively closed. The Jews lived in the same “chamber” within a different ethnic \nenvironment, which is also explained by their diverse confessions and occupation \n(trade, usury, possession, etc.). The territory substantially constrained the life of the Bessarabian settler, \nparticularly in the colony, which had a specific influence on the consciousness of \nboth the individual inhabitant and members of the ethnic-social community. According to the modern researcher Olga S. Lukyanets, the interaction between \nGerman settlers and the local population increased throughout the interwar \nperiod. By the end of the 30s, “Bessarabian Germans more often came into contact \nwith the region’s wealthiest part of the population”.72 The analysed data allows \nconcluding that the conditionally selected stage – the first half of the 19th century \nuntil the 60s – was characterised by episodic information about the Germans of \nthe region. Despite the voluminous and multifaceted works of the authors of that \ntime (Aleksandr Zashchuk, E. H. Busch, Aleksandr F. Rittikh, Peter von Köppen, \nDmitrii Bagalei and many others), the history and everyday life of the Germans in \nBessarabia were covered episodically and superficially. The second stage, lasting 60-80 years, can be characterised as a period of \nheightened interest in many elements of German settlers’ lives and daily lives, \nfollowed by a rise in hostility toward their presence from liberal and Slavophile \nauthors. Negative manifestations can also be seen in the next stage, which spans \nthe 1890s through 1918. Authors, who emphasised the importance of German \nsettlers to the economic growth of Russian regions, argued against such one-sided \nviews (Semion Bondar’, Grigorii Pisarevskii, Jakob Stach, Karl Lindemann and \nothers). Simultaneously, various publications appeared at the time in which Vyacheslav Stepanov 69 researchers attempted to perform a detailed study of the region’s people, \nincluding the lifestyle of the German colonists (Lev S. Berg). Furthermore, these \ngeneralising books often featured less specific content than those expressly \ndevoted to German colonists. At the same time, they attempted to employ a \ncomplex approach (Aleksandr I. Zashchuk laid the foundation in Ethnography of \nthe Bessarabian Region) that was to be developed in the future. CONCLUSIONS It is essential to add that modern researchers from Germany (Ute Schmidt), \nthe Republic of Moldova (Valentina Chirtoagă) and Russia (Sergei Kretinin, Sergei \nKapustin) are increasingly interested in studying the German colonists of \nBessarabia. It gives hope that the study of their historical fate in Bessarabia will \nbe continued. A comprehensive historiographical survey of the German colonists \nof Bessarabia also awaits its time. REFERENCES 1. Babilunga Nikolai, Dinamika gorodskogo naseleniya Bessarabii v epohu \nimperializma [Dynamics of the urban population of Bessarabia in the era of \nimperialism], in “Izvestiya AN MSSR: Seriya obshchestvennyh nauk”, 1981, № 2, \np. 53-61. 1. Babilunga Nikolai, Dinamika gorodskogo naseleniya Bessarabii v epohu \nimperializma [Dynamics of the urban population of Bessarabia in the era of \nimperialism], in “Izvestiya AN MSSR: Seriya obshchestvennyh nauk”, 1981, № 2, \np. 53-61. 2. Babilunga Nikolai, Natsional'nyj sostav gorodskogo naseleniya Bessarabii \nv epohu imperializma [The national composition of the urban population of \nBessarabia in the era of imperialism], in “Izvestiya AN MSSR: Seriya obshches-\ntvennyh nauk”, 1979, № 3, p. 24-33. 3. Bahturina Aleksandra, “Luchshe pust' nemtsy razoryatsya, chem budut \nshpionit'”: nemtsy-kolonisty i rossiiskoe obshchestvo v gody “germanskoi” voiny \n[“Better let the Germans go bankrupt than spy”: German colonists and Russian \nsociety during the “German” war], in https://cyberleninka.ru/article/n/luchshe-\npust-nemtsy-razoryatsya-chem-budut-shpionit-nemtsy-kolonisty-i-rossiyskoe-\nobschestvo-v-gody-germanskoy-voyny/viewer. 4. Becker Jakob, Bessarabien und sein Deutschtum, Bietigheim – \nWürttemberg, Krug, 1966. 5. Berg Lev, Bessarabiya. Strana. Lyudi. Hozyaistvo [Bessarabia. Country. People. Economy], Kishinev, Universitas, 1993. 6. Chirtoagă Valentina, Etnicii germani din Basarabia: evoluție demografică, \nsocial-economică și cultural-spirituală (1814-1917). Teză de doctor în istorie [The German Colonists from Bessarabia in Russian Historical Literature 70 German ethnic groups in Bessarabia: demographic, socio-economic and cultural-\nspiritual evolution (1814-1917). The Doctoral Thesis in History], Chișinău, 2018. German ethnic groups in Bessarabia: demographic, socio-economic and cultural-\nspiritual evolution (1814-1917). The Doctoral Thesis in History], Chișinău, 2018. German ethnic groups in Bessarabia: demographic, socio-economic and cultural-\nspiritual evolution (1814-1917). The Doctoral Thesis in History], Chișinău, 2018. 7. Disendorf Viktor, Istoricheskaya demografiya nemetskogo naseleniya \nRossiii SSSR (XVIII – nachalo XXI vv.) [Historical demography of the German \npopulation of Russia and the USSR (18th – the beginning of the 21st centuries)], \nRottenburg, 2010, http://wolgadeutsche.net/diesendorf/Ist_demograf.pdf. 8. Druzhinina Elena, Yuzhnaya Ukraina 1800-1825 gg [Southern Ukraine \n1800-1825], Moskva, “Nauka”, 1970. 8. Druzhinina Elena, Yuzhnaya Ukraina 1800-1825 gg [Southern Ukraine \n1800-1825], Moskva, “Nauka”, 1970. 9. Druzhinina Elena, Yuzhnaya Ukraina v period krizisa feodalizma. 1825–\n1860 gg. [Southern Ukraine in the period of feudalism crisis. 1825–1860], Moskva, \n“Nauka”, 1981. 9. Druzhinina Elena, Yuzhnaya Ukraina v period krizisa feodalizma. 1825–\n1860 gg. [Southern Ukraine in the period of feudalism crisis. 1825–1860], Moskva, \n“Nauka”, 1981. 10. Golopenția Аnton, Românii de la est de Bug [Romanians from the East \nof the Bug River], București, Editura Enciclopedică, 2006. 10. REFERENCES Golopenția Аnton, Românii de la est de Bug [Romanians from the East \nof the Bug River], București, Editura Enciclopedică, 2006. 11. Grosul Yаkim, Krest'yane Bessarabii (1812-1861) [Peasants of \nBessarabia (1812-1861)], Kishinev, Gosud. Izdatel’stvo Moldavii, 1956. 11. Grosul Yаkim, Krest'yane Bessarabii (1812-1861) [Peasants of \nBessarabia (1812-1861)], Kishinev, Gosud. Izdatel’stvo Moldavii, 1956. 12. Grosul Yakim, Budak Il’ya, Ocherki istorii narodnogo hozyaistva \nBessarabii (1812-1861 gg.) [Essays on the history of the national economy of \nBessarabia (1812-1861)], Kishinev, Kartia moldoveniaske, 1967. 12. Grosul Yakim, Budak Il’ya, Ocherki istorii narodnogo hozyaistva \nBessarabii (1812-1861 gg.) [Essays on the history of the national economy of \nBessarabia (1812-1861)], Kishinev, Kartia moldoveniaske, 1967. 13. Grosul Yakim, Budak Il'ya, Krest'yanskaya reforma 60-70 gg. XIX veka v \nBessarabii [Peasant reform 60-70s of 19th century in Bessarabia], Kishinev, \nGosud. Izdatel’stvo Moldavii, 1956. 13. Grosul Yakim, Budak Il'ya, Krest'yanskaya reforma 60-70 gg. XIX veka v \nBessarabii [Peasant reform 60-70s of 19th century in Bessarabia], Kishinev, \nGosud. Izdatel’stvo Moldavii, 1956. 14. Izhboldina Irina, “Zhivaya starina” v proizvedeniyah russkih literatorov \nBessarabii kontsa XIX - nachala XX vv.: issledovaniya i materialy [“Zhivaya starina” \nin the works of Russian writers of Bessarabia in the late 19th – early 20th centuries: \nresearch and materials], Kishinev, 2012. 15. Kabuzan Vladimir, Narodonaselenie Bessarabskoi oblasti i levoberezh-\nnyh raionov Pridnestrov'ya (konec XVIII – pervaia polovina XIX vv.) [The population \nof the Bessarabian region and the left-bank regions of Transnistria (the end of the \n18th – the first half of the 19th centuries)], Kishinev, “Shtiintsa”, 1974. 15. Kabuzan Vladimir, Narodonaselenie Bessarabskoi oblasti i levoberezh-\nnyh raionov Pridnestrov'ya (konec XVIII – pervaia polovina XIX vv.) [The population \nof the Bessarabian region and the left-bank regions of Transnistria (the end of the \n18th – the first half of the 19th centuries)], Kishinev, “Shtiintsa”, 1974. 16. Kapustin Sergei, Coverage of the history and culture of life support of the \nGerman settlers of Bessarabia and the left bank of the Dniester in the late 20th - \nearly decades of the 21st century in the printed scientific literature of Germany, in \n“ABYSS”, 2019, № 1 (07), http://abyss.su/abyss_issue/7/10 16. Kapustin Sergei, Coverage of the history and culture of life support of the \nGerman settlers of Bessarabia and the left bank of the Dniester in the late 20th - \nearly decades of the 21st century in the printed scientific literature of Germany, in \n“ABYSS”, 2019, № 1 (07), http://abyss.su/abyss_issue/7/10 17. REFERENCES Kapustin Sergei, \nEtnografiya Bessarabii glazami issledovatelej \nnemetskogo proiskhozhdeniya [Ethnography of Bessarabia through the eyes of \nresearchers of German origin], in “Uchenye zapiski Orlovskogo gosudarstvennogo \nuniversiteta”, 2017, no. 3(76), p. 23–28. Vyacheslav Stepanov 71 18. Kretinin Sergei, Bessarabskie nemtsy: poslednie kolonisty Rossiiskoj \nimperii [Bessarabian Germans: the Last Colonists of the Russian Empire], in \n“Rossiya – Germaniya” 2013 No 2(6) p 65–69 18. Kretinin Sergei, Bessarabskie nemtsy: poslednie kolonisty Rossiiskoj \nimperii [Bessarabian Germans: the Last Colonists of the Russian Empire], in \n“Rossiya – Germaniya” 2013 No 2(6) p 65–69 19. Kustryabova Stepanida, Polozhenie trudyashchihsya i demograficheskie \nprotsessy v gorodah Bessarabii (1918–1940) [The situation of workers and \ndemographic processes in the cities of Bessarabia (1918–1940)], Kishinev, \n“Shtiintsa”, 1977. 20. Lukyanets Olga, Kollektivnye predstavleniya narodov Bessarabii drug o \ndruge (pervaya polovina ХХ v.) [Collective ideas of the peoples of Bessarabia about \neach other (the first half of the 20th century)], in Etnograficheskie issledovaniya v \nRespublike Moldova, Kishinev, 2006, р. 179–208. 21. Lukyanets Olga, Russkie issledovateli i moldavskaya etnograficheskaya \nnauka v XIX – nachale XX v. [Russian researchers and Moldovan ethnographic \nscience in the 19th – early 20th centuries], Kishinev, 1986. 22. Lyah Kateryna, Nіmets'komovnі kolonіsty Pіvdnya Ukrayny v mul'tina-\ntsіonal'nomu otochennі: problema vzaemodіy kul'tur (XIX – pochatok XX st.) \n[German-speaking colonists of Southern Ukraine in a multinational environment: \nthe problem of cultural interaction (19th – early 20th centuries)], Donetsk, 2005. 22. Lyah Kateryna, Nіmets'komovnі kolonіsty Pіvdnya Ukrayny v mul'tina-\ntsіonal'nomu otochennі: problema vzaemodіy kul'tur (XIX – pochatok XX st.) \n[German-speaking colonists of Southern Ukraine in a multinational environment: \nthe problem of cultural interaction (19th – early 20th centuries)], Donetsk, 2005. 23. Meshcheryuk Ivan, Sotsial'no-ekonomicheskoe razvitie bolgarskih i \ngagauzskih siol v yuzhnoi Bessarabii (1808–1856) [Socio-economic development \nof the Bulgarian and Gagauz villages in southern Bessarabia (1808-1856)], \nKishinev, Redaktsionno-izdatel’skij otdel AN MSSR, 1970. 23. Meshcheryuk Ivan, Sotsial'no-ekonomicheskoe razvitie bolgarskih i \ngagauzskih siol v yuzhnoi Bessarabii (1808–1856) [Socio-economic development \nof the Bulgarian and Gagauz villages in southern Bessarabia (1808-1856)], \nKishinev, Redaktsionno-izdatel’skij otdel AN MSSR, 1970. 24. Nistor Ion, Aspecte geopolitice şi culturale din Transnistria [Geopolitical \nand cultural aspects of Transnistria], Bucureşti, 1942. 24. Nistor Ion, Aspecte geopolitice şi culturale din Transnistria [Geopolitical \nand cultural aspects of Transnistria], Bucureşti, 1942. 25. Obolenskaya Svetlana, “Nemeckii vopros” i predstavleniya v Rossii o \nnemtsah v gody Pervoi mirovoi vojny [The “German question” and ideas in Russia \nabout the Germans during the First World War], in “Rossiya i Germaniya”, Moskva, \nvyp. 18. Kretinin Sergei, Bessarabskie nemtsy: poslednie kolonisty Rossiiskoj \nimperii [Bessarabian Germans: the Last Colonists of the Russian Empire], in \n“Rossiya – Germaniya” 2013 No 2(6) p 65–69 “Rossiya – Germaniya”, 2013, No 2(6), p. 65–69. REFERENCES 2, 2001, p. 175–197. 26. Savinova Natal'ya, Rossiiskii natsionalizm i nemetskie pogromy v Rossii \nv gody Pervoi mirovoi voiny: 1914-1917 gg. [Russian nationalism and German \npogroms \nin \nRussia \nduring \nthe \nFirst \nWorld \nWar: \n1914-1917], \nhttp://www.dissercat.com/content/rossiiskii-natsionalizm-i-nemetskie-\npogromy-v-rossii-v-gody-pervoi-mirovoi-voiny-1914-1917-g. 26. Savinova Natal'ya, Rossiiskii natsionalizm i nemetskie pogromy v Rossii \nv gody Pervoi mirovoi voiny: 1914-1917 gg. [Russian nationalism and German \npogroms \nin \nRussia \nduring \nthe \nFirst \nWorld \nWar: \n1914-1917], \nhttp://www.dissercat.com/content/rossiiskii-natsionalizm-i-nemetskie-\npogromy-v-rossii-v-gody-pervoi-mirovoi-voiny-1914-1917-g. 27. Schmidt Ute, Bessarabien. Deutsche Kolonisten am Schwarzen Meer, \nPotsdam, Deutsches Kulturforum Östliches Europa, 2008. 27. Schmidt Ute, Bessarabien. Deutsche Kolonisten am Schwarzen Meer, \nPotsdam, Deutsches Kulturforum Östliches Europa, 2008. 28. Schmidt Ute, Bessarabiya. Nemeckie kolonisty Prichernomor'ya: ocherki \n[Bessarabia. German colonists of the Black Sea region: essays], Odessa, Optimum, \n2015. 28. Schmidt Ute, Bessarabiya. Nemeckie kolonisty Prichernomor'ya: ocherki \n[Bessarabia. German colonists of the Black Sea region: essays], Odessa, Optimum, \n2015. German Colonists from Bessarabia in Russian Historical Literature 72 29. Schmidt Ute, Die Deutschen aus Bessarabien. Eine Minderheit aus \nSüdosteuropa (1814 bis heute), 2 durchgesehene Auflage. Köln – Weimar – Wien, \nBöhlau Verlag 2004. 30. Schulz Joachim, Die deutschen Volksgruppen in Südosteuropa, Leipzig – \nBerlin, Teubner, 1940. 30. Schulz Joachim, Die deutschen Volksgruppen in Südosteuropa, Leipzig – \nBerlin, Teubner, 1940. 31. Shcherbinina Ol’ga, Kritika sovremennoi zapadnogeranskoi burzhuaznoi \nistoriografii istorii Moldavii [Criticism of Modern West German Bourgeois \nHistoriography of the History of Moldova], Kishinev, “Shtiintsa”, 1984. 32. Shornikov Igor, Obshchestvenno-politicheskaya i literaturnaya deyatel'nost' \nP. A. Krushevana [Socio-political and literary activities of P. A. Krushevan]. Avtoreferat dissertatsii na soiskanie uchenoi stepeni kandidata istoricheskih \nnauk. Special'nost' 07.00.02 – Otechestvennaya istoriya. Moskva, 2011. 32. Shornikov Igor, Obshchestvenno-politicheskaya i literaturnaya deyatel'nost' \nP. A. Krushevana [Socio-political and literary activities of P. A. Krushevan]. Avtoreferat dissertatsii na soiskanie uchenoi stepeni kandidata istoricheskih \nnauk. Special'nost' 07.00.02 – Otechestvennaya istoriya. Moskva, 2011. 33. Sobolev Ivan, Bor'ba s “nemetskim zasil'em” v Rossii v gody Pervoi miro-\nvoi vojny [The fight against “German dominance” in Russia during the First World \nWar]. Avtoreferat dissertacii na soiskanie uchenoj stepeni kandidata istoricheskih \nnauk. Special'nost': Special'nost': 07.00.02 – Otechestvennaya istoriya. Sankt-\nPeterburg, 1998. 33. Sobolev Ivan, Bor'ba s “nemetskim zasil'em” v Rossii v gody Pervoi miro-\nvoi vojny [The fight against “German dominance” in Russia during the First World \nWar]. Avtoreferat dissertacii na soiskanie uchenoj stepeni kandidata istoricheskih \nnauk. Special'nost': Special'nost': 07.00.02 – Otechestvennaya istoriya. Sankt-\nPeterburg, 1998. 34. Stepanov Vyacheslav, Trudy po etnografii naseleniya Bessarabii XIX – \nnachala XX vv. 38. Ziеbart Alfred, Arzis, Bessarabien. Blick auf 150 Jahre Kolonistenleben \nund –schicksal 1816–1966, Ludwigsburg, Eigenverlag des Verfassers, 1971. REFERENCES Ocherki istorii etnograficheskogo izucheniya bessarabskih ukraintsev \n[Works on the ethnography of the population of Bessarabia in the 19th – early 20th \ncenturies. Essays on the history of the ethnographic study of Bessarabian \nUkrainians], Kishinev, 2001. 35. Suga Alexander, Die völkerrechtliche Lage Bessarabiens in der geschicht-\nlichen Entwicklung des Landes, Bonn, 1958. 36. Sulyak Sergei, Naselenie Bessarabii v XIX v.: chislennost', natsional'nyi \nsostav, religioznaya prinadlezhnost' (Po materialam dorevolyutsionnyh issledo-\nvanii) [The population of Bessarabia in the 19th century: size, ethnic composition, \nreligious affiliation (based on pre-revolutionary studies)], in Trudy Ob’edinennogo \nnauchnogo soveta po gumanitarnym problemam i istoriko-kul’turnomu naslediyu. 2009, Sankt-Peterburg, 2010, p. 54–61. 37. Zelenchuk Valentin, Naselenie Bessarabii i Podnestrov'ya v XIX v. (Etnicheskie i social'no-demograficheskie processy) [The Population of the \nBessarabia and Transnistria in the 19th century (Ethnic and Social-Demographical \nprocesses)], Kishinev, 1979. 38. Ziеbart Alfred, Arzis, Bessarabien. Blick auf 150 Jahre Kolonistenleben \nund –schicksal 1816–1966, Ludwigsburg, Eigenverlag des Verfassers, 1971."
https://openalex.org/W4390748108	https://ojs.itb-ad.ac.id/index.php/JUTECH/article/download/2032/444	Indonesian	null	Sistem Pakar Mendiagnosa Penyakit Gigi Anak Usia Satu Sampai Lima Tahun Menggunakan Metode Certainty Factor	Jutech	2,022	cc-by	3,915	Website: http://ojs.itb-ad.ac.id/index.php/JUTECH E-ISSN: 2797 - 4111 SISTEM PAKAR MENDIAGNOSA PENYAKIT GIGI ANAK USIA SATU SAMPAI LIMA TAHUN MENGGUNAKAN METODE CERTAINTY FACTOR Website: http://ojs.itb-ad.ac.id/index.php/JUTECH E-ISSN: 2797 - 4111 SISTEM PAKAR MENDIAGNOSA PENYAKIT GIGI ANAK USIA SATU SAMPAI LIMA TAHUN MENGGUNAKAN METODE CERTAINTY FACTOR Website: http://ojs.itb-ad.ac.id/index.php/JUTECH E-ISSN: 2797 - 4111 () Korespondensi: fahrulrazi0398@gmail.com (Fahrul Razi) SISTEM PAKAR MENDIAGNOSA PENYAKIT GIGI ANAK USIA SATU SAMPAI LIMA TAHUN MENGGUNAKAN METODE CERTAINTY FACTOR Fahrul Razi1 1ITB Ahmad Dahlan, Jakarta Kata Kunci: Teeth, Expert System, Certainty Factor, PHP Juli – Desember 2022, Vol 3 (2) : hlm 131-140 ©2022 Institut Teknologi dan Bisnis Ahmad Dahlan. All rights reserved. Juli – Desember 2022, Vol 3 (2) : hlm 131-140 ©2022 Institut Teknologi dan Bisnis Ahmad Dahlan. All rights reserved. Abstract Teeth are one of the most important organs as a digestive tool for the human body. Therefore the teeth must be treated as well as possible. If no treatment is taken, the disease will cause the teeth to become infected. Dental diseases in one-to-five-year-olds most often seen are dental caries and gingivitis. Dental Caries can also be called cavities. While Gingivitis is a disease of tartar and gum inflammation. Teeth are one of the most important organs as a digestive tool for the human body. Therefore the teeth must be treated as well as possible. If no treatment is taken, the disease will cause the teeth to become infected. Dental diseases in one-to-five-year-olds most often seen are dental caries and gingivitis. Dental Caries can also be called cavities. While Gingivitis is a disease of tartar and gum inflammation. Expert System is an information system that contains the knowledge of an expert so that it can be used for consultation. In the expert system there are several methods, one of which is Certainty Factor. Cerainty Factor is one technique used to overcome uncertainty in decision making. f f y f dental caries and gingivitis. Dental Caries can also be called cavities. While Gingivitis is a disease of tartar and gum inflammation. Expert System is an information system that contains the knowledge of an expert so that it can be used for consultation. In the expert system there are several methods, one of which is Certainty Factor. Cerainty Factor is one technique used to overcome uncertainty in decision making. f g f Expert System is an information system that contains the knowledge of an expert so that it can be used for consultation. In the expert system there are several methods, one of which is Certainty Factor. Cerainty Factor is one technique used to overcome uncertainty in decision making. Kata Kunci: Teeth, Expert System, Certainty Factor, PHP Jurnal Teknologi Informasi (JUTECH) Juli – Desember 2022, Vol 3 (2) : hlm 131-141 PENDAHULUAN Perkembangan ilmu pengetahuan terutama dibidang teknologi informasi semakin berkembang dan hampir semua bidang menggunakan perangkat komputer sebagai alat pendukung untuk mencari atau menghasilkan informasi (Takariyanti, Batan dan Irawan, 2020). Seiring dengan perkembangan komputer, maka peran dan kegunaan komputer semakin besar bahkan mendominasi kehidupan umat manusia sampai saat ini. Teknik untuk membuat komputer mampu mengolah pengetahuan salah satunya disebut teknik kecerdasan buatan (Nugroho, et al., 2021). Kecerdasan buatan atau Artificial Intelligence (AI) dapat membuat komputer melakukan hal-hal yang dilakukan layaknya seperti manusia. Dalam mengambilan keputusan manusia dapat menjadi komputer sebagai menetapkan keputusan (Sagat dan Purnomo, 2021). Gigi adalah organ tubuh yang berperan penting dalam mengunyah makanan, membantu kelancaran berbicara, dan juga untuk estetika. Oleh karena itu kesehatan gigi sangat penting (Nurlaela, 2013). Beberapa faktor yang menyebabkan timbulnya penyakit gigi dan mulut antara lain yakni kurangnya menjaga kebersihan mulut, adanya jamur, adanya bakteri dan virus (Yuliyana dan Sinaga, 2019). Oleh karena itu, menjaga kesehatan gigi dengan membiasakan diri rajin menggosok gigi penting untuk dilakukan termasuk pergi ke dokter gigi untuk memeriksakan gigi secara berkala setiap enam bulan sekali (Arifin, 2016). Sistem pakar merupakan sistem yang berusaha mengadopsi pengetahuan manusia ke komputer (Hendimi, Pratama dan Misuma, 2019). Sistem ini bisa menyamai atau meniru kemampuan seorang pakar (Hadi dan Diana, 2019). Pakar (Expert) adalah seseorang yang mampu menjelaskan suatu tanggapan, mempelajari hal-hal baru seputar topik permasalahan (domain), menyusun kembali pengetahuan jika dipandang perlu, memecah aturanaturan jika dibutuhkan, dan menentukan relevan tidaknya keahlian mereka (Yuwono. 2010). Sistem pakar mempunyai ciri-ciri diantaranya adanya fasilitas mengenai informasi yang dapat dipercayai, tidak sulit dimodifkasi, penggunaan tidak mengacu pada satu jenis komputer saja, dan memiliki tingkat adaptasi yang baik (Syahputra, Elfitryani dan Maya, 2019). Dengan adanya Sistem Pakar yang akan dirancang, nantinya dapat dijadikan layanan konsultasi untuk dapat membantu dalam pendiagnosaan berdasarkan gejala-gejala klinis yang ada (Rusdiyanto, 2018). Faktor kepastian (certainty factor) diusulkan oleh Shortliffe dan Buchanan pada 1975 untuk mengakomadasikan ketidakpastian pemikiran (inexact reasoning) seorang pakar(Hasibuan, Sunandar dan Suginam, 2017). Certainty Factor ini dapat mengambarkan tingkat keyakinan seorang pakar terhadap masalah yang sedang dihadapi(Aji, Furqon dan Widodo, 2018). Cara kerja metode Certanty Factor ini adalah dengan menunjukkan ukuran kepastian terhadap suatu fakta atau aturan. Metode CF melakukan penalaran layaknya seorang pakar, dan untuk mendapatkan nilai kepercayaan(Santi dan Andari, 2019). Maka akan dilakukan penelitian dalam memanfaatkan kecerdasan buatan dengan menggunakan metode Certainty Factor yang bertujuan untuk mengalisis pemyakit gigi pada anak usia satu sampai lima tahun. 132 METODE Metode dalam penelitian ini disusun melalui sebuah kerangka kerja terleboh dahulu (Ramadhan, 2018). Metode penelitian merupakan prosedur dan teknik penelitian (Ramadhan, 2018). Adapun kerangka kerja penelitian yang digunakan seperti pada gambar 1. Kerangka kerja penelitian Gambar 1. Kerangka Kerja Penelitian Gambar 1. Kerangka Kerja Penelitian Berdasarkan gambar kerangka kerja penelitian di atas dapat dijelaskan sebagai berikut; Berdasarkan gambar kerangka kerja penelitian di atas dapat dijelaskan sebagai berikut; 1. Mengidentifikasi Masalah Sistem Pakar Mendiagnosa Penyakit Gigi Anak Usia Satu Sampai Lima Tahun Menggunakan Metode Certainty Factor (Fahrul Razi) 2. Menganalisa Masalah 2. Menganalisa Masalah Menganalisa masalah ini merupakan suatu langkah yang dilakukan untuk alur selanjutnya dari penelitian, sehingga dalam melakukan penelitian dapat memahami permasalahan yang ada. Menganalisa masalah ini merupakan suatu langkah yang dilakukan untuk alur selanjutnya dari penelitian, sehingga dalam melakukan penelitian dapat memahami permasalahan yang ada. 3. Menentukan Tujuan Tujuan penelitian dapat dikatakan suatu hasil yang diharapkan ketika penelitian telah selesai dikerjakan, dan sesuatu yang akan dicapai dalam melakukan penelitian 5. Pengumpulan Data dan Informasi 5. Pengumpulan Data dan Informasi Tujuan dari pengumpulan data adalah untuk mendapatkan suatu informasi dari data-data yang dibutuhkan untuk penelitian agar mencapai tujuan yang diharapkan. Pengumpulan data dengan melakukan wawancara langsung dengan pakar mengenai hal-hal yang berhubungan dengan objek penelitian. Pengumpulan data ini merupakan sebuah teknik yang dilakukan untuk memperoleh data yang ideal, dilakukan secara langsung melalui penelitian lapangan. 6. Penyelesaian Masalah Dengan Menggunakan Certainty Factor Dilakukan perancangan dan analisa pada permasalahan yang ada sesuai dengan data-data yang sudah terkumpul dengan langkah-langkah yang terdapat pada metode Certainty Factor, sehingga permasalahan-permasalahan tersebut dapat menemukan solusinya. 6. Penyelesaian Masalah Dengan Menggunakan Certainty Factor Dilakukan perancangan dan analisa pada permasalahan yang ada sesuai dengan data-data yang sudah terkumpul dengan langkah-langkah yang terdapat pada metode Certainty Factor, sehingga permasalahan-permasalahan tersebut dapat menemukan solusinya. Certainty Factor didefinisikan sebagai persamaan berikut : 3. Menentukan Tujuan 3. Menentukan Tujuan Tujuan penelitian dapat dikatakan suatu hasil yang diharapkan ketika penelitian telah selesai dikerjakan, dan sesuatu yang akan dicapai dalam melakukan penelitian 3. Menentukan Tujuan Tujuan penelitian dapat dikatakan suatu hasil yang diharapkan ketika penelitian telah selesai dikerjakan, dan sesuatu yang akan dicapai dalam melakukan penelitian Tujuan penelitian dapat dikatakan suatu hasil yang diharapkan ketika penelitian telah selesai dikerjakan, dan sesuatu yang akan dicapai dalam melakukan penelitian Jurnal Teknologi Informasi (JUTECH) Juli – Desember 2022, Vol 3 (2) : hlm 131-140 4. Mempelajari Literatur 4 p j Berdasarkan tujuan yang ingin dicapai, maka dibutuhkan kajian literatur sesuai dengan tujuan penelitian. Sumber dari literatur ini dapat berasal dari artikel, jurnal ilmiah mengenai metode Certainty Factor, serta referensi-referensi yang berhubungan dengan penelitian. 4 p j Berdasarkan tujuan yang ingin dicapai, maka dibutuhkan kajian literatur sesuai dengan tujuan penelitian. Sumber dari literatur ini dapat berasal dari artikel, jurnal ilmiah mengenai metode Certainty Factor, serta referensi-referensi yang berhubungan dengan penelitian. Berdasarkan tujuan yang ingin dicapai, maka dibutuhkan kajian literatur sesuai dengan tujuan penelitian. Sumber dari literatur ini dapat berasal dari artikel, jurnal ilmiah mengenai metode Certainty Factor, serta referensi-referensi yang berhubungan dengan penelitian. 1. Mengidentifikasi Masalah Untuk memberikan penjelasan mengenai penelitian yang akan diteliti, temtukan terlebih dahulu masalah yang akan diteliti. Menentukan masalah pada penelitian merupakan langkah pertama yang harus dilakukan. Sistem Pakar Mendiagnosa Penyakit Gigi Anak Usia Satu Sampai Lima Tahun Menggunakan Metode Certainty Factor (Fahrul Razi) 133 2. Menganalisa Masalah Menganalisa masalah ini merupakan suatu langkah yang dilakukan untuk alur selanjutnya dari penelitian, sehingga dalam melakukan penelitian dapat memahami permasalahan yang ada. 2. Menganalisa Masalah Menganalisa masalah ini merupakan suatu langkah yang dilakukan untuk alur selanjutnya dari penelitian, sehingga dalam melakukan penelitian dapat memahami permasalahan yang ada. 8. Hasil dan Pembahasan. 8. Hasil dan Pembahasan. CF (E, e) = CF (H, E) Dalam aplikasinya, CF(H,E) merupakan nilai kepastian yang diberikan oleh pakar terhadap suatu aturan, sedangkan CF(E,e) merupakan nilai kerpercayaan yang diberikan oleh pengguna terhadap gejala yang dialaminya Dalam aplikasinya, CF(H,E) merupakan nilai kepastian yang diberikan oleh pakar terhadap suatu aturan, sedangkan CF(E,e) merupakan nilai kerpercayaan yang diberikan oleh pengguna terhadap gejala yang dialaminya 7. Pengujian Sistem Tahap ini dilakukan untuk memberikan kepastian terhadap Sistem Pakar yang sudah dibangun apakah sudah sampai pada tujuan yang diharapkan serta sudahkah bisa dilakukan implementasi terhadap pengguna. CF (H, e) = CF (E, e) CF (H, E) CF (H, e) = CF (E, e) * CF (H, E) CF (H, e) = CF (E, e) * CF (H, E) Dimana CF (H, e) adalah certainty factor hipotesis yang dipengaruhi oleh evidence, dan CF (E, e) adalah certainty factor evidence E yang dipengaruhi oleh evidence e, dan CF (H, E) adalah certainty factor hipotesis dengan asumsi evidence diketahui denganpasti,yaituketikaCF(E,e)=1. Jika semua evidence pada antecedent diketahui dengan pasti maka persamaannya akan menjadi: CF (E, e) = CF (H, E) CF (H, E) = MB (H, E) – MD (H, E) Dimana CF (H, E) : Certainty Factor dari hipotesis H yang dipengaruhi oleh gejala (evidence) E. Besarnya CF berkisar antara -1 sampai 1. Nilai -1 menunjukkan ketidakpercayaan mutlak sedangkan nilai 1 menunjukkan kepercayaan mutlak, dan MB (H, E) : ukuran kenaikan kepercayaan (measure of increased belief) terhadap hipotesis H yang dipengaruhi oleh gejala E, dan MD (H, E) : ukuran kenaikan ketidakpercayaan (measure of increased disbelief) terhadap hipotesis H yang dipengaruhi oleh gejala E.Bentuk dasar rumus certainty factor, adalah sebuah aturan JIKA E MAKA H seperti ditunjukkan oleh persamaan 2 berikut: 134 7. Pengujian Sistem 7. Pengujian Sistem Tahap ini dilakukan untuk memberikan kepastian terhadap Sistem Pakar yang sudah dibangun apakah sudah sampai pada tujuan yang diharapkan serta sudahkah bisa dilakukan implementasi terhadap pengguna. 8. Hasil dan Pembahasan. Pada tahap hasil, akan diperlihatkan suatu hal yang sukses dikerjakan oleh sistem yang dibangun. Sedangkan pada tahap pembahasan, akan diperlihatkan apakah hasil yang diperoleh sistem cocok dengan hasil yang diinginkan dari analisa yang sudah dilakukan. 8. Hasil dan Pembahasan. Pada tahap hasil, akan diperlihatkan suatu hal yang sukses dikerjakan oleh sistem yang dibangun. Sedangkan pada tahap pembahasan, akan diperlihatkan apakah hasil yang diperoleh sistem cocok dengan hasil yang diinginkan dari analisa yang sudah dilakukan. HASIL DAN PEMBAHASAN Data yang diperoleh dari hasil wawancara dengan pakar berupa data penyakit, gejala, serta nilai probabilitas masing-masing gejala sesuai dengan penyakit yang bersangkutan. Beberapa jenis penyakit gigi dapat dilihat pada Tabel 1 berikut ini. Sistem Pakar Mendiagnosa Penyakit Gigi Anak Usia Satu Sampai Lima Tahun Menggunakan Metode Certainty Factor (Fahrul Razi) 135 g p j p y g g p p Tabel 1.Penyakit Gigi Product Penyakit P01 P02 P03 Caries Superficialis Caries Median Caries Profunda P04 P05 P06 Caries Pulpilis Caries Mati Gingivitis Tabel 1.Penyakit Gigi Product Penyakit P01 P02 P03 Caries Superficialis Caries Median Caries Profunda P04 P05 P06 Caries Pulpilis Caries Mati Gingivitis Sistem Pakar Mendiagnosa Penyakit Gigi Anak Usia Satu Sampai Lima Tahun Menggunakan Metode Certainty Factor (Fahrul Razi) 1 Selanjutnya akan ditunjukkan beberapa gejala penyakit gigi seperti yang terlihat pada Tabel 2 berikut ini. Tabel 2.Gejala Penyakit Gigi Kode Gejala Penyakit G1 Gigi berlubang mengenai email gigi G2 Gigi ngilu cepat hilang ketika makan makanan dingin atau panas G3 Gigi berlubang mengenai dentin gigi G4 G5 G6 G7 G8 G9 G10 G11 G12 Gigi ngilu dengan rasa nyeri hilang lama ketika makan makanan dingin atau panas Gigi sakit berdenyut Gigi berwarna kecoklatan atau kehitaman Gigi berlubang dalam Gusi bengkak Gusi berwarna merah Gusi berdarah ketika makan atau gosok gigi Terdapat karang gigi Gigi berwarna kuning kecoklatan Berikut adalah nilai probabilitas dari penyakit gigi beserta rule yang telah di dapat dari pakar. Berikut adalah nilai probabilitas dari penyakit gigi beserta rule yang telah di dapat dari pakar. R1: IF gigi ngilu hilang dengan cepat AND gigi berlubang mengenai email gigi THEN Caries Superficialis (CF: 0.7). R2: IF gigi ngilu dengan rasa nyeri hilang lama AND gigi berlubang mengenai dentin gigi THEN Caries Median (CF: 0.75). R2: IF gigi ngilu dengan rasa nyeri hilang lama AND gigi berlubang mengenai dentin gigi THEN Caries Median (CF: 0.75). R3: IF gigi sakit berdenyut AND gigi berwarna kecoklatan atau kehitaman THEN Caries Profunda (CF: 0.8). R3: IF gigi sakit berdenyut AND gigi berwarna kecoklatan atau kehitaman THEN Caries Profunda (CF: 0.8). R4: IF gigi sakit berdenyut AND gigi berlubang dalam AND gigi berwarna kecoklatan atau kehitaman THEN Caries Pulpilis (CF: 0.9). R4: IF gigi sakit berdenyut AND gigi berlubang dalam AND gigi berwarna kecoklatan atau kehitaman THEN Caries Pulpilis (CF: 0.9). Jurnal Teknologi Informasi (JUTECH) Juli – Desember 2022, Vol 3 (2) : hlm 131-140 HASIL DAN PEMBAHASAN R5: IF gigi sakit berdenyut AND gigi berlubang dalam AND gigi berwarna kecoklatan atau kehitaman AND gusi bengkak AND gusi berwarna merah THEN Caries Mati (CF: 0.9). R6: IF terdapat karang gigi AND gusi berwarna merah AND gusi bengkak AND gigi berwarna kekuningan kecoklatan AND gusi berdarah THEN Gingivitis (CF: 0.9). R7: IF gigi sakit berdenyut AND gigi berlubang dalam THEN Caries Pulpilis (CF: 0.8). 136 R8: IF gigi sakit berdenyut AND gigi berlubang dalam AND gigi berwarna kecoklatan atau kehitaman THEN Caries Mati (CF: 0.8). R9: IF terdapat karang gigi AND gusi berwarna merah AND gusi berdarah THEN Gingivitis (CF: 0.7) Contoh kasus dari penyakit mulut pada kucing dengan menjawab beberapa pertanyaan, dan di dapatkan gejala-gejala dengan penyelesaian : Jurnal Teknologi Informasi (JUTECH) Juli – Desember 2022, Vol 3 (2) : hlm 131-140 138 Sistem Pakar Mendiagnosa Penyakit Gigi Anak Usia Satu Sampai Lima Tahun Menggunakan Metode Certainty Factor (Fahrul Razi) 137 p y , p g j g j g p y Rule 1 = IF G1 (CF=0.2) AND G2 (CF=0.1) THEN P01 (CF=0.7) CF1 (P01, G1 ∩ G2 ) = Min[0.2 ; 0.1] * 0.7 = 0.07 Fakta Baru: P01 Hipotesis CF= 0.07 Rule 2 = IF G3 (CF=0.4) AND G4 (CF=0.3) THEN P02 (CF=0.75) CF2 (P02, G3 ∩ G4 ) = Min[0.4 ; 0.3] * 0.75 = 0.225 Fakta Baru: P02 Hipotesis CF= 0.225 Rule 3 = IF G5 (CF=0.86) AND G6 (CF=0.65) THEN P03 (CF=0.8) CF3 (P03, G5 ∩ G6 ) = Min[0.86 ; 0.65] * 0.8 = 0.52 Fakta Baru: P03 Hipotesis CF= 0.52 Rule 4 = IF G5 (CF=0.86) AND G7 (CF=0.83) AND G6 (CF=0.65) THEN P04 (CF=0.9) CF4 (P04, G5 ∩ G7 ∩ G6 ) = Min[0.86 ; 0.83 ; 0.65] * 0.9 = 0.585 Fakta Baru: P04 Hipotesis CF= 0.585 Rule 5 = IF G5 (CF=0.86) AND G7 (CF=0.83) AND G6 (CF=0.65) AND G8 (CF=0.4) AND G9 (CF=0.66) THEN P05 (CF=0.9) CF5 (P05, G5 ∩ G7 ∩ G6 ∩ G8 ∩ G9) = Min[0.86 ; 0.83 ; 0.65 ; 0.4 ; 0.66] * 0.9 = 0.36 Fakta Baru: P05 Hipotesis CF= 0.36 Rule 1 = IF G1 (CF=0.2) AND G2 (CF=0.1) THEN P01 (CF=0.7) CF1 (P01, G1 ∩ G2 ) = Min[0.2 ; 0.1] * 0.7 = 0.07 Fakta Baru: P01 Hipotesis CF= 0.07 Rule 2 = IF G3 (CF=0.4) AND G4 (CF=0.3) THEN P02 (CF=0.75) CF2 (P02, G3 ∩ G4 ) = Min[0.4 ; 0.3] * 0.75 = 0.225 Fakta Baru: P02 Hipotesis CF= 0.225 Rule 3 = IF G5 (CF=0.86) AND G6 (CF=0.65) THEN P03 (CF=0.8) CF3 (P03, G5 ∩ G6 ) = Min[0.86 ; 0.65] * 0.8 = 0.52 Fakta Baru: P03 Hipotesis CF= 0.52 Rule 4 = IF G5 (CF=0.86) AND G7 (CF=0.83) AND G6 (CF=0.65) THEN P04 (CF=0.9) CF4 (P04, G5 ∩ G7 ∩ G6 ) = Min[0.86 ; 0.83 ; 0.65] * 0.9 = 0.585 Fakta Baru: P04 Hipotesis CF= 0.585 Sistem Pakar Mendiagnosa Penyakit Gigi Anak Usia Satu Sampai Lima Tahun Menggunakan Metode Certainty Factor (Fahrul Razi) 137 Rule 5 = IF G5 (CF=0.86) AND G7 (CF=0.83) AND G6 (CF=0.65) AND G8 (CF=0.4) AND G9 (CF=0.66) THEN P05 (CF=0.9) CF5 (P05, G5 ∩ G7 ∩ G6 ∩ G8 ∩ G9) = Min[0.86 ; 0.83 ; 0.65 ; 0.4 ; 0.66] * 0.9 = 0.36 Fakta Baru: P05 Hipotesis CF= 0.36 Sistem Pakar Mendiagnosa Penyakit Gigi Anak Usia Satu Sampai Lima Tahun Menggunakan Metode Certainty Factor (Fahrul Razi) P05 Hipotesis CF 0.36 137 ( ) 9 ( ) ( 4) ( 3) AND G10 (CF=0.2) THEN P06 (CF=0.9) CF6 (P06, G11 ∩ G9 ∩ G8 ∩ G12 ∩ G10) = Min[0.1 ; 0.66 ; 0.4 ; 0.3 ; 0.2 ] * 0.9 = 0.18 Fakta Baru: P06 Hipotesis CF= 0.18 ( ) 9 ( ) ( 4) ( 3) AND G10 (CF=0.2) THEN P06 (CF=0.9) CF6 (P06, G11 ∩ G9 ∩ G8 ∩ G12 ∩ G10) = Min[0.1 ; 0.66 ; 0.4 ; 0.3 ; 0.2 ] * 0.9 = 0.18 Fakta Baru: P06 Hipotesis CF= 0.18 Rule 7 = IF G5 (CF=0.86) AND G7 (CF=0.83) THEN P04 (CF=0.8) CF7 (P04, G5 ∩ G7) = Min[0.86 ; 0.83] * 0.8 = 0.664 Fakta Baru: P04 Hipotesis CF= 0.664 Rule 8 = IF G5 (CF=0.86) AND G7 (CF=0.83) AND G6 (CF=0.65) THEN P05 (CF=0.8) CF8 (P05, G5 ∩ G7 ∩ G6 ) = Min[0.86 ; 0.83 ; 0.65] * 0.8 = 0.52 Fakta Baru: P05 Hipotesis CF= 0.52 Rule 9 = IF G11 (CF=0.1) AND G9 (CF=0.66) AND G10 (CF=0.2) THEN P06 (CF=0.7) CF9 (P06, G11 ∩ G9 ∩ G10 ) = Min[0.1 ; 0.66 ; 0.2] * 0.7 = 0.07 Fakta Baru: P06 Hipotesis CF= 0.07 CF Gabungan dari Rule: - Gabungan R4 dan R7 CF = CF4 + CF7 * (1 - CF4) = 0.585 + 0.664 * (1 - 0.585) 0.415 = 0.86056 138 - Gabungan R6 dan R9 CF = CF6 + CF9 * (1 – CF6) = 0.18 + 0.07 * (1 – 0.18) = 0.2374 - Gabungan R6 dan R9 CF = CF6 + CF9 * (1 – CF6) = 0.18 + 0.07 * (1 – 0.18) = 0.2374 KESIMPULAN Setelah melalui proses perancangan aplikasi sistem pakar mendiagnosis penyakit mulut dan gigi pada manusia yang dirancang dengan menggunakan pemograman PHP dan MySQL serta melakukan proses pengujian dan analisisnya maka dapat ditarik kesimpulan secara umum bahwa aplikasi ini mampu mengetahui nilai kepastian dari penyakit yang dialami oleh user lebih cepat dan efektif. Berdasarkan identifikasi masalah yang tercantum dalam latar belakang masalah, berikut ini adalah beberapa kesimpulannya: p y g dalam latar belakang masalah, berikut ini adalah beberapa kesimpulannya: 1. Sistm pakar tentang penyakit gigi pada anak usia satu sampai lima tahun ini dapat mengetahui jenis penyakit yang dialami oleh user. 1. Sistm pakar tentang penyakit gigi pada anak usia satu sampai lima tahun ini dapat mengetahui jenis penyakit yang dialami oleh user. 2. Sistem dapat memberi informasi yang bermanfaat bagi pakar dan masyarakat tentang penyakit gigi pada anak usia satu sampai lima tahun 2. Sistem dapat memberi informasi yang bermanfaat bagi pakar dan masyarakat tentang penyakit gigi pada anak usia satu sampai lima tahun Kesimpulan: Pasien Menderita penyakit Caries Pulpilis dengan tingkat kepastian 86% Kesimpulan: Pasien Menderita penyakit Caries Pulpilis dengan tingkat kepastia 86% Sistem Pakar Mendiagnosa Penyakit Gigi Anak Usia Satu Sampai Lima Tahun Menggunakan Metode Certainty Factor (Fahrul Razi) Fakta Baru: P04 Hipotesis CF= 0.86056 * 100 % = 86.056% P05 Hipotesis CF= 0.6928 * 100 % = 69.28% P06 Hipotesis CF= 0.2374 * 100 % = 23.74 % Kesimpulan: Pasien Menderita penyakit Caries Pulpilis dengan tingkat kepastian P04 Hipotesis CF= 0.86056 * 100 % = 86.056% P05 Hipotesis CF= 0.6928 * 100 % = 69.28% P06 Hipotesis CF= 0.2374 * 100 % = 23.74 % P04 Hipotesis CF= 0.86056 * 100 % = 86.056% P05 Hipotesis CF= 0.6928 * 100 % = 69.28% P06 Hipotesis CF= 0.2374 * 100 % = 23.74 % Kesimpulan: Pasien Menderita penyakit Caries Pulpilis dengan tingkat kepastian 86% esimpulan: Pasien Menderita penyakit Caries Pulpilis dengan tingkat kepastian DAFTAR PUSTAKA Takariyanti, D. N. R., Batan, I. W., & Erawan, I. G. M. K. Laporan Kasus: Rhinitis Unilateral pada Kucing Lokal yang Mengalami Langit-langit Mulut Bercelah (Cleft Palate). Takariyanti, D. N. R., Batan, I. W., & Erawan, I. G. M. K. Laporan Kasus: Rhinitis Unilateral pada Kucing Lokal yang Mengalami Langit-langit Mulut Bercelah (Cleft Palate). Nugroho, F. A., Solikin, A. F., Anggraini, M. D., & Kusrini, K. (2021). Sistem Pakar Diagnosa Virus Corona Dengan Metode Naïve Bayes. Jurnal Teknologi Informasi dan Komunikasi (TIKomSiN), 9(1), 81-88. Nugroho, F. A., Solikin, A. F., Anggraini, M. D., & Kusrini, K. (2021). Sistem Pakar Diagnosa Virus Corona Dengan Metode Naïve Bayes. Jurnal Teknologi Informasi dan Komunikasi (TIKomSiN), 9(1), 81-88. Sagat, N. A., & Purnomo, A. S. (2021). Sistem Pakar Diagnosa Penyakit Mata Menggunakan Metode Teorema Bayes. Jurnal Pendidikan dan Teknologi Indonesia, 1(8), 329-337. Pratama, E. B., & Hendini, A. (2019). Sistem Pakar Diagnosa Penyakit Kulit Pada Kucing Berbasis Web Menggunakan Metode Decision Tree. Sistemasi: Jurnal Sistem Informasi, 8(2), 254-264. Hadi, F., & Diana, Y. (2019). Sistem Pakar Diagnosa Penyakit Gigi dengan Metode Bayes. Sains dan Teknologi Informasi, 5(2), 44-51. 139 Syahputra, T., & Maya, W. R. (2019). Implementasi Sistem Pakar Untuk Mengidentifikasi Pecandu Narkoba Menggunakan Metode Teorema Bayes. Jurnal SAINTIKOM (Jurnal Sains Manajemen Informatika dan Komputer), 18(2), 111-118. Rusdiyanto, R. (2018). Sistem Pakar Diagnosa Penyakit Akibat Gigitan Nyamuk Dengan Metode Bayes Berbasis Web. JUTIM (Jurnal Teknik Informatika Musirawas), 3(1), 30-40. Yuwono, B. (2015, July). Pengembangan sistem pakar pada perangkat mobile untuk mendiagnosa penyakit gigi. In Seminar Nasional Informatika (SEMNASIF) (Vol. 1, No. 4). Arifin, J. (2016). Sistem Pakar Diagnosa Penyakit Gigi dan MulutManusia Menggunakan Knowledge Base System dan Certainty Factor. Jurnal Ilmiah Teknologi Informasi Asia, 10(2), 50-64. Yuliyana, Y., & Sinaga, A. S. R. M. (2019). Sistem Pakar Diagnosa Penyakit Gigi Menggunakan Metode Naive Bayes. Fountain of Informatics Journal, 4(1), 19-23. Yuliyana, Y., & Sinaga, A. S. R. M. (2019). Sistem Pakar Diagnosa Penyakit Gigi Menggunakan Metode Naive Bayes. Fountain of Informatics Journal, 4(1), 19-23. Nurlaela, F. (2013). Sistem Pakar untuk mendeteksi penyakit gigi pada manusia. Speed-Sentra Penelitian Engineering dan Edukasi, 5(4). Aji, A. H., Furqon, M. T., & Widodo, A. W. (2017). Sistem Pakar Diagnosa Penyakit Ibu Hamil Menggunakan Metode Certainty Factor (CF). Jurnal Pengembangan Teknologi Informasi Dan Ilmu Komputer, ISSN, 2(5), 27-36. Santi, I. H., & Andari, B. (2019). Sistem Pakar Untuk Mengidentifikasi Jenis Kulit Wajah dengan Metode Certainty Factor. Jurnal Teknologi Informasi (JUTECH) Juli – Desember 2022, Vol 3 (2) : hlm 131-140 DAFTAR PUSTAKA INTENSIF: Jurnal Ilmiah Penelitian dan Penerapan Teknologi Sistem Informasi, 3(2), 159-177. Hasibuan, N. A., Sunandar, H., Alas, S., & Suginam, S. (2017). Sistem Pakar Mendiagnosa Penyakit Kaki Gajah Menggunakan Metode Certainty Factor. Jurasik (Jurnal Riset Sistem Informasi Dan Teknik Informatika), 2(1), 29-39. Ramadhan, P. S. (2018). Sistem Pakar Pendiagnosaan Dermatitis Imun Menggunakan Teorema Bayes. InfoTekJar: Jurnal Nasional Informatika dan Teknologi Jaringan, 3(1), 43-48 Ramadhan, P. S. (2018). Sistem Pakar Pendeteksian Varicella Simplex Dengan Menggunakan Teorema Bayes. JURIKOM (Jurnal Riset Komputer), 5(5), 454-459. 140
https://openalex.org/W2587103529	https://biblio.ugent.be/publication/8550368/file/8550373	English	null	Bonding in B 2 and B 2 + : Insights from full configuration interaction and valence bond studies	Computational and theoretical chemistry	2,017	cc-by	2,860	Zahid Rashid a,b, Joop H. van Lenthe c, Remco W.A. Havenith d,e,⇑ a Institute for Advanced Study, Shenzhen University, China or Advanced Study, Shenzhen University, China atory of Optoelectronic Devices, Systems of Ministry of Education Guangdong Province, College of Optoelectronics Engineering, Shenzhen Institute for Advanced Study, Shenzhen University, China b Key Laboratory of Optoelectronic Devices, Systems of Ministry of Education Guangdong Province, College of Optoelectronics Engineering, Shenzhen University, 518060 Sh h Chi b Key Laboratory of Optoelectronic Devices, Systems of Ministry of Education Guangdong Province, College of Optoelectronics Engineering, Shenzhen University, 518060 Shenzhen China c Theoretical Chemistry Group, Department of Chemistry, Debye Institute For Nanomaterials Science, Utrecht University, Princetonplein 1, 3584 CC Utrecht, The Netherlands d Zernike Institute for Advanced Materials and Stratingh Institute for Chemistry, University of Groningen, Nijenborgh 4, 9747 AG Groningen, The Netherlands e Ghent Quantum Chemistry Group, Department of Inorganic and Physical Chemistry, Ghent University, Krijgslaan 281 (S3), B-9000 Gent, Belgium 1. Introduction electronic conﬁguration would be 2rg 22ru 21pu 2 and it would have a triplet ground state. The bond order according to this model would be 1, as the r electrons would not contribute to the bonding. Only the two p electrons in the pu orbitals would contribute to the bonding. This situation is similar to the C2 case, except that in B2 the pu orbitals are half ﬁlled, leading to an even weaker p bond than in the C2 case. The description of bonding in small molecules remains a chal- lenge: although for simple diatomic molecules chemical bonding seems clear, the situation in C2 showed that it can still lead to sur- prises [1–7]. Several theoretical studies, at various levels of theory, have been devoted to describe the chemical bonding in the C2 molecule. There is, however, still an on-going debate within the chemical community regarding the bond-order in this molecule. Considering the carbon-carbon bond length and the relaxed force constant, C2 ﬁts in between the molecules containing double and triple CAC bonds, i.e., ethene and ethyne [4,8]. On the other hand, analyses based on full conﬁguration interaction and valence bond calculations [2] and that based on natural orbital functional theory (NOFT) [9] have shown that C2 and its isoelectronic species (CN+, BN and CB!) all possess a quadruple bond: a normal r bond, two p bonds and one long, weaker, r bond. In this report, we study the nature of the bonding in B2 using full conﬁguration interaction calculations and valence bond calcu- lations, to see if there are similarities in the r bonding situation to C2. We also study the bonding in the ionized B2 + and B2 2+ species, to elucidate the nature of the chemical bonding with only one or no bonding p electrons. Article history: Full Conﬁguration Interaction (Full-CI) and Valence Bond Self-Consistent Field (VBSCF) methods have been used to study the electronic structure and bonding in B2 and B2 + molecules. The bonding analysis based on these calculations shows that the B2 molecule is stabilised due to the formation of a double r bond, one strong r-bond together with one second weaker r-bond, and two weak p bonds. Upon ion- ization one p electron is removed from the system and B2 + is formed, which has a one electron r bond, instead of a p bond. It has been shown that a few carefully chosen VB conﬁgurations are enough to describe the bonding; with these structures, geometrical parameters as well as dissociation energies of these unusual molecular species are in agreement with full-CI results. Keywords: B2 B2 + Valence bond theory Full CI Ab initio calculations ! 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http:// creativecommons.org/licenses/by/4.0/). ⇑Corresponding author at: Zernike Institute for Advanced Materials and Strat- ingh Institute for Chemistry, University of Groningen, Nijenborgh 4, 9747 AG Groningen, The Netherlands. http://dx.doi.org/10.1016/j.comptc.2017.02.001 2210-271X/! 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Computational and Theoretical Chemistry 1116 (2017) 92–95 Computational and Theoretical Chemistry 1116 (2017) 92–95 Contents lists available at ScienceDirect a b s t r a c t Article history: Received 14 December 2016 Received in revised form 1 February 2017 Accepted 2 February 2017 Available online 6 February 2017 Article history: Received 14 December 2016 Received in revised form 1 February 2017 Accepted 2 February 2017 Available online 6 February 2017 Keywords: B2 B2 + Valence bond theory Full CI Ab initio calculations E-mail address: r.w.a.havenith@rug.nl (R.W.A. Havenith). 3.2. Valence bond calculations on B2 and B2 + At ﬁrst, a valence bond calculation was performed with the bonding situation with two electrons in each boron-2s orbital and one electron in the px and one electron in the py orbitals (VB-1, Fig. 1). In total, six structures were used. The energy as a function of the B-B distance is plotted in Fig. 2 (VB-1). As can be seen, no minima can be found using these structures, and the p- bonds alone in the bonding pu orbitals are not capable of keeping the atoms together: the nonbonding electrons repel too much to form a stable molecule. Fig. 2. Dissociation curves for B2, calculated using different valence bond models and full-CI (see text). the calculated dissociation energy is 56.65 kcal/mol, which is in very good agreement with the full CI dissociation energy. The equi- librium bond distance is 1.595 Å. At this bond distance, all 26 con- ﬁgurations contribute to the wavefunction. The most important are those, which describe two r bonds between the two boron atoms. The combined weight of these conﬁgurations is 38%. The second in importance are those, which describe two doubly occupied orbi- tals; their accumulated weight in the wavefunction is 34%. The remaining 28% consists of those conﬁgurations, which have one doubly occupied valence orbital on one of the two boron atoms. The general bonding picture of this more accurate wavefunction is similar to the simpler VB-2 description. A minimum in the potential energy curve (VB-2, Fig. 2) is obtained when structures are added that describe two r bonds between the two boron atoms (VB-2, Fig. 1). Using this (12- conﬁguration) VB wavefunction, the calculated dissociation energy is 44.9 kcal/mol, which is only in fair agreement with the full CI dissociation energy (Table 1 and Fig. 2). Correlation in B2 is appar- ently extremely important to describe the bonding situation accu- rately. The atomic orbitals involved in these r bonds are the s and pz orbitals of each boron atom (Fig. 3). One strong bond is formed between VB orbitals (1) and (8) with an overlap of 0.77, and one weaker bond is formed between VB orbitals (2) and (7) with an overlap of only 0.09. The VB orbitals (3), (4), (5), and (6) form two p bonds; the overlap between the p orbitals is 0.40. Using the delocal model, the dissociation curves have been cal- culated. 2. Computational details Another molecule that may show similar behaviour, as C2, is the B2 molecule. According to a molecular orbital scheme, the valence In all calculations, the cc-pVDZ basis set was used. The full CI calculations were performed using GAMESS-UK [10]. Both 1s orbi- tals were kept frozen. Initial orbitals were taken from a preceding full valence MCSCF calculation. The valence bond (VBSCF [11,12]) calculations were performed with TURTLE [13] as implemented in GAMESS-UK. Strictly atomic VB calculations were performed where the orbitals were kept http://dx.doi.org/10.1016/j.comptc.2017.02.001 2210-271X/! 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Z. Rashid et al. / Computational and Theoretical Chemistry 1116 (2017) 92–95 93 Fig. 1. Different structures considered in the valence bond calculations. strictly atomic, and delocal VB calculations were performed in which the orbitals were allowed to delocalise [14]. 3.1. Full conﬁguration interaction calculations on B2, B2 + and B2 2+ According to the full CI calculations, the ground state of B2 is a triplet with a leading valence electronic conﬁguration 2rg 22ru 21pu, x 11pu,y 1 . At the equilibrium distance of 1.632 Å (Table 1), the weight of this conﬁguration is 76.6%. The next leading conﬁguration with a weight of 9.5% is 2rg 21pu,x 1 1pu,y 1 3rg 2, in which the two electrons from the antibonding 2ru orbital are excited to the bonding 3rg orbital. Fig. 1. Different structures considered in the valence bond calculations. Ionization of B2 leads to a wavefunction with a 84.0% weight for the 2rg 22ru 23rg 1 conﬁguration. Note that upon ionization, the low- est conﬁguration has no p electrons anymore, and one electron occupies the bonding 3rg orbital. As expected, the dissociation energy becomes lower, and the bond distance increases with respect to that of neutral B2 (Table 1). Further ionization to B2 2+ does not lead to a bonded species, and spontaneous dissociation occurs. Fig. 2. Dissociation curves for B2, calculated using different valence bond models and full-CI (see text). The full CI calculations are indicative that the bonding situation in B2 is unusual and shows agreement with the bonding situation in for example C2. A closer analysis of the bonding using valence bond calculations is presented in the next section. 3.2. Valence bond calculations on B2 and B2 + The dissociation energy obtained using the 12- conﬁgurations wavefunction (VB-4, Fig. 2) is 52.9 kcal/mol, which is substantially better than the dissociation energy obtained using the strictly atomic VB-2 model with the same number of conﬁgu- rations. A similar bonding picture is obtained: the wavefunction consists mainly of only one structure describing one strong r bond between the valence bond orbitals (5d) and (6d) (Fig. 4), with an overlap of 0.86, and another weaker r bond between the orbitals (1d) and (2d) (Fig. 4). These orbitals have an overlap of 0.29. Note that the overlap has increased signiﬁcantly in comparison to the overlap obtained with the strictly atomic model. The two pu orbi- tals are both occupied with one electron each. To improve the quality of the VB wavefunction in VB-local, four- teen more conﬁgurations were included in the calculation (VB-3 in Fig. 2). Ten of these have one doubly occupied valence orbital (either orbital 2s or 2pz) on either of the two B atoms and four con- ﬁgurations have two doubly occupied orbitals (2s on one atom and 2pz on the other atom). With this (26-conﬁguration) wavefunction, Table 1 Spin- coupled VB will also give similar results as it is in essence the same as VBSCF, but the latter method is more general. librium bond distance calculated using this wavefunction is 1.606 Å, which is also in excellent agreement with that calculated using full CI. This wavefunction gives a very accurate description, in line with the full-CI results, but the main bonding features are already captured by the simpler 12-conﬁguration VB wavefucn- tion. For comparison, a full valence CASSCF calculation (CASSCF (6,8)) yields a dissociation energy of 59.74 kcal/mol. Smaller active spaces are not well-suited to describe this dissociation. Spin- coupled VB will also give similar results as it is in essence the same as VBSCF, but the latter method is more general. Table 1 Table 1 The full CI dissociation energies (kcal/mol), equilibrium distance (Å), and total energy (Hartree) for B2, and B2 +. Molecule Edis R Etot B2 59.57 1.632 !49.274884 B2 + 42.63 2.195 !48.951664 The dissociation energy calculated using 26-conﬁguration wavefunction (VB-5, Fig. 2) in VB-delocal is 59.60 kcal/mol, which is very close to that obtained with the full CI calculations. The equi- Z. Rashid et al. / Computational and Theoretical Chemistry 1116 (2017) 92–95 Fig. 3. The strictly atomic valence bond orbitals involved in the bonding. Fig. 4. The delocal valence bond orbitals involved in the bonding of B2. Z. Rashid et al. / Computational and Theoretical Chemistry 1116 (2017) 92–95 94 Fig. 3. The strictly atomic valence bond orbitals involved in the bonding. Fig. 3. The strictly atomic valence bond orbitals involved in the bonding. Fig. 3. The strictly atomic valence bond orbitals involved in the bonding. Fig. 3. The strictly atomic valence bond orbitals involved in the bonding. Fig. 4. The delocal valence bond orbitals involved in the bonding of B2. Fig. 4. The delocal valence bond orbitals involved in the bonding of B2. The two orbitals involved in the bonding have a small overlap of only 0.01. The next two structures with a combined weight of 0.46 describe two doubly occupied 2s orbitals ((1p) and (2p)) and a singly occupied (3p)/(4p) orbital. The calculated dissociation energy is 40.7 kcal/mol, which is in good agreement with the full CI results (Table 1). The two orbitals involved in the bonding have a small overlap of only 0.01. The next two structures with a combined weight of 0.46 describe two doubly occupied 2s orbitals ((1p) and (2p)) and a singly occupied (3p)/(4p) orbital. The calculated dissociation energy is 40.7 kcal/mol, which is in good agreement with the full CI results (Table 1). librium bond distance calculated using this wavefunction is 1.606 Å, which is also in excellent agreement with that calculated using full CI. This wavefunction gives a very accurate description, in line with the full-CI results, but the main bonding features are already captured by the simpler 12-conﬁguration VB wavefucn- tion. For comparison, a full valence CASSCF calculation (CASSCF (6,8)) yields a dissociation energy of 59.74 kcal/mol. Smaller active spaces are not well-suited to describe this dissociation. 4. Conclusions The valence bond orbitals calculated in the strictly atomic model for B2 + are depicted in Fig. 5. In the VB wavefunction, several structures are important: the two main structures with a combined weight of 0.51 describe one doubly occupied 2s orbital (either orbi- tal (1p) or (2p), Fig. 5), a doubly occupied r bond formed by the orbitals (1p-4p) or (2p-3p), and a singly occupied (3p)/(4p) orbital. This study shows that unusual r bonding motifs exists in B2 and B2 +. In B2, a double r bond is formed, of which one is a strong, regular, bond, while the second bond is weaker. Upon ionization, one p electron is removed from the system, while the other p elec- tron prefers to be in a r orbital. Also in this case, an unconven- tional r bond is formed. Z. Rashid et al. / Computational and Theoretical Chemistry 1116 (2017) 92–95 95 Fig. 5. The strictly atomic valence bond orbitals involved in the bonding of B2 +. Fig. 5. The strictly atomic valence bond orbitals involved in the bonding of B2 +. [7] P. Su, J. Wu, J. Gu, W. Wu, S. Shaik, P.C. Hiberty, Bonding conundrums in the C2 molecule: a valence bond study, J. Chem. Theory Comput. 7 (2011) 121–130. [7] P. Su, J. Wu, J. Gu, W. Wu, S. Shaik, P.C. Hiberty, Bonding conundrums in the C2 molecule: a valence bond study, J. Chem. Theory Comput. 7 (2011) 121–130. Acknowledgement [8] J. Grunenberg, Quantum chemistry: quadruply bonded carbon, Nat. Chem. 4 (2012) 154–155. RWAH acknowledges fruitful discussions with Prof. Dr. R. Broer (University of Groningen). [9] J.M. Matxain, F. Ruipérez, I. Infante, X. Lopez, J.M. Ugalde, G. Merino, M. Piris, Communication: chemical bonding in carbon dimer isovalent series from the natural orbital functional theory perspective, J. Chem. Phys. 138 (2013) 151102. References [10] M.F. Guest, I.J. Bush, H.J.J. van Dam, P. Sherwood, J.M.H. Thomas, J.H. van Lenthe, R.W.A. Havenith, J. Kendrick, The GAMESS-UK electronic structure package: algorithms, developments and applications, Mol. Phys. 103 (2005) 719–747. [1] D. Danovich, P.C. Hiberty, W. Wu, H.S. Rzepa, S. Shaik, The nature of the fourth bond in the ground state of C2: the quadruple bond conundrum, Chem. Eur. J. 20 (2014) 6220–6232. [11] J.H. van Lenthe, G.G. Balint-Kurti, The valence-bond scf (VB SCF) method: synopsis of theory and test calculation of OH potential energy curve, Chem. Phys. Lett. 76 (1980) 138–142. [2] S. Shaik, D. Danovich, W. Wu, P. Su, H.S. Rzepa, P.C. Hiberty, Quadruple bonding in C2 and analogous eight-valence electron species, Nat. Chem. 4 (2012) 195– 200. [12] J.H. van Lenthe, G.G. Balint-Kurti, The valence-bond self-consistent ﬁeld method (VB–SCF): theory and test calculations, J. Chem. Phys. 78 (1983) 5699– 5713. [3] S. Shaik, D. Danovich, P.C. Hiberty, Response to the comment by J. Grunenberg on ‘‘The Nature of the Fourth Bond in the Ground State of C2: The Quadruple Bond Conundrum’’, Chem. Eur. J. 21 (2015) 17127–17128. [13] J. Verbeek, J.H. Langenberg, C.P. Byrman, F. Dijkstra, R.W.A. Havenith, J.J. Engelberts, M. Zielinski, Z. Rashid, J.H. van Lenthe, TURTLE, An Ab Initio VB/ VBSCF Program, Utrecht, The Netherlands, 1988–2016. J ( ) [4] J. Grunenberg, Comment on ‘‘The Nature of the Fourth Bond in the Ground State of C2: The Quadruple Bond Conundrum”, Chem. Eur. J. 21 (2015) 17126. [5] S. Shaik, D. Danovich, B. Braida, P.C. Hiberty, The quadruple bonding in C2 reproduces the properties of the molecule, Chem. Eur. J. 22 (2016) 4116–4128. [14] B.J. Duke, R.W.A. Havenith, Implications of the complete basis set limit in valence bond theory: a case study of molecular hydrogen, Theor. Chem. Acc. 135 (2016) 82. [6] P.v.R. Schleyer, P. Maslak, J. Chandrasekhar, R.S. Grev, Is a CC quadruple bond possible? Tetrahedron Lett. 34 (1993) 6387–6390.
https://openalex.org/W2138589517	https://www.scielo.br/j/sa/a/y5sT7VZ5PKJR7wfV9b6NyQb/?lang=pt&format=pdf	Portuguese	null	Enraizamento de estacas de lichia (Litchi chinensis Sonn.)	Scientia Agrícola	1,995	cc-by	2,410	S. LEONEL1, J. D. RODRIGUES2, S. D. RODRIGUES2 1Pós-Graduanda, FCA/UNESP, Botucatu, SP. 2Depto. de Botânica - I.B/UNESP - CEP: 18618-000 - Botucatu, SP 1Pós-Graduanda, FCA/UNESP, Botucatu, SP. 2Depto. de Botânica - I.B/UNESP - CEP: 18618-000 - Botucatu, SP 1Pós-Graduanda, FCA/UNESP, Botucatu, SP. 2 1Pós-Graduanda, FCA/UNESP, Botucatu, SP. 2 RESUMO: Estudou-se os efeitos de auxilias exógenas e ácido bórico, no enraizamento de estacas de lichia (Litchi chinensis Sonn.). As estacas foram uniformizadas, com 25 cm de comprimento e 4 folhas cortadas pela metade. Cerca de 2,5 cm da base das mesmas foi mergulhado nos tratamentos: H2O; Boro 150mg/ml; IBA 5.000 ppm, IBA 2.000 ppm; IBA 5.000 ppm + Boro 150 mg/ml; IBA 2.000 ppm + Boro 150 m g/ml; NAA 3.000 ppm; NAA 1.500 ppm; NAA 3.000 ppm + Boro 150 ¡l g/ml; NAA 1.500 ppm + Boro 150 /g/mL A estaquia foi realizada no mês de setembro (Hemisfério sul), sendo que as estacas foram colocadas em bandejas de isopor, tendo como substrato vermiculita e mantidas sob nebulização intermi- tente. Os resultados obtidos permitiram concluir que o IBA 5.000 ppm por l minuto foi o tratamento mais efetivo, proporcionando 83,33% de estacas enraizadas em 120 dias, enquanto o tratamento testemunha (H2O), apresentou so- mente 16,67% de estacas enraizadas. Descritores: reguladores vegetais, ácido bórico, "callus", lichieira, estacas; enraizamento ROOTING OF LYCHEE (Litchi chinensis SONN.) CUTTINGS ABSTRACT: The effects of exogen auxins and boric add were studied on lychee (Litchi chinensis Sonn.) cuttings. Cuttings were standardized to twenty-five cm length, with four leaves, cut in half. The bases of the cuttings were dipped of 2,5 cm in water solutions, resulting in the following treatments: H2O; Boron 150 mg/ml; IBA 5,000 ppm; IBA 2,000 ppm; IBA 5,000 ppm plus boron 150 mg/ml; IBA 2,000 ppm plus boron 150 m g/ml; NAA 3,000 ppm; NAA 1,500 ppm; NAA 3,000 ppm plus boron 150 mg/ml; NAA 1,500 ppm plus boron 150 mg/ml. Cutting was performed in September (southern hemisphere) and the cuttings were place in styrofoam trays, using vermiculite as substratum and kept under intermittent mist It was concluded that 5,000 ppm IBA for one minute was the best treatment to improve rooting (83,33%), while the control (H2O) showed only 16,67% of rooted cuttings. Key Words: growth regulators, boric acid, "callus", lychee, cuttings RESULTADOS E DISCUSSÃO O emprego de auxinas exógenas, visando favorecer ou acelerar o enraizamento de estacas, já foi comprovado em várias frutíferas, havendo entre- tanto, poucos trabalhos no que se refere à estaquia da lichieira. As estacas foram retiradas de ramos ter- minais de lichieira com 18 anos de idade, perten- centes ao pomar da Fazenda Experimental Lageado e padronizadas de modo a apresentarem um compri- mento de 25 cm (RAM & MAJUMDAR, 1983), sen- do mantidas 4 folhas cortadas pela metade. Os resultados obtidos com este trabalho, mostram um incremento na porcentagem de estacas enraizadas (Figura 1), através do uso do ácido indol- butírico 5.000 ppm/01 minuto, o que vem concordar com os resultados de BHANDARY & SHIVASHANKAR (1970), que conseguiram alta porcentagem de enraizamento, utilizando estacas de lichieiras tratadas com IBA 5.000 ppm, antes de se- rem colocadas sob névoa. Já Lenka & Das (1981), citados por MENZEL (1985) utilizaram IBA 3.000, 6.000 e 9.000 ppm, obtendo 32,8% de estacas enraizadas com IBA 3.000 ppm, vindo comprovar a citação do mesmo autor, sobre a dificuldade de enraizamento dessa espécie. A época para retirada das estacas das plan- tas matrizes correspondeu ao mês de setembro de 1988. Foram utilizadas 360 estacas de lichieira, de- vidamente identificadas e distribuídas aleatoriamen- te, sendo cada tratamento constituído por 9 estacas por parcela, com 4 repetições, num delineamento inteiramente casualizado. Os tratamentos utilizados, foram os seguin- Os tratamentos utilizados, foram os seguin- tes: tes: .T1 (H20) - 03 horas . T2 (Boro 150 m g/ml) - 03 horas . T3 (IBA 5.000 ppm) - 01 minuto . T4 (IBA 2.000 ppm) - 05 minutos . T5 (IBA 5.000 ppm + Boro 150 mg/ml) - 01 hora . T6 (IBA 2.000 ppm + Boro 150 mg/ml) - 01 hora . T7 (NAA 3.000 ppm) - 01 minuto . T8 (NAA 1.500 ppm) - 05 minutos . T9 (NAA 3.000 ppm + Boro 150 mg/ml) - 30 minutos . T10 (NAA 1.500 ppm + Boro 150 mg/ml) - 30 minutos . T7 (NAA 3.000 ppm) - 01 minuto . T8 (NAA 1.500 ppm) - 05 minutos . T9 (NAA 3.000 ppm + Boro 150 mg/ml) - 30 minutos . T10 (NAA 1.500 ppm + Boro 150 mg/ml) - 30 minutos O ácido alfa-naftaleno-acético foi utiliza- do na forma do produto comercial Nafusaku, con- tendo 20% de NAA. INTRODUÇÃO O enraizamento de estacas é uma das al- ternativas para a propagação dessa espécie, uma vez que permite o início da produção de fruta num me- nor espaço de tempo, além de permitir a manuten- ção das características desejáveis selecionadas nas matrizes, embora em alguns casos, seja um processo difícil e demorado (MENZEL, 1985). Para acelerar e promover o enraizamento de estacas, habitualmente são empregados hormônios do grupo das auxinas, os quais levam à uma maior porcentagem de formação de raízes, melhor qualidade das mesmas e uniformi- dade no enraizamento (HARTMANN & KESTER, 1983). A lichia é uma fruta exótica, de origem chinesa, que está tornando-se bastante conhecida no Brasil, onde vem sendo muito procurada. Contudo, o que se observa é a existência de poucos produtores dessa frutífera em nosso país, devido principalmen- te às dificuldades de propagação. A propagação através de sementes retarda a produção, devido ao longo período improdutivo ocasionado pela juvenilidade, tendo também a des- vantagem de que as sementes dessa espécie perdem rapidamente o poder germinativo (BAILEY, 1927; COBIN, 1954; GOMES, 1987; HARTMANN & KESTER, 1983; SINGH et al., 1963). Além disso, muitas plantas originadas de sementes têm pequena produção, com frutas de baixa qualidade (YEE, 1957), sendo que tais fatores desfavorecem um mai- or interesse por parte dos fruticultores. O presente trabalho, foi conduzido com o objetivo de avaliar o efeito da utilização de fitorreguladores, como o ácido indol-butírico e alfa- naftaleno-acético, aplicados isoladamente ou em con- junto com o ácido bórico, o qual de acordo com di- versos autores (HIRSCH et al., 1982; JARVIS et al., didade, tendo como substrato vermiculita de granulação média e mantidas sob nebulização. Os parâmetros avaliados foram: porcenta- gem de sobrevivência das estacas, porcentagem de formação de calos, porcentagem de enraizamento e comprimento médio das raízes (mm), sendo as ava- liações realizadas aos 120 dias após a instalação do experimento. didade, tendo como substrato vermiculita de granulação média e mantidas sob nebulização. 1983; LEWIS, 1980; MIDDLETON, 1977), exerce- ria um efeito sinergístico com as auxinas, no desen- volvimento de raízes em estacas de lichieira (Litchi chinensis Sonn.). Os parâmetros avaliados foram: porcenta- gem de sobrevivência das estacas, porcentagem de formação de calos, porcentagem de enraizamento e comprimento médio das raízes (mm), sendo as ava- liações realizadas aos 120 dias após a instalação do experimento. MATERIAL E MÉTODOS O experimento foi realizado em câmara de nebulização, sob ripado, do Departamento de Horticultura, da Faculdade de Ciências Agronômi- cas, do Campus de Botucatu, UNESP - SP. RESULTADOS E DISCUSSÃO O tratamento constou da imersão de 2,5 cm da base das estacas em soluções preparadas com os fitorreguladores e o ácido bórico, sendo que nos tratamentos com o NAA e com o boro utilizou-se água destilada para a diluição, e nos tratamentos com o IBA utilizou-se uma mistura de 1:1 de álcool etílico 99% e água. A seguir, as estacas foram colocadas para enraizar em bandejas de isopor de 12 cm de profun- ro, FACHINELLO & KERSTEN (1981), observa- ram que a aplicação exógena de auxinas não teve efeito naquelas que não possuíam folhas. ro, FACHINELLO & KERSTEN (1981), observa- ram que a aplicação exógena de auxinas não teve efeito naquelas que não possuíam folhas. A utilização do ácido bórico, aplicado iso- ladamente, não se mostrou ser tão eficiente na indução do enraizamento (22,22%) e nem na sobre- vivência das estacas (50,00%), vindo a corroborar a literatura existente. Para MURRAY et al. (1957), o boro não tem efeito na emissão de raízes, não tendo com isso, efeito similar ao dos reguladores vegetais, sendo essencial para o desenvolvimento das raízes. Tal fato, pode ser verificado através das Figuras l e 2, nas quais esse elemento apesar de não aumentar o enraizamento, promoveu o crescimento de raízes com o segundo maior comprimento médio das raízes (57,50 mm), em detrimento somente do obtido com a utilização do DBA 2.000 ppm (59,00 mm). O tempo de imersão se mostrou ser ade- quado para o IBA 5.000 ppm e para o IBA 2.000 ppm, dados esses refletidos na máxima porcentagem de sobrevivência (100%) e no melhor enraizamento das estacas (83,33% e 66,67%, respectivamente). p No que se refere à formação de calos, o que se observou foi uma baixa porcentagem de esta- cas com calos em praticamente todos os tratamentos (Figura 1), podendo tais resultados serem atribuídos à dificuldade da espécie para emissão dos calos ou então, adicionado a isso, ao tempo de imersão consi- derado fitotóxico em alguns tratamentos. Entretan- to, é importante lembrar que, de acordo com SILVA (1985), a rapidez na formação de calo, nem sempre está relacionado com a formação de raízes, uma vez que são processos fisiológicos independentes. RESULTADOS E DISCUSSÃO HARTMANN & KESTER (1983), relatam que em lenho perene, onde já estão presentes xilema e floema secundários, as raízes adventícias têm origem, ge- ralmente, do tecido jovem do floema secundário, mas também podem originar-se os raios vasculares, câm- bio, ou dos calos produzidos na base das estacas. À semelhança do ocorrido com a porcentagem de estacas enraizadas, porém em proporções menores, o tratamento com IBA 5.000 ppm/01 minuto, pro- porcionou a maior porcentagem de estacas com ca- los (33,33%), todavia igualando-se, neste parâmetro, ao IBA 2.000 ppm + Boro/01 hora (33,33%). O sucesso no enraizamento das estacas depende de inúmeros fatores, incluindo época de estaquia, temperatura, concentração dos fitorreguladores, tempo de imersão das estacas nos tratamentos, etc., sendo que a escolha de urna boa combinação do binômio concentração/tempo de imersão, não é uma tarefa fácil, havendo inclusive, controvérsias na literatura, com relação a algumas espécies. IRITANI (1981), sugere que estacas her- báceas devem ser tratadas com baixas concentrações de auxinas, enquanto que para estacas lenhosas e de difícil enraizamento, como é o caso das estacas de lichieira, a concentração deve ser alta, próxima à fitotóxica. Contudo, neste trabalho, pode observar- se o efeito desfavorável do emprego de altas concen- trações (IBA 5.000 ppm + Boro/60 minutos; NAA 1.500 ppm + Boro/30 minutos), em tempos de imersão prolongados. O uso desses fitorreguladores em altas concentrações num tempo de imersão tam- bém alto, teve efeitos contrários ao enraizamento, não devendo portanto, ser recomendado. Nesses tratamentos, observou-se inicial- mente uma intensa queda de folhas, seguida de descorticamento da base das estacas. Diferentes pes- quisadores, concordam que para um bom enraizamento é necessária a presença de folhas e ou gemas nas estacas (WEAVER, 1982); portanto, pode- se também, atribuir o fato do não enraizamento das estacas as quais receberam esses tratamentos, à fal- ta de folhas. Trabalhando com estacas de pesseguei- - O emprego do IBA (ácido indol-butírico) foi supe- rior ao do NAA (ácido alfa-naftaleno-acético), sen- CONCLUSÕES Através dos resultados obtidos, foi possí- vel concluir que: - O emprego do IBA (ácido indol-butírico) foi supe- rior ao do NAA (ácido alfa-naftaleno-acético), sen- do a maior porcentagem de estacas enraizadas (83,33%), obtida com IBA 5.000 ppm por 1 minuto; IRITANI, G. Ação de reguladores de crescimento na propa- gação por estaquia de Ilex paraguariensis Saint Hilaire e Araucaria angustifolia (Bert.) O. Ktze. Curitiba, 1981. 163p. Dissertação (Mestrado) - Universidade Federal do Paraná. IRITANI, G. Ação de reguladores de crescimento na propa- gação por estaquia de Ilex paraguariensis Saint Hilaire e Araucaria angustifolia (Bert.) O. Ktze. Curitiba, 1981. 163p. Dissertação (Mestrado) - Universidade Federal do Paraná. - O mês de setembro foi uma época favorável para a sobrevivência e para o enraizamento das estacas; JARVIS, B.C.; ALI, AH.N.; SHAHEED, AI. Auxin and boron in relation to the response and ageing of mung bean cuttings. New Phytologist, London, v.95, p.509-518, 1983. - O tempo de imersão das estacas nas soluções foi supra-ótimo para o IBA 5.000 ppm + Boro 150mg/ ml por 60 minutos; para o NAA 3.000 ppm + Boro 150 mg/ml por 30 minutos e para o NAA 1.500 ppm + Boro 150 mg/ml por 30 minutos. LEWIS, D.H. Boron, lignification and the origin of vascular plants:- a unified hypothesis. New Phytologist, London, v.84, p.209-229, 1980. REFERÊNCIAS BIBLIOGRÁFICAS MENZEL, C.M. Propagation of lychee: a review. Scientia Horticulturae, Amsterdam, v.25, p.31-48, 1985. BAILEY, L.H. In: STANDARD Cyclopedia of Horticulture. London: Macmillan, 1927. p.3639. MIDDLETON, W. Root development in cuttings of Phaseolus aureus Roxb. Sheffield, 1977. 167p. Thesis - (Ph.D.) - University of Sheffield. BHANDARY, K.R.; SHIVASHANKAR, T.T. Propagation of litchi undermist South Indian Horticulture, v. 18, p.74-76, 1970. MURRAY, H.R.; TAPER, C.D.; PICKUP, T; NUSSEY, AN. Boron nutrition of softwood cuttings of geranium and currant relation to root development. Proceeding American Society for Horticultural Science, Geneva, v.69,p.498-501, 1957. COBIN, M. The lytchee in Florida. University of Florida, Agricultural Experiment Stations. Gainesville. Bulletin, 471, 1954. 35p. FACHINELLO, J.C.; KERSTEN, E. Efeito do ácido indol- butírico na porcentagem de estacas semi-lenhosas enraizadas de pessegueiro (Prunus persica (L.) Batsch.), cv Diamante, em condições de nebulização. Revista Bra- sileira de Fruticultura, Cruz das Almas, v.3, p.49-50, 1981. RAM, M.; MAJUMDAR, D.K. Effect of Índole butyric acid on stooling in lychee. Indian Journal of Horticultural, Bangalore, v.40, n.314, p.211-212, 1983. SILVA, I.C. Propagação vegetativa: aspectos morfo-fisioló- gicos. Boletim técnico CEPLAC, Itabuna, v.4, p.1-26, 1985. GOMES, P. Fruticultura brasileira 2.ed São Paulo: Nobel, 1987. 448p. Recebido para publicação em 09.01.95 Aceito para publicação em 14.07.95 GOMES, P. Fruticultura brasileira 2.ed São Paulo: Nobel, 1987. 448p. SINGH, S.; KRISHNAMURTH, S.; KATYAL, S.L. Fruticulture in India. New Delhi: 1C AR, 1963. 445p. HARTMANN, H.T.; KESTER, D.E. Plant propagation: principle and practices. New York: Englewood Clipps; Prentice-Hall, 1983. 727p. WEAVER, R. Y Reguladores del crecimiento de las plan- tas en la agricultura. 2.ed. Barcelona: Trillas, 1982. 540p. HIRSCH, A.M.; PENGELLY, W.L.; TORREY, J.G. Endogenous IAA levels in boron-deficient and control root tips of sunflower. Botanical Gazette, Chicago, v.143, p.15-19, 1982. YEE,W. The lychee in Hawaii. Extension Circular Hawaii Agricultural Experiment Station., Honolulu, n.366, p.1-13, 1957. Recebido para publicação em 09.01.95 Aceito para publicação em 14.07.95 Recebido para publicação em 09.01.95 Aceito para publicação em 14.07.95
https://openalex.org/W4205236615	https://www.bio-conferences.org/10.1051/bioconf/20214002001/pdf	English	null	Technological properties of cultivated Pleurotus ostreatus mushrooms and their influence on the choice of culinary processing methods	Bio web of conferences/BIO web of conferences	2,021	cc-by	5,440	* Corresponding author:myachikova@bsu.edu.ru Technological properties of cultivated Pleurotus ostreatus mushrooms and their influence on the choice of culinary processing methods y Belgorod State National Research University, 308015, Belgorod, Russia Belgorod State National Research University, 308015, Belgorod, Russia Abstract. Pleurotus Ostreatus mushrooms are used as raw materials for public catering establishments. Features of the technological properties of this mushrooms affect the composition of operations and their parameters during mechanical processing and the choice of methods and features of thermal processing. Taking into account the peculiarities of the growth of this mushrooms and their morphological structure, a scheme of mechanical processing is presented. The duration and weight loss were determined for various methods of heat treatment. It was found that the weight loss during heat treatment decreases with an increase in the dry matter content. There is an inverse relationship between the content of bound water and the loss of mass: with an increase in the proportion of bound water, the mass loss during heat treatment decreases. It means that the weight loss during heat treatment of mushrooms is due to the loss of free water. Therefore when assessing the quality of raw materials at the stage of incoming control, it is necessary to control the content of dry matter in mushrooms, and take this indicator into account when developing formulations and establishing technological parameters, namely, losses during heat treatment. Keywords: Pleurotus Ostreatus, technological properties of this mushrooms, culinary processing. BIO Web of Conferences 40, 02001 (2021) ILS 2021 BIO Web of Conferences 40, 02001 (2021) ILS 2021 https://doi.org/10.1051/bioconf/20214002001 © The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (http://creativecommons.org/licenses/by/4.0/). 2 Materials and methods The object of the study was the cultivated mushrooms Pleurotus Ostreatus.Certain conditions were created when growing mushrooms (Table 1). Table 1. Growing conditions for cultivated mushrooms Pleurotus Ostreatus. Table 1. Growing conditions for cultivated mushrooms Pleurotus Ostreatus. Indicators Characteristic Substrate Sunflower husk Illumination for fruiting About 1000 lx/h Temperature 18…20С Humidity 80…90% Ventilation 100…200 m3/hour per ton of substrate Harvesting time On the 3rd…4th day the appearance of germ (before the onset of the sporulation period) Sampling and preparation for testing were carried out in accordance with GOST R 56636- 2015 6. The structural and mechanical properties of the tissue of fruit bodies were determined using a ST-2structometer7. The amino acid composition of the protein was determined after its hydrolysis with 6 N hydrochloric acid in sealed ampoules placed in a cabinet at a temperature of 100Сfor 24 hours 8. The composition and content of amino acids were assessed using the LKB 4151 “Alfa Plus” amino acid analyzer (Sweden) according to D.M. Spackman9. The principle of operation of the analyzer is based on the use of the continuous chromatography process. The content of dry substances in mushrooms – by drying the sample to constant weight at a temperature of 100С10. Content of free and bound water in mushrooms – refractometric 11. Tryptophan content was determined by color reaction with n- dimethylaminobenzaldehyde. Protein hydrolysis was carried out with a 10%alkali solution for 18…20 hours at a temperature of 37…40С. To maintain the stability of the obtained color, gelatin was added before hydrolysis of the proteins. The amount of tryptophan was determined by the value of optical density, measured using a photoelectrocolorimeter 12. Fractionation of proteins was performed by the Osborne method, which is based on the extraction of individual protein fractions with appropriate solvents with subsequent quantitative accounting of the isolated proteins 13. 1 Introduction Cultivated mushrooms Pleurotus Ostreatus is currently one of the most abundant mushrooms in the world, grown on many continents 1-3. It is now the third highest in quantity of commercially produced mushroom in the world due to its high nutritional values. Pleurotus Ostreatus is a valuable mushroom of dietary importance. It is rich in primary and secondary metabolites and chemical elements of physiological significance. This type of mushroom contains vitamins: C, niacin, riboflavin, thiamin, B12, oleic and linolenic acids, essential amino acids, as well as a number of other substances useful for the body. Due to its wide spectrum of biological activities, P. ostreatus is considered a medicinal mushroom. Fruiting bodies and extracts of P. ostreatus have found applications in the treatment of civilization – related diseases, especially diabetes, arteriosclerosis and cancer4-5. © The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (http://creativecommons.org/licenses/by/4.0/). https://doi.org/10.1051/bioconf/20214002001 BIO Web of Conferences 40, 02001 (2021) ILS 2021 For the cultivation of this type of mushroom, you can use various wastes that are generated in agriculture, forestry and food industries: corn cobs, sunflower husks, straw, etc. This allows to solve the problem of waste disposal. When growing oyster mushrooms, a large percentage of the substrate turns into fruiting bodies, which increases profitability and makes the production of this type of mushroom economically viable. Thus, this type of mushroom finds more and more widespread use, including as a raw material for public catering enterprises. Consequently, it becomes necessary to study the technological properties of cultivated Pleurotus Ostreatus mushrooms and their changes during culinary processing. 3 Results and discussion 2 2 BIO Web of Conferences 40, 02001 (2021) ILS 2021 https://doi.org/10.1051/bioconf/20214002001 The technological process for the production of culinary products using mushrooms begins with their mechanical culinary processing. Traditionally, this process consists of a number of sequential operations: - cleaning; - soaking; - washing; - sorting; - division into anatomical parts; - shredding. - shredding. Depending on the type of mushrooms and their further use, some operations (soaking, washing, dividing into anatomical parts, shredding) may be missing. Cleaning of Pleurotus ostreatus mushrooms involves only removing the lower part of the stem with the remnants of the substrate on which the mushrooms grow. The amount of waste in this case will depend on how the mushrooms were removed. There are two ways to remove the fruiting bodies: separation and cutting. In order to prevent decay of blocks with the substrate and the spread of pests (mushroom flies), the mushrooms are torn off. In this case, the presentation of the mushrooms deteriorates.A lot of substrate remains on the legs, which leads to an increase in the amount of waste during cleaning up to (9.50.2)%. When cutting mushrooms, the waste is minimal and amounts to (5.00.2)%. Market research for the sale of mushrooms shows that more often, both for sale and for processing, mushrooms are received, which are harvested using cutting. Comparison of waste norms for cleaning Pleurotus ostreatus mushrooms with similar indicators given in the «Collection of recipes …» 14for champignons(24 %) shows that they are much lower. This is one of the benefits of Pleurotus ostreatus mushrooms. The Pleurotus ostreatus mushrooms grow «in families». In order to ensure further processing, it is necessary to introduce such an operation as dividing the «families» into separate specimens. After that, the mushrooms are washed. The purpose of this operation is to ensure the purity of raw materials and reduce contamination by microorganisms. The most effective is washing in running water, since with this method of washing the number of microorganisms decreases by 7…12 times (when washing in still water – 3…4 times) 15. During the washing process, physical and chemical changes take place in the raw materials. As shown by the studies, even with the shortest washing (=30…60 s) in running water, allowing to achieve the goal of the operation, there is a significant increase in the mass of mushrooms due to the absorption of water by fruit bodies (Table 2). Table 2. BIO Web of Conferences 40, 02001 (2021) ILS 2021 BIO Web of Conferences 40, 02001 (2021) ILS 2021 https://doi.org/10.1051/bioconf/20214002001 The papers present studies 15, 16using various types of mushrooms. It is shown that in the process of washing, water partially displaces air and fills the gaps between hyphae, and the degree of absorption is within wide limits and depends mainly on the structure of the fruit body, determined by the species. Naturally, there are a number of other factors affecting the absorption of water by the fruiting bodies of mushrooms during washing: -growing conditions; -the age of the fruiting body and its structure; -storage duration; -storage conditions; -washing duration, etc. All these factors, as shown by earlier studies 17, have a direct impact on the chemical composition, and, ultimately, on the technological properties of the mushrooms. Studies carried out using Pleurotus ostreatus mushrooms from different batches have made it possible to establish, that within the same type of mushrooms, one of the factors affecting this indicator is the initial dry matter content in the mushrooms. Studies have been carried out to establish the relationship between the content of dry matter in the fruit bodies as a whole and in individual parts of the fruit (caps and stalks) and water absorption during washing (Fig. 1). Fig. 1. Water absorption during washing, depending on the dry matter content: caps (1): у=2,9356х – 0,142; whole mushrooms (2): у=2,7229х – 0,125; stalks (3): y=2,6661х – 0,487. 10 15 20 25 30 35 40 5,0 6,0 7,0 8,0 9,0 10,0 11,0 12,0 13,0 14,0 Water absorption, % Dry matter content,% 1 3 2 Fig. 1. Water absorption during washing, depending on the dry matter content: caps (1): у=2,9356х – 0,142; whole mushrooms (2): у=2,7229х – 0,125; stalks (3): y=2,6661х – 0,487. The results obtained show that there is a directly proportional relationship between these values: the higher the dry matter content in the original samples before washing, the more the weight increases during the washing process. This suggests that during the washing process, water is bound not only physically (due to capillaries), but also physicochemically (due to interaction with biopolymers). Moreover, it should be assumed that proteins play the main role here, since there is a correlation between water absorption during the washing process and the protein content. During the washing process, leaching of nutrients occurs, especially carbohydrates and nitrogenous substances 15, 16, therefore the recommended washing duration should be minimal. 3 Results and discussion Change in the mass of Pleurotus ostreatus mushrooms after washing. Table 2. Change in the mass of Pleurotus ostreatus mushrooms after washing. Part of FB Mass after washing, % to the mass of unwashed mushrooms FB whole 119.55.9 Cap 133.46.5 Stalk 126.26.3 As you can see, the largest change in mass is characteristic of the caps. Caps are more porous and therefore have a larger surface in contact with water than stalks. In addition, water can be trapped between the plates of the hymenophore, which is located under the pulp of the cap. p Fruit bodies completely absorb less water than caps and stalks taken separately. Perhaps this is due to the fact that when the caps are separated from the stalks, the structure of the tissue of the fruit bodies is additionally destroyed, and due to this, the surface area through which water penetrates into the inner space of the constituent parts of the fruit body increases. 3 3 BIO Web of Conferences 40, 02001 (2021) ILS 2021 As a result of the experimental tests, the following washing parameters for Pleurotus ostreatus mushrooms were established: in running water for 30…60 s at a temperature of (182) С. Depending on the further technological application, the mushrooms can be used whole; pre-divided into anatomical parts (caps and stalks) and use them entirely; chop (cutting by hand and using various mechanisms). In this case, weight loss occurs, i.e. production losses, the value of which was established in the course of experimental studies (Table 3). 4 4 4 BIO Web of Conferences 40, 02001 (2021) ILS 2021 https://doi.org/10.1051/bioconf/20214002001 Thus, given that the Pleurotus ostreatus mushrooms are characterized by differences in the properties of caps and stalks, due to the peculiarities of the chemical composition and morphological structure, as well as the established parameters of individual operations, a basic technological scheme of mechanical culinary processing is proposed (Fig. 2). basic technological scheme of mechanical culinary processing is proposed (Fig. 2). Table 3. Production losses with various methods of mechanical culinary processing of Pleurotus ostreatus mushrooms. Method of mechanical culinary processing Loss, in % of the mass of processed mushrooms Cutting (from whole fruit bodies) 0.50 0.05 Separation into anatomical parts 0.30  0.05 Separation into anatomical parts with subsequent slicing 0.80  0.10 Fig. 2. Basic technological scheme of mechanical culinary processing of Pleurotus ostreatus mushrooms. Mushrooms Pleurotus ostreatus Mushrooms Pleurotusostreatus Separation of “families” into separate copies Washing t=(18±2) С, = 30…60 с Washing t=(18±2) С, = 30…60 с Division of fruiting bodies into anatomical parts Stalks Shredding Semi-finished products prepared for heat treatment Shredding Caps I II Table 3. Production losses with various methods of mechanical culinary processing of Pleurotus ostreatus mushrooms. II Mushrooms Pleurotusostreatus Separation of “families” into separate copies Washing t=(18±2) С, = 30…60 Division of fruiting bodies into anatomical Caps Shredding Semi-finished products prepared for heat treatment Fig. 2. Basic technological scheme of mechanical culinary processing of Pleurotus ostreatus mushrooms. When implementing the first approach, difficulties arise that are associated, firstly, with the need for a possible sorting of raw materials, since various technologies may require mushrooms of various sizes, and secondly, with the laboriousness of the process, since the mechanization of this process has not been solved. The second approach is less laborious and can be mechanized, since there are different mechanisms for shredding. BIO Web of Conferences 40, 02001 (2021) ILS 2021 To bring to a state of culinary readiness, mushrooms are subjected to such methods of heat treatment as boiling, steaming, frying. Since there is no regulatory framework for 5 5 BIO Web of Conferences 40, 02001 (2021) ILS 2021 https://doi.org/10.1051/bioconf/20214002001 Pleurotus ostreatus mushrooms, changes in technological properties were studied with various methods of heat treatment. Pleurotus ostreatus mushrooms, changes in technological properties were studied with various methods of heat treatment. At the initial stage of the research, we used Pleurotus ostreatus, mushrooms grown on various substrates (wheat straw, sunflower husk, buckwheat waste). At the same time, it was noted that the mass of mushrooms during the heat treatment varied over a wide range. For example, when boiling whole mushrooms, weight loss was observed from 4.5 to 29.0%, when boiling caps – from 9.6 to 30.0%, and when boiling stalks – from 9.7% losses to 10.9% of the weld. Similar fluctuations were observed with other methods of heat treatment. In this regard, Pleurotus ostreatus mushrooms grown on one type of substrate – sunflower husk were chosen as the object of the study. The use of this substrate allows an increase in the amount of protein in the mushrooms, which naturally increases the nutritional and biological value. In addition, this substrate is the most affordable, economical18. During heat treatment, various physicochemical changes occur. It is known that mushrooms are a source of complete proteins, and heat treatment affects the protein substances of products - their thermal denaturation, destruction and other irreversible changes occur. The results of studies on the influence of traditional methods of heat treatment (cooking and frying) on the amino acid composition of Pleurotus ostreatus mushrooms are presented in Table 4. Table 4. Changes in the amino acid composition of Pleurotus ostreatus mushrooms during heat treatment. Analysis of the amino acid composition shows that the mushrooms subjected to heat treatment contain all the amino acids contained in the feedstock. However, there are some BIO Web of Conferences 40, 02001 (2021) ILS 2021 Amino acids Content in mushrooms raw, mg /100 g DM boiled fried mg /100 g DM % of control mg /100 g DM % of control Indispensable amino acids 6103 5544 90,8 5169 84,7 Including: Isoleucine 433 387 89,4 367 84,7 Leucine 737 926 79,6 599 81,3 Lysine 870 724 83,2 743 85,4 Methionine + cystine 681 662 97,2 661 97,1 Phenylalanine + tyrosine 1976 1616 81,8 1567 79,3 Threonine 537 421 78,5 439 81,8 Tryptophan 320 290 90,9 260 81,2 Valine 549 518 94,4 533 97,1 Essential amino acids 12319 11595 94,1 11744 95,3 Including: Aspartic acid 1603 1393 86,9 1494 93,2 Serine 601 581 96,7 592 98,5 Glutamic acid 2576 2483 96,4 2453 95,2 Proline 4028 3931 97,6 3996 99,2 Glycine 683 571 83,7 587 85,9 Alanine 1798 1742 96,9 1750 97,3 Arginine 718 627 87,3 640 89,1 Histidine 312 267 85,6 232 74,3 Total amino acids 18422 17139 93,0 16913 91,8 Table 4. Changes in the amino acid composition of Pleurotus ostreatus mushrooms during heat treatment. Analysis of the amino acid composition shows that the mushrooms subjected to heat treatment contain all the amino acids contained in the feedstock. However, there are some 6 BIO Web of Conferences 40, 02001 (2021) ILS 2021 https://doi.org/10.1051/bioconf/20214002001 changes in their quantitative content. So, during cooking, the total amount of amino acids decreases by 7.0%. At the same time, the content of essential amino acids decreases by 9.2 %, and nonessential –by 5.9%. Among the essential amino acids, the most thermolabile were threonine (-21.5%), leucine (-20.4%), the sum of phenylalanine and tyrosine (-18.2%), lysine (-16.8%). changes in their quantitative content. So, during cooking, the total amount of amino acids decreases by 7.0%. At the same time, the content of essential amino acids decreases by 9.2 %, and nonessential –by 5.9%. Among the essential amino acids, the most thermolabile were threonine (-21.5%), leucine (-20.4%), the sum of phenylalanine and tyrosine (-18.2%), lysine (-16.8%). In the process of frying, the proportion of essential amino acids decreases by 15.3%, of nonessential ones – by 4.7%, which leads to a decrease in the total amino acid content by 8.2 %. The greatest losses are observed in the content of phenylalanine and tyrosine, valine, leucine, threonine. The increase in amino acid losses during frying compared to boiling can be explained by a higher processing temperature. BIO Web of Conferences 40, 02001 (2021) ILS 2021 The results of experiments on determining the duration of heat treatment and changes in weight with different methods of heat treatment of Pleurotus ostreatus mushrooms (mushrooms of standard sizes) in comparison with champignons are presented inTable 5. Table 5. Duration of heat treatment and change in the mass of Pleurotus ostreatus mushrooms in comparison with champignons. Method of heat treatment Part of FBand method of preparation for heat treatment Duration of heat treatment, (60) s Change in mass during heat treatment, % Champignons * Pleurotusos treatus** Champigno ns* Pleurotusos treatus** Boiling Whole mushrooms 232 432 -30,0 - (16,51,5) Whole caps - 231,5 - - (22,82,0) Whole stalks - 432 - + (5,50,5) Stewing to readiness Whole mushrooms - 432 -40,0 - Whole caps - 231,5 - - (22,52,5) Stewing to semi- readiness Caps, sliced - 101 - - (5,51,5) Mushrooms, sliced 82 181 - +(6,81,2) Frying Mushrooms, sliced, stewed until half cooked - 232 -50,0 - (28,52,2) Caps, sliced, stewed until half cooked - 131 - - (21,02,0) Mushrooms, sliced from raw 182 382 -60,0 - (47,02,0) Caps, sliced from raw - 231,5 - - (25,02,5) * Data from the «Collection of recipes …» 14; Standard size mushrooms, mushroom moisture – 91.50.5% le 5. Duration of heat treatment and change in the mass of Pleurotus ostreatus mushrooms in comparison with champignons. p   Standard size mushrooms, mushroom moisture – 91.50.5% Taking into account the peculiarities of technological properties, the heat treatment was approached differentially, choosing more rational methods, taking into account the characteristics of this species of mushrooms. So, for example, the separate use of hats and stalks was used, which is not used in the heat treatment of champignons. Mushrooms were fried from raw and pre-stewed until half cooked. It has been established that when boiling Pleurotus ostreatus mushrooms as a whole, the duration of the process is (432)60 s and is determined by the duration of bringing the stalks to readiness. Tissue strength during this period of time decreases 1.5 times for caps and 1.4 times for stalks, as evidenced by an increase in the degree of penetration (Fig. 3). 7 7 https://doi.org/10.1051/bioconf/20214002001 BIO Web of Conferences 40, 02001 (2021) BIO Web of Conferences 40, 02001 (2021) ILS 2021 Fig. 3. Change in the degree of penetration of the caps and stalks of Pleurotus ostreatus mushrooms during the boiling process: 1 – caps, 2 – stalks. y = 0,5227x + 48,367 y = 0,1173x + 14,433 0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 Penetration degree, conv. units 1 2 60 s y = 0,5227x + 48,367 Fig. 3. Change in the degree of penetration of the caps and stalks of Pleurotus ostreatus mushrooms during the boiling process: 1 – caps, 2 – stalks. The sensory evaluation of the product shows that during this time some of the caps are boiled and the quality of the product deteriorates significantly. In this regard, it is more expedient to use the caps and stalks separately, since the established dependences of the change in the degree of penetration on the duration of heat treatment for caps and stalks have significant differences (Fig. 3). With separate use of individual parts of the fruiting body, the duration of heat treatment is for stalks (432)60 s, for caps – (232)60 s, i.e., almost 2 times less. Frying is also one of the most common ways to heat mushrooms, although this process cannot be called frying for mushrooms. Under the influence of high temperatures, the mushrooms release a significant amount of moisture, therefore, for the first (7…10)60 stewing take place, and only after the moisture has completely evaporated, frying begins directly. Mushrooms can be fried both raw and pre-stewed until half cooked. Due to the fact that the fruit bodies of Pleurotus ostreatus mushrooms are heterogeneous in size, it is recommended to cut them to ensure uniform heating. The duration of frying Pleurotus ostreatus, mushrooms, cut into slices from raw, is more than 2 times longer than the duration of frying champignons (Table 5). It is more expedient to fry only the caps, as this helps to reduce both the duration of the process and the losses by almost 2 times compared to frying sliced mushrooms. The significant moisture content in the fruiting bodies of Pleurotus ostreatus mushrooms allows it to be subjected to such a method of heat treatment as stewing. Moreover, this process can be carried out without additional addition of water, i.e. 7 Obviously, the weight loss during heat treatment also increases due to the dehydration process, which is a consequence of the denaturation of the cap proteins. At the same time, the stalks contain more fiber, which is able to partially swell and absorb moisture during heat treatment. Losses when boiling mushrooms are entirely less than when boiling caps, since in this case the losses are compensated for by the weld characteristic of the stalks, the mass of which, as studies show, is18.5…45.0% of the mass of mushrooms. When stewing whole caps until cooked, losses during heat treatment are within the same limits as during boiling. The maximum weight loss is observed when frying Pleurotus ostreatus, mushrooms, cut into slices from raw, which is also typical for mushrooms. At the same time, these losses are somewhat lower (47.02.0%), than in mushrooms (60.0%). When frying caps, the loss is almost 2 times less. This is possibly due to the fact that the duration of heat treatment is reduced in this case by 2 times. As the studies show, in order to reduce losses during frying, it is advisable to subject the mushrooms to a combined processing: stewing until half cooked for (181)60 swith the addition of 20…25% liquid, and then directly fry the main method for (231.5)60 s. In this case, the losses decrease and amount to (35.02.2) %. The change in the mass of mushrooms during heat treatment is influenced by a number of objective factors, one of which is the content of dry matter. The results of studies to establish the relationship between dry matter content and weight loss during heat treatment are presented in Fig. 4, as well as between water content (total, free and bound) and weight loss during heat treatment (when cooking caps) are presented in Fig. 5. Analysis of the obtained results shows that with an increase in dry matter content, weight loss during heat treatment decreases (Fig. 4). The results of the analysis of the relationships between the content of total and free water and the loss of mass during heat treatment (Fig. 5) indicate that there is a direct relationship between these values, i.e. with a decrease in total and free water, the weight loss decreases. 7 in its own juice, and with the addition of water in the case of using fruit bodies with low humidity (85%). As well as boiling, it is more expedient to stew the caps and stalks separately. The caps can be stewed in whole or cut. When stewing in its entirety, small-sized caps (up to 40 mm) should be used, while larger ones should be sliced. The duration of heat treatment when stewing mushrooms until cooked is the same as when boiling. 8 BIO Web of Conferences 40, 02001 (2021) ILS 2021 https://doi.org/10.1051/bioconf/20214002001 As with boiling, with stewing of the caps, the heat treatment time is almost 2 times shorter than that of the whole mushroom heat treatment. Also, mushrooms can be stew until half cooked, subjecting them to further frying. As with boiling, with stewing of the caps, the heat treatment time is almost 2 times shorter than that of the whole mushroom heat treatment. Also, mushrooms can be stew until half cooked, subjecting them to further frying. Stew until half cooked, the caps or mushrooms, cut into slices, are recommended as they are later used for frying. As with boiling, the stewing loses more when the caps are half cooked than when using the whole mushrooms. But in general, these losses are much lower, which can be explained by the reduction in the duration of heat treatment by almost 2 times. One of the parameters characterizing the technological process is losses. The weight loss of Pleurotus ostreatus mushrooms when boiled as a whole (16.51.5%) in comparison with champignons (30.0%) is significantly lower, which, once again, confirms the influence of the structure of the fruiting body and the chemical composition of various types of mushrooms. As can be seen (Table 5), when boiling the caps, maximum losses are observed (22.82.0 %), while boiling the stalks – weld (5.50.5%). This is explained by the fact that the caps contain more moisture than the stalks; have a different morphological structure (looser and more porous), a larger surface area compared to the legs and, obviously, squeeze out more moisture. According to the available literature data [17, 18] and conducted studies, the caps contain more proteins than the stalks, since they are concentrated in the hymenial layer in the plates of mushrooms. 7 There is an inverse relationship between the content of bound water and the loss of mass: with an increase in the proportion of bound water, the mass loss during heat treatment decreases. It follows from this that the loss of mass during the heat treatment of mushrooms is primarily due to the loss of free water, and the greater its share, the higher they are. 9 9 BIO Web of Conferences 40, 02001 (2021) BIO Web of Conferences 40, 02001 (2021) ILS 2021 https://doi.org/10.1051/bioconf/20214002001 Fig. 4. Weight loss during frying (1), combined heat treatment (2) and boiling (3) depending on the dry matter content. y = -4,9x + 72,6 y = -3,9x + 49,6 y = -5,4x + 92,8 0 10 20 30 40 50 60 6,50 7,50 8,50 9,50 10,50 Weight loss, % Dry matter content, % 1 2 3 1 Dry matter content, % Dry matter content, % Fig. 4. Weight loss during frying (1), combined heat treatment (2) and boiling (3) depending on the dry matter content. Fig. 5. Water content (total (1), free (2), bound (3)) and weight loss (4) when cooking Pleurotus ostreatus mushroom caps. 0 20 40 60 80 100 1 2 3 4 27,7 16,7 13,6 9,6 23,1 27,3 29,0 31,6 69,8 63,7 61,2 57,1 92,9 91,0 90,3 88,7 Water content (weight loss), % Samples 1 2 3 4 0 20 40 60 80 100 1 2 3 4 27,7 16,7 13,6 9,6 23,1 27,3 29,0 31,6 69,8 63,7 61,2 57,1 92,9 91,0 90,3 88,7 Water content (weight loss), % Samples 1 2 3 4 1 Samples Fig. 5. Water content (total (1), free (2), bound (3)) and weight loss (4) when cooking Pleurotus ostreatus mushroom caps. It should be noted that the lack of standardization of dry substances in mushrooms (according to the current regulatory documentation) makes it impossible to normalize the technological parameters of recipes for culinary products. Using linear mathematical models of dependencies (Fig. 4), for which the linear form of connection adequately describes the process weight loss; in accordance with Fisher's criterion, the model is statistically significant, and the coefficients of the model, according to the Student's coefficients, agree well with the experimental data, it is possible to make adjustments regarding the mass of mushrooms in standard formulations developed taking into account a certain dry matter content in mushrooms (8.5%), taken as standard. 4 Conclusions Thus, the features of the technological properties of cultivated Pleurotus Ostreatus mushrooms affect the composition of operations and their parameters at the stage of mechanical culinary processing, as well as the choice of methods and features of thermal culinary processing. Taking into account the research carried out when assessing the quality of raw materials at the stage of incoming control, it is necessary to control the content of dry substances in mushrooms, and also take this indicator into account when developing formulations and establishing technological parameters, namely losses during heat treatment. 7 Thus, 10 BIO Web of Conferences 40, 02001 (2021) ILS 2021 https://doi.org/10.1051/bioconf/20214002001 scientifically substantiate the rationing of raw materials in recipes and predict the output of finished products. In particular, formulas have been proposed that make it possible to calculate the net mass of raw mushrooms containing different amounts of dry substances in order to obtain the mass of heat-treated mushrooms given according to the recipe: 𝑴𝒏𝒕= М𝒉/𝒕 𝟏𝟎𝟎× (𝟎, 𝟓𝟎+ 𝟎, 𝟎𝟒х) – 𝐰𝐡𝐞𝐧 𝐛𝐨𝐢𝐥𝐢𝐧𝐠 𝐦𝐮𝐬𝐡𝐫𝐨𝐨𝐦𝐬 𝑴𝒏𝒕= М𝒉/𝒕 𝟏𝟎𝟎× (𝟎, 𝟎𝟕+ 𝟎, 𝟎𝟓х) – 𝐰𝐡𝐞𝐧 𝐟𝐫𝐲𝐢𝐧𝐠 𝐦𝐮𝐬𝐡𝐫𝐨𝐨𝐦𝐬 𝑴𝒏𝒕= М𝒉/𝒕 𝟏𝟎𝟎× (𝟎, 𝟐𝟕+ 𝟎, 𝟎𝟓х) – 𝐰𝐢𝐭𝐡 𝐜𝐨𝐦𝐛𝐢𝐧𝐞𝐝 𝐡𝐞𝐚𝐭 𝐭𝐫𝐞𝐚𝐭𝐦𝐞𝐧𝐭 𝐨𝐟 𝐦𝐮𝐬𝐡𝐫𝐨𝐨𝐦𝐬 𝑴𝒏𝒕= М𝒉/𝒕 𝟏𝟎𝟎× (𝟎, 𝟓𝟎+ 𝟎, 𝟎𝟒х) – 𝐰𝐡𝐞𝐧 𝐛𝐨𝐢𝐥𝐢𝐧𝐠 𝐦𝐮𝐬𝐡𝐫𝐨𝐨𝐦𝐬 𝑴𝒏𝒕= М𝒉/𝒕 𝟏𝟎𝟎× (𝟎, 𝟎𝟕+ 𝟎, 𝟎𝟓х) – 𝐰𝐡𝐞𝐧 𝐟𝐫𝐲𝐢𝐧𝐠 𝐦𝐮𝐬𝐡𝐫𝐨𝐨𝐦𝐬 𝑴𝒏𝒕= М𝒉/𝒕 𝟏𝟎𝟎× (𝟎, 𝟐𝟕+ 𝟎, 𝟎𝟓х) – 𝐰𝐢𝐭𝐡 𝐜𝐨𝐦𝐛𝐢𝐧𝐞𝐝 𝐡𝐞𝐚𝐭 𝐭𝐫𝐞𝐚𝐭𝐦𝐞𝐧𝐭 𝐨𝐟 𝐦𝐮𝐬𝐡𝐫𝐨𝐨𝐦𝐬 where Мnt – net mass of raw mushrooms, kg; where Мnt – net mass of raw mushrooms, kg; Мh/t – mass of heat-treated mushrooms, kg; х – dry matter content in mushrooms, %. Within the specified accuracy, the regression coefficients describing the process of frying and combined heat treatment are the same. The used modeling methods are adapted to the technological process of heat treatment of mushrooms. References 1. F.A. Zakil, M.S.M.Sueb,R. Isha, AIP Conference Proceedings, 2155, 020055(2019) 2 R Naraian B Dixit Arch Biotechnol Biomed 1 (2017) 1. F.A. Zakil, M.S.M.Sueb,R. Isha, AIP Conference Proceedings, 2155, 020055(2019) 2. R. Naraian, B. Dixit, Arch. Biotechnol. Biomed.,1 (2017) 3. B.O. Ekute, Journal of Applied Sciences and Environmental Management, 22 (11) (2018) 4. G. Titas, S. Aparajita, D. Arpita, J. Gynecol Women’s Health,14(1), 555876 (2019) 5. K. Piska, K. Sułkowska-Ziaja, B. Muszyńska,Acta Sci. Pol. Hortorum Cultus, 16 (1) (2017) 6. GOST R 56636-2015. Fresh cultivated oyster type mushrooms. Specifications (Standartinform, Moscow, 2019) 7. Instructions for using the attachments on the structurometer ST-2 8. V.G. Kanarev, Methods of protein and aminoacid analysis of plants (Leningrad, 1973) 9. D.M. Spackman, S. Moor, W.A. Stein, Analytical Chemistry,30(7) (1958) 11 https://doi.org/10.1051/bioconf/20214002001 BIO Web of Conferences 40, 02001 (2021) ILS 2021 10. GOST 28561-90. Fruit and vegetable products. Methods for determination of total solids or moisture (Standartinform, Moscow, 2011) 11. A.I. Ermakova, Methods of biochemical research of plants (Agropromizdat, Leningrad, 1987) 12. A.I. Ermakov, N.P. Yarosh, A.A. Mikhailov, Applied biochemistry and Microbiology, 3(1) (1967) 13. I.I. Vasilenko, V.I. Komarov, Grain quality assessment (Agropromizdat, Moscow, 1987) 14. Collection of recipes of dishes and culinary products for public catering enterprises (Economy, Moscow, 1983) 15. Yu.T. Zhuk, Canning and storage of mushrooms (biochemical bases) (Light and food industry, Moscow, 1982) 16. Yu.T. Zhuk, Research of edible mushrooms in connection with the problem of effective use of natural resources (Moscow, 1975) 17. P.A. Sychev, N.P. Tkachenko, Mushrooms and mushroom farming (Stalker, Donetsk, 2003) 18. N.A. Bisko, I.A. Dudka, Biology and cultivation of edible mushrooms of the genus oyster mushroom (Naukova dumka, Kiev, 1987) 12 12

Subsets and Splits

No community queries yet

The top public SQL queries from the community will appear here once available.